Cotton and Wheat Inter Cropping - Ph.D. Thesis.

Productivity and resource usein cotton and wheat relay intercropping

Lizhen Zhang

Productivity and resource use in cotton and wheat relay intercropping

Promotor: Prof. dr. ir. J.H.J.Spiertz Emeritus-Hoogleraar Gewasecologie, met bijzondere aandacht voor nutriënten- en stofstromen Co-promotor: Dr. ir. W. van der Werf

Universitair Hoofddocent, Leerstoelgroep Gewas- en Onkruidecologie

Promotiecommissie:

Prof. dr. ir. A.H.C. van Bruggen (Wageningen Universiteit) Prof. dr. ir. J. Goudriaan (Wageningen Universiteit) Prof. dr. L.H.W. van der Plas (Wageningen Universiteit) Prof. dr. Fusuo Zhang (China Agricultural University)

Dit onderzoek is uitgevoerd binnen de C.T. de Wit Onderzoekschool: Production Ecology and Resource Conservation

Productivity and resource use in cotton and wheat relay intercropping

Lizhen Zhang

Proefschrift ter verkrijging van de graad van doctor

op gezag van de rector magnificus van Wageningen Universiteit

prof. dr. M.J. Kropff in het openbaar te verdedigen

op woensdag 7 november 2007 des namiddags half twee in de Aula.

Zhang, L. (2007) Productivity and resource use in cotton and wheat relay intercropping Zhang, L. – [S.l.: s.n.]. Ill. PhD thesis Wageningen University. – With ref. – With summaries in English, Dutch and Chinese. ISBN: 978-90-8504-759-9

Abstract Zhang, L., 2007. Productivity and resource use in cotton and wheat relay intercropping. PhD

thesis, Wageningen University, Wageningen, The Netherlands. With summaries in English, Dutch and Chinese, 198 pp.

From the early 1980s onwards, farmers in the Yellow River cotton producing region intercropped cotton and winter wheat; currently on more than 60% of the total cotton acreage. The driving force for intercropping was the need to increase household income by producing a cash crop, while maintaining the production of a major staple food. This study aims at analyzing the productivity and resource use of cotton-wheat relay intercropping systems. Wheat is sown in strips with interspersed bare soil in October and harvested in June of the next year, while cotton is sown in the interspersed space in the wheat crop in April and harvested before the next wheat sowing in October. Crop growth, phenology, productivity, quality, resource use efficiencies and profitability of mono- and intercrops were studied at the plant, field and system levels. The measurements were carried out in field experiments during three consecutive years with monocultures of wheat and cotton and four intercropping designs differing in strip and path width as well as number of rows per strip. The intercrop systems were identified by the number of rows per strip of wheat and cotton, as 3:1, 3:2, 4:2 and 6:2, respectively. All intercropping systems showed an advantage in land productivity compared to growing of monocrops. The fiber quality of cotton was not affected by intercropping. The land equivalence ratio was 1.39 in the 3:1, 3:2 and 4:2 systems, and significantly lower, 1.28, in the 6:2 system. All systems thus provide a substantial land use advantage. Resource (light and nitrogen) use efficiencies of intercropped wheat were similar to the monoculture; however, the resource capture decreased, because part of the land space was assigned to cotton. For intercropped cotton, light use efficiencies were similar to the monoculture; the amount of light intercepted decreased due to a delay in development and growth during the seedling stage and by the extent of canopy closure after the wheat harvest. The relative nitrogen yield total of intercrops was higher than the land equivalence ratio. Nitrogen use efficiency of cotton was decreased. The analysis of the N balance sheet showed that in the intercropping systems N was considerably more prone to losses than in the sole cotton. Conventional N-management in intercrops results in high N-surpluses that pose an environmental risk. Water productivity, both of wheat and cotton, was lower for the intercrops than for monocultures. The lower WP in the intercropping systems compared to the sole crop is a concern for the sustainability of these systems; water productivity needs to be enhanced. A simple mechanistic model for cotton (SUCROS-Cotton) was developed to explore the prospects to optimize intercropping systems. This model simulates cotton development as well for intercrops as for monoculture.

The findings suggest that the productivity and resource use efficiencies of cotton-wheat intercropping can be improved by modifying the conventional management practices and by system optimization. It is concluded that the intercropping systems increase farmers’ income under a wide range of wheat and cotton prices. Keywords: Grain yield; lint yield; phenological delay; light use; nitrogen use; resource use efficiency;

modelling; profitability; water productivity.

Preface The idea to develop a PhD-research project was put forward by Dr. Wopke van der Werf, when he visited the Cotton Research Institute (CRI), Chinese Academy of Agricultural Sciences (CAAS) in the summer of 2000. He was invited by Dr. Jingyuan Xia, the former director of CRI to participate in the programme “Biological control of cotton aphid in wheat and cotton intercropping system in China”. Encouraged and supported by Drs. van der Werf and Xia, I wrote a sandwich-PhD project proposal that was awarded in 2001 by the C.T. de Wit Graduate School of Production Ecology and Resource Conservation (PE&RC) of Wageningen University with a PhD-fellowship. From November 2001 to November 2004, a series of experiments were carried at CRI. The International Agricultural Centre (IAC) provided funding for a 6-month stay, from July to October 2004, in the Group Crop and Weed Ecology (CWE) to work on system analysis and modelling of cotton-wheat intercropping systems. The data analyses and writing of papers was done before and during my final stay, from September 2006 to April 2007, in the Group Crop and Weed Ecology (CWE) at Wageningen. I express my sincere gratitude to the support given by Wageningen University and Research Centre (PE&RC, Group Crop and Weed Ecology (CWE) and IAC) and the Cotton Research Institute of the China Agricultural Academy of Sciences (CRI-CAAS). I am indebted to Prof. Huub Spiertz, my promoter, for his scientific guidance, understanding and support. His interest in the research, insight guidance and rich experience in experimentation and data analysis were extremely helpful for developing the scientific framework of this thesis. His broad and deep thinking and his visit to CRI enabled me to gain more knowledge throughout the programme. His kindness and hospitality gave me the feeling to be at home during my 3-times stay at Wageningen. I am also grateful to his wife, Ms Julienne Spiertz, for serving many times delicious and traditional Dutch food and for her hospitality. I am indebted to Dr. Wopke van der Werf, my co-promotor, for the countless discussions, face-to-face, by email or by Skype, over the past years. His first visit to CRI was of prime importance for the start of my PhD programme and the lay-out of the series of experiments. His professional skills in system analysis and strong back-ground in applied mathematics were very helpful to analyse the experimental data and to develop the simulation model. The inspiring ideas, critical comments, constructive suggestions and word-for-word correction of manuscripts have been very valuable to me. I benefited not only of his scientific skills, but also of his strong ability to use a clear and sound language, of his writing skills and his life philosophy. I also extend my sincere thanks to Ms. Saskia Beverloo for the warmly invitations and hospitality.

I am grateful to Dr. Jingyuan Xia for his irreplaceable help. With out his advice and support, I could not have started this programme. My deep gratitude and heartfelt thanks are due to Kunbo Wang, Zhiyong Hou, Dr. Fuguang Li, the directors of CRI, Xiaoxuan Song, the assistant director of CRI, and Dr. Xuebiao Pan, the director of the meteorological department of China Agricultural University (CAU) for their moral encouragement and hospitality. I am also grateful to Dr. Baoguo Li, the vice-director of the College of Agricultural Resources and Environmental Sciences of CAU, for the fruitful discussions, his comments, encouragement and support. I thank also Prof. Jan Goudriaan (PPS) for many constructive comments and reviewing manuscripts, even after his retirement. His interest in my research has been a constant source of inspiration. My sincere indebtedness is extended to Dr. Lammert Bastiaans and Dr. Xinyou Yin (CWE), both excellent scientists for useful discussions and their critical comments and suggestions. I also wish to express my sincere thanks to Ms. Gon van Laar for all the efforts in getting this thesis printed and for checking the programme code of the model line-by-line. Her editorial skills and valuable advice have been very useful for finalizing the thesis. My profound appreciation goes to the all staff of CWE and PE&RC, especially to Hilde Holleman and Jenny Elwood, the secretaries of CWE, and Dr. Claudius van de Vijver of PE&RC for their hospitality and excellent assistance. I wish to express my earnest thanks and deep appreciation to Siping Zhang for his day-to-day assistance with the tremendous amount of experimental work, which was indispensable for the successful completion of this thesis. I also thank Jiae Hu for her hard work to collect field data. I am also grateful to my home institute for the solid support to my scientific development and my life. Many of my colleagues at the institute helped me in one way or another. My profound appreciation thus goes to Xinhua Zhao, Shoujun Wei, Quanyi Liu, Youlu Yuan, Jinjie Cui, Baosan Zhang and many other people – too many to mention them all individually. I wish to thank my friends in Wageningen, Dule Zhao, Qi Jin, Dong Jiang, Fuyu Ma, Linzhi Li, Huaidong Du, Wen Jiang, Xiulian Sun, Xiaoxiao Zhang, Yuntao Ma, Jianbo Shen, Xiaotang Ju and many others for their help and the enjoyable time. Finally, a special word of thanks goes to my dear wife Qiaoyu Niu and my lovely son Zhenyu Zhang for all their understanding and moral support. My wife Qiaoyu made the biggest contribution to the care for and education of our son and to a well-organized family life. I wish to thank my parents for their love and to my elder sisters and young brothers for their day to day care for our parents. Lizhen Zhang Wageningen, November 2007

Contents

Chapter 1 General introduction 1Chapter 2 Growth, yield and quality of wheat and cotton in relay strip

intercropping systems 11

Chapter 3 Cotton development and temperature dynamics in relay intercropping with wheat

35

Chapter 4 Light interception and radiation use efficiency in relay intercrops of wheat and cotton

55

Chapter 5 Nitrogen economy in relay intercropping systems of wheat and cotton

79

Chapter 6 Development and validation of SUCROS-Cotton: A mechanistic crop growth simulation model for cotton, applied to Chinese cropping conditions

99

Chapter 7 General discussion 121

References

133

Appendix I Listing of light interception model for wheat and cotton in intercropping systems

147

Appendix II Listing of the model SUCROS-Cotton 155 Appendix III List of variables used in the model SUCROS-Cotton 171

Summary 小结 Samenvatting

179183187

List of publications of the author 191 PE&RC PhD Education Certificate 195 Curriculum vitae 197 Funding 198

CHAPTER 1

General introduction

Chapter 1

2

Agriculture in China

China accounts for 22 percent of the world’s population but it has access to only 7 percent of the world’s arable land. Water resources per capita are less than one-fourth of the world’s average. In terms of cultivated land, the occupancy is 0.12 ha per capita, about one-third of the world‘s average (Li and Zuo, 1997). The arable land is not only limited in quantity but also in quality; the cultivated arable land amounts to about 130 Mha (Tianzhi, 2004). Some recent estimates indicate a loss of fertile land of about 40 Mha during the last decades due to industrial development, building new houses and infrastructures. The best arable land is therefore used intensively. Trends in cereal production (rice, wheat and maize) in China during the period from 1961 to 2003 show an enormous increase in production from 100 to about 400 Mtons, while the harvested area stagnated or even decreased. So, the huge increase of cereal production was mainly brought about by an increase of crop yields (Dobermann et al., 2004). This increase was associated with an exponential increase of nitrogen use until the end of the 1990s and a decrease of the agronomic nitrogen use efficiency.

Where heat resources are sufficient, two or more crops are consecutively grown or intercropped in one field. The main types of multiple cropping systems are: inter- or relay-cropping of wheat and corn, double cropping of winter wheat and corn, relay cropping of winter wheat and cotton, and multiple cropping of rice (2 - 3 rice cropping cycles, rice-wheat systems, etc.). In this study, we will only deal with relay intercropping of wheat and cotton.

Cotton production in China

China is the largest cotton producer in the world. Cotton occupies a crucial position in the national economy and the livelihood of many Chinese farmers. China views cotton as a strategic commodity because of its historic importance in clothing its large army, in obtaining foreign exchange, and as a source of state tax revenue (Fang and Babcock, 2003). About 200 million Chinese farmers currently produce cotton. Averaged for the period from 2000 to 2004, the sowing area amounted to 4.8 million ha, 5.2 million tons for lint production and 1087 kg lint ha–1 for yield (National Bureau of Statistic of China (NBC), 2005). Short-fiber, low-yielding varieties (Gossypium arboretum L.) were introduced to south China 2000 years ago and quickly spread throughout China after introduction of upland cotton (Gossypium hirsutum L.). The three major producing regions are: Yellow River Valley, Yangtze River Valley and north-west


3

region (Hsu and Gale, 2001) (Fig. 1), with percentage of 47%, 26% and 21% of the total cotton area, respectively. The major characteristics of those cotton producing regions are listed in Table 1. The average farm size in China is 0.10 to 0.13 ha per person and varies per region. Cotton farming techniques remain very labor intensive compared with U.S. farming practices. Cotton is hand-picked during the harvest period seven to ten times, leaving virtually no fiber unpicked on the plant. If immaturity or rain inhibits picking, the whole plant is cut, brought under shelter in the family living quarters if necessary, where picking is resumed, and the wood is used later for heating. In the provinces of Henan, Shandong, and Hebei, where there is considerable competition among crops for limited land, an increasing proportion of land was reserved year by year for intercropping e.g. cotton with wheat，garlic or onions (Butterworth and Wu, 2004). In the case of intercropping with winter wheat, farmers would have had to set aside land for cotton and other crops during the previous fall. This system has a significant advantage in bio-control of cotton aphid (Xia, 1997).

Fig. 1: Map of major cotton producing regions in China

Cotton Areas in major provinces(% of the total) in 2004Henan 16.7Hebei 7.2Shandong 18.6Jiangsu 7.2Anhui 7.0Hubei 7.2Hunan 3.0Xinjiang 20Others 13.1China (total) 100

Modified after Fang and Babcock (2003)

Chapter 1

4

Table 1: Characteristics of major cotton producing regions in China Regions Characteristics a

Yellow River Yangtze River Northwest

Provinces Shandong, Henan Hebei, Shanxi,Shaanxi

Jiangsu, Anhui, Hubei, Hunan, Jiangxi, Sichuan, Zhejiang

Xinjiang, Gansu

Lint yield (kg ha–1) b 810 947 1,393 Varieties: Maturity classes

Mid, mid-early and early maturing

Mid and mid- early maturing

Mid-early and early maturing

Bt cotton ～95% ～75% None Cropping patterns Intercropping with

winter wheat or mono-cropping

Double cropping with winter wheat or rape seed

Mono crop at high densities

Daily air temperature (°C ) c

19-22 21-24 18-25, plastic films used to protect seedlings

Annual rainfall (mm) 500-800, frequent drought and water shortages

1,000-1,600, frequent flood

below 200, arid conditions

Duration of growing season (days) d

195-220 220-270 South, 185-230 North, 160-190

Annual sunshine duration (hours)

2,200-3,000 1,500-2,500 2,700-3,500

a Only major cotton producing provinces are included in the classification. b Yields are averaged from 1995 to 1999. c Averaged daily air temperatures during cotton growing season from Apr. to Oct. d Threshold values are a minimum temperature of 10 °C. Modified after Fang and Babcock (2003), Economic Research Services, USDA.

Wheat and cotton intercropping

History and region

From the early 1980s onwards, farmers in the Yellow River cotton producing region started to intercrop cotton (Gossypium hirsutum L.) and winter wheat (Triticum aestivum L.) because of the need to increase household income by production of the cash crop cotton, while having to continue the production of wheat as a major staple food. In the three northern/central provinces – Hebei, Henan and Shandong – cotton and wheat intercropping system covered only 116,000 hectares in 1980, but reached


5

more than 1.2 million hectares in 1988 and 1.6 million hectares in 1990 (Zhang and Li, 1997). Since then, the acreage of intercropping has stabilized. The main soil type in this region is a deep clay-loam soil that is very suitable for agricultural production. The annual temperature sum is 4200-5500 degree-days above 0 °C, or 3600-4800 degree-days above 10 °C. The frost-free period amounts to 170-200 days. Annual rainfall varies from 500 to 900 mm, of which two thirds fall in summer. Spring drought is frequent.

Relay in time and strip in space

In the wheat-cotton intercropping system, a relay strip intercropping approach is used to grow both wheat and cotton on the same field in one year. In this system, strips of winter wheat, sown in the fall, are intersown with cotton in the spring. After the wheat harvest in early summer, cotton occupies the whole land. At cotton harvest, in the fall, two crops have been grown in one field, with the seedling phase of cotton and the maturation phase of wheat overlapping in time and space. The cotton harvest is early enough to sow wheat timely, thus closing the annual cycle (Fig. 2).

Several intercropping patterns are used in practice. They are named after the numbers of rows of wheat and cotton that are alternated, e.g. the 3:2 system is an intercropping system consisting of 3 rows of wheat and 2 rows of cotton alternating. Other systems in use include 3:1, 4:2 and 6:2. In addition, some variability in the row distances exists. These systems are characterized by differences in production and competitive relationships among cotton and wheat, but these differences have not been quantitatively documented.

Productivity

Yields of relay intercropped cotton and wheat have not been thoroughly studied. Maximum wheat yields in cotton-wheat intercrops are in the order of 3000 to 3800 kg ha–1 depending on the proportion of wheat in the intercrop. The yield of cotton lint in cotton-wheat intercrop in the Yellow River valley amounts to about 750 kg ha–1, but higher production levels have also been reported. For instance, Zhang and Li (1997) report that in Fugou county, which is located in the province of Henan, the wheat yield in intercrops amounted to 3375-3750 kg ha–1, 75-83% of monocropped wheat, while the cotton yield was almost as high as in the monocrop of cotton: 1350 kg ha–1 lint. This result was achieved under very good management and high external inputs.

Chapter 1

6

a b

c d

e f

Fig. 2: Pictures of wheat-cotton relay strip intercropping systems at different growth stages near Anyang, China (a: the 3:1 system during the intercropping period, May 1, 2002; b: the 3:2 system during the intercropping period, May 15, 2003; c: the 3:2 system at end of intercropping period, June 12, 2003; d: the 3:2 system one day after wheat harvest, June 16, 2003; e: the 6:2 system 10 days before cotton cutout, July 23, 2003; f: the 4:2 system during cotton harvest period, October 12, 2002)


7

Limitation and utilization of resources

China’s water resource per capita is about one-fourth of the world’s average. It varies greatly from south to north. About 80% of the available water is concentrated in southern China along the Yangtze River, whereas the arable land in this region accounts for only 36% of the national total. The northern region has only 20% of the water resources with 64% of the prime agricultural land. In northern China, irrigation combined with industrial water use is responsible for a high rate of lowering the water table – approximately 1 m per annum – and it virtually exhausts the yearly supply of the Yellow River (Cheng et al., 1992; Jin et al., 1999; Liu et al., 2001; Zhen et al., 2005). The lowering of the underground water table has resulted in higher exploitation costs for irrigation and more uncertainties about water availability. Agricultural use of water is therefore constrained.

Application of mineral nitrogen (N) fertilizer in China amounts to 379 kg ha–1 per year on average (Li and Zuo, 1997), which is more than 2.5 times the world average. There are no exact data on how much fertilizer is applied to cotton and wheat intercrops at the national level, but it is presumably as high as the above-mentioned national averages for all crops. As a consequence of the high fertilizer rate, especially of nitrogen, a significant portion of the nutrient cannot be taken up by the plant and is either leached to the underground water or lost to the air (Cheng et al., 1992; Cai and Smit, 1994; Zhen et al., 2006). In addition to environmental cost, farmers spend more money on fertilizer than they would need to do if fertilizer use efficiency could be raised by improved management, e.g. better timing or optimizing application rate (Hou et al., 2007). Improving the resource use efficiency is therefore an objective of both agronomists and farmers.

Improved irrigation and fertilization of intercrops could reduce resource use competition and increase yields (Willey, 1990; Midmore, 1993). Optimized productivity and resource use efficiency could potentially be achieved by improved timing and dosing of irrigation and fertilizer. Such optimization however requires a good understanding of the growth and ecophysiology of the intercrop.

Relay intercropping is a way to produce two crops in one field in the same season when the temperature requirements are too high to allow the second crop to be sown after the first crop is harvested. The second crop is intersown in the first crop, commencing development and early growth before the first crop is harvested. Suboptimal solar radiation and heat accumulation in intercrops are potentially limiting factors for the growth and production of the second crop.

Chapter 1

8

Problem definition

Intercropped cotton and wheat are among the major crops in the North China Plain. Furthermore, monocrops of wheat, cotton, corn and oil seed rape are grown. Resource use and production of the cotton-wheat intercrops in comparison to monocrops have not been well defined, but there are indications that crop production per unit of land is substantially increased, compared to monocrops (Willey, 1979; Reddy and Willey, 1981; Ahmed and Rao, 1982). Relay intercropping does impose some specific constraints. Shading of cotton by the wheat crop affects growth and development of cotton. Competition for light, water and nitrogen is likely to play an important role. These constraints may reduce cotton yield but they could also decrease cotton fiber quality and delay the sowing date of the next wheat crop. No research effort of importance has been devoted to the potential further optimization of wheat and cotton intercropping systems, by way of planting patterns, variety choice, sowing date, nitrogen management and irrigation practices, etc. Agricultural resource use, notably water and nitrogen, has resulted in unsustainability problems. Hence, there is clearly a need for studying productivity and resource use of cotton-wheat intercropping systems.

Research objectives

The main objectives of the research are:

1. to quantify the growth, yield, and quality of wheat and cotton in relay strip intercropping systems;

2. to quantify the phenology of intercropped cotton in relation to temperature and the effect of shading by wheat on the development of cotton;

3. to quantify light interception and use efficiency of wheat and cotton in intercropping and monocropping.

4. to quantify nitrogen uptake and use efficiency of wheat and cotton in the intercropping systems, and to explore the opportunities to improve nitrogen management.

5. to develop an ecophysiological-based simulation model for cotton that can be used to explore options to optimize resource use, e.g. temperature.


9

This study is also of practical significance, due to its potential for indicating opportunities to improve relay intercropping systems of wheat and cotton. It is also an interesting case study how to fine-tune cropping systems and make clever use of resource use in time and space. Such an approach is needed to promote ecologically more diverse cropping systems that are potentially more suited to serve multiple functions.

Thesis outline

The thesis consists of seven chapters:

Chapter 1 provides an introduction to agricultural production, especially cotton production systems, and production regions for cotton in China. The set-up of relay-intercropping (in time and space) of wheat and cotton is introduced. Then, the problems associated with intercropping are defined. The objectives and the research approach are presented.

Chapter 2; lint quality of cotton and the productivity of four intercropping systems and monocultures of wheat and cotton are analysed based on data derived from 3 years field experimentation. The advantage of intercropping on the land use is quantified by analysing land equivalence ratios (LER) and border row effects. The growth delay of intercropped cotton is analysed by fitting expolinear growth equations to periodic harvest data.

Chapter 3; the phenological delay of intercropped cotton is quantified based on the analysis of temperature profiles both in the canopy and in soil layers of the four intercrop systems and the monocrops. The relationship between the delay in development of cotton and the reproductive capacity is studied. The potential of soil covers to improve the thermal environment of the cotton seedlings is studied.

Chapter 4; to quantify the amount of light intercepted, a light interception model suitable for the strip canopy structure is applied. From the accumulated light interception and total dry weight, the light use efficiency of wheat and cotton in intercropping and monoculture is derived. The spatial pattern and diurnal course of photosynthetic active radiation (PAR) in the intercrops are described.

Chapter 5; total nitrogen uptake and nitrogen use efficiency of wheat and cotton in intercropping and monoculture are quantified based on data of two series of field experiments. Furthermore, the nitrogen dynamics of cotton in intercrops and the monocrop are analysed based on physiological characteristics such as N content and

Chapter 1

10

specific leaf nitrogen content. A nitrogen balance sheet is applied to assess potential N-losses and to evaluate the current nitrogen management practices. Options to improve the sustainability of the intercropping systems are explored.

Chapter 6; a mechanistic cotton growth and development simulation model SUCROS-Cotton is developed to analyse temperature effects on phenology and yield. The model is validated for agro-ecological conditions and management practices in China.

Chapter 7 is the general discussion. It links the findings on crop growth, development and productivity with prospects to improve the agronomic performance, economic profit, and efficiencies of natural resources and external inputs. The potential of new practices, e.g. film cover, ridge arrangement, and genetic improvement, are explored and discussed.

CHAPTER 2

Growth, yield and quality of wheat and cotton in relay strip intercropping systems*

* Field Crops Research 103 (2007), 178-188 L. Zhang a,b,c, W. van der Werf b, S. Zhang a, B. Li c and J.H.J. Spiertz b

a Cotton Research Institute, Chinese Academy of Agricultural Sciences, Key

Laboratory for Genetic Improvement of Cotton, Ministry of Agriculture, Anyang, Henan 455004, P.R. China

b Wageningen University, Plant Sciences, Crop and Weed Ecology Group, P.O. Box 430, 6700 AK Wageningen, The Netherlands

c College of Agricultural Resources and Environmental Sciences, Key Laboratory of Plant and Soil Interaction, Ministry of Agriculture, China Agricultural University, Beijing 100094, P.R. China

Chapter 2

12

ABSTRACT

Intercropping of wheat and cotton is practiced at a large scale in northern China, but the productivity of intercrops, compared to monoculture, and the productivity and growth patterns of different alternative intercropping patterns have not been quantitatively documented. In this study, four typical wheat-cotton intercropping patterns were examined as to their growth and productivity in field experiments over three growing seasons in Anyang, Henan province, China. The systems varied in the number of wheat and cotton rows in the alternating strips of either crop, and were labeled accordingly as 3:1, 3:2, 4:2 and 6:2. Dry matter accumulation, yield, land equivalence ratio (LER) and lint quality were determined.

Grain yield of wheat, averaged over three seasons, ranged from 4,600 to 5,200 kg ha–1 in intercropping, corresponding to 70 - 79 percent of the yield in the monoculture (6550 kg ha–1). The 3:1 system gave the highest wheat yield (79% of monoculture), followed by the 6:2 (73%), 3:2 (70%), and 4:2 (70%) systems. Cotton lint yield, averaged over three seasons, ranged from 590 to 740 kg ha–1 in intercropping, corresponding to 54 to 69 percent of the yield in cotton monoculture (1085 kg ha–1). The 3:2 and 4:2 systems gave the highest lint yields (69% and 68% of monoculture, respectively), which was significantly lower than in monoculture but significantly higher than in the 3:1 (58%) and 6:2 (54%) systems. The land equivalent ratio was 1.39 in the 3:1, 3:2 and 4:2 systems, and significantly lower, 1.28, in the 6:2 system. All systems provide a substantial land use advantage.

Cotton growth patterns in monocultures and intercrops were characterized by fitting expolinear growth equations to periodic harvest data. Fitted parameters indicate a growth delay, compared to cotton monoculture, of 11.8 d in the 3:1 system, 6.3 d in the 3:2 system, 6.9 d in the 4:2 system and 5.6 d in the 6:2 system. Estimated growth rate during the linear growth phase was lowest in the 6:2 system (5.9 g m–2 d–1), significantly greater in the 3:1 (7.0 g m–2 d–1), 4:2 (7.7 g m–2 d–1) and 3:2 (8.4 g m–2 d–1) systems, and greatest, but not significantly different from 3:2 and 4:2 systems, in the monoculture (8.9 g m–2 d–1). These results are interpreted in terms of the competitive effect of wheat during the seedling phase of cotton, which is strongest in the 3:1 system, causing a comparatively long growth delay, and the ability of the cotton leaf canopy to intercept radiation after wheat harvest, which is diminished in the 6:2 system due to the large distance between cotton rows, resulting in a comparatively low rate of linear growth.

Effects of intercropping on the quality of cotton were minor and mostly below detection threshold.

Keywords: Crop growth analysis; grain yield, lint yield; land equivalence ratio (LER); fiber quality; expolinear growth equation, competition; growth delay.

Growth, yield and quality of wheat and cotton in relay strip intercropping systems

13

INTRODUCTION

One-third of all cultivated land in China is used for multiple cropping (Zhang and Li, 2003). Cotton based intercropping, such as wheat/cotton, is a major application of multiple cropping, and plays an important role in combining food security and farmer’s income. From 1980 onwards, the acreage of intercropped cotton increased more than four times in Henan, Hebei and Shandong provinces in the Huang-Huai-Hai plain (Yellow River valley), one of major cotton producing regions of China. More than 65% of cotton is currently cultivated as relay strip intercropping with wheat in this region; in Henan province, about 95% of cotton is intercropped, which is around 600,000 ha. The total area of relay-strip intercropped cotton and wheat in China amounts to 1,400,000 ha.

In wheat-cotton intercropping systems (Fig. 1), strips consisting of a few rows of winter wheat, sown in the fall with strips of bare soil interspersed, are intersown with cotton in the spring. The seedling phase of cotton and the reproductive phase of wheat overlap over a period of approximately seven weeks between the sowing of cotton in April and the harvest of wheat in June. Immediately after the harvest of cotton at the end of October, wheat is sown.

Fig. 1: 3:2 wheat and cotton relay strip intercropping system in June, shortly before wheat harvest

Chapter 2

14

Different intercropping patterns are in use. They are named after the numbers of rows of wheat and cotton in the strips of either crop that are alternated. For example, in the 3:2 pattern, the wheat strips consist of three crop rows, and the cotton strips of two crops rows (Fig. 1). Other systems in use include 3:1, 4:2 and 6:2. In addition to the differences in number of rows per strip of the two crops, there are slight differences in row distance. It is to be expected that the differences in crop ratios and row distances among the systems will modify the competitive relationships and the ability of the crops to capture and utilize resources; however, productivity and profitability of alternative systems of intercropping have not been quantitatively documented.

Yield advantages from intercropping are often attributed to complementation between component crops in the mixture, resulting in a better total use of resources when growing together rather than separately. A variety of cotton-based intercropping systems has been introduced worldwide. In India, cotton is intercropped with a variety of other crops, including cowpea, pigeon pea, rice, groundnut, and soybean (Deazevedo et al., 1993; Padhi et al., 1993; Blaise et al., 2005). In eastern Zambia, cotton is intercropped with groundnut (Waterworth, 1994), resulting in a high land equivalent ratio (LER), while in Egypt cotton is intercropped with basil to suppress pests (Schader et al., 2005). In the United States and in China relay intercropping of cotton with winter and spring crops is used to conserve and enhance natural enemies of cotton aphid (Parajulee et al., 1997; Xia, 1997).

Interspecific competition will occur when two crops are grown together (van der Meer, 1989). Such competition usually decreases survival, growth or reproduction of at least one species (Crawley, 1997). In a strip intercropping system, a border row effect has often been found. For instance, in maize/soybean strip intercropping, West and Griffith (1992) observed 26% increase in corn and 27% reduction in soybean border rows in 8-row alternating strips in Indiana, and Ghaffarzadeh et al. (1994) found that strip intercropping had 20 to 24% higher corn yields and 10 to 15% lower soybean yields in border rows of the two crops in Iowa, USA. The higher yield of border rows of the tallest species (corn) suggests increased resource capture which might have gone at the expense of the neighbouring rows of the smaller species (soybean). In corn/soybean and grain sorghum/soybean strip intercropping systems, Lesoing and Francis (1999) reported that corn and grain sorghum border-row yields next to soybean were increased significantly compared with inside rows. Increased yields in border rows of corn were attributed to increases in seed number as well as seed weight. In China, Li et al. (2001b) and Zhang and Li (2003) found that yields in border rows of intercropped wheat were significantly higher than those in inner rows, both in wheat/maize and wheat/soybean intercropping systems.


15

Although wheat-cotton relay strip intercropping systems are widely practiced in China, there is only scant information on crop and land productivity of such inter-cropping systems. In this chapter, we want to determine (i) whether there is yield advantage in intercropping; (ii) whether there is a difference in intercropping advantage between different intercropping systems; (iii) whether intercropping affects cotton quality. Furthermore, we study dry matter accumulation as a dynamic process to determine (iv) differences among systems in the growth process through time. The work described in this chapter has primarily the objective to document, describe and characterize; i.e. it is hypothesis generating rather than hypothesis testing in nature. Nevertheless, the research objectives can be formulated as null hypotheses that this research intends to refute: (i) intercropping does not provide a yield advantage; (ii) different intercropping systems have the same cotton and wheat yields and land equivalent ratio; (iii) intercropping has no effect on quality; and (iv) there are no differences in the time course of dry matter accumulation between systems.

MATERIALS AND METHODS

Experimental years and site

Field experiments were conducted in 2001/02, 2002/03 and 2003/04 at the Cotton Research Institute of Chinese Academy of Agricultural Sciences (CRI, CAAS), Anyang city, Henan province, China at 36°07´ N and 116°22´ E. The soil of the experimental field is a sandy loam, with a pH of 8.0, a bulk density of 1.36 g cm–3, an organic matter content of 13.2 g kg–1, a total N content of 1.02 g kg–1, a total P content of 0.52 g kg–1 and a total K content of 17.3 g kg–1. Weather data were collected on site (Table 1).

Lay-out of experiments

Field experiments comprised six treatments including four different intercropping patterns and monocultures of wheat (Triticum aestivum L.) and cotton (Gossypium hirsutum L.). Distance between rows was 20 cm in wheat monoculture and 80 cm in cotton. The same distances between crop rows in wheat and cotton strips were used in intercrops, while the distance between bordering wheat and cotton rows was chosen according to farmers’ practice. Width of the wheat strip, as measured between the outer wheat rows was 100 cm in the 6:2 system, 60 cm in the 4:2 system, and 40 cm in the 3:2 and 3:1 systems (Fig. 2). The interspersed space for sowing cotton was 100 cm

Chapter 2

16

in the 6:2 system, 90 cm in the 4:2 system, 80 cm in the 3:2 system, and 60 cm in the 3:1 system. Total width of one adjacent wheat and cotton strip (called the “minimum sequence”) was 200 cm in the 6:2 system, 150 cm in the 4:2 system, 120 cm in the 3:2 system and 100 cm in the 3:1 system (Fig. 2). The 3:1 and 3:2 patterns are common in farmer’s practice, the 4:2 system less common, and the 6:2 pattern the least common.

Table 1: Weather data of Anyang, Henan Province, China, in 2002, 2003 and 2004

Month Climate

factor

Year

1 2 3 4 5 6 7 8 9 10 11 12

2002 7.7 12.5 16.2 20.1 25.3 31.8 32.1 30.9 27.2 20.0 11.4 1.7

2003 2.6 7.7 12.3 19.4 26.0 30.9 29.4 28.2 25.4 20.5 10.2 5.7

Tmax

(°C)

2004 5.2 12.3 15.1 21.6 25.9 30.3 30.8 28.6 27.6 18.9 13.3 4.0

2002 –3.5 –0.4 4.8 9.0 13.4 19.4 22.4 21.1 14.4 8.2 –0.1 –3.9

2003 –6.1 –0.7 2.4 8.7 15.1 19.2 21.7 20.2 16.5 8.9 2.8 –3.4

Tmin

(°C)

2004 –5.8 –0.8 4.0 9.3 13.7 18.7 21.9 20.1 15.6 7.7 1.9 –4.1

2002 0 0 8 18 62 54 49 64 36 8 0 19

2003 5 11 23 26 16 57 75 84 81 136 15 10

Rainfalla

(mm)

2004 0 9 5 15 62 73 149 126 29 8 33 8

2002 3.8 4.1 5.4 4.7 5.6 4.0 4.5 5.5 6.0 4.7 3.0 1.1

2003 3.8 3.7 4.0 5.5 6.2 6.7 4.5 4.0 4.6 5.3 2.3 4.5

Sunshine

hours

(h d–1) 2004 4.2 6.7 5.5 7.5 8.4 5.8 5.7 4.6 5.8 4.4 3.9 2.2

a Rainfall amounts are monthly totals; other values represent monthly averages of daily values.

In this chapter, we use two measures to express plant density of either crop in the intercrop. One is the row length density (m m–2). This is the total row length of one crop species, per m2 of intercrop area, averaged over the whole field. The other is homogenized row distance (Table 2). This is the average distance between rows of one of the component crops in the intercrop, when disregarding the other species. Row length density of inner and border rows of wheat, di and db, respectively, were calculated as total row length per unit intercrop area. All of these measures are thus


17

based on the total intercropped land area, including area planted to the other crop. This approach is especially suited in a relay intercropping system because the crops are crowding for the same space only during a short period of time. An advantage of this approach is that it takes away the need to arbitrarily dividing the available space between two crops.

20 2030 30 30

Sole wheat

3W:1C

N

W WWW

W WW CC W

20 20 20

80

C C

Sole cotton

20

100 cm

W

20 2030 30 30

Sole wheat

3W:1C

N

W WWW

W WW CC W

20 20 20

80

C C

Sole cotton

20

100 cm

W

3W:2C

CW WW CC W

20 20 2020 2040

C CC WWWWW

20 20 2520 254025

150 cm

C CC WWW WW

20 20 2020 3040302030

W W

200 cm

4W:2C

6W:2C

120 cm3W:2C

CW WW CC W

20 20 2020 2040

C CC WWWWW

20 20 2520 254025

150 cm

C CC WWW WW

20 20 2020 3040302030

W W

200 cm

4W:2C

6W:2C

120 cm

Fig. 2: Layout of wheat-cotton intercropping systems and monocultures

Chapter 2

18

The six treatments were arranged in four randomized blocks with a plot size of 180 m2. Wheat was sown on 4 November 2001, 2 November 2002 and 3 November 2003. Cotton was sown on 26 April 2002, 25 April 2003 and 25 April 2004. The wheat cultivar ‘Zhongyu 5’ was used, which was especially developed for wheat-cotton intercropping systems by CRI. Cotton cultivars were Chinese bred middle maturity upland Bt cotton ‘Shiyuan 321’ in 2002 and ‘CCRI45’ in 2003 and 2004.

Table 2: Layout of cropping systems

Row length density (m m–2)a

- Wheat - - Cotton -

Homogenized row width (cm)bCropping pattern

Inner Border Total Wheat Cotton

Sole wheat 5 0 5 - 20 -

Sole cotton - - - 1.25 - 80

3:1 (3W:1C) 1 2 3 1 33.3 100

3:2 (3W:2C) 0.83 1.67 2.5 1.67 40 60

4:2 (4W:2C) 1.33 1.33 2.67 1.33 37.5 75

6:2 (6W:2C) 2 1 3 1 33.3 100 a Row length density is total row length of a component crop per unit intercrop area (m m–2,

or rows per meter cross-row). b Homogenized row width is calculated as total width of the minimum sequence (cf. Fig. 2)

divided by the number of rows of one of the crop species.

Flood irrigation was applied in 2002 and drip irrigation was used in 2003 and 2004; the quantity of irrigation amounted to 200-250 mm water per season. Total nitrogen use amounted to 300-400 kg ha–1 applied as organic material (dung and cotton cake) and compound fertilizers each year. The amounts of irrigation water and N-fertilizer were chosen to meet the requirements of high-yielding wheat and cotton crops.

Measurements

To assess the aboveground biomass, 1 m of each row of each plot was sampled once per two weeks. First, the number of plants was counted, then a sample was selected for detailed analyses; the wheat sample consisted of twenty plants while the cotton sample


19

consisted of ten seedlings during early growth or three plants later on. To determine dry matter distribution, wheat plants were subdivided in leaf, stem and spike, while cotton plants were subdivided in leaf, stem, squares, flowers, and bolls. In the cotton samples, the number of leaves, branches, nodes and fruits per plant were counted, and leaf area and plant height were measured. After measurements, the samples were oven-dried at 65 °C to constant weight to determine dry matter (DM).

Final grain and lint yields were determined by harvesting all plants in a sampling area of 5 m row length by 2 m width in the 3:1 system, 5 m length by 2.4 m width in the 3:2 system, 3.5 m row length by 3 m width in the 4:2 system, 5 m row length by 2 m width in the 6:2 system, 5 m row length by 2 m width (10 rows) in sole wheat and 5 m row length by 2.4 m width (3 rows) in sole cotton. In 2002, wheat samples were taken separately in border rows and inner rows to determine presence of a border row effect. Wheat grain was sun-dried and weighed after threshing. Reported grain yields (dry matter) are based on accounting for 12% water content in the sun-dried grain. Cotton was picked by hand at about 10 day intervals after opening of the first bolls. The yields before frost and after frost were determined to calculate percentage of lint before frost (LBF - %), which is a yield component indicating timeliness of yield formation. Lint cannot be harvested for fiber after frost. Lint was separated from seed using a small ginner. Ginning out turn (GOT - %) is calculated as the ratio of lint weight after ginning to lint plus seed weight before ginning. A low ginning out turn means that there is relatively low yield of fiber per unit seed cotton (lint + seed).

A sub-sample of 50 open bolls, harvested before the first frost from each plot, was taken to measure fiber quality. Measurements were made at the Seed Quality Test Center of Ministry of Agriculture in Anyang, using a high volume instrument (HVI-900), according to the internationally accepted ICC standard (Anonymous, 2001). Quality depends on the length, strength, fineness, and uniformity of the fibers (Saville, 2004). Fiber length is quantified by “2.5% fiber length”, i.e. the average length of the 2.5% longest fibers, as measured in a high volume instrument. Fiber strength is measured by measuring the force at which a standard fiber sample will break. It is measured in cN (centiNewton) per tex, where “tex” is a standard quantity of fiber (Saville, 2004). Fineness is characterized by the parameter “micronaire”, which is a complex outcome of fiber fineness, strength and maturity (Saville, 2004). Uniformity is the ratio between the mean length of all fibers in a sample and the mean length of the longest 50%. Large values for strength and length are indicators for good quality. Good quality cotton is characterized by micronaire values between 3.8 and 4.9, with larger micronaire indicating lower quality. Uniformity is considered high if the parameter is between 83 and 85% (Anonymous, 2004).

Chapter 2

20

Data analysis

Treatment effects on yield, LER, and quality traits, were calculated with ANOVA, in SPSS 11.0, using cropping system as fixed effect and block, year and the interactions between block and year and between cropping system and year as random effects. Least significant differences (LSD) were used to separate treatment means at P<0.05.

To assess intercropping yield performance compared to monoculture, the land equivalent ratio (LER) was used (Willey, 1985).

sc

ic

sw

iw

YY

YY

,

,

.

,LER += (1)

where Yw,i, Yw,s, Yc,i and Yc,s are grain yields of intercropped and sole wheat, and lint yields of intercropped and sole cotton, respectively. LER is similar in meaning to relative yield total, i.e. the sum of the two component crops in the intercrop, each scaled by the monocrop yield.

To quantify the relationship between yield and row length density of component crops, a power relationship was used:

bY ax= (2)

where Y is yield (g m–2), x is row length density (m m–2) and a, b are fitted parameters. The parameter b determines the curvature of the response, while the parameter a determines the scale along the y-axis.

To model the growth pattern of intercropped cotton, expolinear equations (Goudriaan and Monteith, 1990) were fitted using non-linear least squares regression:

( )ln 1 expmt m b

m

cW r t tr

⎡ ⎤= + −⎡ ⎤⎣ ⎦⎣ ⎦ (3)

where Wt is dry mass (g m–2), t (d) the day after sowing (DAS), cm (g m–2 d–1) the maximum absolute growth rate, rm (d–1) the initial relative growth rate and tb (d) the time at which the growth rate is cm/2. The time course of the growth rate is logistic with growth parameter rm and final value cm. The parameter rm measures the initial exponential rate of growth of the cotton plants, tb measures the time at which the exponential growth goes over into linear growth, and cm measures the rate of linear growth, when the crop has attained maximum soil cover.

Parameters were estimated per treatment and per year following a two-step procedure. In step 1, the parameter rm and tb were estimated using logarithmic transformation of the observations and of Eq. 3. The logarithmic transformation results in greater


21

emphasis on the early observations, resulting in a more accurate estimation of the parameters shaping the early growth, rm and tb. In step 2, rm and tb were fixed at their thus estimated values and cm was estimated using untransformed observations and the untransformed Eq. 3. The use of untransformed data in the second step results in a better estimate of the linear growth rate cm. R2 and root mean square error obtained in the second step are reported. An ANOVA with cropping system and year as main effects was then conducted for each parameter, to determine whether there were significant differences between treatments and between years in the estimated parameter values. Only main effects were used in the ANOVA.

RESULTS

Yield performance of crops

Wheat

Wheat yields were significantly different between intercrops and monoculture (P<0.01; Table 3). Moreover, among the intercropping systems, the 3:1 system gave significantly higher wheat yields, averaged over three years, than the 3:2, 4:2 and 6:2 systems (P<0.05; Table 3). Differences in grain yield between years were significant (P<0.01) as well as the interaction between intercropping systems and years (P<0.01). Among intercropping treatments, no interaction between year and system was found. The interaction was thus wholly due to a comparatively large year effect in wheat monoculture, as a result of the high yield, in relation to the full land cover with the wheat crop. Averaged over three seasons, the grain yield in intercrops ranged from 70 to 79 percent of the yield in the monoculture (6550 kg ha–1). The 3:1 system gave the highest wheat yield (79% of monoculture), followed by the 6:2 (73%), 3:2 (70%), and 4:2 (70%) systems.

Cotton

Cotton lint yields were significantly lower in intercrops than in monoculture (P<0.01; Table 3). The yields of the monoculture ranged from 933 kg ha–1 in 2003 to 1170 kg ha–1 in 2004, with a significant year effect (P<0.05). Cotton lint yield, averaged over three seasons, ranged from 586 to 744 kg ha–1 in intercropping, corresponding to 54 to 69 percent of the average yield in cotton monoculture (1085 kg ha–1). Significant differences were found between intercropping systems. The 3:2 and 4:2 systems gave the highest lint yields (69% and 68% of monoculture, respectively), which was

Chapter 2

22

Table 3: Observed yields, density and LER of the intercrops and the monocultures in 2002, 2003 and 2004

Homogenized density Observed yield1 Year Cropping pattern

Wheat Cotton Wheat Cotton

(ear m–2) (plant m–2) (g m–2) (g m–2)

LER

2002 3:1 -2 4.2 d§ 552 b 60.4 b 1.25 a 3:2 - 6.8 a 500 c 66.0 b 1.23 a 4:2 - 6.0 b 476 c 77.4 b 1.30 a 6:2 - 4.3 d 514 b 60.9 b 1.20 a Sole cotton - 5.1 c - 115.2 a - Sole wheat - - 761 a - - SE - 0.1 15 6.1 0.04

2003 3:1 509 b 5.2 c 416 b 57.4 b 1.46 a 3:2 378 d 7.6 a 362 c 67.0 b 1.41 a 4:2 452 bc 6.4 b 392 bc 58.1 b 1.40 a 6:2 431 cd 4.4 d 395 bc 49.2 b 1.30 a Sole cotton - 6.7 b - 93.3 a - Sole wheat 712 a - 521 a - - SE 23 0.2 11 7.5 0.07

2004 3:1 540 b 5.3 c 585 b 69.5 bc 1.46 ab

3:2 472 b 8.1 a 513 c 90.2 b 1.53 a 4:2 538 b 6.6 b 502 c 87.4 bc 1.48 ab 6:2 511 b 5.2 c 516 c 65.6 c 1.33 b Sole cotton - 6.6 b 117.0 a - Sole wheat 738 a - 683 a - - SE 24 0.2 13 7.4 0.05

Mean 3:1 524 b 4.9 c 517 b 62.4 c 1.39 a 3:2 425 d 7.5 a 459 c 74.4 b 1.39 a 4:2 495 bc 6.3 b 457 c 74.3 b 1.39 a 6:2 471 cd 4.7 c 475 c 58.6 c 1.28 b Sole cotton - 6.1 b - 108.5 a - Sole wheat 725 a - 655 a - - SE 18.2 0.09 7.9 4.1 0.04

1 Wheat yield includes 12% water. Cotton yield is lint after ginning. 2 No densities were measured in wheat in 2002. § Same lettering in the same column subdivision indicates no significant difference in LSD test

at P = 0.05.


23

significantly lower than in monoculture but significantly higher than in the 3:1 (58%) and 6:2 (54%) systems. There was no significant difference in cotton yield between the 3:2 and 4:2 system and between the 3:1 and 6:2 system.

Land equivalent ratio

The LER of the intercrops varied from 1.20 to 1.25 in 2002, from 1.30 to 1.46 in 2003 and from 1.33 to 1.53 in 2004. The LER of the 6:2 system was the lowest in all years; on average 1.28 (Table 3). LER of intercrops differed significantly between the 6:2 and other systems (P<0.03). The 3:1, 3:2 and 4:2 systems had a similar LER (1.39) over the three years. It is concluded that wheat and cotton intercropping systems do have a substantial yield advantage in dry mass compared to single crop systems, with the arrangements 3:1, 3:2 and 4:2 providing a significantly higher LER than the 6:2 system. The lower LER in the 6:2 system is in part due to a low relative yield of cotton, which might be due to incomplete radiation interception (see below).

Cotton quality

Quality traits were affected very little by intercropping. No significant differences were detected in staple length, fiber strength and uniformity (Table 4). Slight but significant differences in micronaire point to an improvement of cotton quality in the 3:1, 3:2 and 4:2 systems, as compared to the monoculture.

Table 4: Cotton quality parameters and yield components Length1 Strength2 Micronaire Uniformity GOT3 LBF4 Cropping

system mm cN/tex - % % %

3:1 28.5 a§ 27.1 a 3.8 a 83.7 a 38.4 b 86.5 b

3:2 28.5 a 27.0 a 3.7 a 83.6 a 38.3 b 89.6 ab 4:2 28.4 a 26.6 a 3.9 ab 83.5 a 38.7 ab 91.4 a 6:2 28.6 a 27.3 a 4.0 ab 83.6 a 38.9 ab 90.4 ab Monoculture 29.0 a 27.4 a 4.2 b 84.0 a 39.4 a 90.5 ab SE 0.25 0.56 0.13 0.26 0.23 1.50 1 Length is 2.5% staple length (mm). 2 Strength is fiber strength of ICC standard (cN/tex). 3 GOT is ginning out turn (%). 4 LBF is percentage of lint harvested before frost over total (%). § Same lettering in the same column indicates no significant difference in LSD test at P = 0.05.

Chapter 2

24

Cotton yield components – lint before frost and ginning out-turn

No significant differences in the percentage of lint before frost were found between monoculture and intercrop treatments (Table 4). LBF was significantly lower in the 3:1 systems than in the 4:2 system, pointing to a delayed development in 3:1 (see below). Ginning out turn was slightly but significantly lower in the 3:1 and 3:2 systems than in monoculture (Table 4).

The relationship between yield and cropping pattern

Effect of row length density

First, the existence of a relationship between yield and row length density was tested by determining whether the slope in a linear model of yield versus row length density was different from 0. The resulting slope was 80.8 ± 6.3 g m–1 for wheat in intercrops or monoculture (t56 = 12.7; P<0.001) and 22.5 ± 7.3 g m–1 (t44 = 3.08; P=0.004) for cotton intercrops, indicating a significant response in both cases. Next, parameters of Eq. 2 were determined in non linear least squares regression. Data of all years were combined in the analysis, and models with common parameters over all years and year-specific parameters were fitted.

Wheat yields in all treatments were fitted satisfactorily with Eq. 2, using a common curvature parameter for all years (b = 0.54 ± 0.04) and year-specific scale parameters (a = 300 ± 15 g m–2 in 2002; a = 222 ± 11 g m–2 in 2003; a = 297 ± 15 g m–2 in 2004). Thus, wheat yield is closely related to row length density and is characterized by a common relationship for intercrops and monoculture (Fig. 3a). The scale parameter a is different between years, reflecting differences in wheat yield between years (cf. Table 3).

Cotton yields in intercrops were fitted adequately with Eq. 2, using a common curvature parameter for all years (b = 0.44 ± 0.31) and year-specific scale parameters (a = 60.1 ± 3.7 g m–2 in 2002; a = 52.9 ± 3.6 g m–2 in 2003; a = 71.5 ± 3.9 g m–2 in 2004). The R2 of the overall model was 0.41, and the root mean square error 13.9 g m–2. Monoculture yield did not fit the relationship between row length density and yield as found in intercrops (Fig. 3d-f). Significance of the difference was tested in a z-test comparing observed monocrop yield to yield predicted from the regression equation (Eq. 2) as parameterized with the intercrop data. The standard error of the difference was calculated as the square root of the sum of the squared standard error of the mean observed yield and the MSE of the regression. The difference was significant in all years (2002: z = 3.04; P < 0.001; 2003: z = 2.04; P = 0.02; 2004: z = 0.007). The


25

results show that cotton yield in intercrops has a shallow response to homogenized row length density, while monocrops have markedly higher yields than intercrops with similar density of rows (Fig. 3d-f).

Effect of border rows

Wheat grain yield was significantly (61%) higher in border rows than in inner rows (P<0.01) (Fig. 4). By separating row length density into border row length density (db) and inner row length density (di), the relationship between grain yield (Y, g m–2), and db and di (m m–2) was expressed as:

Y = 202.8 db + 131.8 di (R2 = 0.997)

here, 202.8 is the yield per meter of border row (g m–1) and 131.8 is the yield per meter of inner row (g m–1). Grain yields per meter row length in the monoculture were not significantly different from those of inner rows in intercrops.

0

20

40

60

80

100

120

140

0.0 0.5 1.0 1.5 2.0Row length density (m m-2)

Cot

ton

lint y

ield

(g m

-2)



0

150

300

450

600

750

900

0 1 2 3 4 5 6

Whe

at g

rain

yie

ld (g

m-2

)

0 1 2 3 4 5 6 0 1 2 3 4 5 6

2002 2003 2004

2002 2003 2004

a b c

d e f

Fig. 3: Yield response to row length density in wheat (a, b and c) and cotton (d, e and f) in 2002 (a, d), 2003 (b, e) and 2004 (c, f). Open symbols are intercrop yields and filled symbols monoculture yields

Chapter 2

26

0

50

100

150

200

250

300

0 1 2 3 4 5 6 7Row number

Gra

in y

ield

(g m

-1)

3:1 3:2 4:2 6:2 Monoculture

Fig. 4: Grain yield per meter row at different row positions in four different intercropping systems and wheat monoculture in 2002

Crop growth dynamics

Wheat

Dry matter accumulation in wheat proceeded in an almost linear fashion from mid-March till the end of May (Fig. 5). No further dry matter was accumulated during the last two weeks before harvest. Sole cotton showed a higher rate of dry matter accumulation (expressed per m2 total area, including bare space in the intercrops) than intercrops. Trends in 2003 and 2004 (not shown) were similar as in 2002.

Cotton

There were major differences in the size of cotton plants between intercrop and monoculture (P<0.01) at the time of wheat harvest. Individual plant dry weight at the time of wheat harvest was 1.39 ± 0.20 gram in monoculture, 0.28 ± 0.03 g in the 3:1 system, 0.37 ± 0.06 g in the 3:2 system, 0.40 ± 0.07 g in the 4:2 system, and 0.36 ± 0.04 g in the 6:2 system. Individual plants weight was not significantly different between the 3:2, 4:2 and 6: 2 systems, but plant weight was significantly lower in the 3:1 system (P<0.05) and very much and significantly higher (P<0.001) in monoculture. The growth of cotton seedlings in intercrops was thus severely suppressed; indicating competitive effects, especially in the 3:1 system.


27

Dry matter accumulation in cotton attained a constant rate approximately 20-30 days after the harvest of wheat (Fig. 5). The time course of dry matter accumulation in intercrops was somewhat delayed in intercrops, compared to monoculture, especially in the 3:1 system. Crop growth patterns in 2003 and 2004 were similar to those shown for 2002.

0

500

1000

1500

2000

2500

70 90 110 130 150 170 190 210 230 250 270Julian day

Abov

e-gr

ound

dry

mas

s (g

m-2

)

3:1 3:2 4:2 6:2 monoculture

wheat

cotton

ScHw

Fig. 5: Growth of aboveground dry mass in wheat and cotton in monoculture and intercrops in 2002. Arrows indicate sowing time of cotton (Sc) and harvest time of wheat (Hw)

Expolinear growth curves of cotton

Expolinear growth equations (Eq. 3; Fig. 6) were fitted to data from periodic harvests in three seasons to summarize the growth of intercropped and monoculture cotton in three parameters: a parameter rm characterizing the initial relative rate of growth during the exponential growth phase of the seedlings, a parameter cm characterizing the linear growth rate of a mature canopy after maximum light interception is attained, and a time parameter tb characterizing the time of transition from the exponential to the linear growth phase. R2 of fitted equations ranged from 0.914 to 0.999 indicating that the expolinear curve was suitable for description of the dry matter accumulation (total as well as fruits) in intercropped cotton as well as monoculture.

Chapter 2

28

Fig. 6: Expolinear growth curves fitted to data of aboveground dry weight in intercropped cotton and monoculture in 2002, 2003 and 2004

Fitted values of rm for total dry matter showed a significant difference between the 6:2 system and the monoculture, but not between any other treatments (Table 5). The time of transition from exponential to linear growth was earliest in the monoculture (66.7 ±

0100200300400500600700800900

40 60 80 100 120 140 160

DAS

Abo

ve-g

roun

d dr

y m

ass

(g m

-2)

0100200300400500600700800900

Abo

ve-g

roun

d dr

y m

ass

(g m

-2)

40 60 80 100 120 140 160

DAS

0100200300400500600700800900

Abo

ve-g

roun

d dr

y m

ass

(g m

-2)

200220032004

a 3:1 b 3:2

c 4:2 d 6:2

e Sole cotton f

3:16:2

4:23:2Sole


29

1.2 DAS), significantly later in the 3:2, 4:2 and 6:2 systems (respectively 73.0, 73.6 and 72.3 ± 1.2 DAS) and last (and significantly different from all other treatments) in the 3:1 system (tb = 78.5 ± 1.2 DAS). The parameter values point to a growth delay of 5.6 to 6.9 days in systems with wide cotton strips (3:2, 4:2 and 6:2) and a longer delay of 11.8 d. in the system with the narrower cotton strips (3:1). The smallest delay was found in the 6:2 system, i.e. the system with the widest space between the wheat strips. The largest delay was found in the 3:1 system, the system with the narrowest space for cotton and the most severe shading by wheat. The largest estimate for linear growth rate was obtained for the monoculture (8.9 ± 0.51 g m–2 d–1), but estimates for the 3:2 and 4:2 systems were not significantly smaller: 8.4 and 7.7 ± g m–2 d–1, respectively. Linear growth rate in the 3:1 system was 7.0 ± 0.51 g m–2 d–1, which was significantly smaller than in the monoculture, but not significantly smaller than in the 3:2 and 4:2 systems. The 6:2 system had the smallest linear growth rate, significantly smaller than all other systems, except the 3:1.

Table 5: Parameters of expolinear growth equation for growth of above-ground dry matter and fruit dry mass in intercropped cotton and cotton monoculture

rm tb cm Cropping pattern d–1 DAS g m–2 d–1

Above-ground dry mass 3:1 0.124 ab§ 78.5 a 7.0 bcd 3:2 0.129 ab 73.0 b 8.4 abc 4:2 0.130 ab 73.6 b 7.7 abc 6:2 0.138 a 72.3 b 5.9 d Sole cotton 0.124 b 66.7 c 8.9 a SE 0.004 1.2 0.51

Fruit dry mass 3:1 0.124 b§ 113.1 a 8.1 a 3:2 0.176 ab 97.2 bc 5.7 b 4:2 0.149 b 101.6 b 6.2 b 6:2 0.221 a 92.6 c 3.5 c Sole cotton 0.164 ab 93.9 bc 6.3 b SE 0.020 2.4 0.51

§ Same lettering in the same column subdivision means no significant difference according in LSD test at P = 0.05.

Chapter 2

30

Similar but more variable results were obtained for fruit dry mass (Table 5). The initial exponential growth rate was largest in the 6:2 system (0.221 ± 0.02 d–1), significantly different only from the 3:1 system, which had the lowest exponential growth rate (0.124 ± 0.02 d–1). Trends in the time parameter tb were similar to those found for total biomass, but more outspoken. The latest time of transition to linear growth was again found in the 3:1 system (td = 113.1 ± 2.4 DAS), compared to 93.9 ±2.4 DAS in monoculture, i.e. a time delay of 19.8 d. (P<0.001). The 3:2, 4:2 and 6:2 systems did not show a significant delay in comparison to the monoculture, but they all had a significantly earlier transition to linear growth than the 3:1 system. After the late transition to linear growth, the 3:1 system showed the largest linear rate of all systems, 8.1 ± 0.51 g m–2 d–1, significantly greater than in all other treatments, including the monoculture. The 3:2 and 4:2 systems had rates of linear growth that were not different from monoculture, and the rate of linear growth of fruit biomass was lowest, 3.5 ± 0.51 g m–2 d–1, and significantly different from all other systems, in the 6:2 system.

DISCUSSION AND CONCLUSIONS

The results of the study allow the following evaluation of null hypotheses as formulated in the introduction: (i) relay intercropping of cotton and wheat in a strip design provides substantial yield advantage compared to monocultures – the first hypothesis was thus rejected; (ii) two of the four tested intercropping systems were not significantly different in any trait, viz. the 3:2 and 4:2 systems, but the other systems showed markedly different traits; the 3:1 system had significantly greater wheat yield and significantly lower cotton yield than the other intercropping systems, but a similar LER as the 3:2 and 4:2 systems, while the 6:2 system had a similar wheat yield as the 3:2 and 4:2 systems (slightly but not significantly higher), and a markedly and significantly lower cotton yield, resulting in the lowest LER of all tested systems – the second null hypothesis was thus also rejected; (iii) results of the study did not show significant deleterious effects of intercropping on cotton lint quality; on the contrary, a slight but significant increase in micronaire (related to fiber fineness) was found in three systems (3:1, 3:2 and 4:2), compared to monoculture – the third null hypotheses was thus only partially rejected, but contrary to expectation, quality improvement rather than quality loss was observed; (iv) there were very substantial and significant differences in the growth pattern of different intercropping systems. The main differences are a long growth delay in the 3:1 system, and a low rate of linear growth in the 6:2 system – thus the fourth null hypothesis was also rejected.


31

The four tested systems differ in important ways in how production is divided between the component crops. Three systems, 3:2, 4:2 and 6:2, show similar production of wheat, amounting to 70, 70 and 73% of wheat production in monoculture, respectively. The 3:1 system has a greater wheat production, amounting to 79% of monoculture. The differences in wheat production are closely related to the number of wheat rows per m cross row (Fig. 3), and the interception of light by these rows (chapter 4). Outer rows of the wheat strips can capture extra resources (Fig. 4) and thus mitigate some of the yield reduction resulting from incomplete land coverage in intercropping. Production of cotton is highest in the 3:2 and 4:2 systems, and amounts to respectively 68 and 69% of monoculture. Cotton production is substantially lower in the 3:1 and 6:2 systems, respectively 58 and 54% of monoculture. The causes for the lower productively in these two systems may be different and warrant further investigation. The parameterization of expolinear growth curves suggests that the lower yield in the 3:1 system is mostly due to a severe growth reduction in the early growth phase of the crop, as a result of which the transition to linear growth is delayed, as evidenced by large values of tb (Table 5). In the case of the 6:2 system, it seems likely that the cotton plants are unable to capture all incident radiation later in the season, due to the large distance between the rows (Fig. 2), resulting in radiation loss on the soil (Chapter 4) and a low linear rate of growth (Table 5).

This study is to our knowledge the first one to apply the expolinear growth equation for analysing growth patterns in intercrops. The expolinear model is a suitable device for summarizing crop growth processes and extracting key parameters from data sets, enabling an incisive analysis of intercropping effects in different growth phases.

The similarity of LERs in the 3:1, 3:2 and 4:2 systems, in combination with the finding that an increase in wheat production in 3:1 goes at the expense of cotton yield, suggests that these systems utilize all resources and that it will be difficult to obtain even higher land use ratios. Evidence from this study suggests that the 6:2 system does not fully capture all available resources. This may be related to the large width of the wheat strips, resulting in a wide gap that cotton plants need to bridge before they can form a closed canopy intercepting all incoming radiation. Light interception and utilization in different relay strip intercropping systems of cotton and wheat will be further analysed elsewhere.

Under the climatic conditions of Northern China, intercropping systems give a substantial advantage in land productivity over rotations of sole cotton and sole wheat. Further south, warmer temperatures allow sowing of cotton after wheat harvest, however in Northern China, the heat sum does not allow cotton to mature if the cotton is not sown into the maturing wheat. Therefore, in this region, intercropping is the only

Chapter 2

32

option to harvest a crop of wheat and a crop of cotton from one parcel of land in one year. The economic profitability of different intercropping systems depends critically on product prices, with lint being approximately ten times more valuable per unit weight than wheat. As long as the price ratio stays above six, the intercropping systems 3:2 and 4:2 are the most profitable. The current commonness of intercropping in the Huang-Huai-Hai plain of China is a reflection of the economic feasibility of the systems, but it is also affected by government regulations that require farmers to produce enough wheat, i.e. there is no completely free market in which producers can make purely economic decisions.

The high lint yield of cotton monoculture, in comparison to lint yields from intercropping at similar row length densities, shows the competitive effects of the wheat during the intercropping phase, lasting approximately seven weeks. The common relationship between row length density and yield in wheat monoculture and intercrop indicates that the response of wheat can be seen as a pure response to crop density, while competitive effects from the later sown cotton are negligible. The data in Fig. 3 thus indicate that competition between wheat and cotton is asymmetric. This asymmetry is not surprising because wheat is the taller species of the two. In intercropping systems, the taller species is often found to be dominant, capturing resources, notably light, at the expense of the smaller species. For instance, in maize/soybean intercropping, maize border row yield was increased at the expense of yield of border row soybean (Iragavarapu and Randall, 1996). The much greater wheat yield in border rows (Fig. 4) confirms that wheat profits from resources that “belong” to the space between the wheat strips.

The greater wheat yield in border rows can be attributed to greater light interception (Chapter 4) and a better acquisition of nutrients in the border rows. This capability of a wheat crop was also found by Xiao et al. (2004). We conclude that the number of border rows is a major factor in determining the advantage of wheat productivity in intercropping systems. The border row effect found in this study is twice as large as reported for other intercrops, e.g. 20%-26% increase of maize yield in border rows in a maize/soybean intercrop (West and Griffith, 1992; Ghaffarzadeh et al., 1994). Li et al. (2001b) reported that yield increase of intercropped wheat is due not only to yield increase of border rows but also of inner rows in the wheat/maize intercropping systems. Our results do not show such a yield increase in inner rows, and we find it difficult to envisage a causal mechanism for such an increase.

We conclude that relay strip intercropping of cotton and wheat greatly increases land productivity. The 3:1, 3:2 and 4:2 systems are similar in productivity, whereby the 3:1 system gives a relatively greater wheat yield and a relatively smaller cotton yield than


33

the 3:2 and 4:2 systems. The 6:2 system is suboptimal. Development of intercropped cotton is delayed by up to two weeks, compared to cotton grown in monoculture. The crop physiological background of this delay, and methods to overcome it, merit further study.

CHAPTER 3

Cotton development and temperature dynamics in relay intercropping with wheat*

* Field Crops Research (2007), Submitted L. Zhang a,b,c, W. van der Werf b, S. Zhang a, B. Li c and J.H.J. Spiertz b

a Cotton Research Institute, Chinese Academy of Agricultural Sciences, Key

Laboratory of Cotton Genetic Improvement, Ministry of Agriculture, Anyang, Henan 455004, P.R. China



Chapter 3

36

ABSTRACT

In the Yellow River valley of China, more than 1.4 million ha of cotton are grown as relay intercrops with wheat. Cotton is sown in April when winter wheat is already in the reproductive phase; thus, a wheat crop with a fully developed canopy will compete for resources with cotton plants in the seedling stage. Yields of cotton are lower in relay intercropping systems than in a monocrop, but the aggregate yield of the cotton wheat system is greater than of monocultures of the component crops. In this study, we tested the hypothesis that the lower yield of intercropped cotton is a consequence of delayed development of the cotton as a result of a modified microclimate in the intercrop.

Field experiments were conducted in three subsequent years in Anyang, Henan, China. Wheat and cotton were grown as monocrops and as intercrops. Four intercrop layouts were investigated, differing in arrangement of wheat and cotton rows: 3:1, 3:2, 4:2 and 6:2. Developmental stage of the cotton was recorded at regular intervals during the growing cycle while air and soil temperatures were measured with thermocouples at several soil depths and cross-row positions in the canopy.

Temperatures at and near the soil surface were substantially (on average 3 degrees) lower in intercrops than in monoculture, especially on sunny days, thus lowering the rate of temperature accumulation for cotton seedlings in intercrops, compared to monocultures. Cotton in intercrops showed a pronounced delay in early development, compared to monocrops. No significant differences were found between different intercropping patterns; the duration from planting to first square, expressed in thermal time, lasted 531 and 638-670 °Cd for the monoculture and the intercrops, respectively. As a result the open boll stage in intercrops was delayed by 10-15 calendar days, compared to cotton monoculture, and the number of fruit nodes per plant was reduced from 30.3 in monocrops to 19.9 in intercrops, averaged over three years. This decrease in reproductive capacity reduced cotton lint yield considerably.

A plastic film cover increased temperatures in a 3:2 intercrop at the soil surface by 1.9 °C and at 5 cm soil depth by 2.7 °C, thus restoring the thermal conditions to levels common in monoculture. A cover with straw, however, decreased the temperature at the soil surface by 2.9 °C and at 5 cm depth by 1.3 °C.

We conclude that the thermal climate in wheat-cotton intercrops is suboptimal for the cotton seedlings. The resulting delay in development of cotton culminates in a lower reproductive capacity and as a consequence in a lower lint yield. Agronomic measures like a plastic film cover, growing cotton on ridges or choice of semi-dwarf wheat or early maturing cotton cultivars may be employed to overcome this constraint.

Keywords: Air temperature; soil temperature; soil cover; thermal time; physiological time; phenology.

Cotton development and temperature dynamics in relay intercropping with wheat

37

INTRODUCTION

Intercropping of wheat and cotton is practiced at a large scale in China (Zhang and Li, 1997). In the Yellow River valley of China, more than 1.4 million ha of cotton are grown as relay intercrops with wheat. In this system, wheat and cotton are grown as a relay cropping system, in which the two component crops overlap partially in time. The wheat is sown in autumn (October or November) in strips that consist of three up to six plant rows, with strips of bare soil interspersed. The cotton is sown in the open spaces between wheat strips in spring (April); approximately seven weeks later, in June, the wheat is harvested. Subsequently, the cotton crop can develop a full canopy covering the available space like a monoculture. The cotton is harvested in October, after which wheat or another winter crop can be sown.

Cotton has an indeterminate growth pattern. The subsequent stages are described in detail by Kohel and Lewis (1984) and by Munger et al. (1998). The young plant first produces a taproot and main stem. The first four to nine side branches are vegetative while the subsequent side branches are generative. The generative branches produce flowers and fruits, one on each node. The number of fruit nodes per branch varies according to the conditions. The development of the fruits goes through the subsequent stages of square (flower bud), flower and boll. Mature bolls open up, exposing the lint. During the development of the plants, more and more bolls are produced, up to a point where these bolls monopolize all the available assimilates such that further flowers abort. Before this happens, growers “cut out” the terminal bud to prevent formation of further generative branches that would bear flowers that would not be able to produce mature fruit. Thus, the production of harvestable bolls is maximized.

Temperature has profound effects on many growth and development processes in cotton, including leaf growth, photosynthesis and respiration, fruit development and fibre quality (Reddy et al., 1991a; Reddy et al., 1991b; Reddy et al., 1992; Reddy et al., 1993; Braden and Smith, 2004). Low temperatures during the seedling phase slow down growth and development (Constable, 1976). Low temperatures, soil as well as air, delay the emergence and reduce seedling vigour, which often causes weak canopy establishment, risk of plant diseases and poor root growth.

In Chapter 2, we showed that the land equivalent ratio of cotton-wheat relay intercrops is greater than 1, indicating that productivity per unit land is enhanced in intercrops. However, the yield of cotton, the most valuable component of the intercrop in monetary terms, is lower in intercrop than in monoculture. In wheat-cotton strip intercropping systems, the cotton seedlings experience shading by the taller wheat

Chapter 3

38

plants, and this will affect both canopy and soil temperatures.

In this study, we explore the hypothesis that the lower yield of intercropped cotton is a consequence of delayed development of the cotton as a result of a modified microclimate in the intercrop. We hypothesize that the developmental delay caused during the early stages of seedling growth results in a reduced production of bolls, and as a consequence in a decreased lint yield. Understanding how the development of the cotton plant is affected by intercropping is essential to suggest avenues for optimizing intercropping systems.

The objectives of this study are to (i) quantify the developmental delay and the development of flowers and bolls in cotton-wheat intercrops in comparison to cotton monoculture; (ii) quantify air and soil temperature in cotton-wheat intercrops in comparison to cotton monoculture; and (iii) explore the ameliorative effect of soil covers on soil and air temperature in intercrops.


Experimental years and site

Field experiments were carried out at the Cotton Research Institute (CRI), Anyang city, Henan province, China during four subsequent growing seasons: 2001/2002, 2002/2003, 2003/2004 and 2004/2005. The research site is located in the Yellow River region, at 36°07´ N and 116°22´ E.

Main field experiment

The main experiment, to measure development, fruit growth and temperature in intercrops of wheat and cotton, was conducted in 2001/2002, 2002/2003, and 2003/2004. Six treatments were carried out in a randomized block design with four replicates. The treatments were: monocrop wheat (Triticum aestivum L.), monocrop cotton (Gossypium hirsutum L.), and four intercropping systems: 3:1 (i.e. 3 rows of wheat alternating with a single row of cotton), 3:2, 4:2 and 6:2 (i.e. 3, 4 or 6 rows of wheat alternating with 2 rows of cotton). The layouts of the four intercropping patterns are presented in detail by Zhang et al. (Chapter 2). A summary is given in Table 1. Wheat was sown on 4 November 2001, 2 November 2002 and 3 November 2003, and cotton was sown on 26 April 2002, 25 April 2003 and 25 April 2004. The wheat cultivar was ‘Zhongyu 5’ throughout. The cotton cultivars were mid-early maturing insect resistant Bt cotton ‘Shiyuan 321’ in 2002, and ‘CCRI 45’ in 2003 and 2004,


39

respectively. Maximum and minimum temperatures during the cotton growing season in each year are presented in Fig. 1.

Table 1: Characteristics of wheat-cotton strip intercropping systems and the monocultures

Strip width (m)b Homogenized density (plants per m2)c

Cropping system

Total width (m)a Wheat Cotton Wheat Cotton

3:1 1 60 40 524 ± 18 4.9 ± 0.1 3:2 1.2 60 60 425 ± 26 7.5 ± 0.2 4:2 1.5 85 65 495 ± 27 6.3 ± 0.2 6:2 2.0 130 70 471 ± 22 4.7 ± 0.1 Sole wheat 0.2 20d - 725 ± 28 - Sole cotton 0.8 - 80d - 6.1 ± 0.2

a Total width refers to the width of a minimum set of component crop rows. b Strip width is the sum of all row spacings assignable to one component crop. The row

spacing between adjacent cotton and wheat rows is thereby equally divided between the two crops.

c Homogenized density (± SE) is the number of wheat ears per m2 or, respectively, the number of cotton plants per m2. Data for wheat were taken on the 15th of March in 2003 and 2004 and those for cotton on the 15th of September in 2002, 2003 and 2004.

d Row distance is 20 cm in wheat monoculture and 80 cm in cotton monoculture.

Effect of soil cover on temperature in the intercrop

This experiment was conducted in a 3:2 intercrop to quantify the effect of cover materials on soil temperature during the intercropping phase. The two cover materials tested were (i) a 0.008 mm thin white plastic film and (ii) a layer of wheat straw of about 1 to 2 cm thickness. In the control intercrop, no cover material was used. Wheat cultivar ‘Zhongyu 5’ was sown on 1 November 2004 and the cotton cultivar ‘CRI45’ on 27 April 2005.

Measurements

Development

After the establishment of the seedlings, ten plants in one row per plot were labelled to monitor development stages. Stage descriptions of Munger et al. (1998) were used. Particular attention was given to the time at which 50% of the plants had emerged (emergence), when 50% of the plants had produced a square (squaring), when 50% of

Chapter 3

40

the plants had produced at least one open flower (flowering), and when 50% of the plants had produced at least one open boll (open boll). A square was considered “appeared” when its subtending leaf had unfolded. Observations were made at least three times per week.

Organ number

Counts were made every two weeks of the number of leaves on the main stem, leaves on fruit branches, fruit branches, fruit nodes, squares, flowers, young bolls, big bolls and open bolls.

Temperature

Air temperatures were measured in a Stevenson screen on site. Hourly measurements were made in 2003, 2004 and 2005. Temperatures in the canopy and in the soil were measured with T-107 sensors (Campbell Sci., Logan, UT), consisting of a thermistor encapsulated in a cylindrical aluminium housing. Placement of sensors is shown in Fig. 2. Temperatures were measured every 30 minutes and an hourly average was recorded, with an automatic data logger CR23X (Campbell Sci., Logan, UT).

Quantification of heat units

Two systems are used to record the accumulation of heat units; (i) degree-days Monteith (1977), and (ii) physiological time. The degree-day

0

5

10

15

20

25

30

35

40

45

Tem

pera

ture

(o C)

0

5

10

15

20

25

30

35

40

45

Tem

pera

ture

(o C)

0

5

10

15

20

25

30

35

40

45

0 20 40 60 80 100 120 140 160

DAS

Tem

pera

ture

(o C)

TmaxTmin

a 2002

b 2003

c 2004

Fig. 1: Maximum and minimum air temperatures in Anyang during three cotton growing seasons Time axis is days after sowing. Sowing dates are (a) 26 April 2002; (b) 25 April 2003; and (c) 25 April 2004


41

concept is based on a linear relationship between development rate and temperature above a threshold (Bonhomme, 2000; Yang et al., 2004; Thompson and Clark, 2006). The physiological time concept (Diekmann, 1996) assumes also this linear increase above a threshold, but in addition, it lets the development rate decrease linearly with temperature above an optimum temperature (Campbell and Norman, 1998; Soltani et al., 2006).

W5 W3 C110 cm5 cm soil depthSoil surface Bottom

Above wheat

W2W4

N

E2 E3 E4

Below soil surface

Row direction

Fig. 2: Placement of temperature sensors in a 4:2 wheat-cotton intercrop to measure spatial variability cross-row in temperature in the intercrop. The placement between the two cotton rows is considered the centre (C1). Positions E2 (east-2) and W2 (west-2) are underneath the plants in the cotton row. Position E3 and W3 are midway between cotton and wheat rows. E4 and W4 are below the border wheat rows. Positions W5 and E5 (not shown) are in the middle of the wheat strip, i.e. midway between the second and third wheat row. The coding of sensor placement in other systems was analogous; 1 = centre, 2 = beneath cotton row, 3 = midway between cotton and wheat, 4 = beneath border wheat row, 5 = in the centre of the wheat strip. Row orientation in the intercropping system was approximately –15 degrees

Thermal time (TT) is calculated by accumulation of the daily mean temperature T above a base temperature Tb:

( )TT Max 0, bT T= −∑ (1)

where, T is calculated as the mean of the measured minimum and maximum temperature.

Physiological time (PT) is calculated (Hanninen, 1995; Cao and Moss, 1997) as by:

Chapter 3

42

0

( )

0

b

bb o

o b

mo m

m o

m

T TT T T T TT T

f TT T T T TT T

T T

<⎧⎪ −⎪ ≤ ≤

−⎪⎪= ⎨ −⎪ ≤ ≤⎪ −⎪

>⎪⎩

(2)

where, T is air temperature, Tb the base temperature, To the optimal temperature and Tm the maximum temperature. For cotton, Tb = 12 °C, To = 30 °C, and Tm = 35 °C (Anonymous, 1982; Hearn and da Roza, 1985; Zhang et al., 2003a; Zhao et al., 2005). PT is calculated as the sum of daily values of f(T). Development is completed when PT attains a value that is equal to the number of days needed for development under optimal conditions, i.e. the number of days that development lasts at To (Soltani et al., 2006).

Data analysis

Data were analysed using analysis of variance (ANOVA) in SPSS 11.0. Least significant differences (LSD) were used to separate treatment means at α = 0.05.

RESULTS

Phenology

Calendar time

In cotton monoculture, the time needed for the crop to reach the open boll stage was 138, 148 and 151 days in 2002, 2003 and 2004, respectively. In intercrops, the time until open-boll was invariably longer than in monoculture by 11-14 d., by 15 d. and by 9-12 d. in 2002, 2003 and 2004, respectively (Table 2). Significant differences in time to open-boll were found between different intercrops in 2002 and 2004, but the ranking among intercrops was not consistent over the years (Table 2). Duration till the open boll stage was related to temperature; development was fastest in the warmer year 2002.


43

Table 2: Period from sowing to open-boll stage in inter-cropping and monoculture, 2002-2004

Duration from sowing to open boll §(d) Cropping pattern

2002 2003 2004

3:1 (3W:1C) 149 b1 163 a 163 a

3:2 (3W:2C) 152 a 163 a 163 a

4:2 (4W:2C) 149 b 163 a 161 b

6:2 (6W:2C) 148 b 163 a 160 c

Monoculture 138 c 148 b 151 d

SE 1.1 0.7 0.2 § Open boll stage indicates that 50% of the plants have

produced at least one open boll. 1 Means followed by the same letter in same column do

not differ significantly at P < 0.05. Table 3: Physiological time (PT) and Thermal time (TT) accumulated during three development phases in intercrops and monoculture, 2002-2004

PT (d)a TT (°C d)a Cropping pattern P-Sb S-Fc F-Od P-S S-F F-O 3:1 32.5 a§ 14.8 a 41.3 a 670 a 301 a 803 a 3:2 31.2 a 15.7 a 41.3 a 638 a 325 a 803 a 4:2 31.2 a 15.7 a 41.4 a 638 a 325 a 806 a 6:2 31.2 a 15.7 a 41.7 a 638 a 325 a 815 a Monoculture 26.8 b 15.0 a 41.5 a 531 b 333 a 828 a SE 0.6 0.5 0.4 15 11 9

a PT indicates physiological time and TT is thermal time. Physiological time is the accu-mulation of the daily “effective development”, normalized to 1 for optimal development. Thermal time is the sum of daily mean temperature above a threshold of 12 °C.

b P-S is the duration from planting to 50% first square. c S-F is the duration from 50% first square to 50% first flower. d F-O is the duration from 50% first flower to 50% open boll. § Means followed by the same letter in same column do not differ significantly at P < 0.05.

Chapter 3

44

Thermal time

The average duration from sowing to open boll in cotton monoculture was 1692 °Cd, when measured in thermal time, and 83.3 d, when measured in physiological time. Both TT and PT had similar values in the three years, confirming the usefulness of these quantities to account for temperature effects on development. The duration from sowing to open boll in intercrops, expressed as TT or PT on the basis of temperatures measured in a Stevenson screen, was longer than in the monoculture. The delay was caused during early development, from planting to first square (P-S), which lasted 531 °Cd in monoculture and 638-670 °Cd in intercrops, or – expressed in physiological time – 26.8 days in monoculture and 31.2-32.5 days in intercrops (Table 3). The duration of subsequent developmental phases was similar in monoculture and intercropping (Table 3).

Rate and duration of organ development

Number of leaves

At the time of wheat harvest, the number of leaves on the main stem of cotton in intercrops was on average two leaves less than in the monoculture (Fig. 3). The number was lowest in the 3:1 system. The leaf appearance rate in monoculture was 0.36 d–1. The rate of leaf appearance on the main stem in the 3:1, 3:2, 4:2 and 6:2 systems was 0.25, 0.26, 0.26 and 0.27 d–1, respectively, before the wheat was harvested and 0.39, 0.37, 0.37 and 0.37 d–1 thereafter.

Number of fruit branches and nodes

There were consistent and significant differences between monoculture and intercropped cotton in the number of fruit branches and the number of fruit nodes per plant (Table 4). The average number of fruit branches per plant in monoculture was 11.1, 12.6 and 9.1 in 2002, 2003 and 2004, respectively, while in intercrops, it varied

0

2

4

6

8

10

12

14

16

35 40 45 50 55 60 65 70 75

DAS

Leav

es (p

lant

-1)

3:13:24:26:2Sole cotton

Fig. 3: Number of leaves on the main stem in intercrops and monoculture, 2003


45

from 8.0-9.3, from 10.4-11.6 and from 6.8-8.0 in the three subsequent years of the study. Averaged over the three years, plants in monoculture had 10.9 fruit branches, and those in the intercrops 9.0; a difference of 1.9 fruit branches (17%) per plant. The average number of fruit nodes per plant in monoculture was 34.8, 34.9 and 21.2 in 2002, 2003 and 2004, respectively. In intercrops, it varied from 14.5-22.3, from 24.0-30.1, and from 10.5-16.2 in successive years. Averaged over the three years, plants in monoculture had 30.3 fruit nodes, and those in the intercrops 19.9; a difference of 10.4 fruit nodes (34%) per plant.

Table 4: Number of fruit branches and fruit nodes per plant in intercropped and sole cotton one week before cut out§ in 2002-2004※

Cropping pattern SE Plant organ Year 3:1 3:2 4:2 6:2 Sole

2002 8.0 a1 9.3 a 8.4 a 9.0 a 11.1 b 0.56 2003 10.4 a 10.6 a 11.5 ab 11.6 ab 12.6 b 0.46

Fruit branch

2004 6.8 a 7.5 ab 8.0 b 7.3 ab 9.1 c 0.28 2002 14.5 a 22.3 a 19.9 a 22.2 a 34.8 b 3.40 2003 24.0 a 24.3 a 28.1 ab 30.1 b 34.9 b 1.87

Fruit node

2004 10.5 a 13.3 ab 16.2 b 13.0 ab 21.2 c 1.24 ※ Observations were made on 25 July 2002, 25 July 2003, and 15 July 2004. § Cut out is the removal of the terminal bud, to stop further terminal growth and increase in

the number of fruit branches. 1 Means followed by the same letter in same row do not differ significantly at P < 0.05.

Number of fruits

The formation of squares, flowers and bolls was delayed in intercropping, compared to monoculture (Fig. 4). In 2003, the number of bolls did not change after 120 DAS (days after sowing) due to the removal of the terminal bud (‘cut out’). However, in 2004, the number of bolls in the 3:2 and 4:2 intercrops kept increasing after cut-out due to further elongation of fruit branches. The number of fruit branches per unit land area was higher in these two systems due to the higher plant densities of cotton (Table 1).

Chapter 3

46

Fig. 4: Trends in number of squares, flowers and bolls in intercropping and monoculture in 2003 and 2004. Densities are expressed per unit area of the whole system

Temperature dynamics

Air temperature below cotton canopy

The temperatures experienced by the cotton seedlings were lower in the intercrops than in the monoculture. Measurements in a 3:2 intercrop in 2003 are given as an example (Fig. 5). Averaged over the final 12 days of the intercropping period (from 33 DAS to 45 DAS in 2003), temperatures in this intercrop were on average 1.5 °C lower than in the monoculture. The difference in temperature varied with the level of incoming radiation; from –2.7 °C (e.g. at 35 DAS in the 3:2 intercrop in 2003; Fig. 5)

0

25

50

75

100

125

150

175

200

Num

ber o

f squ

ares

and

flow

ers

(no

m-2

)

0

20

40

60

80

100

120

140

60 75 90 105 120 135 150DAS

Num

ber o

f bol

ls (n

o m

-2)

3:1 3:2 4:2 6:2 Sole cotton

60 75 90 105 120 135 150DAS

a 2003 b 2004

c 2003 d 2004


47

on sunny days to negligible on overcast days (e.g. at 39 DAS in the 3:2 intercrop in 2003; Fig. 5). An example of the difference in air temperatures between monoculture and intercrop is given in Fig. 5. After the wheat harvest, air temperatures did not differ between intercrops and cotton monoculture.

Fig. 5: Air temperature below a cotton plant for a 3:2 intercrop and sole cotton during the intercropping period in 2003

Fig. 6: Soil temperature at 5 cm depth below a cotton row in three different intercropping systems during three different phases of the interaction (a) around sowing; (b) during intercropping; (c) around the harvest of wheat. The time of sowing is indicated by an arrow pointing downward (↓). The time of wheat harvest is indicated by an arrow pointing upward (↑)

15

20

25

30

30 35 40 45 50

DAS

Tem

pera

ure

(o C)

3W:2Csole C

10

13

16

19

22

25

28

-12 -9 -6 -3 0 3 6 9 12DAS

Soil

tem

pera

ture

(o C)

3:1 3:2 6:2 Sole cotton

21 24 27 30 33 36 39 43 46 49 52 55

a b c

Chapter 3

48

Soil temperature

Soil temperature differed significantly between intercrops and monoculture. The difference was greatest, up to 4 °C, before the sowing of cotton, and amounted to 3 °C during the intercropping period. The difference in soil temperature vanished after the harvest of the wheat. Examples of soil temperatures at 5 cm depth during different periods are given in Fig. 6. There was little difference in soil temperature between different intercropping systems. On days with low soil temperatures due to lack of direct radiation, soil temperatures were the same in the intercrops and the monoculture. During approximately two weeks after the date of cotton sowing, soil temperatures in intercrops were close to those in the monoculture due to soil tillage.

Diurnal course in soil temperature

Measurements were made at different placements in the intercrop: in the centre (C1) of the cotton strip, that is, between the cotton rows in the 3:2, 4:2 and 6:2 systems, and underneath the cotton row in the 3:1 system; within a cotton row in the 3:2, 4:2 and 6:2 systems (E2 and W2), midway between adjacent cotton and wheat rows (E3 and W3), underneath the border wheat rows (E4 and W4) and in the centre of the wheat strip (E5 and W5), i.e. between the centre rows in 4:2 and 6:2, and underneath the centre wheat row in 3:1 and 3:2. Measurements in the monocrops were made in the rows. The following patterns were found (Fig. 7):

Daily amplitudes of soil and air temperature near the soil surface on sunny days were vast: in the order of 20 °C;

Amplitudes of daily variation in air and soil temperature were much smaller on cloudy days than on sunny days;

Sole cotton had the highest soil and air temperatures, both at night and day time;

The differences in temperature between systems were much larger during daytime then at night;

Sole wheat and the centre of the wheat strip had much lower daytime temperatures than the cotton; night time temperatures underneath the wheat canopy were marginally higher than in intercropped cotton, but lower than in sole cotton;

Different placements in the cotton strip showed marginally different courses of temperature that appear consistent with differences in direct radiation reaching the soil, in relation to shading by the wheat or by the cotton seedlings.


49

Fig. 7: Diurnal temperature patterns in cotton mono-culture and at different placements in three different intercropping systems. Coding of placements (cf. Fig. 2): C1 = centre (beneath cotton row in 3:1), W2 = beneath western cotton row, W3 = midway between cotton and wheat, W5 = in the centre of the wheat strip

14

16

18

20

22

24

26

28

30

Soil

tem

pera

ture

(o C)

14

16

18

20

22

24

26

28

30

0 2 4 6 8 10 12 14 16 18 20 22 24Hour

Soil

tem

pera

ture

(o C)

C1 W2W3 W5Sole cotton Sole wheat

10

15

20

25

30

35

40

45A

ir te

mpe

ratu

re (o C

)

a

b

c

3:2, May 30, 2003

At bottom

3:1, May 18, 2004

At 5 cm depth

6:2, May 18, 2004

At 5 cm depth

Chapter 3

50

Fig. 8: Temperature at the soil surface and at 5 and 10 cm depth, underneath a soil cover of plastic film or straw in a 3:2 intercrop within the cotton rows in 2005

10

12

14

16

18

20

22

24

26

28So

il te

mpe

ratu

re (o C

)

10

12

14

16

18

20

22

24

26

28

Soil

tem

pera

ture

(o C)

10

12

14

16

18

20

22

24

26

28

-10 -5 0 5 10 15DAS

Soil

tem

pera

ture

(o C)

Bare soil Film Straw

Soil surface

5 cm depth

10 cm depth

b

c

a


51

For instance, the middle of the cotton strip in the 3:2 system reached a lower maximum temperature than the sole cotton (Fig 7a), while the sensor placed midway between the cotton and wheat rows reached the lowest maximum, which could be readily explained by shading from the wheat. Likewise, in the 6:2 system (Fig. 7c) soil temperature between adjacent cotton and wheat rows reached higher afternoon temperatures than soil directly underneath the cotton row. The data presented in Fig. 7b clearly illustrate the diminished potential for temperature accumulation by cotton in intercrops with wheat, as compared to monoculture.

Modifications of heat resource

Soil cover had substantial effects on soil temperature in the 3:2 intercrop (Fig. 8). Average soil temperature under a film cover increased by 1.9 °C at the surface, by 2.7

°C at 5 cm depth and by 2.5 °C at 10 cm depth, while soil temperature under a straw cover decreased by 2.9 °C at the surface, by 1.3 °C at 5 cm depth and by 0.7 °C at 10 cm depth.

DISCUSSION

Phenology and heat requirement

Cotton development in wheat-cotton relay strip intercropping systems was delayed by 9-15 calendar days, compared to cotton in monoculture. This delay (from planting to the first square) corresponded with 4.7 Physiological days defined by the PT method and 115 degree-days expressed as thermal time. Our results are similar to those reported for wheat-soybean intercropping, where the development of the relay intercropped crop was also affected at an early stage (Wallace et al., 1992). However, when comparing the development delay to the results reported for other intercropping systems (Matthews et al., 1991; Gethi et al., 1993; Bukovinszky et al., 2004), the delay in wheat-cotton intercropping is large and has serious consequences. The yield components such as fruit numbers of cotton in wheat-cotton intercropping were much affected by the developmental delay, while the yield of suppressed crops in the reported intercropping were mostly affected by competing natural resources with competitive crops, i.c. light, nitrogen, water.

Chapter 3

52

Temperature and development

During the intercropping, air and soil temperatures below cotton rows in intercrops were lower than in monoculture of cotton. Roussopoulos et al. (1998) reported that when a cotton growing season is only 1 °C cooler on average, it will considerably delay maturity. Reddy et al. (1992) found that main stem elongation, leaf area growth and biomass accumulation were all very sensitive to temperature, at 3 weeks after emergence of the cotton. Leaves expanded at least three times more rapid at 31 °C than at 19 °C in pima cotton (Reddy et al., 1993). Our results show that the leaf appearance rate decreased by 28% during the intercropping period. Likewise, in maize, increased soil temperatures accelerated the rates of leaf tip appearance and full leaf expansion, enabling the crop to attain maximum leaf area index more rapidly (Stone et al., 1999). Similarly, we found that a lower soil temperature was an important factor in delaying growth and development of cotton seedlings. Although intercropping decreased air and soil temperatures during the shading period, a chilling effect or ‘cold shock’, as defined by Michael and Milroy (2004), did not play a role, because the minimum temperatures did not differ much between intercrops and sole cotton.

Our results indicate that the decrease of both air and soil temperature was mainly caused by shading of wheat, because temperatures were not different between intercrops and monoculture when the air temperature was low due to the absence of direct radiation. A soil cover consisting of plastic film effectively ameliorated soil temperature in intercropping. Licht and Al-Kaisi (2005) found an increased soil temperature as a result of tillage. Thus, tillage may also provide an option to raise soil temperature in intercrops. An improvement of temperature may be also achieved by measures that diminish shading, such as the use of shorter wheat cultivars, or planting cotton on ridges.

Spatial heat variation and system design

Among the tested intercropping systems, the 3:2 and 4:2 systems seem optimal, while cotton in the 3:1 system suffers a greater developmental delay than in other systems and cotton in the 6:2 system has difficulty to completely cover the soil after wheat harvest, due to the wide gap that is left between two adjacent cotton strips after the harvest of the interspersed wheat strip consisting of six rows. Any increase in heat loading and radiation penetration into the cotton strip by widening the strip width for cotton is expected to be accompanied by a reduction in land cover and radiation interception by the wheat, and – depending on the width of the wheat strip – a reduction in land cover percentage by cotton after wheat harvest, due to larger row


53

distance. Moreover, as strips get wider, it becomes more difficult to achieve high use efficiencies of other resources, like nutrients and water, because part of the land is not covered. Thus, options to improve system design by modifying strip width and the number of plant rows per strip seem limited.

Improvement of agronomic practices

Intercropping delayed the stage at which the number of fruit branches and nodes reached their maximum, and it decreased the number of bolls, thus decreasing lint yield. The results suggest that there are several ways to enhance yield of intercropped cotton: (i) breeding of early maturing varieties to advance the time of squaring and reach the peak of fruit formation before ‘cut out’, (ii) agronomic measures that diminish shading, such as the use of shorter wheat varieties and growing cotton on raised beds; (iii) increasing plant density to obtain more fruits per unit of land. The 3:2 and 4:2 systems were superior in this respect to the 3:1 and 6:2 systems (Table 1).

Modification of heat resource

The soil temperature can be significantly increased by a soil cover with plastic film. The magnitude of temperature increase by a film cover can compensate for the decrease caused by intercropping. A cover with straw, which is usually applied to reduce weed growth or to maintain soil moisture (Bilalis et al., 2003) cooled the soil. In aerobic rice systems, daily mean soil temperature at 2 cm depth below the soil surface was 29 °C underneath a film cover, 25 °C underneath a straw cover and 25.3 °C underneath bare soil (Tao et al., 2006). These trends confirm our results. Before introducing soil cover practices at a large scale, an economic and environmental impact analysis should be carried out.

CHAPTER 4

Light interception and radiation use efficiency in relay intercrops of wheat and cotton*

* Field Crops Research (2007), Submitted L. Zhang a,b,c, W. van der Werf b, L. Bastiaans b, S. Zhang c, B. Li a and J.H.J. Spiertz b

a China Agricultural University, College of Agricultural Resources and Environmental

Sciences, Key Laboratory of Plant and Soil Interaction, Beijing 100094, P.R. China b Wageningen University, Plant Sciences, Crop and Weed Ecology Group, P.O. Box

430, 6700 AK Wageningen, The Netherlands c Cotton Research Institute, Chinese Academy of Agricultural Sciences, Key

Laboratory of Cotton Genetic Improvement, Anyang, Henan 455004, P.R. China

Chapter 4

56

ABSTRACT

In China, a large acreage of cultivated land is devoted to relay intercropping of winter wheat and cotton. Wheat is sown in strips with interspersed bare soil in October and harvested in June of the next year, while cotton is sown in the interspersed space in the wheat crop in April and harvested before the next wheat sowing in October. This paper addresses the question how strip width and number of plant rows per strip of wheat or cotton affect light interception (LI) and light use efficiency (LUE) of both component crops.

The measurements were carried out in field experiments during three consecutive years from 2002 to 2004 with monocultures of wheat and cotton and four intercropping designs differing in strip and path width as well as number of rows per strip. The intercrop systems were identified by the number of rows per strip of wheat and cotton, respectively, as 3:1, 3:2, 4:2 and 6:2. Total light interception over a season was calculated from LAI measurements, using a model for light interception in a row crop. Light interception and distribution of light during the intercropping phase was measured with sensors.

Wheat monocrops intercepted 618 MJ m–2 photosynthetically active radiation (PAR) from Mar-18 to harvest in 2002. Averaged over three years, wheat in the four intercrops (3:1, 3:2, 4:2 and 6:2, respectively) intercepted 83, 71, 73 and 75% as much PAR as the sole wheat. Cotton monocrops intercepted on average 444 MJ m–2 PAR from sowing to harvest in three years. Cotton in the four intercrops (3:1, 3:2, 4:2 and 6:2, respectively) intercepted 73, 93, 86 and 67% as much PAR as the sole cotton. No differences in LUE of wheat or cotton were found between systems.

The analysis indicates that the high productivity of intercrops, compared to monocultures, can be fully explained by an increase in light interception by the component crops per unit cultivated area. The model results indicate that light interception can be modified by choice of the number of crop rows per strip and strip width. The best distribution of light is attained in systems that have narrow strips. The potential for system improvement is discussed.

Keywords: Leaf are index (LAI); light use efficiency (LUE); photosynthetic active radiation (PAR); intercropping; competition.

Light interception and radiation use efficiency in relay intercrops of wheat and cotton

57

INTRODUCTION

China has three cotton growing areas, and in one of those, the Yellow River valley, the majority of the cotton (1.4 million ha) is grown in relay intercropping with wheat. In these systems, wheat is sown in autumn and harvested in early summer of the following year. Space is left open in the wheat crop to enable sowing of cotton before the harvest of wheat, resulting in a strip-based wheat crop that covers the land incompletely. From the sowing of cotton in April, till the harvest of wheat, in June, the cotton and wheat are growing simultaneously, competing for light, water and nutrients. During this phase, which lasts about seven weeks, the wheat crop shades the cotton plants that are still in the seedling stage. After wheat is harvested, the whole space is available for cotton, and the gaps that appear after the wheat harvest have to be bridged by the expanding leaf canopy of cotton. The spatio-temporal architecture of the intercropping system determines the pattern of light capture of both component crops.

Light interception (LI) and light use efficiency (LUE) are powerful concepts for characterizing the resource capture and use efficiency of cropping systems, including intercrops. Crop light use efficiency is defined by the often found linear relationship between accumulated biomass and cumulative intercepted PAR (Monteith, 1977; Russell et al., 1989); it represents the slope of the relation. Improved productivity can result from either greater interception of solar radiation, a higher light use efficiency, or a combination of the two (Willey, 1990). Light interception as a result of mixing two species and growing them together in stead of alone is sometimes increased, either as a result of a lengthening of the period of soil coverage (temporal advantage), or as a result of a more complete soil cover (spatial advantage) (Keating and Carberry, 1993). Resource use efficiency is not likely to be much affected in intercropping systems with component crops that differ in growing period, since competition between component crops is weak (Fukai and Trenbath, 1993).

Present-day wheat-cotton relay intercrops are strip-based rather than row-based. In a wheat-cotton intercropping system, wheat is cropped in strips with a spare path for intercropping cotton. After wheat is harvested, the cotton crop is also strip structured because the plants may not be able to bridge the space freed up by wheat. The whole cropping season includes three phases: (i) a wheat phase; (ii) an intercropping phase and (iii) a cotton phase (Fig. 1). Different intercropping designs are used in practice and their spatial configuration has been characterized by the number of wheat rows per strip and the number of cotton rows per strip; 3:1, 3:2, 4:2 and 6:2 (Chapter 2). Yield

Chapter 4

58

analyses demonstrated high land equivalence ratios: 1.28 in the 6:2 system, and 1.39 in each of the other three systems. To what extent the increased productivity of relay-intercropping of wheat and cotton is determined by light interception, by light use efficiency, or by a combination of those two is still unknown.

i

ii

iii

Wheat strip

Cotton strip

W120cm

P80 cm

W120 cm

H

H

P80 cm

i

ii

iii

Wheat strip

Cotton strip

W120cm

P80 cm

W120 cm

H

H

P80 cm

Fig. 1: Conceptual representation of the cross-row profile of a wheat-cotton relay intercrop as used in calculations of radiation interception by cotton and wheat with the row crop model (6:2 system). Crop phases: (i) wheat phase from February/March to end of April; (ii) intercropping phase from end of April to middle of June; (iii) cotton phase from middle of June to October

The fractions of the incoming PAR which are absorbed by canopies of component crops in intercrop systems mainly depend on leaf area index and canopy structure (Spitters and Aerts, 1983; Lantinga et al., 1999; Bastiaans et al., 2000). Although the principles are understood, Willey (1990) noted that it is a challenge to determine light capture by component crops in intercrops. Measurement is difficult, especially over a whole growing season, but several modelling approaches have been suggested to


59

calculate light interception in heterogeneous canopies. Wilkerson et al. (1990) describes an empirical approach based on a competitive factor and an ‘area of influence’. Detailed three-dimensional light interception models have also been developed (Whitfield, 1986; Gijzen and Goudriaan, 1989; Rohrig et al., 1999). A simplified approach, based on a block-shaped strip crop structure, was suggested and elaborated by Goudriaan (1977) and Pronk et al. (2003). This approach is used here to calculate light distribution in wheat-cotton intercrops, because it truthfully represents the geometry of this system (Fig. 1; cf. Fig. 1 in Chapter 2).

The objectives of this study are: (i) to characterize the spatial distribution of PAR in a cotton-wheat intercrop system, based on measurements during the intercropping phase; (ii) to estimate PAR interception by component crops and by both monocultures over a growing season, using a model for light interception in a row crop; (iii) to calculate LUE, based on the relationship between dry matter accumulation and cumulative intercepted radiation; and (iv) to explore options for improving cropping arrangements and geometry. Null hypotheses that pertain to this work are: (i) all systems have the same light capture; (ii) all systems have the same light use efficiency. Clearly, given the high LER of these systems, one of these hypotheses must be incorrect.

MATERIALS AND METHOD

Field experiments

Field experiments were conducted in 2001/2002, 2002/2003 and 2003/2004 at the Cotton Research Institute (CRI), Anyang, Henan province, China, 36°07´N and 116°22´E. The experiments comprised six cropping systems with wheat (Triticum aestivum L.) and cotton (Gossypium hirsutum L.), two of which were monocultures, while four were intercropping systems. The four intercropping patterns were characterized by the number of wheat and cotton rows that were alternated: 3 wheat rows: 1 cotton row (3:1), 3 wheat rows: 2 cotton rows (3:2), 4 wheat rows: 2 cotton rows (4:2), and 6 wheat rows: 2 cotton rows (6:2). Just as in the monoculture, distance between wheat rows in a strip was 20 cm. Systems that contained wheat strips with a larger number of wheat rows were also characterized by a larger gap-width between wheat strips. Consequently, the row length density of wheat, compared to monoculture, was very similar among intercropping systems varying between 50% (3:2) and 60% (3:1 and 6:2; Table 1). Gaps between wheat strips were planted with either 1 (3:1) or 2 (all other systems) cotton rows. In these last systems this resulted in

Chapter 4

60

an uneven distribution of cotton rows, with a distance of 40 cm for the rows planted in the same gap and distances varying from 80 cm (3:2) to 160 cm (6:2) between adjacent cotton rows of neighbouring gaps. The row length density of cotton in intercropping systems were either smaller (1.00 m m–2 for 3:1 and 6:2) or larger (1.33 m m–2 for 4:2 and 1.67 m m–2 for 3:2) than in monoculture (1.25 m m–2; row distance 80 cm). Detailed information on the design parameters of all systems are presented in Table 1 and 2. Maps of the various systems are given in Chapter 2. Cotton development is presented in Chapter 3 and wheat development is expressed according to the decimal code of (Zadoks et al., 1974).

Table 1: Geometry of wheat and cotton strips at three growth phases in wheat-cotton relay intercropping systems

Wheat phase Intercrop phase Cotton phase Cropping system

Width of single unit

Wheat rows1

Wheat rows lost1

Gap-width between wheat rows

Distance of cotton to nearest wheat row

Cotton rows1

Distance between cotton rows2

(cm) # # (cm) (cm) # (cm)

Sole wheat 20 1 0 - - 0 - Sole cotton 80 0 4 - - 1 80 3:1 100 3 2 60 30 1 100 3:2 120 3 3 80 20 2 40/80 4:2 150 4 3.5 90 25 2 40/110 6:2 200 6 4 100 30 2 40/160

1 indicates number of crop rows in a width of single unit. 2 indicates uniform row distance in cotton monoculture and 3:1 system and narrow-wide row

distance in a single unit of 3:2, 4:2 and 6:2 systems.

Determination of above-ground dry matter

To assess the above-ground biomass, 1 m of each row within an adjacent wheat plus cotton strip (“minimum combination” as defined in Chapter 2) of each plot was sampled once per two weeks. For wheat, the dry matter (DM) measurements were made from 18 March to 10 June 2002, from 20 April to 11 June 2003, and from 13 April to 31 May 2004. DM measurements in cotton were conducted from 14 June to 18 September 2002, from 29 May to 15 September 2003, and from 17 June to 7


61

September 2004. First, the number of plants was counted, then a sample was randomly selected for detailed analyses; the wheat sample consisted of twenty plants while the cotton sample consisted of ten seedlings during early growth and three plants later on. The samples were oven-dried at 65 °C to constant weight to determine dry matter (DM). Details on the measurements of dry weights are presented in Chapter 2.

Table 2: Parameters describing density, strip width, path width and maximum plant height of component crops in wheat-cotton strip intercropping systems, as used in the model for calculating light interception

Row length density※ Parameters of a strip structured canopy§ wheat Cotton Wheat Cotton (before wheat harvest) (after wheat harvest) W1 P2 H3 W P H

Intercropping pattern

m m–2 m m–2 cm cm cm cm cm cm 3:1 3 1 60 40 65 80 20 90 3:2 2.5 1.67 60 60 65 120 0 90 4:2 2.67 1.33 80 70 65 120 30 90 6:2 3 1 120 80 65 120 80 90 Monoculture 5 1.25 20 0 65 80 0 90

※ Row length density (RLD) is total row length of a component crop per unit intercrop area (m m–2, or rows per meter cross-row).

§ Whole width indicates the width of a minimal combination in wheat-cotton strip intercropping systems.

1 W indicates strip width of each crop, for wheat, is defined as the sum of the product of number of rows and row space (20 cm); for cotton, is defined as a narrow row distance (fixed 40 cm) plus each 40 cm for two sides of a cotton strip.

2 P indicates path width of each crop, is the whole width of intercropping minus strip width. 3 H indicates maximum height of wheat and cotton, the measured dynamics of wheat and

cotton height are used in models.

Light interception

Incoming radiation

In 2002, 2003 and 2004, daily global solar radiation was derived from sunshine hours, using Ångström’s equation (1924) with coefficients applicable to China (Zhou et al., 2005). Sunshine hours were measured at the experimental site. Daily incoming global

Chapter 4

62

radiation in 2002, 2003 and 2004 is presented in Fig. 2. Photosynthetically active radiation (PAR) was computed as 50% of global radiation.

In 2004, global solar radiation was measured with a pyranometer (LI-200SZ, LI-COR, Lincoln, NE, USA) and datalogger (CR10X, Campbell Sci., Logan UT). A comparison of daily incoming global radiation estimated with Ångström’s formula and measured with a pyranometer was made using data collected in 2004 (Fig. 3). The association yielded a coefficient of determination of R2 = 0.84, a bias of –0.6 MJ m–2 d–1, and a RMSE (root mean square error between observed and estimated values) of 2.9 MJ m–2 d–1. Based on this result, the prediction of the Ångström equation was deemed adequate.

Measurement of Leaf Area Index

Leaf area index of wheat was determined by using the same samples as used for determining DM weight. The number of plants per m row length over the width of a minimum combination of one wheat plus cotton strip, was counted. Homogenized plant density, i.e. density of one component species expressed per unit intercropping area, was then calculated by dividing the counted number of plants by total sample area. Next, length and width of each leaf were determined on ten randomly selected plants from each plot. Following (Miralles and Slafer, 1991), leaf area was calculated as:

Area = 0.835 × Length × Width

0

5

10

15

20

25

30

Rad

iatio

n (M

J m

-2 d-1

)

0

5

10

15

20

25

30

Rad

iatio

n (M

J m

-2 d-1

)

0

5

10

15

20

25

30

0 40 80 120 160 200 240 280

Day number

Rad

iatio

n (M

J m

-2 d-1

)

a 2002

b 2003

c 2004

Fig. 2: Seasonal trends in daily global radiation at Anyang, China, in 2002, 2003 and 2004


63

The total leaf area per plant was determined, and LAI was determined by multiplying plant density and leaf area per plant.

Leaf area index of cotton was also based on the 2 weekly sampling for dry weight measurements. Length and width of each leaf were measured as indicated in Fig. 4, and leaf area was then estimated as:

Area = c × Length × Width

The coefficient c was estimated by collecting 59 leaves of four plants on 20 August 2004 and 29 leaves of 2 plants on 26 August 2004 from two cultivars ‘CRI45’ and ‘33B’. Maximum length and maximum width of all leaves of sampled plants were measured and the areas of the individual leaves were measured with a Leaf Area Meter (AM200, ADC BioScientific Ltd. UK). The coefficient c was estimated as the slope of the relationship between measured area (y-variable), and the product of length and width of the leaf (x-variable). A good linear relationship through the origin was found (R2 = 0.98). The estimate of the coefficient c was 0.81 ± 0.006 (Fig. 5).

Leaf area index (LAI) was expressed as m2

leaf per m2 total mono- and intercrop area. The leaf area of a component crop is thus “homogenized” over the whole intercrop area.

Calculation of light interception

Cumulative light interception was computed from daily incoming radiation and the calculated fraction of intercepted radiation. The fraction of PAR intercepted daily,

0

5

10

15

20

25

30

35

0 5 10 15 20 25 30 35

Measured radiation (MJ m-2 d-1)

Est

imat

ed ra

diat

ion

(MJ

m-2

d-1

)

Bias:- 0.6RMSE=2.9Line: 1:1

Fig. 3: Relationship between measured and estimated global radiation at Anyang, China, in 2004, as measured with a pyranometer (x-axis) and as estimated from measured sunshine hours using Angstroms equation (y-axis)

L

W

Fig. 4: Leaf length (L) was measured from the point of attachment of the petiole to the leaf tip. Width (W) was measured as the greatest distance

Chapter 4

64

both for wheat and cotton, was calculated with the strip crop model of Goudriaan (1977), based on measurements of leaf area index, and parameters describing the geometry of the system and the height growth of the plants (Table 2). For all phases of the relay intercropping it was assumed that a strip-path geometry existed. Light interception in monocultures was calculated on the assumption of a homogeneous canopy.

)exp(1 iii Lkf −= (1)

where f is the fraction of light intercepted by the canopy, L the leaf area index, and i is component crop, k the light extinction coefficient. The value of k is 0.7 for wheat (Yunusa et al., 1993; Olesen et al., 2004) and 0.95 for cotton (Sadras, 1996).

Radiation interception by the wheat canopy was estimated from pseudo-stem erection (Zadoks scale 30) until harvest in 2002; however, light interception of wheat was calculated based on measurements started from beginning of anthesis (Zadoks scale 60) in 2003 and from flag-leaf sheath extension (Zadoks scale 41) in 2004.

The output of the model is the time course of the cumulative light interception by each component crop.

The equations for light interception in a strip crop are (Goudriaan, 1977; Pronk et al., 2003):

, , ,,

,

( )( )1 exp( )

i comp i p i s iint i i

i comp i

f f S Sf f

K L− −

= −−

(2)

)exp(1()( ,, icompi

ii

iicomp LK

PWW

f −+

= (3)

i

iiiicomp W

PWLL

)(,

+= (4)

)exp()1(, iiiiip LKaaS −−+= (5)

)exp()1()exp( ,, iiiicompiiis LKbLKbS −−+−= (6)

y = 0.8099xR2 = 0.9806

0

50

100

150

200

250

300

350

0 50 100 150 200 250 300 350

Product of length and width (cm2)

Mea

sure

d le

af a

rea

(cm

2 )

Fig. 5: Relationship between measured leaf area and product of leaf length and width in cotton


65

i

iiii P

HPHa

−+=

22

(7)

i

iiii W

HWHb

−+=

22

(8)

where, int refers to intercropping, and i is a component crop in the intercrop: wheat or cotton. fint,i is the fraction of light intercepted by a component crop with reference to the radiation incident on the whole intercropping system, fi is the proportion of incoming light intercepted by crop i in monoculture, and fcomp,i is the proportion of light incident per unit of component crop area in an intercrop that is intercepted by that component crop. Wi and Pi are widths of the crop strip and the path for wheat or cotton, which are taken to be constants, and Hi is plant height, which is input to the model according to interpolation between two-weekly measurements. Li represents the LAI of species i in the monocrop, while Lcomp, represents the LAI of species i per unit area of strip in the intercrop. Sp,i, Ss,i and ai and bi are intermediate variables, which are described in detail by Pronk et al. (2003).

During the intercropping phase, the young cotton plants are shaded by the wheat crop. The fraction of LI by cotton in this phase is therefore proportional to the amount of light transmitted to the strip with cotton seedlings.

', , ,int c p w int cf S f= (9)

where fint,c′ is LI of cotton during intercropping period, fint,c is Eq. 2 for cotton crop, and Sp,w is calculated according to Eq. 5 for wheat crop.

Calculation of light use efficiency (LUE)

LUE was calculated by regressing measured cumulative dry matter on cumulative intercepted PAR, for each plot, year and system separately.

Measurement of cross-row profiles of transmitted radiation in intercrops

Horizontal, cross-row profiles of transmitted photosynthetically active radiation (PAR, 0.4-0.7 μm) were determined by placing a 1.0 m long quantum meter (LI-190SB line quantum sensor, LI-COR, Lincoln, NE, USA), at different positions along a transect across the rows. The sensor was always placed at soil level and parallel to the crop rows. Measurements were carried out in the middle of the wheat strip, underneath the border wheat rows, midway between cotton and wheat rows, underneath cotton rows

Chapter 4

66

and in the centre of the cotton strip. Moreover, a reading was taken above the canopy. Placement of sensors in the 4:2 system is illustrated in Fig. 6. Measurements were done during the intercrop phase with a ripening wheat canopy and cotton seedlings in rows between the wheat strips on 1 May, 26 May, and 12 June 2002.

1.00

0.75

0.50

0.250.00

W5 W4 W3 W2 C1 E2 E4E3Placement

Can

opy

dept

h

NRow direction

Fig. 6: Setting of PAR sensors in the wheat-cotton strip intercropping system (4:2). Letters refer to the orientation: west (W), central (C) and east (E), and the numbers indicate the sequence from the central position of cotton strip to central position of wheat strip

The diurnal course of PAR was measured by individual quantum sensors (LI-190SZ, LI-COR, Lincoln, NE, USA) on 29 May and 12 June 2003, and on 6 May and 9 June 2004. Measurements were made while both wheat and cotton were present in the intercrop. Two measurements were made per hour. Hourly averages were recorded with a datalogger (CR23X, Campbell Sci., Logan, UT).

RESULTS

Dynamics of leaf area index

The homogenized LAI of wheat in intercrops was less than in monoculture (Fig. 7a, c and e). Expressed as a proportion of the LAI in monoculture, the LAI of wheat in intercrops was closely associated with the row length density. Thus, the LAI of wheat in intercropping systems was largely determined by the relative width of the space left for cotton.


67

Fig. 7: Growth pattern of leaf area index (LAI) for wheat and cotton in 2002-2004. Filled symbols indicate the monocultures of wheat or cotton. Arrows indicate the dates of cotton ‘cut-out’ (31 July in 2002, 29 July in 2003 and 1 August in 2004) Maximum LAIs of cotton were reached around 105 DAS (days after sowing) in all systems, 5 to 10 days after cotton ‘cut-out’ (Fig. 7b, d and f). The rate of increase in leaf area in the intercropping systems after the wheat harvest was lowest in systems with a low RLD (row length density), such as 3:1 and 6:2, and highest in the system with the greatest RLD (3:2). Maximum LAIs of intercropped cotton in the 3:1 and 6:2 systems were much lower than in monoculture, reflecting the low row length density

0123456789

70 80 90 100 110 120 130 140 150 160 170Julian day

LAI


0123456789

LAI

0123456789

LAI

0

0.5

1

1.5

2

2.5

3

3.5

0

0.5

1

1.5

2

2.5

3

3.5

0

0.5

1

1.5

2

2.5

3

3.5

160 180 200 220 240 260

a 2002 Wheat

c 2003 Wheat

e 2004 Wheat

b 2002 Cotton

d 2003 Cotton

f 2004 Cotton

Chapter 4

68

in these systems. Maximum LAIs of the 3:2 and 4:2 systems were at least as high as in monoculture, indicating that the high row length densities, compared to cotton monoculture, made up for the initial, shade-induced, delay in the growth of LAI.

Fig. 8: Cumulative light interception in wheat and cotton in the intercropping systems and the monocultures during the measuring periods in 2002 to 2004

-100

0

100

200

300

400

500

600

700

PAR

inte

rcep

ted

(MJ

m-2

)

-100

0

100

200

300

400

500

70 90 110 130 150 170Julian day

PAR

inte

rcep

ted

(MJ

m-2

)


0

100

200

300

400

500

600

0

100

200

300

400

500

0

100

200

300

400

500

140 160 180 200 220 240 260

-100

0

100

200

300

400

PAR

inte

rcep

ted

(MJ

m-2

)

a 2002 Wheat from Mar-18 to Jun-10

c 2003 Wheat from Apr-29 to Jun-11

e 2004 Wheat from Apr-13 to May-31

b 2002 Cotton from sowing to Spt-18

d 2003 Cotton from sowing to Spt-15

f 2004 Cotton from sowing to Spt-7


69

Light interception

The course of cumulative LI for wheat and cotton in the six different systems in the three years is given in Fig. 8. Monoculture wheat intercepted more light than wheat in intercrops. Wheat in the 3:1 system captured the most light of the four intercropping systems. The pattern was the same in the three years. The LI course of cotton showed an initial delay in all intercropping systems and this delay was never fully compensated for. From the wheat harvest onwards, the LI increased faster in the 3:2 and 4:2 than in the 3:1 and 6:2 systems, reflecting differences among these systems in RLD and LAI.

The total amount of light intercepted by wheat in the different systems is given for all three years separately in Fig. 9a, c and e. The amount of PAR intercepted in intercrop systems from stem elongation stage to harvest in 2002 ranged from 422 to 491 MJ m–2. This amount corresponds to 68-79% of the LI in the monocrop (618 MJ m–2). Differences between cropping systems were significant (P<0.01) except for the 3:2 and 4:2 system in 2002 (P=1.0) and the 3:2, 4:2 and 6:2 system in 2004 (P=0.57-0.95). Averaged over the three years, the PAR intercepted by wheat differed significantly among all systems (P<0.05) except for the 4:2 and 6:2 system (P=0.2). The amount of LI in 3:1, 6:2, 4:2 and 3:2 system was 83, 75, 74 and 71% of the monoculture, respectively. The lowest LI of wheat was found in the 3:2 system, being the intercrop with the lowest row length density of wheat. Light interception by wheat was significantly higher in the 3:1 system than in the 6:2 system. In both configurations 60% of the area was planted with wheat, i.e. the row length density was 60% of that in the monoculture, but in the 3:1 system the unplanted area was distributed over twice as many gaps, resulting in twice as many border rows that were able to intercept sideways incident radiation, thus compensating for the gaps that were only half as wide as in the 6:2 system (Chapter 2).

The amount of PAR intercepted in monoculture cotton from sowing to the open boll stage (September) ranged from 415 to 491 MJ m–2 in 2002, 2003 and 2004 (Fig. 9b, d and f). LI in monoculture was always significantly higher than in intercrops, except for LI in the 3:2 system in 2002, which was not significantly lower than LI in the monoculture. LI in the 6:2 system was always the lowest, though not significantly different from the 3:1 system in 2002. LI in the 3:2 system was always the highest of the intercrops, though not significantly different from the 4:2 system in 2002 and 2003. The 4:2 and 3:1 system held an intermediate position, but in all years, LI of the 4:2 system was significantly higher than LI in 3:1. Averaged over three years, the LI in

Chapter 4

70

the 3:2, 4:2, 3:1 and 6:2 systems amounted to 93, 86, 73 and 67% of the monoculture. The LI differed significantly among all cropping systems (P<0.01).

0

200

400

600

800

PAR

inte

rcep

ted

(MJ

m-2

)

0

100

200

300

400

PAR

inte

rcep

ted

(MJ

m-2

)

0

100

200

300

400

500

3:1 3:2 4:2 6:2 MonoCropping system

PAR

inte

rcep

ted

(MJ

m-2

)

0

100

200

300

400

500

600

0

100

200

300

400

500

600

0

100

200

300

400

500

3:1 3:2 4:2 6:2 MonoCropping system

a 2002 Wheat from Mar-18 to Jun-10

c 2003 Wheat from Apr-29 to Jun-11

e 2004 Wheat from Apr-13 to May-31

b 2002 Cotton from sowing to Spt-18

d 2003 Cotton from sowing to Spt-15

f 2004 Cotton from sowing to Spt-7

a

bd d c

c

ab b

c

a

a

be

c d

ab

c c c

c

b b

d

a

d

bc

e

a

Fig. 9: Amount of PAR intercepted by wheat and cotton in the intercropping systems and the monocultures (mono) during the measuring periods in 2002, 2003 and 2004

Thus, the monocrop cotton intercepted more light than the intercrops, which will be partly due to the observed growth delay of cotton in the intercropping phase of these systems. Increased planting densities, present in the 3:2 and 4:2 systems helped to minimize the losses in LI, but were not completely able to compensate for the losses


71

caused by the delay in canopy closure in intercrops. The 3:1 and 6:2 system, having a row length density below that of the monoculture, never reached full light interception and therefore intercepted considerable less radiation. The even distribution of cotton rows in the 3:1 system resulted in a smaller reduction in LI than the uneven row distribution in the 6:2 system.

Light use efficiency

The LUE of wheat was estimated as the slope of a fitted linear relationship between intercepted PAR and DM for each plot and each year and system. ANOVA of LUE values, taking into account data collected in all the three years, showed that wheat in intercrops and monoculture did not significantly differ in LUE (P=0.74). Differences Table 3: Light use efficiency (LUE) of wheat and cotton in intercropping systems and monoculture in 2002-2004

LUE Wheat Cotton

From stem elongation to harvest

During reproduction period

From sowing to boll open

Year Cropping system

g DM MJ (PAR)–1) 2002 3:1 2.71 b 2.58 a 1.37 a 3:2 2.83 b 2.59 a 1.29 a 4:2 2.85 b 2.65 a 1.55 a 6:2 2.85 b 2.36 a 1.55 a Monoculture 3.43 a 3.21 a 1.33 a SE 0.16 0.30 0.12 2003 3:1 2.05 a 1.16 ab 3:2 2.28 a 1.23 ab 4:2 1.90 a 1.20 ab 6:2 2.02 a 1.04 b Monoculture 1.41 b 1.32 a SE 0.14 0.07 2004 3:1 1.77 ab 1.44 ab 3:2 1.99 a 1.46 ab 4:2 1.64 ab 1.31 ab 6:2 1.45 b 1.20 b Monoculture 1.96 ab 1.52 a SE 0.16 0.08 Mean 3:1 2.13 a 1.32 a 3:2 2.29 a 1.33 a 4:2 2.06 a 1.35 a 6:2 1.94 a 1.26 a Monoculture 2.19 a 1.39 a SE 0.13 0.05 The same letter in same column subdivision means no significant difference according to LSD0.05.

Chapter 4

72

between years were significant (P=0.004). The interaction between cropping systems and years was borderline significant (P=0.058), indicating that the year effects were different among systems.

The LUE of wheat was significantly higher in the monoculture than in the intercrops (P=0.02) in the first year, but not in the other two years. Averaged over three years, the LUEs of wheat for the intercrops and the monoculture ranged from 1.94 to 2.29 g DM MJ (PAR)–1 during the reproductive period, and from 2.71 to 3.43 g DM MJ (PAR)–1 from stem elongation (Zadoks scale 30) to harvest in 2002 (Table 3).

The LUE of cotton in intercropping systems did not significantly differ from that in monoculture in 2002 to 2004 except for a lower value in the 6:2 system in 2003 and 2004 (Table 3). Averaged over three years, the LUE of cotton in intercropping systems and the monoculture did not differ significantly (P>0.05), ranging from 1.26 to 1.39 g DM MJ (PAR)–1.

Fig. 10: Fraction of light intercepted at different placements in intercropping systems from sunrise to sunset during the intercropping period (averages of data collected on 1 May, 26 May and 12 June, 2002)

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Frac

tion

of li

ght i

nter

cept

ed

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 20 40 60 80 100

Distance from central of wheat strip (cm)

Frac

tion

of li

ght i

nter

cept

ed

0 20 40 60 80 100 120 140

W5 W4 W3 C1 E3 E4 W5 W4 W3 W2 C1 E2 E3 E4

a 3:1

c 3:2

b 4:2

d 6:2

W5 W4 W3W2 C1 E2 E3E4 W5 W4 W3 W2 C1 E2 E3 E4


73

0

200

400

600

800

1000

1200

1400

1600PA

R d

ensi

ty (u

mol

m-2

s-1)

0

200

400

600

800

1000

1200

1400

1600

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Hour

PAR

den

sity

(um

ol m

-2 s-1

)

E3 E2 C1 W2 W4 W5 Sole cotton

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Hour

a 3:1 May 29, 2003 b 4:2 June 9, 2004

c 6:2 May 6, 2004 d 6:2 June 12, 2003

East row West row

Fig. 11: Daily courses of PAR density at soil surface at different placements in wheat-cotton intercrop systems and in monoculture in 2003 and 2004. The measurements taken in sole cotton were underneath cotton rows

Spatial distribution and diurnal course of PAR

During the intercropping phase, the fraction of PAR intercepted varied considerably across the rows in the system (Fig. 10). At the middle of the wheat strip (W5) the fraction of light intercepted ranged from 0.83±0.05 to 0.91±0.03, which was somewhat less than in the monoculture of wheat (0.94±0.01). From the wheat border row (E4 or W4) to the cotton row (E2 or W2; C1 in the 3:1 system), the fraction of light

Chapter 4

74

intercepted decreased considerably, but still light intensities were considerably lower than in the monoculture cotton: averaged from May 1 to June 12, the fraction of light intercepted by the monoculture of cotton was only 0.13±0.03. In the intercropping systems, measured close to the cotton row, the fraction of light intercepted was 0.68±0.06 in the 3:1 system, 0.53±0.07 in the 3:2 and 4:2 systems, and 0.50±0.06 in the 6:2 system, illustrating the strong reduction in light intensity compared to the monoculture cotton, and the effect of gap width between the wheat rows on shading, with comparatively strong shading in the 3:1 system, and relatively mild shading in the 6:2 system. The measurements were qualitatively in agreement with results of the row crop model.

Diurnal courses of PAR density are shown in Fig. 11. During the intercropping phase, the lowest PAR densities occurred in the middle of the wheat strips (W5; Fig. 11 b and c), due to the higher leaf area index of the wheat. Low PAR densities at the placements C1 (Fig. 11 a) and E2 (Fig. 11 b and c), illustrate the severe shading effects of wheat on cotton seedlings. Only in the middle of June in 2003, the PAR density in the sole cotton (underneath cotton rows) was during part of the day lower than the PAR density measured close to the cotton row in the 6:2 intercropping system. This probably reflects the evolving shading of cotton seedlings in the monoculture, where, at the end of the intercropping phase, the seedlings were already quite advanced. The PAR density reaching cotton had a minimum in the morning for east rows (E2; Fig. 11 a, b and c) and in the afternoon for west rows (W2; Fig 11 d), illustrating the effect of north-south oriented rows on the diurnal course of the shade cast by the wheat plants.


Results in this paper show that intercropping of wheat and cotton, using a relay-strip intercropping approach, increases the total capture of radiation, compared to monocultures of either crop. It was shown that this increase in resource capture can be held solely responsible for the high land equivalence ratios in these systems, as no significant differences in light use efficiency were found between monocrops and intercrops nor between the different intercropping systems. This result deviates from results of an analysis on nitrogen in the same system (Chapter 5), which demonstrated that in intercropped cotton, a higher total capture of N was combined with a lower nitrogen use efficiency, as compared to the monoculture of cotton.


75

Light use efficiency of wheat and cotton

Light use efficiencies of both wheat and cotton were not affected by intercropping. This is consistent with the literature, which indicates that the LUE of dominant component crops in intercropping systems is generally not affected. Examples include e.g. millet in millet/groundnut (Willey, 1990), sorghum in sorghum/groundnut (Matthews et al., 1991), and maize in maize/cowpea (Watiki et al., 1993). For the lower, shade crops, e.g. groundnut and cowpea, the LUE was sometimes slightly affected, potentially as a result of competition for other resources, i.e. water, nitrogen. The relay nature of the current intercrop makes that both crops can be considered dominant for most of the time; i.e. the wheat crop doesn’t receive any shading at all from the cotton, and although the cotton is strongly shaded during its seedling stage, the majority of light capture and dry matter production occurs in the period after harvest of the wheat, when cotton is the only and therefore dominant crop in the system.

The LUE of wheat in intercrop systems and monoculture during reproductive period ranged from 1.94 to 2.29 g DM MJ (PAR)–1 and ranged from 2.71 to 3.43 g DM MJ (PAR)–1 from stem elongation to harvest in 2002. This value is close to the value of 2.6 to 3.1 g DM MJ ( PAR)–1 reported by Kiniry et al. (1989) and within the range from 1.8 to 4.2 g DM MJ ( PAR)–1 reported by Olesen et al. (2002) and O’Connell et al. (2004). A lower LUE of wheat in the grain filling phase was also found for oilseed rape (Justes et al., 2000). The lower value of LUE at the end of the growing period could partly be explained by a lower photosynthetic capacity of reproductive organs compared to the leaves, while the photosynthetic rates of the leaves also decline due to N-reallocation and senescence (Justes et al., 2000).

The LUE of cotton in intercropping systems and monoculture ranged from 1.26 to 1.39 g DM MJ (PAR)–1, without significant difference between intercropping and monoculture. The measured range of LUE is consistent with the range of LUE values, 1.2 to 1.7 g MJ–1, reported for various genotypes of upland cotton (Rosenthal and Gerik, 1991; Pinter Jr et al., 1994; Sadras and Wilson, 1997).

Light interception and crop geometry

Increased light capture was shown to be the sole factor responsible for the high productivity of the wheat-cotton intercropping systems, reflected in the earlier reported high LER-values (1.28-1.39; Chapter 2) of these systems. The widths of wheat and cotton strips were found to have effects on total light interception as well as the distribution of captured light over both component crops. To some extent this might

Chapter 4

76

come as a surprise, as throughout the entire growing season of nearly 12 months, both crops are only simultaneously present for seven weeks. In this period competition, defined as mutual negative effect of species on one another, hardly occurs as, due to the huge size differences, their relation is largely unilateral, with wheat affecting cotton through shading. More importantly, both crops influence one another in a rather implicit manner. The width of the wheat strips and the distances between them determine the potential configurations of the cotton crop, after the harvest of the wheat. Differences in light capture and distribution of light over the two crops thus mainly result from the fraction of land area planted by the each crop (reflected in the row length density), the width of individual strips and the number of rows planted per strip.

In this study, width of wheat and cotton strips were closely related, resulting in systems in which the fraction area planted with wheat was always between 50 and 60% of that in monoculture. Patterns of LAI closely resembled these values relative to the LAI of the wheat monoculture. LI of wheat in the intercropping systems was however much closer to that of the monoculture and varied between 71-83%. Not surprisingly, systems with a higher fraction area planted with wheat were also found to have a higher fraction light interception by wheat. The high values of LI (71-83%), compared to the relative row length density of 50-60%, mainly result from additional light interception by border rows. Indeed, the yield is higher in border rows than in middle rows (Chapter 2).Comparison of the LI of the 3:1 and 6:2 system, both with an area planted of 60%, showed that a distribution of wheat in narrower strips increases light interception. This observation confirms the important role of border rows as a compensatory mechanism for LI.

The presence of wheat at the sowing of cotton, and in the seven weeks following from that, had a clear influence on the cotton seedlings. Detailed observations on light distribution in the intercrop showed that the shading effect on cotton seedlings was most severe in systems with a narrower cotton strip (3:1 system; light interception 68%), but even in the system with the widest cotton strip (6:2 system) half of the light was lost. Shading was found to have a significant effect on leaf area development of the cotton seedlings. At the time of wheat harvest this resulted in a delay of about 10 days, with modest differences between intercropping systems. Widening of the gap between wheat strips from 60 cm (3:1) to 100 cm (6:2) did not offer much relief from this problem. An increase in density of the cotton in the 3:2 and 4:2 systems, compared to monoculture, was a more effective way to enhance the LAI growth of cotton in intercrop.


77

After harvest of the wheat crop, the cotton starts to quickly develop its leaf area. Observations showed that in this phase row length density was the most important determinant of leaf area development. In the 3:2 and 4:2 system, being systems with a plant density exceeding that of the monoculture, the increased leaf area development resulted in maximum LAIs identical or even higher than in the monoculture cotton. Cumulative light interception reached values as high as 93% (3:2 system) and 86% (4:2 system) of that of the monoculture. Contrary, in systems with a lower row length density than the monoculture, leaf area development stayed markedly behind and cumulative LI reached values of only 73% (3:1 system) and 67% (6:2 system) of that of the monoculture. The significant difference in LI between the two last intercropping systems, which both were characterized by a row length density of 1 m m–2, demonstrates that a more even distribution of cotton rows improves LI, though to a much lower extent than an increase in row length density. Light capture of cotton in intercropping systems is thus markedly favoured by systems that allow for the creation of a high row length density. This asks for systems with relatively narrow wheat strips.

The light interception model used in this study provides a tool to evaluate systems for the most efficient capture of PAR. A sensitivity analysis was conducted to investigate the effect of parameter choices on calculated light interception of intercropped cotton, by using a homogeneous canopy model and a row structured with presented parameters. The result showed that the LI of intercropped cotton by using a row structured model was very close to that by using an homogeneous model except for 6:2 system, which is the most heterogeneous canopy of all systems, with wide gaps during the monoculture phases i and iii before and after the 7 weeks of intercropping.

This study focused on the effects of strip and path width and the number of crop rows per strip on light interception and productivity. Orientation of rows and strips may also affect light penetration. For instance, the growth of intercropped mungbean, that was shaded by a dominant tall maize intercrop, was favoured when rows were planted in the N/S direction (Dhingra et al., 1991). A narrower path between strips (or rows) of a taller crop might reduce this row orientation effects (Midmore, 1993). In our study, we found a PAR density difference between west row and east row in a near N/S orientated intercropping; thus it is necessary to be evaluated in the future.

The synthesis of results leads to the following conclusions with respect to optimal wheat and cotton intercropping systems: (i) wheat-cotton intercropping systems enjoy increased light interception in comparison to monocrops, partially by utilizing PAR during winter and spring by the wheat crop which would otherwise be ‘wasted’ when growing only a monoculture of cotton, and also by a comparatively high light interception by wheat in the strip crops with bare soil interspersed. (ii) total light

Chapter 4

78

interception is determined by the width of the strips of the component crops in relay intercropping systems; narrowing the path between the wheat strips increased the shading of cotton seedlings during the intercropping period; (iii) the currently used intercropping designs 3:2 and 4:2 are already optimal, and there is very little space for improvement, except by techniques that would increase early development and harvest index of the cotton crop; (iv) the most likely systems that could further improve resource capture and enable a rapid early growth and development of the cotton are systems with wider cotton strips that could allow for higher radiation interception and higher temperature environment for the cotton, or systems that use cultural techniques (raised beds, plastic film) to improve the light and temperature environment for the cotton. Above-mentioned options for system improvement will be investigated in future work.

CHAPTER 5

Nitrogen economy in relay intercropping systems of wheat and cotton*

* Plant and Soil (2007), Accepted L. Zhang a,b,c, J.H.J. Spiertz b, S. Zhang a, B. Li c and W. van der Werf b

a Cotton Research Institute, Chinese Academy of Agricultural Sciences, Key Laboratory for Genetic Improvement of Cotton, Ministry of Agriculture, Anyang, Henan 455004, P.R. China



Chapter 5

80

ABSTRACT

Relay intercropping of wheat and cotton is practiced on a large scale in China. Winter wheat is thereby grown as a food crop from November to June and cotton as a cash crop from April to October. The crops overlap in time, growing as an intercrop, from April till June. High levels of nitrogen are applied. In this study, we analysed the N-economy of the monocultures of cotton and wheat, and of four relay intercropping systems, differing in number of rows per strip of cotton or wheat. Field experiments were carried out from 2001/02 to 2003/04 in the Yellow River region in China. We quantified the nitrogen uptake and nitrogen use efficiency of wheat and cotton in relay intercropping systems to test if intercrops are more resource use efficient in comparison to monocrops.

Nitrogen (N) yields of wheat per unit area in the four intercropping systems were lower than in the monocrop, which ranged from 203 to 288 kg ha–1. The total N-uptake per unit biomass was similar between wheat in mono- and intercrops. On average, the N-yield of cotton per unit area was lower in intercrops than in monocrops, which ranged from 110 to 127 kg ha–1, but the total N-uptake per unit biomass was higher in intercropped cotton, as dry matter production was reduced to a greater extent by intercropping than N-uptake. The N-uptake of cotton was diminished during the intercropping phase, but recovered partially during later growth stages. The physiological nitrogen use efficiency (IE) of wheat was not much affected by intercropping, but it was reduced in cotton, due to delayed flowering and less reproductive growth.

Total N-efficiency of the system was assessed by comparing the relative nitrogen yield total (RNT), i.e. the sum of the ratio’s of total N-uptake by a component crop in the intercrop relative to the N-uptake in the monocrop, to the relative yield total. RNT ranged from 1.4 to 1.7, while the relative yield total (RYT) ranged from 1.3 to 1.4, indicating that intercrops used more nitrogen per unit production than monocrops.

An analysis of the crop nitrogen balance showed that the nitrogen surplus of sole crops amounted to 220 kg ha–1 for wheat and 140 kg ha–1 for cotton, while in the intercropping systems, the annual N surplus exceeded 400 kg ha–1. Conventional N-management in intercrops thus results in high N-surpluses that pose an environmental risk. The N management could be improved by means of a demand-based rate and timing of N applications.

Keywords: N content; internal N use efficiency (IE); crop N dynamics; relative N yield total (RNT).

Nitrogen economy in relay intercropping systems of wheat and cotton

81

INTRODUCTION

Wheat-cotton relay intercropping is widely practiced by farmers in the Yellow River cotton producing region of China (Chapters 1 and 2), because it meets the need of farmers to grow a profitable cash crop as well as to secure food supply. Relay intercropping enables the farmers to grow cotton, in a rotation with winter wheat, when the heat resource does not match the heat requirement of cotton if sown after harvesting winter wheat. The cotton is sown in April, approximately seven weeks before the harvest date of wheat. Strips are left open in the wheat crop at sowing (October/November) to provide space for the cotton plants during their seedling stage (April, May and June). After the wheat harvest in June, cotton plants can exploit the full space, above-ground as well as below-ground. A cotton-wheat relay intercropping system is thus characterized by three main phases: (i) winter wheat (vegetative stage) grown in strips from November till April; (ii) intercropping of wheat (reproductive stage) and cotton (seedling stage) from April till June, and (iii) sole cotton (vegetative and reproductive stage) from June till October (cf. Fig.1 in Chapter 4). The two component crops in the system interact directly only during the second phase; however the physiology, ecology and productivity of the relay strip intercropping system are determined by the spatial architecture and temporal dynamics of the leaf canopy and the root systems during the whole growing cycle. There is hardly any information available on the N utilization and requirement of cotton and wheat in relay intercropping systems. This information is needed to develop profitable and sustainable systems.

Willey (1990) suggested that intercrops may well improve the total nutrient capture by taking up nutrients that might be leached in a monocropping system. As a result intercropping could make greater demands on the soil. Still, even though the intercropping system as a whole may consume more nitrogen, each component crop in an intercropping system is likely to take up less nitrogen than in a monocrop situation, due to competition with the other crop. This, for instance, was found by Blaise et al. (2005), who reported that total N uptake in strip intercropping of cotton and pigeon pea in rainfed regions of central India was lower than in sole cotton because of a lower crop productivity.

Efficiency of N-use per each component crop is diagnosed by nitrogen use efficiency (NUE) and is commonly measured by internal (or physiological) efficiency (IE), expressed as kg yield per kg N uptake (Haefele et al., 2003). Several studies show that intercrops use soil nutrients more efficiently than sole crops, because of a higher N-recovery and increased dry matter yields (Hauggaard-Nielsen et al., 2001; Zhang et

Chapter 5

82

al., 2004a). However, agronomic N-use efficiency may also be decreased by intercropping. For instance, Aggarwal et al. (1992) found that upland rice intercropped with grain legumes took up the same amount of N as a monocrop, but produced significantly less dry matter.

To determine the land use advantage in terms of N yield for intercrops, a modification of the land equivalent ratio (Willey, 1979), expressed as a relative N yield total (RNT) (Baumann et al., 2001) or land equivalent ratio (LER) for N yield (NLER) (Szumigalski and Van Acker, 2006), can be employed. The RNT is calculated as the sum of the ratios of N-uptakes by component crops in the intercrop to their respective N-uptakes in monoculture. A value of RNT exceeding the LER suggests that intercropping is not nitrogen-efficient.

In many crops, N uptake is related with the accumulation of biomass and growth of leaf area over time (Booij et al., 1996; Lemaire et al., 2007). In case of N shortage, the actual uptake will fall short of N demand, and the concentration of N in the leaves will decrease more rapidly than in a well-fertilized crop. The robust relationships between N uptake, biomass accumulation and leaf area growth thus enable prediction of N uptake in time and provide a basis for the optimization of N application, both in quantity and in timing. Leigh and Johnston (1987) suggested a close association between leaf N concentration, rate of photosynthesis and biomass growth. To assess the balance between N uptake and demand, and to diagnose N deficiency the observed N dilution in the biomass can be compared to the minimal concentration of N in shoots necessary to produce the maximum dry mass (Justes et al., 1994). The trend in N dilution with increasing biomass can be represented by a power relationship (Flenet et al., 2006). The N dilution curve can be used to diagnose N deficiency, to manage N fertilization and for modelling N allocation (Lemaire et al., 2007). N dilution curves have been widely studied for monocrops but not for crops in intercropping systems. Considering the relatively high cost of nitrogen (N) fertilizer, and environmental concerns associated with excessive N application, increasing the N use efficiency of cropping systems is needed urgently.

The N-economy of a crop is the result of many processes occurring in the soil, crop and atmosphere and can be quantified by a wide variety of parameters (Spiertz and Vos, 2005; Peng and Bouman, 2007). In this study, we determined N-uptake in relation to DM-yield, physiological N use efficiency, and N-dilution in biomass over time. Furthermore, a nitrogen balance sheet analysis was made based on estimates of N-input and -output in the systems. We used different indicators at the crop and system level to analyse the nitrogen economy and use efficiency (Table 1).


83

Table 1: Indicators for analysis of N use in relay intercropping systems and monocultures1 Indicator/Symbol Parameter Dimension/Unit At crop level N uptake (Nu) Rate of N uptake per unit land area

per day g m–2 d–1

N content (Nc) N concentration in plant organs N content per unit biomass N content per unit leaf area

% ( or g kg–1) kg kg–1 kg m–2

SLN Specific leaf N content per unit leaf area

g m-2

N yield (Ny) Total N uptake per unit land area at final harvest

kg ha–1

Nitrogen use efficiency (NUE) expressed as internal efficiency (IE)

N utilization efficiency for grain or lint production; kg grain or lint per kg N uptake

kg kg–1

At system level Relative N yield total

(RNT) Combined N yield of intercrops relative to the monocrops

None

N balance N input minus N output kg ha–1 1 All indicators except RNT and N balance are per crop species.

The specific objectives are: (a) to quantify the nitrogen uptake and nitrogen use efficiency of wheat and cotton in relay intercropping systems; (b) to determine the relationships between N uptake and accumulation of biomass and growth of leaf area; (c) to quantify N dynamics of cotton in relation to the effects of intercropping; and (d) to explore the opportunities for a more effective nitrogen management of wheat-cotton intercropping systems.


Field experiments

Field experiments were conducted in 2001/02, 2002/03 and 2003/04 consecutively on the same field at the Cotton Research Institute of Chinese Academy of Agricultural Sciences (CRI, CAAS), Anyang city, Henan province, China at 36°07´North and 116°22´East.

Chapter 5

84

Soil parameters of the field are: sandy loam, pH 8.0, bulk density 1.36 g cm–3, and organic matter content 13.2 g kg–1. At the start of the experiments, in October 2001, the soil contained 1.02 g kg–1 total N, 0.52 g kg–1 P and 17.3 g kg–1 K. The amount of precipitation in 2002, 2003 and 2004 was 318, 539 and 517 mm, respectively.

For details on crop development we refer to Chapters 2 and 3.

Experiment 1: N economy of intercropping and monoculture systems

This experiment was conducted in 2001/02, 2002/03 and 2003/04, and comprised six treatments including four different intercropping patterns and monocultures of wheat (Triticum aestivum L.) and cotton (Gossypium hirsutum L.). The four intercropping patterns were all strip intercrops, with strips of cotton and wheat alternating. The number of wheat vs. cotton rows per strip in the four systems was 3:1, 3:2, 4:2 and 6:2. Row distance in wheat was 20 cm and between cotton rows 40 cm, but distance between cotton and wheat was varied (cf. Fig. 2 in Chapter 2). The experimental systems are characterized by the width of a “minimal combination”, i.e. an adjacent wheat and cotton strip, the row length density (total row length per m2, or simply the number of rows per m, expressed over the whole intercropped area) and homogenized densities, i.e. densities of either crop, expressed per unit total intercrop area (Table 2).

Table 2: Characteristics of relay intercropping systems and monocultures (Exp. 1)

Row length density b (m m–2)

Homogenized density c

(# m–2) Cropping system

Total width (m) a

Wheat Cotton Wheat Cotton 3:1 1 3 1 524 ± 18d 4.9 ± 0.11 3:2 1.2 2.5 1.67 425 ± 26 7.5 ± 0.22 4:2 1.5 2.67 1.33 495 ± 27 6.3 ± 0.20 6:2 2.0 3 1 471 ± 22 4.7 ± 0.11 Sole wheat 0.2 5 - 725 ± 28 - Sole cotton 0.8 - 1.25 - 6.1 ± 0.16

a Total width refers to the minimum width of a combination of component crops (cf. Fig. 2 in Chapter 2).

b Row length density is the total row length of a component crop per unit area (m m–2 or rows m–1).

c Homogenized density is the number of ears per m2 for wheat averaged for 2003 and 2004 and number of plants per m2 for cotton averaged from 2002 to 2004;.

d Standard error.


85

The six treatments were arranged in four randomized blocks with a plot size of 180 m2. Wheat was sown on 4 November 2001, 2 November 2002 and 3 November 2003. Cotton was sown on 26 April 2002, 25 April 2003 and 25 April 2004. The wheat cultivar was ‘Zhongyu 5’ and cotton cultivars were middle maturing upland Bt cotton ‘Shiyuan 321’ in 2002 and the Verticillium-tolerant variety ‘CRI45’in 2003 and 2004.

Irrigation water was applied by flooding (350 mm) in 2002 and by drip application in 2003 (342 mm) and 2004 (182 mm). The total amount of nitrogen applied to wheat and cotton ranged from 302 to 412 kg N ha–1 y–1. Information on amount, composition and application dates of fertilizer materials is provided in Table 3. Fertilizers were applied evenly over the whole area of the plots, not discriminating between crop strips and paths.

Table 3: Type, amount and timing of nitrogen applied in Exp. 1

Nitrogen applied (kg N ha–1) Nitrogen source1 2001/2002 2002/2003 2003/2004

Organic material Natural dried dung 20 (BS)2 - - Cottonseed cake 46 (BS) 57 (BS) - Subtotal 66 57 -

Fertilizers Wheat growing season

Compound fertilizer 53 (BS) 47 (BS) 100 (BS) Ammonium phosphate 50 (BS) - - Urea 30 (Dec 17) 77 (Mar 29) 111 (Feb 17)

Urea 26 (Feb 23) Subtotal 159 124 211 Cotton growing season

Ammonium phosphate 20 (June 17) - - Urea 77 (June 17) 52 (June 20) 51 (June 12)

Urea 90 (July 22) 69 (July 14) 96 (July 8) Subtotal 187 121 147 Total organic and fertilizer N 412 302 358 1 Nutrient contents are: 1.1 % N, 0.3 % P, 1.2 % K for dried dung, 4.3 % N, 0.5 % P, 0.8 %

K for cotton seed cake, 12 % N, 28 % P2O5, 15 % K2O for compound fertilizer, 18 % N, 46 % P2O5 for ammonium phosphate, and 46 % N for urea.

2 Dates in brackets are dates of application; BS indicates before sowing

Chapter 5

86

Experiment 2: Response of the intercropping system to nitrogen input

The response of the 3:2 relay intercropping system to three nitrogen doses was investigated in 2002/03 and 2003/04. Cultivars and planting dates were the same as in Expt 1, but the minimum combination width was 20 cm wider. The amount of irrigation water was 177 mm over the whole wheat and cotton growing period. As in Expt 1, fertilizer doses were applied evenly on strips with and without a crop. Nitrogen was applied at full dose (180 kg ha–1), at half dose, and at 150% dose. Before wheat was sown in 2002, 57 kg N ha–1 was applied as cotton seed cake and 47 kg N ha–1 as compound fertilizer; no base fertilizer was applied in 2003. During the growing season, urea (46% N) was applied on wheat and cotton by drip irrigation according to the designated dose level (50, 100 or 150%).

Measurements

To determine crop yields, plants were harvested at maturity in each plot; the sampling area covered 5 m row length by 2 m width in the 3:1 system, 5 m length by 2.4 m width in the 3:2 system, 3.5 m row length by 3 m width in the 4:2 system and 5 m row length by 2 m width in the 6:2 system.

To assess above-ground dry mass (DM) and nitrogen uptake, 1 m row of each plot was sampled once per two weeks. First the number of plants was counted; next a sub-sample was selected for analysis. The subsamples consisted of twenty randomly selected wheat plants, and – in the case of cotton – of ten seedlings, or, later on, three plants.

The leaf area of the subsampled cotton plants was determined by measuring length and width of individual leaves, and using the formula (Chapter 4):

leaf area = 0.810 × length × width

The samples were subsequently oven-dried at 65 °C to determine dry matter weight.

Nitrogen content of plant and soil samples was determined by the Micro-Kjeldahl method (Ogg, 1960). Leaves and stems were weighted and analysed separately in 2002. In 2003 and 2004, analyses were made using the total vegetative above-ground biomass. N-content of reproductive organs (i.e. the spike of wheat and the squares, flowers and bolls of cotton) was determined separately.

Soil samples were taken before wheat sowing in October 2001, 2002 and 2003, taking five samples per plot with a 3 cm diameter auger. Soil samples from 0-30 cm depth and from 30-60 cm depth was analysed separately.


87

Data analysis

Nitrogen use efficiency

Internal efficiency of nitrogen use (IE) is defined as:

IE ii

i

YN

= (1)

Where Yi is the yield of crop i (g harvestable product m–2) and Ni is nitrogen uptake (g total N-uptake m–2) by crop i.

Relative nitrogen yield

The relative nitrogen uptake is calculated as:

, ,

, ,

RNT RNW RNC W I C I

W S C S

N NN N

= + = + (2)

In this equation RNW and RNC are the relative values for wheat and cotton N yield, respectively. NW,I, NW,S, NC,I and NC,S represent nitrogen uptake of intercropped and sole wheat, and of intercropped and sole cotton, respectively.

Nitrogen dilution effect

The N dilution curve is described using a power function (Flenet et al., 2006):

% cbcN a W= (3)

where N% is nitrogen concentration, W is above-ground dry mass, ac is the N percentage at W=1 g m–2 and bc is a curvature parameter.

Statistical analysis

Data on grain and lint yield, harvest index, nitrogen content, and total nitrogen uptake were analysed by ANOVA in SPSS 11.0, using a randomized block design with cropping system as fixed effect. Least significant differences (LSD) were used to separate treatment means (P<0.05).

Chapter 5

88

RESULTS

Crop yields and total N-uptake

Crop yields, harvest indices and total nitrogen uptake by the intercropped wheat and cotton, compared to monocultures, are shown in Table 4. The nitrogen yield of wheat in the intercropping systems ranged from 130 to 205 kg ha–1, which was significantly lower (P<0.01) than the 3-years average N yield of sole wheat (253 kg ha–1). Intercropping reduced the nitrogen yield of wheat proportionally to grain yield, reflecting mainly the differences in homogenized plant densities. The reduction in nitrogen yield of wheat by intercropping was greatest in the 3:2 and 4:2 systems, and smallest in the 3:1 system. The response of the 6:2 system varied between years.

The effect of intercropping on nitrogen uptake of cotton differed between years. In 2002, the nitrogen yield of cotton in intercrops ranged from 120 to 140 kg ha–1 compared to an N yield in the monoculture of 127 kg ha–1, with no significant difference (P>0.05) between the monocrop and intercrops. However, in 2003 and 2004, cotton N yields in intercrops were significantly lower than in monocrops, except for the 3:2 system. Especially, the total N uptake of cotton in the 6:2 system was extremely low (Table 4), which was caused by an incomplete canopy closure and low biomass and lint yield.

The relative nitrogen yield of cotton (RNC) in intercropping systems was much higher than the relative lint yield (RYC), indicating that the lint yield was much more diminished by intercropping than nitrogen uptake (Fig. 1). RNC values did not differ significantly among the intercropping patterns 3:2, 4:2 and 3:1; but the value was significantly lower for the 6:2 system (P<0.05). For wheat there were hardly any differences between the relative nitrogen yield of wheat (RNW) and the relative grain yield of wheat (RYW) (Fig. 1). Both parameters, RNW and RYW, were closely associated with row length densities (total row length of one crop species per m2, averaged over the whole field, m m–2). The values were somewhat higher in the 3:1 and 6:2 systems and lower in the 3:2 system.

N-use efficiency of wheat and cotton

The physiological nitrogen use efficiencies (IE) of wheat varied from 24.4 to 30.9 kg grain per kg N uptake for intercrops and from 25.5 to 27.7 kg kg–1 for the monocrop. In 2002 and 2003, the IE of wheat was not significantly affected by intercropping (Table 4), but in 2004, wheat in the 3:2 system showed a significantly higher IE-value than in the other systems including the monocrop.


89

Table 4: Homogenized yield, harvest index (HI), nitrogen yield1 and internal use efficiency (IE) for three cropping cycles (Exp. 1)

Wheat Cotton

Grain HI N yield IE Lint HI N yield IE

Cropping pattern

g m–2 g g–1 g m–2 g g N–1 g m–2 g g–1 g m–2 g g N–1

- 2001/2002 - 3:1 551.5 a 0.45 a 19.9 a 27.9 a 60.4 a 0.12 a 13.7 a 4.5 a 3:2 500.1 b 0.45 a 17.8 a 28.0 a 66.0 a 0.14 ab 14.0 a 5.5 a 4:2 475.8 b 0.43 a 18.6 a 26.1 a 77.4 a 0.13 a 13.2 a 6.1 a 6:2 514.4 ab 0.44 a 19.7 a 26.2 a 60.9 a 0.12 a 12.0 a 5.1 a Monoculture 760.7 c 0.39 a 28.8 b 27.7 a 115.2 b 0.20 b 12.7 a 9.7 b SE 15.0 0.02 0.9 1.7 6.1 0.02 2.0 1.0

- 2002/2003 - 3:1 415.6 a 0.40 a 17.1 ac 24.4 a 57.4 a 0.16 a 7.9 ab 7.3 a 3:2 362.0 b 0.43 a 13.0 b 28.0 a 67.0 a 0.17 a 9.7 ac 7.2 a 4:2 391.7 ab 0.47 a 14.7 bc 26.8 a 58.1 a 0.15 a 10.1 ac 5.8 a 6:2 395.3 ab 0.46 a 16.2 c 24.7 a 49.2 a 0.17 a 5.2 b 9.4 a Monoculture 520.8 c 0.45 a 20.3 d 25.7 a 93.3 b 0.20 a 11.0 c 9.4 a SE 11.4 0.03 0.7 1.5 7.5 0.03 0.9 1.5

- 2003/2004 -

3:1 584.8 a 0.45 ab 20.5 a 29.0 a 69.5 ab 0.16 a 10.0 a 7.3 a 3:2 513.3 b 0.46 a 17.0 b 30.9 b 90.2 a 0.18 a 12.2 b 7.4 ab 4:2 502.3 b 0.44 ab 18.8 ab 26.8 a 87.4 ab 0.22 a 8.5 a 10.5 b 6:2 515.8 b 0.49 a 17.4 ab 29.7 a 65.6 b 0.21 a 6.5 c 10.1 ab Monoculture 682.9 c 0.39 b 26.9 c 25.5 a 117.0 c 0.20 a 11.7 b 10.0 ab SE 14.5 0.02 1.0 1.6 7.4 0.02 0.5 1.0 1 N yield indicates the total nitrogen uptake of cotton at the open boll stage. a, b and c: a common letter in one column subdivision means no significant difference at LSD0.05.

The effect of intercropping on the IE of cotton varied between years. In 2002, it ranged from 4.5 to 6.1 kg lint per kg N uptake in intercropping systems, which was significantly (P<0.02) lower than in sole cotton (9.7 kg kg–1). Due to the large variation, IE did not significantly differ between systems in 2003 and 2004. Averaged over three years, the IE values amounted to 6.4, 6.7, 7.5, 8.2 and 9.7 for the 3:1, 3:2, 4:2, 6:2 systems and the monoculture, respectively. Lower IE’s of cotton in intercrops, compared to monoculture, were mainly due to lower DM accumulation and lower (15-30%) harvest indexes (HI).

Chapter 5

90

0

0.5

1

1.5

2

RN

T

0

0.5

1

1.5

2

RN

T

0

0.5

1

1.5

2

3:1 3:2 4:2 6:2


RN

T

CottonWheat

0

0.5

1

1.5

2

LER

0

0.5

1

1.5

2

LER

0

0.5

1

1.5

2

3:1 3:2 4:2 6:2


LER

a 2002 b 2002

c 2003 d 2003

e 2004 f 2004

Fig. 1: Relative N yield total (RNT=RNW+RNC; a, c and e) and relative agronomic yield (LER=RYW+RYC; b, d and f) of wheat and cotton in the intercropping systems, in three years experimentation. SE’s pertain to the contributions of wheat and cotton to the RNT and LER


91

Crop yields and total N-uptake under reduced input

In experiment 2 (2003/2004), the nitrogen fertilizer input to a 3:2 intercrop was varied between 38 and 150 kg ha–1 during the wheat phase and between 53 and 210 kg ha–1 during the cotton phase in intercropping systems. Grain yields of wheat were somewhat lower under the highest nitrogen application rate (210 kg ha–1) than with the other two N-inputs, due to a lower HI at higher N-input (Table 5). Lint yields of cotton differed strongly between the two seasons but, however, did not respond to the reduced N-rates. Despite the substantial difference in N-inputs between treatments, the uptake of nitrogen did not differ significantly (Table 5), indicating ample soil N-reserves. Estimated apparent N recoveries in 2002/2003 and 2003/2004 amounted to 115 and 141 kg ha–1 in the wheat crop, and 80 and 87 kg ha–1 in the cotton crop.

Table 5: Homogenized yield, above-ground dry mass (DM) and nitrogen uptake (Nu) under different N application rates (Na) in the 3:2 system (Exp. 2)

Wheat Cotton Na Yield1 DM Nu IE Na Yield2 DM Nu IE g m–2 g m–2 g m–2 g m– g g N–1 g m–2 g m–2 g m–2 g m–2 g g N–1

- 2002/2003 - 14.1 350 730 13.9 26.0 5.3 61.0 417 7.9 7.6 17.8 369 724 14.8 25.0 10.5 62.2 399 8.1 7.7 25.3 333 747 13.7 24.7 21.0 51.2 363 7.9 6.5 SE 6 35 1.4 2.5 - 7.1 21 0.5 0.8

- 2003/2004 - 3.8 369 693 12.0 32.7 5.3 99.6 567 9.1 11.2 7.5 385 713 10.7 36.0 10.5 96.9 491 9.2 10.5

15.0 375 772 11.9 31.5 21.0 98.2 454 7.7 12.9 SE 28 33 1.2 3.5 - 4.3 48 0.7 0.9

1 Yield of wheat is based on grain weight with 12% water content. 2 Yield of cotton is lint dry weight. The experiment was irrigated and fertilized by drip irrigation after June 2003; the total amount of water including rainfall was 717 mm in 2002/03 and 693 mm in 2003/04.

Cotton nitrogen dynamics

During the early growth stages, nitrogen uptake was slower in intercropped cotton than in the monocrop (Fig. 2), but uptake caught up in intercrops from 80 days after sowing

Chapter 5

92

(DAS). N-uptake levelled off from 110 DAS. In all the three years, the rate of N-uptake was most fast and total N-uptake reached the highest level in the 3:2 system and both were lowest in the 6:2 system. In each of the years, cotton in the 3:1 system had a low rate of N-uptake early on, but cotton in this system accumulated N at the fastest rate of all systems at the end of the season. Cotton in the 4:2 system showed similar patterns of N-uptake as cotton in the 3:2 system, but at a somewhat slower rate and lower final level, reflecting the lower density of cotton plants in this system, compared to the 3:2 system (Table 2). Likewise, the low level of nitrogen uptake in the 6:2 system can be explained by the lower canopy density and biomass yield in comparison to other systems (Tables 2 and 4).

Major influences of intercropping on the cumulative N-uptake by the cotton seedlings are shown at the end of the intercrop period, i.e. June 14 (49 DAS) in 2002, June 10 (46 DAS) in 2003 and June 17 (53 DAS) in 2004 (Fig. 3). All intercropping systems show large and significant differences in relative N-yield per plant with the monoculture, ranging from 0.18 to 0.48 over three years, 2002-2004. There are no significant differences between intercropping systems, except for the significantly lower N-accumulation in the 3:1 system compared to the other intercropping systems in 2003.

N accumulation in cotton at the end of the wheat growing period is a

02468

10121416

Nitr

ogen

upt

ake

(g m

-2)

02

46

81012

1416

Nitr

ogen

upt

ake

(g m

-2)

0

2

4

6

8

10

12

14

16

30 50 70 90 110 130 150

DAS

Nitr

ogen

upt

ake

(g m

-2)

3:1 3:24:2 6:2Monoculture

a 2002

b 2003

c 2004

Fig. 2: Nitrogen uptake by cotton in the intercropping systems and the monoculture in 2002-2004


93

reflection of the competitive effects of the wheat on the cotton, reflected in the N-content of the above-ground biomass (Fig. 4). The competitive effect is closely related to the width of the space between the wheat strips, which is narrowest in the 3:1 system (60 cm), wider in 3:2 (80 cm) and 4:2 (90 cm), and widest in 6:2 system (100 cm; cf. Fig. 2 in Chapter 2). Correspondingly, the N-content of cotton seedlings is the least in the 3:1 system, greater in 3:2, still greater in the 4:2 and 6:2 systems, and greatest in the cotton monoculture (Fig. 4). Averaged over three years the N content of cotton seedlings in intercrops ranged from about 3.2 to 3.4%, which is significantly lower (P<0.01) than in monoculture (3.6%). Cotton had a significantly lower N-content in the 3:1 system than in any other intercropping system (P<0.05). Thus wheat exhibited significant effects on N-uptake of cotton seedlings during intercropping.

Nitrogen dilution curves for wheat and cotton

Nitrogen dilution curves of wheat and cotton are shown in Fig. 5a and 5b. Wheat showed an ongoing decline in leaf nitrogen content with increase in total dry weight. The dilution was less rapid in the monocrop of wheat than in intercrops, possibly resulting from a lesser dry weight increase as a result of a stronger competition for light compared to wheat grown in intercrop strips, where there is sideways incidence of light. Leaf nitrogen content in cotton decreased rapidly while the crop progressed from the seedling stage to an above-ground dry mass of 200 g m–2, but it remained relatively constant from then on. The level at which N

0

0.1

0.2

0.3

0.4

0.5

0.6

2002 2003 2004Year

Rela

tive

N yi

eld

per

plan

t

3:13:24:26:2

Fig. 3: Relative nitrogen yield of cotton per plant in the intercropping systems at the time of wheat harvest in 2002-2004

3

3.2

3.4

3.6

3.8

4

3:1 3:2 4:2 6:2 MonoCropping pattern

Nitr

ogen

con

tent

(%)

Fig. 4: Nitrogen content of above-ground cotton biomass in the intercropping systems and the monoculture at the time of the wheat harvest in 2002-2004

Chapter 5

94

content levelled off was higher in the cotton than in the wheat, and the dilution was somewhat less in the cotton plants of the 3:2 system than of the other systems.

0

1

2

3

4

5

6

0 500 1000 1500 2000

Above-ground dry mass (g m-2)

Nitr

ogen

con

tent

(%)


0 200 400 600 800 1000

a by=32.6x-0.43 R2=0.88 (Intercrops)

y=55.0x-0.48 R2=0.94 (Monoculture)

y=4.1x-0.07 R2=0.76 (3:1)

y=3.7x-0.05 R2=0.66 (3:2)

y=4.7x-0.12 R2=0.81 (4:2)

y=4.3x-0.09 R2=0.76 (6:2)

y=5.7x-0.15 R2=0.85 (Monoculture)

Fig. 5: Relationship between above-ground dry mass and nitrogen content for wheat (a) and cotton (b) in the intercrops and the monocultures in 2002

Trends in specific leaf nitrogen content

The seasonal dynamics in crop nitrogen content of intercropped and monocrop cotton were reflected in the specific leaf nitrogen content (SLN); three phases are distinguished (Fig. 6). The first phase of cotton seedling growth; because of retarded growth, SLN at 50 DAS (time of wheat harvest) was significantly lower in intercrops than in sole cotton. During the second phase, from 50 to 75 DAS, the SLN of the intercrops recovered, exceeding the SLN of the monoculture. During the third phase, from DAS 75 to cotton harvest, the SLN of intercrops stayed higher than

0

0.5

1

1.5

2

2.5

3

40 50 60 70 80 90 100DAS

SLN

(g m

-2)

3:13:24:26:2Monoculture

Fig. 6: Specific leaf nitrogen content (SLN) of cotton before anthesis in the intercropping systems and the monoculture in 2002


95

in the monocrop, because more nitrogen was accumulated in the vegetative organs of intercropped cotton and crop senescence was delayed. The delayed senescence of leaves yielded some increase of dry mass, but did not compensate for the delay in growth during the early stages. These findings show that intercropped cotton accumulates more N per unit of biomass than sole cotton.


Nitrogen use at the crop level

N yield of wheat ranged from 130 to 205 kg ha–1 in intercrops, which is 62 to 84% of the monoculture. The total N uptake of wheat ranged from 203 to 288 kg ha–1 in the monoculture which corresponds to the reported 257 kg ha–1 N uptake of wheat at about 8 t ha–1 grain yield, averaged over five regions in China (Liu et al., 2006).

The values for total N uptake of sole cotton, ranging from 110 to 127 kg ha–1 in 2002, 2003 and 2004, were within the range (67-403 kg ha–1) reported by Rochester (2007), but lower than found by Ishaq et al. (2001) and Sainju et al. (2006). Total N uptake did not differ from the monoculture in any intercropping system in 2002; however, in 2003 similar N uptake as in monoculture was only found in intercropping systems 3:2 and 4:2 and a lower uptake in the 3:1 and 6:2 systems. In intercropping system 6:2 the crop failed to close the canopy and therefore light interception and crop growth were reduced; as a consequence the N yield in 2003 and 2004 was significantly less (48%) than in the monoculture.

The different responses between seasons and within intercropping systems are strongly associated with the degree of competition during the cotton seedling stage and with canopy development after the wheat harvest when the full space can be used by the cotton crop. The nitrogen uptake of intercropped cotton was only 15% to 45% of the sole cotton at the time of the wheat harvest. After the wheat harvest, the nitrogen uptake in the intercropping systems recovered. These findings suggest that during the reproductive phase, the delayed fruiting and boll formation resulted in a weaker sink, and therefore more assimilates were retained in the vegetative parts. As a consequence the leaves could stay green for a longer period of time.

The physiological nitrogen use efficiency (IE) of wheat was not much affected by intercropping, but the N and grain yield per unit homogenized land area were strongly associated with the plant densities. Internal efficiencies (IE) of wheat in intercrops and the monocrop were similar except in the 3:2 system that showed higher values; IE of

Chapter 5

96

wheat ranged from 26.5 to 29.0 kg kg–1 on average over three years. IE was higher in 2004 due to a higher harvest index (HI). A higher HI in intercrops may be explained from a border row effect as a result of border plants receiving more light. Liu et al. (2006) reported that IE of wheat ranged from 19.8 to 66.4 kg kg–1 in China based on estimates by the QUEFTS model. Our results were within the range of 25.5 to 30.5 kg kg–1.

Based on lint yields IE of cotton ranged from 6.4 to 8.2 kg lint per kg N uptake in intercrops in three years, which was significantly lower (P<0.05) than in monoculture (9.7 kg kg–1) except in the 6:2 system. Compared to the monoculture, the lower IE of intercropped cotton was due to a lower HI (18%) as a result of the delay in growth and development during the early growth stages (Chapter 2 and 3). The variation in IE of intercropped cotton under various environmental condition (years) indicates that genotype × environment interactions may play an important role. Thus, a more detailed analysis of the N dynamics of various genotypes of cotton under contrasting environmental conditions is needed.

Nitrogen dynamics at the plant level

Plant N uptake is co-regulated by soil N supply and shoot growth, as concluded by Lemaire et al. (2007) based on studies under different environments (temperate and subtropical) with various crop species (C3 and C4). The relationship between N uptake and biomass accumulation reflects the feed-back regulation of N absorption capacity of roots by shoot growth itself under non-limiting N supply. During the intercropping period, the nitrogen content of cotton seedlings was significantly lower than in the monoculture. We conclude that intercropping decreased N uptake of cotton seedlings strongly during the intercropping period. Cotton plants compensate with extra N uptake at later growth stages for a delay in development as a consequence of shading in the seedling phase.

The coefficient bc of the N dilution curve in sole wheat was 0.48, which was consistent with the study of Flenet et al. (2006). The N demand per unit biomass of cotton in intercrops was greater than in the monoculture. The specific leaf nitrogen content (SLN) of cotton in intercrops ranged from 2.0 to 2.3 g m–2. The relationship between SLN and radiation use efficiency (RUE) for cotton is consistent with other species (Milroy and Bange, 2003). In cotton a linear increase of RUE was found for leaf N contents ranging from 2 to 5% (Sadras, 1996). Milroy and Bange (2003) reported also that leaf assimilate rates were about 3 times higher at SLN-values within a range from 2.78 to 4.32 g m–2 compared to 1.45 g m–2. This corresponds with the findings of


97

Reddall et al. (2004); they reported that leaf assimilate rates rapidly increased within a range from 0.9 to 2.5 g m–2. Thus the lower SLN values in the 3:1 system are likely to have resulted in a reduced light use efficiency. The recovery of the nitrogen content in cotton biomass after the intercropping period, is in good agreement with compensatory effects documented for cotton after loss of reproductive organs (Sadras, 1995). Our results confirm the compensatory effect for cotton with a delay in development in wheat-cotton intercropping systems.

Relative N yields and N balance at the system level

We found that the relative nitrogen yield total (RNT) in the intercropping systems ranged from 1.4 to 1.7; only the RNT of the 6:2 system was significantly lower. Compared to the land equivalent ratios (LER), which ranges on average from 1.3 to 1.4 in intercropping systems (Chapter 2), the RNT is 8% to 21% higher. It was concluded that cotton wheat relay intercropping systems utilize N less efficiently than the monocrops at the system level.

In our experiment the N-doses applied in the intercropping systems (range: 302-412 kg ha–1) and in the monocultures (for wheat: 124-211 kg ha–1 and for cotton 121-187 kg ha–1) were less than in farmers’ practice. The conventional N-fertilizer doses in farmers’ practice in the Yellow River region amounted to 375 kg ha–1 in wheat and 360 kg ha–1 in cotton (Zhen et al., 2006). The soil indigenous N estimated in this study ranged from 115 to 141 kg ha–1 for wheat, which corresponds quite well with the 135 kg ha–1 reported for the same region (Liu et al., 2006). The N-excess (Table 6) derived from the N balance sheet – calculated as N input (applied plus soil indigenous N) minus N removal by wheat grain and seed cotton – ranged from 400 to 410 kg ha–1 in the intercropping systems, which was much higher than in sole cotton (147 kg ha–1) and wheat (205 kg ha–1). Thus, more N in the intercropping systems was prone to losses by leaching or other processes than in the sole cotton system. In both systems, excessive N is being applied in practice, resulting in unnecessarily high input costs and environmental pollution.

The total N content of the soil in a 60 cm soil profile increased from 0.82-0.89 g kg–1 in November 2001 (the beginning of experiment 1) to 0.91-0.96 g kg–1 in December 2004 at the end of the 3-years experimental period. This corresponds with an average increase in total soil N of about 600 kg ha–1 over a period of three years. So, a substantial part of the excess N accumulated in the soil. The increase of soil N might be due to the full return of residues and immobilization of fertilizer N. We concluded that the wheat-cotton intercropping system enriched soil N-status more than a sole

Chapter 5

98

cotton or wheat crop. Based on these findings, the N management of the intercropping systems as well as of monocrops should be improved by means of proper timing and demand-oriented dosing of N fertilization. Simple and robust crop simulation models and GIS can help to improve the decision support to farmers in developing more profitable and sustainable cotton cropping systems (McKinion et al., 2001).

Table 6: N balance sheet including the N residues and N excess in the intercropping and monocropping systems in two years (Exp. 1)

N residue (g m–2)a N excess (g m–2)b Cropping system Wheat Cotton Total Wheat Cotton Total

- 2002/03- 3:1 8.9 ac 4.5 ab 13.4 a 24.0 a 16.8 ab 40.8 a 3:2 5.9 b 5.3 ab 11.2 a 25.1 b 15.7 a 40.8 a 4:2 7.0 ab 6.2 a 13.2 a 24.5 ab 16.2 ab 40.7 a 6:2 8.4 ac 2.4 b 10.8 a 24.4 ab 17.3 b 41.7 a Monoculture 10.0 c 4.9 ab - 21.9 c 14.0 c - SE 0.8 0.9 1.3 0.2 0.4 0.4

- 2003/04- 3:1 8.9 a 5.1 ab 14.0 a 21.1 a 18.6 a 39.7 a 3:2 6.9 a 6.0 b 12.9 ab 22.5 b 17.1 b 39.6 a

4:2 8.9 a 3.2 ac 12.1 ab 22.7 b 18.1 ab 40.8 b 6:2 7.3 a 1.7 c 9.0 b 22.4 b 18.6 a 41.0 b Monoculture 13.4 b 3.7 ac - 19.1 c 15.4 c - SE 1.0 0.7 1.3 0.3 0.4 0.3

a N residue＝N yield – N removal1 b N excess＝(N applied＋soil indigenous N2) minus N removal

1 N removal was derived from wheat grain and cotton seed yield and an estimated N content of 1.97% and 2.5% for wheat and cotton, respectively.

2 Soil indigenous N was estimated by extrapolation of the linear relationship between N applied and total N uptake.

CHAPTER 6

Development and validation of SUCROS-Cotton: A mechanistic crop growth simulation model for cotton, applied to Chinese

cropping conditions* * Agricultural Systems (2008), Submitted L. Zhang a,b,d, W. van der Werf b, W. Cao c, B. Li a and J.H.J. Spiertz b a Laboratory for Plant and Soil Interaction Processes, College of Natural Resources

and Environmental Sciences, China Agricultural University, Beijing 100094, P.R. China


c MOA Key Laboratory of Crop Growth Regulation, Nanjing Agricultural University, Nanjing, Jiangsu 210095, P.R. China

d Cotton Research Institute, Chinese Academy of Agricultural Sciences, Key Laboratory of Cotton Genetic Improvement, Ministry of Agriculture, Anyang, Henan 455004, P.R. China

Chapter 6

100

ABSTRACT

A mechanistic model for the development, growth and production of cotton (SUCROS-Cotton) was developed following the concepts of the Wageningen SUCROS models. Particular attention was given to the phenological development of the plant and the plasticity of fruit growth in response to temperature, day length, variety traits, and management influences such as plastic film cover and intercropping with wheat. The model is characterized by comparatively simple code and transparent algorithms, and it can be easily adapted to other abiotic conditions or cultivars. It was parameterized for Chinese cotton varieties and validated for cotton crops grown in monoculture and relay cultivation with wheat in two regions with contrasting climates: the Yellow River region and the continental Northwestern province Xinjiang. Model validation with seven years weather and crop growth data showed that the phenology, growth and yield are simulated satisfactorily; the root mean square error (RMSE) for date of emergence, date of flowering, date of open boll stage and duration from sowing to boll opening was less than 4 d both in the monoculture and wheat-cotton intercropping systems (less than 6.7% of the observed). The RMSE% over the observed total for dry matter was less than 6.6%, for lint yield 6.6%, and for number of harvestable bolls 10.0%.

The SUCROS-Cotton model provides a tool: (1) to assess production opportunities of cotton in various ecological zones in response to temperature, incoming radiation and management; (2) to identify optimal cotton ideotypes for different agro-ecological conditions and to guide breeding efforts; and (3) to explore resource-use-efficient cropping systems, including intercropping options, and crop management practices such as mulching and sowing date.

Keywords: Lint yield; growth; development; intercropping; physiological development time.

Development and validation of SUCROS-Cotton

101

INTRODUCTION

Simulation models of crop growth and production provide a widely accepted tool for assessing agricultural production opportunities in different agro-ecological zones in response to weather and management, for identifying ideotypes that are well adapted to certain agro-ecological conditions, for better understanding interactions between genotypes, environment and management (Kropff and Goudriaan, 1994; Yin et al., 2004), and for deriving optimal management strategies in adaptation to uncertain weather and changing climate (Meinke et al., 2001).

Cotton is a major cash crop in China, with a total production area of more than 4.8 million ha. There are three major cotton production regions: (i) the Yellow River basin; (ii) Xinjiang, and (iii) the Yangtze River basin (Hsu and Gale, 2001). Climatic conditions in these areas are quite different. The Xinjiang area is in the Northwest. Summers are short, hot and arid, and major production constraints are cold night temperatures in the spring and in the fall. Due to the short growing season, as defined by the occurrence of frosts, the area is only suitable for early and mid-early maturing varieties, and plastic film cover is used after seeding to increase temperature in the seedbed and accelerate early seedling development. All cotton in Xinjiang is irrigated. The Yangtze River basin is the most southern and warmest region of the three. Here, it is possible to grow a cotton crop and a wheat crop in sequence within a 12 month period, enabling two harvests per year. Irrigation is generally not needed because there is sufficient rainfall. The Yellow River basin valley has lower temperatures than the Yangtze River basin but it has warmer spring and autumn temperatures than Xinjiang. A large proportion (60%) of the cotton in this area is sown within the wheat crop several weeks before wheat harvest, which enables the cotton to accumulate enough temperature to complete its development before the first night frosts in autumn. To enable the young seedlings to grow in the taller wheat, the wheat is grown in strips of three to six rows, with spare paths interspersed for seeding the cotton (Chapter 2). Despite the reduction in wheat yield that results from the open space in the wheat, and the reduction in cotton yield that results from the shading by the wheat during the initial cotton development and the wide spacing between cotton rows in some systems, the aggregate yield in the relay intercropping system of cotton and wheat vastly exceeds the yields in cotton and wheat monocrops, as shown by land equivalent ratios of up to 1.39 (Chapter 2). Application of a plastic film cover is an effective method to increase temperature in the seedbed and accelerate seedling development, but this is little used by farmers in the Yellow River region. Plant densities vary widely between the three production areas; from 37,500 plants per ha in the Yangtze River basin, to

Chapter 6

102

60,000 plants/ha in the Yellow River basin and 225,000 plants per ha in Xinjiang. The low plant density in the Yangtze River basin is made possible by the vigorous growth of the cotton under the hot humid conditions of this region. The high plant density in Xinjiang compensates for the shortness of the growing season. Different varieties are used in the three regions.

Early maturing varieties (CRI16, CRI36) and mid-early varieties (CRI32, CRI35) are used in Xinjiang, and mid-early varieties (hybrid CRI29, 38, CRI41) or early varieties (CRI27) in the Yellow River basin and mid maturing varieties (Simian2) in the Yangtze River basin. In the Yellow River basin, the terminal growing point of the main stem is removed at the end of July (‘cut-out’) to curtail vegetative development and formation of additional flowers, and direct assimilates to the growing bolls, In the Yangtze River basin and Xinjiang, the time of ‘cut-out’ is in mid July. Most cotton varieties used in China are bred by the China Cotton Research Institute (CRI) in Anyang, province Henan, in the Yellow River basin.

Numerous models exist for cotton (Baker et al., 1983; Duncan, 1972; Hanan and Hearn, 2003; Hearn, 1994; Jallas et al., 2000; Jones et al., 1980; Lemmon and Chuk, 1997; McKinion et al., 1975; Mutsaers, 1984; Pan et al., 1997; Wall et al., 1994). These models differ in their modeling objective, modeling concepts, and domain of applicability. The prediction quality of some of these models has been extensively assessed in experiments to justify their application. For instance, GOSSYM has been validated and applied in many locations (Boone et al., 1993; Khorsandi and Whisler, 1996; McKinion and Baker, 1989; Reddy, 1988). It was applied to analyse the effects of weather on cotton yields (Reddy, 1990) and to evaluate economic strategies of growth regulator application for cotton production (Watkins et al., 1998). Based upon GOSSYM, the model COTGROW has been developed. This model is adapted to Chinese varieties, but it does not account for characteristic elements in Chinese cotton cultivation, such as relay intercropping with wheat or use of plastic film cover. Moreover, GOSSYM and COTGROW are based on complex and incompletely documented code. These models are therefore difficult to adjust and develop further for exploration of options for cotton production in China. A concise and transparent new model for cotton under Chinese growing conditions is needed to enable studies, relevant to China, on land use, ideotyping, cotton management, sustainability and consequences of global change.

The objectives of this chapter are: (1) to describe the conceptual structure of a model, SUCROS-Cotton, that meets the above-formulated needs; (2) to evaluate the model using extensive data sets on the phenology, growth and productivity of cotton under different combinations of genotype, environment, management and cropping system.


103

DESCRIPTION OF THE MODEL

General

The model is developed based on concepts of the SUCROS class of crop growth simulation models, which are characterized by concise and transparent code that has been extensively documented (Goudriaan and van Laar, 1994). SUCROS uses algorithms for calculating radiative climate, distinguishing direct and diffuse radiation.

The model simulates development and growth of cotton under the influence of temperature and radiation. The model consists of modules for phenology, photosynthesis, morphogenesis, fruit formation and abscission, dry matter partitioning, yield, and management practices. It uses the physiological development time concept (Soltani et al., 2006) to keep track of development. Compared to SUCROS (Goudriaan and van Laar, 1994), SUCROS-Cotton includes new algorithms for quantifying effects of film cover, cut-out, genetic traits, and it implements the boxcar train approach (Goudriaan and van Roermund, 1999) to simulate cotton fruit dynamics, abscission, single boll weight and yield. SUCROS-Cotton includes the concept of a supply-demand ratio (Gutierrez et al., 1984) to model fruit abscission and fruit filling. SUCROS-Cotton is programmed in the Fortran Simulation Translator (FST) (van Kraalingen et al., 2003). A list of symbols is provided in Table 1.

Model structure

The state variables in SUCROS-Cotton are Physiological Development Time (τ, days), Leaf Area Index (L; m2 leaf area m–2 surface area), total biomass (B; g m–2), carbohydrate reserve pool, and the biomasses of leaves, stems, fruits, roots and lint. Development is computed on the basis of temperature, photoperiod and a cultivar coefficient. Photosynthesis, dry matter partitioning, and LAI are calculated as in SUCROS (Goudriaan and van Laar, 1994). Potential growth of the fruits is simulated based on genetic traits (e.g. maximum weight of single boll) to calculate demand. The ratio of supply and demand for dry matter is computed to scale growth processes of fruits, which may be sink or source limited (Gutierrez et al., 1984). Fruit abscission is based on supply/demand ratio. Lint yield is calculated from fruit mass by using allometric relationships.

Chapter 6

104

Table 1: List of symbols in SUCROS-Cotton Acronym1 Symbol2 Units Meaning AGE a d Calendar age of a fruit BOLL Ai # Number of fruits in boxcar train i. IE c - the conversion factor from increment of soil

temperature to increment of air temperature DAYL D h Day length TDWW DM g m–2 or kg ha–1 Dry matter RET f(T) - Thermal effect TPLUS ΔT °C Increment of air temperature by increase of

soil temperature under film cover RPE g(D) - Photoperiod effect VI h(v) - Cultivar effect LAI L m2 leaf m–2 area Leaf area index PDT τ d Physiological development time, measured as

equivalent number of days with conditions allowing maximum rate of development

FALL1 R1 d–1 Fruit abscission rate caused by assimilate stress

FALL2 R2 d–1 Fruit abscission rate caused by pest injury FALL3 R3 d–1 Fruit abscission rate caused by extreme

temperature RBOLL rfruit d–1 Relative growth rate of a single fruit RWBOLL Pi g boll–1 d–1 Potential growth rate of fruits in boxcar train i STRBOL ρ kg kg–1 Supply demand ratio: DM resulting from net

photosynthesis per kg DM ‘demanded’ by sinks

WBOLL Bi g m–2 or kg ha–1 Total biomass of reproductive structures- fruits: squares, flowers and bolls in boxcar train i

WBAGE Wi g boll–1 Potential weight of a single fruit WBMAX Wmax g boll–1 Maximum weight of a single boll 1 Acronym in FST-listing (see Appendix II). 2 Symbol in text.

Simulation of development

The development rate depends on temperature, photoperiod and genetic traits (Cao and Moss, 1997; Robertson, 1968). The integral of the daily development increments, τ, is computed as:

( ) ( ) ( ) df T g D h V t= ⋅ ⋅ ⋅∫τ (1)

where the term f(T) represents the influence of temperature, g(D) represents the influence of day length, and h(V) indicates the influence of variety. f(T) and g(D) are time-varying functions, while h(V) is a constant for a given variety. Development is


105

completed when equals 60 days, being the minimum number of days needed by the earliest maturing cultivar (CRI127) to reach 50% open bolls under optimal conditions for development. Development stage is calculated by dividing by 60.

Influence of temperature

The temperature effect on cotton development, f(T), is calculated as:

0 if

if( )

if

0 if

b

bb o

o b

mo m

m o

m

T TT T T T TT T

f TT T T T TT T

T T

<⎧⎪ −⎪ ≤ <

−⎪⎪= ⎨ −⎪ ≤ ≤⎪ −⎪

>⎪⎩

(2)

where, T is daily average air temperature, except from sowing to emergence, when soil temperature is used. Tb is the biological base temperature, To is the optimum temperature and Tm is the maximum temperature for development. Tb was set to 12 °C, To to 30 °C and Tm to 35 °C (Anonymous, 1982; Baker et al., 1983; Zhao et al., 2005).

The physiological time concept should ideally operate on an hourly time scale. As the model is run with meteorological variables that are commonly measured at an agricultural meteorology station, i.e. daily minimum and maximum temperature, an approximation is applied. As suggested in Goudriaan and van Laar (1994), a daily value for f(T) is calculated by determining the function value at the daily minimum temperature, at the average temperature, and at the maximum temperature, and calculating a weighted average function value, according to respective weights of 0.25, 0.50 and 0.25.

Cotton radicles start to elongate when temperature in the top soil, where the seeds are, rises above 15 °C. Cotyledons start elongation when soil temperature exceeds 16 °C (Xu and Xu, 1989). The optimum range of temperatures for emergence is from 25 to 30 °C (Anonymous, 1982).

Influence of day length

Generally, cotton (Gossypium hirsutum L.) is considered to be daylength-neutral. The photoperiod effect is therefore not included in most cotton models. However, daylength sensitivity exists when large differences occur in latitude, as is the case in China. In SUCROS-Cotton, an empirical relationship (Goudriaan and van Laar, 1994) is used to describe the response of development rate to daylength:

Chapter 6

106

b

mb m

m b

max

1 if

( ) if

0 if

D DD Dg D D D DD D

D D

⎧ ≤⎪

−⎪= < ≤⎨ −⎪⎪ >⎩

(3)

where, D is actual day length, Db is base daylength below which development is most rapid, and Dm is maximum daylength, above which there is no development. The following parameter values are used: Db = 14 hours per day and Dm = 16 hours per day (Anonymous, 1982).

Influence of variety

The variety factor h(V) is calculated from observations as the ratio between the sum of f(T) at 50% open boll for a reference cultivar (i.c. CRI27) and the accumulation of f(T) for the variety under consideration:

( )( )

( ) reff Th V

f T= ∑∑ (4)

h(V) ranges from 0.75 to 1.00 in the available varieties (Table 2). A higher value means earlier maturation.

Some indicative values of τ for subsequent development stages are 2.5 at emergence, 17.5 at squaring, 27.5 at flowering, and 60 at open boll. The crop has reached a given stage if 50% of the plants have reached it.

Effect of plastic film on development

Temperature underneath a plastic film cover is higher than air temperature as measured in a Stevenson screen. Temperature above the soil underneath the film cover is modelled with an empirical relationship. A temperature increase, ΔT, to be added to the measured daily average air temperature, is calculated as follows:

( ) bsf s

s b

T TT c T TT T−

Δ = −− (5)

where Ts is soil temperature without mulching, which is obtained from a linear regression with air temperature (Ts = 0.890 + 1.017 × T, R2 = 0.975; Zhang et al., 2003a). Tsf is also obtained from a linear regression with air temperature: Tsf = 7.5725 + 0.8303 × T, R2 = 0.805 (Zhao et al., 1996). c is the increase in soil temperature resulting from a 1 °C increase in air temperature (Anonymous, 1988; Zhang et al., 2003a); this coefficient c is 0.51 from emergence to squaring, 0.22 from squaring to flowering, and is zero after flowering, because the film cover will be removed or is heavily shaded.


107

Morphogenesis

Cotton develops first a stem and two cotyledons; next monopodials will grow, the branches of which are initially dormant. Typically, there are four to nine monopodial nodes on the main stem depending on genetic earliness. Fruit nodes are located in the sympodial branches. Squares are formed at the fruit nodes. The rate at which new organs such as main stem leaves, fruit branches and fruit nodes are computed in the model is by the use of linear regressions against f(T), based on experimental data. The functions are variety independent. Plant height is modelled as a linear function of f(T) until growth is terminated by removal of the top bud (“cut-out”).

Leaf area

The calculation of intercepted PAR is based on the leaf area index (L). In the model, leaf area index is computed by source-limited growth (Goudriaan and van Laar, 1994); i.e. the availability of assimilates determines the formation and growth of new leaves. SUCROS-Cotton uses a Developmental-stage dependent specific leaf area, which is linearly interpolated between values of 16.4 m2 kg–1 DM at the 3-leaf stage, 22.0 m2

kg–1 DM at 50% squaring, and 13.6 m2 kg–1 DM at the 50% open boll, based on the observations (L. Zhang, unpublished data).

Photosynthesis

The calculation of canopy photosynthesis in SUCROS-Cotton uses the algorithms for SUCROS, described in Goudriaan and van Laar (1994) and Thornley and Johnson (1990). Five-point Gaussian integration is used to integrate the PAR interception over the vertical canopy profile. Daily PAR interception is calculated using 3-point Gaussian integration.

Respiration

Respiration is divided into two parts: growth respiration and maintenance respiration. SUCROS-Cotton uses algorithms for calculating plant respiration described by Goudriaan and van Laar (1994).

Dry matter partitioning

It is assumed that 70% of the daily produced assimilates are directly available for the growth while 30% are allocated to a storage pool. When daily photosynthesis does not satisfy the demand of maintenance and growth respiration, dry matter is reallocated from the storage pool. If the daily net photosynthesis rate is zero or negative, the

Chapter 6

108

relative consumption rate of the reserve pool is 0.2 d–1, otherwise it is 0.1 d–1 (Pan et al., 1997). The daily available dry matter therefore consists of one component from current photosynthesis and another component from the long term storage pool. Dry matter partitioning to root and shoot differs among development stages. The stage-dependent partitioning coefficients were based on values provided by van Heemst (1988).

Fruit development and abscission

Formation of reproductive organs proceeds from squaring to flowering, boll filling and boll opening. All development stages of the reproductive structure will be referred to as fruit. The fixed boxcar train methodology (Goudriaan and van Roermund, 1999) is used to simulate the dynamics of fruits development and abscission caused by assimilate stress, pest injury and extreme temperature. The boxcar train is essentially a population model that keeps track of advancement in development, and changes in numbers due to recruitment of new fruits and abscission (i.e. death) of existing fruits. Successive boxcars represent developmental stages, and the gradual accession of fruits in the boxcar train represents development. Seventy four boxcars are used to represent the changes in the developmental stage distribution of reproductive structure through the season as, across a wide range of genotypes, approximately 74 calendar days are required for the development of a fruit from just initiated square to open boll. Boxcars 1-22 represent squares (flower buds), boxcars 23-25 flowers, boxcars 26-33 small bolls, and boxcars 34-74 large bolls. The number of boxcars that is used for each of these stages represents relative ages of fruits. Bolls in boxcars 60-74 are mature enough to be harvested, although they are not yet open.

Two coupled boxcar train state variables are involved in the calculation of the number and weight of fruits. The first boxcar train represents the numbers in each development class. The second boxcar train represents the total biomass of the reproductive structures within each class. Fruit age is represented by boxcar number. The terminal outflows of two coupled boxcar trains are the numbers and total biomass of the opening bolls (Fig. 1).

Abscission

One of the factors driving abscission is the ratio between assimilate supply and demand (ρ). The supply demand ratio is calculated as the ratio between the supply from daily photosynthesis and reallocation from storage pool and the total demand for assimilates from all age classes of fruits. Reproductive structures of different age differ in their response to assimilate stress. Small fruits i.e squares (boxcars 1-22) and small


109

A 1-5 A 6-20 A 21-25 A 26-33 A 34-74 A outOpen boll

B 1-5 B 6-20 B 21-25 B 26-33 B 34-74 B out

Yield boll

weight of yield boll

P

R1

R3

R2

T

Rain

State variables Rate variables

Flow by which a state variable is change

Flow of information

Single boll weight

⍴

Auxiliary variables

Pestpopulation

Harvestable immature boll

(A 60-74)

Weight of harvestable

immature boll(B 60-74)

Input variables

A 1-5A 1-5 A 6-20A 6-20 A 21-25A 21-25 A 26-33A 26-33 A 34-74A 34-74 A outOpen boll

B 1-5B 1-5 B 6-20B 6-20 B 21-25B 21-25 B 26-33B 26-33 B 34-74B 34-74 B out

Yield boll

weight of yield boll

P

R1

R3

R2

T

Rain

State variables Rate variables

Flow by which a state variable is change

Flow of information

Single boll weight

⍴

Auxiliary variablesAuxiliary variables

Pestpopulation

Harvestable immature boll

(A 60-74)

Weight of harvestable

immature boll(B 60-74)

Input variablesInput variables

Fig. 1: Relational diagram of the simulation of cotton fruit development and fruit growth in SUCROS-Cotton. A 1-5 are the numbers of very small squares (flower buds), A 6-20 bigger squares, A 21-25 biggest squares and flowers, A 26-33 small bolls, A 34-74 larger bolls, A 60-74 harvestable immature bolls. B i-j are the total biomass of the fruits in the classes i-j in the array B. R1,R2 and R3 is rate of fruit abscission causing by assimilate stress, pest injury and extreme temperature, respectively. P is potential growth rate of fruits from a class to the next. ρ is ratio of supply and demand of assimilate. T is air temperature. B out is weight of open bolls bolls (boxcars 26-33), are shed when there is assimilate stress. The rate of abscission, R1 equals 0 when ρ=1 and linearly reaches 1 when ρ=0. Assimilate stress will cause losses of redundant fruits in order to reach maximum boll weight. Bolls that age are older than 8 days (boxcars 34-74) do not abscise.

Other factors causing abscission are excessively high temperature and pest injury. Because extreme high temperatures cause abscission, the fruit abscission rate from extreme temperature, R3 is zero below 32 °C and increases linearly to 1.0 d–1 at 45 °C.

Chapter 6

110

The rate of abscission from pest injury, R2 is estimated from 0 to 0.001 d–1 based on the pest population.

Calculating fruit number

Equations 6-9 are used to calculate the number dynamics of fruits in each boxcar. Equation 6 defines the developmental width of each boxcar:

Ng

=λ (6)

where, λ is developmental width of one boxcar, g is the developmental width of the whole boxcar train, and N is the total number of boxcars (N = 74). g is set to 70 d. Equation 7 defines Ci, the number of fruits per unit of development within each boxcar.

ii

ACλ

= (7)

where, Ai is the number of fruits per boxcar. Transfer of fruiting structures from one boxcar to the next, or out of the boxcar train is proportional to development rate, υ, taken equal to 1 d–1:

1i iQ C+ = ⋅υ and out NQ C= ⋅υ (8)

( ) ( )1 1 2 3dd

ii i i

A A A R R R At −= − − + +

υλ

(9)

Equation 9 represents the number dynamics in each boxcar, taking into account the development and mortality of fruits. The coefficients R1, R2 and R3 represent the effects of the three main causes of fruit abscission: assimilate stress, pest injury and extreme temperature, respectively.

Calculation of boll weight

The growth rate of fruit biomass in each boxcar is calculated as:

( ) ( )1 1 2 3dd

ii i i i i

B B B R R R B P At −= − − + + +

υ υρλ

(10)

where, B is the biomass (g m–2; dry weight) of all fruits within a certain boxcar, ρ is the supply/demand ratio, and Pi is the potential growth rate of individual fruits (g per fruit) in boxcar number i. The first term in the equation, υ/λ(Bi–1 – Bi), ascertains that developing bolls in the boxcar train take their weight along. The second term allows for fruit mortality. The third term is biomass increase of all the fruits in boxcar i under the influence of assimilate supply and demand, and the potential dry matter uptake for


111

growth, calculated for the specific developmental class i. The actual growth rate is computed as the product of fruit number, potential growth rate per fruit, and the supply/demand ratio. Up to boxcar number 33, if assimilate demand of the developing structure is stressed the cotton decreases number of fruit Ai to adjust, therefore, ρ for these boxcars is set to one, irrespective of the availability of assimilates. As the fruits grow bigger (boxcar 34 and higher), bolls will not be shed when assimilates are limited, but their growth rate is decreased by a proportion of ρ.

Demand for assimilates is based on potential fruit growth, derived from a logistic growth equation (Zhang et al., 2004b):

fruit

max

1 r a

WWk e− ⋅=

+ ⋅ (11)

where, Wmax is the maximum weight of a single fruit, as a genetic input parameter, varying from 7-8 g per single boll (Anonymous, 1982); rfruit is the relative rate of fruit growth, a is the age of the fruit (equivalent to boxcar number) in calendar days, and k is a parameter that can be derived from Wmax and W0; we found a value of 250. The relative growth rate of the fruits, rfruit is genotype independent, 0.114 d–1.

The potential dry matter requirement of all fruits in a class i for developing to the next boxcar, is then calculated as:

( )1i i iP W W+= − (12)

Calculating weight of boll shell

In the mature cotton boll, around 25% of dry matter is allocated to the shell, which does not contribute to the yield, around 45% to seed and around 30% to fiber. The absolute shell weight does not significantly vary among genotypes, but it does vary with age of the fruit and temperature. Parts of dry matter in the shell will be relocated to the seed during boll maturation if the temperature is greater than 15 °C (Zhang et al., 2003a). The dynamics of shell weight can be represented by a logistic growth curve within the first 30 days (boxcars 26-55) until it reaches a maximum; thereafter (boxcar 56-74), it shows a exponential decrease by a relative rate 0.0038 d–1 (Zhang et al., 2004b).

Chapter 6

112

Cotton relay intercropped with wheat

When cotton is relay intercropped with wheat, the air temperature experienced by the cotton seedlings is substantially lower than in the monoculture, especially on sunny days (Chapter 3). In SUCROS-Cotton, the measured air temperature (Chapter 3) in the canopy of cotton shaded by wheat during intercropping period was used to replace the air temperature from the weather station. After wheat harvest, the air temperature used is the same as in cotton monoculture, as there is no temperature effect in relay intercrops, as compared to cotton monoculture, after the harvest of wheat (Chapter 3). The light interception is computed by a model for homogeneous canopy, because sensitivity analysis showed that a homogeneous canopy model for intercropped cotton gave almost the same results as a row structured model, except for a small overestimation of light interception and production in the 6:2 system (Chapter 4).

MODEL PARAMETERIZATION

Parameters in SUCROS-Cotton are based on literature (Anonymous, 1982; Baker et al., 1983; Pan et al., 1997; van Heemst, 1988) and on field measurements and the national cotton variety tests conducted from 1998 to 2001 at the China Cotton Research Institute in Anyang, Henan, China (Yellow River region). The cultivars used included normal, not genetically modified varieties (e.g. CRI35), as well as Bt-varieties that were genetically modified for insect resistance using toxin genes from Bacillus thuringiensis (e.g. CRI38), and ‘double gene’ varieties (e.g. CRI41), with insect resistance based on two complementary genes conferring insect resistance: a Bt-gene and a gene coding for Cowpea Trypsinase Inhibitor, derived from cowpea, Vigna unguiculata (Li et al., 2001a). In total, 13 cultivars were used with classified genotypic differences in earliness. The four classes are:

1) early maturing; sown in summer, such as CRI27;

2) mid-early maturing; sown in spring, such as WM 11;

3) mid-early maturing hybrid; such as CRI38;

4) mid-maturing; such as Handan333.

Cultivar coefficients were derived from the national variety test; differences among genotypes are listed in Table 2. For cultivars whose earliness coefficient was not included in Table 2, h(V) was estimated based on similarity to test cultivars.


113

Table 2: Phenological development and cultivar coefficients for 13 cotton cultivars Development duration (days) Cultivar Year

S-E1 E-F2 F-O3 S-O4 h(V)a

Zhong142 (early) 1998 6 52 47 105 0.96 Yuzao472 (early) 1998 5 52 48 105 0.96 Lu458 (early) 1998 5 47 49 101 1.00 Han241 (early) 1998 6 49 48 103 0.98 CRI27 (early) 1998 6 46 48 100 1.00 ZKZ5 (Bt) 2001 10 63 56 129 0.83 LuH9513 (Bt) 2001 14 61 51 126 0.80 Jiza566(Bt) 2001 11 62 53 126 0.80 CRI38 (hybrid Bt) 2001 14 59 53 126 0.80 Nankang2 (Bt) 2001 12 60 50 122 0.85 WM11 2001 14 58 52 124 0.82 Jiwu538 2001 10 60 48 118 0.86 Handan333 2001 12 63 51 126 0.80 1 S-E: sowing to emergence; 2 E-F: emergence to flowering; 3 F-O: flowering to open boll; 4 S-O: total duration from sowing to open boll. a h(V) is cultivar coefficient.

MODEL VALIDATION

Data and method

Validation data were obtained from six experiments conducted from 1998 to 2004 near Anyang (36°07Ń and 116°22É) using various genotypes, management (e.g. film cover), sowing dates and cropping systems, and from experiments conducted in 1998 and 1999 near Arkesu, Xinjiang (40°22Ń and 80°04É), using film cover. These two locations are representative for the Yellow River and the Xinjiang cotton producing regions.

Experiment 1 was conducted near Anyang in 1998 and 2000. Data was collected from farmers’ fields. Early maturing cultivar CRI27 was used in 1998, mid-early hybrid CRI29 in 2000.

Chapter 6

114

Experiment 2 was conducted near Arkesu, Xinjiang. The crop was managed in accordance with common practice in this area; the cultivar was CRI23 in 1998 and CRI35 in 1999. Plastic film was used and plant density was about 225,000 plants/ha.

Experiment 3 was conducted near Anyang in 2001 and comprised four cotton varieties: mid-early hybrid CRI29, double gene GM mid-early cultivar CRI41, early maturing conventional CRI37, and mid-early maturing conventional CRI32. The experiment had three replicates, plot area was 6.3 × 14.5 m2, row spacing was 0.7 m, and plant density was 60,000 plants per ha.

Experiment 4 was conducted near Anyang in 2001, and comprised three management practices: (1) removing vegetative branches, performing ‘cut-out, and no plastic film cover; (2) removing vegetative braches, performing ‘cut-out, and with plastic film cover; (3) not removing vegetative branches, not performing ‘cut-out’ and no plastic film cover. Hybrid CRI29 was used. Plant density was 60,000 plants per ha.

Experiment 5 was conducted in Anyang in 2002 and comprised four sowing dates – 15 April, 5 May, 25 May, and 14 June. The treatments were carried out with three cultivars: a mid-early conventional cultivar CRI35, early conventional cultivar CRI36 and the double gene GM-cultivar CRI41. There were three replicates, plot area was 9.6 × 14.5 m2, and plant density was 60,000 plants per ha.

Experiment 6 was conducted near Anyang from 2002 to 2004 and its main focus was to compare productivity of cotton in monoculture to that in four wheat-cotton relay intercropping systems: 3:1, 3:2, 4:2 and 6:2 according to the number of wheat and cotton rows (Chapter 2). Row distance in sole cotton was 80 cm. Width of the wheat strip, as measured between the outer wheat rows was 100 cm in the 6:2 system, 60 cm in the 4:2 system, and 40 cm in the 3:2 and 3:1 systems. The interspersed space for sowing cotton was 100 cm in the 6:2 system, 90 cm in the 4:2 system, 80 cm in the 3:2 system, and 60 cm in the 3:1 system. Total width of one adjacent wheat and cotton strip was 200 cm in the 6:2 system, 150 cm in the 4:2 system, 120 cm in the 3:2 system and 100 cm in the 3:1 system. Experiment 6 had four replicates in randomized blocks with a plot size of 180 m2. Middle maturity Bt cotton ‘Shiyuan321’ (h(V)=0.85) was used in 2002 and ‘CRI45’ (h(V)=0.79) in 2003 and 2004.

Weather data of Anyang was obtained from the weather stations of the Cotton Research Institute (CRI), Chinese Academy of Agricultural Sciences (CAAS). Weather data of Arkesu was from national standard weather station.

The root mean square error RMSE was used for measuring the difference between observed and simulated results:


115

∑=

−×=n

iii SO

n 1

2)(1RMSE (13)

where Oi and Si are observed and simulated results respectively, and n is the number of independent samples from different years, experiments or treatments.

Validation of the simulation of development

Monoculture of cotton

The validation of development in sole cotton system is shown in Fig. 2. The RMSE of observed and simulated results, with large difference in ecological zones and years, was 1.1 d for sowing to emergence, 1.5 d for emergence to flowering (2.5% of the observed), 3.7 d for flowering to boll open (6.7% of the observed), and 4.2 d for the period from sowing to 50% open boll (3.3% of the observed).

Cotton in wheat-cotton intercropping systems

The delay in development of intercropped cotton, calculated by the model was 4.7 physiological days. Observations and simulations on the duration from sowing to flowering and from flowering to open boll in intercropped and monocropped cotton showed good agreement in all three years (Fig. 3). The RMSE for durations from sowing to flowering and flowering to open boll ranged from 2.1 to 2.9 d (3.0% to 3.4% of the observed). It is concluded that the model simulations of phenology are satisfactory, both in monoculture cotton and in intercrops.

Film cover

Film cover is widely used in Xinjiang to enhance early development. Fig. 4 presents the phenological development of cotton grown under film and non-film conditions in the Yellow River region and in Xinjiang. The results show that cotton with film cover emerged 4-5 days earlier, flowered 6-10 days earlier, and took around 7 days less to get to the 50% open boll stage. Effects of film cover differ between years and locations. There was a close agreement between simulations and observations (Fig. 4).

Validation of fruit growth and development

The simulation of fruit growth and development under different sowing dates was validated for the cultivars CRI41 and CRI35 (Fig. 5). The two cultivars are both mid-

Chapter 6

116

early and have the same value of h(V): 0.82. The RMSE of simulated and observed values for number of squares per plant was 3.36. The RMSE was 0.976 for the number of flowers, 1.57 for the number of young bolls, 1.57 for the number of big bolls (10.6% of the observed) and 1.04 for the number of open bolls per plant (10.0% of the observed). Dynamics are presented in Fig. 5 and show good correspondence between simulations and observations. The results show that the dynamics of fruit development is captured satisfactorily by the algorithms in SUCROS-Cotton.

Fig. 2: Observed and simulated durations of phenological stages (d, days) in Anyang in 1998 and 2000 (Expt 1), in 2001 (Expt 3) and in Arkesu, Xinjiang in 1998 and 1999 (Expt 2)

40

50

60

70

80

40 50 60 70 80

Observed (d)

Sim

ulat

ed (d

)

Anyang 2001 Anyang 2000 Anyang 1998 Arksu 1998 Arksu 1999

0

4

8

12

16

20

0 4 8 12 16 20

Sim

ulat

ed (d

)

40

50

60

70

80

40 50 60 70 80

90

100

110

120

130

140

150

160

90 100 110 120 130 140 150 160

Observed (d)

a Sowing to emergence

RMSE=1.1

Line: 1:1

b Emergence to flowering

RMSE=1.5

Line: 1:1

c Flowering to open boll

RMSE=3.7

Line: 1:1

d Sowing to open boll

RMSE=4.2

Line: 1:1


117

50

60

70

80

90

100

50 60 70 80 90 100Observed (d)

Sim

ulat

ed (d

)

3:1 3:2 4:2 6:2 Sole cotton

50 60 70 80 90 10050

60

70

80

90

100

50 60 70 80 90 100Observed (d)

Sim

ulat

ed (d

)

3:1 3:2 4:2 6:2 Sole cotton

50 60 70 80 90 100

a Sowing to flowering

RMSE=2.1

Line: 1:1

b Flowering to open boll

RMSE=2.9

Line: 1:1

Fig. 3: Observed and simulated durations of phenological stages (d) of cotton in monoculture and in three different relay intercropping systems with wheat, Anyang, 2002-2004 (Expt 6)

0

10

20

30

40

50

60

70

80

90

80 110 140 170 200 230 260 290

Julian day

PDT

Film cover (Observed) Film cover (Simulated)Bare soil (Observed) Bare soil (Simulated)

80 110 140 170 200 230 260 290

a

Arkesu, Xinjiang, 1998

Mid maturing cultivar CRI23

b

Anyang, Henan, 2001

Mid maturing hybrid CRI29

Fig. 4: Observed and simulated physiological development time (PDT, τ) with and without film cover conditions in (a) Arkesu, Xinjiang, 1998 (Expt 2), and in (b) Anyang, Yellow River region, 2001 (Expt 4)

Chapter 6

118

0

3

6

9

12

15

18

21

24

27

170 190 210 230 250 270 290Julian day

Num

ber o

f fru

it pe

r pla

nt

0

3

6

9

12

15

18

21

24

27

Num

ber o

f fru

it pe

r pla

nt

170 190 210 230 250 270 290

a b

c d

Fig. 5: Dynamics of number of fruits in cotton: squares and flowers ( , ), young bolls ( , ), big bolls ( , ) and open bolls ( , ) in cotton sown on: (a) 4 April, (b) 5 May, (c) 25 May and (d) 14 June 2002 near Anyang (Expt 5). Lines are simulated results and symbols observations. Cultivars CRI41 (open symbols, genetically modified insect resistant variety) and CRI35 (closed symbols, conventional variety) are similar in maturity

Validation of dry matter accumulation and yield

The simulation of dry matter accumulation was evaluated using results of experiments conducted in 2001 (Expt 3, comparison of four cultivars) and 2002 (Expt 5, comparison of four sowing date) in Anyang. The RMSE ranged from 371 to 444.5 kg ha–1 total dry matter, 4.3 to 6.2% of the observed. The RMSE of dry matter under three different forms of management (Expt 4), ranged from 283 to 447 kg ha–1, 4.7 to 6.6 % of the observed.


119

The simulation of yield, as affected by sowing date, showed a good agreement for three cultivars with a RMSE of 61 kg ha–1 for lint yield (6.6% of the observed) and a RMSE of 312 kg ha–1 for total open boll weight (9.8% of the observed; Fig. 6). The simulated results were higher than observed in most occasions (Fig. 6-a), indicating that simulation may be improved by accounting for growth limiting factors, notably water.

0

100

200

300

400

0 100 200 300 400

Observed weight (g m-2)

Sim

ulat

ed w

eigh

t (g

m-2

)

0

20

40

60

80

100

120

0 20 40 60 80 100 120 Observed yield (g m-2)

Sim

ulat

ed y

ield

(g m

-2)

CRI41CRI36CRI35

n=20 RMSE=31.2Line: 1:1

a Open bolls (shell and seed cotton)

b Lint

n=20 RMSE=6.1 Line: 1:1

Fig. 6: Comparison of simulated and observed weight of open bolls (a) and lint yield (b) in cotton sown at different dates near Anyang, 2002 (Expt 5). n indicates the number of samples

DISCUSSION

The results of the model validation for the years 1998 to 2004 at locations in Xinjiang and Yellow River region showed a good performance of SUCROS-Cotton. The highest observed error was less than 6.7% of observed values for phenological stages, 6.6% for dry matter, 6.6% for lint yield, and 10.0% for numbers of final yield fruit.

Currently available crop growth simulation models have shortcomings for application to questions related to cotton agriculture in China. First, the growing degree days (GDD) approach assesses the effect of supra-optimal temperatures on crop growth incorrectly because it assumes a linear relationship between development rate and temperature, even at high temperatures that can occur in China and are detrimental to the plant. In reality, development rate reaches a plateau, and it may even decrease at temperatures that are supra-optimal (Yin et al., 1995). Furthermore, photoperiod effects are largely not accounted for in the major cotton models.

Chapter 6

120

New features of SUCROS-Cotton, relative to existing cotton models, are: (1) physiological development time is used in stead of degree days, thus, the thermal, photoperiod and genetic effects are included; (2) Chinese agronomical practices such as film cover and cut-out are well simulated; (3) diffuse radiation depending on canopy layers is taken into account following the approach of SUCROS; (4) fruit growth and development are simulated by using the boxcar methodology with advantages in model transparency and conciseness; (5) the model code is concise and transparent, which makes adaptation straightforward. The conceptual structure allows parameterization under a wide range of conditions with respect to temperature, radiation, genotypes and cropping systems i.e. monoculture and intercropping systems.

SUCROS-Cotton is the first physiological process-based mechanistic cotton crop growth model that is applicable in China. It provides a tool (1) to assess production opportunities in various ecological zones in response to weather and water availability; (2) to derive optimal cotton ideotypes under different agro-ecological conditions to guide breeding efforts; and (3) to analyse crop management, e.g. effects of mulching and sowing date, or interactions with intercrops, such as wheat. SUCROS-Cotton can also be applied to estimate yield losses caused by pest injury and the magnitude of cotton compensatory growth. Using the model can add value and insight to field work by enabling a process-oriented interpretation of experimental findings, and by predicting ex ante the outcomes of experimental treatments. The model can be readily extended with algorithms for the uptake and utilization of water and nutrient (Bouman and van Laar, 2006) to broaden its application potential to a wider range of cropping systems and conditions.

CHAPTER 7

General discussion

Chapter 7

122

Research questions and scope of the study

The aim of this study is to answer general and specific research questions related to cotton and wheat relay strip intercropping. The following general questions are addressed:

Do intercropping systems show advantages in yield and profitability? Which intercropping systems are the most productive and resource use efficient?

What are the major characteristics and processes that determine the performance of the intercropping systems?

Do intercropping systems use resources efficiently and sustainably?

How can farmers benefit from a more quantitative understanding of the functioning of intercropping systems?

The main findings of this study include quantifications of productivity, quality, utilization of resources (e.g.: land, light and nitrogen) based on the results of field experiments and crop modelling. Crop growth and development of cotton in the monocropping and intercropping systems are quantified in relation to micrometeorological and agronomical factors. The prospects of improving cotton wheat relay-intercropping systems are explored by discussing the results of our study in a wider context of management options, crop improvement, sustainability issues and profitability.

Productivity and cropping arrangements

Which intercropping pattern is most productive and to what extent component crops gain or suffer in specific intercropping systems are questions of interest for agronomists, biologists as well as farmers (Connolly et al., 2001; Gibson et al., 1999; Jolliffe, 2000). In relay and strip intercropping systems, the land equivalence ratio (LER) is commonly used to assess the impact of one crop on the performance of the other and the combined yield of the various intercropping systems (Baumann et al., 2001; Ghosh, 2004; Wallace et al., 1992; Yadav and Yadav, 2001). Our study showed that the LER of the four tested wheat-cotton intercropping systems ranged from 1.28 to 1.39 (Chapter 2). Thus, all intercropping systems showed a clear advantage in land productivity based on crop yields. Among the four intercropping systems, the 3:1, 3:2 and 4:2 were the most productive systems with regard to crop yield; the 6:2 system was the least.

Intercropped cotton with a higher row length density (m row length per m2 total land area), e.g. the 3:2 and 4:2 systems, showed a faster and greater dry mass accumulation

General discussion

123

and lint yield than systems with a lower row length density, such as the 3:1 and 6:2 systems. Thus, an increase in plant density compensated to a large extent for the developmental and growth delays caused by competition for resources and modification of the thermal environment by the interaction with wheat during the seedling stage.

In intercropping systems the grain yield of wheat was 61% higher in border rows than in inner rows. This yield advantage resulted from more light capture (Chapter 4) and a better acquisition of nutrients by wheat (Chapter 5). This yield increase took place at the expense of the dry mass growth of cotton during the intercropping period (Chapter 2). This confirms the findings reported for strip intercropped soybean, corn and wheat systems (Iragavarapu and Randall, 1996; Li et al., 2001b).

Development delay and growth of intercropped cotton

In wheat-cotton intercropping systems, cotton is sown in April in the path assigned at wheat sowing in November of the previous year. The duration of the intercropping period is relatively short, only seven weeks, but a fully developed wheat canopy competes for light and nutrients with cotton seedlings. Therefore, the utilization of resources such as light interception, nitrogen and water uptake by intercropped cotton during this period is affected by the competitive strength of the wheat crop. As a result the development rate, canopy size, amount of light interception and total N uptake of intercropped cotton were decreased (Chapter 3 to 5).

Cotton development in the intercropping systems was delayed by 10-15 calendar days (Chapter 3). This delay corresponded with 4.7 physiological days (days with optimal temperature conditions) or 115 degree-days expressed as thermal time for the duration from sowing to the first square. The magnitude of the delay was the same in all tested intercrops (Chapter 3). The growth of cotton in the intercropping systems was correspondingly retarded by 6-12 days (Chapter 2). The developmental delay of cotton in wheat-cotton intercropping systems is long compared to the developmental delays reported for other intercropping systems (Bukovinszky et al., 2004; Gethi et al., 1993). It is associated with a 2.7 °C decrease of air temperature on a sunny day during the intercropping period, as a consequence of shading by wheat. The delay in development reduced the number of fruit branches, nodes and fruits before ‘cut out’ (removal of topmost buds of main stems), thus, decreasing harvest index and lint yield, which was also reported in literature (Lei and Gaff, 2003; Sadras, 1995)

Chapter 7

124

Resource capture and use efficiencies

Light

Light use efficiencies (LUE) of both wheat and cotton were not affected by intercropping. Wheat and cotton intercropping systems captured more light than monoculture of cotton. Compared to a cropping system of sole cotton, in relay intercropping systems a substantial amount of light (420 to 490 MJ m–2 from early spring to wheat harvest) is captured by the wheat crop before cotton plants emerge. From a spatial point of view, the wheat strips in the intercrops captured about 20% more light than the canopy of monocropped wheat per unit of strip area, thus compensating in part for the spare paths between the wheat strips (Chapter 2). Light capture by wheat strips was markedly influenced by row length density, reflecting the relative area planted with wheat. Narrow strips (60 cm) give greater light interception and higher yields than wider strips (120 cm) due to a greater number of border rows with an increased compensation ability.

The light intercepted by intercropped cotton ranged from 67% (6:2) to 93% (3:2) of the amount intercepted in the monoculture. Compared to monoculture, the amount of light intercepted by intercropped cotton was decreased by 12 MJ m–2 during the intercropping period and by 100 MJ m–2 during the period from wheat harvest to open boll stage in the 6:2 and 3:1 systems. Thus, the direct effect of shading was small compared to the indirect effect through a lower LAI (incomplete canopy closure). To increase light capture by cotton, the intercropping systems could be improved by alleviating the delay in cotton development by applying a film cover or by increasing plant densities to reduce ‘spatial’ losses in light interception.

The higher productivity of intercrops, compared to monocultures, can be fully explained by an increase in total light interception. It also shows that light interception and distribution can be modified by strip width and the number of crop rows per strip.

Nitrogen

Total nitrogen uptake of wheat in the intercropping strips was approximately 15% higher, expressed per unit strip area, than in monoculture. The increased nitrogen uptake is due to additional growth of wheat in the border rows, and the capture of extra nitrogen from the strips allocated to cotton.

The N uptake of cotton in the intercropping systems was reduced during the intercropping period. However, the loss of N uptake by intercropped cotton at the seedling stage recovered during the post-wheat period due to ‘compensatory growth’ of the vegetative plant parts resulting from the reduced fruit setting (Chapter 3). A

General discussion

125

similar compensatory growth was reported for loss of reproductive organs (Lei and Gaff, 2003; Sadras, 1995). The relative N uptake of cotton in the intercropping systems, compared to monoculture, was 8 to 21% higher than the relative lint yield, compared to monoculture (Fig. 1 in Chapter 5).

The total N uptake of both wheat and cotton can be explained by the relationship between the rate of N uptake and accumulated biomass. For wheat, no difference between intercrops and monocrops was found and the relationship did also not differ from most other C3 crops (Lemaire et al., 2007). The N dilution pattern of wheat corresponds with a study on C3 crops reported by Flenet et al. (2006). For cotton, the intercrop required more N than the monoculture from flowering onwards. Therefore, intercropped cotton showed a different trend; it accumulated more N in vegetative organs than cotton in monoculture.

Internal N use efficiencies (IE) of wheat did not differ between intercropping systems and monoculture. The IE of cotton was significantly (P<0.05) lower in intercrops than in monoculture (Chapter 5). The lower IE of intercropped cotton is due to a decrease of harvest index (HI). The cotton in the 3:2 and 4:2 intercrops showed the highest lint yield and therefore utilized nitrogen more efficiently than in other intercropping systems.

It was concluded that cotton intercropping systems utilize N less efficiently than cotton monocrops. As a consequence, more N will stay in crop residues and contribute to N-accumulation in the soil or N-losses. As suggested by Zhang et al. (2004a), more research is needed to improve the N management in the rhizosphere ecosystem to enhance nutrient use efficiency and crop productivity in intercropping.

Water

Due to the typical continental monsoon climate, the temporal distribution of annual rainfall in the North China Plain is extremely variable, with more than 80% concentrated in the cotton growing season (April to October). As a result, rainfall during the wheat growing season (November to June) ranges from 100 to 180 mm, which can only meet approximately 25 to 40% of the water requirement of a wheat crop. The evapotranspiration (ET) of wheat-cotton intercropping systems is more than 1200 mm per cycle in North China Plain (Zhang and Li, 1997). In traditional high yielding wheat cropping systems, wheat is usually irrigated four or five times during the growing season (Li et al., 2005) while cotton is irrigated only one or two times (Zhang and Li, 1997). Flood irrigation is applied with about 75 mm water each time. Consequently, more than 70% of irrigation water resources are used for winter wheat. Because most of the irrigation water comes from underground sources with wells

Chapter 7

126

deeper than 30 m, irrigation for wheat production is seriously threatening groundwater resources and the sustainability of agricultural production systems (Zhang et al., 2003b).

The water productivity (WP), kg grain or lint yield per unit delivered rainfall plus irrigation, ranged from 0.95 to 1.28 kg m–3 for wheat, and from 0.11 to 0.22 kg m–3 for cotton in the intercropping systems, 27% and 40% lower than the WP in the monocultures of wheat and cotton, respectively (Table 1). The lower WP of wheat in the intercropping systems was due to a decline in yield per unit of homogenized land area under ‘full’ irrigation. The lower WP of cotton in the intercropping systems was due to a lower biomass yield and a reduced harvest index.

Table 1: Crop yield, delivered water and water productivity (WP) in the intercropping systems and monoculture in 2002 to 2004

Wheat Cotton

Grain Water1 WP2 Lint Water WP

Year Cropping pattern

g m–2 mm kg m–3 g m–2 mm kg m–3

2002 Intercropping3 510 467 1.09 66 302 0.22

Monoculture 761 1.63 115 0.38

2003 Intercropping 391 413 0.95 58 526 0.11

Monoculture 521 1.26 93 0.18

2004 Intercropping 529 414 1.28 98 495 0.16

Monoculture 683 1.65 117 0.24 1 Water indicates total rainfall plus irrigation during the growing season. 2 WP is calculated as grain or lint yield per unit delivered water1. 3 indicates the data were averaged by four tested intercropping patterns.

Profitability and sustainability of wheat-cotton relay intercropping

Profitability

To evaluate the economic profitability of wheat-cotton intercropping systems, one needs to take into account the fluctuation of price ratio between lint and grain (L/G). The price per unit lint is at present about 10 times higher than per unit grain in China. The price of grain tends to increase because of current shortages on the international

General discussion

127

food, feed and biofuel market (Cassman and Liska, 2007). The gross income, which is the product of yield and price for the intercropping systems was not significantly (P>0.05) higher than for sole cotton in 2002 and 2003; but in 2004 it was significantly (P<0.01) higher in the 3:2 and 4:2 systems due to a lower price ratio in L/G (9.3) (Fig. 1 a, b and c). However, the gross incomes of the intercropping systems (averaged from 2002 to 2004, Henan province) were higher than of wheat/corn double cropping system (National Bureau of Statistic of China (NBC)). The intercropping system will gain more income than sole cotton and wheat/corn double cropping system if the L/G is between two points: shift for grain (SG) to shift for cotton (SC) (Fig. 1 d). The SG and SC are thresholds of L/G. The intercropping will be less profitable than wheat/corn when the L/G < SG and it will be less profitable than monocropping cotton when the L/G > SC. Compared to sole cotton and wheat/corn double cropping system, the gross income is higher for the wheat/corn system and sole cotton within a L/G range from 9 (SG) to 14 (SC). This profitable L/G window is much narrower for the 3:1 system than for 3:2. The 6:2 system shows hardly any advantage to increase farmer’s income. Zhang (2004) analysed the effect of L/G ratio on the shift from sole cotton to wheat/corn system in China by using Cobb-Douglas production model (Curve Expenditure System); this model is based on the behaviour of households as producer, consumer and labour force (Chambers and Quiggin, 2001). He found a L/G threshold value of 11.7 for a cotton price of 14 Yuan kg–1. It is concluded that the intercropping increases farmers’ income under a wide range of prices.

Sustainability The adoption of Bacillus thuringiensis (Bt) cotton cultivars have strongly contributed to a decrease in the use of pesticides and as a consequence increased the profitability, and the ecological safety of cotton production by smallholder farmers (Huang et al., 2002). In our study, we focused on the abiotic (nitrogen and water) aspects, because the effect of relay intercropping on pest incidence in cotton was studied by Xia (1997). Nitrogen The analysis of the N balance sheet (Chapter 5) showed that N was considerably more prone to losses in intercrops than in sole crops. In current farming practices the N dose in intercropping systems is quite high. Compared to the monoculture, the wheat-cotton intercropping system will likely enrich the soil fertility because of the extra N returned by crop residues. Therefore, first the N management of the intercropping system should be improved by means of proper timing and dosing of N applications. Chen et al. (2006) concluded that a large amount of basal N fertilization before sowing of

Chapter 7

128

0

5000

10000

15000

20000

25000

Gro

ss in

com

e (Y

uan

ha-1

)

0

5000

10000

15000

20000

25000

3:1 3:2 4:2 6:2 Wheat& corn

Solecotton

Cropping pattern

Gro

ss in

com

e (Y

uan

ha-1

)

Wheat Corn Cotton

4 6 8 10 12 14 16 18

Price ratio of L/G

6:23:13:2Sole cottonWheat & corn

a 2002 b 2003

c 2004 d

SG

SC

Fig. 1: Gross income of wheat-cotton intercropping systems, a monocrop of cotton and a wheat/maize double cropping system (a, b and c) and the relationship between the gross income and the price ratio of lint and grain (L/G) from 2002 to 2004 (d, yield was averaged over three years and lint price was fixed to 14 Yuan per kg). Arrows point to a shift for grain (SG) or a shift for cotton (SC)

wheat in the North China Plain was unnecessary, when soil mineral N in the 0–30 cm layer was higher than 30 kg N ha–1. Otherwise, the risk of N leaching and gaseous losses increases when the wheat crop is irrigated before winter. Second, a precise recommendation requires dedicated experiments with different cropping systems at the field and regional levels. Furthermore, the opportunity of using GIS integrated with decision-support based on crop simulations to derive site specific fertilizer recommendations can help to develop more profitable and sustainable wheat-cotton intercropping systems (Kersebaum et al., 2005).

General discussion

129

Water

The lower WP in the intercropping systems compared to the sole crop is a concern for the sustainability of these systems; water productivity needs to be enhanced. The opportunities are (a) site-specific (strip-specifically) irrigation based on actual requirements of each component crop by using drip or sub-drip irrigation technologies; (b) to develop intercropping systems with alternating ridges and furrows. Taking 3:2 system as an example, with ridges of 60 cm alternated with a furrow of 60 cm width; then, two rows of cotton can be sown on the ridges and 3 rows of wheat in the furrows. By irrigating wheat only in furrows, water can be saved (Li et al., 2007; Schneekloth et al., 2006); and (c) to reduce irrigation water volumes according to the level of local aquifers with respect to groundwater resources and to minimize yield loss from soil water deficits (Jalota et al., 2006; Singh and Singh, 1996; Xue et al., 2003).

Prospects for optimization of intercropping systems

Modifying the microclimate of cotton seedlings

Modification of the crop environment may have a significant impact on crop growth and yield. The most obvious modification of the environment in relay wheat-cotton intercropping is shading of cotton by the wheat crop; as a consequence, the reduced capture of light results in a reduced growth and yield of cotton. Shading also modifies other environmental conditions, e.g. air and soil temperatures (Midmore, 1993; Midmore et al., 1988). By environmental modification intercrops can even obtain higher yields than sole cropping (Midmore et al., 1988); for example, under hot conditions reduced soil temperature and moisture by intercropping may favor multiplication and growth of some soil micro-organisms (Singh et al., 1986). The environment can also be modified by other management practices, e.g. soil cover by film or straw (Rana et al., 2004). The productivity of cotton in the intercropping systems will be enhanced when the heat load can be increased during the early growth phase. The possible management practices to modify microclimate factors for the intercropping systems are:

(i) a film cover to increase soil temperature during early cotton growing stages. The results derived from simulation runs with the cotton crop growth model SUCROS-Cotton showed that the duration from sowing to flowering of cotton in intercropping system would have been 11 days shorter in 2002 under film cover.

Chapter 7

130

(ii) a ridge-furrow cultivation to decrease the shading of wheat. A ridge of 10 cm height for cotton in the intercropping systems, which has the same effect as decreasing plant height of wheat, will increase the light interception of cotton during intercropping period, and the air and soil temperatures can thus be increased.

(iii) seedling transplanting. The cotton is first sown in a seedling bed and then the seedlings with three true leaves are transplanted to the field. This management practice will shorten the duration of the intercropping period; thus, the intercropped cotton will have sufficient time to mature. The disadvantage of this practice is the extra labour requirement.

Genetic selection for ‘earliness’

An important aspect of an intercropping system is the extent of competition between the crops. The plant types of the intercropped species should be selected to complement each other (Davis and Woolley, 1993; Smith and Francis, 1986).

For wheat, the most important traits with respect to wheat-cotton intercropping systems are plant height and earliness. Shortening the height of the wheat canopy will decrease shading of cotton. Early maturing cultivars clear the field in time for the sowing of wheat; occasionally however, farmers need to pull out cotton plants to clean the field for wheat sowing at the expense of the yield and quality of cotton.

For cotton, the genetic traits related to maturity and resistance to diseases should be taken into account when breeding objectives are set to the intercropping systems. The intercropped cotton shows often a weak vigor under the shading environment, so seedling diseases may become a problem. Inadequate management of plant density can clearly be detrimental to the intercrop (Davis and Garcia, 1987). For example, the optimum density for cotton in monoculture is approximately 6 plants per m2 in the Yellow River region in China, but it is about 25% higher in 3:2 intercrop and 20% lower in 3:1 and 6:2 intercrops, depending on system arrangement. Thus, the cultivars with shorter fruit branches may be suitable for the 3:2 system and those with longer branches may be advantage in 3:1 and 6:2 systems.

The earliness of cultivars is most critical for intercropped cotton because of development and growth delay. The ultimate objective of breeding early maturing cultivars could be to select genotypes which can be sown directly after wheat is harvested. Then, a shift from relay intercropping to double cropping of wheat and cotton would be possible. However, this objective seems to be not feasible due to the

General discussion

131

lack of relevant genes and an unacceptable yield penalty (Bednarz and Nichols, 2005; Echekwu and Alabi, 1994; May and Bridges, 1995).

Use of crop models to simulate crop growth and development

Simulation models to study the behaviour of natural and agricultural ecosystems have been developed and used in the past three decades. They have been extremely helpful in integrating knowledge from various disciplines in one framework and helped to improve insight into complex ecosystems. Several models have been developed for the purpose of modelling interaction of intercropping system, such as INTERCOM (Kropff and Goudriaan, 1994) to simulate the crop and weed competition relations, and GAPS, an object-oriented dynamic simulation model for modelling plant competition (Rossiter and Riha, 1999). In most of the models, the competition of intercropped species for light, water, and nitrogen is based on the population of each component and its ecophysiological traits like plant height and relative growth rate.

For cotton, a wide range of models have been developed over the last 30 years in many countries. Some major models are GOSSYM in U.S. (Baker et al., 1983) and OZCOT in Australia (Hearn, 1994). COTGROW (Pan et al., 1997) is a good Chinese model based on GOSSYM. However those models are complex and can not be easily adapted to various varieties and cultivation practices, such as mulching and intercropping. SUCROS-Cotton, an ecophysiological model for cotton development, growth and production, was developed by Zhang et al. (Chapter 6) following the concepts of the SUCROS-class (Goudriaan and van Laar, 1994) of crop growth models and gave particular attention to the phenological development of the plant and the plasticity of fruit growth in response to temperatures which could be affected by covering plastic film and intercropping.

The validation of SUCROS-Cotton, after the model was parameterized by the findings of field experiments for the intercropping systems (Chapter 3), showed a good applicability for assessing phenology of cotton in intercropping systems. The RMSE was less than 2.9 d. It is concluded that the SUCROS-Cotton performed satisfactory to predict cotton development for monocrops as well as for intercropping systems. Thus, it can be used as a tool to optimize the effects of system design on temperature and phenology. Further improvements should focus on incorporating, the shading effect during the intercropping period and the heterogeneity of the canopy of cotton in intercropping systems by integrating the light interception model for strips (Chapter 4). Then the competition for resources can be systematically explored.

Chapter 7

132

Integration of knowledge on relay intercropping of wheat and cotton

The aggregate results presented in this thesis, lead to the following interpretation of growth processes in cotton-wheat intercropping: (i) dry matter accumulation is driven by light interception as determined by crop leaf area; (ii) the growth rate per unit of intercepted light is similar in all systems; (iii) wheat-cotton systems with a narrow path between the wheat strips results in a developmental delay in the cotton, probably as a result of lower temperature, that is not compensated for after wheat harvest, and which results in a delay and decrease in fruit set, fruit production and harvestable yield; (iv) intercropping systems with a high density of cotton, such as the 3:2 and 4:2 system, show a more rapid recovery and increase of LAI after wheat harvest than systems with a lower density of cotton; (v) there are no developmental effects of intercropping on the wheat; (vi) systems with a wide path between the wheat strips enable cotton to advance quickly in development and produce a high fruit biomass; these systems have a higher nitrogen use efficiency, because the nitrogen is more efficiently transformed into harvestable yield than in systems that result in a low fruit yield of cotton; (vii) the nitrogen use efficiency of wheat is unaffected by intercropping, as wheat is the first and dominant crop, and resource uptake and use efficiency in wheat are a pure response to density, and are not affected by interference from the second, submissive crop, cotton. Although not experimentally tested, it can be easily seen that a very wide path width for cotton would lead to a lower radiation interception, and lower yield.

The factors determining the performance of the cotton crop such as: a decreased light interception, a reduced fruit formation and N uptake were mainly caused by a combination of developmental delay and retarded growth.

The nitrogen requirement of cotton was increased by compensatory vegetative growth when the formation of fruit organs was reduced. It is concluded that the developmental delay of cotton was one of the most important factors determining productivity in wheat-cotton intercropping systems.

133

References

Aggarwal, P.K., Garrity, D.P., Liboon, S.P., Morris, R.A., 1992. Resource use and plant interactions in a rice-mungbean intercrop. Agronomy Journal 84, 71-78.

Ahmed, S., Rao, M.R., 1982. Performance of maize soybean Intercrop combination in the tropics - results of a multi-location study. Field Crops Research 5, 147-161.

Ǻngström, A., 1924. Solar and terrestrial radiation. Quart. J. Roy. Meteor. Soc. 50, 121-126.

Anonymous, 1982. Chinese cotton farming system and cultivation. Shanghai Sci-Tech press, Shanghai (in Chinese).

Anonymous, 1988. Chinese cotton cultivation by plastic film cover. Shandong Sci-Tech Press, Jinan (in Chinese).

Anonymous, 2001. Guidelines for HVI testing. Revised July 2001. United States Department of Agriculture, Agricultural Marketing Service, Cotton Program, 3275 Appling Road, Memphis, Tennessee 38133. http://www.am-s.usda.gov/cotton /pdf%20forms/HVIguidejuly2001.pdf (accessed on May 28, 2007)

Anonymous, 2004. Cotton classification; understanding the data. United States Department of Agriculture, Agricultural Marketing Service, Cotton Program. http://www.utexas.edu/conters/nfic/fc/documents/classing.pdf (accessed on May 28, 2007)

Baker, D.N., Lambert, J.R., MacKinion, J.M., 1983. GOSSYM: A simulator of cotton crop growth and yield. Clemson University, Clemson.

Bastiaans, L., Kropff, M.J., Goudriaan, J., van Laar, H.H., 2000. Design of weed management systems with a reduced reliance on herbicides poses new challenges and prerequisites for modelling crop-weed interactions. Field Crops Research 67, 161-179.

Baumann, D.T., Bastiaans, L., Kropff, M.J., 2001. Competition and crop performance in a leek-celery intercropping system. Crop Science 41, 764-774.

Bednarz, C.W., Nichols, R.L., 2005. Phenological and morphological components of cotton crop maturity. Crop Science 45, 1497-1503.

Bilalis, D., Sidiras, N., Economou, G., Vakali, C., 2003. Effect of different levels of wheat straw soil surface coverage on weed flora in Vicia faba crops. Journal of Agronomy and Crop Science 189, 233-241.

Blaise, D., Bonde, A.N., Chaudhary, R.S., 2005. Nutrient uptake and balance of cotton plus pigeonpea strip intercropping on rainfed vertisols of central India. Nutrient Cycling in Agroecosystems 73, 135-145.

Bonhomme, R., 2000. Bases and limits to using ‘degree-day’ units. European Journal

References

134

of Agronomy 13, 1-10. Booij, R., Kreuzer, A.D.H., Smit, A.L., van der Werf, A., 1996. Effect of nitrogen

availability on dry matter production, nitrogen uptake and light interception of Brussels sprouts and leeks. Netherlands Journal of Agricultural Science 44, 3-19.

Boone, M.Y.L., Porter, D.O., McKinion, J.M., 1993. Calibration of GOSSYM: Theory and practice. Computers and Electronics in Agriculture 9, 193-203.

Bouman, B.A.M., van Laar, H.H., 2006. Description and evaluation of the rice growth model ORYZA2000 under nitrogen-limited conditions. Agricultural Systems 87, 249-273.

Braden, C.A., Smith, C.W., 2004. Phenology measurements and fiber associations of near-long staple upland cotton. Crop Science 44, 2032-2037.

Bukovinszky, T., Trefas, H., van Lenteren, J.C., Vet, L.E.M., Fremont, J., 2004. Plant competition in pest-suppressive intercropping systems complicates evaluation of herbivore responses. Agriculture Ecosystems and Environment 102, 185-196.

Butterworth, J., Wu, X., 2004. China, Peoples Republic of cotton and products annual 2004. Gain Report Number: CH4013. USDA Foreign Agricultural Service, Beijing, China.

Cai, Y.L., Smit, B., 1994. Sustainability in Chinese agriculture – challenge and hope. Agriculture Ecosystems and Environment 49, 279-288.

Campbell, G.S., Norman, J.M., 1998. An introduction to environmental biophysics. Springer-Verlag New York, Inc., New York.

Cao, W., Moss, D.N., 1997. Modelling phasic development in wheat: A conceptual integration of physiological components. Journal of Agricultural Science 129, 163-172.

Cassman, K.G., Liska, A.J., 2007. Food and fuel for all; realistic or foolish. Biofuels, Bioproducts and Biorefining 1, 18-23.

Chambers, R.G., Quiggin, J., 2001. Decomposing input adjustments under price and production uncertainty. American Journal of Agricultural Economics 83, 20-34.

Chen, X., Zhang, F., Romheld, V., Horlacher, D., Schulz, R., Boning-Zilkens, M., Wang, P., Claupein, W., 2006. Synchronizing N supply from soil and fertilizer and N demand of winter wheat by an improved Nmin method. Nutrient Cycling in Agroecosystems 74, 91-98.

Cheng, X., Han, C.R., Taylor, D.C., 1992. Sustainable agricultural development in China. World Development 20, 1127-1144.

Connolly, J., Goma, H.C., Rahim, K., 2001. The information content of indicators in intercropping research. Agriculture, Ecosystems and Environment 87, 191-207.

Constable, G.A., 1976. Temperature effects on the early field development of cotton. Australian Journal of Experimental Agriculture and Animal Husbandry 16, 905-910.

References

135

Crawley, M.J., 1997. Plant ecology. Blackwell Science, Cambridge. Davis, J.H.C., Garcia, S., 1987. The effects of plant arrangement and density on

intercropped beans (Phaseolus Vulgaris) and maize. 1. Traits related to dry matter and seed productivity. Field Crops Research 16, 105-115.

Davis, J.H.C., Woolley, J.N., 1993. Genotypic requirement for intercropping. Field Crops Research 34, 407-430.

Deazevedo, D.M.P., Beltrao, N.E.D., Vieira, D.J., Danobrega, L.B., 1993. Annual cotton cowpea intercropping. 1. Effect of row arrangement. Pesquisa Agropecuaria Brasileira 28, 813-822.

Dhingra, K.K., Dhillon, M.S., Grewal, D.S., Sharma, K., 1991. Performance of maize and mungbean intercropping in different planting patterns and row orientations. Indian Journal of Agronomy 36, 207-212.

Diekmann, M., 1996. Relationship between flowering phenology of perennial herbs and meteorological data in deciduous forests of Sweden. Canadian Journal of Botany-Revue Canadienne De Botanique 74, 528-537.

Dobermann, A., Buresh, R.J., Peng, S., Witt, C., Wang, G., Huang, J., Yang, J., Zhong, X., Zou, Y., 2004. Overview of China’s cropping systems. Resource management for sustainable intensive agricultural systems. ERSEC Ecological Book Series-1 UNESCO publication. Tsinghua University Press and Springer, pp. 149-170.

Duncan, W.G., 1972. SIMCOT: A simulator of cotton growth and yield. In: Murphy, C.M. (Ed.), Proceedings of a Workshop for Modeling Tree Growth. Duke Univ., Durham, NC, pp. 115-118.

Echekwu, C.A., Alabi, S.O., 1994. A diallel analysis of earliness in interspecific crosses of cotton. Discovery and Innovation 6, 241-243.

Fang, C., Babcock, B.A., 2003. China’s cotton policy and the impact of China’s WTO accession and Bt cotton adoption on the Chinese and U.S. cotton sectors. Working Paper 03-WP-322. Center for Agricultural and Rural Development, Iowa State University, Ames, Iowa 50011-1070, p. 30.

Flenet, F., Guerif, A., Boiffin, J., Dorvillez, D., Champolivier, L., 2006. The critical N dilution curve for linseed (Linum usitatissimum L.) is different from other C3 species. European Journal of Agronomy 24, 367-373.

Fukai, S., Trenbath, B.R., 1993. Processes determining intercrop productivity and yields of component crops. Field Crops Research 34, 247-271.

Gethi, M., Omolo, E.O., Mueke, J.M., 1993. The effect of intercropping on relative resistance and susceptibility of cowpea cultivars to Maruca testulalis Geyer when in mono and when intercropped with maize. Insect Science and Its Application 14, 305-313.

Ghaffarzadeh, M., Prechac, F.G., Cruse, R.M., 1994. Grain yield response of corn,

References

136

soybean, and oat grown in a strip intercropping system. American Journal of Alternative Agriculture 9, 171-177.

Ghosh, P.K., 2004. Growth, yield, competition and economics of groundnut/cereal fodder intercropping systems in the semi-arid tropics of India. Field Crops Research 88, 227-237.

Gibson, D.J., Connolly, J., Hartnett, D.C., Weidenhamer, J.D., 1999. Designs for greenhouse studies of interactions between plants. Journal of Ecology 87, 1-16.

Gijzen, H., Goudriaan, J., 1989. A flexible and explanatory model of light distribution and photosynthesis in row crops. Agricultural and Forest Meteorology 48, 1-20.

Goudriaan, J., 1977. Crop micrometeorology: A simulation study. Simulation Monographs, Pudoc, Wageningen.

Goudriaan, J., Monteith, J.L., 1990. A Mathematical function for crop growth based on light interception and leaf-area expansion. Annals of Botany 66, 695-701.

Goudriaan, J., van Laar, H.H., 1994. Modelling potential crop growth processes. Kluwer Academic Publishers, Dordrecht, The Netherlands.

Goudriaan, J., van Roermund, H.J.W., 1999. Modelling of ageing, development, delays and dispersion. In: Leffelaar, P.A. (Ed.), On systems analysis and simulation of ecological process with examples in CSMP, FST and FORTRAN. Kluwer Academic Publishers, Dordrecht, pp. 89-126.

Gutierrez, A.P., Pizzamiglio, M.A., Dos Santos, W.J., Tennyson, R., Villacorta, A.M., 1984. A general distributed delay time varying life table plant population model: Cotton (Gossypium hirsutum L.) growth and development as an example. Ecological Modelling 26, 231-249.

Haefele, S.M., Wopereis, M.C.S., Ndiaye, M.K., Barro, S.E., Isselmod, M.O., 2003. Internal nutrient efficiencies, fertilizer recovery rates and indigenous nutrient supply of irrigated lowland rice in Sahelian West Africa. Field Crops Research 80, 19-32.

Hanan, J.S., Hearn, A.B., 2003. Linking physiological and architectural models of cotton. Agricultural Systems 75, 47-77.

Hanninen, H., 1995. Effects of climatic change on trees from cool and temperate regions: An ecophysiological approach to modelling of bud burst phenology. Canadian Journal of Botany-Revue Canadienne de Botanique 73, 2030-2030.

Hauggaard-Nielsen, H., Ambus, P., Jensen, E.S., 2001. Interspecific competition, N use and interference with weeds in pea-barley intercropping. Field Crops Research 70, 101-109.

Hearn, A.B., 1994. OZCOT: A simulation model for cotton crop management. Agricultural Systems 44, 257-299.

Hearn, A.B., da Roza, G.D., 1985. A simple model for crop management applications for cotton (Gossypium hirsutum L.). Field Crops Research 12, 49-69.

References

137

Hou, Z., Li, P., Li, B., Gong, J., Wang, Y., 2007. Effects of fertigation scheme on N uptake and N use efficiency in cotton. Plant and Soil 290, 115-126.

Hsu, H.H., Gale, F., 2001. Regional shifts in China’s cotton production and use. Cotton and Wool Situation and Outlook. U.S. Department of Agriculture, Economic Research Service, CWS-2001.

Huang, J., Hu, R., Rozelle, S., Qiao, F., Pray, C.E., 2002. Transgenic varieties and productivity of smallholder cotton farmers in China. The Australian Journal of Agricultural and Resource Economics 46, 367-387.

Iragavarapu, T., Randall, G., 1996. Border effects on yields in a strip-intercropped soybean, corn, and wheat production system. Journal of Production Agriculture 9, 101-107.

Ishaq, M., Ibrahim, M., Lal, R., 2001. Tillage effect on nutrient uptake by wheat and cotton as influenced by fertilizer rate. Soil and Tillage Research 62, 41-53.

Jallas, E., Martin, P., Sequeira, R., Turner, S., Cretenet, M., Gerardeaux, E., 2000. Virtual COTONS, the firstborn of the next generation of simulation models. Virtual Worlds 1834, 235-244.

Jalota, S.K., Sood, A., Chahal, G.B.S., Choudhury, B.U., 2006. Crop water productivity of cotton (Gossypium hirsutum L.) - wheat (Triticum aestivum L.) system as influenced by deficit irrigation, soil texture and precipitation. Agricultural Water Management 84, 137-146.

Jin, M., Zhang, R., Sun, L., Gao, Y., 1999. Temporal and spatial soil water management: A case study in the Heilonggang region, P.R. China. Agricultural Water Management 42, 173-187.

Jolliffe, P.A., 2000. The replacement series. Journal of Ecology 88, 371-385. Jones, J.W., Brown, L.G., Hesketh, J., 1980. COTCROP: A computer model for cotton

growth and yield. In: Hesketh, J.D., Jones, J.W. (Eds.), Predicting photosynthesis for ecosystem models. CRC Press, Boca Raton, FL, pp. 209-241.

Justes, E., Mary, B., Meynard, J.M., Machet, J.M., Thelierhuche, L., 1994. Determination of a critical nitrogen dilution curve for winter-wheat crops. Annals of Botany 74, 397-407.

Justes, E., Denoroy, P., Gabrielle, B., Gosse, G., 2000. Effect of crop nitrogen status and temperature on the radiation use efficiency of winter oilseed rape. European Journal of Agronomy 13, 165-177.

Keating, B.A., Carberry, P.S., 1993. Resource capture and use in intercropping - solar radiation. Field Crops Research 34, 273-301.

Kersebaum, K.C., Lorenz, K., Reuter, H.I., Schwarz, J., Wegehenkel, M., Wendroth, O., 2005. Operational use of agro-meteorological data and GIS to derive site specific nitrogen fertilizer recommendations based on the simulation of soil and

References

138

crop growth processes. Physics and Chemistry of the Earth 30, 59-67. Khorsandi, F., Whisler, F.D., 1996. Validation of the soil-temperature model of

GOSSYM-COMAX. Agricultural Systems 51, 131-146. Kiniry, J.R., Jones, C.A., O’Toole, J.C., Blanchet, R., Cabelguenne, M., Spanel, D.A.,

1989. Radiation use efficiency in biomass accumulation prior to grain-filling for five grain crop species. Field Crops Research 20, 51-64.

Kohel, R.J., Lewis, C.F., 1984. Cotton. American Society of Agronomy, Madison. Kropff, M.J., Goudriaan, J., 1994. Competition for resource capture in agricultural

crops. In: Resource capture by crops. Eds J.L. Monteith, R.K. Scott & M.H. Unsworth, Nottingham University Press, Loughborough, Leicestershire, pp. 233-253.

Lantinga, E.A., Nassiri, M., Kropff, M.J., 1999. Modelling and measuring vertical light absorption within grass-clover mixtures. Agricultural and Forest Meteorology 96, 71-83.

Lei, T.T., Gaff, N., 2003. Recovery from terminal and fruit damage by dry season cotton crops in tropical Australia. Journal of Economic Entomology 96, 730-736.

Leigh, R.A., Johnston, A.E., 1987. The usefulness of expressing nitrogen concentrations in crops on the basis of tissue water. Journal of the Science of Food and Agriculture 38, 317-318.

Lemaire, G., Oosterom, E., Sheehy, J., Jeuffroy, M.H., Massignam, A., Rossato, L., 2007. Is crop N demand more closely related to dry matter accumulation or leaf area expansion during vegetative growth? Field Crops Research 100, 91-106.

Lemmon, H., Chuk, N., 1997. Object-oriented design of a cotton crop model. Ecological Modelling 94, 45-51.

Lesoing, G.W., Francis, C.A., 1999. Strip intercropping effects on yield and yield components of corn, grain sorghum, and soybean. Agronomy Journal 91, 807-813.

Li, F., Cui, J., Liu, C., Wu, Z., Li, F., Zhou, Y., Li, X., Guo, S., Cui, H., 2001a. Study on insect-resistant transgenic cotton harbouring double-gene and its resistance to insect pests. Agricultural Sciences in China 1, 70-76.

Li, H.M., Inanaga, S., Li, Z.H., Eneji, A.E., 2005. Optimizing irrigation scheduling for winter wheat in the North China Plain. Agricultural Water Management 76, 8-23.

Li, L., Sun, J.H., Zhang, F.S., Li, X.L., Yang, S.C., Rengel, Z., 2001b. Wheat/maize or wheat/soybean strip intercropping I. Yield advantage and interspecific interactions on nutrients. Field Crops Research 71, 123-137.

Li, X., Zuo, C., 1997. Sustainable agriculture and rural development in China. Part 1: The agro-ecosystem and China’s rural economy. Promotion of sustainable agricul-ture and rural development in China: An element for a policy framework and a national agenda 21 action programme. FAO/UNDP/Ministry of Agriculture, China.

References

139

Li, X.L., Su, D.R., Yuan, Q.H., 2007. Ridge-furrow planting of alfalfa (Medicago sativa L.) for improved rainwater harvest in rainfed semiarid areas in Northwest China. Soil and Tillage Research 93, 117-125.

Licht, M.A., Al-Kaisi, M., 2005. Strip-tillage effect on seedbed soil temperature and other soil physical properties. Soil and Tillage Research 80, 233-249.

Liu, C.M., Yu, J.J., Kendy, E., 2001. Groundwater exploitation and its impact on the environment in the North China Plain. Water International 26, 265-272.

Liu, M.Q., Yu, Z.R., Liu, Y.H., Konijn, N.T., 2006. Fertilizer requirements for wheat and maize in China: The QUEFTS approach. Nutrient Cycling in Agroecosystems 74, 245-258.

Matthews, R.B., Azamali, S.N., Saffell, R.A., Peacock, J.M., Williams, J.H., 1991. Plant growth and development in relation to the microclimate of a sorghum groundnut intercrop. Agricultural and Forest Meteorology 53, 285-301.

May, O.L., Bridges, B.C., 1995. Breeding Cottons for Conventional and Late-Planted Production Systems. Crop Science 35, 132-136.

McKinion, J.M., Baker, D.N., 1989. Application of the GOSSYM/COMAX system to cotton crop management. Agricultural Systems 31, 55-65.

McKinion, J.M., Baker, D., Hesketh, J., Jones, J.W., 1975. SIMCOT II - a simulation of cotton growth and yield. Computer Simulation of a Cotton Production System. USDA-ARS-S-52, p. 27.

McKinion, J.M., Jenkins, J.N., Akins, D., Turner, S.B., Willer, J.L., Jallas, E., Whisler, F.D., 2001. Analysis of a precision agriculture approach to cotton production. Computers and Electronics in Agriculture 32, 213-228.

Meinke, H., Baethgen, W.E., Carberry, P.S., Donatelli, M., Hammer, G.L., Selvaraju, R., Stockle, C.O., 2001. Increasing profits and reducing risks in crop production using participatory systems simulation approaches. Agricultural Systems 70, 493-513.

Michael, P.B., Milroy, S.P., 2004. Impact of ‘cold shock’ on early development of cotton. 11th Australian Cotton Conference proceedings, Queensland, Australia.

Midmore, D.J., 1993. Agronomic modification of resource use and intercrop productivity. Field Crops Research 34, 357-380.

Midmore, D.J., Roca, J., Berrios, D., 1988. Potato (Solanum Spp) in the hot tropics.4. Intercropping with maize and the influence of shade on potato microenvironment and crop growth. Field Crops Research 18, 141-157.

Milroy, S.P., Bange, M.P., 2003. Nitrogen and light responses of cotton photosynthesis and implications for crop growth. Crop Science 43, 904-913.

Miralles, D.J., Slafer, G.A., 1991. A simple model for nondestructive estimates of leaf area in wheat. Cereal Research Communications 19, 439-444.

References

140

Monteith, J.L., 1977. Climate and efficiency of crop production in Britain. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences 281, 277-294.

Munger, P., Bleiholder, H., Hack, H., Hess, M., Stauss, R., van den Boom, T., Weber, E., 1998. Phenological growth stages of the cotton plant (Gossypium hirsutum L.): Codification and description according to the BBCH scale. Journal of Agronomy and Crop Science-Zeitschrift für Acker und Pflanzenbau 180, 143-149.

Mutsaers, H.J.W., 1984. KUTUN: A morphogenetic model for cotton (Gossypium hirsutum L.). Agricultural Systems 14, 229-257.

National Bureau of Statistic of China (NBC), Agriculture. In: China Statistical yearbook: 2003, 2004, 2005. Beijing, China.

National Bureau of Statistic of China (NBC), 2005. Agriculture. In: China Statistical yearbook. Beijing, China.

O’Connell, M.G., O’Leary, G.J., Whitfield, D.M., Connor, D.J., 2004. Interception of photosynthetically active radiation and radiation-use efficiency of wheat, field pea and mustard in a semi-arid environment. Field Crops Research 85, 111-124.

Ogg, C.L., 1960. Determination of nitrogen by the micro-Kjeldahl method. Journal of the Association of Official Agricultural Chemists 43, 689-693.

Olesen, J.E., Berntsen, J., Hansen, E.M., Petersen, B.M., Petersen, J., 2002. Crop nitrogen demand and canopy area expansion in winter wheat during vegetative growth. European Journal of Agronomy 16, 279-294.

Olesen, J.E., Hansen, P.K., Berntsen, J., Christensen, S., 2004. Simulation of above-ground suppression of competing species and competition tolerance in winter wheat varieties. Field Crops Research 89, 263-280.

Padhi, A.K., Sahoo, B.K., Das, K.C., 1993. Production potential, economic and energetics of upland cotton (Gossypium hirsutum) based intercropping systems under upland, rainfed situation. Indian Journal of Agricultural Sciences 63, 160-165.

Pan, X., Han, X., Dong, Z., 1997. A cotton growth and development simulation model COTGROW: I. Photosynthesis and dry matter partitioning. Cotton Science 9, 132-141 (in Chinese with English abstract).

Parajulee, M.N., Montandon, R., Slosser, J.E., 1997. Relay intercropping to enhance abundance of insect predators of cotton aphid (Aphis gossypii Glover) in Texas cotton. International Journal of Pest Management 43, 227-232.

Peng, S., Bouman, B.A.M., 2007. Prospects for genetic improvement to increase lowland rice yields with less water and nitrogen. In: Spiertz, J.H.J., Struik, P.C., van Laar, H.H. (Eds.), Scale and Complexity in Plant Systems Research: Gene-Plant-Crop Relations. Springer, Wageningen UR Frontis Series Vol 21, pp. 251-266.

Pinter Jr, P.J., Kimball, B.A., Mauney, J.R., Hendrey, G.R., Lewin, K.F., Nagy, J.,

References

141

1994. Effects of free-air carbon dioxide enrichment on PAR absorption and conversion efficiency by cotton. Agricultural and Forest Meteorology 70, 209-230.

Pronk, A.A., Goudriaan, J., Stilma, E., Challa, H., 2003. A simple method to estimate radiation interception by nursery stock conifers: A case study of eastern white cedar. Netherlands Journal of Agricultural Science 51, 279-295.

Rana, G., Katerji, N., Introna, M., Hammami, A., 2004. Microclimate and plant water relationship of the ‘overhead’ table grape vineyard managed with three different covering techniques. Scientia Horticulturae 102, 105-120.

Reddall, A., Sadras, V.O., Wilson, L.J., Gregg, P.C., 2004. Physiological responses of cotton to two-spotted spider mite damage. Crop Science 44, 835-846.

Reddy, K.R., Reddy, V.R., Hodges, H.F., 1992. Temperature effects on early season cotton growth and development. Agronomy Journal 84, 229-237.

Reddy, K.R., Hodges, H.F., McKinion, J.M., 1993. Temperature effects on pima cotton leaf growth. Agronomy Journal 85, 681-686.

Reddy, M.S., Willey, R.W., 1981. Growth and resource use studies in an intercrop of pearl millet/groundnut. Field Crops Research 4, 13-24.

Reddy, V.R., 1988. Estimation of parameters for the cotton simulation model GOSSYM: Cultivar differences. Agricultural Systems 26, 111-122.

Reddy, V.R., 1990. Application of GOSSYM to analysis of the effects of weather on cotton yields. Agricultural Systems 32, 83-95.

Reddy, V.R., Baker, D.N., Hodges, H.F., 1991a. Temperature effects on cotton canopy growth, photosynthesis, and respiration. Agronomy Journal 83, 699-704.

Reddy, V.R., Reddy, K.R., Baker, D.N., 1991b. Temperature effect on growth and development of cotton during the fruiting period. Agronomy Journal 83, 211-217.

Robertson, G.W., 1968. A biometeorological time scale for a cereal crop involving day and night temperatures and photoperiod. International Journal of Biometeorology 12, 191-223.

Rochester, I.J., 2007. Nutrient uptake and export from an Australian cotton field. Nutrient Cycling in Agroecosystems 77, 213-223.

Rohrig, M., Stutzel, H., Alt, C., 1999. A three-dimensional approach to modeling light interception in heterogeneous canopies. Agronomy Journal 91, 1024-1032.

Rosenthal, W.D., Gerik, T.J., 1991. Radiation use efficiency among cotton cultivars. Agronomy Journal 83, 655-658.

Rossiter, D.G., Riha, S.J., 1999. Modeling plant competition with the GAPS object-oriented dynamic simulation model. Agronomy Journal 91, 773-783.

Roussopoulos, D., Liakatas, A., Whittington, W.J., 1998. Controlled-temperature effects on cotton growth and development. Journal of Agricultural Science 130, 451-462.

References

142

Russell, G., Jarvis, P.G., Monteith, J.L., 1989. Absorption of radiation by canopies and stand growth. In: Russell, G. (Ed.), Plant canopy: their growth, form and function. Cambridge University Press, Cambridge, pp. 21-39.

Sadras, V.O., 1995. Compensatory growth in cotton after loss of reproductive organs. Field Crops Research 40, 1-18.

Sadras, V.O., 1996. Cotton responses to simulated insect damage: Radiation-use efficiency, canopy architecture and leaf nitrogen content as affected by loss of reproductive organs. Field Crops Research 48, 199-208.

Sadras, V.O., Wilson, L.J., 1997. Growth analysis of cotton crops infested with spider mites.1. Light interception and radiation-use efficiency. Crop Science 37, 481-491.

Sainju, U.M., Whitehead, W.F., Singh, B.P., Wang, S., 2006. Tillage, cover crops, and nitrogen fertilization effects on soil nitrogen and cotton and sorghum yields. European Journal of Agronomy 25, 372-382.

Saville, B.P., 2004. Physical testing of textiles. Woodhead Publishing/CRC Press, 314 pp. http://www.knovel.com/knovel2/Toc.jsp?bookID=925 (accessed on May 26, 2007)

Schader, C., Zaller, J.G., Kopke, U., 2005. Cotton-basil intercropping: Effects on pests, yields and economical parameters in an organic field in Fayoum, Egypt. Biological Agriculture and Horticulture 23, 59-72.

Schneekloth, J.P., Klocke, N.L., Davison, D.R., Payero, J.O., 2006. Furrow irrigation management with limited water. Applied Engineering in Agriculture 22, 391-398.

Singh, N.B., Singh, P.P., Nair, K.P.P., 1986. Effect of legume intercropping on enrichment of soil nitrogen, bacterial activity and productivity of associated maize crops. Experimental Agriculture 22, 339-344.

Singh, R., Singh, J., 1996. Irrigation planning in cotton through simulation modeling. Irrigation Science 17, 31-36.

Smith, M.E., Francis, C.A., 1986. Breeding for multiple cropping systems. In: Francis, C.A. (Ed.), Multiple cropping systems. Macmillan, New York, pp. 219-249.

Soltani, A., Hammer, G.L., Torabi, B., Robertson, M.J., Zeinali, E., 2006. Modeling chickpea growth and development: Phenological development. Field Crops Research 99, 1-13.

Spiertz, J.H.J., Vos, J., 2005. Nitrogen use efficiency and productivity of food crops in cereal-based agroecosystems. Contributed papers (Eds.: Zhaoliang Zhu, Katsu Minami and Guangxi Xing), 3rd Intern Nitrogen Conf, October 12-16, 2004, Nanjing, China. Science Press USA Inc., pp. 259-265.

Spitters, C.J.T., Aerts, R., 1983. Simulation of competition for light and water in crop-weed associations. Aspects of Applied Biology 4, 467-483.

Stone, P.J., Sorensen, I.B., Jamieson, P.D., 1999. Effect of soil temperature on

References

143

phenology, canopy development, biomass and yield of maize in a cool-temperate climate. Field Crops Research 63, 169-178.

Szumigalski, A.R., Van Acker, R.C., 2006. Nitrogen yield and land use efficiency in annual sole crops and intercrops. Agronomy Journal 98, 1030-1040.

Tao, H.B., Brueck, H., Dittert, K., Kreye, C., Lin, S., Sattelmacher, B., 2006. Growth and yield formation of rice (Oryza sativa L.) in the water-saving ground cover rice production system (GCRPS). Field Crops Research 95, 1-12.

Thompson, R., Clark, R.M., 2006. Spatio-temporal modelling and assessment of within-species phenological variability using thermal time methods. International Journal of Biometeorology 50, 312-322.

Thornley, J.H.M., Johnson, I.R., 1990. Plant and crop modeling, a mathematical approach to plant and crop physiology. Clarendon Press, Oxford.

Tianzhi, R., 2004. Overview of China’s cropping systems. Resource management for sustainable intensive agricultural systems. ERSEC Ecological Book Series-1 UNESCO publication. Tsinghua University Press and Springer, pp. 96-101.

van der Meer, J., 1989. The ecology of intercropping. Cambridge University Press, New York, 237 pp.

van Heemst, H.D.J., 1988. Plant data values required for simple crop growth simulation models: Review and bibliography. Simulation Report CABO-TT nr.17. Wageningen University, Wageningen, The Netherlands, pp. 67-69.

van Kraalingen, D.W.G., Rappoldt, C., van Laar, H.H., 2003. The Fortran simulation translator, a simulation language. European Journal of Agronomy 18, 359-361.

Wall, G.W., Amthor, J.S., Kimball, B.A., 1994. COTCO2: A cotton growth simulation model for global change. Agricultural and Forest Meteorology 70, 289-342.

Wallace, S.U., Whitwell, T., Palmer, J.H., Hood, C.E., Hull, S.A., 1992. Growth of relay intercropped soybean. Agronomy Journal 84, 968-973.

Waterworth, J.V., 1994. Intercropping cotton and groundnut in low and high rainfall areas in eastern Zambia. Experimental Agriculture 30, 461-465.

Watiki, J.M., Fukai, S., Banda, J.A., Keating, B.A., 1993. Radiation interception and growth of maize/cowpea intercrop as affected by maize plant density and cowpea cultivar. Field Crops Research 35, 123-133.

Watkins, K.B., Lu, Y.C., Reddy, V.R., 1998. An economic evaluation of alternative pix application strategies for cotton production using GOSSYM/COMAX. Computers and Electronics in Agriculture 20, 251-262.

West, T.D., Griffith, D.R., 1992. Effect of strip-intercropping corn and soybean on yield and profit. Journal of Production Agriculture 5, 107-110.

Whitfield, D.M., 1986. A simple model of light penetration into row crops. Agricultural and Forest Meteorology 36, 297-315.

References

144

Wilkerson, G.G., Jones, J.W., Coble, H.D., Gunsolus, J.L., 1990. SOYWEED: A simulation model of soybean and common cocklebur growth and competition. Agronomy Journal 82, 1003-1010.

Willey, R.W., 1979. Intercropping: Its importance and research needs. I. Competition and yield advantages. Field Crops Abstract 32, 1-10.

Willey, R.W., 1985. Evaluation and presentation of intercropping advantages. Experimental Agriculture 21, 119-133.

Willey, R.W., 1990. Resource use in intercropping systems. Agricultural Water Management 17, 215-231.

Xia, J., 1997. Biological control of cotton aphid (Aphis gossypii Glover) in cotton (inter)cropping systems in China: A simulation study. PhD thesis Wageningen University, Wageningen, The Netherlands.

Xiao, Y.B., Li, L., Zhang, F.S., 2004. Effect of root contact on interspecific competition and N transfer between wheat and faba bean using direct and indirect N15 techniques. Plant and Soil 262, 45-54.

Xu, X., Xu, Y., 1989. The influence on cotton growth and development by environmental factors: 2. The relationship of cotton vegetative growth with temperature. Jiangxi Cotton 1, 9-11 (in Chinese).

Xue, Q., Zhu, Z., Musick, J.T., Stewart, B.A., Dusek, D.A., 2003. Root growth and water uptake in winter wheat under deficit irrigation. Plant and Soil 257, 151-161.

Yadav, R.S., Yadav, O.P., 2001. The performance of cultivars of pearl millet and cluster bean under sole cropping and intercropping systems in arid zone conditions in India. Experimental Agriculture 37, 231-240.

Yang, H.S., Dobermann, A., Lindquist, J.L., Walters, D.T., Arkebauer, T.J., Cassman, K.G., 2004. Hybrid-maize: A maize simulation model that combines two crop modeling approaches. Field Crops Research 87, 131-154.

Yin, X., Kropff, M.J., McLaren, G., Visperas, R.M., 1995. A nonlinear model for crop development as a function of temperature. Agricultural and Forest Meteorology 77, 1-16.

Yin, X., Struik, P.C., Kropff, M.J., 2004. Role of crop physiology in predicting gene-to-phenotype relationships. Trends in Plant Science 9, 426-432.

Yunusa, I.A.M., Siddique, K.H.M., Belford, R.K., Karimi, M.M., 1993. Effect of canopy structure on efficiency of radiation interception and use in spring wheat cultivars during the preanthesis period in a mediterranean-type environment. Field Crops Research 35, 113-122.

Zadoks, J.C., Chang, T.T., Konzak, C.F., 1974. A decimal code for the growth stages of cereals. Weed Research 14, 415-421.

Zhang, F., Li, L., 2003. Using competitive and facilitative interactions in intercropping

References

145

systems enhances crop productivity and nutrient-use efficiency. Plant and Soil 248, 305-312.

Zhang, F., Shen, J., Li, L., Liu, X., 2004a. An overview of rhizosphere processes related with plant nutrition in major cropping systems in China. Plant and Soil 260, 89-99.

Zhang, L., 2004. Quantitative analysis of agricultural household response to ratio of grain and cotton price change in China. Cotton Science 16, 49-54.

Zhang, L., Cao, W., Zhang, S., 2004b. Dynamic simulation on dry matter partitioning and yield formation in cotton. Scientia Agricultura Sinica 37, 1621-1627 (in Chinese with English abstract).

Zhang, L., Cao, W., Zhang, S., Luo, W., 2003a. Simulation model for cotton (Gossypium hirsutum L.) development stages based on physiological development time. Cotton Science 15, 97-103 (in Chinese with English abstract).

Zhang, L., Li, Y., 1997. The technical approach on wheat-cotton double cropping system sustainable development in Huanghuaihai plain. Proceedings of International Symposium of Sustainable Agricultural Technologies (ISSAT’97), Beijing.

Zhang, X.Y., Pei, D., Hu, C.S., 2003b. Conserving groundwater for irrigation in the North China Plain. Irrigation Science 21, 159-166.

Zhao, D., Reddy, K.R., Kakani, V.G., Koti, S., Gao, W., 2005. Physiological causes of cotton fruit abscission under conditions of high temperature and enhanced ultraviolet-B radiation. Physiologia Plantarum 124, 189-199.

Zhao, Z., Liu, D., Nan, J., 1996. Modelling and application of phenological stages in cotton. Journal of Shanxi Agricultural University 16, 92-96 (in Chinese).

Zhen, L., Routray, J.K., Zoebisch, M.A., Chen, G.B., Xie, G.D., Cheng, S.K., 2005. Three dimensions of sustainability of farming practices in the North China Plain: A case study from Ningjin county of Shandong province, P.R. China. Agriculture, Ecosystems and Environment 105, 507-522.

Zhen, L., Zoebisch, M.A., Chen, G.B., Feng, Z.M., 2006. Sustainability of farmers' soil fertility management practices: A case study in the North China Plain. Journal of Environmental Management 79, 409-419.

Zhou, J., Wu, Y.Z., Yan, G., 2005. General formula for estimation of monthly average daily global solar radiation in China. Energy Conversion and Management 46, 257-268.

147

Appendix I Listing of row-structured light interception model for wheat and

cotton in intercropping systems TITLE A row-structured light interception model for intercropping * ********************************************************************** * Lizhen Zhang * * * * The row structured model used for calculation of * * light interception by wheat and cotton in intercropping systems * * based on the approach of Goudriaan (1977) and Pronk et al. (2003) * * * * i number of wheat rows in intercrop * * j number of cotton rows in intercrop * * For example, RD31 is RD for i=3 and j=1; * * indicates RD (cumulative PAR) in 3:1 system * * ********************************************************************** INITIAL INCON ZERO = 0. PARAM KC = 0.95 PARAM KW =0.7 PARAM WHARVE =163. PARAM CSOW =116. * KC light extinction coefficient of cotton * KW light extinction coefficient of wheat * WHARVE day of wheat harvest * CSOW day of cotton sowing WEATHER WTRDIR='E:\WEATHER\ANYANG\';CNTR='AY';ISTN=1;IYEAR=2002 * Reading weather data: * RDD Daily global radiation J/m2/d * TMMN Daily minimum temperature oC * TMMX Daily maximum temperature oC * VP Vapour pressure kPa * WN Wind speed m/s * RAIN Precipitation mm * LAT Latitude of the site degree

Appendix I

148

TIMER STTIME =1.; FINTIM = 300.; DELT = 1.; PRDEL =1. PRINT RDC,RDC31,RDC32,RDC42,RDC62,RDW,RDW31,RDW32,RDW42,RDW62 DYNAMIC * Light interception * RADIATION INTERCEPTION (MJ/M2/D) RDR = RDD/1000000./2. RD = INTGRL(ZERO,RDR) * RDR: daily incoming PAR; RD: cumulative PAR * PAR intercepted by wheat and cotton in the intercrops * RDRW : daily PAR intercepted by mono-wheat MJ m-2 d-1 * RDW : cumulative PAR intercepted by mono-wheat MJ m-2 * RDRC : daily PAR intercepted by mono-cotton MJ m-2 d-1 * RDC : cumulative PAR intercepted by mono-cotton MJ m-2 * RDRWij: daily PAR intercepted by wheat in intercrops MJ m-2 d-1 * RDWij : cumulative PAR intercepted by wheat in intercrops MJ m-2 * RDRCij: daily PAR intercepted by cotton in intercrops MJ m-2 d-1 * RDCij : cumulative PAR intercepted by cotton in intercrops MJ m-2 RDRW=RDR*FINTW RDRC=RDR*FINTC RDRC31=RDR*FCI31 RDRC32=RDR*FCI32 RDRC42=RDR*FCI42 RDRC62=RDR*FCI62 RDRW31=RDR*FW31 RDRW32=RDR*FW32 RDRW42=RDR*FW42 RDRW62=RDR*FW62 RDW=INTGRL(ZERO,RDRW) RDC=INTGRL(ZERO,RDRC) RDW31=INTGRL(ZERO,RDRW31) RDW32=INTGRL(ZERO,RDRW32) RDW42=INTGRL(ZERO,RDRW42) RDW62=INTGRL(ZERO,RDRW62) RDC31=INTGRL(ZERO,RDRC31) RDC32=INTGRL(ZERO,RDRC32) RDC42=INTGRL(ZERO,RDRC42) RDC62=INTGRL(ZERO,RDRC62) RDW1=NOTNUL(RDW) DASC=INSW(TIME-CSOW,0.,TIME-CSOW) *DASC: day after sowing of cotton

Listing of row structured light interception model for wheat and cotton

149

* LAI * Measured leaf area index (LAI), * As an example, LAI data (treatment averages) of 2002 are used in * the conclusions. Light interception were calculated based on LAI per * Plot in this thesis. * LAIW : leaf area index of mono-wheat m2 leaf m-2 ground * LAIC : leaf area index of mono-cotton m2 leaf m-2 ground LAIW=INSW((0.000001-LW),LW,0.000001) LW =AFGEN(LWTAB,TIME) FUNCTION LWTAB=1.,2.0,77.,2.78,94.,6.53,106.,7.7,120.,6.58,... 137.,6.03,148.,4.26,165.,2.,166.,0.,365.,0.0 LAIC=INSW((0.000001-LC),LC,0.000001) LC=AFGEN(LCTAB,TIME) FUNCTION LCTAB=1.,0.0,115.,0.0,125.,0.006,165.,0.10,176.,0.46,... 190.,1.13,207.,2.32,219.,2.63,233.,2.33,246.,2.04,261.,0.91,... 282.,0.33,290.,0.0,365.,0.0 * 3:1 system (3 rows wheat and 1 row cotton) * LAI31W: leaf area index of wheat in 3:1 system m2 leaf m-2 ground * LAI31C: leaf area index of cotton in 3:1 system m2 leaf m-2 ground LAI31W=INSW((0.000001-LW31),LW31,0.000001) LW31=AFGEN(LW1TAB,TIME) FUNCTION LW1TAB=1.,1.2,77.,1.84,94.,3.6,106.,4.3,120.,3.54,... 137.,3.62,148.,2.82,165.,1.2,166.,0.,365.,0.0 LAI31C=INSW((0.000001-LC31),LC31,0.0000001) LC31=AFGEN(LC1TAB,TIME) FUNCTION LC1TAB=1.0,0.,115.,0.,125.,0.006,165.,0.023,176.,0.078,... 190.,0.46,207.,1.06,219.,1.65,233.,1.57,246.,1.41,261.,1.,... 282.,0.18,290.,0.,365.,0. * 3:2 system (3 rows wheat and 2 rows cotton) * LAI32W: leaf area index of wheat in 3:2 system m2 leaf m-2 ground * LAI32C: leaf area index of cotton in 3:2 system m2 leaf m-2 ground LAI32W=INSW((0.00001-LW32),LW32,0.000001) LW32=AFGEN(LW2TAB,TIME) FUNCTION LW2TAB=1.,1.0,77.,1.26,94.,2.84,106.,3.99,120.,3.3,... 137.,2.9,148.,2.33,165.,1.,166.,0.,365.,0. LAI32C=INSW((0.000001-LC32),LC32,0.000001) LC32=AFGEN(LC2TAB,TIME) FUNCTION LC2TAB=1.,0.,115.,0.,125.,0.006,165.,0.04,176.,0.17,... 190.,0.74,207.,1.63,219.,2.94,233.,2.42,246.,2.37,261.,1.32,... 282.,0.18,290.,0.,365.,0. * 4:2 system (4 rows wheat and 2 rows cotton) * LAI42W: leaf area index of wheat in 4:2 system m2 leaf m-2 ground * LAI42C: leaf area index of cotton in 4:2 system m2 leaf m-2 ground

Appendix I

150

LAI42W=INSW((0.000001-LW42),LW42,0.0000001) LW42=AFGEN(LW3TAB,TIME) FUNCTION LW3TAB=1.,1.06,77.,1.33,94.,3.04,106.,3.75,120.,3.12,... 137.,2.73,148.,2.57,165.,1.06,166.,0.,365.,0. LAI42C=INSW((0.00001-LC42),LC42,0.000001) LC42=AFGEN(LC3TAB,TIME) FUNCTION LC3TAB=1.,0.,115.,0.,125.,0.006,165.,0.042,176.,0.15,... 190.,0.64,207.,1.72,219.,2.42,246.,2.33,261.,1.87,... 282.,0.13,290.,0.,365.,0. * 6:2 system (6 rows wheat and 2 rows cotton) * LAI62W: leaf area index of wheat in 6:2 system m2 leaf m-2 ground * LAI62C: leaf area index of cotton in 6:2 system m2 leaf m-2 ground LAI62W=INSW((0.000001-LW62),LW62,0.000001) LW62=AFGEN(LW4TAB,TIME) FUNCTION LW4TAB=1.,1.2,77.,1.58,94.,4.03,106.,4.55,... 137.,4.12,148.,2.4,165.,1.2,166.,0.,365.,0. LAI62C=INSW((0.000001-LC62),LC62,0.0000001) LC62=AFGEN(LC4TAB,TIME) FUNCTION LC4TAB=1.,0.,115.,0.,125.,0.006,165.,0.03,176.,0.14,... 190.,0.46,207.,1.33,219.,1.79,233.,1.83,246.,1.81,261.,0.96,... 282.,0.11,290.,0.,365.,0. * Plant height HW=AFGEN(HWTAB,TIME) HC=AFGEN(HCTAB,TIME) FUNCTION HWTAB=1.,10.,70.,10.,130.,64.,165.,64. FUNCTION HCTAB=1.,0.,130.,0.,148.,5.,178.,27.8,184.,37.5,... 190.,43.8,206.,85.1,211.,89.8,217.,91.2,365.,91.2 * HW: measured plant height of wheat cm * HC: measured plant height of cotton cm * Fraction of PAR interception of wheat PARAM W31=60.; W32=60.; W42=80.0; W62=120. PARAM P31=40.; P32=60.; P42=70.; P62=80. * Wij : strip width of wheat in intercrops cm * Pij : path width of wheat in intercrops cm * FINTW: fraction of PAR intercepted by mono-wheat - * FWij : fraction of PAR intercepted by wheat in intercrops - * Mono-wheat FINTW=1.-EXP(-KW*LAIW) * 3:1 wheat HFW31=1.-EXP(-KW*LAI31W) L31WR=LAI31W*(W31+P31)/W31 FI31WR=W31/(W31+P31)*(1.-EXP(-KW*L31WR))


151

IPBW31=(SQRT(HW*HW+P31*P31)-HW)/P31 SPW31=IPBW31+(1.-IPBW31)*EXP(-KW*LAI31W) TI31=EXP(-KW*L31WR) IRBW31=(SQRT(HW*HW+W31*W31)-HW)/W31 SRW31=IRBW31*TI31+(1.-IRBW31)*EXP(-KW*LAI31W) FIA31=HFW31-FI31WR SIA31=SPW31-SRW31 FR31W=HFW31-((FIA31*SIA31)/(1.-TI31)) FW31=INSW((0.-FR31W),FR31W,0.) *3:2 wheat HFW32=1.-EXP(-KW*LAI32W) L32WR=LAI32W*(W32+P32)/W32 FI32WR=W32/(W32+P32)*(1.-EXP(-KW*L32WR)) IPBW32=(SQRT(HW*HW+P32*P32)-HW)/P32 SPW32=IPBW32+(1.-IPBW32)*EXP(-KW*LAI32W) TI32=EXP(-KW*L32WR) IRBW32=(SQRT(HW*HW+W32*W32)-HW)/W32 SRW32=IRBW32*TI32+(1.-IRBW32)*EXP(-KW*LAI32W) FIA32=HFW32-FI32WR SIA32=SPW32-SRW32 FR32W=HFW32-((FIA32*SIA32)/(1.-TI32)) FW32=INSW((0.-FR32W),FR32W,0.) *4:2 wheat HFW42=1.-EXP(-KW*LAI42W) L42WR=LAI42W*(W42+P42)/W42 FI42WR=W42/(W42+P42)*(1.-EXP(-KW*L42WR)) IPBW42=(SQRT(HW*HW+P42*P42)-HW)/P42 SPW42=IPBW42+(1.-IPBW42)*EXP(-KW*LAI42W) TI42=EXP(-KW*L42WR) IRBW42=(SQRT(HW*HW+W42*W42)-HW)/W42 SRW42=IRBW42*TI42+(1.-IRBW42)*EXP(-KW*LAI42W) FIA42=HFW42-FI42WR SIA42=SPW42-SRW42 FR42W=HFW42-((FIA42*SIA42)/(1.-TI42)) FW42=INSW((0.-FR42W),FR42W,0.) * 6:2 wheat HFW62=1.-EXP(-KW*LAI62W) L62WR=LAI62W*(W62+P62)/W62 FI62WR=W62/(W62+P62)*(1.-EXP(-KW*L62WR)) IPBW62=(SQRT(HW*HW+P62*P62)-HW)/P62 SPW62=IPBW62+(1.-IPBW62)*EXP(-KW*LAI62W)

Appendix I

152

TI62=EXP(-KW*L62WR) IRBW62=(SQRT(HW*HW+W62*W62)-HW)/W62 SRW62=IRBW62*TI62+(1.-IRBW62)*EXP(-KW*LAI62W) FIA62=HFW62-FI62WR SIA62=SPW62-SRW62 FR62W=HFW62-((FIA62*SIA62)/(1.-TI62)) FW62=INSW((0.-FR62W),FR62W,0.) * Fraction of PAR interception of cotton PARAM C31=80.; C32=120.; C42=120.0; C62=120. PARAM PC31=20.; PC32=0.0001; PC42=30.; PC62=80. * Cij : strip width of cotton in intercrops cm * PCij : path width of cotton in intercrops cm * FINTC: fraction of PAR intercepted by mono-cotton - * FCij : fraction of PAR intercepted by cotton in intercrops * after wheat harvest - * FCij*SPWij: fraction of PAR intercepted by cotton in intercrops * during intercropping period (from sowing to wheat harvest) - * FCIij: fraction of PAR intercepted by cotton in intercrops * for whole growing season - * Mono-cotton FINTC=1.-EXP(-KC*LAIC) * 3:1 cotton HFC31=1.-EXP(-KC*LAI31C) L31CR=LAI31C*(C31+PC31)/C31 FI31CR=C31/(C31+PC31)*(1.-EXP(-KC*L31CR)) IPBC31=(SQRT(HC*HC+PC31*PC31)-HC)/PC31 SPC31=IPBC31+(1.-IPBC31)*EXP(-KC*LAI31C) TC31=EXP(-KC*L31CR) IRBC31=(SQRT(HC*HC+C31*C31)-HC)/C31 SRC31=IRBC31*TC31+(1.-IRBC31)*EXP(-KC*LAI31C) FCA31=HFC31-FI31CR SCA31=SPC31-SRC31 FR31C=HFC31-((FCA31*SCA31)/(1.-TC31)) FC31=INSW((0.-FR31C),FR31C,0.) *3:2 cotton HFC32=1.-EXP(-KC*LAI32C) L32CR=LAI32C*(C32+PC32)/C32 FI32CR=C32/(C32+PC32)*(1.-EXP(-KC*L32CR)) IPBC32=(SQRT(HC*HC+PC32*PC32)-HC)/PC32 SPC32=IPBC32+(1.-IPBC32)*EXP(-KC*LAI32C)


153

TC32=EXP(-KC*L32CR) IRBC32=(SQRT(HC*HC+C32*C32)-HC)/C32 SRC32=IRBC32*TC32+(1.-IRBC32)*EXP(-KC*LAI32C) FCA32=HFC32-FI32CR SCA32=SPC32-SRC32 FR32C=HFC32-((FCA32*SCA32)/(1.-TC32)) FC32=INSW((0.-FR32C),FR32C,0.) *4:2 cotton HFC42=1.-EXP(-KC*LAI42C) L42CR=LAI42C*(C42+PC42)/C42 FI42CR=C42/(C42+PC42)*(1.-EXP(-KC*L42CR)) IPBC42=(SQRT(HC*HC+PC42*PC42)-HC)/PC42 SPC42=IPBC42+(1.-IPBC42)*EXP(-KC*LAI42C) TC42=EXP(-KC*L42CR) IRBC42=(SQRT(HC*HC+C42*C42)-HC)/C42 SRC42=IRBC42*TC42+(1.-IRBC42)*EXP(-KC*LAI42C) FCA42=HFC42-FI42CR SCA42=SPC42-SRC42 FR42C=HFC42-((FCA42*SCA42)/(1.-TC42)) FC42=INSW((0.-FR42C),FR42C,0.) * 6:2 cotton HFC62=1.-EXP(-KC*LAI62C) L62CR=LAI62C*(C62+PC62)/C62 FI62CR=C62/(C62+PC62)*(1.-EXP(-KC*L62CR)) IPBC62=(SQRT(HC*HC+PC62*PC62)-HC)/PC62 SPC62=IPBC62+(1.-IPBC62)*EXP(-KC*LAI62C) TC62=EXP(-KC*L62CR) IRBC62=(SQRT(HC*HC+C62*C62)-HC)/C62 SRC62=IRBC62*TC62+(1.-IRBC62)*EXP(-KC*LAI62C) FCA62=HFC62-FI62CR SCA62=SPC62-SRC62 FR62C=HFC62-((FCA62*SCA62)/(1.-TC62)) FC62=INSW((0.-FR62C),FR62C,0.) * Fraction of PAR intercepted by intercropped cotton * in whole growing season FCI62=INSW((TIME-WHARVE),FC62*SPW62,FC62) FCI32=INSW((TIME-WHARVE),FC32*SPW32,FC32) FCI42=INSW((TIME-WHARVE),FC42*SPW42,FC42) FCI31=INSW((TIME-WHARVE),FC31*SPW31,FC31)

Appendix I

154

* Total Fraction of PAR intercepted by wheat and cotton in intercrops FT31=FCI31+FW31 FT42=FCI42+FW42 FT32=FCI32+FW32 FT62=FCI62+FW62 TRANSLATION_GENERAL DRIVER = 'EUDRIV' END STOP

155

Appendix II

Listing of the model SUCROS-Cotton DEFINE_CALL TOTASS(INPUT,INPUT,INPUT,INPUT,INPUT, ... INPUT, INPUT,INPUT,INPUT, ... OUTPUT,OUTPUT,OUTPUT,OUTPUT,OUTPUT) TITLE SUCROS-Cotton ************************************************************************ * * * SUCROS-Cotton * * A simple physiological based simulation model for * * potential cotton growth and development * * * * (FST-version 2.0) * * Lizhen Zhang * * Cotton Research Institute (CRI),CAAS, * * Anyang city, Henan province, 455112, P.R.China * * Email:[email protected] * * * * ********************************************************************** ARRAY BOLL(1:N) ,BOLLI(1:N) ,RBOLL(1:N) ,CBOLL(1:N),NETFLO(1:N),... FLOW(1:N+1),AGE(1:N) ,WBAGE(1:N) ,RWB(1:N) ,... STRESS(1:N),RWBAGE(1:N) ,WBOLL(1:N) ,... CWBOLL(1:N),WFLOW(1:N+1),WNETFL(1:N),RWBOL(1:N),... SHELL(1:N) ,SEDCOT(1:N) INITIAL INCON ZERO =0.0 INCON IW =7.8 ;IWLV=4.2;IROOT=1.8;ISTEM=1.8;ISHOT=5.5661 INCON ILAI =0.006 ;IWBOLL=1.92 INCON BOLLI=0. ;BOLL0I=0. PARAM GF =70. ;PLANTS=60000. ARRAY_SIZE N=74 WEATHER WTRDIR='E:\WEATHER\ANYANG\';CNTR='AY';ISTN=1;IYEAR=2002 * Reading weather data: * RDD Daily global radiation J.m-2.d-1 * TMMN Daily minimum temperature oC * TMMX Daily maximum temperature oC * VP Vapour pressure kPa * WN Wind speed m.s-1 * RAIN Precipitation mm * LAT Latitude of the site degree TIMER STTIME =116.; FINTIM =360.; PRDEL =1. ; DELT = 1. PRINT PDT ,LAI ,BIOMUP,WSHOOT,TDRW ,WLEAF ,WSTEM, BOLLW, ROOT,... HEIGHT,LEAFM ,FBA ,FN ,SQUARE,FLOWER,SBOLL, BBOLL, BOLOUT,...

Appendix II

156

GREBL ,COTTON,COTOUT,COTGRE,LINT ,LINTO ,LINTG, RLINT, LINTT,... WSOUT ,COTR ,STRBOL,PISHOT,PIROOT,FLT ,FALL1, FALL2, FALL3,... GLAI ,GTW ,RWBT ,RWBL ,RWLV ,RWSM ,FALLBL,RTE,... RPBO ,POOL ,RGRM ,RGUP ,MAINT ,... TAV ,TINT ,TMMN ,TNINT ,TMMX ,TMINT DYNAMIC * 1. Physiological development RPDT=MAX(0., RTE*RPE*VI) PDT =INTGRL(ZERO, RPDT) * PDT is physiological development time, including effectiveness of * thermal, photoperiod and variety maturity * 1.1. Thermal effectiveness (RTE) * Temperature table * Temperature TBase Toptimal TMax * - Emergence 15 25-30 40 (soil temperature) * - Square 12 30 35 * - Flowering 12 30 35 * - Boll opening 10 26-30 35 * Variables: * RTE relative thermal effectiveness * RTEB temp. table for air temperature * RTESB temp. table for soil temperature * TSAV soil temperature in degree c * TSCAV soil temperature with film cover in degree c * TAV average daily air temperature in degree c * TPLUS intermediate variable for temp. compensation with film cover * IE compensation coefficient before square is 0.51, flowering 0.22, * after flowering 0 (from field experiments in China) * Assume daily temperature are 50% of average and 25% of maximum * and 25% of minimum FUNCTION RTETB = -20.,0., 12.,0., 30.,1.0, 35.,0.,45.,0. FUNCTION RTESTB= -20.,0., 15.,0., 20.,0.3, 25.,1., 30.,1., 40.,0.,50.,0. * 1.1.1. Temperature of bare soil in monoculture of cotton TAV =(TMMN+TMMX)/2. TSAV=0.890 + 1.017*TAV * Soil temperature is only used from sowing to emergence * 1.1.2. Effect of film mulching * FILM=0.: without film cover; FILM=1.: with film cover PARAM FILM=0. TSCAV=7.5725 + 0.8303*TAV TPLUS=MAX(0.,((TSCAV-TSAV)/NOTNUL((TSCAV-12.)/NOTNUL(TAV-12.)))*IE) IE =INSW(PDT-17.5, 0.51, TETMP) TETMP=INSW(PDT-27.5, 0.22, 0.) * Effect of film mulching in the intercropping systems TIPLUS=MAX(0.,((TISCAV-TISAV)/NOTNUL((TISCAV-12.)/... NOTNUL(TINT-12.)))*IE)

Listing of the model SUCROS-Cotton

157

* 1.1.3. Temperature in wheat-cotton intercropping systems * INTERC=0.: sole cotton; INTERC=1.: strip intercropped with wheat PARAM INTERC=1. TMIN =TMMX-5.4 TNIN =TMMN TIN =(TMIN+TNIN)/2. TMINT =INSW(TIME-165.,TMIN,TMMX) TNINT =INSW(TIME-165.,TNIN,TMMN) TINT =INSW(TIME-165.,TIN,TAV) TISAV =TSAV-3.0 TISCAV=TSAV+2.7 * RTE calculations RFEE =AFGEN(RTESTB , TSCAV) RIES =AFGEN(RTESTB , TISAV) RIEF =AFGEN(RTESTB , TISCAV) RTEE1 =AFGEN(RTESTB , TSAV ) RTEI =INSW(INTERC-1., RTEE1,RIES) RTEIF =INSW(INTERC-1., RFEE, RIEF) RTEE =INSW(FILM-1. , RTEI, RTEIF) TAVI =INSW(INTERC-1., TAV,TINT) TAVIF =INSW(INTERC-1., TAV+TPLUS,TINT+TIPLUS) TAV1 =INSW(FILM-1. , TAVI ,TAVIF) TMMNI =INSW(INTERC-1., TMMN,TNINT) TMMNIF=INSW(INTERC-1., TMMN+TPLUS, TNINT+TIPLUS) TMMN1 =INSW(FILM-1. , TMMNI,TMMNIF) TMMXI =INSW(INTERC-1., TMMX,TMINT) TMMXIF=INSW(INTERC-1., TMMX+TPLUS,TMINT+TIPLUS) TMMX1 =INSW(FILM-1. , TMMXI,TMMXIF) RTEFSB=0.25*( 2.*AFGEN(RTETB,TAV1) + AFGEN(RTETB,TMMN1) + ... AFGEN(RTETB,TMMX1) ) RTE =INSW(PDT-2.5, RTEE, RTEFSB) PT =INTGRL(ZERO,RTE) * 1.2. Photoperiod effectiveness (RPE) * RPE: relative photoperiod effectiveness due to daylength. RPE=AFGEN(DLTB,DAYL) FUNCTION DLTB=8.0,1.,12.,1.,14.,1., 16.,0. * 1.3. Variety index (VI) * VI be valued according to genetic factor of cultivars * - Early maturing, 0.96-1.00, such as CRI37

Appendix II

158

* - Early-mid maturing, 0.82-0.84, such as CRI32 * - Middle maturing, 0.80-0.81, such as CRI12 PARAM VI=0.85 * 2. Cotton morphogenesis * CUT: Cut-out practice * Normally, farmer removes terminal bud of main stems to * avoid useless young fruits, the date of cut-out is at * cotton with 12-14 fruit branches, in 60000 plant.ha-1, * cut-out is practiced generally on early Aug. * in Yellow River region, China * FFB: the sequence of node on main stem, above FFB, * the fruit branches are initialized PARAM FFB=9. PARAM CUT=220. CUTOFF=INSW(TIME-CUT, 0., 1.) * 2.1 Plant height * HEIGHT plant height cm * RHR rate of height growth cm.d-1 * FLT temperature factor RHR1 =1.5*FLT RHR2 =INSW(PDT-2.5 ,0. ,RHR1) RHR =INSW(CUTOFF-1.,RHR2,0.) HEIGHT=INTGRL(ZERO, RHR) * 2.2. Leaf number in main stem * RLR rate of leaf development no.d-1 * LEAFA actual leaf number no * RLDR leaf abscission rate no.d-1 * LEAFD differentiation leaf number no * LEAFP potential total leaf number no * LEAFM total appearing leaf number no RLR1 =0.5998*RTE-0.0594 RLR2 =INSW(PDT-2.5,0.,MAX(0.,RLR1)) RLR3 =INSW(CUTOFF-1.,RLR2,0.) RLDR1=1./70.*LEAFA RLR =RLR3-RLDR1 LEAFM=INTGRL(ZERO,RLR3) LEAFA=INTGRL(ZERO,RLR) LEAFD=AFGEN(LDTB,LEAFM) LEAFP=LEAFM+LEAFD FUNCTION LDTB= 0.,4., 1.,5., 5.,9., 11.,9., 12.,10., 30.,11., ... 50.,15. * 2.3 Fruit branches * RFBR rate of fruit branch development no.d-1 * FBA actual fruit branches calculated by RFBR no * FBAV actual fruit branches derived from LEAFM no


159

* FBAD differentiation fruit branches no * FBAP potential fruit branches no RFBR1=1.5459*RTE-0.7712 RFBR2=INSW(LEAFM-FFB,0.,MAX(RLR3,RFBR1)) RFBR =INSW(CUTOFF-1.,RFBR2,0.) FBA =INTGRL(ZERO,RFBR) * For check number of fruit branch by two ways FBAV1=LEAFM-FFB FBAV =INSW(LEAFM-FFB, 0., FBAV1) FBAD =INSW(LEAFM-FFB, 0., LEAFD-2.) FBAP =FBAV+FBAD * 2.4. Fruit nodes * RFNR rate of fruit node development no.d-1 * FN actual fruit nodes no * FNAVM maximum fruit nodes after cut-out no RFNR1=5.8746*RTE-3.2508 RFNR2=INSW(LEAFM-FFB,0.,MAX(RFBR,RFNR1)) RFNR =INSW(FN-FNAVM,RFNR2,0.) FNAVM=AFGEN(FNTB,FBAP) FUNCTION FNTB=0., 0., ... 1., 1., 2., 2., 3., 3., 4., 5., 5., 7., 6., 9., 7.,12.,... 8.,15., 9.,18.,10.,22.,11.,26.,12.,30.,13.,35.,14.,40.,... 15.,45.,16.,51.,17.,57.,18.,63.,19.,70.,20.,77.,21.,84.,... 22.,92., 23.,100., 24.,108., 25.,117., 26.,126., ... 36.,234., 60., 500. FN=INTGRL(ZERO,RFNR) * 3. Calculation of leaf area index (LAI) * GLAI growth rate of LAI ha.ha-1.d-1 * FLV partitioning index to leaves * RWLV growth rate of leaf weight kg.ha-1.d-1 * RGUP dry matter to daily shoot growth kg.ha-1.d-1 * RLDR LAI abscission rate ha.ha-1.d-1 * SLA specific leaf area ha.kg-1 GLAI=INSW(PDT-2.5, 0., RWLV*0.0022*FLT - RLDR*LAI) RLDR=AFGEN(RLDRTB,PDT) SLA =AFGEN(SLATB ,PDT) LAI =INTGRL(ILAI ,GLAI) FUNCTION SLATB =0.,0., 2.5,0.00164, 27.5,0.0022, 80.,0.00136, ... 100.,0.00136 FUNCTION RLDRTB=0. ,0., 17.5,0.0, 27.5,0.0, 50.,0.014, 60.,0.05,... 100., 0.05 * FLT: temperature effect on growth FLT1=1. - 0.003*(TAV-30.)**2. FLT2=MAX(0. ,FLT1) FLT =MIN(FLT2,1. )

Appendix II

160

* 4. Dry matter (DM) partitioning * 4.1. Long-term pool * Assuming there is a long-term pool, which maintains cotton * growth when daily photosynthesis rate is insufficient * SCR daily DM taken from long-term pool kg.ha-1.d-1 * RGCHO dry matter from daily DM kg.ha-1.d-1 * RPUP rate of DM growth for long term pool kg.ha-1.d-1 * GTW daily dry matter kg.ha-1.d-1 * RPBO daily DM remained when supply bigger than demand, * stored in the long-term pool too * RWBL DM supply to boll growth kg.ha-1.d-1 * RWBT DM demand for boll growth kg.ha-1.d-1 * POOL long-term pool (DM reservoir) kg.ha-1 SCR =INSW(GTW-0.,0.2*POOL,0.1*POOL) GROWTH=0.7*GTW+SCR RPUP =0.3*GTW-SCR+RPBO RPBO =MAX(0.,(RWBL-RWBT)) POOL =INTGRL(ZERO,RPUP) * 4.2. Partitioning to organs * 4.2.1. Root * RGRA actual growth rate of root kg.ha-1.d-1 * RGRM maximum growth rate of root kg.ha-1.d-1 * ROOT total DM partitioning to root kg.ha-1 * ROOTA actual root weight kg.ha-1 RGRM =AFGEN(ROOTTB,PDT)*GTW RDROOT=AFGEN(RDTB,PDT) ROOT =INTGRL(IROOT,RGRM) RGRA =RGRM-RDROOT*ROOTA ROOTA =INTGRL(IROOT,RGRA) FUNCTION ROOTTB=0.,0.33, 2.5,0.33, 17.5,0.33, 27.5,0.0, 100.,0. FUNCTION RDTB =0.,0., 2.5,0., 17.5,0., 70.,0.02, 100.,0.02 * 4.2.2. Shoot * RGUP daily DM for shoot growth kg.ha-1.d-1 * BIOMUP total DM in shoot kg.ha-1 RGUP =GTW-RGRM BIOMUP=INTGRL(ISHOT,RGUP) * 4.2.3. Leaves, stems and fruits * RWSM daily DM for stem growth kg.ha-1.d-1 * RWLV daily DM for leaf growth kg.ha-1.d-1 * RWBL daily DM for fruit growth kg.ha-1.d-1 * WLEAF leaf weight kg.ha-1 * WSTEM stem weight kg.ha-1 * BOLLW total fruit weight including abscission kg.ha-1


161

* BOLL: fruits including square, flower and boll RWSM =AFGEN(FSMTB,PDT)*GROWTH FLV =AFGEN(FLVTB,PDT) RWLV =FLV*GROWTH RWLVA=RWLV-RLDR*WLEAF RWBL =AFGEN(FBLTB,PDT)*GROWTH FUNCTION FLVTB= 0.0,0., 2.5,0.65, 10.,0.7, 17.4999,0.5, 17.5,0.1,... 27.5,0.1, 40.0,0.10, 60.,0.0, 100.,0. FUNCTION FSMTB= 0.0,0., 2.5,0.35, 10.,0.3, 17.4999,0.5, 17.5,0.1,... 27.5,0.1, 40.0,0.00, 60.,0.0, 100.,0. FUNCTION FBLTB= 0.0,0., 2.5,0.00, 17.4999,0.0, 17.5,0.8,... 27.5,0.8, 40.0,0.90, 60.,1.0, 100.,1.0 WLEAF=INTGRL(IWLV ,RWLVA) WSTEM=INTGRL(ISTEM ,RWSM) * Dry matter calculated by measured partitioning index (as a check) PIROOT=1.-PISHOT PISHOT=-4E-05*PDT**2. + 0.0055*PDT + 0.7136 ASHOOT=TDRW*PISHOT WSHOOT=WLEAF+WSTEM+BOLLW * 5. Fruit growth, development and abscission * Modelled by using a fixed boxcar train method * 5.1. Fruit number * BOLOUT number of open bolls no.plant-1 * BOLTOT number of squares, flowers, bolls no.plant-1 * FALL1 abscission due to insufficient dry matter supply * FALL2 abscission due to pest * FALL3 abscission due to high temperature * RTEBOL RTE specially in cotton boll stage on * base temperature at 10 oC * STRBOL ratio of supply and demand * FALLBL total shed bolls * SQUARE number of squares no.plant-1 * FLOWER number of flowers no.plant-1 * SBOLL number of small bolls no.plant-1 * BBOLL number of big bolls no.plant-1 * GREENB number of green bolls (with certain yield) no.plant-1 GAMMA =GF/REAL(N) INFL =MAX(0.,RFNR) CBOLL(1:N) =MAX(0.,BOLL(1:N))/GAMMA FLOW(1) =INFL; ... FLOW(2:33) =CBOLL(1:32)*RTE; ...

Appendix II

162

FLOW(34:N+1)=CBOLL(33:N)*RTEBOL NETFLO(1:N) =FLOW(1:N)-FLOW(2:N+1) RBOLL(1:5) =NETFLO(1:5) -BOLL(1:5) *FALL2-BOLL(1:5) *FALL3; ... RBOLL(6:20) =NETFLO(6:20)-BOLL(6:20)*FALL1-BOLL(6:20)* ... FALL2-BOLL( 6:20)*FALL3;... RBOLL(21:25)=NETFLO(21:25)-BOLL(21:25)*FALL2-BOLL(21:25)*FALL3;... RBOLL(26:33)=NETFLO(26:33)-BOLL(26:33)*FALL1-BOLL(26:33)* ... FALL2-BOLL(26:33)*FALL3;... RBOLL(34:N) =NETFLO(34:N) -BOLL(34:N) *FALL2-BOLL(34:N) *FALL3 OUTFL =FLOW(N+1) RTEBOL=AFGEN(RTEBTB,TAV) FALL1 =AFGEN(FALLTB,STRBOL) FALL2 =AFGEN(FAL2TB,PDT) FALL3 =AFGEN(FAL3TB,TAV) FUNCTION RTEBTB=-10.,0., 0.,0., 10.,0., 22.,1., 35.,1. FUNCTION FALLTB= 0.,1., 1.,0. FUNCTION FAL2TB=-10.,0., 0.,0.,2.5,0., 27.5,0.,60.,0.001, 80.,0.,100.,0. FUNCTION FAL3TB=-10.,0., 0.,0., 30.,0., 45.,1. BOLL =INTGRL(BOLLI ,RBOLL) BOLOUT=INTGRL(BOLL0I,OUTFL) SQUARE=ARSUMM(BOLL, 1,22) FLOWER=ARSUMM(BOLL,23,25) SBOLL =ARSUMM(BOLL,26,33) BBOLL =ARSUMM(BOLL,34,74) GREENB=ARSUMM(BOLL,60,74) FALLBL=FN-ARSUMM(BOLL,1,N) * 5.2. Boll weight simulation * WBOLL weight of total fruits g.plant-1 * WBOUT weight of open bolls g.plant-1 * WBTOT weight of total fruit excluding WBOUT g.plant-1 * WSOUT weight of an open boll g.boll-1 * RWBAGE weight growth rate of boxes g.boll-1. * WBSUM weight of total fruits kg.ha-1 * WSQU weight of squares kg.ha-1 * WFLW weight of flowers kg.ha-1 * WSBO weight of small bolls kg.ha-1 * WBBO weight of big bolls kg.ha-1 * WGREEN weight of green bolls kg.ha-1 PARAM BOLLW0=0.019 WINFL =INFL*BOLLW0 CWBOLL(1:N) =MAX(0.,WBOLL(1:N))/GAMMA WFLOW(1) =WINFL; ... WFLOW(2:33) =CWBOLL(1:32)*RTE; ... WFLOW(34:N+1)=CWBOLL(33:N)*RTEBOL


163

WNETFL(1:N) =WFLOW(1:N)-WFLOW(2:N+1) WOUTFL =WFLOW(N+1)

RWBOL(1:5) =WNETFL(1:5) + BOLL(1:5)*RWBAGE(1:5)- ... BOLL(1:5)*FALL2*WBAGE(1:5)- ... BOLL(1:5)*FALL3*WBAGE(1:5); ... RWBOL(6:20) =WNETFL(6:20)+ BOLL(6:20)*RWBAGE(6:20)- ... BOLL(6:20)*FALL1*WBAGE(6:20)- ... BOLL(6:20)* FALL2*WBAGE(6:20)- ... BOLL( 6:20)*FALL3*WBAGE(6:20); ... RWBOL(21:25) =WNETFL(21:25)+ BOLL(21:25)*RWBAGE(21:25)- ... BOLL(21:25)*FALL2*WBAGE(21:25)- ... BOLL(21:25)*FALL3*WBAGE(21:25); ... RWBOL(26:33) =WNETFL(26:33)+ BOLL(26:33)*RWBAGE(26:33)- ... BOLL(26:33)*FALL1*WBAGE(26:33)- ... BOLL(26:33)*FALL2*WBAGE(26:33)- ... BOLL(26:33)*FALL3*WBAGE(26:33); ... RWBOL(34:N) =WNETFL(34:N)+BOLL(34:N)*RWBAGE(34:N)*STRESS(34:N)- ... BOLL(34:N) *FALL2*WBAGE(34:N)- ... BOLL(34:N) *FALL3*WBAGE(34:N) WBOLL =INTGRL(BOLLI,RWBOL) WBOUT =INTGRL(ZERO,WOUTFL) WSOUT =WBOUT/NOTNUL(BOLOUT) WSQU =ARSUMM(WBOLL,1,22)*PLANTS/1000. WFLO =ARSUMM(WBOLL,23,25)*PLANTS/1000. WSBO =ARSUMM(WBOLL,26,33)*PLANTS/1000. WBBO =ARSUMM(WBOLL,34,74)*PLANTS/1000. WGREEN=ARSUMM(WBOLL,60,74)*PLANTS/1000. BOLLW =WSQU+WFLO+WSBO+WBBO+WBOUT*PLANTS/1000. *5.3. Potential weight of single boll * AGE fruit age: square,1-22, flower, 23-25, small boll, 26-33, * big boll, 34-74, green boll, 60-74 * WBOLLP potential weight of single boll kg.ha-1 * WBMAX maximum weight of single boll, is a param of cultivar g * WBAGE weight of single boll for each age(1-74) * RWBT daily demand of total fruit growth kg.ha-1.d-1 * WBP total demand of total fruit growth kg.ha-1 AGE(1) = 1.;AGE(2) = 2.;AGE(3) = 3.;AGE(4) = 4.;AGE(5) = 5.;... AGE(6) = 6.;AGE(7) = 7.;AGE(8) = 8.;AGE(9) = 9.;AGE(10)=10.;... AGE(11) =11.;AGE(12)=12.;AGE(13)=13.;AGE(14)=14.;AGE(15)=15.;... AGE(16) =16.;AGE(17)=17.;AGE(18)=18.;AGE(19)=19.;AGE(20)=20.;... AGE(21) =21.;AGE(22)=22.;AGE(23)=23.;AGE(24)=24.;AGE(25)=25.;... AGE(26) =26.;AGE(27)=27.;AGE(28)=28.;AGE(29)=29.;AGE(30)=30.;... AGE(31) =31.;AGE(32)=32.;AGE(33)=33.;AGE(34)=34.;AGE(35)=35.;... AGE(36) =36.;AGE(37)=37.;AGE(38)=38.;AGE(39)=39.;AGE(40)=40.;... AGE(41) =41.;AGE(42)=42.;AGE(43)=43.;AGE(44)=44.;AGE(45)=45.;... AGE(46) =46.;AGE(47)=47.;AGE(48)=48.;AGE(49)=49.;AGE(50)=50.;... AGE(51) =51.;AGE(52)=52.;AGE(53)=53.;AGE(54)=54.;AGE(55)=55.;... AGE(56) =56.;AGE(57)=57.;AGE(58)=58.;AGE(59)=59.;AGE(60)=60.;... AGE(61) =61.;AGE(62)=62.;AGE(63)=63.;AGE(64)=64.;AGE(65)=65.;...

Appendix II

164

AGE(66) =66.;AGE(67)=67.;AGE(68)=68.;AGE(69)=69.;AGE(70)=70.;... AGE(71) =71.;AGE(72)=72.;AGE(73)=73.;... AGE(74:N)=REAL(N) PARAM RMBOLL=0.1143 PARAM WBMAX =7. WBAGE(1:N) =(WBMAX/(1.+219.*EXP(-RMBOLL*AGE(1:N)))) RWBAGE(1) =WBAGE(2)-WBAGE(1) ;... RWBAGE(2:N-1)=(WBAGE(3:N)-WBAGE(2:N-1));... RWBAGE(N) =0. RWB(1:N) =RWBAGE(1:N)*MAX(0.,BOLL(1:N))*PLANTS/1000. RWBT =ARSUMM(RWB,1,N) STRESS(1:33)=1.;... STRESS(34:N)=STRBOL * Yield bolls and weight including open bolls, green bolls YIELDW =(WGREEN+WBOUT) *PLANTS/1000. GREBL =GREENB*PLANTS * 5.4. Boll shell weight * Used for calculating seed cotton * SHELL: weight of boll shell for bolls at each age g.boll-1 SHELL(1:25) =0. ;... SHELL(26:55)=2.3/(1.+21753.5329*EXP(-0.28475*AGE(26:55)));... SHELL(56:N) =2.2195*EXP(-0.0038*AGE(56:N)) * 5.5 Seed cotton yield * SEDCOT seed cotton per plant g.plant-1 * COTOUT seed cotton of open bolls kg.ha-1 * COTGRE seed cotton of total green bolls kg.ha-1 * COTTON seed cotton yield kg.ha-1 * COTR ratio of seed cotton and shell SEDCOT =WBOLL-SHELL*BOLL COTOUT =(WBOUT-SHELL(N)*BOLOUT)*PLANTS/1000. COTGRE =ARSUMM(SEDCOT,60,74)*PLANTS/1000. COTTON =(COTGRE+COTOUT) COTR =MAX(0.,COTTON/NOTNUL(YIELDW)) * 5.6. Lint * GOTOUT lint percentage for open bolls * GOTGRN lint percentage for green bolls * LINT total lint yield kg.ha-1 * LINTO lint of open bolls kg.ha-1 * LINTG lint of green bolls kg.ha-1 * RLINT averaged lint percentage * LINTM lint of open bolls in Chinese mu kg.mu-1 * LINTT total lint in Chinese mu kg.mu-1 * (1 ha equals 15 mu)


165

PARAM GOTOUT=0.38; GOTGRN=0.34 LINTO =COTOUT*GOTOUT LINTG =COTGRE*GOTGRN LINT =LINTO+LINTG RLINT =LINT/NOTNUL(COTOUT+COTGRE) LINTM =LINTO/15. LINTT =LINT /15. * 5.7. Ratio of supply and demand * STRBOL: supply/demand ratio, 1., no stress, 0., maximum stress STRBL1 =MIN (1., RWBL/NOTNUL(RWBT)) STRBOL =INSW(PDT-17.5, 1., MAX(0.,STRBL1) ) * 6. Photosynthesis * 6.1. Leaf CO2 assimilation PARAM AMX=50. FUNCTION AMPDTT=0.0,0.0, 2.5,1.0, 60.,1.0, 70.,0.1, 100.,0. AMAX =AMX*AMPDT*AMTMP AMPDT =AFGEN(AMPDTT,PDT) AMTMP =MAX(0.0, 1.-0.003*(TAV-30.)**2.) * 6.2. Daily gross CO2 assimilation PARAM EFF =0.45 PARAM RHOS=0.2 PARAM KDF =1.0 PARAM SCP =0.2 CALL TOTASS (TIME, LAT , RDD, SCP,RHOS,AMAX, EFF, KDF, LAI,... DAYL, DTGA, FGROS,IABSD,BALNCE) * 6.3. Carbohydrate production GPHOT=DTGA*30./44. * 6.4. Maintenance MAINT=MINTS*MNTE*MNPDT MINTS=WSQU *0.038 + WFLO*0.076 + WLEAF*0.0264 + ROOTA*0.038 + ... WSTEM*0.006 + WSBO*0.038 + WBBO *0.032 + WBOUT*0.032 MNTE =2.**((TAV-28.0)/10.) MNPDT=AFGEN(MNPDTT, PDT) FUNCTION MNPDTT=0.0,0.0, 2.5,1.0, 27.5,0.9, 60.,0.5, 70.,0., 100., 0. * 6.5. Growth GTW =GPHOT - MAINT - GPHOT*(1.-1./ASRQ) ASRQ =AFGEN(ASRQTB, PDT) FUNCTION ASRQTB=0.,1.42, 2.5,1.42, 17.5,1.42, 27.5,1.64, 100.,1.64

Appendix II

166

* 6.6. Dry matter TDRW =INTGRL(IW, GTW) TRANSLATION_GENERAL DRIVER='EUDRIV' END PARAM INTERC=0. END PARAM INTERC=1.; FILM=1. STOP * 7. SUBROUTINES *----------------------------------------------------------------------* * SUBROUTINE ASTRO * * Purpose: This subroutine calculates astronomic daylength, * * diurnal radiation characteristics such as the daily * * integral of sine of solar elevation and solar constant. * * * * FORMAL PARAMETERS: (I=input,O=output,C=control,IN=init,T=time) * * name type meaning units class * * ---- ---- ------- ----- ----- * * DOY R4 Daynumber (Jan 1st = 1) - I * * LAT R4 Latitude of the site degrees I * * SC R4 Solar constant J.m-2.s-1 O * * SINLD R4 Seasonal offset of sine of solar height - O * * COSLD R4 Amplitude of sine of solar height - O * * DAYL R4 Astronomic daylength (base = 0 degrees) h O * * DSINBE R4 Daily total of effective solar height s O * * * *----------------------------------------------------------------------* SUBROUTINE ASTRO (DOY, LAT, SC , SINLD, COSLD, DAYL, DSINBE) IMPLICIT REAL (A-Z) *-----PI and conversion factor from degrees to radians PI = 3.141592654 RAD = PI/180. *-----declination of the sun as function of daynumber (DOY) DEC = -ASIN (SIN (23.45*RAD)*COS (2.*PI*(DOY+10.)/365.)) *-----SINLD, COSLD and AOB are intermediate variables SINLD = SIN (RAD*LAT)*SIN (DEC) COSLD = COS (RAD*LAT)*COS (DEC) AOB = SINLD/COSLD *-----daylength (DAYL) DAYL = 12.0*(1.+2.*ASIN (AOB)/PI) DSINBE = 3600.*(DAYL*(SINLD+0.4*(SINLD*SINLD+COSLD*COSLD*0.5))+ & 12.0*COSLD*(2.0+3.0*0.4*SINLD)*SQRT (1.-AOB*AOB)/PI) *-----solar constant (SC) and daily extraterrestrial radiation SC = 1370.*(1.+0.033*COS (2.*PI*DOY/365.)) RETURN END


167

*----------------------------------------------------------------------* * SUBROUTINE TOTASS * * Purpose: This subroutine calculates daily total gross * * assimilation (DTGA) by performing a Gaussian integration * * over time. At three different times of the day, * * radiation is computed and used to determine assimilation * * where after integration takes place * * * * FORMAL PARAMETERS: (I=input,O=output,C=control,IN=init,T=time) * * name type meaning units class * * ---- ---- ------- ----- ----- * * DOY R4 Daynumber (January 1 = 1) - I * * LAT R4 Latitude of the site degrees I * * DTR R4 Daily total of global radiation J.m-2.d-1 I * * SCP R4 Scattering coefficient of leaves for visible * * radiation (PAR) - I * * AMAX R4 Assimilation rate at light saturation I * * kg CO2. ha leaf-1.h-1 * * EFF R4 Initial light use efficiency kg CO2.J-1 I * * KDF R4 Extinction coefficient for diffuse light I * * LAI R4 Leaf area index ha.ha-1 I * * DAYL R4 Astronomic daylength (base = 0 degrees) h O * * DTGA R4 Daily total gross assimilation kg CO2.ha-1.d-1 O * * * SUBROUTINES and FUNCTIONS called : ASTRO, ASSIM * *----------------------------------------------------------------------* SUBROUTINE TOTASS (DOY, LAT , DTR, SCP,RHOS,AMAX, EFF, KDF, LAI, & DAYL, DTGA,FGROS,IABSD,BALNCE) IMPLICIT REAL(A-Z) REAL XGAUSS(3), WGAUSS(3) INTEGER I1, IGAUSS DATA IGAUSS /3/ DATA XGAUSS /0.112702, 0.500000, 0.887298/ DATA WGAUSS /0.277778, 0.444444, 0.277778/ PI = 3.141592654 CALL ASTRO(DOY,LAT,SC,SINLD,COSLD,DAYL,DSINBE) *---assimilation set to zero and three different times of the day (HOUR) DTGA = 0. IABSD= 0. DO 10 I1=1,IGAUSS *-------at the specified HOUR, radiation is computed and used to compute * assimilation HOUR = 12.0+DAYL*0.5*XGAUSS(I1) *-------sine of solar elevation SINB = MAX (0., SINLD+COSLD*COS (2.*PI*(HOUR+12.)/24.)) *-------diffuse light fraction (FRDF) from atmospheric * transmission (ATMTR) PAR = 0.5*DTR*SINB*(1.+0.4*SINB)/DSINBE ATMTR = PAR/(0.5*SC*SINB)

Appendix II

168

IF (ATMTR.LE.0.22) THEN FRDF = 1. ELSE IF (ATMTR.GT.0.22 .AND. ATMTR.LE.0.35) THEN FRDF = 1.-6.4*(ATMTR-0.22)**2 ELSE FRDF = 1.47-1.66*ATMTR END IF FRDF = MAX (FRDF, 0.15+0.85*(1.-EXP (-0.1/SINB))) *-------diffuse PAR (PARDF) and direct PAR (PARDR) PARDF = PAR * FRDF PARDR = PAR - PARDF CALL ASSIMS (SCP,RHOS,AMAX,EFF,KDF,LAI,SINB,PARDR,PARDF, & FGROS, IABS,BALNCE) *-------integration of assimilation rate to a daily total (DTGA) IABSD=IABSD + IABS*WGAUSS(I1) DTGA = DTGA + FGROS*WGAUSS(I1) 10 CONTINUE DTGA = DTGA * DAYL RETURN END *----------------------------------------------------------------------* * SUBROUTINE ASSIMS * * Purpose: This subroutine performs a Gaussian integration over * * depth of canopy by selecting three different LAI's and * * computing assimilation at these LAI levels. The * * integrated variable is FGROS. * * # Soil refection is included * * # with 5-point Gausssian integration * * * * FORMAL PARAMETERS: (I=input,O=output,C=control,IN=init,T=time) * * name type meaning units class * * ---- ---- ------- ----- ----- * * SCP R4 Scattering coefficient of leaves for visible * * radiation (PAR) - I * * RHOS * AMAX R4 Assimilation rate at light saturation I * * kg CO2.ha leaf-1.h-1 * * EFF R4 Initial light use efficiency kg CO2.J-1 I * * KDF R4 Extinction coefficient for leaves I * * LAI R4 Leaf area index ha.ha-2 I * * SINB R4 Sine of solar height - I * * PARDR R4 Instantaneous flux of direct radiation (PAR) W.m-2 I * * PARDF R4 Instantaneous flux of diffuse radiation(PAR) W.m-2 I * * FGROS R4 Instantaneous assimilation rate of O * * whole canopy kg CO2.ha soil-1.h-1 * *----------------------------------------------------------------------* SUBROUTINE ASSIMS (SCP,RHOS,AMAX,EFF,KDF,LAI,SINB, $ PARDR,PARDF, FGROS, IABS,BALNCE)


169

IMPLICIT REAL(A-Z) REAL XGAUSS(5), WGAUSS(5) INTEGER I1, I2, IGAUSS *-----Gauss weights for three point Gauss DATA IGAUSS /5/ DATA XGAUSS /0.0469101,0.2307534,0.5, 0.7692465,0.9530899/ DATA WGAUSS /0.1184635,0.2393144,0.2844444,0.2393144,0.1184635/ *-----reflection of horizontal and spherical leaf angle distribution SQV = SQRT(1.-SCP) REFH = (1.-SQV)/(1.+SQV) REFS = REFH*2./(1.+1.6*SINB) ETA1 = (REFH-RHOS)/(RHOS-1./REFH) ETA2 = (REFS-RHOS)/(RHOS-1./REFS) *-----extinction coefficient for direct radiation and total direct flux CLUSTF = KDF / (0.8*SQV) KBL = (0.5/SINB) * CLUSTF KDRT = KBL * SQV CORRV1 = ETA1*EXP(-2.*KDF *LAI) CORRV2 = ETA2*EXP(-2.*KDRT*LAI) DENOM1 = 1.+CORRV1 DENOM2 = 1.+CORRV2 *-----selection of depth of canopy, canopy assimilation is set to zero FGROS = 0. IABS = 0. DO 10 I1=1,IGAUSS LAIC = LAI * XGAUSS(I1) *--------absorbed fluxes per unit leaf area: diffuse flux, total direct * flux, direct component of direct flux. VISDF = (1.-REFH)*PARDF*KDF * & (EXP(-KDF*LAIC) + EXP(KDF*LAIC)*CORRV1/REFH)/DENOM1 VIST = (1.-REFS)*PARDR*KDRT * & (EXP(-KDRT*LAIC) + EXP(KDRT*LAIC)*CORRV2/REFS)/DENOM2 VISD = (1.-SCP) *PARDR*KBL *EXP (-KBL *LAIC) *--------absorbed flux (J/M2 leaf/s) for shaded leaves and assimilation of * shaded leaves VISSHD = VISDF + VIST - VISD IF (AMAX.GT.0.) THEN FGRSH = AMAX * (1.-EXP(-VISSHD*EFF/AMAX)) ELSE FGRSH = 0. END IF *--------direct flux absorbed by leaves perpendicular on direct beam and * assimilation of sunlit leaf area VISPP = (1.-SCP) * PARDR / SINB FGRSUN = 0. IABSUN = 0. DO 20 I2=1,IGAUSS VISSUN = VISSHD + VISPP * XGAUSS(I2)

Appendix II

170

IF (AMAX.GT.0.) THEN FGRS = AMAX * (1.-EXP(-VISSUN*EFF/AMAX)) ELSE FGRS = 0. END IF FGRSUN = FGRSUN + FGRS * WGAUSS(I2) IABSUN = IABSUN + VISSUN * WGAUSS(I2) 20 CONTINUE *--------fraction sunlit leaf area (FSLLA) and local assimilation * rate (FGL) FSLLA = CLUSTF * EXP(-KBL*LAIC) FGL = FSLLA * FGRSUN + (1.-FSLLA) * FGRSH IABSL = FSLLA * IABSUN + (1.-FSLLA) * VISSHD *--------integration of local assimilation rate to canopy * assimilation (FGROS) FGROS = FGROS + FGL * WGAUSS(I1) IABS = IABS + IABSL * WGAUSS(I1) 10 CONTINUE FGROS = FGROS * LAI IABS = IABS * LAI * Warning: in the expression for ISOIL * the use of REFS and REFH is not a mistake IREFL = PARDR*(REFS + CORRV2/REFS)/DENOM2 + & PARDF*(REFH + CORRV1/REFH)/DENOM1 ISOIL = (1.-REFS)*PARDR*(EXP(-KDRT*LAI)-EXP(KDRT*LAI) & *CORRV2/REFS)/DENOM2 + & (1.-REFH)*PARDF*(EXP(-KDF*LAI)-EXP(KDF*LAI) & *CORRV1/REFH)/DENOM1 ITOT = PARDF + PARDR BALNCE = 100. * (ITOT-IREFL-ISOIL-IABS)/ITOT RETURN END ENDJOB

171

Appendix III

List of variables used in the model SUCROS-Cotton Variable Description UnitAFGEN FST function for linear interpolation AGE

Fruit age: square, 1-22; flower, 23-25; small boll, 26-33; big boll, 34-74; green boll, 60-74

d after squaring

AMAX

Actual CO2 assimilation rate at light saturation for individual leaves

kg CO2 ha–1 leaf h–1

AMPDT

Factor accounting for effect of physiological development time (PDT) on AMX

-

AMPDTT Table of AMPDT as function of PDT -AMTMP Factor accounting for effect of daytime temperature on AMX -AMX

Potential CO2 assimilation rate at light saturation for individual leaves


AOB Intermediate variable -AROOT Actual root weight kg DM ha–1

ARRAY FST function for an array -ARRAY_SIZE FST function for array size -ARSUMM FST function for summation of array elements -ASHOOT Actual shoot weight kg DM ha–1

ASIN Fortran function arcsine -ASRQ

Factor accounting for physiological development temperature (PDT) on assimilate (CH2O) requirement for dry matter production

kg CH2O kg–1 DM

ASRQTB Table of ASRQ as function of PDT ASSIMS

Subroutine to perform a Gaussian integration over depth of a canopy by selecting three different LAIs and computing assimilation at these LAI levels.

ASTRO Subroutine to calculate day length ATMTR Atmospheric transmission coefficient -BALNCE Check -BBOLL Number of Big boll per plant no plant–1

BIOMUP Dry matter of shoot kg DM ha–1

BOLL

Array of fruit number, from first stages(1,square) to the last stage (74, boll at one day before open)

no plant–1

BOLL0 Initialization of number of open boll noBOLLI Initialization of number of fruit array noBOLLW

Array of fruit weight, from first stages(1,square) to the last stage (74, boll at one day before open)

BOLLW0 The initial weight of fruit g fruit–1

BOLTOT Total number of fruit per plant no plant–1

CBOLL Concentration of number of boll -CLUSTF Cluster factor -

Appendix III

172

CNTR FST variable for country name (in WEATHER) -CORRV1 Intermediate variable -CORRV2 Intermediate variable -COS Fortran function cosine -COSLD Intermediate variable in calculating solar height -COTGRE Seed cotton yield of green boll kg seedcotton ha–1

COTOUT Seed cotton yield of open boll kg seedcotton ha–1

COTR

Ratio of total seed cotton yield in total boll weight, another part is shell weight

-

COTT Seed cotton of total fruit from 1 to 74 stages kg seedcotton ha–1

COTTON Seed cotton yield including open boll and green boll kg seedcotton ha–1

CUT The day that cut the top of plant to stop growth dCUTOFF Switch of Cut, 0 is not cut , 1 is cut the top -CWBOLL Concentration of fruit weight -DAYL Day length h d–1

DEC Declination of the sun radiansDELT Time interval of integration dDENOM1 Intermediate variable -DENOM2 Intermediate variable -DLTB Table for effect of relative photoperiod in relation to day length -DOY Day number of year since 1 January (is day 1) dDSINBE

Integral of SINB over the day, with a correction for lower atmospheric transmission at lower solar elevations

s d–1

DTGA Daily total gross CO2 assimilation of the crop kg CO2 ha–1 ground d–1

DTR Daily solar radiation J m–2 d–1

DYNAMIC FST function for dynamic part of the simulation EFF

Initial light conversion factor for individual leaves

kg CO2 ha–1 leaf h–1 (J m–2 leaf s–1)–1

ENDJOB FST function ends the simulation ETA1 Intermediate variable -ETA2 Intermediate variable -EXP Fortran for exponential function FAL2TB

Table for effect of abscission because of pests in relation to PDT

-

FAL3TB

Table for effect of abscission in relation to high temperature (TAV)

-

FALL1 Abscission due to dry matter stress -FALL2 Abscission due to pest injury -FALL3 Abscission due to high temperature -FALLBL Total number of abscission fruit per plant no plant–1

FALLTB Table for effect of abscission in relation to dry matter stress -FBA Actual fruit branch number no plant–1

FBAD Differential fruit branch number no plant–1

FBAP Potential fruit branch number no plant–1

FBAV

Actual fruit branch number calculated according to the relation with leaf number

no plant–1

List of variables used in the model SUCROS-Cotton

173

FBAV1 Intermediate variable -FBLTB

Table of fraction of shoot dry matter allocated to fruit as function of PDT

-

FFB

The number of main stem leaf after which the first fruit branch initiates

-

FGL CO2 assimilation rate at one depth in the canopy kg CO2 ha–1 leaf h–1

FGROS Instantaneous canopy CO2 assimilation kg CO2 ha–1 ground h–1

FGRS

Intermediate variable for calculation of assimilation of sunlit leaves

-

FGRSH

CO2 assimilation rate at one depth in the canopy for shaded leaves


FILM Effect of film mulching, 0 is non film, 1 is with film -FINTIM FST variable for finish time of simulation FLOW Flow by which a state variable is change FLOWER Number of flower per plant no plant–1

FLT Effect of temperature -FLT1 Intermediate FLT2 Intermediate FLV Fraction of shoot dry matter allocated to leaves -FLVTB Table of FLV as function of PDT -FN Number of fruit node per plant no plant–1

FNAVM Maximum number of fruit node per plant no plant–1

FNTB Table of fruit node as a function of number of fruit branch -FRDF Fraction diffuse in incoming radiation -FSLLA Fraction of sunlit leaf area -FSMTB

Table of fraction of shoot dry matter allocated to stem as function of PDT

-

GAMMA Developmental width of fruit in boxcar train -GF Total Development stage of fruit in boxcar train -GI Initialization of G -G Actual development stage of fruit in boxcar train -GLAI Net growth rate of leaf area index ha leaf ha–1 ground d–1

GOTGRN Lint percentage of green boll -GOTOUT Lint percentage of open boll -GPHOT Daily total gross CH2O assimilation of the crop kg CH2O ha–1 ground d–1

GREBL Number of green boll per ha no ha–1

GREBLW Total weight of green boll per ha kg DM ha–1 groundGREENB Number of green boll per plant no plant–1 GROWTH

Available dry matter for daily growth related long term pool

kg DM ha–1 ground d–1

GTW

Gross growth rate of crop dry matter, including translocation

kg DM ha–1 ground d–1

HEIGHT Plant height cmHOUR Selected hour during the day hI1, I2 Do-loop counters -IABS Instantaneous radiation absorbed J m–2 ground s–1

Appendix III

174

IABSD Irradiation absorbed of a daily total J m–2 ground d–1

IABSL Instantaneous radiation absorbed per unit leaf area J m–2 leaf s–1

IABSUN

Direct flux absorbed by leaves perpendicular on direct beam and assimilation of sunlit leaf area

J m–2 leaf s–1

IE

Coefficient for air temperature compensation due to soil temperature increase under film mulching

-

IGAUSS Do-loop counter -ILAI Initial leaf area index ha leaf ha–1 groundINFL Inflow in the boxcar INITIAL FST function for initialization INSW FST function for input switch INTERC

Wheat-cotton intercropping system (1: intercropping; 0: monoculture)

INTGRL FST function for integration IREFL Reflected flux by leaves J m–2 s–1

IROOT Initialization of root dry matter kg DM ha–1

ISHOT Initialization of shoot dry matter kg DM ha–1

ISOIL Reflected flux by soil J m–2 s–1

ISTEM Initialization of stem dry matter kg DM ha–1

ISTN FST variable for station number (in WEATHER) -ITOT Instantaneous total radiation J m–2 s–1

IW Initialization of total dry matter kg DM ha–1

IWBOLL Initialization of fruit dry matter kg DM ha–1

IWLV Initialization of stem dry matter kg DM ha–1

IYEAR FST variable for initial year (in WEATHER) -KBL

Extinction coefficient for direct component of direct PAR flux

ha ground ha–1 leaf

KDF Extinction coefficient for leaves ha ground ha–1 leafKDRT Extinction coefficient for total direct PAR flux ha ground ha–1 leafLAI Leaf area index ha leaf ha–1 groundLAIC Leaf area index above selected height in canopy ha leaf ha–1 groundLAT Latitude of the weather station degreesLDTB Table effect of daylength as a function of daylength -LEAFA Actual leaf number no plant–1

LEAFD Differential leaf number no plant–1

LEAFM Maximum leaf number after plant cut no plant–1

LEAFP Potential leaf number no plant–1

LINT Lint yield per ha kg lint ha–1

LINTG Lint yield per ha contributed by green boll kg lint ha–1

LINTM Lint yield of open boll per Chinese mu, 1 ha=15 mu kg lint mu–1

LINTO Lint yield per ha contributed by open boll kg lint ha–1

LINTT Total Lint yield per mu including green and open boll kg lint mu–1

MAINT Maintenance respiration rate of the crop kg CH2O ha–1 d–1

MAX Fortran function to choose maximum value -MIN Fortran function to choose minimum value -MINTS Maintenance rate of plant in different organs kg CH2O ha–1 d–1


175

MNPDT Effect of Maintenance rate on related to PDT -MNPDTT Table of effect to maintenance as a function of PDT -MNTE Effect of maintenance rate on related to temperature -N Array size NETFLO Net inflow to boxcar NOTNUL FST function to avoid division by zero -OPENBL Open boll number per ha no ha–1

OPENBW Open boll dry matter per ha kg DM ha–1

OUTFL Out flow of boxcar -PAR Instantaneous flux of photosynthetically active radiation J m–2 ground s–1

PARDF Instantaneous diffuse flux of incoming PAR J m–2 ground s–1

PARDR Instantaneous direct flux of incoming PAR J m–2 ground s–1

PDT

Physiological development time: Physiological day under optimal condition

-

PI Ratio of circumference to diameter of circle -PIROOT Partitioning index for total dry matter allocated to root -PISHOT Partitioning index for total dry matter allocated to shoot -PLANTS Plant density plantPOOL Long-term pool kg DM ha–1

PRDEL FST variable for printing interval PRINT FST function for output RAD Factor to convert degrees to radians radians degree–1

RBOLL The rate of development fruit array in boxcar -RCGRE Ratio of seed cotton to dry matter of green boll -RCOUT Ratio of seed cotton to dry matter of open boll -RDD Total daily global radiation (from WEATHER) J m–2 d–1

RDROOT Relative root decease rate d–1

RDTB Table of RDROOT as a function of PDT -REFH Reflection coefficient for diffuse PAR -REFS Reflection coefficient for direct PAR -RFBR Initiate rate of fruit branch no plant–1 d–1

RFBR1 Intermediate variable -RFBR2 Intermediate variable -RFEE Relative thermal effectiveness under film mulching -RFNR Initiate rate of fruit node no plant–1 d–1

RFNR1 Intermediate variable RFNR2 Intermediate variable RGRA Actual rate of root dry matter growth related to RDROOT kg DM ha–1 d–1

RGRM Potential rate of root dry matter growth kg DM ha–1 d–1

RGUP Ra te of shoot dry matter growth kg DM ha–1 d–1

RHOS Fraction of soil reflection of radiation flux -RHR Rate of plant height growth cm d–1

RHR1 Intermediate variable -RHR2 Intermediate variable -RLDR Relative LAI decease rate d–1

RLDR1 Leaf decease rate no plant–1 d–1

Appendix III

176

RLDRTB Table of RLDR as a function of PDT -RLINT Average lint percentage -RLR Actual rate of leaf number development no plant–1 d–1 RLR1 Intermediate variable RLR2 Intermediate variable RLR3 Intermediate variable RMBOLL Relative growth rate of fruit to the age of fruit age–1

ROOT Potential root weight related to RGRM kg DM ha–1

ROOTA Actual root weight related to RGRA kg DM ha–1

ROOTTB Table of RGRM as a function of PDT -RPBO

The dry matter accumulation due to supply is higher than demand of fruit organs

kg DM ha–1 d–1

RPDT Rate of PDT development PDT d–1

RPE Relative photoperiod effectiveness -RPUP Rate of long term pool growth kg DM ha–1 d–1

RTE Relative thermal effectiveness -RTEBOL RTE effect in boll maturing stage in the boxcar -RTEBTB Table of RTEBOL as a function of temperature -RTEE RTE before emergence -RTEE1 RTE related soil temperature without film -RTEFSB Table of RTEE1 as a function of soil temperature -RTESTB

Table of RFEE as a function of soil temperature with film mulching

-

RTETB Table of RTE after emergence as a function of air temperature -RWB Array of dry matter demand of fruit kg DM d–1

RWBAGE Rate of fruit dry matter growth against fruit age g DM age–1

RWBL Rate of fruit dry matter growth kg DM ha–1 d–1

RWBOL Array of fruit dry matter growth rate -RWBT Daily dry matter demand of total fruit growth -RWLV Rate of leaf dry matter growth kg DM ha–1 d–1

RWSM Rate of stem dry matter growth kg DM ha–1 d–1

SBOLL Number of small bolls per plant no plant–1

SC

Solar constant, corrected for varying distances between sun-earth

J m–2 s–1

SCP Scattering coefficient of leaves for PAR SCR Rate of dry matter used from long term pool kg DM ha–1 d–1

SEDCOT Array of seed cotton dry matter against fruit array g plant–1

SHELL Array of shell dry matter against fruit array g plant–1

SIN Fortran function for sine -SINB Sine of solar elevation -SINLD Intermediate variable in calculating solar declination -SLA Specific leaf area m2 (ha) leaf g–1 leafSLATB Table for relationship SLA and PDT SQRT Fortran function for square root SQUARE Number of squares per plant no plant–1

SQV Intermediate variable in calculation of reflection coefficient -


177

STOP FST variable to stop the simulation STRBL1 Intermediate variable STRBOL

Ratio of dry matter supply and demand for fruit growth and development

-

STRESS Array of stress of dry matter boll filling -STTIME FST variable to start the simulation TAV Average air temperature degree CTAV1

Average air temperature plus compensation effectiveness due to film mulching

degree C

TDRW Total dry matter kg DM ha–1

TETMP Intermediate variable in calculation of IE -TIME FST variable of time -TIMER FST function for time -TIN

Daily average temperature in wheat-cotton intercropping systems

degree C

TIPLUS

The compensation to air temperature due to film mulching and increase of soil temperature in wheat-cotton intercropping systems

degree C

TISAV Soil temperature in wheat-cotton intercropping systems degree CTISCAV

Soil temperature with covering plastic film in wheat-cotton intercropping systems

degree C

TMMN Daily minimum temperature (from WEATHER) degree CTMMN1 Daily minimum temperature under film mulching degree CTMMX Daily maximum temperature (from WEATHER) degree CTMMX1 Daily maximum temperature under film mulching degree CTMIN

Daily maximum temperature in wheat-cotton intercropping systems

degree C

TNIN

Daily minimum temperature in wheat-cotton intercropping systems

degree C

TOTASS

FORTRAN subroutine to calculate gross CO2 assimilation of the crop

-

TPLUS

The compensation to air temperature due to film mulching and increase of soil temperature

degree C

TSAV Average soil temperature degree CTSCAV Average soil temperature under film mulching degree CVI Varity maturity index -VISD

Absorbed direct component of direct flux per unit leaf area (at depth LAIC)

J m–2 leaf s–1

VISDF Absorbed diffuse flux per unit leaf area (at depth LAIC) J m–2 leaf s–1

VISPP Absorbed light flux by leaves perpendicular on direct beam J m–2 leaf s–1

VISSHD

Total absorbed flux for shaded leaves per unit leaf area (at depth LAIC)

J m–2 leaf s–1

VISSUN

Total absorbed flux for sunlit leaves in one of three Gauss point classes

J m–2 leaf s–1

VIST Absorbed total direct flux per unit leaf area (at depth LAIC) J m–2 leaf s–1

WBAGE Array of single fruit weight g DM fruit–1

Appendix III

178

WBBO Weight of open boll kg DM ha–1

WBMAX Maximum single boll weight g DM fruit–1

WBOLL Array of fruit weight in boxcar g DM plant–1

WBOLLP Array of potential weight of fruit in boxcar g DM plant–1

WBOUT Weight of open boll per plant g DM plant–1

WBP Total dry matter demand kg DM ha–1

WBSUM Total dry matter of fruit kg DM ha–1

WBTOT Total dry matter of fruit except open boll g DM plant–1

WEATHER FST function to call weather subroutine -WFLO Weight of flowers per plant g DM plant–1

WFLOW Flow of fruit dry matter in boxcar WGAUSS Array containing weights to be assigned to Gauss points -WGREEN Weight of green boll kg DM ha–1

WINFL Inflow of fruit dry matter in boxcar WLEAF Total dry matter of leaves kg DM ha–1

WNETFL Net flow of fruit dry matter in boxcar WOUTFL Out flow of fruit dry matter in boxcar WSBO Total dry matter of small bolls kg DM ha–1

WSOUT Total dry matter of open bolls kg DM ha–1

WSQU Total dry matter of squares kg DM ha–1

WSTEM Total dry matter of stem kg DM ha–1

WTRDIR FST variable for weather directory -XGAUSS Array containing Gauss points -YIELDB Total number of boll those contribute to yield no fruit ha–1

YIELDW Total dry matter of boll those contribute to yield kg DM ha–1

ZERO Initial value of zero in an integration

179

Summary Cotton occupies a crucial position in the national economy and the livelihood of many Chinese farmers. China accounts for 22 percent of the world’s population but it has access to only 7 percent of the world’s arable land; thus, the best arable land is used intensively. From the early 1980s onwards, farmers in the Yellow River cotton producing region started to intercrop cotton and winter wheat on more than 60% of total cotton area. In wheat-cotton intercropping, relay and strip intercropping systems are commonly used to grow both wheat and cotton on the same field in one year. In these systems, strips of winter wheat sown in the fall, are intersown with cotton in the spring. From April to June, two crops are grown together on one field, with the seedling phase of cotton and the maturation phase of wheat overlapping in time and space. After the wheat harvest in early summer, the whole space is occupied by cotton. In farmers’ practice several intercropping patterns are used. They are named after the numbers of rows of wheat and cotton that are alternated, e.g. the 3:2 system is an intercropping system consisting of 3 rows of wheat and 2 rows of cotton alternating. Other systems in use include the 3:1, 4:2 and 6:4 patterns. In addition, some variability in row distances and strip width exists. This study aimed at analysing productivity and resource use efficiency of cotton wheat relay intercropping systems. The growth, phenology, resource capture, use efficiencies, productivity and quality in intercrops were determined at field level and compared to those of monocultures of wheat and cotton. The field experiments were conducted at the Cotton Research Institute at Anyang, Henan Province, China. An ecophysiological-based simulation model of cotton growth, SUCROS-Cotton, was developed to help understand system performance and explore opportunities for optimization at the system level. We found that the land equivalence ratio (LER) ranged from 1.28 in the 6:2 intercropping system to 1.39 in the 3:1, 3:2 and 4:2 systems. All intercropping systems thus showed a substantial advantage in land productivity, compared to monoculture. The 3:1 system gave the highest wheat yield (79% of monoculture), followed by the 6:2 (73%), 3:2 (70%), and 4:2 (70%) systems. The wheat yields were closely related to the row length density of wheat, homogenized over the whole system. Cotton lint yields were highest in the 3:2 and 4:2 intercrops (69% and 68% of monoculture, respectively), which was significantly lower than in monoculture but significantly higher than in the 3:1 (58%) and 6:2 (54%) systems. The lower cotton yield in intercrops, compared to monoculture, was associated with a delay in development and retarded growth of cotton seedlings due to shading by the taller wheat. Surprisingly,

Summary

180

this delay in intercrops did not affect cotton fiber quality. Growth patterns of dry mass in monocultures and intercrops were quantified by fitting expolinear growth equations. The parameters indicated a growth delay of 6-12 days in intercropped cotton, compared to the monoculture. The developmental delay of cotton is an ecophysiological ‘bottle neck’ to productivity in wheat-cotton intercropping systems. The delay in the duration from sowing to first square of intercropped cotton amounted to 115 ‘degree-days’ (base-temperature of 12 °C) or 4.7 ‘physiological days’ (days with optimal temperature conditions) in comparison to sole cotton, due to a temperature decrease at the level of the seedlings of several degrees, due to shading during the intercropping period. The formation of reproductive cotton organs was therefore delayed. Fruit set in the intercropped cotton was reduced: by 1-3 branches and 5-20 nodes at the time of ‘cut out’. Among the intercrops, there were no differences in phenology. A soil cover with plastic film increased soil temperature at 5 cm soil depth by 2.7 °C on average; thus, applying a film cover is a way to accelerate crop development and seedling growth. From explorations with the cotton growth model SUCROS-Cotton, it was concluded that the effect of a film cover can eliminate the development gap between intercropped and monocropped cotton. Light use efficiencies (LUE) ranged from 1.9 to 2.3 g DM MJ–1 (PAR) for wheat during the reproductive period and from 1.3 to 1.4 g DM MJ–1 (PAR) for cotton calculated for the whole growing season. LUEs of both crops were not affected by intercropping. Wheat monocrops intercepted 618 MJ m–2 photosynthetically active radiation (PAR) from Mar-18 to harvest in 2002. Averaged over three years, wheat in the four intercrops (3:1, 3:2, 4:2 and 6:2, respectively) intercepted 83, 71, 73 and 75% as much PAR as the sole wheat. At a spatial scale, the strip intercropped wheat captured on a daily base about 20% more light than monocropped, comparing to row length density. Cotton monocrops intercepted on average 444 MJ m–2 PAR from sowing to harvest in three years. Cotton in the four intercrops (3:1, 3:2, 4:2 and 6:2, respectively) intercepted 73, 93, 86 and 67% as much PAR as the sole cotton. Thus, the total light interception was substantially increased in the intercropping systems, in comparison to sole crops. The internal N use-efficiency (IE), expressed as kg grain (wheat) or kg lint (cotton) per kg N uptake, ranged in the intercropping systems from 24 to 31 kg kg–1 for wheat and from 4.5 to 10 kg kg–1 for cotton. The IE of wheat was similar in intercropping systems and in the monoculture (26 to 28 kg kg–1); but the IE of cotton was lower for intercrops than in monoculture (9 to 10 kg kg–1). N-uptake per unit land area by wheat was lower in intercrops than in the monocrop, but the uptake per unit biomass was

Summary

181

similar in monocrops and intercrops. The N-uptake per unit land area by cotton was also lower in intercrops, but the uptake per unit biomass was higher in intercropped cotton, as dry matter production was reduced more by intercropping than N-uptake. The relative total nitrogen uptake (RNT), which was calculated in a similar way as LER, but based on N-uptake rather than yield, ranged from 1.4 to 1.7; thus, RNT was higher than LER, demonstrating that intercrops utilized N comparatively less efficient than land area. The conventional N-management in intercrops results in high N-surpluses that can cause an environmental risk. The N management could be improved by means of a demand-based rate and timing of N applications. Besides nitrogen water is an important resource for cotton production. It was estimated that water productivity (WP) of intercrops was 26% and 41% lower than in the monocultures of wheat and cotton, respectively. The lower WP for wheat in the intercropping systems was due to a ‘partial’ use of the land area, but a ‘full’ supply of water. The lower WP of cotton in the intercropping systems was mainly due to decreased lint yields. To integrate the effects of environment, genotypic traits and management practices on the phenology and crop performance of cotton the model SUCROS-Cotton was developed. The evaluation of SUCROS-Cotton for intercrops showed a RMSE value of less than 2.9 days (3.4% of the observed) for phenology, indicating that the model SUCROS-Cotton was able to account for environmental influences on development of cotton in monoculture as well as intercrops. Thus, it can be used as a research tool to optimize the system management for relay intercropped cotton. The findings of this study suggest that there are prospects to improve cotton wheat relay intercropping systems when taking into account the temporal and spatial environment by management interactions. The most promising management practices for accelerating crop development are: soil cover and ridge-furrow cultivation. In relation to resource use efficiency and sustainability of cotton wheat intercropping systems it was shown that major improvements can be made in fine-tuning the nitrogen management. The study may also give guidance to define improved traits in cotton breeding programs and to prioritize future research activities.

183

小结

棉花是关系中国国计民生的主要经济作物。中国人多地少，人口占世界总人口的

22%，可耕地却只占世界的 7%。因此，中国的可耕地利用集约程度很高。自 1980 年

以来，中国黄河流域棉区逐步实行了麦棉两熟种植方式。目前 60%以上的棉田都是冬

小麦跟棉花套种两熟。麦棉两熟在中国是一种广泛推广应用的周年全田冬小麦和棉花

带状间套种植的耕作制度。在麦棉两熟耕作制度中，冬小麦在秋天带状种植，并预留

棉行；棉花在次年春天套种到预留棉行中。小麦和棉花的共生期从 4 月到 6 月共约 7

周左右。共生期后，套种棉花利用全部土地和资源。农民普遍采用的套种模式有 3:1

式、3:2 式、4:2 式和 6:2 式。( 数字分别表示小麦和棉花的行数，如 3:2 式为 3 行小麦

和 2 行棉花的配置模式 ) 。同一种套种模式下，行距和带宽也存在一定差异。

本研究旨在研究分析麦棉两熟小麦和棉花的生产力和资源利用效率。通过与单作

小麦和棉花对照相比较的大田试验，研究两种套种作物的生长、发育、资源截获量、

利用效率、产量和品质等。田间试验在中国河南省安阳市的中国农业科学院棉花研究

所试验地进行。同时，还研究建立了一个基于生理生态学的棉花生长发育模拟模型

SUCROS-Cotton，在系统水平上分析和优化麦棉两熟的配置。

研究表明，麦棉两熟的土地利用率(Land equivalent ratio, LER)在 6:2 式种植条件

下为 1.28，在 3:1 式、3:2 式和 4:2 式中皆为 1.39。和一熟种植比较，所有的参试套

种模式都表现出较强的土地生产优势。3:1 式的小麦产量最高，相当于单作小麦的

79%。依次为 6:2 式、3:2 式和 4:2 式，分别为单作小麦的 73%、70%和 70%。小麦

产量取决于小麦全田平均的行长密度(row length density，RLD)。棉花皮棉产量以 3:2

式和 4:2 式最高，分别为单作棉花的 69%和 68%，显著低于纯作棉花，但显著高于

3:1 式（58%）和 6:2 式（54%）。麦棉两熟的皮棉产量较单作棉花低，主要是由套种

后棉花生育期延迟和共生期小麦对棉苗遮荫造成的生长缓慢所导致的。与普遍认同的

观点不同，麦棉套种不影响棉花的纤维品质。

套种和单作棉花的干物质生长可以用指数-线性生长模型 (expolinear model) 来定

量。对模型参数的拟合表明，套种棉花的生长比单作棉花延缓 6-12 天。

在麦棉两熟种植条件下，棉花的生育期延迟是限制生产潜力提高的生理生态学瓶

颈。生育延迟发生在棉花播种到现蕾这一时期。用大于 12℃有效积温表示，套种棉花

的生育过程较单作棉花延迟 115 度·日；用生理发育时间表示，则延迟 4.7 个生理发育

小结

184

日数。生理发育日数指棉花在最适温度条件下所需要的发育天数。和单作棉花相比，

套种棉花的生育延迟主要是棉苗在共生期受到小麦的遮荫造成冠层温度降低引起的。

由于生育延迟，套种棉花的成铃模式也相应较单作棉花迟缓。到棉花打顶，套作棉花

比单作少 1-3 台果枝和 5-20 个果节。但是, 几种套种模式的生育期没有显著差异。套

种棉田覆盖塑料地膜能提高土壤温度。地膜覆盖 5 cm 地温平均提高 2.7℃。因此，地

膜覆盖是一项解决套种棉花迟发晚熟的关键农艺措施。应用 SUCROS-Cotton 对地膜

覆盖效应的分析表明，地膜覆盖可弥补套种引起的生育期延缓。

研究表明小麦和棉花的光能利用率(LUE)不受套种影响。光能利用率小麦在生殖生

长阶段为 1.9-2.3 g DM MJ-1(PAR),棉花在整个生育期为 1.3-1.4 g DM MJ-1(PAR)。

单作小麦在 2002 年 3 月 18 日到收获这一时期共截获光合有效辐射（PAR）618

MJ m-2。三年平均，3:1 式、3:2 式、4:2 式和 6:2 式小麦的 PAR 截获量分别为单作小

麦的 83%、71%、73%和 75%。在空间层面上，带状套种小麦比单作小麦平均每天在

相同 RLD 条件下多截获 20%的 PAR。单作棉花三年平均全生育期截获 PAR 444 MJ

m-2。3:1 式、3:2 式、4:2 式和 6:2 式棉花的 PAR 截获量分别为单作棉花的 73%、

93%、86%和 67%。周年全田麦棉套种比单作小麦和棉花的光截获量显著增多。

麦棉套种条件下，N 肥内在生理利用效率 IE（每吸收 1 kg 纯氮生产的 kg 小麦籽

粒或皮棉产量）小麦为 24-31 kg kg-1，棉花为 4.5-10 kg kg-1。套种小麦的 IE 与单作

小麦（26-28 kg kg-1）相当。但套种棉花的 IE 却低于单作棉花（9-10 kg kg-1）。套种

小麦在单位土地面积上吸收的总氮量低于单作小麦，但单位干物质吸收的总氮量与单

作小麦相同。套种棉花在单位土地面积上吸收的总氮量也低于单作棉花，但是，单位

干物质吸收的总氮量高于单作棉花。套种棉花干物质产量降低程度比吸收量大。几种

麦棉套种模式的相对氮素吸收量 RNT（与 LER 的计算相似）为 1.4-1.7，明显高于

LER，表明麦棉套种的氮素利用率比土地利用率低。常规的氮肥管理导致了大量的氮

素过剩，增加了环境负荷。根据作物的实际需求改善施氮量和施氮时间可以提高麦棉

套种的氮肥利用率。

除了氮素营养外，水分也是影响麦棉套种的重要因素。麦棉套种的作物水分生产

率（WP）比单作小麦和棉花分别低 26%和 41%。套种小麦 WP 的降低是因为套种模

式只是部分利用土地资源，灌溉却是针对整个土地面积。套种棉花 WP 降低的原因主

要是因为皮棉产量降低。

综合环境效应、遗传特征和农艺措施，研究建立了棉花生长发育的模拟模型

SUCROS-Cotton。对模型的评估表明，麦棉两熟的生育期观察值和模拟值的均方差根

小结

185

RMSE 小于 2.9 天，误差只相当于观察值的 3.4%。因此，SUCROS-Cotton 模型能较

好地模拟环境因子对套种和单作棉花生育期的影响，可作为一个有效的研究工具用于

优化麦棉套种配置和管理。研究发现，考虑环境因子的时空特征及与田间管理的互作效应，能够改良和提高麦棉两熟种植制度。地膜覆盖和垅作种植可缓解麦棉套种的迟发晚熟，精细氮肥管理能够有效提高麦棉两熟的资源利用效率和可持续发展。研究结果还有助于棉花遗传改良和确定将来的研究重点。

187

Samenvatting Katoen is een belangrijk gewas voor de economie van China en voor het inkomen van de Chinese boer. In China woont thans 22% van de wereldbevolking, maar het areaal voor akkergewassen bedraagt slechts 7% in vergelijking tot de rest van de wereld. Sinds de jaren tachtig wordt in het stroomgebied van de Gele Rivier katoen geteeld in een volgmengteelt met tarwe. De tarwe wordt in het najaar gezaaid in stroken met enkele gewasrijen, afgewisseld door niet-ingezaaide stroken, waarin in april van het volgende jaar katoen gezaaid wordt. De korrelvulling- en afrijpingfase van tarwe en de begingroei en -ontwikkeling van katoen vallen samen; in dit teeltsysteem concurreren de twee gewassen gedurende een periode van ca zeven weken – van april tot juni – om ruimte. In juni wordt de tarwe geoogst, en van juni tot oktober heeft het katoengewas het hele veld beschikbaar. Het telen van tarwe en katoen in een jaarcyclus op hetzelfde perceel is in de praktijk ontstaan. De agro-ecologische mogelijkheden en beperkingen van dit mengteeltsysteem zijn nog onvoldoende bekend.

In dit proefschrift wordt de productiviteit en de benutting van zonlicht, water en stikstof in enkele volgmengteeltsystemen van tarwe en katoen onderzocht. In het onderzoek uitgevoerd op het Nationale Onderzoeksinstituut voor Katoen in Anyang, China werden de groei, ontwikkeling, benutting van groeifactoren, gebruiksefficiëntie, productiviteit en kwaliteit bepaald in veldexperimenten. Daarbij werd een vergelijking gemaakt van vier systemen die verschillen in het aantal gewasrijen van tarwe en van katoen, namelijk 3:1 (3 rijen tarwe tegen 1 rij katoen), 3:2, 4:2 en 6:2, met de twee monocultures: tarwe en katoen. Alle vier teeltsystemen komen in de praktijk voor. Naast de variatie in het aantal rijen per gewas verschillende de mengteeltsystemen ook in de ruimte tussen de tarwestroken. Deze bedraagt 40 cm in het 3:1 systeem, 60 cm in het 3:2 systeem, 70 cm in het 4:2 systeem, en 80 cm in het 6:2 systeem. Bij tarwe is de rijafstand binnen een strook in alle systemen 20 cm. Ten gevolge van variaties in het aantal gewasrijen en de breedte van de vrijgehouden strook voor de katoen, verschillen de systemen in geaggregeerde rijdichtheid. Deze wordt gemeten als het aantal rijen per meter dwarsdoorsnede over het hele systeem. De geaggregeerde rijdichtheid van tarwe bedraagt dan 3 m–1 in het 3:1 en het 6:2 systeem, 2.5 m–1 in het 3:2 systeem, 2.67 m–1 in het 4:2 systeem, en 5 m–1 in de monocultuur. De geaggregeerde rijdichtheid van de katoenrijen bedraagt 1 m–1 in het 3:1 en 6:2 systeem, 1.67 m–1 in het 3:2 systeem, 1.33 in het 4:2 systeem, en 1.25 in de monocultuur.

Allereerst werd de productiviteit van de verschillende systemen gekarakteriseerd door uit proeven de Land Equivalent Ratio (LER) te berekenen. Deze parameter geeft weer welke oppervlakte aan monocultures van de samenstellende gewassen van een

Samenvatting

188

mengcultuur nodig zou zijn om dezelfde opbrengst te verkrijgen als van een eenheid mengcultuur. De LER wordt berekend als:

w,i c,i

w,m c,m

LERY YY Y

= +

waar Yw,i de korrelopbrengst is van tarwe in mengteelt, Yw,m de opbrengst van tarwe in monocultuur, en waar Yc,i en Yc,m de vezelopbrengst van katoen in mengteelt en monocultuur is. De LER bedroeg 1.28 in het 6:2 systeem, en 1.39 in de overige drie mengteeltsystemen. Alle systemen gaven dus een grote winst in landgebruikefficiëntie t.o.v. monocultuur. De hoogste tarweopbrengst werd gehaald in de monocultuur; de opbrengst van het 3:1 systeem bedroeg 79% van de monocultuur, gevolgd door 6:2 (73%), 3:2 en 4:2 (beide 70%). De tarweopbrengst in alle systemen hield nauw verband met de geaggregeerde rijdichtheid. De opbrengst van katoenvezel, gemeten t.o.v. monocultuur, was het hoogst in de 3:2 en 4:2 systemen (69 en 68% t.o.v. monocultuur), wat significant lager was dan in monocultuur, maar hoger dan in de 3:1 (58%) en 6:2 (54%) systemen.

De begingroei van katoen werd goed beschreven met een expolineaire vergelijking. Uit de berekende parameters kon een vertraging in de drogestoftoename van katoen met 6 tot 12 dagen in mengteelten t.o.v. de monocultuur afgeleid worden. De lagere opbrengst van katoen in mengcultuur was enerzijds het gevolg van een tragere beginontwikkeling, die toe te schrijven is aan beschaduwing door tarwe, en anderzijds van de geringere lichtonderschepping in de brede stroken na de tarweoogst. Het uiteindelijke gevolg was een latere vruchtvorming, 115 graaddagen, en een reductie van het aantal vruchten (bolvorming). Als gevolg van deze effecten daalde in de mengteelten de oogstindex van de katoen, d.w.z. de hoeveelheid katoenvezel per eenheid totale biomassa. Tussen mengteeltsystemen waren de verschillen in ontwikkeling van de katoen echter niet significant. Het afdekken van de grond met doorzichtig plastic folie resulteerde in een verhoging van de bodemtemperatuur van 2 à 3 ºC. Uit modelberekeningen bleek dat de toepassing van folie perspectief biedt om de vertraagde ontwikkeling en groei van katoen in mengteelt te compenseren.

De lichtonderschepping in monocultures en mengteelten werd geschat met een wiskundig model voor rijgewassen. De lichtonderschepping door tarwe in de mengteelten 3:1, 3:2, 4:2 en 6:2 bedroeg ten opzichte van de monocultuur resp. 83, 71, 73 en 75% van de inkomende straling. Voor katoen bedroegen deze percentages respectievelijk 73, 93, 86 en 67%. De berekende percentages lichtonderschepping laten voor tarwe een goed verband zien met de gemeten opbrengsten in de mengteelten. Voor beide gewassen is de lichtonderschepping in de mengteelt substantieel gereduceerd t.o.v. de monocultuur. De lichtgebruiksefficiëntie, d.w.z. de

Samenvatting

189

verhouding tussen opbrengst en lichtonderschepping, was hoger voor tarwe dan katoen, maar er werden zowel voor tarwe als voor katoen geen verschillen tussen mengteelt en monocultuur aangetoond. De verschillen in productiviteit tussen mengteelt en monocultuur konden dus volledig worden toegeschreven aan verschillen in lichtonderschepping.

Vanuit het oogpunt van ecologische duurzaamheid en kwaliteit van bodem en grondwater is het stikstofgebruik belangrijk. De verhouding tussen productie en opname van stikstof uit de bodem werd in de tarwe niet beïnvloed door mengteelt, maar in de katoen werd een lagere stikstofgebruiksefficiëntie gevonden in mengteelt dan in monocultuur. De lagere gebruiksefficiëntie van stikstof in katoen in mengteelt komt doordat de opbrengst in sterkere mate wordt verlaagd dan de totale biomassa, terwijl de stikstofopname vooral verband houdt met de biomassaproductie. De lage stikstofgebruiksefficiëntie van katoen in mengteelt is daarom een rechtstreeks gevolg van de lagere oogstindex, welke weer samenhangt met de vertraagde ontwikkeling in de beginfase en het lagere aantal vruchten. Voor de mengteeltsystemen werd analoog met het concept van LER een relatieve totale stikstof opbrengst (RNT) berekend. De hogere waardes van deze parameter, tussen 1,4 en 1,7 in verschillende systemen, ten opzichte van de LER wijzen ook in de richting van een lagere gebruiksefficiëntie van de stikstof in mengteelt t.o.v. monocultuur.

Naast stikstof is water een belangrijke groeifactor. In mengteelten lag de waterproductiviteit (kg opbrengst per kg water) van tarwe 26% lager en van katoen 41% lager dan in monocultuur. De lagere efficiëntie van watergebruik door tarwe is het gevolg van de onvolledige bedekking van het grondoppervlak in een volgmengteelt, bij gelijkblijvende watergift door irrigatie en regenval. De lagere gebruiksefficiëntie van water door katoen wordt vooral verklaard door de lagere lintopbrengsten in mengteelt.

Voor de integratie van effecten van klimaatsfactoren, genotype en management op de gewasontwikkeling en -opbrengst is een gewasgroeimodel voor katoen, SUCROS-COTTON, ontwikkeld. Dit model is ontwikkeld volgens principes en concepten van de Wageningse groep van SUCROS-modellen en gekalibreerd en gevalideerd voor Chinese katoenrassen en groeiomstandigheden. Het model verschilt van andere modellen voor katoen door een relatief eenvoudige code. Met het model wordt beoogd opties te verkennen om de groei en ontwikkeling van katoen in volgmengteeltsystemen te optimaliseren.

De analyses in dit proefschrift geven aan dat de volgmengteelt van tarwe en katoen leidt tot een efficiënter landgebruik, wat voor China van groot belang is. Tevens geven analyses aan dat de mengteelt voor de Chinese boer voor een brede band van prijzen voor katoen en tarwe economisch profijtelijk is. Wat het gebruik van stikstof en water

Samenvatting

190

betreft is de mengteelt echter meer verspillend dan de monocultuur. Er worden verschillende mogelijkheden aangegeven waarmee deze gebruikefficiënties, o.m. voor licht en stikstof, in tarwe-katoensystemen verder geoptimaliseerd zouden kunnen worden. Door katoen op ruggen te telen zou de efficiëntie van het watergebruik door tarwe kunnen verbeteren. Tevens zou de begingroei van katoenplanten verbeterd worden door hogere bodem- en luchttemperaturen als gevolg van een betere lichtonderschepping. Gebruik van plastic folie is een andere mogelijkheid om de bottleneck van de trage begingroei van katoen te doorbreken. De lage efficiëntie van gebruik van stikstof in mengteelt kan waarschijnlijk verbeterd worden door in de dosering rekening te houden met de lagere relatieve behoefte van gewassen in volgmengteeltsystemen.

191

List of publications of the author Submitted papers in this thesis Zhang, L., van der Werf, W., Zhang, S., Li, B., Spiertz, J. H. J., 2007. Growth, yield

and quality of wheat and cotton in relay strip intercropping systems. Field Crops Research 103, 178-188.

Zhang, L., Spiertz, J. H. J., Zhang, S., Li, B., van der Werf, W., 2007. Nitrogen economy in relay intercropping systems of wheat and cotton. Plant and Soil (accepted)

Zhang, L., van der Werf, W., Zhang, S., Li, B., Spiertz, J. H. J., 2007. Cotton development and temperature dynamics in relay intercropping with wheat. Field Crops Research (submitted)

Zhang, L., van der Werf, W., Bastiaans, L., Zhang, S., Li, B., Spiertz, J.H.J., 2007. Light interception and radiation use efficiency in relay intercrops of wheat and cotton. Field Crops Research (submitted)

Zhang, L., van der Werf, W., Cao, W., Li, B., Spiertz, J. H. J., 2007. Development and validation of SUCROS-Cotton: A mechanistic crop growth simulation model for cotton, applied to Chinese cropping conditions. Agricultural Systems (submitted).

Recent publications (after 2001) Zhang, L., Li, B., Yan, G., Van der Werf, W., Spiertz, J.H.J., Zhang, S., 2006.

Genotype and planting density effects on rooting traits and yield in cotton (Gossypium hirsutum L.). Journal of Integrative Plant Biology 48, 1287-1293.

Zhang, L., Cao, W., Zhang, S., Zhou, Z., 2005. A simulation model for boll growth, development and abscission in cotton. Scientia Agricultura Sinica 31, 70-76 (in Chinese with English abstract).

Zhang, L., Cao, W., Zhang, S., Zhou, Z., 2005. Characterizing root growth and spatial distribution in cotton. Acta Phytoecologia Sinica 29, 226-273 (in Chinese with English abstract).

Zhang, L., Cao, W., Zhang, S., Zhou, Z., 2004. A simulation model for morphogenesis and LAI in cotton. Cotton Science 16, 77-83 (in Chinese with English abstract).

Zhang, L., Cao, W., Zhang, S., 2004. Dynamic simulation on dry matter partitioning and yield formation in cotton. Scientia Agricultura Sinica 37, 1621-1627 (in Chinese with English abstract).

Zhang. L., 2004. Quantitative analysis of agricultural household response to ratio of grain and cotton price change in China. Cotton Science 16, 49-54.

List of publications

192

Zhang, L., Cao, W., Zhang, S., Luo, W., 2003. Simulation model for cotton development stages based on physiological development time. Cotton Science 15, 97-103 (in Chinese with English abstract).

Zhang, L., Cao, W., Zhang, S., Luo, W., 2003. A process model of photosynthetic production and dry matter accumulation in cotton. Cotton Science 15, 138-145 (in Chinese with English abstract).

Long, T., Zhang, L., 2006. Study and application of GIS based crop cultivation expert system. Computer Engineering 32, 176-178 (in Chinese).

Meng, Y., Wang, Y., Wang, L., Chen, B., Zhang, L., Shu, H., Zhou, Z., 2006. Effect of the composite root population of wheat-cotton intercropping system on cotton root growth. Scientia Agricultura Sinica 39, 2228-2236 (in Chinese with English abstract).

Long, T., Zhang, L., 2005. The design of a cotton production management system based on GIS. China Cotton 32, 4-6 (in Chinese).

Wang, L., Meng, Y., Zhou, Z., Zhang, L., Chen, B., Bian, H., Zhang, S., Wang, Y., 2005. Temporal and spatial dynamic distribution of cotton-wheat composite root system under condition of cotton-wheat double cropping system. Acta Agronomica Sinica 31, 888-896 (in Chinese with English abstract).

Meng, Y., Wang, L., Zhou, Z., Wang, Y., Zhang, L., Bian, H., Zhang, S., Chen, B., 2005. Dynamics of soil enzyme activity and nutrient content in intercropped cotton rhizosphere and non-rhizosphere. Chinese Journal of Applied Ecology 16, 2076-2080 (in Chinese with English abstract).

Meng, Y., Wang, L., Zhou, Z., Chen, B., Wang, Y., Zhang, L., Bian, H., Zhang, S., 2005. Effect of the composite root population of wheat-cotton double cropping system on soil enzyme activity and soil nutrient content at the cotton rhizosphere and non-rhizosphere zones. Scientia Agricultura Sinica 38, 904-910 (in Chinese with English abstract).

Ma, F., Cao, W., Zhang, L., Zhu, Y., Li, S., Zhou, Z., Li, C., Xu, L., 2005. A physiological development time-based simulation model for cotton development stages and square and boll formation. Chinese Journal of Applied Ecology 16, 626-630 (in Chinese with English abstract).

Zhang, H., Zhu, Y., Cao, W., Zhou, Z., Zhang, L., 2004. A dynamic knowledge model for nitrogen and water management of cotton. Chinese Journal of Applied Ecology 15, 777-781 (in Chinese with English abstract).

Zhou, Z., Meng, Y., Chen, B., Zhang, L., Sun, X., Wang, L., Shi, P., 2004. Effect of shading in wheat-cotton double cropping symbiotic period on photosynthetic performance of leaves during cotton seedling stage. Scientia Agricultura Sinica 37, 825-831 (in Chinese with English abstract).

List of publications

193

Zhang, H., Zhu, Y., Cao, W., Zhang, L., 2003. A dynamic knowledge model for yield target and yield components in cotton. Cotton Science 15, 279-283 (in Chinese with English abstract).

Zhang, H., Cao, W., Zhou, Z., Zhu, Y., Zhang, L., 2003. A dynamic knowledge model for optimal LAI in cotton. Cotton Science 15, 151-154 (in Chinese with English abstract).

Yang, C., Zhang, L., Lin, E., 2003. Study on photosynthetic active radiation (PAR) and soil temperature of monoculture cotton and intercropping of wheat-cotton plantation. Cotton Science 15, 193-194 (in Chinese with English abstract).

Li, Y., Xu, H., Zhang, L., Wang, J., Miao, Y., Yang, Z., 2002. A study on cotton root development and distribution of different cotton varieties. Acta Agriculturae Boreali-Sinica 17, 109-113 (in Chinese with English abstract).

Yang, C., Li, Y., Lin, E., Zhang, L., 2001. Differences of transpiration and root distribution among cotton varieties under different planting types. Cotton Science 13, 372-376 (in Chinese with English abstract).

6W:2C

6 rows wheat

2 rows cotton

Sole cotton

3W:1C

3 rows wheat 1 row cotton

195

PE&RC PhD Education Certificate

With the educational activities listed below the PhD candidate has complied with the educational requirements set by the C.T. de Wit Graduate School for Production Ecology and Resource Conservation (PE&RC), which comprises a minimum total of 32 ECTS (= 22 weeks of activities) Review of Literature (4.2 credits)

- Cropping systems, resource capture and use efficiency (2002) Writing of Project Proposal (5 credits)

- Productivity and resource use in cotton and wheat intercropping systems (2002) Post-Graduate Courses (4.2 credits)

- Agro-ecological approaches for rural development; PE&RC (2001) - Advanced statistics; PE&RC (2001) - How to manage diversity in living systems; PE&RC (2002)

Deficiency, Refresh, Brush-up and General Courses (11.4 credits)

- Basic statistics; PE&RC (2001) - Simulation of ecological processes; PPS&BFS (2002) - Farm household economics; GDE (2002) - Simulation of crop growth; PPS (2002)

Competence Strengthening / Skills Courses (1.4 credits)

- English; PE&RC (2002) Discussion Groups / Local Seminars and other Scientific Meetings (2.8 credits)

- Discussion group of crop and weed ecology (2001-2007) - CRI seminars (2002-2007)

PE&RC Annual Meetings, Seminars and the PE&RC Weekend (1.5 credits)

- PE&RC conference: Agriculture and nature (2001) - PE&RC day: Food insecurity (2001) - Winter school: Functional biodiversity for sustainable crop protection (2001)

International Symposia, Workshops and Conferences (3.4 credits)

- Symposia on cotton, 2004 and 2005; China (2004) - The 2006 beltwide cotton conference; San Antonio, Texas, USA (2006)

3W:2C

2 rows cotton 3 rows wheat

4W:2C

2 rows cotton 4 rows wheat

197

Curriculum vitae Lizhen Zhang was born in Jingtai county, Gansu province, China on December 30, 1967. He graduated from China Agricultural University (CAU) and received his BSc degree majoring in Agronomy in 1989. He worked at the Cotton Research Institute (CRI) of the Chinese Academy of Agricultural Sciences (CAAS) from 1989 up to now. During the first 7 years, he worked in the Division of Scientific Management, and became a vice director of the division during the last 2 years. From 1995, he became vice director of the Agronomy Department of CRI, and started his research on cotton cultivation, water saving and cropping systems as a leader of the Cotton Physiology and Ecology Research group. He became an associate professor of CRI in 2000. In 1998, he enrolled in the Graduate School of CAAS as a part-time MSc-student; after finishing all courses, he enrolled as a part-time PhD-student at Nanjing Agricultural University (NAU) and gained his PhD degree in 2003. The topic of the PhD was: simulation of growth and development in monocropping cotton. In 2004, he worked together with the Agricultural Meteorology Department of the College of Agricultural Resources and Environmental Sciences, CAU as a part-time cooperation researcher, and taught a course on advanced agricultural meteorology for MSc students. In 2001, he got an opportunity to start a sandwich-PhD programme at the Crop and Weed Ecology group (CWE) of Wageningen University, under auspices of the C.T. de Wit Graduate School for Production Ecology and Resource Conservation (PE&RC). Since 2001, he has carried out field experiments at CRI, China, and took courses, analysed data, and wrote the papers at Wageningen University.

198

Funding The work presented in this thesis was financially supported by the sandwich-PhD programme of Wageningen University. Further support was provided by the Program for Changjiang Scholars and Innovative Research Teams in Universities (IRT0412), and by the ‘948’ project of Chinese Ministry of Agriculture. Printing of the thesis was financially supported by ‘Stichting Fonds Landbouw Export Bureau 1916-1918’ (LEB Foundation, Wageningen, The Netherlands).

Cotton and Wheat Inter Cropping - Ph.D. Thesis.

Documents

Transcript of Cotton and Wheat Inter Cropping - Ph.D. Thesis.