Explore on Farm: Adoption of GAP for Wheat

in North Africa

TABLE OF CONTENTS

Page List of Authors………………………………………………………………………………..iii Forward……………………………………………………………………………………….vi Agriculture in the North Africa Region and Constraints to Sustainable Productivity M. El Mourid, E. De-Paw and F. Taher ]……………………………………………………1 Climate Risk Analysis and Crop Management Decision Making Tools H. Benaouda, M. El Mourid, M. Boutfirass…………………………………………………10 Constraints to Cereal-Based Rainfed Cropping in North Africa and Methods to Measure and Minimize their Effects H. Gomez-Macpherson, A.F. Van Herwaarden, H.M. Rawson …………………………… 14

Why do any of your research on-farm? H. M. Rawson ……………………………………………………………………………30

Optimizing Variety x Sowing Date for the Farm M. Boutfirass……………………………………………………………………………38

Optimizing Plant Population, Crop emergence, Establishment and Sowing Rate M. Boutfirass, M. Karrou………………………………………………………………..50

Optimizing Tillage Systems On-Farm H. Boulal, R. Mrabet……………………………………………………………………..60

Optimizing Nitrogen Use on the Farm S. Rezgui, M. Fakhfakh…………………………………………………………………37

Cropping Sequencing

K. El Mejahed, O. Zaghouane, F. Djennadi………………………………………………….86

i

List of Authors

Helena Gómez Macpherson FAO Crop and Grassland Service Viale delle Terme di Caracalla 00100 Rome Italy helena.gomezmacpherson@fao.org Howard M. Rawson 28 Bambridge St Weetangera ACT 2614 Australia rawson@netspeed.com.au Fawzi Taher Regional Crop Production Officer FAO Regional Office for the Near East P.O. Box 2223 Cairo, Egypt Phone: ++202-3316000 (ext. 2815) Fax : ++202-7495981 E-mail : fawzi.taher@fao.org Mohammed El Mourid Regional Coordinator North Africa Regional Program ICARDA - Tunis n° 1 rue des Oliviers - El Menzah V - 2037 Tunis - Tunisia Post Office Box : 435 El Menzah 1- 1004 Tunis - Tunisia Phone : ++ 216 71 752 134/099 Fax : ++ 216 71 753 170 GSM : ++ 216 98 464 104 secretariat@icarda.org.tnm.elmourid@cgiar.org

iii

Eddy De-Pauw Head, Geographic Information Systems Unit International Center for Agricultural Research in the Dry Areas (ICARDA) P.O.Box 5466, Aleppo, SYRIA Tel. +963-21-2213433 Fax +963-21-2213490 E.DE-PAUW@CGIAR.ORG Hassan Benaouda hbenaouda_inra@yahoo.fr Mohamed Boutfirass boutfirass@hotmail.com Hakim Boulal h.boulal@caramail.com Rachid Mrabet Head of Centre Régional de la Recherche Agronomique de Meknès BP 578 Meknès 50000 Maroc tél: +21255300243 fax: +21255300244 mobile: +21261430768 mrabet.site.voila.fr Mohamed Karrou karrou@inra.org.ma K. EL Mejahed Fadhila Djennadi Omar Zaghouane Institut Technique des Grandes Cultures- ITGC BP16 El Harrach Algiers, Algeria Tel : ++ 213 21 524432/ 524431/521691 (Direct) Fax : ++ 213 21 523529/521691 zaghouaneo@yahoo.fr

iv

Salah Rezgui Maître de Conference Institut National Agronomique de Tunis 43, Avenue Charles Nicole, Tunis 1082 Phone : ++ 216 71 289 431 Fax : ++ 216 71 799 391 GSM : ++ 216 96 513 598 rezgui.salah@iresa.agrinet.tnsalahrezgui08@yahoo.fr Moez Fakhfakh Centre Technique des Céréales Bousalem Tel : ++ 216 78 602 964 fakhfakhmoez@yahoo.fr

v

Forward

The FAO Regional Office for the Near East wishes to acknowledge the cooperation and partnership between FAO and ICARDA for the preparation of this guideline on “Explore on Farm for Adaptation and Adopting of GAP for wheat in North Africa”. This work aims mainly at improving the sustainable production of rainfed wheat based farming systems through increasing the understanding of the effects of localized environmental factors on crop and varietal performance. This guideline is intended to be written to agricultural researchers, extension officers and farmers to stimulate a cycle of learning in order to work on farm by conducting on farm trails. The above guideline will bring farmers and facilitators together, on farm to explore agronomic options through guided and collaborative crop trails. These guidelines at the end will improve management of wheat cropping systems and thus sustainable yield in North Africa Countries mainly Morocco, Algeria, Tunisia and Libya. These guidelines comprise nine chapters addressing the following topics:

• Agriculture in the North Africa Region and Constraints to Sustainable Productivity.

• Climate Risk Analysis and Crop Management Decision Making Tools. • Constraints to Cereal-Based Rainfed Cropping in North Africa and Methods to

Measure and Minimize their Effects. • Why do any of your research on-farm? • Optimizing Variety x Sowing Date for the Farm. • Optimizing Plant Population, Crop emergence, Establishment and Sowing Rate. • Optimizing Tillage Systems On-Farm. • Optimizing Nitrogen Use on the Farm. • Cropping Sequencing.

The FAO Regional Office for the Near East would like to thank all authors; Helena Gómez Macpherson, Howard M. Rawson, Fawzi Taher, Mohammed El Mourid, Eddy De-Pauw, Hassan Benaouda, Mohamed Boutfirass, Hakim Boulal, Rachid Mrabet, Mohamed Karrou, K. EL Mejahed, Fadhila Djennadi, Omar Zaghouane, Salah Rezgui, and Moez Fakhfakh; for their excellent contributions. These guidelines have been translated into Arabic through funds from FAO language implementation programme by a good team composed of Dr. Youssef Hamdi, Chairman of the Egyptian Centre of Organic Agriculture, Dr. Said Abdel Maksoud, Professor, Agricultural Economics, Agric. Institute Economics and Master Facilitator of the Institute of Cultural Affairs, and Dr. Mosad Abdel Aleem, Head, Wheat Research Department, Field Crops Agric. Research Institute, Agric. Research Centre.

vi

Special thanks and appreciation go to the excellent work of Ms. Helena Gomez Macpherson (former Agriculture Officer, Crop and Grassland Service, FAO) and Mr. Howard M. Rawson (Australia) for their major contribution in preparing this work and who previously prepared a similar work on this subject in rainfed Mediterranean areas and Central Asia environments. Thanks and appreciations go to Mr. Fawzi Taher (Crop Production Officer at FAO Regional Office for the Near East) and Mr. Mohamed ElMourid (ICARDA Regional Coordinator for North Africa) who conceived and made possible the preparation and release of this publication. We hope that this work which is presented in both languages (Arabic and English) will be of interest to persons concerned in the Arab countries and else where. Mohamad Albraithen ADG/Regional Representative, RNE

vii

Agriculture in the North Africa Region and constraints to sustainable productivity Agricultural production in North Africa is constrained not only by its harsh dry environment and shallow soils but also by a diversity of sociological, country and global pressures. This chapter depicts the shifting kaleidoscope of these highly complex interacting factors as a background to the Explore on farm approach.

The Regional Setting Climate The countries of North Africa are characterized by diverse but generally dry climates, in which evaporation exceeds precipitation. Most of the region is arid (88%) with aridity tending to increase with distance from the coast (Table 1). Rainfall is concentrated in winter from October-April.

Within an overall Mediterranean climatic pattern, the climates of the region show great diversity. Moisture and temperature conditions can differ markedly as a result of differences in local topography, exposure to maritime influence or nearness to desert areas. In fact, areas with the typical Mediterranean climate only occur in a fairly narrow belt around the Mediterranean Sea, surrounded by vast areas with quite different Mediterranean zone is evidenced by the map of agroclimatic zones, prepared in accordance with the UNESCO classification (Fig.1). Table 1. Extent of aridity in North Africa

Country Hyper-arid

Arid

Semi-arid

Sub-humid

Humid

Square km

Algeria 71.7 15.9 8.6 3.8 0.0 2,381,741

Egypt 91.5 8.5 0.0 0.0 0.0 997,739

Libya 80.9 17.6 1.5 0.0 0.0 1,757,000

Morocco (without Sahara)

0.0 41.8 55.5 2.7 0.0 458,730

Tunisia 14.4 52.3 30.3 3.0 0.0 164,150

Source: ICARDA Geographical Systems Laboratory

Excluding the hyper-arid zones, which cover the driest deserts and have no potential for agriculture, less than 30% of the region, or about 1,700,000 sq km, is ‘dryland’. These are the areas with some potential for either dryland farming (sub-humid and semi-arid zones) or for extensive rangeland agriculture (arid zone).

Owing to the high degree of aridity and variability of rainfall in most of North Africa, agriculture is particularly vulnerable to drought. Fig. 2 shows the large amplitude of inter-annual precipitation variations for Tunis, which is typical for the region.

Severe drought is common in North Africa.

Fig. 1. Agroclimatic zones of North Africa. Data source: ICARDA Geographical Systems Laboratory 1

Prolonged droughts were reported in Morocco in 1979-84 and 1994-95. In Tunisia droughts occurred recently in 1988-89 and 1994-95. Starting in 1998 and continuing into 2000, North Africa experienced the worst regional drought in decades. As shown by Figure.3, using the case of the barley crop, the region's crop production has fluctuated widely in recent years (Figure 3 excludes Egypt. There much of the crop is irrigated).

The causes of drought in North Africa are complex and still poorly understood. Most of the precipitation in the region is generated as a result of depressions which are steered southward from the North Atlantic during blocking episodes by mid-latitude high-pressure cells. Variations in precipitation are possibly linked to fluctuations of the North Atlantic Oscillation (NAO), a large-scale mode of climate variability in the northern

hemisphere. However, unlike the El Nino

Southern Oscillation (ENSO), this phenomenon can as yet not be used as a predictor of rainfall patterns for the next growing season.

Fig. 2. Inter-annual precipitation fluctuations for Tunis 1878-1990. Source FAOCLIM database.

Soils Dry climatic conditions and calcareous parent materials are the main factors controlling soil formation in the region. The Calcisols are a typical product of a climatic pattern characterized by precipitation concentrated within one season and a major annual precipitation deficit. In addition, the common calcareous parent materials, such as limestones, marls and calcareous sandstones, predominate in this region and impart some of their characteristics to the various soil types.

A significant feature of this region is the relative abundance of stony and shallow soils, especially in the more arid parts and the mountainous areas. The general aridity, in conjunction with factors related to topography, is also responsible for the widespread occurrence of soils with poor or no profile development (Regosols), and dark cracking clays (Vertisols), and for a general shortage of prime agricultural soils.

Saline soils occur in the more arid parts of the region where they are associated with specific topographical positions (depressions without an external drainage outlet), geological formations (e.g., gypsum outcrops) and irrigated agriculture.

Soil textures are usually heavy, with clay and silt dominating the fine fraction. In drier areas and on parent materials derived from marls, soils with high silt content may be common. These soils tend to have poor structural properties, which makes them vulnerable to water and wind erosion and compaction under heavy machinery use. Mobile sands are common in the desert areas in the form of sand sheets and dunes, as a product of wind erosion.

Fig. 3. Relative fluctuations in barley production in the Maghreb countries, Algeria, Morocco and Tunisia, 1961-2001. Source: FAOSTAT, 2002

Despite their heavy textures, the soils of the region are generally well drained. The high base content ensures a high nutrient retention capacity. Organic matter levels are highly variable and site-specific, but tend at a regional scale to be low and deplete rapidly under constant use.

Agriculture In North Africa, most agriculture takes place within “less-favored areas”. These areas are significantly constrained by moisture stress due to low and highly variable rainfall, extreme temperatures, short cropping season, shallow

2

soils, soil nutrient depletion, steep slopes and by socioeconomic factors and lack of infrastructure.

The crops and livestock Wheat and barley represent the main components of annual rainfed cropping systems, although crops such as cotton, especially in Egypt under irrigation, are also important. Food legumes: faba beans, chickpeas, and lentils also represent major components of annual crop rotations.

Other annual crops such as potatoes, summer crops (melons), oilseeds and sugar beet, are gaining in importance, especially where irrigation is possible. Perennials include olive, almond, fig and pistachio trees and fodder trees.

Livestock, small ruminants in particular, represent a major component of the farming system. Due to its extensive nature, livestock production mainly relies upon grazing on communal lands. These lands scarcely provide the minimum nutrient requirements because they can be overstocked and degraded. By and large, the agricultural area devoted to forage and fodder crops remains inadequate, showing no sign of growth in spite of a real potential.

Difficult climatic and edaphic conditions for crop and livestock production are reflected in the low and unstable productivity levels of rainfed production systems. In the future, because arable land and water are in short supply, sustainable increases in agricultural production will have to come from sustainable increases in productivity per unit area, in particular through a significant improvement in water-use-efficiency at the farm level (more crop per drop).

The combination of high population growth (more than 2%), negative trade balance, and growing food deficits are exerting increasing pressure on government treasuries, hence precipitating the need for a significant and rapid increase of food production. Over the next two decades, agricultural development of the region will be confronted with an enormous challenge calling for significant increases in agricultural production to ensure significant and sustainable poverty reduction, both rural and urban, while preventing further degradation of the natural resource base. Significant investments in agricultural research, technology transfer, and extension will be therefore required to enable the NARSs of the region to meet the challenge. To further complicate matters, such crucial investment needs have come at a time of economic difficulties,

which have already led to cuts, in real terms, in domestic research and extension budgets exacerbated by donors’ fatigue and retrenchment. Moreover, the tremendous challenge ahead will have to be met with less resource in particular less land and much less water as growing urbanization and a dynamic industrial sector are increasingly diverting the supply of these resources from agriculture, not to mention a shrinking rural labor force increasingly attracted by the “urban dream”.

Rangelands, small herders and pastoralists A considerable proportion of the land in North Africa is considered extensive rangelands. The rangelands are of great economic value to the small herders who depend on them for their livelihood. A generation ago, the native rangeland pasture vegetation provided most of the feed needs of the small ruminant population but today that feed is inadequate. The absolute level of this feed source is steadily falling as a result of overgrazing, removal of vegetation through plowing and/or for fuel wood, and soil erosion and desert encroachment.

Small herders and pastoralists are often forced to aim for short-term requirements using production strategies that are destructive and unsustainable in the longer term. Inappropriate land use policies and the absence of secure property and access rights on rangelands in many countries have exacerbated the problem. Moreover, traditional local institutions governing access to grazing lands have been disrupted, resulting in a system of “open access”, with no regulatory mechanism to control the extent and intensity of grazing. To make things worse, pastures and pastoral resources are increasingly being appropriated by individuals or encroached upon by herders and farming communities. As a result of the degradation and shrinking of common pastures, livestock owners are increasingly relying on crop residues and purchased feeds.

In the recent past, small herders relied a great deal on institutional and market access-options of feed resources to complement rangeland grazing. Now institutional access-options are eroding. They used to provide small herders with access to free or highly subsidized feed during critical times. In addition to its detrimental effect on the natural resource base, there is evidence that the decline in the rangeland productivity is contributing to rural poverty and out-migration.

At the region level, the steady demand for livestock products provides clear market

3

opportunities. Opportunities are coupled to consumer preferences for high quality local livestock products, especially fresh meat of specific local breeds, which have fostered the increase of livestock population, However, marketing of livestock products is characterized by a lack of discriminatory labeling . This does not enable producers to capture the value-added potential. High transportation costs, poor storage facilities, and shortage of effective veterinary services also diminish the potential.

Food security and investment in agriculture Over the past two decades the world has gone through dramatic changes that have set the path for economic globalization, leading to significant political, social, and economic implications throughout the planet. Moreover, the economic and institutional reforms stemming from structural adjustment programs have shaped a whole new environment characterized by the global, and often abrupt, opening of markets. In the wake of such fundamental changes, agricultural development is emerging as the strategic asset for overall economic growth and poverty alleviation in North Africa. As a result of a relatively low crop productivity, and lack of income generating opportunities (on farm as well as off-farm), especially in the less-favored areas where most of the resource-poor smallholders live, household food security remains a serious concern in the region. This in turn is exerting tremendous pressure on most governments of the region to increase food production, at least to maintain politically acceptable levels of food security.

In spite of all these challenges, most of the countries in the region have experienced a steady decline in public investment in agricultural development even though agricultural development remained a priority in the political will. For example public finances directed to agriculture in Egypt have declined from 20% of the state budget in the 1960s to only 10% in the mid 1980s. Morocco witnessed the same trend decreasing from 30% of public funds in the 1960-70s to 15% in 1990s.

Sustainable food security is a fundamental objective of all the countries in North Africa where, may be more than in any other region, it is linked to grain production (Belaïd, 2002). Indeed, the consumption-production gap for cereals in most of the countries has widened and the region is now a net importer of cereals and even the world’s leading durum wheat importer (50% of

the world market). Production of wheat and particularly durum wheat declined in most countries in the region during the last 30 years: Algeria fell by 41 kg per capita, Morocco by 53 kg. The same trend is observed for food legumes for which all the Maghreb countries became net importers in the 1990s whereas they were world exporters in the 1980s. Table 2 shows imports of wheat by selected countries.

Table 2. Trends in total wheat and durum wheat imports in selected countries of WANA (1992-2001).

Bread wheat imports (000 tons)

Year Egypt Algeria Morocco

1992 5807 3399 1552

1993 6004 3643 2811

1994 5866 4800 2403

1995 5856 5814 1256

1996 5932 3782 2336

1997 6893 3630 1592

1998 7166 5221 2591

1999 7430 4250 2819

2000 5973 4750 3100

2001 5800 5000 3300

Mean 6273 4429 2376

Durum wheat imports (000 tons)

Year Tunisia Algeria Libya Morocco

1992 22 2481 317 44

1993 12 2381 166 260

1994 33 2265 212 292

1995 516 3523 289 55

1996 449 1457 147 337

1997 123 1758 247 353

1998 551 2658 217 520

1999 225 1935 113 477

2000 248 2018 223 529

2001 450 2200 250 670

Mean 263 2268 218 354

4

The Constraints to Agricultural Development and Food Security Agricultural development in North Africa is constrained by many factors that go beyond the standard ones of climate and soils. Foremost, in most countries there is a declining interest in agricultural and rural development. Additionally the rural communities in the region are facing a number of converging trends that threaten the future of this poorest sector of society and through them agricultural production of the region.

Among the most important are:

Global Climate Change: North Africa is predicted to become warmer and drier with a greater frequency of droughts. Yields of cereals will decline and imports and prices will increase in proportion.

Water Scarcity: The region (with West Asia) is already one of the most water scarce and this is predicted to worsen markedly over the next 25 years. As a result, the food security situation will also likely worsen.

Currently the region is the largest importer of grain (with West Asia). The problems in the dry areas are even more acute: water resources are being mined causing declining water tables and the salinization of good agricultural lands.

Land degradation (desertification): as much as 45% of the total land area dedicated to agriculture and rangelands is degrading, further reducing the already low potential productivity of the land. Moreover, rangeland area in the Maghreb countries has decreased by 10–13% since the mid 1970s.

Poverty: despite progress, poverty remains a threat mainly in the rural areas where 60-70 % of the poor are concentrated. It particularly strikes women and young people who are the most unemployed. This phenomenon is shown in the low ranking in HDI of North African Countries (64 to 123);

Low diversification of crops and livestock: the systems rely primarily on wheat and barley and small ruminants (mainly sheep);

Highly fragmented holdings: In the Maghreb countries, the small and medium land holders account for 80% of all farms (3 million). This fragmentation may seriously hinder the impact of agricultural development projects;

Low rainfall lowland and mountain areas have been largely ignored by research and policy makers. Consequently, appropriate technologies to improve farm productivity in these areas sustainably are not available or rare;

Inefficient seed production and distribution systems: Sufficient high quality affordable seed is not available for farmers to plant on time;

Over production and land degradation: The combination of relatively high population growth (more than 2%) leading to increased food demand and low land productivity is exerting tremendous pressure on smallholders to opt for over cultivation and/or cultivation of fragile and erosion prone areas;

High population growth and trends toward a younger population is placing pressure on labor markets, with likely increased unemployment and out migration, often illegal in Europe.

Limiting infrastructure: Inadequate policy and institutional support is very often exacerbated by a lack of basic infrastructure (roads, social services) that greatly limit prospects for improving the welfare of resource-poor households;

Because of the remoteness of these areas and lack of industrial investment, agriculture remains the nearly sole source of income as off-farm income-generating opportunities are virtually non-existent;

Shortage of feed for livestock both in volume and quality represents the primary constraint to improving livestock production in the region;

Mountainous areas have generally low crop and animal productivity and poverty is more acute, leading to out-migration in view of the very limited off and on-farm employment opportunities.

The smallness of farms and fragmentation of high land represent serious impediments to agricultural growth. Wheat and barley are the most common crops with livestock being an integral component of the farming system. Much agricultural activity is conducted on sloping land with high risk of erosion by water run-off with little, if any, conservation measures in place;

Poor linkages between research, extension, the farming community, especially women, and policy makers: Poor interaction has constrained technology transfer activities. Commonly

5

Technology and management options for high potential areas

extension programs are top-down approaches that consider farmers as passive recipients.

Technology and management options for less-favored areas

Most promoted technology highlights input intensification rather than input efficiency. Farmers’ input in setting research agendas remains the exception rather than the rule. In the very rare instances where participatory mechanisms are in place, they tend, by and large, to benefit high-resource farmers;

People-centered development

Investing in science and technology

Building regional and international partnerships

Subsidies on feed, fuel, and machinery have provided strong incentives to farmers to increase their flock size and expand cereal cultivation on fragile lands.

Looking Ahead Sustainable development in North Africa will depend on commitment, capacity, and good governance of the key actors involved. Even in areas with high potential production,

degradation of the resource base as a result of intensive cropping, especially mono-cropping, and overuse of inputs such as irrigation water, inorganic fertilizers, pesticides, and herbicides, has been a major contributing factor to the dwindling crop yields.

A focus on improving the well-being of rural people, and reducing rural poverty in the widest possible sense, will mean improved quality of life, not just increased average income of rural populations. Investing in agricultural research will always pay off.

At the same time governments must invest in education, health, clean water and rural infrastructure. Policies should provide incentives for sustainable natural resource management, such as secure property rights for smallholders.

Broad priorities In an era of increasingly open markets and constant communication/information revolutions, the fate of the region agricultural sector will undoubtedly depend on its capacity to “grow in place,” that is, by adopting a regional research/development approach that would create the right conditions and incentives to significantly improve agricultural productivity while preserving the environment.

Above all, poor people must participate in making decisions and implementing programs that affect them. Finally, policy research is required from the household to national levels to help farmers cope wilh global changes and challenges.

Where does Explore On-Farm fit in? The priorities of agricultural/rural development in the region should focus on: In its current form, Explore has a role to play in

taking agriculture technologies to the farm and adapting and optimizing them into sustainable models for use in local regions with their peculiar constraints.

fostering broad-based rural economic growth

improving social well being, managing and mitigating risk, and reducing vulnerability,

enhancing sustainability of natural resource management. In the process it makes farmers key decision-

making partners with funding agencies, researchers and extension workers. It links governments to the land. It is one hands-on tool for bringing rural communities together in trials to develop their local agriculture through heritage, experiment and knowledge.

The aim should be to devise and implement strategies (options) that would ensure reasonable complementarity (trade-offs) between the goals of growth, equity, and preservation. Such options should include:

6

General Information relating to agriculture in North Africa

Table 1A: Distribution of population

(1) (2) (3) (4) (5) (6) (7)

mn inhtts % % % % % %

2000 1965-2000 2000

Algeria 30,29 2.96 60 40 24 24 12

Libyan A.J. 5,29 3.76 88 12 6 6 49

Mauritania 2,66 2.82 58 42 53 53 4

Morocco 29,88 2.56 56 44 37 36 7

Tunisia 9,46 2.26 66 34 25 25 10

Egypt 67,88 2.42 45 55 37 33 8

Portugal 10,02 0.34 64 36 14 13 15

Spain 39,91 0.69 78 22 7 7 31

(1): Total population in millions of inhabitants (2): Average annual demographic growth rate in period 1965-00 (%) (3): Part of urban population in the total population (%) (4): Part of rural population in the total population (%) (5): Part of agricultural population in the total population (%) (6): Part of agricultural labour force in the total labour force (%) (7): Number of inhabitants per agriculture employee

Table 2A: Gross Domestic Product, economic growth, agriculture ratio to the GDP

GDP GDP /Inhtts

Exchange Rate*

GDP growth rate

AGDP /GDP

AGDP /Agr. E.

mns $ $ $ P 1 M U % % $

1990-99 1999

Country Year

(1) (2) (3) (4) (5) (6)

Algeria 1999 47015 1580 0,015 1.6 12.3 2341

Libyan A.J. 1999 27683 5349 1,851 na na na

Mauritania 1997 1058 437 0,007 na 25.24 464

Morocco 1999 34999 1193 0,105 2.3 13.0 1073

Tunisia 1999 21031 2223 0,904 4.6 14.1 3340

Egypt 2000 98364 1449 0,271 4.4 17.0 1552

Portugal 2000 103876 10371 0,922 2.5 3.17 5408

Spain 2000 580297 14720 0,922 2.2 3.0 11903

Adapted from MEDAGRI in Development and food policies in the Mediterranean region (CIHEAM, 2002)

(1): Gross Domestic Product in millions of $ US (2): Gross Domestic Product per inhabitant in $ US (3): Exchange rate, $ US per 1 local monetary unit

7

(4): Average annual growth rate of GDP on period 1990-99 (%) (5): Part of agricultural GDP in the total GDP (%) (6): Agricultural GDP per agricultural employee (1 $ US) Table 3A: Cultivated areas, irrigated areas, means of production, 1999

Arab land, perm. crops

Cult. Land/ htts

Cult. land /Agr. E.

Irrig. land /Cult. land

Cult. land /tract

Fert /Cult. land

1000 ha ha ha % ha / tract. kg / ha

Country

(1) (2) (3) (4) (5) (6)

Algeria 8215 276 3,3 7 88 18

Libyan A.J. 2150 420 19,55 470 63,24 29

Mauritania 500 190 0,82 49 1315 -

Morocco 9445 322 2,2 14 219 35

Tunisia 5100 545 5,5 8 145 22

Egypt 3300 49 0,4 100 38 360

Portugal 2705 271 4,0 24 16 95

Spain 18530 465 13,8 20 21 125

Adapted from MEDAGRI in Development and food policies in the Mediterranean region

(1): Arable land and permanent crops, 1000 ha (2): Cultivated land per inhabitant, ha (3): Cultivated land per agricultural employee, ha (4): Part of irrigated land in the cultivated land (%) (5): Cultivated land per tractor, ha (6): Fertilizers per hectare, kg

Table 4A: Main agricultural products, 2000

Cereals Vegetables Fruit Milk Meat Sugar Olive oil Country

1000 T

Algeria 1226 2580 1490 1376 509 0 50

Libyan A.J. 251 882 376 206 203 - 9

Mauritania 263 12 25 307 57 - -

Morocco 2082 3615 25929 1266 540 475 62

Tunisia 1095 2154 933 920 219 19 200

Egypt 20046 13563 6575 3831 1391 1400 -

Portugal 1686 2429 1713 1983 704 60 47

Spain 24794 11982 15044 6530 5071 1146 788

Adapted from MEDAGRI in Development and food policies in the Mediterranean region

8

Table 5A: Self-sufficiency ratios for main food products, 1999

Self-sufficiency ratio= production *100 / (production + import – export)

Cereals Sugar Milk Olive oil Meat Country

%

Algeria 17.33 0 86.33 99.92 95.82

Libyan A.J. 10.52 0 na na na

Mauritania 20.87 0 na 0 0

Morocco 32.55 50.42 99.37 130.64 99.53

Tunisia 36.56 6.04 99.06 624.53 98.88

Egypt 68.23 53.72 98.81 0 88.52

Portugal 36.69 21.05 102.11 67.14 88.38

Spain 82.31 85.71 96.00 116.22 100.42

Adapted from MEDAGRI in Development and food policies in the Mediterranean region

Table 6A: Labour Productivity level and human development

Country Productivity level Human development Labour Productivity index

Ranking 96-98 HDI-98 France =100

Algeria 7 0,683 5,3

Libyan A.J. na na na

Mauritania na 0,418 na

Morocco 10 0,589 5,0

Tunisia 6 0,703 8,0

Egypt 13 0,623 3,2

Spain 4 0,899 36,6

Source: MEDAGRI in Development and food policies in the Mediterranean region

9

Climate Risk Analysis and Crop Management Decision Making Tools Tools are described enabling farmers to decide on their best planting date, the likelihood of drought and the length of their growing season. The tools inform the farmer of the element of risk they face when taking management decisions involving their predictions of climate. Using the tools provides farmers with some reason for selecting early, medium or late varieties of cereals or even other species to grow to enable them to work optimally within the climatic constraints of their farms.

Introduction Water is the factor most limiting agricultural production all over the world. In arid and semi-arid regions of the Maghreb countries, water is extremely limited. Agriculture there depends on very low annual rainfall which is highly erratic in timing and amount and erratic in distribution throughout the region. Compounding the problem, soils are shallow and have low water storage capacity. The risks of water shortage in agriculture there are extreme.

Crop production in general and cereal production specifically are inherently risky businesses. The role of the agricultural scientist, in support of the farmer, is to develop strategies for coping with and working around climatic constraints to optimize crop production for a region. In harsh and variable environments any strategies have to be tested over a longer period of time than is required in more favourable areas, in order to be sure that once recommended, the practices will be productive, sustainable and economically viable.

The first step in developing practices is to describe and understand the physical environment, its likely limitations to the crop in degree and in time. The second step is to frame that knowledge into a suite of simple and friendly tools that will help farmers to make time-based decisions about how to manage their crops and farms.

This paper describes some of the background to developing these tools. They are based on the analysis of risk. The procedure has been first to collect information spanning many years on climate, soils and crops and to consolidate it into a data base. Then the long-term rainfall records in the database have been used to generate a cumulative probability and rainfall variability analysis for the whole year as well as for monthly data. This is to allow characterisation of any cropping season to see how it compares to the long series of seasons.

The first farmer tool coming from these analyses is the Standardized Precipitation index (SPI). It tells the farmer what the risk of drought will be in his area. How many years in ten his crop may fail using his standard practices. The next tells the farmer when his best planting date is likely to be. It is based on the records of timing of the first significant rain (FSR) for each region, sufficient to establish a crop and follow-up rains to maintain growth. A further tool tells the farmer the likely length of the growing period (LGP) on his farm taking into account recent trends towards shorter seasons. This provides the information needed to decide what crop he can plant and which variety (cultivar): should it be early, medium or late.

Farmers need to know the risk of early, medium season and late droughts in their region so they can make fundamental management decisions during the cropping season: when to apply nitrogen, to irrigate if water is available (supplemental irrigation) or to graze their animals on the crop.

All this information is the basis for a Global Information System (GIS). This enables us to have an easy means of checking what is happening with Mahgreb crops at any time and in any location.

The tools are described in the following paragraphs

Rainfall Probability and variability Analysis The climatic analysis is based on the cumulative rain probability (frequencies); using real data from many years assuming that these cumulative rain values are random and repetitive in time. Different probabilities of occurrence are assigned to these values. The deciles that allow us to divide our series of data into ten are given in figure 1 using data from Sidi El Aidi as an example. Figure 1 can be explained as follows: D1: nine years out of ten our data will follow at least this curve (the most likely situation with low rainfall).

10

11

D7: three years out of ten data will follow at least this curve (less probable situation with high rainfall). Annual rainfall of any specific season being studied is generally compared to the D5 curve (red) which represents the median distribution of rain throughout the year for the location.

As an example comparison, the 2003-2004 cropping season at Sidi El Aidi (Settat-Morocco (blue curve)) was wet early in the season. But, a period of drought occurred during January and early February. Rain resumed during April and May. However, this rain was mostly unavailable for winter sown cereals that had already matured especially in shallow soils. This comparison between the long term D5 smoothed probability shows that while this type of analysis can be very useful for retrospective analysis, explaining why crops in particular years fared well or poorly, it can have weaknesses when predicting climate for the forthcoming season.

Rainfall variability In arid and semi-arid areas, the annual rainfall and its distribution are highly erratic and vary from year to year and from region to region. There is no

clear cyclic phenomenon of rainy and dry years. However, dry years are becoming more and more frequent during the last decades.

Figure 1. Annual rainfall probability. Sidi El Aidi. Lat 36°N

In the region of Aleg (Mauritania) for instance, the decrease in the annual rainfall is obvious (see Figure 2). During the example 80-year period, rainfall declined by 2 mm per year on average but the variability from year to year was very high with a variation coefficient of 46%.

The moving average shows a significant reduction in annual rainfall of 100 mm and reduction in year to year variability at Aleg since 1974.

Standardized Precipitation Index (SPI) The SPI is an index that describes the level of drought within a given area or region. It represents the difference between the average rainfall over a given period divided by the mean square of the same period. The computation of the SPI is based on a series of monthly rainfall data for at least 30 years. Drought occurs when the SPI is negative. Drought is moderate for SPI -1.5 to 0, severe for -2 to -1.5 and extreme for values lower than -2.

The graph of SPI over 70 years in Figure 3 shows that there has been continuous drought since 1990 in the eastern part of Morocco (Taouriret region). There the index reached its lowest values during the last decade (-2.3 in 1998 and 2000).

Figure 3. Standard Precipitation Index (SPI) in Taouriret, 1930 - 2000

Figure 2. Annual rainfall at Aleg in Mauritania 1921 to 2002. The red curve is a linear regression fitted to the data. The blue curve is the moving average.

In Sidi El Aidi region (Settat- Central part of Morocco) the SPI has been calculated for two periods of the growing season: October-December (cereal crop establishment) and January-March (main growth period). The following graph (Figure 4) shows that during the last four years the first period has been wetter than the second, when

12

previously it was more often drier during establishment. This pattern, if it continues, and the grain growth period becomes relatively even drier and shorter will require changes in the management of the crops in this region.

First Significant Rainfall In arid and semi-arid regions, sowing is perhaps the most crucial step in crop establishment and crop production. If it is done correctly and on time it will frequently lead to good production in winter sown cereals. However, deciding on planting time is not easy especially under limiting and erratic conditions of rainfall.

Given the shortness of the rainy period, any delay in planting time can affect crop yield negatively. Any significant delay exposes the crop to late season water stress and high temperatures. The first significant rain (FSR) can be considered a

dependable tool for deciding on the best planting time. The FSR is defined as the period after October 1 during which there is enough rain to ensure crop germination and stand establishment.

Figure 4. Standard Precipitation Index (SPI) at Sidi El Aidi 1980 – 2004. Green curve January-March, blue curve October-December

Analysis of the FSR in Settat and Berrechid regions (central part of Morocco) is shown in Figure 5. There 25 mm rain is enough to allow successful planting (this amount differs depending on soil type and depth). The figure shows that in the wetter region (Berrechid) farmers can start planting their cereal crop earlier from the first of November, whereas in Settat they have to wait until the tenth of November (figure 5).

FSR can be used to set planting time in areas where no demonstration trials have been conducted.

Length of the Growing Period (LGP) The length of growing period is the period of the season where temperature and water availability allow the crop to grow. It represents the period where precipitation (P) exceeds half of

evapotranspiration (ET). The analysis of LGP in Khouribga region (central part of Morocco) shows a shift of the start of this period from October in 1960-65 to November in 1995-2000. The whole period is also shortened from 180 days to 110 days because the end of season moved from early March to late April (figure 6).

Figure 6. Length of the growing period at Kouribga in 1960-65 (blue curve) and 1995-2000 (red curve). Starting and end points are where the curves cross the month axis

Figure 5. Probability of receiving the first significant rain of 25 mm by particular dates after October 1 at Settat (green curve) and Berrechid.

Later possible planting date followed by a significantly shortened growing season call for a rethink of cereal management strategies, particularly a move to shorter duration cultivars or even other species to best utilise the resources available. As an example, winter cereal crops are no longer suitable in the eastern part of Morocco now the LGP has reduced to 80 days.

By knowing the LGP of different regions at the national level, we can more reliably recommend suitable species, varieties and crop management approaches for farmers to use in advance, as climate continues to change. We can also put breeding objectives in place in anticipation of further changes based on observed climate patterns

Early and late-season drought risks The knowledge of drought occurrence risks during the early growth season is very important for planting date assessment. In Settat region, the analysis of late drought risks based on rainfall data from year 1970 to 2000 shows that October is the most risky period. Despite some fluctuations in time November and December are less risky and more stable. In this region then, winter cereal planting should be avoided during October and targeted to November.

The knowledge of drought occurrence risks during the late growth season is very important for planting date assessment and the choice of varieties. In Settat region, the analysis of late drought risks based on rainfall data from year 1970 to 2000 shows that the risk starts early February. Under these conditions, the LGP is shortened and the production potential is reduced. The choice in terms of species and varieties is also reduced.

Conclusion and Recommendations These tools help better management of the growing of cereals in highly variable environment. They are easy to use.

References: El Mourid, M., A. El ouali, A. Ambri, M.El oumri, and W. Goebel. 1996. Caracterisation agro-écologique: outils de gestion et d’aide à la prise de decision en agriculture aléatoire. Pp 41-51. AlAwmia INRA Morocco

El mourid, M. and D.G. Watts. 1993. Rainfall patterns and Probabilities in the Semi arid cereal production region of Morocco. Pp.59-80. In. M. Jones et al. editors. 1993: The Agrometeorology of rainfed barley-based farming systems. ICARDA Aleppo Syria

13

Constraints to Cereal-Based Rainfed Cropping in North Africa and Methods to Measure and Minimize their Effects The main constraints to crop production in Mediterranean environments are shortage of water and extreme temperatures. This chapter describes how they impact at particular stages on the growth and development of the crop and on building of the components of yield. It introduces in general terms how the farmer can work around them by avoiding having sensitive yield-generating stages and periods of stress occurring simultaneously.

Methods of measuring crop yield components, water use and water use efficiency and how they might be used to assess what might have gone wrong with the crop are described.

Finally, it introduces trials that might be required to optimize on-farm productivity in Mediterranean environments of North Africa. Such optimization trials are the subjects of the “Explore On-farm” series of chapters.

Water shortage is the major abiotic factor limiting growth and yield of cultivated rainfed crops in North Africa. The damaging effect of a drought period will depend on its duration, on how much water is stored in the soil and the proportion that the crop can access, how fast it is used or lost, and on the crop developmental stage at that time.

Drought duration and occurrence is quite variable. The chapter “Climate Risk Analysis and crop management decision making tools” explains how to study this variability and estimate the risk of occurrence.

Crop management should aim to make as much of the season’s rainfall available to the crop as possible and at the right times to form the yield components that maximize grain production. Above all, crop growth and development should be manipulated so that during the final stages of the crop cycle enough rainfall or stored water remains to fill the grain.

Other factors besides water shortage that may reduce yield significantly are high temperatures that accelerate transpiration and development, and any predominant damaging pests, diseases or weeds. Late frosts are important in more continental type of weather and in the highlands. Understanding the relationships between all these factors and crop growth and yield is a first critical step towards developing management strategies for the special conditions of any farm.

Before launching into any trials to optimize production on any farm, ensure the collaborating farmer has some basic idea of how the crop grows and develops and what processes s/he can manipulate. The following gives a simple outline.

Crop growth and development How it all works It is probably worth explaining that crops both need to grow (get bigger) and develop (add more parts of the same type and more parts that are different).

The simplest way of thinking about development is to split it into two categories, vegetative development that includes production of leaves, tillers and roots, and reproductive development, meaning the parts that finish up as the stalk and ear (head or spike) with its grains.

There are many steps within each of these processes of vegetative and reproductive development but there is one big step that switches the plant from predominantly vegetative to predominantly reproductive development. Without that switch wheat would continue to produce only leaves and tillers for several seasons.

The switch involves length of day (longer day for a faster switch) and temperature (warmer conditions make the plant’s internal clocks go faster towards the switch but semi-winter wheat

11

requires low temperature first to accelerate the switch).

The plant senses day length and temperature continuously and records and sums them until it reaches the numbers required by the variety to make the switch. More details of these processes are in the chapter “Optimizing variety x sowing date for the farm”, where the concept of heat sums is explained and the factors that make different varieties develop at different rates. Growing bigger Plants grow bigger by absorbing carbon dioxide (CO2) from the air through tiny controlled apertures (stomata) in their green surfaces.

The carbon of the CO2 is built into sugars inside the plant using energy from sunlight in a process called photosynthesis. These sugars, in different forms and combinations with inorganic nutrients extracted from the soil via the roots, become the bricks from which the plant is built. Some sugars go straight into growth while others are put into a storage bank for later withdrawal and use, particularly for filling the grains. The stems can be a good sugar store (see the chapter Optimizing nitrogen use on the farm for how N can affect this and yield).

It follows that as long as the crop has enough sunlight and CO2, water and inorganic nutrients, and it is warm enough, all organs will grow as big as their potential allows.

There is one drawback to photosynthesis, water leaks out through open stomata as the CO2 diffuses in. The hotter, drier and windier the air is, the faster the precious water is lost for each unit of CO2 that enters. Though some passage of water from the roots to the leaves is necessary to extract nutrients from the soil, the plant must retain adequate water for essential growth processes.

A severely droughted crop keeps its stomata closed to stop water loss, but this prevents CO2 from entering which concomitantly prevents

growth. This is why crops that run out of water soon die. They starve as well as dehydrate.

Hopefully, the farmer will appreciate from the previous discussion that the plant must use water to grow and the hotter it gets the less efficiently it uses available water to grow. When the crop is under water stress, the organs growing at that time will be proportionately less able to reach their potential size.

Plants continually sense their environment and adjust their overall growth and the blend of the components of growth and yield to current weather. They also continually adjust their growth to their neighbours in the crop. Developing: more parts of the same type To a degree, the more sunlight that the crop can absorb to drive photosynthesis, the better it grows. As more of the ground is covered by green leaf the more sunlight is intercepted.

So before a crop can grow at its greatest absolute rate, it must have many leaves. Every leaf takes approximately the same time to expand (measured in temperature time) and the next one queued on any shoot cannot start to expand until the previous one is well under way.

Development is all about waiting turn in a sequence. The size that any unit achieves is dependent on how much plant substrate is available when its turn comes to grow.

The crop can produce many shoots (tillers) that eventually grow fairly independently and each tiller has its own leaves. However, tillers are subject to the same queuing rules that apply to leaves and even more so to the amount of substrate the plant has available when its turn comes to grow.

If the plant is short of substrate (because of a drought or low light or nitrogen shortage) when a tiller’s turn comes to grow, it will remain dormant and possibly never grow. The next tiller in the sequence will start if conditions improve again.

Cross section of a wheat leaf

The picture that the farmer should have in mind now is that the crop is continuously sensing its environment and adjusting its size, shape and types of components to match the constraints. The farmer’s job is to ensure that the crop’s composition leads to best yield within the constraints.

12

Stepping from vegetative to reproductive development

13

Crops do not really stop one process and then start another.

There is always an overlapping progression: so new leaves are still expanding long after the crop starts to grow its microscopic ears. This can be seen in the next figure where the green bar labelled leaf appearance extends beyond the picture of the ear at terminal spikelet.

However, just like each leaf and tiller queues to develop and have their final size set by plant substrate at the time of their growth, the reproductive parts, ears, spikelets, florets and finally grains, are subject to those rules. Whether the next floret starts to grow depends on the state of the previous floret and whether substrates are

available to support it.

The figure attempts to show this interplay. It may seem complex but it is very much simplified compared with what the plant in the crop does. It shows when the plant components that bear the yield structures are forming.

It also identifies when the processes are potentially most sensitive to stresses and might require higher inputs. This is when the farmer can have an impact on yield.

Discuss the figure with the farmer and use it to introduce how the crop must fit its key stages within the limiting characteristics of the rainfed Mediterranean environment.

Some of these special characteristics are highlighted now.

HM Rawson

How the crop develops, continually sensing and responding to its environment. The top section shows pictures of the crop and names of some stages possibly familiar to the farmer. The next section shows pictures of two internal flowering stages, double ridges and terminal spikelet and when they occur. Below again are the components of growth, overlapping in their timing and competing for resources. Next are the yield components and when they form in relation to everything else. Finally at the bottom on a green background are stages when the crop is particularly sensitive to certain stresses or lack of resources.

14

Growing season – understanding the environment of the region and farm Minimizing risk In rainfed agriculture, farmers have an intimate relationship with the prevailing weather because they depend on it. After generations of farming under erratic weather, local experience is the base of crop management decisions.

Memories are stronger for events that caused greater crop losses. Fear of those events recurring can lead to over conservative management decisions. Minimizing risks ensures a crop will be harvested every year. However, minimizing risks every year for the occasional bad year also reduces the opportunities to capitalize on the good years in a region of erratic weather.

Traditional practices may be reliable but they leave little room for advance. New technologies and improved varieties may require different management decisions to realize their potential for increased yield.

Merging the wealth of on-farm experience with an improved understanding of when in crop growth and by how much the weather impacts yield is the route to reliable improvements in yield through adapted methods of management.

Read the chapter “Climate Risk Analysis and crop management decision making tools” for detailed environment description and an explanation of risk analysis. Water constraint In Mediterranean environments seasonal rainfall is limited and in most years crops suffer a terminal drought during grain filling. If this description fits the region of your interest discuss the following questions with local farmers.

Could the traditional sowing date be advanced so terminal drought is partially avoided? Which date actually corresponds to the average break of season when rainfall is sufficient to work the ground and sow? Is it earlier than the traditional sowing date? If sowing were advanced, would flowering date also be advanced? Would then the risk of frost during flowering be increased? When is the last probable frost date? Break of season The break of season or first significant rain (FSR) can be defined as the period after the 1st October during which rainfall is enough to ensure germination and stand establishment of the crop.

Though the farming calendar may be set by it in advance, it is very variable from year to year.

The FSR date (at least 25 mm in one or two consecutive days) for the past 29 years in Jemaa Shaim (Morocco) are shown in the following table. The FSR occurred in October in 10 out of 29 years. However, in 5 of these 10 years (in red) there was no follow-up rain (less than 20 mm in the following 25 days after FSR). The numbers in blue show the next significant rain after FSR.

The FSR occurred in November in 10 years, of which 3 out of 10 suffered a drought after FSR. On the other hand, sowing had to be delayed after 1 January in 3 out of the 29 years.

Planting in October seems risky, however, planting at a much safer date (e.g. December) will

When break of season rains fell in Jemaa Shaim and seasonal rainfall. Year refers to that when the crop was sown. The numbers are the date in the month

Year Oct Nov Dec Jan Feb Rain

1975 14 310 1976 28 19 262 1977 17 502 1978 17 368 1979 19 16 268 1980 4 210 1981 12 209 1982 7 11 159 1983 15 224 1984 7 357 1985 26 271 1986 15 15 207 1987 25 514 1988 9 434 1989 26 > 237 1990 17 202 1991 9 28 227 1992 16 30 209 1993 4 336 1994 10 132 1995 11 640 1996 12 19 475 1997 21 388 1998 31 192 1999 26 229 2000 21 182 2001 11 355 2002 20 14 294 2003 22 392

N° 10 10 6 1 2

waste days to weeks of potential crop growth. Planting in December rather than earlier also increases the risk of running the crop into hot weather and drought at the end of the crop cycle.

What is required is for the farmer to be able to adapt approaches each year to the prevailing weather and maybe even split approaches on farm between higher and lower risk options. Dry sowing of the crop is a possible choice in some years.

For example, it may be necessary to plant a different variety onto opening rains in October than in December to take full advantage of the early season break. When date of last frost sets earliest anthesis In zones with cool winters in which frosts are not unusual, attention should be paid to possible frost damage during anthesis/flowering, particularly when earlier sowing of a cultivar means earlier flowering.

The earliest flowering date that can be planned is determined by the probability of frost. The relationships between the level and duration of low temperature and crop damage around anthesis are not fully understood, but a good policy is that frost should be avoided.

Historic data of daily minimum temperature will help to determine the probability of occurrence. The approach will be similar to that used in the FSR analysis for Jemaa Shaim, and as for FSR, the dates of last frost are spread in various months.

Going for the safest date may excessively increase the chance of running the crop into terminal drought. Advancing anthesis one or two weeks may be considered a low risk.

Once you know the likely dates of the two constraints of opening rains and anthesis frosts you can start to plan for the variety you might use for the area. This is dealt with in the chapter “Optimizing variety x sowing date for the farm”. Rainfall during the season Just as opening rains and frosts are variable in timing and amount, so too is total season rainfall variable in timing and amount between seasons. All farmers are aware of this whether they maintain records of weather or not. Total water available to the crop is the first determinant of yield while its distribution in relation to the formation of crop yield components is next.

Wheat requires at least 200 litres of water during grain filling to produce each 1 kg grain and up to twice that amount in very hot, dry areas.

Scaling that ratio up, a 3 000 kg (3 t ha-1) crop will need 600 000 litres either from stored soil moisture or from rain. This is equivalent to 60 mm rain (1 mm rain falling on 1 m2 is 1 litre or over 1 ha is 10 000 litres).

The crop also needs water to grow prior to anthesis. For a 3 t ha-1 yield the crop will need to have more than 6 t ha-1 of biomass (total dry matter) produced by anthesis, requiring around 160 mm rain or equivalent stored moisture (a clay near field capacity could contain this amount).

If long-term weather data show this amount of water is not available in most years, yields will be less than 3 t ha-1 and a short duration variety might be most efficient for the area.

There is no point producing a lot of biomass before anthesis, as is possible with a long duration variety, if there is no water remaining after anthesis to fill the grains.

The table with data from Jemaa Shaim illustrates the variability of rainfall in the Mediterranean region. Over 29 years the average seasonal rainfall has been 303 mm, most falling between November and April and sufficient for good crops. However, there have been severe droughts. In 1994, the accumulated rainfall was 132 mm with the FSR not occurring until February. Terminal drought is the most common: for example, April rainfall was less than 20 mm in 15 out of 29 years. Droughts may occur also within the season. In February for example, average rainfall is 37 mm but this has ranged between 0 and 99 mm. Depending on interactions with developmental stage of the crop, yield in those years could be 0 and 4 t ha-1. The soil as a bank for rainfall The soil provides a store for rain water that the crop can access between rainfall events. If the soil is porous and deep, rainfall variability has less impact on specific yield components. Stress will develop more slowly and the plant will have time to adjust its structure without death of parts.

If stress develops quickly during grain filling, as happens in shallow soils, the plant may not even have time before it dies to move stored carbohydrates from stems to help fill the developing grain. This is often the case in the vast areas of shallow soils in Algeria.

15

It will be worth assessing with local farmers the depth and structure of their soils. Soils may change over short distances particularly where topography is undulating and steep. Discuss the farmers’ observations of crop performance across the region that might bear on soil type. Do some digging or soil tests if necessary to determine whether soils can be improved in their ability to capture and carry available water better between rainfall events. Water at the right time for optimizing yield The pattern of water availability and use within the season can alter not only biomass production and grain yield but also the apportioning of dry matter between them (Harvest Index). The sliding relationships are shown in the figure.

You need to be careful how you assess harvest index (HI). A high value towards 0.50 (50 percent of crop weight being as grain) is usually considered to be good. High HI can however, occur when yield is poor as well as when yield is high. The following circumstances illustrate the possibilities. Check the figure as you read on. How apportioning of the season’s water before and after anthesis can change yield, biomass (the dry weight of the crop), and the ratio of yield to biomass (harvest index)

16

To the left of the optimum, low yield results from the crop growing poorly before anthesis. It is unable to use its water supply fully. Though this leaves ample water for grain filling, the crop has been unable to produce enough grain sites by anthesis to exploit water to generate a big yield. Though this is an undesirable outcome, harvest index will be high and thus, low biomass can lead to high HI.

To the right of the optimum the crop has grown vigorously and used up its water supply before anthesis in making biomass. Although it has adequate grain sites for a large yield, the grains are not filled because water is exhausted too early. Then the harvest index is low.

What the farmer should aim for is a balance between the previous two situations, aiming to have just enough water remaining in the soil at anthesis to completely fill all grain sites. This will produce a high HI in association with high yield.

HI is used later in association with values of water use efficiency to assess how well crops have performed and identify what might have gone wrong. The farmer may find these concepts too difficult.

However, measuring the proportions of different yield components at maturity can also help to identify when and where the problems of a crop occurred.

The farmer should have little difficulty understanding yield components if the ideas behind the earlier figure of plant development are clear.

Harvesting and how to measure and use yield components When should you analyse yield components? Wheat crops are quite plastic in the way they can allocate growth to different components, putting more or less into particular structures depending on current weather.

This flexibility aims to preserve yield potential despite intermittent weather constraints. The composition of yield components at maturity is a reflection of the history of the plant’s coping with the weather. Consequently, analysing the yield components at maturity can help unravel what went right or wrong during growth of the crop. It helps to show what an experimental treatment did.

Yield component analysis is not something that you always need to do. It can be very time consuming. Plot grain yield and biomass may be all you need. However, in case you and the farmer decide that it is necessary for your study, you can follow the detailed procedures below. You will also find a shortened form of harvesting and crop analysis procedures in “Optimizing plant population, crop emergence, establishment and sowing rate” under the heading “Maturity Harvest”.

Measurements of fresh crop material to determine yield and yield components The least measurements to take at maturity are plot grain yield and harvest index. Also take 100-grain weights from the grain threshed from the harvest index sample. From these you compare kernel (individual grain) weights and estimate kernel number across your treatments.

Record the information for each plot, together with the plot identifier and date, on a table that was prepared in advance of harvest time. The chapter “Optimizing variety x sowing date for the farm” explains how to design tables.

For these measurements it is vital to measure and record the plot areas actually harvested to work out production per unit area. Otherwise the comparisons between treatments become nonsense.

It is also vital to bring all samples to similar water content before making comparisons, as water is very heavy. Spreading out and quickly drying all samples on a hot day in the sun, at the same time before weighing, may achieve this satisfactorily. Preferably use a dehydrator oven at a set temperature as this standardizes the water content of samples.

The sooner this is done after harvest the better, as any losses of material to respiration or accidents will negate the work of the studies. At anthesis (a hand-harvest) Collecting samples for a detailed study 1. As each treatment reaches anthesis, walk down one side of the plot and in three places grab four to eight culms including the stem, leaves and spike and cut them off at ground level with a serrated knife or secateurs. Repeat down the other side of the plot keeping the bundle of culms in one hand and adding to it with each grab. Try and retain as much dead leaf as possible. You should end up with about 35-55 culms. Do not sample from edge rows. Sampling in the early morning or late in the day is best.

2. If you have large plastic bags, put each sample into a bag with a dated identifying tag and go onto the next plot ready to harvest, take grabs, and put into another bag with a tag and so on until all plots that are ready have been sampled. Move all bags to a cool room with a bench. Keep samples out of the sun until they can be processed. It is possible to sample 30 plots per hour per person using this method.

3. Count fertile culms (stems with heads expected to produce grain) and sterile culms (stems which will not produce grain) in each grab sample. Do not lose dead leaf. Measure average height (not including the awns) of the bundle and estimate the average head length from four to five heads. Record plot and treatment identifiers and estimates and measurements on your previously prepared sheets.

4. Fold the sample several times, put in a paper bag and staple closed. Put into a dehydrator at 70-80°C. You can process 20 to 30 samples per hour per person using this method.

5. Depending on the capacity of your dehydrator, samples will dry in two to three days. If empty bags representative of those containing the samples are also placed into the dehydrator, these can be used as tare bags when weighing samples as they come out of the dehydrator. At maturity (a hand-harvest) 6. Follow the collecting methodologies 1 and 2 exactly as described for the anthesis harvest.

7. Count fertile culms (stems with heads containing grain) and any sterile stems in the grab sample. Do not lose dead leaf. Measure average height of the bundle not including the awns and estimate average head length from four to five heads. Record treatment identifiers and measurements.

8. Cut the fertile spikes from the sample and place into a paper bag. Fold the remaining part of the sample (straw) several times, place into another paper bag and staple closed. Put into a dehydrator at 70-80°C. You will process 10-20 samples/hour/person using this method.

9. After two to three days the straw sample bags can be removed from the dehydrator and weights recorded using a tare bag as for the anthesis sampling. The spikes should be kept in the dehydrator for seven days or until the dry weight is stable. Water loss from the grain is slow while spikes are intact. At maturity (a machine harvest) 10. As soon as possible after the latest maturing plots are harvest ripe, trim the ends of plots and measure and record the harvestable plot length.

Do not harvest the edge rows of plots. They overestimate yield due to access to additional nutrients and soil water from the pathways between plots.

17

If you do not have the option to exclude the edge rows during machine harvest then some estimate of their contribution to yield must be made. Grain yield of edge rows can be 1.4 to 1.75 times the yield of inner rows.

11. As the total grain harvested from each plot is weighed, a grain sample of 500 g needs to be separated and retained for determination of water content, hectolitre weight and screenings.

These 500 g subsamples are important because together with the harvest index samples they are used to calculate spike density, biomass and yield components for both anthesis and maturity.

The rest of the bag can be disposed of unless seed is needed for later trials. Calculating yield and yield components The calculations use some shorthand as below:

strawdw = dry weight of straw sample

spikedw = dry weight of sample of spikes

spikes = number of spikes in sample

graindw = dry weight of grain threshed from spike sample

100 kw = dry weight of 100 kernels

hsubgraindw = dry weight of 500 g subsample of grain from plot harvester

hgrainfw = fresh weight of grain harvested from plot harvester

18

The calculations of yield components follow:

subsample dry weight (subdry) = strawdw + spikedw

culm weight (culm) = subdry/spikes

grain weight/spike = graindw/spikes

harvest index (HI) = graindw/subdry

kernel weight (kernel) = (100 kw + 100 kw)/0.2

kernels/spike = (graindw/(kernel x 0.001))/spikes

dry yield = ((hsubgraindw/500) x hgrainfw)/(plot length x cutting width of harvester) i.e. yield for 70–80°C oven dry weight expressed as g/m²

kernel number = dry yield/(kernel x 0.001)

biomass = dry yield/HI

spikes/m² = biomass/culm

Calculating anthesis biomass Anthesis culm = anthesis subdry/anthesis spikes

Anthesis biomass=anthesis culm x spikes/m²

Measurements of water and how to estimate crop water use and water use efficiency As yield in Mediterranean environments is so commonly limited by water, it is vital to measure not only the total amount of rainfall but also its distribution within the season, as this affects the composition of the yield components.

Furthermore, it will be very useful to know how much water gets into the soil store and is extracted from the soil.

These measurements will enable you to calculate crop water use and crop water use efficiency (WUE).

They will tell you whether the available rainfall is being converted optimally into yield or whether some is being wasted.

They will also show what impact your experiment treatments are having on the use of available water to produce yield.

The figure gives an example of how nitrogen applications can alter the amount of yield that can be produced per unit of water the crop uses.

How N can change the relationship between yield and water use

Blue line - 10 kgha-1 N at sowing top dressed with an additional 80 kgha-1 N top dressed with an additional 240 kgha-1 N

Measurements needed to calculate crop water use and water use efficiency Before sowing a rain gauge should be installed near the farmer’s dwelling or at the trial site to measure rain throughout the season. Readings should be taken on a regular (weekly) basis, recorded and the gauge emptied. The gauge is estimating the amount of rain the crop receives so it should be well away from overhanging trees or other structures. This basic measurement takes only five minutes each week and is vital.

19

So you can estimate the amount of water stored in the soil at the beginning of the crop cycle for possible use by the crop and unused water at the end you need to take soil samples. These measurements could take up to half a day twice in the season.

At sowing and harvest use a hand auger or thin walled tube to sample to 0.6 m deep. Bulk the sample into a plastic bag. Take two or more samples from each plot. Record their plot location on the bag.

Weigh the extracted ‘wet’ soil samples and then mix each sample well within its plastic bag before subsampling. Weigh the subsamples ‘wet’ and then dry at about 104°C before reweighing.

Measure the diameter of the cutting surfaces of the auger so that bulk density of the soil and volumetric water content can be calculated as below. Write down all weights and measurements in prepared tables as you go.

Assuming that no water drains below the soil depth tested or that run off occurs during high intensity rainfall events (see the chapter on “Optimizing tillage systems on-farm” for how to measure runoff), these on-site soil and rainfall measurements enable you to estimate crop water use as follows. Calculating crop water use These calculations may seem complex, but try them and you will find they look more difficult than they are.

• Gravimetric water content (%)= ((weight wet soil subsample –weight dry soil subsample)/weight of dry soil subsample)*100.

The three parts of a rain gauge

• Dry weight of soil extracted from hole (g)= (weight dry soil subsample/weight wet soil subsample) x weight all wet soil taken from hole.

• Volume of soil extracted from hole (cm³) = (diameter of soil auger/2) ² x 3.14 x depth of hole (e.g. for a 10 cm auger and a 60 cm hole the volume would be 4710 cm3).

• Bulk density of soil (g/cm³) = Dry weight of soil extracted from hole/volume of soil extracted from hole. Expect about 1.45 for loam and 1.37 g/cm³ for deep sands.

• Volumetric water content of soil = Gravimetric water content (%) x bulk density (g/cm3) x depth of sampling (cm) x 0.1 (multiplier to give mm water).

• Crop water use (mm) = (Volumetric water content at sowing – volumetric water content at maturity) + rainfall between two periods of soil sampling.

You can get a rough idea of crop water use by totalling rainfall in mm between sowing and three weeks before the crop is ripe minus 110 mm for losses to evaporation.

Also you can estimate water use efficiency (WUE) roughly by dividing grain yield in kg by the above-mentioned crop water use in mm.

You should get WUE values between 10 (poor) and more than 15 kg ha1 per mm water (good).

Calculating crop water use efficiency Water use efficiency is a measure of grain yield expressed per mm of water that the crop uses.

You can get a first approximation of WUE by dividing harvested grain yield by rainfall totalled over the season (as measured by the farmer using the rain gauge). It is usually suggested that you subtract 110 mm from your total rainfall to allow for unmeasured losses. You would use this calculation if you had not had time to collect any soil samples.

Estimating crop cover

Values of more than 15 kg grain per hectare per mm of water used means that the crop has performed well. Less than 10 kg ha-1 per mm of water used indicates that the crop has been limited by something other than water such as disease, poor nutrition, or frost.

If you have these numbers for your crops you can calculate the potential yields for the region as limited by water. For example, the Jemaa Shaim area (Morocco, see earlier) should produce yields around 3 t ha-1 from its average season rainfall of 303 mm. Calculating WUE more accurately Though the simple way to estimate WUE is to divide harvested grain yield by crop water use as above, such a calculation includes the water that the crop cannot access and thus underestimates efficiency.

This inaccessible water includes:

1. Water that is lost through evaporation from bare soil between the crop plants, sometimes assumed to be about 110 mm;

2. Water that is so tightly held by the soil that the crop cannot extract it. This lower limit of extractable water varies between soil types.

Assuming a 60 cm deep soil, the approximate lower limits of water remaining after extraction are 110 mm (red-brown earth), 80 mm (loam red earth), and 30 mm (deep sands).

If you can more closely estimate (1) and (2) you can improve your assessment of WUE, although, if you measure water in the soil at sowing and again at maturity, (2) is not important.

Estimating evaporation from bare soil (1) Do this by first estimating crop ground cover at anthesis. Look at the crop in several places around the field at an angle of 45° through the hole created by putting the tip of your index finger and thumb together and estimating the percentage of

the hole covered by crop, not soil. A denser crop at flowering intercepts more sunshine over the whole season and consequently has lower total soil evaporation (see figure of soil evaporation).

Use your percentage crop cover in association with the soil evaporation figure and for the nearest soil type, read off your value of soil surface evaporation for the season. For example if crop cover at anthesis was 70 percent and your soil was a sandy loam, evaporation will be about 90 mm. This is the value you subtract from seasonal rainfall (instead of 110 mm) to give a more accurate crop water use in (1).

Once you have adjusted your crop water use for crop cover at anthesis, recalculate the more accurate estimate of WUE. Use this new estimate to assess how efficiently your crop really is using its water.

Season evaporation from the soil surface for three soil types as related to percent crop cover at anthesis

Diagnosing problems with WUE and HI You may have some data from an old trial that allows you to roughly calculate WUE from seasonal rainfall.

Try the diagnostics key on the next page to see how well it identifies any problems you observed. The key also highlights in green, some problem areas that are addressed in the chapters accompanying the current chapter.

20

21

On-farm trials to identify and overcome yield constraints in a region A set of trials addressing major yield determining crop management issues is proposed. These are designed with the expectation that they will need modifications to adapt them to local questions and circumstances. They do not cover all possible questions but are seen as a core set that will be added to by other experimenters. Contributions are invited from anyone who feels they can cover a topic of their choice. Guidelines are available explaining how to write a chapter.

Due to climate variability, trials designed to optimize crop production through changed farm methodologies should be repeated for at least three years. In fact, as farmers adopt changed technologies resulting from the trials, further experiments are likely to be needed to readapt the farm to the new situation. Changed economics, considered within the trials, may play a part. This is a dynamic building process.

The proposed trials are independent but at the same time are related showing where addressed issues interact. For example, a change from conventional to zero tillage will require a re-evaluation of varieties and sowing dates.

Climate Risk Analysis and Crop Management Decision Making Tools This chapter provides a background against which to plan any trials. Tools are described enabling farmers to decide on their best planting date, the likelihood of drought and the length of their growing season. The tools inform the farmer of the element of risk they face when taking management decisions involving their predictions of climate. Using the tools provides farmers with some reason for selecting early, medium or late varieties of cereals or even other species to grow to enable them to work optimally within the climatic constraints of their farms.

Optimizing variety x sowing date for the farm The combinations ensure that the varieties reach anthesis at the optimum date for the location. The experiment also assesses the loss in yield due to straying from those dates. It will help to answer questions like: i) when is the optimum period for anthesis? ii) could a slightly longer or shorter cycle variety be better suited to the farm than conventional ones and when should it be sown? iii) what is the yield penalty for sowing later than the conventional date?

Optimizing plant population, crop emergence, establishment and sowing rate

Farmers may reduce their costs of seed and weed control through adjustment of seed rate. Furthermore, crop yields can be increased through improved seeding methods. The trial uses techniques to work out optimum plant population. It is of particular interest in areas using seed rates over 100 kg ha-1 yet failing to achieve good plant populations. It will be useful where there are weed problems, particularly where weeds are emerging before the crop and reducing grain yields.

Optimizing tillage systems on-farm This trial compares crops grown under three tillage systems: full or conventional tillage, minimal tillage and no tillage. It focuses on long- term protection of the soil resource; reduced machinery wear and savings in fuel costs and labour. It is a potentially important trial in areas where water is a major limitation to production and where soil structure is poor and erosion a problem. It could also be important where the optimal sowing date is frequently missed because of time lost through conventional tillage.

Optimizing nitrogen use on the farm Growers should aim to match the supply of nitrogen with the requirements of the crop. Crop requirements constantly change depending on available soil water and rainfall.

The trial will help to identify if nitrogen is a prime limitation to yield. It explains how to calculate how much N should be applied and when it should be applied to ensure that yield responses are more consistent from year to year. It is of particular interest where little or no nitrogen is applied to crops either directly or via rotation crops; where lower yields than reported from regional research trials are obtained; where symptoms of haying-off appear and where rainfall during the season and across years is very variable.

Cropping sequencing (Rotations) This trial will be of particular interest where wheat has been cropped continuously for many seasons; where yields have declined inexplicably; where root diseases or nematodes are a problem; or where there are markets for crops other than wheat. The trial shows how to introduce alternative species (break crops) into the cropping sequence. It explains that interactions between the wheat and break crops, including the roles of nitrogen, water, disease and weather are complex.

22

At each visit, write down the date in the first column opposite the plant growth stage that you have identified. Notes in the second column and pictures on the sample sheets will help you to decide on the plant growth stage. The third column is for your estimate of ground cover as it increases from seedling emergence towards flowering and green leaf number reductions thereafter. The fourth column is for a summary of your management decisions as the crop advances and for dates of your actions. The rows that cut across the full table width are for your calculations of the state of your crop.

It examines these complexities and explores how and why they might influence the performance of wheat in different sequences.

Finally:

Why do any of your research on-farm?

If you are not convinced that you should do your trials on-farm, read this chapter that explains why. It also defines trials, how to approach and design them, and how to interact with collaborating farmers.

Field sheets. A guide to recording your observations The field sheet calculations assume that you have

drilled your crop in 18 cm rows but they will be equally useful if you have used other row spacings or broadcast your crop. Check in the chapter on crop establishment for comparisons between drilled and broadcast crops.

You will need structured records to remind you of any special observations made of your trial crop and any management actions taken. The usual way to make such records is on a date basis in a diary. If you can make them on both a date and a crop development basis on field sheets, this will make them even more useful when the time comes to interpret the trial. Basing observations on development allows you to make comparisons between crops in different locations. You can compare them at the same stage of development.

The example numbers in the field sheets assume also that you take measurements based on several samples within your crop using 1 metre length of rows as the basic sample. You will need a stick or other measuring tool 1 m long as your standard to lay down next to rows so you can count plants or spikes along that measured section. It is good policy to return to the same row sections for subsequent measurements, so preferably mark the random measured sections on your first visit. At least ten measured samples give a good average, but go for a number you can reasonably achieve, and preferably more than ten.

You will find that field sheets are an efficient way of keeping records. They are altogether in just two pages. Imagine unscrambling a whole season’s worth of data spread throughout a diary with several trials all jumbled together. Even worse is a set of data written on your palm; handy but biodegradable.

The field sheets are a way of keeping a crop diary, but you may have a number of harvests that cannot be added to those field sheets. The chapter on “Optimizing variety x sowing date for the farm” provides guides to designing tables for harvests while the detailed section in this introductory chapter on yield components and in “Optimizing plant population, crop emergence, establishment and sowing rate” tells how to sample your crop and calculate yield components from the samples.

You will find example field sheets on the following pages. They already have example entries in red (whiteout those entries and then photocopy the cleaned sheets for your trials if you like). The blue entries are descriptions of good crops that you might like to aim for as the experiment progresses. You can use your observations to calculate whether you are ahead of or behind the blue targets. If your calculated numbers disagree substantially with the blue targets, things that you might like to check in your crop are also listed. Use the pictures of the Zadoks growth stages on later pages to help you enter the stages of development correctly.

The blue targets may be different in your region. Change them if you know the values. If not, do the required trials to estimate them

Take field sheets with you to the field. Use the same field sheet each time you visit that field. At each visit, make notes of unusual conditions of the crop such as herbicide damage, frosted leaf tips, frost damage on spikes and any leaf curling. An accurate record of crop management as the season progresses (sowing, fertilizers, weed and disease control, irrigations, harvest) will help you to identify which factors limit yield most.

23

24

farmer name: field: variety: Abdelkader 407 Massira

date - GROWTH STAGES - MANAGEMENT DECISIONS Zadok

Visual description PRE-EMERGENCE MANAGEMENT

previous crop: chickpea stubble: burnt / retained / other…… soil preparation: chisel plough pre-emergence herbicides:………... date …2 November………….. date / rate /……………………………

18/11 0.0 Sowing sowing method: drilled / broadcast row spacing: … /other. 18 cm sowing rate (kg/ha): 100 / other seed treatment: Vitavax 200……… sowing depth (cm): 0 / 5 / 10 / seed size: small / medium / large 0.3 germination, seed swollen sowing fertilizer: seedbed: / very dry/ dry / moist / wet / 0.5 radicle emerged from seed type:..15-15-15 tilth: / poor / good / excellent / 0.7 coleoptile emerged from seed date: 18 Nov/ crust after sowing: no / moderate / bad

2/12 1.0 Emergence rate: 200 kg/ha

your individual seedling counts : 23 25 29 26 34 32 16 23 24 21 34 32 35 26 27 23 32 38 36 29 35 39 33 32 37 39/m row

(A) your average establishment = 30 target: 25 – 30 plants/m row length at 18 cm or about 150 plants/m2

POOR CROP STAND? CHECK… crust / water stress / low seeding rate / too deep sowing / bad seed / bad seedbed / weeds / / herbicide toxicity / bad incorporation of residues / nutrient stress / waterlogging / birds / insects / diseases

LEAVES ON MAIN SHOOT Ground cover

POST-EMERGENCE MANAGEMENT

1.1 1st leaf more than half visible weeds: species and % cover disease: species and % infected: 14/12 1.2 2nd leaf more than half visible 15% …………………………………….. …………………………………….. 3/1 1.3 3rd leaf more than half visible 25% If herbicide used: If fungicide used:

1.4 4th leaf more than half visible type or brand:……………………. type or brand:……………………. 25/1 1.5 5th leaf more than half visible 40% date applied / rate /……………… date applied / rate /……………… 5/2 1.6 6 or more leaves visible and

stem not elongating 70% results: / poor / good / excellent / results: / poor / good / excellent /

aim for 100 shoots/ m row length your target number of tillers/plant is therefore ( 100 / (A) ) - 1 = 2.3 example for 30 plants / m row length the target is 2 or more tillers per main shoot if using 18 cm row spacing Tillering MANAGEMENT DURING TILLERING AND STEM ELONGATION

3/1 2.1 main shoot and 1 tiller fertilizer/type/: ….urea... fertilizer/type/: ….....…………….. 2.2 main shoot and 2 tillers date / rate / 10 Jan/ 200 kg/ha date / rate / ….......……………....

25/1 2.3 main shoot and 3 tillers 5/2 2.4 main shoot and 4 tillers pest: species and % infected: disease: species and % infected:

2.5 main shoot and 5 tillers …………………………………….. ……………………………………. Stem elongation If pesticide used: If fungicide used:

30/1 3.1 1st node detectable type or brand:……………………. type or brand:……………………. 17/2 3.2 2nd node detectable 85% date applied / rate /……………… date applied / rate /………………

3.3 3rd node detectable results: / poor / good / excellent / results: / poor / good / excellent / 3.4 4th node detectable

ground cover at this stage should be over 90% (= effectively complete) N° of shoots with nodes is a first approximation of final spike number. Target is at least 70 of these shoots/m row

POOR GROUND COVER? LOW SHOOT NUMBER? CHECK too deep sowing / frost / water stress / nutrient stress / / diseases / insects / waterlogging / weeds

CROP NOT HEALTHY? CHECK / frost / water stress / nutrient stress / diseases / insects / waterlogging / lodging / salinity

farmer name: field: variety: Abdelkader 407 Massira

date - GROWTH STAGES - MANAGEMENT DECISIONS Zadoks Visual description Ground MANAGEMENT FROM BOOTING TO HEADING

Booting stages cover

3.9 flag leaf collar just visible Lodging: date…………. slight/ moderate/ severe/…… 4.1 early boot, flag leaf sheath extending 90% % field affected …………….. 27/2 4.3 mid boot, boot opposite collar of 2nd last leaf 4.5 late boot, boot above collar of 2nd last leaf 4.7 flag leaf sheath opening 4.9 first awns visible

Heading stages 5.0 1st spikelet of spike just visible

11/3 5.5 50% of spike visible, mid heading 95% 6.0 full heading, but not flowering

your individual spike counts : 43 66 87 75 56 65 80 75 78 85 /m row length (average = 71) (B) your average spike number:……71…………. target: is 65 spikes /m row at 18 cm or 350 spikes/m2

LOW SPIKE NUMBER? frost / insects / water stress / nutrient stress / diseases Flowering Green

leaf N° MANAGEMENT FROM FLOWERING TO HARVEST

6.2 20% of spikes are flowering, early flowering Lodging: date…………. slight/ moderate/ severe…… 6.5 50% of spikes are flowering, mid flowering 3.5 % field affected ……………. 18/3

6.8 80% of spikes are flowering, late flowering Kernel extending 7.02 kernels near middle of spike extended 20% 7.05 kernels extended 50%

2/4 7.1 kernels watery ripe, clear liquid 3.0 Milk development 7.3 early milk, liquid off white

9/4 7.5 mid milk, contents mostly milky liquid 2.1 7.9 very late milk, half liquid, half solids

your individual kernel counts : 29 33 34 31 35 30 32 33 30 31 (average = 32) depending on variety

(C) your average kernel N° per spike..32. (B) × (C) = 2272 kern/m row (or 12600 /m2) target 7500 – 10000 /m2

SPIKE DAMAGED? LOW KERNEL NUMBER? CHECK / frost / insects / water stress / nutrient stress / diseases / birds Spike distribution in field Dough development

14/4 8.1 v early dough, mostly solids if kernels crushed 1.2 uniform density & height ? no 8.3 early dough, kernels soft and almost dry irregular, no rows missing? yes

8.5 soft dough, finger nail impression not held irregular, missing rows? no 23/4 8.7 hard dough, finger nail impression held 0.2

Ripening Harvest date: 12th May

27/4 9.0 kernels hard, difficult to divide by thumb nail 9.2 harvest ripe, cannot be dented by thumb nail 9.3 kernels loosening in day time

(D) final average kernel weight (mg) …28 mg………….. SMALL SHRIVELLED KERNELS? CHECK / water stress / nutrient stress / diseases / frost / insects / high temperature and wind

Grain yield est (at 10% moisture) = Grain yield actually harvested =

(B) x (C) x (D) / (row spacing in cm) x 1100 = 3.2 t/ha = 2.9 t/ha

Field sheets adapted from Maarten Stapper and David Murray (SIRAGCROP)

25

26

Why do any of your research on-farm? Crops and their environments are highly complex systems with a multitude of variables that change from location to location to different degrees and on various time scales. Because of this complexity, practices optimised for a Research Station might not be so successful when transferred to another location. Though the new location may appear similar to the Research Station, there may be an undefined key limitation or combination of minor but different limitations that constrain production there. In many cases doing a small-scale trial, actually at the new location, will lead to an optimal local farming practice more rapidly than trying additional sub treatments at the Research Station. This article is concerned with the practicalities of on-farm research. It identifies when on-farm research is needed, the benefits and traps of the approach and the types of trials that are likely to be appropriate. It and associated Explore On-Farm articles are intended to provide general concepts, basic designs of trials, methods and sample data plus ways of interpreting that data to improve farm management. An aim is to stimulate ideas. The trials will need to be adapted significantly to match the requirements of the location.

When are local trials needed? • When practices recommended by the

Research Station or central authorities are being followed, but yields remain low

• When a location has special attributes that do not fit the general pattern for the area – the location may be on a hill, in a steep-sided valley, on more rocky ground than the norm, perhaps with wetter or more saline soil

• When a farm or group of farms is less productive than neighbouring farms despite best efforts by farmers

What is a trial? Trials may vary but have the common feature that they test whether changing something in a system alters the variable of interest, and by how much. An example might be checking whether adding a fertiliser to a crop at a particular time increases grain yield.

Good trials have these elements: • A question and, hopefully, an answer • A hypothesis, which is your expected

outcome of the trial and why you collectively think that way

• Treatments, that is, variations to the normal procedure

• A control, which is what the farmer normally does – this control must always be included to compare or check against any treatment

• A design, that is, where treatments are positioned in the field and in relation to each other – the design is very important and may be the difference between getting an answer or not

• A method, that is, the steps taken to get the answer to the question; and

• Measurements and recording of effects by counting, linear or volume assessment, or weighing – these give the trial objectivity and show how big the effects are.

Recording the weather during the study is also very important. Weather changes every season and may affect how well the treatment works. For example, the effect of added fertiliser will be very different depending on the timing and amount of rainfall. Having weather information helps to put the trial into the context of other years. A trial that is not done properly is often worthless. It may seem to give an answer though the answer may in fact be quite wrong. If forward planning or changed farming practice is based on that wrong answer, the farmer could lose income on future crops. No trial is preferable to a bad trial.

Is a trial worth the effort?

Some benefits of trials • The trial may show how to increase economic

returns by a changed practice that increases yield. The change could be quite minor and cost little

27

• The trial may show how to reduce costs by reducing the use of fertiliser, machinery, labour or water without loss in yield

• In the long term there may be indirect benefits of any changed technology that improves sustainability. An example might be a lime application which raises pH in the topsoil this season, thereby releasing previously bound nutrients, improving nitrogen fixation, flocculating clays and increasing earthworm populations over time leading to better incorporation of surface organic matter, improved aeration, better drainage, and finally, higher yield

• Unexpected positive things may be discovered about the farm because the farmer has been thinking about it differently and observing it more closely and critically

• There is close interaction between the researcher and the farming community as they work together. The farmer learns research methods and concepts from the researcher and the researcher learns more about the local cropping systems and about limitations other than the technological ones

• Neighbours will be interested in the trial, will doubtless give their opinions, may even join in and will be the first to adopt then adapt the changed technology if it works

• Communal trials build up a feeling of community and benefits spread beyond the prime aims

Some costs of trials • The trial land will be out of normal use during

the trial, so normal production may not be achieved – including the trial as part of the farm’s normal cropping pattern and keeping the trial small and manageable can minimise this cost

• The yield from the trial could be appreciably lower than normal yield, even perhaps a total loss – can the farm afford that risk?

• Extra inputs will be a cost – items like fertilisers, labour, machinery and perhaps most importantly, time

• There may be a personal and social cost of possible failure on the farmer, the family and friends. Expectations of the study should be moderate, not exaggerated. If anything, it is better to undersell expectations

About the design of your trial

Small and simple is best • Don’t plan too big an experiment – discuss

demands on the farms’ resources and time as well as your own – aim to complete the study in full. Be realistic

• Have a well-defined central question that can be clearly understood by collaborators and answered by your approach

• If a factor other than the main one appears to be strongly influencing the answer, then formulate with the farmer a subsidiary question by adding sub-treatments but

• Do not add so many sub-treatments that the farmer is lost in the complexity

• Don’t use fancy statistics to confuse yourself and collaborating farmers. Don’t be misled by averages. Examine the detail in the data to assess the trends and the significance of any ‘outliers’.

Don’t be fooled about outcomes The farm is about to commit resources to a trial hopefully leading to a long-term benefit. There are some traps the farmer must also be wary of. • Do not subconsciously design the study and

select the information to give the expected answer - the expectations may be wrong

• When collecting and analysing information, have an open mind. One should look for little things that are unexpected - they may be the most important; the obvious things have probably been seen and addressed already

• The trial must be assessed in a detached way. Everything should be written down so that current conclusions can be reassessed later and new ones reached

Use variation within and between fields Many fields have good and less good parts and there are often gradients between the two. Explain to the farmer that if bits of the different parts can be included as sub trials without jeopardising the main trial done on the dominant land type, the trial will probably provide much more information However, exclude the areas that are very unrepresentative of the whole field, or unimportant.

Choosing which varieties to plant In all trials include the variety most used on the farm or locally. If the trial is about comparing

28

varieties, discuss with the farmer which varieties to include and why. Talk about their pest and disease resistance, lodging tolerance including propensity for tillering, fertiliser and water requirements, grain quality and yield, and likely crop duration from different planting dates.

How many plots, how big, where and when? If practicable, use a strip design to answer the question posed by the trial. This is the simplest possible form of design that farmers will readily understand. It is a rectangular strip a few metres wide, marked out within a normal farm crop. The treatments are applied in bands along the strip like the stripes across a scarf, and replicated as blocks at intervals. This design is appropriate for simple trials like checking for fertiliser response of a crop whereas a research layout is better if there are many variables and sub treatments.

29

Plot size – The edge effect Plot size will depend on the question you are asking but larger is usually better. Explain to farmers why observations and measurements should only be on plants towards the central part of treatment plots, avoiding the outer rows that are subject to ‘edge effects’. Plants at the edges of experimental plots have more water, light, and nutrients available to them than the plants within the plots so they often yield better than the norm. These ‘edge effects’ are exacerbated if a clean cultivated border surrounds the plot, so consider surrounding the trial with a border of additional crop if it is not already located within a crop. Examples of the two approaches are shown in the above diagram. Plot size, therefore, depends on balancing the need to eliminate edge effects

against using bigger plots that may make the trial more costly and harder to maintain.

The control plot and how many plots are needed for a trial? There will always be a control plot (C in the diagram), which is the plot planted with the usual variety and with normal farm practice applied. Additionally, there will be at least one treatment plot, the main variation to normal practice. Trials usually have several treatments that represent degrees of variation to normal practice. A fertiliser study might have four treatments (T1-T4) of 0, 40, 80, 120 kg/ha plus a control of 60 kg/ha that represents normal practice. This would give a study of five plots. These would be T1, T2, C, T3 and T4 in the illustration Clarify for farmers why it is wise to have three replicates of all treatments, or more if resources permit. Replicates are copies of the same thing and when compared show how much the same thing varies. So in the fertiliser trial, the basic block would be repeated twice giving three blocks or replicates of five plots each, a total of 15 plots. Only two replicates, block 1 and block 2, are shown in the illustration for simplicity.

Strip plot positions (green) marked out in a farmer’s crop (yellow) and in a fallow field (grey). Controls (C) and treatments (T1-T4) are repeated in two blocks. Harvest areas are the blue-edged squares around the treatment letters.

Depending on the complexity of the question the smallest study would be 6 plots (a control, one treatment and three replicates of both) and the largest would depend on available resources.

Laying out the plots In a strip design of the above fertiliser trial the treatments could be applied end to end along the strip and after a gap of a few metres, repeated in block 2 and then again in block 3. Plots can be of any width within the strip as long as they are wide enough to permit the cutting of at least one harvest in their centre plus a wide enough border of plants around the harvest areas to avoid fertiliser interference from adjacent plots (wash on, root exploration etc). The harvest areas are shown as blue-edged squares in the illustration. Check the descriptions of specific trials for a guide to actual plot size and design. If the question is more complex involving two or more interacting factors, like ‘What is the best variety by sowing date for the farm’, requiring 3 varieties and 3 sowing dates, a conventional research design is generally required.

Location of plots Locate the trial in a uniform average area that typifies the farm or field. Avoid areas that are

likely to be useless for cropping or are in some way abnormal. If the field has small outstandingly good or substandard parts, also exclude these extreme types from the main trial. You could consider adding them as sub trials. Explain to the farmer that responses on the extreme parts will not apply to the farm in general, but could provide useful secondary information about the farm. The trial plots should preferably be within the normal crop. Mark the locations and dimensions of each plot clearly right from the outset and label the treatments. White sticks work well.

30

When should plots be prepared? The trial plots should be cultivated at the same time and in the same way as the normal crop and as part of the normal crop preparation. There are two exceptions. If the trial is about cultivation methods obviously there will be differences. If sowing date is a variable, plots for each sowing date must be cultivated and otherwise prepared separately, each at a similar period in advance of its sowing date. This complicates the study because of the need to get machinery to small plots past other plots and through the normal crop.

Observing the crop as it grows

Seedling emergence Variable results in trials start with variable plant stands due to uneven cultivation and problems during sowing. These problems are unavoidable in some cases, but if they are noted or measured, their influence on the final data can be allowed for. Ten days to three weeks after planting, when all seedlings are emerged, lay a metre-long measuring stick next to the row and count the

number of seedlings along its length. Do this for one sample in every plot in every block. Note down all the numbers with their plot identifier and determine how homogeneous is the density among plots. Is one treatment or one block poorer than others? If the seed has been broadcast, put down four sticks of one metre to form a metre square or quadrat and count seedlings inside the square. A larger or smaller area can be enclosed but its size should be known and the same area enclosed for each plot. If preferred, the metre square approach can also be used in crops that are precision sown in rows by machine.

Treatment plots marked out in a farmer’s crop with white pegs and string Flowering

Flowering is a most critical time for the crop, both in terms of the crop’s sensitivity to the environment and because flowering marks the beginning of the grain filling phase. For each plot make a record of the date of flowering and of the weather at that time. A useful, though less important and less precise date, is that of crop maturity but, if possible, note it down too.

Harvesting The common question in trials is, “How will the treatment affect final yield?” So final yield must be measured accurately and in an unbiased fashion. Discuss some basic rules with the farmers: • Do not leave plants in the field after they are

ripe while waiting for the next treatment to be ready. Without doubt an unexpected event will occur and the ripe treatment will be lost to birds, mice, rain, hail or straying cattle

• Use the same method for harvesting all plots. Do not harvest some by machine and others by hand unless there is an emergency and machinery is unavailable at harvest time

• If a small combined harvester is available, first trim away plot edge rows or borders and then harvest the remainder, plot by plot, collecting the grain (and trash) of each plot separately. The area harvested should be identical for all plots and must be a known area. If it cannot be identical because of driver errors or other factors, then measure and write down the areas actually harvested

• If harvesting by hand, ignore edge rows. Do not select individual plants in plots. Use the measuring stick or metre square quadrat to indicate a metre of crop row or crop area to

31

harvest from the central part in each plot. Avoid parts of centre rows that you know established poorly from earlier observations and notes. In general do not include in the sample sections that have been positively or negatively affected by something other than the treatment

• Cut off the plants at ground level but do not include any soil

• Tie up and clearly mark the plot sample with its plot identification, block, treatment, and harvest date. Bag if possible so that parts are not lost

• Keep the samples together in a vermin proof area

• A main source of error in assessing biological and grain yield is that the amount of moisture in samples may differ between treatments. Dry the samples together in the hot sun, a hot greenhouse, a plastic house or in an oven.

Measuring the results • Weigh the-dried samples. Weigh in the bags if

the bags are uniform weight and then subtract the bag weight from the total. Write the numbers down.

• If samples are hand harvested, thresh and weigh the grain. If they are machine harvested, just weigh. Write down the weight.

• If hand harvesting, calculate Harvest Index (HI) by dividing the weight of the threshed grain by the weight of the whole sample. The result will commonly be between 0.25 and 0.55. If machine harvesting, divide the weight of the threshed grain by the sum of the trash and grain weights

• Refer to the companion Explore-On-Farm pamphlet “Constraints to cereal-based cropping” for details of how to harvest the crop, calculate yield components and estimate water use and water use efficiency.

Recording the weather on the farm If you are collaborating with literate farmers, and there is no weather station close by, think seriously about the possibility of recording temperature and rainfall on the farm.

Temperature should be recorded regularly to pinpoint frosts, heat waves and calculate mean temperatures. All are needed to interpret why a crop yielded well or poorly and matured early or late, and to plan changes in management for future years.

Weekly maximum and minimum temperature over a year at a cold location showing frost-free months

Preferably the farmer should use a maximum and minimum thermometer hung in a slat-sided box painted white or in the shade. Explain how it is used, especially the need to reset it after reading. It is best to read and reset it at least once a week, preferably on the same day each week, and all the year round. The data in the figure were collected using such a thermometer. The amount of rainfall and its seasonal distribution will be an indicator of both potential yield for the farm and the variability of potential between years. It will also enable the farmer to work out water use efficiency. If available, use a standard rain gauge to collect rain. Otherwise use a large straight-sided can and measure by dipping a ruler into the water. If rainfall is very low, the water should be tipped into another much smaller can, calibrated to the large can, for measurement. Once a week measure then discard accumulated rainfall, preferably on the same day each week. Locate the measuring devices in the trial field away from buildings or overhanging trees.

Using long-term weather information

Timing of frost If long-term weather data are available either from the farm or from a local meteorological station, the farmer should note down the beginning and end of the frost-free period for the area and check whether it changes from year to year. Frost can have a devastating effect on yield particularly if it occurs around flowering. On the example figure from a high latitude site, the frost-free period marked as a yellow zone is less than four months.

Amount of water Ask collaborating farmers to assess from long-term weather data, if available, how much water is available in the average season, and how much rainfall varies among years. Explain that a crop needs a certain amount of water during its growth to reach its potential. If that water is not available when needed, yield will be lower than potential. A shorter duration variety might be more appropriate for the area as that will require less water but it may also have lower potential yield.

Writing it all down

Keeping accurate and frequent records is vital Records must be collected and set out in a structured way following patterns determined largely before the trial begins. Structured data and worksheets lead to clear and structured thinking. Haphazard trials with haphazard observations and recording are generally worthless. Discuss and design the worksheets with collaborating farmers before the trial starts. Provide them with hard copies of the final design in a workbook produced for the specific trial. Encourage them to use the workbook for all data and notes.

Discuss some rules for recording data: • Write things down as they occur. Memories

play tricks. • Do not write important information on scraps

of paper. Someone will throw them away. Back of the hand is OK as a temporary measure.

• Dedicate one workbook or worksheet to one trial. Using a diary that mixes experimental data with appointments and shopping lists is messy and can be confusing.

• Enter information collected on different dates in sequential date order. Avoid mixing up dates in the workbook.

• Always label data with its plot identifier and date. What seems obvious today will be confusing next month.

• Organise data in tables (with rows and columns) so comparisons between replicates or treatments can be made by eye.

• Avoid long single columns of data such as used in adding machines that print on paper rolls.

• Where possible, keep a standard design for tables. See the pamphlet on Optimising variety x sowing date for examples.

• When possible, enter the data as it is collected straight into tables so that they can be easily summarised as totals or means at the bottoms of columns and the ends of rows. Rewriting data copied from scraps of paper, sample bags or distributed notes in a workbook can be very time consuming and prone to mistakes.

Defining some characteristics of the farm Start your discussions about the farm(s) itself by trying to define with farmers the extent of the available growing season. This will be an introduction to thinking about the interplay of the environment and any varieties used. Explain to him/her that knowing the growing season is important for using any cropping area to its full, sustainable potential, particularly in a rainfed environment. The critical factors are how long is the local environment warm enough, moist enough and sunny enough to start and support the growth of the crop; and to what extent can different varieties and management approaches alter the length of the season? If available, long-term weather records from the local meteorological agency could well be useful for calculating the average starting and finishing dates for the area and to illustrate the discussion.

Limits to growing season length Discuss with collaborating farmers what you all see as the main limitations to the growing season. Its length is likely to be limited by a cold season such as winter and by the start and end of the rainfall period. It can also be affected by one-off catastrophic weather events, like frosts, hailstorms, rainstorms, gales, and hot dry winds. There are other catastrophic risks like locust plagues and some diseases that are more likely to occur at specific times in the year and may limit the crop season. If

32

such events occur every year they should be avoided by having the crops at stages that are not damaged during these events.

• Was the previous crop harvested early to take advantage of the best market price? Would the current crop yield more if planted early?

In some perennially warm, damp areas, both starting and finishing dates may be largely under the farmers’ control, but in all areas farmers can have a degree of control over season length either through farm management or choice of crop.

Are activities during crop growth timely? • Are fertilisers and irrigation, where available,

being applied at the most beneficial time for the crop and are they used efficiently?

• Are pests, diseases and weeds being controlled at the right time and effectively? Could the outcomes be achieved by using less effort or fewer chemicals in smaller amounts?

Debate with farmers the following questions relating to the growing season: • Could local crops yield better if sown earlier

or later? Why? • Is machinery available when needed? • Are crops maturing fully or are they drying

off before their grain is completely filled? Finding out which, and to what extent, these environmental, crop variety and management factors impact on best use of the growing season is the first reason for doing on-farm trials.

• Are local temperatures too high or too low at any stage for any of the crops? When?

The second reason for trials is to test ideas for overcoming the constraints to production that have been identified.

• Is there sufficient water available at the right times for the crops? When is it limiting?

• Is it possible to modify soil structure to enable crops to use the full potential growing season? What to include in your trial

• Is it possible to grow one crop species after another, perhaps using shorter duration varieties, to make more productive use of the potential growing season?

Debating these questions should have improved your knowledge and the farmers’ awareness of local limitations. Hopefully in your discussions you will have identified the key constraints to productivity on the farm and have debated ways that the constraints might be overcome.

These questions are intended to lead to the conclusion that two main factors determining the potential growing season are the environment and how well the plant material chosen matches the environment. These possible solutions could become the

treatments in a trial. A third all-important factor splicing the first two together is the farmer’s management of the crop. Activities must accurately timed around the changes occurring in the environment and in the crop.

Design the trial with collaborating farmers including these solutions as treatments. It may take several meetings before the trial is finalised to the satisfaction of all parties but it can be an exciting and profitable time.

Some issues associated with timing follow: Throughout, encourage the free-flow of ideas. Do not dominate or be elitist in these interactions. Is the land ready for planting at season start?

• Is the land too dry or too wet to allow the crop to be planted at the optimal time? Can a change in tillage or planting methods overcome the problem?

33

• Are previous crop residues causing delays? Can they be better managed in advance, removed or incorporated or used for mulch?

• Was the previous crop harvested in time to allow planting of the current crop at the optimum date or should a shorter duration variety have been used?

How to discuss the results when the trial is finished When you talk with collaborating farmers about the trial results do it in such a way that the numbers discussed are paralleled with a mental picture of the crop. Stick initially to raw data and simple averages and trends. Get the farmers to think about why a particular plot has done better or worse. Debate its state of weediness, whether it was wetter, whether a carcass of an animal was left there a few years ago and so on. Use the farmer’s knowledge and notes in the workbook to involve him/her in the numbers and guide him/her to do the interpretation. The farmer’s confidence and enthusiasm will grow rapidly. This will build up a much more reliable picture of what the trial means than a statistical analysis. If appropriate, use that analysis yourself to substantiate the farmer’s interpretation and if you decide to publish the results of the study. Follow up these discussions about interpreting the data with considerations of what these interpretations mean to future practices on the farm. Always link the data back to practicalities. As the discussions proceed, get the literate farmers to note down the conclusions, but make your own notes. Even rough notes are very useful for jogging the memory and can save a lot of thinking time later when the memories of the discussions have dimmed. Send a copy of your conclusions to the farmer as soon as you have tidied them up. Do this soon while details of the study are still fresh. Consider following up the study with an on-farm field day to involve other local farmers in the results. Encourage them to air their views on the weaknesses and strengths of the project and to suggest ideas for future collaborative studies.

34

Optimizing Variety x Sowing Date for the Farm The trial finds the best combinations of sowing date and variety for the farm chosen. The combinations ensure that the varieties reach anthesis at the optimum date for the location. The trial also assesses the loss in yield due to straying from those dates. This information enables farmers to decide whether the loss is acceptable in the light of other priorities on the farm. The approach outlined guides trial managers through designing a study suited to their special requirements. It then explains how to conduct the trial and finally interpret the results. Real life trials are used to illustrate likely responses. Companion chapters you will find useful include “Why do any of your research on-farm” for general information on trials, “Constraints to cereal-based rainfed cropping in Mediterranean environments and methods to measure and minimize their effects” and “Optimizing plant population, crop emergence and establishment” for harvesting methodologies, “What is the best cropping sequence for the farm” if yields are declining on-farm, , and “Risk Analysis and decision making tools for North Africa” for weather analysis.

Which farms could benefit from these trials? ☺ Those farms where the planting date to

achieve best yield with the present variety is not known. Optimum planting date may not have been tested but just guessed;

☺ those farms where it is important to know how much yield is lost by missing the optimum planting date. Farms that might commonly plant late because other crops have still to be harvested or the ground cannot be ready on time;

☺ those farms that may benefit from varieties with longer or shorter duration or where this option has never been thought about or tested;

☺ those farms in a region that has new or promising varieties but the varieties have not been tested locally or on the farm.

The best variety Farmers know that choosing the best varieties is the first and probably most important step towards high yield on the farm. But what may not be recognized are the characteristics that are needed in that best variety. What should the farmer look for? Most importantly, the best variety must exactly fit around the weather constraints of the farm. Different weather constraints or the same constraints at different times will possibly require a different variety for best yield.

The farmer must know what these constraints are, what they do to the crop and when they occur in the season. Before starting any trials, check through the following questions on constraints. Step 1. Work out the dates of constraints Does weather limit season start? • On what date approximately is it the right temperature to allow seeds to germinate and grow? When does it warm up enough or cool down enough? • Is the location dependent on opening rains before the land can be worked or before the crop can be sown? Or is sowing usually held up because the ground is too wet? • Is the weather limitation always the same and what is the approximate date when it is no longer a constraint? Is this the usual sowing date? How much does that date vary between years? Is weather limiting during the season? • Are there weather limitations like very overcast conditions, mists and fogs in the period just prior to normal anthesis date? They can result in low grain numbers and poor yield. • Can temperatures be very low or very high around anthesis leading to low grain numbers? How many years in ten? If anthesis were slightly earlier or later would it miss these conditions? Weather at normal anthesis date • Is there always enough water stored in the soil to fill the grain? Or is there a good probability of sufficient rain late in the season? Do current

35

varieties have problems with pinched grain indicating that water for grain filling may be lacking? • Have there been problems with small grain? How many years in ten? This is often due to excessively high temperatures late in the season that shorten the grain filling period. What are the temperatures at that time? • If anthesis were earlier would the constraints be avoided? Diseases or pests • Are these likely to be prevalent at particular times? When, on what dates? Are certain stages of plant development more sensitive to the constraint? Is the variety normally used resistant to those threats? Step 2. Time development to fit around the constraints Once the approximate dates of the constraints to growth and yield have been identified, the next step is to choose a variety that times its development to avoid or minimize the negative effects of those constraints. The best variety is one that can be started at or after the opening of the season and then time its anthesis date to avoid frosts and too dull or too hot conditions, and finally and often most importantly, to start and complete grain filling when water and temperature are not constraints. On top of all these timing requirements, the best variety must have resistance to local diseases, high yield potential and good quality grain.

36

Early and late varieties Early wheat (Triticum aestivum) varieties are those that reach flowering in a relatively short time from planting while late varieties take longer to complete their life cycle.

Early variety = short duration Late variety = long duration.

How quickly a variety develops through its vegetative stage into flowering and then through kernel filling depends on its inbuilt responses to particular components of the weather. The most important response is that to temperature. All varieties develop faster and mature earlier as temperature rises. The very earliest varieties available respond little to changing day length and have for most practical purposes no vernalization response. They only speed up or slow down as mean temperatures change. In general they can be

planted later because they can complete their life cycle quickly. As they go through their cycle so quickly they have limited time before anthesis to build up a lot of vegetative growth and the associated structures for very high potential yield. Late varieties might have either or both day length and cool-temperature (vernalization) response. They are best suited to early planting because they can take a long time to complete their life cycles. They take a long time to get to anthesis so can generate a lot of leaves and tillers and high potential yield. Whether this yield can be realized depends on components of the environment, particularly water and nutrition.

The best sowing date Does yield change with sowing date? Though there might be no choice on the first sowing date for the season because it is dependent on first rains or other season break factors, there are choices after that time. Researchers and farmers over decades have examined the effects on yield of changing planting dates. Invariably they have identified the optimum planting time for their chosen variety at their place and found that delaying planting beyond the optimum time results in a decline in yield. Experimental results from Algeria and Morocco showed that grain yields averaged 24 percent less when planting was four weeks after the optimum (early) time (see the figure) .This was because late planting pushed anthesis

How sowing date at the optimum time and 4 weeks later affects yield of different types of varieties (late, medium and early varieties)

and grain filling into hotter and drier conditions thereby shortening the grain filling period and resulting in smaller grains. Many researchers have argued that the correct sowing time for a variety is the date that gets the crop to anthesis at the optimum time. In the above-mentioned case, this would be before hot conditions constrain yield. As different varieties take different times between sowing and anthesis the planting date required to reach optimum anthesis date on time must differ between varieties. A new variety should be calibrated to its new area to ensure that it is sown on the date that gets it to anthesis at the optimum date for the area. Does yield change with anthesis date? Just as yield can change with sowing date, yield can change with anthesis date. In an area in southern Australia where yield is high, Stapper and Fischer (1990) found that anthesis date was optimum over only two weeks. Reaching anthesis prior to that period could reduce yield to half that because of frost damage. Reaching anthesis after the optimum period reduced yield by more than 5 percent for every week’s delay.

This 5 percent reduction with delay was kept small because they could irrigate. By contrast, missing the optimum anthesis date might be disastrous in areas such as North Africa depending on stored soil moisture or spasmodic rain to fill the grain. The same would apply if temperatures and rates of evaporation were rising rapidly with advance of the season. So, for low-yielding Australian crops, the reduction in yield due to delaying anthesis beyond the optimum date was 24 percent per week’s delay in a dry year and from 13 to 9 percent per week’s delay in a less demanding wheat zone. These trends are more dramatic than those for Algeria and Morocco but in the same direction. In the chapter on optimizing N a delay in anthesis of only eight days after the optimum reduced yield from 3.4 to 2.1 t ha-1 or 38 percent.

Make opportunities to discuss when yield is formed in the crop life cycle. The progression is plants/m2, spikes per plant overlapping with spikelets per spike, grains per spikelet and finally the size of each grain. Talk about how weather will affect yield at particular stages of development. The figure will help with the discussion. It shows when parts are developing in relation to how the crop looks. It also shows (in red on green) the danger periods for some types of stress. For example, if the variety grown on-farm is exposed to frost during the critical period around anthesis when grains (kernels) are just forming, it will yield badly. That variety should be replaced with one that reaches anthesis well after there is a likelihood of frost, or the sowing date should be adjusted.

37

How are sowing date and anthesis date related? An example trial Farmers know they control the sowing date but may believe they cannot change anthesis date. The following example (reworked from Stapper and Fischer, 1990) shows how anthesis date may be changed. Anthesis date can be altered by two means, by changing the date of sowing to a period of different weather and by changing the variety or species to one that develops faster or slower. In this trial there were two varieties. One was early (i.e. fast) and one was late (i.e. slow), needing long days to develop quickly. Of the four sowings, one to three were made over progressively declining temperature and the last sowing (sow 4) during rising temperature. (To see this check the positions of the pink vertical arrows on the weather figure immediately below. Month ‘a’ is the month of the first sowing.

• It took plants from the first sowing (sow 1) almost twice as long to reach anthesis as those from the last sowing (sow 4). To see this compare the numbers on the development bars on the figure and the lengths of the bars; e.g. 140d versus 80d. This is primarily because photoperiod was longer throughout growth of late-sown plants and temperatures were rising (see the temperature and photoperiod curves).

How sowing date and weather changed anthesis date. There were four sowings over four months (sow 1-4) of an early variety (red/pink development bars) and a late wheat (blue development bars). Anthesis date is shown on the bars as circled symbols. Number of days from sowing to anthesis is shown by the value on each development bar (e.g.140d). The optimum anthesis period (period for best yield) in the region is a vertical yellow band. Note that some sowings missed that anthesis period. Weather and bar development figures have the same time scale.

• The late variety was always later to reach anthesis than the early one from equivalent sowing dates. The difference ranged from 30 days from sow 1 to 20 days from sow 4. • Though sowing dates spread over 4 months, anthesis spread over less than two and a half months. In short, as it gets warmer and day length increases, everything goes faster and varieties become more similar in real time. Despite this, variety and sowing date still give a great deal of flexibility when aiming for a particular anthesis date. In this example with extreme varieties, the optimum period of anthesis for the location, shown as a yellow band, could be targeted by sowing the early variety 100 to 120 days before that period and equally well by sowing the late variety 150 to 170 days before the period. So there was a ten-week window for sowing date to hit the optimum anthesis date by using these extreme varieties. Many varieties are available with rates of development being intermediate between the extreme varieties described, so that a variety could be selected to reach the optimum anthesis period from a wide range of sowing dates.

38

Designing and doing your trial

General aims and approaches By using varieties contrasting for life cycle length (early, normal and late) and planting at times that span the normal planting date, the trial aims to check whether there are better combinations of variety and sowing date than normally used on the farm. Primarily it aims to identify the optimum date of anthesis. The trial also provides an opportunity to try out any promising new varieties for fitting to the farm’s constraints.

Designing for the constraints of the farm The design shown here has three varieties (early, normal and late) sown on three dates (early, normal and late) and replicated three times (i.e. across three blocks) making a total of 27 plots. For a farmer who is not used to experimental approaches it can be reduced to 9 plots by ignoring replication and harvesting each treatment strip as one.

A realistic number of treatments On a farm with more resources the trial can be increased to five varieties by three dates (15 plots). The trial must go right through to completion so must be a size that is both manageable and economically feasible. That said, a study with more varieties and planting dates will identify the best combinations more precisely. Plot area The simplest design is to treat the whole strip as the unit for harvest. In that case the strip should be at least one drill run wide and be harvested by machine taking the whole strip. The strip then should be at least 10 m long as the harvesting machine will have to be emptied after completing each treatment.. Ignore edge effects. If hand harvests are contemplated, each plot should be large enough to allow for a harvest at maturity of at least a 2 m² area surrounded by an uncut border 0.5 m wide. Three cuts within each treatment should be done. Details for how to take such harvests and how to manage the samples are in the pamphlet “Optimizing plant population, crop emergence, establishment and sowing rate”. A plot layout for determining the best combination of

sowing date and variety for the farm. Variety 2 and second sowing are those normally used on the farm. Because the varieties are sown as adjacent strips the three replicates or blocks are just nominal separations across the treatments. Replicates can be ignored if preferred and the whole strip harvested as one

Direction of drilling the crop If the example design is being followed and if the crop is being drilled, each variety is sown lengthwise in a run from the start of the first block right through to the end of the third block (three plots). The three varieties are sown in adjacent strips. The nominal blocks can be ignored if a simple trial is preferred. At the first sowing drill the first outside border using the first variety and the last outside border with the last variety. This means there will be two adjacent drill runs of the first and last varieties (see the red arrows on the diagram). If this is a problem, perhaps because of shortage of seed, use the normal variety for all borders. Follow the general approach at the second and third dates of sowing, except that the varieties are sown in a different order. Plant the trial within the farmer’s normal crop; small studies, particularly in a fallow field, can produce unrealistic results because of ‘oasis effects’. The farmer’s crop should abut the borders of the trial. Choosing varieties and sowing dates For the three varieties used in this trial, choose a very early variety, the normal variety for the farm and a late variety.

39

For sowing dates choose the normal one for the farm and have one about two weeks before and another two weeks after that date. Decide whether a 28-day span is too short or too long to be useful and adjust the time accordingly.

40

Preparing the land and sowing Plots for each sowing date must be cultivated in the same way. If at all possible, cultivate each at a similar period in advance of its sowing. Cultivating at the one time might disadvantage later sowings because turning the soil accelerates denitrification and more importantly wastes stored water. This complicates the study because of the need to get machinery to plots past other plots and on some farms may not be a realistic requirement. Apply fertilizer with the seed in each case and if top dressing with fertilizer, this should be done at an equivalent crop stage in each sowing treatment, not at an equivalent time after sowing. For consistency the same amount should be used in each treatment even though this may not seem sensible.

Observations to make during the trial • Weather readings. These should be taken on a regular basis, preferably on the same day each week, not just when something extreme occurs (see introductory chapters and the example

worksheet later in this chapter). Also check for availability of a maximum-minimum thermometer and a rain gauge. • Crop observations during the season. The critical observation to record is the date of anthesis then the date of maturity. Counts of seedlings emerged are also important as variation in numbers between treatments may complicate interpretation of the trial. Counts of spikes are also important because they are directly related to yield. See the pamphlet “Optimizing plant population, crop emergence, establishment and sowing rate” for some details. • Weed infestations, diseases and pests should be noted and controlled as in a normal crop. If possible a record should be kept of their severity for all plots using a rough score out of ten. Different yields between treatments may be due to factors like these that are not intended as part of the trial and differ between seasons. Such observations help in a final assessment. Harvest Methods for harvesting trials at grain maturity are in the chapters “Optimizing plant population, crop emergence, establishment and sowing rate” and “Constraints to cereal-based rainfed cropping in Mediterranean environments and methods to measure and minimize their effects”.

Example worksheets Whenever possible, worksheets within a trial should use a common design. The first worksheet is for entering counts of seedlings per square metre in an experiment with three sowing dates, three varieties and three blocks. If there were a fourth sowing it would be added as four more columns at the right hand end of the table. Numbers for sowing 1 are examples. If you do not use replicate blocks, use the space in the table for three areas of observations within the treatment. Mark each area and return to the same areas when you do counts of spikes before heading. Use the field sheets in “Constraints…” pamphlet to summarise activities. If seedlings have been counted per metre row length rather than per square metre, convert the values shown in the table to those equivalents by multiplying by row spacing in cm/100. So for 18 cm drill rows, 222 m-2 becomes 40 seedlings per metre row. The layout of the table is intended to make comparisons between data sets easy. It can be seen straight away in the example data that emergence was poor in Block/Sample 3 and that variety 3 was poor in all samples. The worksheet for grain yield below copies that for seedling emergence.

GRAIN YIELD PERSQUARE METRE Experiment name: variety x

sowing date Recorded by: JockMc Tavish

Sowing 1 (S1) Sowing 2 (S2) Sowing 3 (S3)

Variety (v) v1 v2 v3 av v1 v2 v3 av v1 v2 v3 av

Block 1 89 100 189 126

Block 2 111 133 200 148

Block 3 100 111 150 120

average 100 115 180 131

variety names v1: early sowing dates S1: early harvest dates S1: Jan 20v2: normal S2: normal S2: Jan 26v3: late S3: late S3: Jan 31

SEEDLINGS PERSQUARE METRE Experiment name: variety x

sowing date Recorded by: JockMc Tavish

Sowing 1 (S1) Sowing 2 (S2) Sowing 3 (S3)

Variety (v) v1 v2 v3 av v1 v2 v3 av v1 v2 v3 av

Block 1 206 189 128 174

Block 2 222 250 156 209

Block 3 156 172 133 154

average 194 204 139 179 variety names v1: early sowing dates S1: early counting dates S1: Oct 24

v2: normal S2: normal S2:S3: v3: late S3: late

A table for biomass (weight of grain, straw and leaves) would be the same again as would that for anthesis dates. A familiar design leads to fewer mistakes and writing down the numbers is faster as you get to know the positions of cells on the sheet. As farmers will be using these worksheets, explain that having a standard layout with a box, or cell, for each number not only makes entering data easier but also makes your job of transferring data onto a computer spreadsheet easier, if this is an option and your preference. It is always useful to have a summary worksheet so that you can see the relationships between the main variables of the trial. Its data will be extracts from the raw worksheets, usually the averages taken from the bottom rows. It might look something like the following table that shows the first sowing.

41

The worksheets for the rainfall and temperature numbers could be set out as below with a cell for recording each variable on a standard day each week. The average temperature (av °C) is worked out as the maximum plus the minimum divided by two. In its simplest form with a base of 0°C, thermal time summation for each week is the average temperature multiplied by seven (days in the week) and the units are °C d.

sowingdate

wheatvariety

seedlingnumber

(/m2)

anthesisdate

grainyield(g/m2)

totalbiomass

(g/m2)

harvestindex

sowing -anthesis(days)

sowing -anthesis

(°Cd)

1 Sept v 1 194 1 Nov 100 286 0.35 61 1000

v 2 204 10 Nov 115 245 0.47 70 1200

v 3 139 15 Nov 180 462 0.39 75 1300

Once you have the results Using graphs to calculate optimum sowing and anthesis date and yield loss for being late

Association between sowing date and grain yield (left) and anthesis date and grain yield (right). The example has two varieties (late and early) and four sowing dates (S1 – S4)

Once the data are available, the first things collaborating farmers will want to know is what

was the best sowing date and the best variety? There will have been an indication of the answers

Rain, max & min temperature& heat sums each week Experiment name: Variety x

sowing date Recorded by: JockMcTavish

date 18Nov

25Nov

2Dec

9Dec

16Dec

23Dec

30Dec

6Jan

13Jan

20Jan

27Jan

3Feb

10Feb

17Feb

24Feb

3Mar

10Mar

17Mar

24Mar

31Mar

rain (mm) 5 37 0 0 12 0 0 40 22 7 0 0 0 0 74 22 0 0 0 0

max °C 30 31 30 28 23 25 26 22 20 27 25 25 23 18 23 30 32 29 34 38min °C 14 13 10 12 9 11 8 10 8 9 9 11 13 6 11 12 14 15 14 18av °C 22 22 20 20 16 18 17 16 14 18 17 18 18 12 17 21 23 22 24 28

week °Cd 154 154 140 140 112 126 119 112 98 126 119 126 126 84 119 147 161 154 168 196sum °Cd 154 308 448 588 700 826 945 1057 1155 1281 1400 1526 1652 1736 1855 2002 2163 2317 2485 2681

when the plots were harvested, but no detailed idea of the trends. Commonly the best way to show trends is on a time-based graph. Find the sowing dates, anthesis dates and yield data for each plot on the worksheets and then calculate the average yields and anthesis dates across blocks. Put those numbers on a figure like the left hand one, scaling it to the appropriate dates and yields. The conclusion from the example left hand figure is that: the late variety is best as it yielded more than the early variety from all sowing dates; the best general sowing date is date 2, though the late variety did equally well from date 1; sowing dates 3 and 4 are past the peak for both varieties with yield declining at about 270 kg per week delay in sowing (around 7 percent per week). Interpret the farms’ data similarly. If three sowing dates were used the pattern may be less clear than in the example. Return to the example. Sowing 1 is too early for the early variety. The right hand figure of the same yield data, plotted against anthesis date instead of sowing date, shows why. According to that figure the optimum anthesis date for the area is about 25 January (when the curves for both varieties reach their peaks). The early variety from sowing 1 reached anthesis two weeks too early for this date and may have lost yield through frost or by having too little biomass accumulated by anthesis to support high yield. Together the curves indicate that the late variety might have yielded best from a sowing date between sowings 1 and 2. See whether there is any indication of an optimum anthesis date amongst varieties for your data by drawing curves through the data for each variety separately. Do the curves peak somewhere near the same date? The anthesis date figure is very useful for forecasting best sowing dates for varieties that were not included in the trial. Whatever the variety, they must aim to have anthesis at the optimum time. An earlier variety than used in the example would need to be sown possibly later than sowing 2 to reach anthesis at the optimum time. A really late variety might have to be sown before sowing 1. The common peak anthesis time is what should be aimed for each year unless the weather data collected by

farmers indicate that the year of the trial was abnormal.

42

Using heat sums to calculate latest sowing date once you know optimum anthesis date The following calculation is approximate but can indicate what is the very latest date that a short duration variety can be sown to reach anthesis at the optimum anthesis date. What this calculation needs is mean temperatures (average of maximum and minimum) for the area and base-temperature for the crop. Base temperatures change with development and between genotypes (Slafer and Rawson 1995), but for this type of study a base of 0°C will provide meaningful conclusions Crops set their rate of development primarily by their inbuilt temperature clock. Each day during growth this clock adds the current day’s average temperature to its running total and when the total reaches particular sums the crop can start its next stage of development. These totals differ between stages and between varieties. They are referred to as heat sums, thermal time or day degrees and the units are °C d (°C each day accumulated over a number of days). For example, on a day with an average

temperature of 20°C a crop accumulates 20°C d>0°C to be added to its running total. The very shortest duration wheat varieties need to accumulate about 1 200°C d to develop between sowing and anthesis. That could be 100 days with an average temperature of 12°C (100 days x 12°C = 1 200°C d) or 60 days at 20°C (60 days x 20°C = 1 200°C d) or some other combination. To work out the approximate latest sowing date from the optimum anthesis date, successively total the day degrees for the days prior to the optimum anthesis date until the desired total for the variety is reached (1 200°C d in this case). The date at which the total heat sum (1 200°C d) is reached is the latest sowing date for a short duration wheat. Collaborating farmers will probably have taken temperatures weekly, so will accumulate mean °C x 7 weekly. The example in the table is based on weather data from the example trial. The optimum anthesis date there was end of month “f”. To reach anthesis at that time a 1 200°C d variety would need to be sown early in month “c” (in the table see the backward summation of °C d from month “f” that reaches 1 225 °C d in month “c”.

Working out sowing date from anthesis date using temperatures collected for the areaThese numbers are from the example trial “How are sowing date and anthesis date related?”

month: b c d e f

1 average temperature (°C) 12.9 8.6 7.9 10.6 13.12 days in the month 31 30 31 31 33 °Cd in the month (1 x 2) 400 258 245 329 3934 °Cd totalling back from month "f" 1624 1225 967 722 393

0

Old weather records could save you work If local farmers or neighbours have been collecting weather information (temperature and rainfall) over the years, or if you have access to long-term weather data for the region, all of the trial may not be needed. You can calculate best sowing and anthesis dates from such data if you know enough about the varieties. You can also assess how often, e.g. how many years in ten, yield is likely to decline because of variations in weather.

You should check the pamphlet “Risk Analysis and decision making tools for North Africa”. This may have the weather data you need for your region already analysed. The necessary frost-free period The frost-free period should extend from when the spike starts to emerge preferably until well into grain filling. If temperatures go down to 0°C for even one night during this time some florets will be sterilized and yield will decline. Several

43

consecutive nights of such temperatures or just one -5°C could lose much of potential yield. You can work out roughly how long this frost-free period needs to be. Depending on average temperature the crop needs two to four weeks before and two to six weeks after anthesis without frost (warmer equals shorter periods). Using the long-term weather information, note down the beginning and end of the frost-

free period and check if it applies in all years. In how many years is the frost-free period too short? If the frost free period is commonly much longer than the required four to ten weeks, still note down the starting and finishing dates. These observations will give you your frost-free period and its variability in starting and finishing date from year to year. Amount of water for grain and biomass Wheat requires at least 200 litres of water during grain filling to produce each 1 kg grain and up to twice that amount in very hot, dry areas. Scaling that ratio up, a 3 000 kg (3 t ha-1) crop will need 600 000 litres (0.6 ML) either from stored soil moisture or from rain. This is equivalent to 60 mm rain (1 mm rain falling on 1 m2 is 1 litre or over 1 ha is 10 000 litres). The crop also needs water to grow prior to anthesis. For a 3 t ha-1 yield the crop will need to have more than 6 t ha-1 of biomass produced by

anthesis, requiring around 160 mm rain or equivalent stored moisture (a clay near field capacity could contain this amount). If the long-term weather data show this amount of water is not available in most years, yields will be less than 3 t ha-1 and a short duration variety might be most efficient for your area. There is no point producing a lot of biomass before anthesis, as is possible with a long duration variety, if there is no

water remaining after anthesis to fill the grains.

Evaporation If the air is very dry and it is hot, sunny and windy, crops potentially lose a lot of water, up to 10 mm a day. If it is cool and sunny losses are much lower. Under the former conditions crops are inefficient in their use of water, under the latter they can produce more biomass or yield for each mm of water they transpire. Consequently, if water is a critical limitation, it is better to avoid these periods of high evaporation during the anthesis and grain filling period. In warm areas where frost seldom occurs during the cooler period of the year, this cool period of lower evaporation might be the best time for grain filling. The coolness also extends the period of grain filling.

Working with numbers in a trial The following trial is entirely imaginary, designed as an introduction to working with numbers and for looking at, thinking about and interpreting a moderately complex set of data that you can discuss with collaborating farmers. Description of the imaginary field The example small farm has no uniform areas. The best field slopes from a small hill where the effective soil depth is about 40 cm, down to a shallow hollow with deep soil that can be damp for some time after rain. The hollow is prone to frost in some years and collects mist after cool nights late in the season. The imaginary trial was planted in a strip to include the hollow (Block 1) the slopes (Block 2) and the exposed hill (Block 3). Due to the variability of the field the blocks could also be thought of as subtrials 1, 2 and 3.

The varieties The varieties included: a very short duration one (early), not normally used in the region (v1); two local high yielding mid season varieties (one flowering 5-12 days later than the other depending on planting date (v2 and v3); and a long duration (late) variety known to be variable in yield from year to year (v4). The sowing dates Of the four sowing dates used (S1 to S4), the third (S3) is considered best for the region; any later sowing (S4) would expose the crop to high temperatures and hot winds during grain filling. S1 is probably too early.

44

The yield data The large table for grain yield with its four varieties by four sowings by three blocks will appear very complicated to farmers at first sight. So start your discussion by looking at this small table. It summarizes the trial by averaging its

three blocks (i.e. three sub-trials). These summary data show that the general expectations for the imaginary farming area were supported as follows: • the best sowing date was sowing 3 as it gave the highest yield on average for the four varieties (last column, 3.4 t ha-1). The first sowing was too early with only 1.3 t ha-1; • a local variety, the medium late v3, was the best across sowings, yielding 2.5t ha-1 (bottom row). • However, there were new and interesting findings: • the long duration variety (v4) did best from the early sowing (1.6 t ha-1), but very badly from the last sowing (0.7 t ha-1 ); • the short duration variety (v1) did acceptably well and by far the best from the final sowing (2.2 t ha-1). This indicated that a late sowing using a short-duration variety would be an option if the ground was not ready at the normal date of planting.

The individual data from the blocks or sub-trials (now see the large workbook table) also reveal some interesting trends that tell something about crop responses to different conditions in the blocks. Recall that Block 1 is in a dip in the field, Block 2 is on a slope and Block 3 is at the top of a rise. Overall, the blocks were very similar in performance (2.2, 2.3 and 2.4 t ha-1; you have to calculate these numbers from the big table) but the patterns across blocks from different sowings were very different. The block differences can be interpreted as follows.

Grain yield (tha-1) for 4 sowings of 4 varieties (v)

v 1 v 2 v 3 v 4 varietyaverage

Sowing 1 1.0 1.2 1.5 1.6 1.3Sowing 2 1.9 2.7 3.5 3.6 2.9Sowing 3 3.0 3.4 3.7 3.5 3.4Sowing 4 2.2 1.8 1.3 0.7 1.5sowings-average 2.0 2.3 2.5 2.3 2.3

The early sowing: severe frost damage caused the short duration variety to yield very poorly from sowings 1 and 2 in the low-lying area of Block 1 (0.8 and 1.4 t ha-1). On the windy exposed hill of Block 3, frost damage to these plantings was far less pronounced (1.2 and 2.3 t ha-1). By contrast the long duration variety did better in Block 1 than Block 3 from the early sowing (1.7 and 1.3 t ha-1). This was because it flowered after the frost events could cause damage, but an unrelated problem of high wind caused the exposed crop in Block 3 to lodge, so that it yielded less well. The last sowing: a contrast between Blocks 1 and 3 also occurred at the last sowing because of the difference in terrain. In this case, Block 1 was superior because, being in a dip with deeper soil, it had more water to take it through the rapidly drying period of grain filling. The late variety, because it filled its grain late in the season, was a disaster from the last sowing in Block 3 because of water shortage (0.3 t ha-1). Being baffled with too many numbers The above-mentioned analysis is concocted to illustrate that there can be a lot more information in a data set than shows in averages. Indeed, farmers might find the host of numbers

Grain yield converted to t ha-1 for all four varieties (v1 early, med. early, med. late, v4 late) four sowings (S1 early, med. Early, med. late, S4 late) and three blocks as would be recorded in the workbook. Due to the variability of the field the blocks could have been treated as sub-trials

GRAIN(t/ha) Experiment name: Imaginary Recorded by: U.Tom Cobbleigh

Sowing 1 (S1) Sowing 2 (S2) Sowing 3 (S3) Sowing 4 (S4)

Variety (v) v1 v2 v3 v4 av v1 v2 v3 v4 av v1 v2 v3 v4 av v1 v2 v3 v4 av

Block 1 0.8 0.9 1.3 1.7 1.2 1.4 1.9 3.2 3.2 2.4 3.1 3.1 3.4 3.2 3.2 2.4 2.0 1.6 1.2 1.8

Block 2 1.0 1.2 1.6 1.8 1.4 1.9 2.8 3.8 3.7 3.1 2.9 3.5 3.8 3.7 3.5 2.0 1.8 1.3 0.5 1.4

Block 3 1.2 1.4 1.7 1.3 1.4 2.3 3.5 3.6 3.9 3.3 3.0 3.5 3.9 3.6 3.5 2.2 1.5 0.9 0.3 1.2

Average 1.0 1.2 1.5 1.6 1.3 1.9 2.7 3.5 3.6 2.9 3.0 3.4 3.7 3.5 3.4 2.2 1.8 1.3 0.7 1.5

45

collected in a trial baffling, just because there are so many. If so, condense the numbers into a few familiar averages, like average grain yield for the whole trial. This will help digest a big data set. This can be compared against normal yield last year and so on. You can then work back into the slightly more complex but still condensed data looking at, for example, average yields of different varieties. Once you have thought and discussed this and its meaning you should drop into the next level of detail until you get right back to the raw data. Explain that bit by bit you can unravel the full story. It is useful, however, to start at the gross level of overall averages. Weave in the detailed observations made by the farmer as you go. Working with harvest index Grab samples were taken at crop maturity in the imaginary trial just as described in the pamphlet “Optimizing plant population, crop emergence, establishment and sowing rate”. This meant that the data could be used to work out harvest index (HI). HI is the proportion of grain to total above-ground matter (grain, straw and leaves) and indicates how efficient crops are in producing grain. An efficient crop produces around 50 percent of its above-ground dry weight as grain. This can fall below 30 percent (HI of 0.3) in frosted, lodged or droughted crops and lower still if some problem like boron deficiency has sterilized some ears. In the current trial the usefulness of working out HI is that it will indicate if there were any problems, shown as depressed HI, during the growth of the crop. The first step is to have the total above-ground dry weight data in a worksheet table like that containing grain weight data. Then make a blank table in which to write down the HI numbers. The HI table shows grain yield, biomass yield and HI for sowings 1 and 3 of Block 1 in the fictitious trial. The grain yield data have been shown in the earlier large table. The bottom two values in the last column of the HI table are average harvest index for sowings 1 and 3. They demonstrate that there was very little grain produced for the amount of biomass in sowing 1 but that sowing 3 was very efficient. As discussed, this was because frosts reduced fertility more than biomass in early and medium early crops of sowing 1 and the medium late and the late crop was lodged.

Check the Diagnostic Key diagram in the chapter “Constraints to cereal-based rainfed cropping in Mediterranean environments and methods to measure and minimize their effects” to see what might cause low HI in wheat crops.

Working out harvest index from grain yield & biomass (t ha-1) for sowings 1 and 3 in Block 1

(v1)early (v2) (v3) (v4)

late av

grain yieldSowing 1 0.8 0.9 1.3 1.7 1.2Sowing 3 3.1 3.1 3.4 3.2 3.2biomassSowing 1 4.7 5.0 6.8 10.6 6.8Sowing 3 5.7 6.0 6.7 6.7 6.3HI %Sowing 1 17 18 19 16 17Sowing 3 54 52 51 48 51

Further reading Bouzerzour, H. et Oudina, M. 1986. Effet des dates de semis et densité de semis sur le rendement du blé et de l’orge dans la région de Settif. Céréaliculture 15. Karrou M. 2003. Identification of potential growth characteristics to be used in breeding durum wheat under semi-arid Mediterranean type of environment. Slafer, G.A. & Rawson, H.M. 1995. Base and optimum temperature vary with genotype and stage of development in wheat. Plant, Cell and Environment 18, 671-679 Stapper, M. & Fischer, R.A. 1990. Genotype, sowing date and plant spacing influence on high-yielding irrigated wheat in southern New South Wales. III Potential yields and optimum flowering dates. Australian Journal of Agricultural Research 41, 1043-1056

46

Optimizing Plant Population, Crop Emergence, Establishment and Sowing Rate These wheat trials aim to help farmers decide whether they can reduce their costs of seed and weed control through adjustment of seed rate, and if their yields can be increased through improved seeding methods. The on-farm trials use techniques to determine optimum plant population for the farm. Best seeding rates differ according to other procedures used on the farm. Comparisons indicate gross returns from contract drilling could double those from the farmer’s broadcasting approaches. The trials also propose including different varieties and different approaches to harvesting depending on the requirements of the trial. Refer to the chapters on tillage, optimizing crop sequences, optimizing variety and sowing date, and to the introductory chapter “Constraints to cereal-based rainfed cropping in Mediterranean environments and methods to measure and minimize their effects” for general harvesting methods and data analysis

Farms that could benefit from these trials include: ☺ those farms currently using seed rates over

100 kg ha-1 yet failing to achieve good plant populations;

47

☺ those farms that plant by broadcasting seed, and do not achieve uniform and correct seed depth;

☺ those farms that have problems with low quality and contaminated seed;

☺ those where seedling emergence is poor; ☺ those with weed problems, particularly where

weeds are emerging before the crop and severely reducing grain yields;

☺ those farms that have poor establishment, and slow early growth.

About establishing a wheat crop The best seed rate for different conditions Establishing a vigorous crop starts with selecting good quality seeds. Then they have to be placed in the soil using a method that will ensure that: • they germinate rapidly; • most of them emerge; and • they produce the optimum number of

competitive plants for the land. Farmers need to know what seed rate they should use under their conditions. They also need to know how soil type, sowing practice and weather influence crop establishment and plant population. The best seed rate is that which produces the highest economic returns. This seed rate may be close to that which produces the highest grain yield. However, in practice grain yield hardly

changes with further increases in seed rate once maximum yield is approached (see how the curves in the figure flatten and even decline at higher seed rates). Seed sown above that needed to reach

the flat part of the curve, is money wasted.

Figure 1. How yield might change with seed rate for a drilled crop and a broadcast crop (after Bouchoutrouch 1986)

0

1

2

3

4

0 30 60 90 120 150

Seed rate (kg/ha)

Gra

in y

ield

(t/h

a)

DrillBroadcast

Think about the cost of seed: When working out seed rates take into consideration that seed for sowing can cost around 5 times the value of grain produced. So close to the flat part of the curve each additional 1 kg ha-1 of seed sown should always produce at least 5 kg ha-1 extra of grain yield. Think about soil type. Under the best conditions, using the very best seed, farmers can expect 55-65 percent seeds will establish on hard-setting clay soils and up to 90 percent on sandy-surfaced soils. However, at higher seed rates (>100 kg ha-1 for drilled crops)

and under any soil conditions, establishment falls to about half those values.

48

Think about seed quality. Where seed is hand harvested and stored on-farms, quality can be poor because of storage at high temperature (above 30°C) and moisture (greater than 12% seed moisture). Sticks, stones and weed seeds reduce quality further. With such poor seed, farmers may have to use very high seed rates to obtain adequate plant populations. The only apparent benefit of such poor seed is that it may cost nothing. This may be false economy Broadcasting or drilling? For broadcast sowing, poor establishment percentages are common (Figure 1), often falling below 50 percent. Part of this is due to rough seedbeds, poor seed covering and poor contact between seed and moist soil. Consequently, the optimum seed rate for broadcast crops can be twice that for drill-sown crops. Maximum yield in broadcast crops is also likely to be lower. This in part is because applied fertilizer is mixed through the soil rather than placed near the seed as in drilling, so is less directly accessible to plant roots. Primarily though, poor yields in broadcast crops are due to lack of uniformity in planting depth and failure to plant at the best depth. Seed to seed competition As a further complication, the establishment percentage in a drilled crop also decreases as seed rate increases. This is because seedlings get closer and closer together in the row and compete more for resources (see Figure 2).

Under ideal soil conditions with drilled crops

establishment will be about 95 percent at a seed rate of 30 kg ha-1, 80 percent if a 60 kg ha-1 seed rate is used and 60 percent if rates are more than 90 kg ha-1 assuming normal sized seed are sown.

Good seed and poor seed

Where seed is broadcast, the seedlings are not spaced as closely together as when drilled, so there is less competition between them at high seed rates during establishment. Consequently, establishment is less affected by seed rate. The main problem with broadcast systems is the uneven distribution and uneven depth of the seed. (See the photo of plants of sown at the same time but at different depths.) And weeds If the crop gets off to a poor start it seldom recovers to reach its yield potential. Plump seeds have high germination vigour, emerge quickly from the soil and have a competitive advantage over weeds. As a general rule, if weeds emerge before or at the same time as the crop, they severely reduce crop yield. However, if the crop emerges before the weeds, its yield is barely reduced by competition. Shrivelled wheat seeds, or seeds that experience drought or water logging during germination will develop slowly, produce fewer leaves and tillers, compete poorly with weeds, be more susceptible to damage from herbicides and yield less. Working out how many seeds to sow to get optimum plant number: those influencing factors

Figure 2 As more viable seeds are drilled per m² proportionately less grow into established plants because adjacent seedlings have to compete more for resources

Given that no yield penalty is likely if the plant population exceeds the optimum for maximizing yield, farmers often use seed rates that result in higher populations than necessary as a means of competing with weeds and as an insurance measure. What about seed depth? If water is not a limitation, the adage is the shallower the better as long as the seed is covered with sufficient soil to protect it from birds and other scavengers. The photo shows plants sown at the same time, but at different depths (see the seed position in relation to the line showing the soil

surface). The shallow sown plants have many more tillers and leaves. Through covering the soil faster they compete better with weeds and reduce

soil surface water loss and ultimately yield better.

49

If water is a limitation, sowing depth may be determined by the depth of water sufficient to support germination. The depth then is decided by the variety. Many varieties will not emerge and establish strong plants if sown deeper than 5 cm. If the crop is sown dry with an expectation of later rain to trigger germination, the aim again is to sow shallowly but deep enough to protect the seeds from scavengers. A fast-emerging crop is the aim.

Working out sowing rate The best seed rate for any farm is influenced by the many factors already discussed. Consequently, broadly recommended rates may differ considerably from the optimum rate for a farm. But the starting point when calculating a sowing rate in kg/ha for a farm is the size of the seeds and the percentage of those seeds that can germinate (and potentially can grow into established plants). Number and size The common range of wheat seed size is 25 to 50 mg and crop establishment varies between 40 and 95 percent of sown seeds depending on soil type, soil moisture, sowing depth, seed quality, diseases and insects. Considering these variables a seed rate of 100 kg ha-1 could result in an established plant population that may vary up to threefold from 360 to 120 plants m-2 even in a drill sown crop. And it is established plant number that the farmer is aiming towards. So even though the recommended rate for an area is 100 kg/ha, the

optimum rate for a particular farm may be nowhere near this. Check these calculations: 100 kg ha-1 sown x 90 percent established/25 mg seed size = 360 plants m-2 (100 x 90/25 = 360)

Plants of the same age from seed sown at 3 depths (note the shallow-sown plants are best).

compared to 100 kg ha-1 sown x 60 percent established/50 mg seed size = 120 plants m-2 (100 x 60/50 = 120) Estimating seed size You can estimate average seed size by counting and weighing a few samples of the seed lot (say 100 seeds in each lot) then dividing the weight by the number. If 100 seeds weigh 3.5 g, individual seeds in the lot each weigh 35 mg. A germination test to estimating seed viability Do a germination test by counting out several lots of preferably 100 seeds taken from well inside the seed sacks keeping each seed lot separate. Dampen squares of paper or towel and spread each group of seeds on a towel so the seeds are not touching each other. Cover them with a second damp towel. Roll up each sandwich of seeds and put in a plastic bag to prevent the towels from drying out. Keep the bags at room temperature. After 4-5 days count how many seeds have germinated in each lot. Percentage germination is the number of germinated seeds divided by the number of seeds in the sample * 100. Adjust the amount of seed sown upwards to allow for the percentage of bad seed in the sample. Sown amount = required amount * 100 divided by percentage germination. A germination test gives the potential of the seed lot to produce established plants. However, in a drill-sown crop, establishment in farmers’ fields will commonly be 10-20 percent less than this (10 percent less in very good field conditions, 20 percent less in rough or dry seed beds). Adjust the sowing rate up again to allow for this.

Is variety important? Choosing the right variety can be very important if sowing date is determined by an event like the timing of rainfall, and the end of the growing season is determined by some other constraint like rising temperature or lack of water,. A short duration (early) variety is best if the available season length is short, while a long duration (late) variety may yield best if there is plenty of water available and the season is long.

Variety will not impact on establishment, but it will impact on yield through how well it fits into the time slot between best sowing date and best harvest date. Consider including two to three varieties in any trials planned to examine optimum sowing rate for a farm. This is an opportunity to include a new variety to compare with the farmer’s traditional variety.

50

Trials to work out optimum seed rate This proposed trial aims to work out the optimum plant population for the farm and by back calculation, the best sowing rate to get the best yield. As variables it can include combinations of seeding method, seeding density, weed competition, sowing date and variety. The choice of variables will depend on the requirements of the farm and the space and facilities available.

Before starting The first step before deciding on treatments and a trial site is to discuss the reasons for the trial with the prospective collaborating farmer(s) and the likely benefits to productivity on the farm. The farmer will have clear practical opinions on how best to achieve any objectives and whether they are worthwhile. The trial should be carried out at an agreed point in the farmer's cropping sequence, consensus should be reached about the costs involved, allocation of costs and responsibilities, and the ownership of the grain that is produced. Check out the possible site(s) with the farmers while the previous crop is still growing. Look for and mark good or bad areas that should be avoided when laying out the trial. The area planted to the trial should be uniform with no residual effects of any previous small plot trials or crop treatments.

The design and treatments The design depends on the farmer’s needs As the trial is presented it is aimed primarily at drill-sown crops. Whatever you do, keep the trial relatively simple so that the farmer can see the results in the field as well as on paper at the end. Seed rates in broadcast crops are frequently set by factors that are not easily fine-tuned for economic yield. Commonly there will be no choice in quality of seed available and ways to improve yield by other means will be very limited. However, if the farmer normally uses poor seed, stored from the previous year’s crop, this may be an opportunity to demonstrate how the use of

quality seed more than pays for itself in increased yield. Consider a trial with his seed and quality seed as the starting point, and if possible, include a drilled treatment to compare with broadcasting. Where there is an opportunity to discuss broadcasting with collaborating farmers, start by looking at the gross returns analysis in the table.

This shows that contract drilling (if available) could provide twice the profits of a broadcast crop. Ask farmers to supply real values so you can properly compare broadcasting and contract drilling.

A comparison of returns from two commonly used seeding systems i.e. broadcasting seed of poor quality at a high rate versus drilling seed of good quality, at a lower rate

broadcast drilled A grain yield (t ha-1)

Choices for treatments The following are choices that you should include or exclude depending on what you and the farmer consider is needed to improve the farm’s production. Wherever possible use large strip plots at least 10 m long and at least one drill width across and

1.5 2.5 B grain (US$/tonne) 200 200 C grain sales (US$)

AxB 300 500

D sowing rate (kg ha-1) 300 100 E seed value (US$/kg) 0.40a 1.00 F seed cost (US$/ha)

DxE 120 100

G cultivation cost

(US$/ha) 0b 0a

H seeding contractor (US$)

0 50

I total costs (US$)

F+G+H 120 150

Gross benefit (US$/ha) C-I

US$180 US$350

a There is a cost of storing seed on-farm and a cost associated with not having sold it as grain. b There is a cost associated with cultivation

always include the farmer’s conventional practice as the standard (control or check) treatment. Always put any trial within the farmer’s crop so he can see and compare any effects against his crop. Keep the plots large and replication will then be less important. Sowing depth For sowing depth, discuss the impacts of sowing depth on yield and why, and ask the farmer to explain the local determinants of sowing depth. Together agree what the optimum depth might be for the area. Use this sowing depth for all treatments unless an aim of the trial is to show how conventional wisdom on sowing depth may be wrong. Plant population: With the farmer’s rate as standard, increase or decrease the rate depending on what you see as faults in his approach. If the fault is poor seed and you will supply good seed, use his seed rate and two seed rates that are less than that rate. Maybe use 200, 100 and 50 kg/ha with separate plots for his seed and your seed. With a very smart farmer, you may decide that you will aim for particular plant populations in the trial. Maybe 50, 100 and 200 plants m-2) will be appropriate. In that case you will first have to determine germination percent for each seed lot (as already described). Then work out sowing rate using seed size (work that out for his and your seed following the method described earlier), germination percentage and a 10 or 20% upscale depending on whether you are drilling or broadcasting. Variety and sowing date Use a local variety for the study. If a second locally adapted variety is available or a new variety that shows promise for the area is available, include one or both within the trial if space allows. Remember that this will double or treble the number of strip plots needed, so you must have good reason to include them. A reason may be that one has rapid early growth that potentially is a better competitor with weeds. If this is the case, the optimum population may be lower than with a standard variety. A further reason may be that a changed sowing date is required to control an aggressive local weed. This means that the new variety should have a longer or shorter duration. This trial would then have to include sowing date as a component.

Weeds If the area has major weed problems, propose as a possibility, a second or combined trial that aims to find the plant population needed to control those weeds and still achieve high yield. The study might use two levels of weed control (sprayed and not sprayed) and the three target plant population chosen under ‘plant population’. Then the trial would be 6 strip plots all using the local variety X 3 sowing rates x 2 weed control levels. With two varieties (12 plots) this is still a manageable size for most farms if it is done as long strips without replication. Number of participants and years for the trial Suggest repeating the trial for three seasons to provide trends of response to the environment. Duplicate the trial on different soil types if appropriate. Try to get several nearby farms doing the trial especially if several questions regarding the optimum plant population need to be answered for the area. Keep the most important treatment(s) on all farms but add variables considered of specific importance, to the extra farms. Add only one special treatment to each farm. This has the multiple benefit that base data are collected more quickly, the farms provide a measure of variation in response to the treatments and perhaps most importantly, the farmers will discuss the trials and learn from each other. You may have to encourage and facilitate this interchange initially.

Some general procedures Sow a predetermined number of viable seeds in each plot after measuring seed size (weight) and checking germination percentage. No chemical or hand weed control is allowed after sowing (except when the subtreatment of chemical weed control is added) so that the effect of seed vigour and plant population on weed growth can be assessed. If the trial compares broadcast with drill-sown treatments, the plots must be large enough to use the local broadcasting technique at sowing and to sow both treatments on the same day. Locate the trial within the farmer’s normal crop and if the trial includes broadcasting versus drill or different varieties, sow the whole trial at the same time. This allows direct visual comparisons through the season and comparisons for yield at the end. Where there is a second planting date, do all subsidiary treatments on that date also.

51

The trial plots should be at least a drill width wide and at least 10 m long at sowing. At harvest the whole strip should be machine harvested if possible (measure the area harvested) or at least six rows x 8 m. If harvesting a broadcast crop, make sure the same area is sampled in that and the drilled treatments.

52

Row spacing should be what farmers normally use and can be 15 to 25 cm, but the choice must be known in advance so that the seed rate needed for each plot can be calculated Basal N and K fertilizers, if used, should be separated from the seeds, but P fertilizers can be mixed with the seed if convenient. The seed should be treated immediately before sowing with standard rates of fungicide to control seed-borne diseases such as smuts and bunt. Leaf diseases and insect pests should be sprayed if they threaten the validity of the trial or if they are atypical for the region.

Measurements Rainfall and temperature

If there is no weather station nearby, put a rain gauge and max/min thermometer at the trial site or near the farmer’s dwelling for convenience. If the farmer makes the measurements and recordings, clarify that this should be weekly throughout the season, when instruments should be emptied or reset. Details of what is required are in the introductory chapters. Supply the farmer with a table in which to record the data. Explain the importance of being able to relate the trial

results to seasonal conditions, particularly when the trial is repeated over several sites and seasons. Results may vary considerably depending on whether it was a hot, cool, dry or wet year. The aim is to be able to decide what agronomic techniques arising from the trial should be used in the average year. Germination tests Use the method already described. Seed rate to be used for sowing can then be calculated using the following formula.

Seed = target plants/m2 x seed weight (mg)

rate (kg/ha)

germination percent.

For example: for a target population of 250 plants/m2 and using seed of average weight (35 mg) and a laboratory germination test of 95 percent. Seed rate =250 x 35/95 or 92 kg ha-1 . For a plot size of 10 m x eight rows (at 20 cm spacing), or 0.0016 ha, the weight of such seed (35 mg) to sow per plot is 147 g i.e = 92 kg ha-1 x 1 000 x 0.0016 ha. Plant counts Count plants in three places in all plots on the same day one week after seedlings begin to emerge. Count all plants in one-metre-long sections of row in these places. Select and mark the places when you sow, avoiding the outside rows and the ends of the plots.

In the case of a broadcast treatment, count within a square of 0.5 m2 at three pre-marked locations in each plot. If you are comparing drilled and broadcast treatments, be sure to also measure the row width.

Calculate emergence percentage as follows: if the target population is 250 plant/m2, 20 cm row spacing is used and seeds have 95 percent germination, this is: 250 x 0.2 x 0.95 = 48 viable seeds per m of row. So if six seedlings are counted on day 7 the emergence percentage is 6/48= 13 percent. The most relevant measure of seed vigour in the field is the speed of emergence. It is related to later crop development and growth. It does need someone to do the counts on a few occasions over the first two weeks after sowing. The person can stop counting and writing down the numbers each day when the 50% emergence is reached. For the example of 250 plant/m2 or 48 viable seeds per metre of row, that day will be when 24 seedlings are out. And for a 50 x 50 cm square it will be the day when 59 plants are counted. From the 50% values you will be able to rank the treatments for initial success two weeks after planting. Do not forget to do a later count of plants, maybe at 3-4 weeks, just to be sure what 100% plant population really is. Get the farmers involved in these counts and writing down the numbers.

Maturity Harvest What do you want from the data? Decide what you want the data to say. Do you just want to know the yield in the various treatments? Or are you also interested in explaining why the treatments produced different results? Sometimes knowing why can help you make decisions about how to further optimise production on the farm. If you just want yield data, go straight to the following section on plot machine harvesting. If you want to understand what happened, read about the grab samples. Grab samples Before you start the machine harvest at the grain ripe stage, go through every plot and collect grab samples of 30 culms from three locations in the plot. These will be used to work out yield parameters such as spikes per unit area, grain yield per culm and so on. The grab samples can be taken the day before harvest if necessary. Cut the grab samples off at ground level but do not include soil as that is very heavy. If you prefer, the 30-culm grab samples can be made up of three samples of 10 culms as that number easily fits a hand. Bundle each 30 culms, tie up and then tie together the three bundles from the plot. Tag them with the plot identifier, treatment, and the date. If you have large plastic bags, put each plot bundle of 90 culms into a bag. This prevents physical damage when you stack all you bundles together to take them back to the farm or laboratory for drying and weighing. Plot machine harvest Now harvest the whole of the strip for each treatment and do this by machine if this is available. You will need the grain from each plot (strip) separately. If the strips are large enough (say at least 10m by a drill width) and all are abutting the next door strip or the farmer’s crop, the edge effects are relatively small. There will be a similar percentage error in each treatment. Make sure you measure the area actually harvested and that you collect all the grain from the machine after each plot. A three-sided quadrat If the machine cannot be cleaned satisfactorily between plots and if grain losses are high, you will have to consider doing 3 x 2m2 hand samples within each plot for this maturity harvest. This is not as much work as it sounds if you make 1m2 quadrats in advance from steel rod to define each

area during cutting. These quadrats should have one open side (3 closed sides) to allow the person cutting to push the quadrat into the crop and then forward to mark the second 1m2 of the harvested area. Cutting each 2m2 area with a sickle will take about 4 minutes. For a 12 plot trial, that is just over 2 hours. If you are assessing weeds, you may have to consider this quadrat method for harvesting. Then the weed samples can be separated from the crop samples as the samples are cut. These are bundled separately, and dried and weighed separately. This method allows you to readily calculate weed infestation per m2. Alternatively you can estimate weed cover by eye using the methods in Rawson and GomezMacpherson (2000).

What to do with the harvested material If you want a detailed description of how to use your harvest material to calculate and use all yield components, look in the introductory pamphlet “Constraints to cereal-based rainfed cropping in Mediterranean environments and methods to measure and minimize their effects”. An analysis of yield components is time consuming but it can explain why things happened. For example, if ears are small, stress or competition from weeds occurred early in growth when number of spikelets is determined. If grains are small the stress occurred at the end of the season during grain filling. Read on for a summary of what to do with the harvested material. The grab samples As soon as possible after the harvest, take the grab samples (and quadrats if you had to do a hand harvest), untie and separate the bundles and lay them in the sun or in a hot glasshouse to air and dry. Keep the tags with the samples and do not mix the samples. If you can dry samples in an oven at 70-80°C do so. This is preferable as it can be controlled. When the grab samples are dry, weigh them. Weigh the 30 culm bundles intact. Write down the weight in a prepared table. Thresh the seed out of each bundle and weigh that. Count 100 seed samples and weigh them. The full plot or quadrat samples Weigh the air dried seed that was machine harvested from each full plot. Measure its moisture content. Adjust the weight of all grain samples to constant moisture content, possibly 11%, but depending on the local standard.

53

Take random samples of grain from these samples and oven dry them as you did the grab samples. Now you have enough data to work out all the yield components. Working out yield components Do all this on a computer spread sheet set out with the treatments with their grab samples, and the full plot and/or quadrat names as rows. The components you have measured and the yield components will be columns. Treatment grain yield is the sum of full plot grain plus grain yield of the grab samples brought to equivalent moisture content. Divide this number by the measured full plot harvested area to express the yield as per m2. Harvest index is the weight of grain of the dried grab samples divided by the full weight of the grab samples. Culm number per unit area is the number of culms in the grab sample (30) multiplied by treatment grain yield/m2 divided by grab sample grain yield. Replicate grab samples should provide similar estimates of culm number. Individual kernel weight can be estimated from 100 seed samples from the full plot harvests. Kernels per m2 is treatment grain yield/m2 divided by individual kernel weight, while kernels per culm is kernels /m2 divided by culms/m2. Do not forget to separate, dry and weigh the

weeds that you took with your grab samples or quadrat cuts. Estimate their weight in relation to the grab sample weights and calculate their proportions. Upscale the weight of the weeds to the full plot. Because you took three grab samples within each plot, you have some estimate of variation in the components within each plot. It would not satisfy a statistician, but will give you some confidence in your numbers. Grain quality If you have the resources, measure grain test weight (weight per unit volume in kg/hl) and small grain sievings (using a 2 mm slotted sieve) on a subsample of grain from the whole plot. These values are related to the yield of flour that can be milled from the grain.

How to analyse and interpret your data The example here will differ from your trial that you have designed to match the requirements of the farms you are trying to optimise. It was designed to determine the optimum sowing density and the optimum sowing depth on a farm that also had problems with weeds. It actually included two varieties, but they produced the same trends so the data presented were all averaged across varieties. Does sowing depth affect yield? When seeds were drilled deeper in the example

Average grain yield, total crop biomass at maturity and weed biomass as affected by sowing depth and plant population of wheat (from Anderson, W.K.)

Sowingdepth Plants/mz Establish-

mentGrainyield Biomass (g/m2)

(mm) Target Actual Actual % t/ha Wheat Weeds

50 25 25 100 1.97 563 237

50 49 99 2.25 608 192

100 98 98 2.49 655 145

200 151 76 2.51 661 83

400 259 65 2.50 658 42

100 25 23 92 1.19 340 460

50 40 80 1.61 447 353

100 72 72 1.89 511 289

200 123 62 2.13 576 224

400 201 50 2.20 595 205lsd (D x P)

5% 0.23 45 64

54

trial shown in the table below (100 vs 50 mm), yield was always less at equivalent sowing densities. When 50 emerging seeds per m² were sown for example, yield was reduced by almost 30 percent. The maximum yield that could be achieved irrespective of numbers of seeds sown was also less from deeper planting, down by around 12 percent (2.51 to 2.20 t ha-1). ).

In fact, almost twice as many seeds were needed to achieve the optimum number of plants if they were planted deep than when they were planted shallow. Put simply, many more seeds need to be planted, if they are planted deeply, to produce the best yield (see the figure). There is no universally correct seed rate. Seed rate must be adapted to other practices used by the farmer.

In fact, almost twice as many seeds were needed to achieve the optimum number of plants if they were planted deep than when they were planted shallow. Put simply, many more seeds need to be planted, if they are planted deeply, to produce the best yield (see the figure). There is no universally correct seed rate. Seed rate must be adapted to other practices used by the farmer.

Talking with collaborating farmers Talking with collaborating farmers When showing the results of your cooperative trial to farmers, discuss not only your equivalent to the above figure but also how to work out from your table the equivalent increase in grain yield for each increase in seed rate.

When showing the results of your cooperative trial to farmers, discuss not only your equivalent to the above figure but also how to work out from your table the equivalent increase in grain yield for each increase in seed rate. Seeing the increments in grain yield for the

increments in seed sown will be more persuasive than a graph, though the graph is vital for exact calculation of optimum seed rate.

Seeing the increments in grain yield for the

increments in seed sown will be more persuasive than a graph, though the graph is vital for exact calculation of optimum seed rate. For example, in the table for the 50 mm sowing depth, increasing plant population from 50 to 100 plants/m² took an additional 19 kg seed ha-1 but gave 240 kg ha-1 more yield (a 12-fold benefit for the extra seed).

For example, in the table for the 50 mm sowing depth, increasing plant population from 50 to 100 plants/m² took an additional 19 kg seed ha

However, increasing from 100 to 200 plants/m² took an additional 40 kg seed ha-1 but gave only 20 kg ha-1 more yield.

However, increasing from 100 to 200 plants/m² took an additional 40 kg seed ha

Superficially this last increment was an economic loss. However, then you have to consider together what the effect of that additional population was on weed growth.

Superficially this last increment was an economic loss. However, then you have to consider together what the effect of that additional population was on weed growth. Together you may conclude from your data that 40 kg seed ha-1 is a cheap means of weed control. Together you may conclude from your data that 40 kg seed haWeed control Weed control Now look at the following pair of figures showing that increasing plant population of wheat has very much decreased the growth of weeds, but more so when the crop was sown shallowly.

Now look at the following pair of figures showing that increasing plant population of wheat has very much decreased the growth of weeds, but more so when the crop was sown shallowly. If you examine your culm counts they will also show that crop plants when sown shallowly are better competitors with weeds.

If you examine your culm counts they will also show that crop plants when sown shallowly are better competitors with weeds. This is because they emerge more quickly and tiller earlier (see the sowing depth photo). This is because they emerge more quickly and tiller earlier (see the sowing depth photo). However, farmers often use seed rates that result in plant populations that far exceed the level required for maximum yield. This may be a conscious use of extra plants to assist in smothering weeds and may be an economically viable solution where herbicides are unavailable or not favoured.

However, farmers often use seed rates that result in plant populations that far exceed the level required for maximum yield. This may be a conscious use of extra plants to assist in smothering weeds and may be an economically viable solution where herbicides are unavailable or not favoured. Some farmers may be concerned that increasing the seed rate will increase the proportion of small grain sievings (passing a 2 mm slotted screen) or decrease the test weight (kg hl-1) of the grain, and that the value of the grain will be reduced.

Some farmers may be concerned that increasing the seed rate will increase the proportion of small grain sievings (passing a 2 mm slotted screen) or decrease the test weight (kg hl

Both measures are related to the yield of flour that can be obtained from a given weight of grain, an Both measures are related to the yield of flour that can be obtained from a given weight of grain, an

-1 but gave 240 kg ha-1 more yield (a 12-fold benefit for the extra seed).

-1 but gave only 20 kg ha-1 more yield. Change in yield with plant population at two

depths of sowing. Arrows show the populations needed to achieve optimum yield

-1 is a cheap means of weed control.

-1) of the grain, and that the value of the grain will be reduced.

For a wheat crop sown shallowly, plant population effects on biomass of the crop and the weeds

For a wheat crop sown deeply, plant population effects on biomass of the crop and weeds

55

important consideration for farmers who use their grain to make flour for their own consumption. If appropriate, examine grain in your trial to check the levels of these grain defects against locally relevant standards (international standards are <5 percent for small grains and >74 kg hl-1 for bread wheat). When discussing this type of trial with local farmers you must consider their attitudes to sowing practice and seed rate. For example, the farmer may not be prepared to sow shallowly for fear of losing seed in seasons when the early rains fail. The use of a different variety with a very long coleoptile may then be a more acceptable option, even if some yield is lost in good seasons. In that case, be sure to include such an alternative variety in your trial to compare with his usual variety. Before you begin your trial and again when the trial is underway, make some measurements of plant populations in nearby farmers’ fields. Assuming your trial actually demonstrates improvements in economic yield, use the contrast in populations to strengthen your arguments for changed practice. In the example, sowing at 100 mm depth compared to 50 mm reduced crop establishment, increased the plant population required to reach the maximum grain yield, reduced the maximum grain yield level itself, and reduced the ability of the crop to suppress weed growth. It cannot be concluded that this result is generally applicable to all rainfed wheat crops, since factors such as soil water, soil temperature, soil texture, weed species and crop varieties can all influence the result. If the results of this trial are supported by other results in the region suggest that farmers try to control sowing depth to about 50 mm and that a target plant population of 250 plants/m2 will considerably assist in suppressing weeds. Substantial savings in costs of seed and weed control should result, as well as grain yield increasing. Supplementary trials may be required to test the effects of the main variables in the region on your crop establishment package.

Typical critical variables are rainfall, temperature and soil type. To see how these will affect your proposed package, run the trial on farms over several seasons, refining it as you go in light of the results, new varieties, and discussions with farmers. With time, a suite of optimized packages can be compiled that will cover the main requirements of all the farms in your region.

Further reading Anderson, W.K. 1986. Some relationships between plant population, yield components and grain yield of wheat in a Mediterranean environment. Australian Journal of Agricultural Research 37, 219-233. Anderson, W.K. & Impiglia, A. 2002. Management of dryland wheat. In B.C. Curtis, S. Rajaram & H Gómez-Macpherson, eds. Bread wheat: improvement and production, p. 407-32. FAO, Plant Production and protection Series 30, Rome, Italy. Bouchoutrouch M. 1986. Effet de la dose de semis et du désherbage sur le blé tendre Nesma. Rapport d’activité 1985-1986. Programme Aridoculture. Settat, Maroc. Pp 54-55. Fawcett, R.G. 1964. Effect of certain conditions on yield of crop plants Nature 204, 858-60. Karrou, M. 1998. Observations on effect of seeding pattern on water use efficiency of durum wheat in semi-arid areas of Morocco. Field Crops Research. 59 (175-179). Puckridge, D.W. & Donald, C.M. 1967. Competition among wheat plants sown at a wide range of densities Australian Journal of Agricultural Research 18, 193-211. Rawson, H.M. & Gomez Macpherson, H. 2000. Irrigated wheat: Managing your crop. FAO. Rome, Italy Shackley, B.J. 2000. Crop management. In W.K. Anderson & J.R. Garlinge, eds. The Wheat Book - principles and practice, p. 137-145. Agriculture Western Australia, Bulletin 4443. Australia.

56

Optimizing Tillage Systems On-farm This chapter describes trials for conducting on-farm that compare the performance of wheat crops and soil quality changes under three tillage systems: conventional, minimum and zero tillage. The primary aim of these trials is to find the best way to protect the soil resource in the long term while maintaining or increasing crop production. Secondary aims are to reduce machinery wear and achieve savings in fuel costs and labour through shifting to conservation or no-tillage systems. Such tillage trials are extremely important in areas where water is a major limitation to production and where soil structure is poor and erosion a problem.

The trials check whether yield can be maintained using standard sowing dates but with the reduced effort and expenditure associated with fewer tillage operations. They also use zero tillage to open up the possibility for timelier sowing. Trials compare tillage systems in affecting soil quality, mainly organic carbon content. They introduce a simple infiltrometer method for assessing how much of the rain falling on the crops is wasted or leads to soil erosion and how much goes towards producing crop growth. It uses the method to compare the three tillage systems for water efficiency.

These trials should not be attempted if aggressive local weeds cannot be controlled other than by full tillage.

Other chapters you might need to refer to are those on choosing the right variety x sowing date, optimizing a cropping sequence for the farm, nitrogen use and crop establishment practices and . “Risk Analysis and decision making tools for North Africa” for weather analysis.

Tillage practices: some background Tillage has been used for millennia to prepare the soil prior to sowing many of the annual grain crops. It involves applying power to break up and rearrange the topsoil. It has the primary aim of destroying weeds and pests but is also important for incorporating, redistributing or releasing nutrients and making the soil suitable for seed sowing and germination and for easy penetration of seedling roots.

The English word “tillage” is derived from the Old English “tillen” which means “to toil”. With only human or animal power available, it took a long time and much toil to till even moderate-sized areas of land. When tractors became available, larger areas could be cultivated per person.

Tractors and their increasing power also made it possible to expand cropping areas into more difficult soils. In time this created problems in less robustly structured soils with many loamy and fine-textured soils weakening within a few years of tractors replacing animal teams for tillage.

Surface water runoff can increase following tillage on many soils, causing increased water erosion particularly on sloping cropland (Mrabet et al., 1993). Erosion by wind is also increased by tillage because the topsoil is left bare and loose.

Other potentially undesired effects of tillage include reducing soil organic matter through oxidation and deleterious effects on soil micro flora and fauna, also leading to reduced soil structural stability and increased surface runoff and water or wind erosion.

If it were possible to retain the desired effects of tillage while reducing or removing the problems it can generate, that would be a major step forward.

Definitions used in this chapter

Conventional multi-pass tillage

Conventional or full tillage rearranges the entire topsoil. It may require several passes to first "turn" the soil and then break it down into a friable seedbed prior to sowing.

57

No-tillage or zero-till involves one pass with a no-till drill during which a part of the soil surface is disturbed or “opened” and the seeds are placed concurrently in that disturbed zone. The seeder opener may be a knife-point as little as 5 mm-wide on a tine, or a single, double or triple-disc set at a slight angle to the direction of travel. A weedicide may be applied during this operation

Minimum tillage as defined here is generally a one-pass tillage operation at sowing synchronous with seed placement, typically achieved using full cut-out points, or full cut-out one-way or offset discs to break up the entire soil surface. It may include ashallow cultivation between seasons to control weeds. Then it may be called reduced tillage.

Conservation tillage is a generic term that covers any tillage system that reduces loss of soil and water compared with conventional tillage. Some have defined it more tightly to include treatment of residues specifying that at least 30 percent of the soil surface should be covered with residues after sowing so as to reduce erosion by water. It is likely to include zero tillage within the definition.

No-tillage: A Win-Win system No-till cultivation is a major step forward for agricultural development in many countries but it can have its problems mainly if crops are not well managed. It also disturbs the soil but limits that disturbance to rows or slots in which the crop seeds and fertilizer are placed.

A reduction in the number of cultivation passes as is the case with no-till also means less wear on machinery, less use of fuel or animal power, less time devoted to soil preparation by the farmer thus a possible overall improvement in gross returns for the farm.

A no-till crop emerging uniformly

Better water use efficiency with no-till A further possible benefit of no-till over multi pass tillage is that rain, particularly heavy rain, is more likely to concentrate in the seeder slots and thereby penetrate directly to the crop’s root zone. This could improve not only water harvesting but also crop water use efficiency. In Morocco, Mrabet (1997, 2000a) observed that WUE in no-till and chisel tillage are high compared to traditional tillage. Table 1: Effect of tillage systems in Morocco on water use efficiency (WUE) and water use (WU) of wheat (Mrabet 2000a)

Tillage systems

WUE kg. mm-1.ha-1

WU mm

No-till 9.5 429

Chisel plow 10.7 408

Deep plow 9.0 430

Rotary tillage 7.9 432

Off-set disk tillage 6.7 415

Traditional 7.7 411

Stubble mulch with sweep

7.4 401

Mean 8.4 418

LSD P≤ 0.05 1.6 1.5

Reduced erosion, improved yield with no-till A major plus of no-till sowing is that it can reduce erosion to low rates. Sustainable soils under no-tillage cropping are therefore possible in the long-term, a difficult goal using conventional tillage particularly on slopes.

As well as making soil systems sustainable, no-till, when used intelligently, can also lead to

58

improved yields (Mrabet, 2000a). An example of this is where no-till has allowed earlier sowing than would have been possible if seeding were delayed by the time taken for multi pass tillage.

Table 2: Effect of tillage systems on the average grain yield and total dry matter (from 95/96 to 98/99) of wheat in semi-arid area of Morocco (Mrabet 2000a). Average rainfall 362 mm.

Tillage systems

Grain yield (t.ha-1)

Total dry matter (t.ha-1)

No-till 2.47 9.03

Chisel plow 2.41 8.19

Disk plow 2.36 8.73

Rotary tillage 2.09 9.18

Off-set disk tillage 1.97 8.12

Traditional 1.93 7.26

Stubble mulch with sweep

1.91 8.21

Mean 2.16 8.40

LSD P≤ 0.05 0.43 NS

Increased soil organic matter with no-till With no–till the soil organic matter (SOM) content is improved in the top soil horizon. Data from a 11-year study under no–till versus conventional till of a Calcixeroll soil in Morocco showed that soil under no–till contained more SOM in the top 0 to 20 cm layer than that under conventional till (Table 3).

Mrabet et al. (2001) also found a 13.6% increase in SOM in the 0-20 cm horizon under no-tillage systems. Under conventional tillage the increase was only 3% over 11 years.

Table 3. Tillage effect on total soil organic matter (g/100g) of 0-20 cm depths of an Calcixeroll soil, Sidi El Aydi, Morocco. (Saber and Mrabet, 2002)

Depth

Tillage systems

0-2.5 cm

2.5-7 cm

7-20 cm

No-till 3.98a 2.57a 2.13a

Conventional 2.51b 2.45b 2.10a

Mean 3.25 2.51 2.11

In each column of Table 3, values followed by the same letter are not significantly different at P≤ 0.05 using LSD Test.

Use of residues with no-till In semi-arid areas grain yield under no-till systems is also considerably increased by a cover of residue on the soil surface (Figure 1). Crop residues reduce soil moisture evaporation leaving more stored water for crop growth and yield. They also limit erosion.

However, where yields of the previous crop are high and large quantities of crop residues are left in the field after harvest, problems of disease and residue handling can become limitations to the no-till approach. Mrabet (2002) reported that 70% cover is sufficient to improve yields of wheat in dry areas. Straw from the previous crop in excess of this requirement can be used for animal feeding.

Figure 1. Effect on wheat yield of the percentage cover of the soil surface by crop residues under no-tillage system in semiarid area of Morocco (Mrabet, 1997; Mrabet, 2002).

Rain water infiltration under different tillage systems: more under no-till A trial by K.J. Bligh in Western Australia compared conventional, minimal and no-till systems for rainfall penetration over three years in the 1980s.

Using infiltrometers, described later by K.J. Bligh for possible inclusion in your on-farm trial, it was found that it took three years before the no-till treatment had settled and matured sufficiently to show benefits. Then, almost all (96 percent) of the 253 mm growing season rainfall penetrated the zero-till plots for use by the crop.

This was more infiltration than under minimum tillage (86 percent) but substantially more than

59

under conventional tillage (79 percent) where 21 percent of the rain ran off, was lost to the crop and caused soil erosion. Interestingly, earthworm numbers were increased tenfold by reducing levels of tillage as the soil micro fauna and flora rebuilt towards the situation in permanent pasture.

These data agreed with Bouzza (1990) who found higher water storage under no-tillage fallow than under clean fallow.

Your trial: Issues to consider before deciding on the design No-till could reduce yield It is possible that minimum or no tillage approaches will give lower yield in the first years after adoption than the conventional tillage operations used currently by the farmer. This could be because soil quality is initially low and residues are not retained to conserve soil water. No till and minimal till operations, while apparently very simple, to be optimized require high technical skills for crop management, establishment and disease protection and to control weeds in diversified rotations.

A review of over 30 tillage trials (see further reading) found that tillage method (conventional or zero-till) did not change yield when data were averaged across studies, locations and years. However, in some places and some years either method could be far better (+1 t ha-1) or far worse (-1 t ha-1) than the other. The reasons for the variability were not always clear but in some cases reduced yield in the no-till treatments could be associated with slow early crop growth, more weeds, or increased soil disease particularly in wetter years. Increased yield in no-till treatments could sometimes be linked with improved aggregate stability, water infiltration, organic matter and, where soil fertility was high, increased water storage and efficiency of water use.

The clear message is that there are many possible variables altering the relationships between tillage method and yield. The aim of the researcher and farmer should be to identify the limiting variables and adopt a tillage system or suit of systems that minimizes their effects.

Likely limitations to yield The lesson for the design of your current trial is that you should ascertain with the collaborating farmer what the likely limitations are to the

different systems of tillage. Discuss weeds, diseases, nutrition, problems of residues from previous crops, varieties and any soil factors that may be limiting establishment.

When planning your design, consider including different levels of the likely limiting factors as sub-treatments within the main tillage treatments. If you decide to use a strip design, you could include these levels as extra strips within the farmer’s field.

The crop rotation is another important factor to consider in your on-farm trials for a suitable tillage system. Trials conducted in semi-arid area of Morocco have shown that Wheat/Fallow rotations permit the highest grain yield in the crop season mainly by utilizing water more efficiently. Table 4: Effect of rotation and tillage practice on wheat grain yield (t.ha-1) in semi arid area of Morocco (Bouzza, 1990; Mrabet, 2000b)

Sidi El Aydi* Jemaa Shaim**

Tillage System

Wheat/Wheat

Wheat/Fallow

Wheat/Wheat

Wheat/Fallow

No–till 1.9 3.5 1.7 3.0 Minimum tillage

1.6 3.4 1.5 3.0

Convent- ional tillage

1.4 2.4 1.6 2.4

*Average grain yield from 1983 to 1992 ** Average grain yield from 1983 to 1998

Note in Table 4 that the no–till system had highest yields regardless of rotation. For areas receiving more than 300 mm; a three year rotation including fallow, wheat and a food legume or another cereal are advised for farmers.

Problems of stubble Plant residues on or above the soil surface are often referred to as stubble. Most is likely to be dead straw remains of the previous wheat crop. Stubble handling can be an intractable problem of no-till and minimum-tillage seeder operation. The problem increases with increase in yield.

In theory, clean stubble on the soil surface should reduce water loss and when incorporated should increase soil organic matter, aeration and water retention and increase yield. Sometimes, however, retaining rather than removing stubble results in reduced yields.

60

Strategies to capitalize on the positives of stubble (maximizing the time that stubble provides soil protection against wind and water erosion) while dealing with any negatives (disease, allelopathic and seeding problems) include straw spreading at harvest followed by flattening to speed break down during the off season and heavy grazing.

Seeding through heavy stubble is difficult. Burning to remove the stubble makes seeding easier but it does waste crop nutrients and cause air pollution. Leaving standing stubble short at harvest may make it easier to sow into. Standing stubble wraps around tines less than straw lying on the soil. Disked seeders may cut through prostrate stubble on firm soil, but if the surface soil becomes soft, such as after rain, discs tend to push straw into the soil rather than cutting.

Methods of sowing into stubble include widening tine spacing on seeders. This allows greater amounts and lengths of straw to pass. If stubble is going to be a major problem in the prospective trial, consider introducing stubble management approaches appropriate for the farm as sub-treatments or additional strip treatments.

Problems of sowing, fertilizer and weed control Seed depth Soil disturbance prior to or during seed placement aims to ensure that sowing depth is optimal and consistent and that the seedling coleoptile can penetrate its covering of soil and the roots rapidly explore the soil to depth. Seeds sown at the wrong depth or at irregular depth because of poorly prepared cloddy soils or inadequate seeding equipment may produce poor yields. The other aim of soil disturbance is to set back or kill any weeds and pathogens that will restrain growth of the developing wheat seedlings.

Fertilizer For no or minimum tillage operations fertilizer is generally placed and not broadcast during sowing.

Position is important. Fertilizer is best placed deeper than the seed through separate delivery tubes behind each sowing point. Fertilizer toxicity to the cereal seedlings is thereby minimized and weed seeds germinating on the soil surface have their access to the deeply placed fertilizer delayed.

Double disc seeder opener

The farm should be aware that for deep fertilizer placement power requirements on no-till seeders might be as high as with conventional seeders sowing into loosened tilled seedbeds. Typically, 3-6 kW may be required per knifepoint for a knifepoint to penetrate over 100 mm into the soil at a speed of about 8 km/hr. This may reduce the number of rows that the farmer can plant at once with normal machinery or animal power.

In your trial, type of fertilizer, timing and placement will depend on local constraints. As with other variables it may be appropriate to include a fertilizer sub-treatment in the study.

Weeds Unfortunately, without herbicides no-till crops can become dominated by weeds and so yield poorly. Herbicides have their associated costs and may have unknown side effects or long-term impacts on the environment. Furthermore, some weeds have developed resistance to some herbicides, leading to a need to rotate both crops and herbicide groups in order to keep crops weed-free.

On the positive side, a no-till system buries fewer fresh weed seeds and brings many fewer dormant weed seeds to the surface where they can germinate, so with increasing time under no-till, weed problems and herbicide applications can be reduced.

Putting fertilizers below the seed using a deep fertilizer point, followed by a shallower seeder point then a press wheel (Photo BOULAL, Morocco)

61

Effective weed control is critical if minimum or no-till practices are to compete with conventional tillage in the production of yield. For your trial it may be advantageous to apply residual herbicides before sowing. In that case, the farmer should add knives to the seeder when sowing to throw soil and herbicide out of the sown row so the seedling can grow in relatively non-toxic soil. Weeds germinating between the sown rows receive the toxin from the thrown soil.

Depending on the farm’s previous experience with weed control methods, it may be necessary to include sub-treatments in the trial with different types or levels of herbicide.

Designing and doing your trial to compare three tillage systems Which farms would benefit from this tillage experiment? • Those farms who don’t adapt adequate tillage techniques to agro-ecological conditions

• Those farms with low rainfall where normal cultivation results in high evaporative or runoff losses and therefore reduced soil moisture;

• Those farms where rain falls in bursts of high intensity leading to run-off;

• Those farms on steeply sloping land with associated problems of water erosion;

• Those farms with poorly structured soils that readily turn to dust when cultivated under dry conditions or form large clods when cultivated wet;

• Those farms with soils that readily form plough pans that are impenetrable to roots or with soils that become compacted under heavy wheeled traffic;

• Those farms with weeds and pests problems that necessitated specific tillage techniques

Questions before you start The first question to answer with the farmer is why are you planning to do tillage trials? The design of the trial will depend on the answer.

Consider whether there are opportunities to:

• reduce erosion problems on the farm where there is steeply sloping land;

• get the crop in earlier, closer to the optimum window for the best variety, if tillage and seeding could be done more quickly. Would earlier

planting lead to increased yield? Would this be an overall economic benefit to the farm? Work out approximate amounts;

Minimum tillage took less time than conventional so the wheat crop (on right) could be sown earlier and in this case yielded more

• cut back tillage passes without loss in yield thus saving money by reducing fuel and implement wear and operator time on tillage activities. Calculate how much might be saved;

• reduce a common problem of soil compaction associated with continuous movement of heavy machinery over the land;

• minimize loss of scarce water stored during fallow by tilling and planting in one pass. Would this lead to increased yield by making more water available to the crop in the establishment phase? Estimate how much.

Next you need to consider what problems are likely to arise from any changes in tillage technology and how they might be overcome.

Amongst these might be:

• dealing with crop residues;

• increases in disease associated with residues and reduced soil disturbance and whether a changed crop sequence might be an option to overcome the problem;

• increases in pests until the system reaches an equilibrium between pests and predators;

• alternative ways of dealing with weeds and the costs of herbicides and their potential problems. It should always be recognized that any chemicals, though very useful in the short-term, may create problems of herbicide resistance in weeds if herbicide types are not rotated in the longer-term;

• establishment problems associated with poorer tilth of seedbeds.

The final issue to think about is that the trial must be long-term, possibly more than three years.

When moving from full tillage to zero-till the micro-flora and fauna take a long time to re-

62

establish themselves and reach equilibrium in a relatively undisturbed soil profile. This duration may discourage the farmer from starting the trial even though the potential benefits of reducing the level of tillage are well understood.

Going ahead with a trial If the farmer considers that the benefits of change outweigh the disadvantages and risks, start to put together a plan that is visionary but cautious in size and resources. Give the farmer a commitment on your time through the study.

Plot size Keep proposed areas for the study to a minimum unless the benefits of reducing tillage are very obvious (like steeply sloping land). Allow for large enough plots at least 10 m long for later division into sub treatments, for things that unfold as obvious limitations on the farm as the study progresses (see the plot and subplot diagrams in the chapter on “What is the best cropping sequence for the farm?” sequences ‘a field layout and general methods to use’ and read there how to layout plots).

Alternatively, if you plan to do a trial with strip plots, put in treatments and sub treatments as separate strips. Size them so they can be planted and harvested entirely by machine. Make them at least 10 m long and a harvester wide. Issues are discussed in “Optimizing plant population, crop emergence, establishment and sowing rate”.

Issues and measurements Address only the issues raised in the above and associated questions. Restrict proposed measurements in the first two years to establishing that yield does not decline with the changed methodologies. However, if yield does seriously decline in the first year, modify the trial to include sub-treatments to determine why. If the reason is obvious and insurmountable, consider curtailing the study at that time.

General aims of the trial The trial compares three systems of tillage initially for their effects on yield. The longer-term aim, once the systems have been through an establishment phase or preliminary testing for two to four years is to define a tillage management package for the farm and region. This might include proposals for crop sequences, time of planting of specific varieties, weed and disease control strategies and plans for handling of residues.

There is no doubt that to produce an optimized system several other variables will need to be changed when the tillage method is changed.

Choice of tillage systems for the trial The multi pass conventional system used in the trial should be that normally used on the farm or in the area.

A minimum tillage system should disturb the soil sufficiently in terms of the cultivated width and depth of each row, to give some control of weeds and an appropriate tilth for seedling emergence. This disturbance should be in one pass concomitant with seeding. A light cultivation may be needed between seasons to control weeds or incorporate stubble. You will need to discuss with the farmer which implements already in use on-farm will be appropriate.

The no-till treatment will be largely dependent on herbicides for weed control so soil disturbance should be that required to open a slot just wide enough for seed and fertilizer placement. This may be only a few millimetres wide. Check on photos in this chapter for ideas on what machinery might suit the farmer’s requirements or how current machinery might be modified. Choice of herbicide and timing of application(s) will vary with location. The aim should be to minimize applications yet achieve sufficient control of weeds to avoid curtailing wheat seedling emergence and early growth.

Spatial design of the study A strip design In the simplest design the three tillage treatments selected for study can be sown as strips next to each other, conventional (C), minimum (M), zero-till (Z), each one or two seeder widths wide and more than 30 m long. The strips can then be marked out as plots each at least 10 m long to serve as three replicates.

If the plots are within a field that is cultivated and sown using the farmer’s normal procedures, the C strip can be part of the larger crop (see Figure showing a simple layout for a tillage experiment). Choose a part of the farm for the trial where the tillage treatments are most likely to give a benefit. Remember that the location of plots is to be a fixture for at least three years.

Sub-treatments If you have identified other issues that should be examined from the outset (options for how to control weeds; options for handling stubble such

63

as burning, incorporation or leaving on the surface; variations on when to sow the crop, etc.) then make the nine plots bigger than 10 m long and split each one to include the important sub-treatment (see the example for how to do this in the chapter on optimizing crop sequences). You may not need to include the sub-treatment on all tillage treatments, but always include it within all replicates of the particular tillage treatment.

Alternatively, make the sub-treatments full treatments by running extra strip plots for each one.

If you are planning to have two sowing dates, then mark out the plots two seeder widths wide and plant the first seeder width as the first date. Plant the second date next to it. Plant in the direction of the arrows in the layout figure in a continuous run from replicate 1 through replicate 3. The layout is repeated for each phase of the rotation and hence for fallow or for crops other than wheat.

If the crops will be harvested by machine, each sub-treatment (subplot) will need to be about 10 m long. If harvests will be by hand, the subplots could be as short as 2 m.

Measuring rainfall infiltration A simple layout for a tillage experiment with three replicates within the farmer’s crop. The conventional tillage strip is part of the farmer’s crop. The minimum and zero till strips are added within the crop.

If the farm has erosion problems, you may decide to include measurements of rainfall run off in the second or later years of the study. Infiltration is the difference between rainfall measured in a rain gauge and rain measured as run off from the tillage treatment. Erosion is assessed as the degree of turbidity of the run off water (how much sediment it contains). These measurements are made with homemade infiltrometers placed within one replicate of each tillage treatment. You may decide to add a fourth replicate plot for the infiltrometers, three per tillage treatment. The plot need only include the main treatments. Building an infiltrometer and how it works

An infiltrometer is a means of collecting and measuring the run off during rainfall events. It is a watertight wall enclosing a measured area of the field, but the wall has one leakage hole at its lowest point. Any water that does not filter down into the soil will temporarily collect on the surface within this enclosure and run out of the hole. The escaping water is caught in a pipe and collected at a lower point in a measuring container, usually a drum.

Infiltrometer walls can be made out of anything that will not absorb water and can be any shape. Galvanized steel sheeting 1 mm thick by 200 mm wide is suitable. Three 2.4 metre lengths bolted together with small overlaps will provide a circular four square metre infiltrometer area. A hole is required in the steel wall to fit a 25 mm diameter or larger drainage hose. Either you or the farmer should have little difficulty making such infiltrometers.

Measuring the runoff from infiltrometers. The grey galvanized iron walls of six infiltrometers enclose the remaining stubble of a zero-till wheat crop. Drums storing runoff are in a trench so they are slightly lower than the infiltrometer walls

64

After construction the circular infiltrometer wall is positioned in its plot on the soil surface and pushed vertically into the soil until the drainage outlet is level with the soil surface.

After rain, the runoff flows under gravity into a container in a trench, one container per infiltrometer. The volume of runoff is measured after each runoff-producing rain if possible. Percentage runoff can then be calculated by subtracting the collected amount from rainfall measured in a rain gauge located on-site. A 60 litre container will collect 15 mm runoff from a four square metre infiltrometer (1 mm of rain falling on 1 m² is 1 litre).

You may decide to put your containers in trenches dug in the paths separating the zero and minimum tillage plots and minimum and conventional tillage plots (white strips in the layout diagram).

The amount of sediment in the runoff can be used to approximate erosion by measuring turbidity of the runoff water by eye. First make a visual or descriptive scale of water clarity from 1 to 4. Each time volume measurements are made score each run-off sample after agitating it. Record the scores. A pattern will become apparent over time.

Infiltrometer rings inserted immediately after sowing can be temporarily removed at harvest if necessary then replaced until immediately before sowing the following season’s crop, in order to record infiltration amounts during any out-of-season rains. The crop inside the infiltrometer must be treated exactly the same as the crop in the remainder of the plot.

Observing and measuring the crop Observations of the crop should show which tillage system produces the best economic yield (allowing for costs of inputs) and the most sustainable yield as well as explaining or at least indicating why. Check the introductory chapters for how to collect measure and analyse plant samples and determine yield. Those chapters also give a guide to measuring crop water use and water use efficiency that could be an important part of your trial.

At seedling and spike emergence: counts of seedlings during the two weeks after sowing show whether the method of seedbed preparation limited early growth while counts of spikes at first spike emergence indicate the degree to which later vegetative growth was constrained. Greater water infiltration would be likely to increase spike numbers and potential grain production.

Counts of seedlings and spikes can be made on metre lengths of row that are selected and marked at sowing. These row lengths should be within the crop, that is, bordered on all sides by plants of the same tillage treatment. Two such lengths within each replicate plot are sufficient. While these are quick and easy measurements that can be taken in an hour or two they can be a useful guide to explaining yield differences between treatments. So if possible, collect the numbers with the farmer, find the average and then discuss them.

If soil pathogens or nematodes are likely to be a factor amplified by the main treatments, their occurrence should be estimated. Check with your

Weekly rainfall and infiltration data. These data are from the third year of a study after multiple, minimum and no-till treatments began. No-till used a double disc seeder opener at sowing such as pictured earlier. Minimum tillage used full cut-out tine points 50 mm deep, 3-pass cultivation used wide tine points 80 mm deep. Rainfall is the full height of each histogram. Sowing and anthesis dates are marked.

65

research colleagues as to how this might be done.

At maturity: grain yield and associated kernel weight are essential measures. Biomass measured at maturity allows harvest index to be calculated (a useful estimate of crop efficiency).

Check the pamphlet “Optimizing plant population, crop emergence, establishment and sowing rate” for harvesting methods for full length strip plots, particularly the use of grab samples to rapidly assess harvest index and other yield components.

Measures at final harvest should be based on crop samples of at least 1 m² area cut to soil level. The 1 m² sample areas should be surrounded by non-harvested border rows. It is essential to measure the area of crop sampled in each plot. In infiltrometer plots the whole enclosed 4 m² area should be harvested.

If the farm has suitable machinery, preferably harvest the whole sub-plots or whole strips. The area harvested must be measured.

Weeds: a major difference between the tillage treatments may be weed production, particularly if there is no prior experience on-farm with the use of weedicides in reduced tillage situations. Rough estimates of weed effects can be made by estimating the proportion of ground covered by weeds at first tillering. Crop yield losses will approximate percentage weed cover. Make these estimates on each sub-plot separately and write down the numbers. Discuss them with the farmer.

A better though more demanding way to estimate yield loss due to weeds is to cut a known area (1 m²) of each plot to soil level at spike emergence and then weigh the weeds and the crop separately once dry. Yield without weeds would have been actual grain yield increased by the proportion of biomass produced by weeds. Weeds germinating after spike emergence have little negative effect on crop production. For future decisions about types of herbicides to use, it is worthwhile identifying the major species of weeds in the biomass samples.

Working with the results Using the crop measurements Seedling emergence: until the systems stabilize there may be no benefits seen from the reduced tillage systems. You may decide to make only cursory observations during the first two years after the treatments are commenced.

Once your detailed study begins, take the counts of seedlings emerged from each of the treatments and graph them separately against time from sowing following the methods described in the chapter on crop establishment.

These counts and curves will show in which treatments seedlings emerged first and what proportion of sowed seeds actually became established seedlings.

In a well-established no-till system, seedlings may emerge slightly earlier and tiller more strongly when hoe type no-till drill is used than under full cultivation. This is because under a no-till system less of the soil water stored over the off-season is lost to evaporation so more water may be available for early growth. Furthermore, rain tends to concentrate in the seeding slots rather than dissipating over the whole width of the seedbed and thus soaks more quickly to the developing seedlings. Earlier emergence can lead to greater yield in short-duration crops.

However, on the contrary, seedlings may emerge less quickly through mulch, as the soil surface tends to be cooler. So where no-till is associated with heavy mulch, emergence may be late and tillering and potential yield curtailed. This is more likely to happen when a disk type no-till drill is used.

If you have used three or more replicates, compare tillage treatments within each block separately after comparing performance averaged for the whole trial. Each replicate should give a similar ranking for the tillage treatments. If that is not the case, the effects are possibly due to random variation. Nevertheless, try hard with the farmer to link differences between replicates to observations you and the farmer have made and the measurements you have on hand. There are always reasons for different behaviours.

Spike numbers: compare the data of spike numbers similarly. Do the tillage treatments rank the same across blocks? Do the rankings for spike numbers match those for seedling emergence with early emergence equating with more spikes? Do your estimates of weed infestation link to the numbers of spikes?

If a biomass harvest has been taken at this stage, dry and weigh the crop and weed fractions separately for each plot. Add the numbers for the two fractions together to work out biomass per plot. Calculate the percentage potential crop losses due to weeds by

66

100 X weed weight

crop weight

Multiplying actual yield at the final harvest by that percentage figure approximates yield losses due to weeds. This enables the farmer to work out crop losses in money terms and to equate that loss with the cost of herbicide or labour to reduce the weed problem.

The final harvest: this is the one that matters most to the farmer. If grain yield per m² is well down on the no-till plots compared with conventional cultivation, and money has been spent on herbicides, the trial may seem a failure. The savings in fuel and time in doing less tillage will have to be factored into the gross returns analysis to assess the degree of economic failure or success.

First, if different areas of crop have been harvested in different plots, convert all the results to a metre square basis or to t/ha. Does ranking of cultivation systems for grain yield remain constant across blocks? Can any differences in ranking across blocks be linked with earlier measured differences in weed infestations, poor emergence or low tiller numbers in specific plots? Are the differences in yield between treatments so small that they are really insignificant? Can it be concluded that tillage system has no effect on yield?

Try throughout to explain why things happened so you can find ways to fix the problem later.

If total crop dry weight (grain, straw and trash) was measured for each plot at harvest, work out harvest index (HI) as grain divided by total. A very good crop should have a value of about 0.5 i.e. 50 percent of the crop’s above-ground weight should be grain. If the reduced tillage treatments were collecting and storing rainwater better than the conventional tillage system, particularly late in growth, harvest index will be higher. If early growth was excessive and water ran out for grain filling, HI could fall to 30 percent or less.

Check average kernel weight in the different treatments by taking a handful of grain from each plot then counting and weighing 100 kernels. If any weights are much below 3 g, the treatment was short of water during grain filling.

Rank the tillage treatments in each block separately. Are the rankings consistent or differences absent?

Rainfall and infiltration data Data from the study described under ‘Measuring rainfall infiltration’ are used here to indicate how your infiltration numbers might appear. Depending on your soil type and previous use of the land, differences between treatments may not appear until the third year of the study or even later.

Once the soils under the minimal or no-till systems stabilize you would expect to see more infiltration and less runoff and erosion occurring than under multi-pass tillage. This occurred in the histograms of the figure. The red bars showing infiltration (the difference between rainfall and runoff) for the no-till treatment are taller than the blue bars for multi-pass cultivation.

However, when the soil profiles become saturated runoff from all three treatments should be similar. Equally, if there are high-intensity rainfall events, differences will be small or even absent.

Check to see whether the infiltrometer data from your study are similar. Also check your measurements of turbidity. If the runoff water was always clear from all tillage treatments there are no problems. If the water was consistently more turbid from the multi-pass system you have an erosion problem that should be fixed. Does zero- tillage fix the problem?

When performing all these measurements, check to see if there is any indication in the data that suggest greater infiltration leads to better growth and yield?

Factors to consider in selecting the best tillage system The trial does have the major drawback that it must be continued over several seasons if it is going to give a definitive answer about which tillage system is right for the farm or for the site on-farm where it is tested. In some soils and environments it may be five years before soil flora and fauna stabilize under minimum and no-till systems.

It may be most efficient on effort when doing this study to establish the different tillage systems and make only cursory observations and measurements of the crops in the first year.

In the second year, place the infiltrometers and make only learning measurements on them. Save detailed observations for the third or later years.

67

The farm will have to gain experience during the first year on how or whether weeds can be controlled by chemical means (no-till) or by limited till (minimal tillage plots). Residues

When applying no-tillage systems, there will be a learning curve for how to manage residues. Is it possible to drill through residues of the previous crop? Should residues alternatively be removed, windrowed, burnt or incorporated. Are they a source of disease?

If the area is dry for most of the season, residues are less likely to be a disease source than if it is wet. If left on the surface, residues will act as a mulch to keep soil surface temperature down and conserve moisture.

While this is likely to be beneficial in hotter regions, in cool areas it may delay crop emergence and development too much with negative consequences to yield. If it is a warm and wet decaying surface, residues can tie up nitrogen and result in the crop looking nitrogen deficient. Incorporated residues act differently to surface residues.

The interactions between residues and weather are numerous and it is not possible to give blanket suggestions for how they should be handled. An approach must be refined for the region.

A common conclusion is that they are more trouble than they are worth, so they should be burnt. This is a simple but wasteful and polluting solution. Furthermore, if retained they can reduce or prevent erosion caused by heavy rain. In areas with sloping cropland this could be important. For example, erosion can be stopped on 15 percent slopes if the surface is covered by straw (about 4 t ha –1 is needed) and the soil moderately permeable to water. When using any level of tillage on slopes it is always wise to work along the contours. Weeds

If weeds emerge before the crop, they will rapidly dominate the use of light, water and nutrients and crop yield will be small. During the early years of establishing a no-till system, it will probably be essential to control aggressive established weeds with herbicides. It may take as long as five seasons of spraying before the soil seed bank of weeds with long dormancy is completely cleared.

Annuals with little dormancy are soon cleared as long as those weeds do not surround the cropped

area. Some grass weeds that can regenerate from vegetative fragments can be best cleared by herbicides or by an initial deep ploughing. Shallow normal tillage done poorly can exacerbate the problem.

The positive side of no-till is that once the seed bank of weeds is at a low level, weeds can be controlled by limited chemical applications or by occasional minimal tillage. Erosion

The benefit of measuring runoff water from plots is that it provides information throughout on that very important aspect of sustainable farming, erosion control. Even before the minimum or no-till systems stabilize, the trial will allow decisions to be made about the level of danger in the present system. Is erosion a problem over the whole farm or of no consequence? And are the sloping areas of the farm under threat of washing away? Timing

A major advantage of minimum or no-till systems, speed of land preparation and timeliness of crop turn around and sowing, and the reduced costs of tillage, are aspects that should be considered when deciding on a tillage system for the farm.

A mix of no-till, minimum till and conventional cultivation will spread the tillage workload through time. By selecting crop varieties of varying duration matched to the systems, farm production can be optimized in hand with increased sustainability.

Some conclusions Transferring new tillage systems to the farmer's field to replace traditional systems is a time consuming effort that needs a well considered strategy and full commitment of researchers and farmers. Depending on type of soils, crop rotation and climate, the trial may need to be developed for at least 6 years or even as a permanent system. It is well known that when appropriately managed, no-tillage systems improve water dissipation in soil, prevent erosion and ameliorate soil quality attributes. Grain yield increase is more likely related to the management of crop residues and the control of weeds and diseases under no-tillage. Judicious used of a rotation is of paramount important of the success of no-tillage in any farm–transfer program. In trials conducted on the research station and in farmers’ fields, combining suitable rotations with an adequate

68

level of crop residue and applied herbicides, yields of wheat were always higher under no-tillage than conventional tillage systems, more so in dry years.

The most important point is that no–till or minimum till must always be seen as one component of a farm management system. Different ecoregions, different soils and toposequence and different climates will require components of the system to be tuned to optimize the whole package. The “Explore” types of trials that ask questions as they collect data through the cropping cycles, and respond by updated management, are a critical tool in the optimization process.

Further reading Bouzza, A. 1990. Conservation tillage in wheat rotations under several management and tillage systems in semi-arid areas. Ph.D. Diss. University of Nebraska, Lincoln, USA, 200p

Cook, R.J. 1990. Diseases caused by root-infecting pathogens in dryland agriculture. Advances in Soil Science 13, 215-39.

FAO. 2000. Manual on integrated soil management and conservation practices. FAO Land and Water Bulletin N° 8, Rome, Italy.

FAO. 2001. Conservation agriculture - Case studies in Latin America and Africa. FAO Soil Bulletin N° 78, Rome, Italy.

FAO. 2002. Training modules on Conservation Agriculture. CD-ROM AGL Series N° 22. FAO, Rome, Italy.

Freitas, V.H. 2000. Soil management and conservation for small farms. FAO Soil Bulletin N° 77, Rome, Italy.

Hamblin, A.P. 1987. The effect of tillage on soil physical conditions. In ‘Tillage: New Directions for Australian Agriculture.’ (Eds P.S. Cornish and J.E. Pratley.) pp. 128-170. (Inkata Press: Sydney).

Kirkegaard, J.A., Angus, J.F., Gardner, P.A. & Muller, W. 1994. Reduced growth and yield of wheat with conservation cropping. 1. Field studies in the first year of the cropping phase. Australian Journal of Agricultural Research 45, 511-528.

Kirkegaard, J.A. 1995. A review of trends in wheat yield responses to conservation cropping in Australia. Australian Journal of Experimental Agriculture 35, 835-848.

Landers, J. 2001: Zero Tillage Development in Tropical Brazil - The Story of a Successful NGO Activity, FAO-AGS Bulletin N° 147, Rome, Italy.

Mrabet R., Bouzza A., Peterson G.A., 1993. Potentiel reduction of soil erosion in Morocco using no-till systems. Agronomy Abstract, American Soc. Agronomy, Madison, WI, USA. p.323.

Mrabet R., 1997. Crop residue management and tillage systems for water conservation in a semiarid area of Morocco. Ph.D Thesis, Colorado State University, USA. 207p.

Mrabet R., 2000a. Differential response of wheat to tillage management systems under continuous cropping in a semi-arid area of Morocco. Field Crops Research 66, 2: 165-174

Mrabet R., 2000b. Long-term No-tillage influence on soil quality and wheat production in semiarid area of Morocco. In: Morison, J.E. (Ed.). proceeding of 15th ISTRO Conference Tillage at the Threshold of the 21st Century: Looking Ahead, Fort Worth, Texas USA July 2-7, 2000.

Mrabet, R., K. Ibno-Namr, F. Bessam & N. Saber. 2001. Soil Chemical Quality Changes and implications For Fertilizer Management after 11 Years of No-Tillage Wheat Production Systems In Semiarid Morocco. Land Degradation & Development 12:505-517.

Mrabet, R. 2002. Wheat yield and water use efficiency under contrasting residue and tillage management systems in a semiarid area of Morocco. Experimental Agriculture 38 :237-248.

Saber, N. & Mrabet, R. 2002. Impact of no tillage and crop sequence on selected soil quality attributes of a vertic calcixeroll soil in Morocco. Agronomie 22, 451-459.

69

Optimizing Nitrogen Use on the Farm When aiming to optimize yield and protein of wheat grown under rainfed conditions, farmers must supply sufficient, not excessive, nitrogen. What works one year may not necessarily be right the next year. The variability of rainfall amount and distribution results in highly variable responses to a set rate of N-fertilizer. Farmers should aim to match the supply of nitrogen, mineralized from soil organic matter or from a bag of fertilizer, with the requirements of the crop. Crop requirements constantly change throughout growth depending on available soil water and rainfall. Both rainfall and crop stage need to be measured and monitored to assess how much N to top-dress and when. This chapter details trials that help identify if nitrogen is a prime limitation to yield on a farm. Secondarily, it explains how to target yield and change grain protein percentage in the crops. It explains how to calculate how much N should be applied and when it should be applied to ensure that yield responses are more consistent from year to year. For introductory methodologies you should check “Constraints to cereal–based rainfed cropping in Mediterranean environments and methods to measure and minimize their effects”. Check the chapter on cropping sequences for effects of rotations, using legumes and non legumes, on nitrogen.

Which farms could benefit from these nitrogen trials? ☺ Those where little or no nitrogen is applied to

crops either directly or via rotations; ☺ those farms with lower yields in most years

than surrounding farms; ☺ those farms with significantly lower yields

than reported from local research trials; ☺ those where symptoms of haying-off appear; ☺ those with variable rainfall during the season

and across years; ☺ those farms that consistently produce grain

with protein contents lower than required.

Too little nitrogen and low yield Many farms in the Mediterranean region have low wheat yields of around 1 t/ha. For a high proportion of these farms yield could be substantially boosted by the application of nitrogen from a bag or via crop rotations. One possible response to N is shown in the figure. It goes to much higher levels of nitrogen application and yield than would apply in North Africa where the rain-limited potential is probably closer to 3.5 to 4 t/ha, but it is intended to demonstrate the likely shape of the response. Other responses appear later. In North Africa, farmers use year-long fallows as a way of making nitrogen available to following crops through mineralization of soil N. This can

be equivalent to 40-60 kgN/ha, a useful amount but still limiting yield. In these fallow years the land and stubble remaining from the previous crop is grazed.

Too much nitrogen and haying-off Farmers growing cereals where rainfall is both limited and variable in distribution, the normal situation for much of North Africa, must also be very careful not to apply too much nitrogen. They must also be careful not to apply it at the wrong stages in crop development. When water is adequate, nitrogen stimulates the crop to grow faster and accumulate more biomass and set a potentially high yield. If water then becomes inadequate to support that increased biomass as the season progresses, the crop responds by shedding leaf and other tissues solely

How grain yield and protein could change with applied N on a low-N farm. See later examples also

70

to survive. This is called ‘haying-off’. Depending on its severity it can cause serious economic loss to growers on three fronts; it decreases yield, the harvested crop is poor quality because of pinched or shrivelled grain and the applied nitrogen was wasted. Indeed, the costly applied nitrogen was more than wasted because it was the cause of the yield losses.

Very high levels of nitrogen can over-stimulate tillering which locks up the carbohydrates in structural materials rather than leaving them in storage to be used later to fill the grains.

Balancing nitrogen to water availability In those seasons when there is sufficient rain during spring to ensure there is little or no drought stress during the later stages of stem elongation through to grain filling, a higher nitrogen crop has a higher yield potential. It can retain more green leaves for longer and use them to fill the grain from current photosynthesis. It only needs to call on stored stem sugars during short periods of very high requirement. By contrast, in the event of post-flowering drought, these crops of high-N status suffer the combined stresses of the drought (haying-off) and shortage of stem sugar reserves. Crops of low nitrogen status grow less before flowering thus having relatively fewer grains to fill per unit land area and less above ground

biomass to supply with water after flowering. Consequently, during a late drought they lose relatively little of their green leaf. This green leaf maintains current photosynthesis and this assimilate, when combined with the high reserves of stem sugars, fills the grains satisfactorily through the drought. If the drought is particularly

severe causing green leaves to die, the stem reserves alone can go a long way towards filling the grains. How should farmers decide what level of nitrogen they should add to boost yield in good years but not cause haying-off in bad years? The solution is a series of collaborative on-farm trials with inputs measured and matched by calculation to requirements. Trials should span a region and be carried out preferably over two or more years. Spanning a region through time enables questions to be answered about the impact of soil type on nitrogen requirements and conjointly the impact of different environmental factors.

Too much N: Normal grain (left) and grain shrivelled due to over application of nitrogen leading to haying-off (right)

Crop management to maximize the benefit of supplementary nitrogen Wheat can be managed to produce grain at or close to the yield ceiling set by water availability in each season. If water availability is not measured or estimated through modelling accurate management is difficult. Management requires knowledge of what factors may be affecting yield other than nitrogen. It requires nitrogen budgeting before sowing and further budgeting as the season progresses. Each decision about whether or not to add N with time depends on cumulative and current rainfall and what targets have been set for yield and protein. Nitrogen will have little impact on crop yield if other factors present a greater limitation. For a good yield response to nitrogen attend first to the following

Criteria to satisfy first to realize full benefits from N applications • Sow locally adapted cultivars at the optimum

time (see chapter on “Optimizing variety, sowing date and crop establishment for the farm”);

When there is adequate water throughout grain filling, applied nitrogen boosts yield. However, when a drought occurs, then extra N may diminish yield.

• ensure there is a low risk of disease by using resistant cultivars and be aware that inappropriate N application may stimulate disease (e.g. Septoria) (see chapter on “What is the best cropping sequence for the farm?”);

• ensure nutrients such as phosphorus, zinc and trace elements are not limiting growth. If there is no knowledge about responses to these elements in the region, consider including them as a simple sub-treatment within the main N trial. This will take little effort and could bring large economic benefits;

• ensure there are no subsoil constraints to root growth such as sodicity or trace element toxicity or physical constraints such as a plough pan.

Discussions with collaborating farmers A good place to start your discussions with collaborating farmers is to find out, in general terms, to what extent the above-mentioned conditions have been met over recent years. As the proposed trials are about nitrogen, ask about direct applications of nitrogen to wheat in the last three seasons. Even without details you will soon be able to categorize a farmer as a non-user, a light haphazard user, a heavy user, a user who monitors the crop’s requirements and matches N applications to them. Your aim is to persuade the associated farmer group into that last category. Ask about yields and whether farmers are happy with their farms’ output. If yields approximate 1 t/ha, the farm is likely to be one of low input and a good candidate for a very basic N study, perhaps broadcasting different levels of N in strips within the farmer’s crop (described later) with a subsidiary test of other elements. If the farm produces higher yields, you might have to contemplate the slightly more complex N trials also described later. Expand your discussions to the other criteria above. It is likely that the haphazard N-user will

have paid little attention to other nutrients and other constraints. You might need to help with a soil test. Find out whether farmers use crop sequences that include a legume for harvesting or a legume for ploughing back in to improve soil organic matter and fertility. You may have discussed the earlier figure of N-response to demonstrate how yield and protein can change with N application, but it might be worth explaining that this is not the only possible response. The shape and level of the curves is dependent on many other interacting factors that may or may not be present on the farm. Data from two other farms contrasting with that in the earlier figure are shown here for comparison. You might like to use these figures to point out how important it is to do a basic low-effort study to categorize the likely response to N on that farm as no exact pattern can be assumed, even when soil tests have been done. Note also the different responses in protein because an input for that will be required when you do your N budgeting.

Nitrogen budgeting for a farm When you have made your assessment of the nutrition state of the farm and have some idea of the likely main constraints to yield, the next step is to construct your nitrogen budget for the farm. If you can arrange a N soil test before planting and complete the following budgets, you will be able to calculate N fertilizer requirements at sowing that will be accurate enough to meet targets of 70-80 percent of average yield.

Response to N applications at two farms contrasting with the farm in the earlier figure

More accurate updated assessments of nitrogen requirements can be made as the season progresses. Actions to rebalance N applications to crop requirements might, for example, be taken when the stems of the crop begin to elongate. This should be done to increase yield and protein but only as a tactical response to favourable soil water supply and low crop N status. It should not be done as a matter of course. A second tactical application to further increase protein can be made up to flowering, but this will also depend on water supply. Rainfall measurement is always an essential, simple and inexpensive part of N management. You should show your farmers how to do this. But first, work through this next simple budget calibrated for North Africa..

The following 3-step calculation estimates the nitrogen requirement and possible supply for a wheat crop yielding 2 t ha

How much N does the crop require? A simple method calibrated for North Africa

-1 of grain at 11 percent protein. It uses a correction factor of 2.34 that converts percent protein to kilograms of nitrogen and assumes that 25 percent of crop nitrogen is held in the straw.

As a first level of understanding of what N will be needed by the current crop, ask the farmer what yield he would like to reasonably achieve. Then work out how much N will be taken off in grain and straw for that yield. This is approximately how much N will have to be supplied. Step 1) Target crop yield x target protein percent

x correction factor = nitrogen required over the season e.g. 2 t ha

Help him through the following list that shows the N requirement for each 100 kg of grain yield for various cereal crops. -1 x 11 percent x correction factor

= 52 kg N ha-1 required.

So a 2.0 t/ha crop of bread wheat would require 60 kg N per ha (3.0*2000/100=60). If that were all supplied as urea that contains approximately 47% N, his application rate would be 125 kg urea per ha. If the farmer is unimpressed with these large application rates and is worried about costs, next point out the cost of the fertilizer against the additional value of the crop he will produce. The additional value should be at least twice the additional cost, otherwise the farmer will not be prepared to take the risk of this additional input. A guide as to what N is left by previous crops in North Africa Now tell him that there might be some cost-free N

in the soil he does not need to buy. It is left over from the previous activity in the farm rotation. Explain that these numbers are indicative and depend on whether a recommended amount of N was applied in these previous years or not. Alternatively, you can do a budget used in many countries. It has more flexibility than the North African approach, but gives similar answers How much N does the crop require? A more flexible calculation Start by calculating overall crop nitrogen requirements assuming a target for yield that is the average for recent years on the farm.

Similarly, a 1 t ha-1Durum wheat, 3.5 kg N per 100 kg yield Bread wheat, 3.0 kg N per 100 kg yield Barley, 2.4 kg N per 100 kg yield Oats, 2.2 kg N per 100 kg yield..

crop would require 1 t ha-1 x 11 percent x 2.34 = 26 kg N ha-1 from somewhere. You will see that this budget method gives similar answers to the North African method just described, but it gives the flexibility to adjust for protein content. How much N is already in the soil? This method also works out how much nitrogen may be in the soil. For best accuracy, test for existing soil mineral nitrogen by taking several soil cores to 60 cm from around the field. This is the extent to which many roots penetrate. Tests to 10 cm are of limited use but if it has been dry since the last harvest, 30 cm may be deep enough to allow for most mineral N. Send the samples to a testing agency or use the method of Wetselaar and colleagues for estimating soil nitrate levels in the field (see Further Reading). For the example calculation it is assumed that the tests indicated 20 kg N ha-1 of mineral nitrogen in the soil to a depth of 60 cm. After a fallow: 60 kg N/ha is in the soil

After Lucerne, 50 kg N/ha is present After potato, 20 kg N/ha is present After oats, no N remains

Remember that mineralization is a continuing process so will make additional nitrogen available throughout the growing season. Nitrogen mineralization varies between 60 kg ha-1 for infertile soils with less than 0.9 percent organic carbon to 100 kg ha-1 on fertile soils (more than 1.8 percent organic carbon). The calculation assumes the soil is infertile. Step 2) Existing soil mineral N + N mineralized during the season = total season supply in the soil e.g. 20 kg N ha-1 + 60 kg N ha-1

= 80 kg N ha-1 potentially available to crop. Step 3) However, actual recovery of nitrogen from soil supplies is often closer to 50 percent than 100 percent e.g. 0.5 x 80 kg N ha-1 = 40 kg N ha-1 actually available to the crop during the season.

The deficit in soil N available for the crop is (required N – available soil N) or ((1) – (3)). In the example for a crop of 2 t ha

What does N do when applied at different crop stages? Why you should split N applications

-1 this is 52 kg N ha-1 – 40 kg N ha-1 = 12 kg N ha-1.

The proportions of the components of crop yield can be changed by managing nitrogen fertilizer during the crop’s growth. This change is possible because N most affects those parts of the crop that are growing at the time it is applied.

As only about 50 percent of applied fertilizer is available in the first season, for our example, the estimated requirement for fertilizer to be applied is 12 kg N ha-1 x 2 = 24 kg N ha-1.

-1So, for a 1 t ha crop the entire requirement could be met by soil mineralization. In fact, mineralization alone could support a crop of 1.5 t ha

So when N is applied at sowing it helps early root growth and leaf production. After the three leaf stage it encourages tillering and tiller roots, during initial node elongation it increases spike numbers while in later elongation it makes more florets capable of producing grain. N applied close to anthesis primarily increases grain protein content.

-1 providing all other management options were optimized (e.g. correct seedbed preparation leading to good crop establishment).

Putting all N on at sowing can be a poor approach because it encourages the crop to put most effort into early growth. It is particularly poor if little remains for ensuring spike and grain fertility.

Budget to work out how much nitrogen the crop needs and its sources

calculation example

Crop N needs But remember, if management decisions or the environment seriously limit early growth, N applied late may not compensate for low yield potential. An adequate plant stand must first be established to make it worthwhile applying N during later growth.

target yield for crop

Encouraging the crop to arrange its form into the yield components that the farmer wants while ensuring growth is balanced to the environment, is a continuing exercise throughout the season. The figure at the end of this chapter explains when yield components are set during the crop’s development. Also read “Constraints to cereal-based rainfed cropping in Mediterranean environments and methods to measure and minimize their effects”. Water and nitrogen Available soil water should be assessed at the time a decision needs to be made about whether to apply supplementary N-fertilizer. As a guide, if rainfall from sowing up to decision time exceeds the potential evaporation then the storage of water in the soil is likely to have increased. Then the chance of getting a positive response to N-fertilizer applied at this time is high. If rainfall has been low or absent then the response will probably be small. It should be borne in mind that nitrogen is used much more efficiently if applied just before a rainfall event.

A 2 t ha-1

B target protein % 11% correction factor of 2.34 C

52 kg N haD crop N needs (A x B x C) -1

Soil N supply 20 kg N

hamineral N at sowing to 60 cm E -1

mineralization of soil N during season

60 kg N haF -1

80 kg N haG gross N supply (E + F) -1

net N supply assuming 50% available to crop

40 kg N haH (G) x 0.5 -1

N the farmer should add

12 kg N haJ extra N needed (D – H) -1

only 50% applied N is available to crop so N to add is -

24 kg N haJ x 2 K -1

Interpreting the effects of N fertilizer applied at sowing: A Case Study The following trial shows that nitrogen applied at sowing can have positive or negative effects on yield depending on the availability of water at different stages of development. It shows that you have to be very careful with nitrogen. It was carried out in a dry region. The six wheat cultivars used differed very little in duration, all reaching anthesis over an eight-day period (anthesis dates 1 to 8 on the figure), but this turned out to be very important to yield. On average across the six cultivars, an application of 37 kg N ha-1 at sowing increased yield by about 200 kg ha-1. This ranged from a 350 kg ha-1 yield increase for the two earliest flowering cultivars to a yield decrease of 200 kg ha-1 for the two latest cultivars flowering one week later (check the green up and down arrows showing response to N in the figure).

Why did N not always increase yield? A yield increase would normally be expected in all six cultivars from this small amount of N fertilizer, but the weather complicated the response. All crops were exposed to a hot, very dry period from immediately preceding anthesis and into early grain filling so their soil water reserves were significantly depleted. Water, not N, became the major constraint to yield. Cultivars that flowered early in the drought had six more days of water to use on filling grain than the cultivars that flowered later. The early cultivars could fill the extra grain sites that had

resulted from the 37 kg ha-1 N applied at sowing. By contrast the latest two cultivars had inadequate water to complete grain filling and ran out earlier when N was applied than when N was absent. They ran out earlier because the added N had stimulated tillering and leaf area production. This in turn had led to faster and greater total use of stored soil moisture before anthesis. This trial was chosen to demonstrate that there are many complex interactions occurring in crops. In this case, a failure to respond to N was nothing to do with the genotype’s inherent responsiveness to N, but due to a third factor that was dominant in that particular season. Always be alert for such additional variables. Repeating trials over several seasons is often the only way to average out their impact when optimizing a farming package for a location.

Setting up your own on-farm trials

A nominal N response curve If you are dealing with collaborating farmers who have not applied N previously, used it in very small amounts, or used it sporadically without knowing why, you need to determine a nominal response curve to N for the location. An on farm N trial is often more reliable than a series of soil analyses and is more convincing for the farmer. Make this a low effort study repeated within the crops of several farmers in a region and apply the N to areas of their crops that were planted on different dates to different varieties if those are available.

Response to nitrogen by six cultivars of wheat that reached anthesis on different dates. The red line is the high N treatment and the blue line low N. It shows a positive response to N by those cultivars flowering early (cv 1 to cv 4), and a negative response by the two latest cultivars (cv 5 and cv 6).

There can be farmer’s future. With this basic trial, apply phosphorus at sowing using 150 kg super phosphate (46%P). This should be sufficient P to allow the crop to respond to N fertilizer. Apply N as ammonium nitrate (33.5%N). Using superphosphate to supply P instead of the more common diammonium phosphate (which also contains N) means that a real zero-N treatment can be included in the trial. This is a low-cost trial that will indicate whether, and at approximately what level, N application is profitable. You will need to show farmers that the additional profit from grain production far exceeds the additional cost of fertilizer. Construct a simple cost-benefit analysis to do this. The second level: more N and fine tuning This trial is more suitable for farms that already have used N fertilizer and found it increases crop yield. It is aimed to determine whether the farms are using their N most profitably. Here there should be 5 strip plots in each replicate, not 4 as in the layout figure, making 10 plots if there are two replicates. There will be 5 N treatments, N0 to N4. At sowing apply di-ammonium phosphate (DAP) to all plots at 100 kg/ha. Because DAP contains 18% N (and 46% P), this means that all plots receive a basal dose of 18 kgN/ha and 46 kgP/ha. At tillering, when there are 4 leaves on plants, the plots labelled N0, N1, N2, N3 and N4 receive 0, 9, 18, 27, and 36 kgN/ha respectively. This would generally be applied as ammonium nitrate (33.5%N) or urea (about 46%N). Again at stem elongation apply 0, 9, 18, 27, and 36 kgN/ha to the N0, N1, N2, N3 and N4 plots as long as there is sufficient water available. This brings the total amount of N applied in the respective treatments to 18, 36, 54, 72 and 90 kgN/ha. If rain does not fall during the season and the stem elongation application is not made, the total amounts will be 18, 27, 36, 45 and 54 kgN/ha. You can calculate how much fertilizer you need by dividing the N application needed by the percent N content of the fertilizer. So 18 kg N/ha requires 53 kg ammonium nitrate ha-1 (34 percent N, i.e. 18 divided by 0.34) or 39 kg urea/ha (46 percent N, i.e. i.e. 18 divided by 0.46)). To work out the actual weight of fertilizer to broadcast on a plot, multiply this calculated per hectare weight by the metre length x breadth of the plot divided by 10,000 (a hectare is 100 x 100

= 10,000 m²). If you have problems broadcasting these small amounts of fertilizer accurately, increase the bulk of the sample by mixing it with sand and broadcast the mixture. If topsoil has a neutral to alkaline pH then urea should not be applied to the surface, as there is a high chance of losses due to ammonia volatilization. Prior to sowing, urea can be drilled into the soil. When topdressing, use nitrate-containing forms of nitrogen fertilizer to minimize losses. The third level: Are trace elements limiting? Application of N-fertilizer may induce a nutrient deficiency through its stimulation of growth. If you suspect that trace elements might be limiting the response of the crops to N and P, apply a full trace element cocktail to a section of each plot. A continuous area 2m wide through all strips should be sufficient to visually assess a response. Use a foliar spray if available. If a visual assessment close to grain maturity indicates a response and you want to scientifically gauge the response and any interactions with N, you should harvest and thresh this area of each plot separately. Do not forget to accurately measure the actual areas harvested. If you get a response to the complete trace element spray then in later trials you can apply treatments with single elements to determine which of the trace elements is needed. A fourth variant: other elements You can use exactly the same approach with strip plots marked out in the farmer’s crop to test whether other elements are limiting production. Just broadcast different levels of the chemical to check for visual or measured response. If supplementary irrigation water is being used, you might need to test whether potash (K) should be added and in what amount. What you do will depend on your observations of nutrient deficiency symptoms in the farmer’s crops.

Some rules to follow with these trials: • Make your decisions about what trials to run

in discussion with the farmer(s). Their enthusiastic support is critical to current and on-going success. They must understand why the trials are being done and understand the steps in the N budgets presented earlier. They should also understand a simple cost-benefit analysis of the trials;

• locate the nitrogen plots in step fashion following the direction of the seeder as in the

figure; this avoids variability between plots due to inaccurately spaced tynes and allows the farmer to manage weeds in all plots with a single pass of machinery;

• make plots at least 4.2 m wide by 10 m long. Length and breadth selected will depend on the farmer’s machinery and the farmer’s preference;

• mark out the plots initially with handfuls of lime or with painted sticks tightly hammered into the ground so the farmer can see them;

• if possible, use two replicates; • put brightly coloured corner posts around the

trial area at anthesis and make sure the plot division markers are still in place and readily visible. This will prevent accidental harvesting later by the farmer;

• Top-dress the nitrogen after sowing following rain. In alkaline soils if using urea, preferably drill it before sowing and then sow at 90°.

As a general rule, if the season is dry the application of supplementary nitrogen at stem elongation may result in haying-off and poor quality grain. However, if the season has been wet and there is adequate soil moisture, a positive response to more supplementary nitrogen is likely. If N is top dressed during the season then the highest efficiencies of crop uptake are achieved if the nitrogen is applied just prior to rainfall. Rain takes the N into the soil quickly for root uptake. Remember that the trials described are only a guide. Modify them in any way to solve the problem you are addressing. Again, do this in discussion with the farmer(s).

Measurements during your study Measuring water As emphasized earlier, the impact of nitrogen on crop growth and yield is dependent on the amount of water available as the crop progresses through its developmental stages. Consequently, to be able to interpret the effects of nitrogen correctly requires two things. First, you need a measure of rainfall during the season. It is very desirable to have a rain gauge at the site of each trial and for the farmer to measure, record and empty the rain collected each week. Rain gauges are cheap and farmers very quickly learn to read them and take an interest in the information they provide. They learn that a certain amount of rain means they can profitably

broadcast a certain amount of fertilizer. Information on rainfall from a nearby meteorological station is acceptable but far less useful because the farmer will not bother to access it or use it for fertilizer scheduling. Second, but far less important, you need soil moisture measurements at the start and end of the trial. Together, these numbers will help with scheduling fertilizer application and enable you to work out crop water use and water use efficiency. How to do this is detailed in the pamphlet called “Constraints to cereal-based rainfed cropping in Mediterranean environ-ments and methods to measure and minimize their effects”. Measuring the crop to determine yield and yield components The measurements that follow are designed to help interpret the results of the trials. If you and the farmers are concerned solely with recording the responses to nitrogen or other fertilizers without understanding in detail why they happened you only need to measure grain yield. This will usually be based on a machine harvest of the whole area of each plot. Do not forget to measure the exact area harvested. However, if you want to get an idea of what went wrong or right with the crop during growth you need some more detailed measurements just before the machine harvest. These are not very onerous. Before the machine harvest you should remove about 5 grab samples from each plot. Just reach into the plot at random and grab and cut off at ground level about 10 culms. Repeat this 5 times in separate places to make a bundle of 50 culms. Tie and mark the bundle with the treatment name and date. If you have 5 N levels, you will have to collect 5 bundles. After they have been spread out and well dried in the sun or in a dehydrator/oven, these bundles can provide you with 1. harvest index (weight of bundle grain divided

by weight of dried bundle before threshing), 2. approximate number of spikes per m² (weight

of whole plot grain divided by weight of bundle grain x 50 divided by full plot area in m²; if there are 50 spikes in the bundle),

3. number of grains per spike (count 100 grains and weigh them and divide that weight into the weight of all bundle grains, then divide that number by 50)

4. average kernel weight (the weight of 100 grains divided by 100)

5. number of grains per m² (you can do that calculation yourself) and

6. biomass per m² (result of (2) divided by 50 multiplied by weight of dried bundle)

That is a lot of information for not much work, but it must be done properly (detailed procedures are in “Constraints to cereal-based rainfed cropping in Mediterranean environments and methods to measure and minimize their effects”).

When you have the numbers, how do you interpret them? The following trial assumes you have applied four nitrogen treatments (0, 25, 50 and 75 kg ha-1) in strip plots, and have duplicated the trial over two farms (0 is the farmer’s normal practice). The aim of the trial is to determine whether the farmers would gain an economic benefit from applying more nitrogen and if so, how the crop changed its growth pattern to give that benefit. Crop grab samples were taken in the same way as just described. But they were taken at anthesis as

well as at maturity. Additionally, crop height was measured. As in the previous description, each complete plot was harvested at maturity for plot yield

The anthesis harvest N-fertilizer kg/ha 0 25 50 75

What does the anthesis table say? Farm 1

About Farm 1 Biomass (t/ha) 6.80 7.80 8.21 8.40 • There was a 1.6 t ha-1 growth response (6.80–

8.40 t ha-1) to a 75 kg ha-1 increase in N; Crop height (cm) 80 83 84 85 • crop height increased with fertilizer level. Farm 2 These observations together indicate that in this season, when the farmer’s normal practice was followed, N limited growth up to flowering.

Biomass (t/ha) 6.90 7.20 7.35 7.30

Crop height (cm) 77 79 79 77 About Farm 2 • N-fertilizer increased growth by 0.4 t ha-1 the

effect being almost saturated by an increment

of 50 kg N ha

-1; On Farm 2 it seems that the farmer’s normal practice is close to optimum prior to flowering. Water, not N, was probably limiting biomass production at higher rates of N.

The maturity harvest What do the maturity data say for Farm 1 Check the yield table as you read. • N-fertilizer increased yield overall by

0.70 t ha-1, but the first 25 kg ha-1 increment of N produced the most benefit;

• N also increased biomass and spike density presumably via more tillering, but a greater proportion of these additional shoots were sterile at high nitrogen (10 percent versus

In the example, season late drought was so severe that when higher increments of N were applied, kernels were unable to fill even to 30 mg and harvest index fell dramatically. In effect the additional N was harmful for the crop.

6 percent). This indicates that adding N intensified intra-crop competition for resources on this farm;

• Increases in N-fertilizer resulted in marginally taller crops. This indicates that conditions following flowering were satisfactory. That all crops also grew taller between anthesis and maturity supports this suggestion;

Which treatments produced a profit? Gross margins calculations help you to make decisions about whether N applications are worth the cost and effort. They have been worked out for the two farms of this imaginary study. Your prices will differ from these examples quoted in $ and based on costs overseas, but trends will remain the same. Furthermore, your grain receiver may not be concerned with grain quality (protein content or screenings), only with weight. Re-work the examples in accordance with local regulations.

• Indications are that yield increase due to N was primarily via an increase in spikes m-2, (leading to more kernels), but also through an increase in the number of kernels per spike. There was the commonly observed trade-off however that kernel weight and harvest index both decreased in response to N-fertilizer.

In summary, maturity data for Farm 1 indicate that in the farmer’s normal system N is a limitation to growth (and eventually yield) prior to flowering.

Assumptions in the calculations • That N fertilizer cost $1 per kg applied and that uptake efficiency was 50 percent. Harvest index for nitrogen was 0.78; So, depending on the relative cost of fertilizer and

the grain price, adding further N could be beneficial.

• crops under farmer practice (‘0’ treatment) had 10 percent grain protein with a selling price of $240/t when there are 0 percent screenings; Almost certainly the increment in yield of

530 kg ha-1 for 25 kg N ha-1 (just over one bag urea) would be an economic proposition.

• grain price increases with protein and decreases with increased screenings (pinched and shrivelled grains passing a 2 mm slotted screen); However, a further increment of 680 kg grain

associated with two to three bags of urea may not be worth the parallel risk of haying-off.

• the maximum tolerance of screenings is 10 percent. Above this grain is unacceptable for milling and is considered animal feed quality, worth only $150/t instead of $240/t.

What do the maturity data say for Farm 2? • N-fertilizer actually reduced yield by

0.46 t ha The Gross Margins -1 and there was no response in biomass production; Farm 1

• more spikes were produced with additional N, but not progressively at higher increments;

N-fertilizer kg/ha 0 25 50 75 Grain yield (t/ha) 3.25 3.78 3.93 3.95

• crop height was unaffected by changed N and plants failed to grow taller between anthesis and maturity;

Input costs ($/ha) 150 175 200 225 Grain protein (%) 10.0 10.1 11.1 13.6 Screenings (%) 2.1 2.1 3.0 4.5 • more sterile stems at high nitrogen (13 versus

6 percent at low N) indicates that competition for resources increased with N on this farm;

Value ($/t) 208 208 213 244 Crop value ($/ha) 676 786 837 964 Gross margin ($/ha) Value-costs

• the yield decrease resulted primarily from a reduction in kernel weight though kernel number increased.

526 611 637 739

$Return per $N 3.4 2.2 2.8 On Farm 2 it seems that it was fruitless to apply more N than in the farmer’s standard regime as growth was primarily limited by lack of water both prior to and after flowering. Though water limitation prior to flowering was severe enough to halt height growth it did not limit floret fertility. Consequently, there could still be arguments for delayed N applications in seasons with late rain.

For Farm 1 the gross margin in dollars per hectare increased with increasing amounts of supplementary N-fertilizer (the red figures in the table). However the greatest return on money spent on N fertilizer was at the lowest rate of N application, being $3.4 for every $1 spent.

For Farm 2 the greatest gross margin was for the farmer practice, the ‘0’ treatment. Additional N-fertilizer wasted money especially at the highest N rate due to haying-off and associated unacceptable levels of small grains screenings. At that rate, $4.10 was lost for every $1 spent on fertilizer.

$Return per $N -1.5 -0.5 -4.1

Some conclusions The prime conclusion from these trials is that nitrogen positively impacts on growth, yield and profits when used thoughtfully, but reduces yield when used in excess. It must be used in balance with other potential limitations to production, particularly water.

Farm 2

0 25 50 75 N-fertilizer kg/ha If farm management does not first attend to the

basics of good seedling establishment with optimum plant populations, with the best variety sown at the optimum time, with adequate control of weeds and disease, nitrogen applications may provide no economic benefits. Nitrogen applied to weedy crops is nitrogen wasted in growing weeds.

Grain yield (t/ha) 2.96 2.75 2.65 2.50 Input costs ($/ha) 150 175 200 225 Grain protein (%) 10.0 12.7 15.3 18.5 Screenings (%) 3.0 6.5 9.0 11.0 Value ($/t) 206 217 240 150 Crop value ($/ha) 610 597 636 375 Good agronomy based on knowledge of principles

and knowledge of the farm and its environment is the secret to success.

Gross margin ($/ha) Value-costs 460 422 436 150

The on-farm trials suggested here provide an opportunity for researchers to help farmers to budget for a correct nitrogen balance, to work with them to monitor water availability through the season and match nitrogen and water inputs for optimum productivity.

This diagram shows what plant parts are developing and growing throughout the crop cycle. Look for stem elongation (first node of stem visible), as this is a good time to add a second dressing of N after the basal application made before or at sowing. This second application should only be done if there is adequate soil moisture. Notice that the first node is when the terminal spikelet is laid down and that then all components of yield are forming apart from kernel weight.

Further reading Angus, J.F. & van Herwaarden, A.F. 2001. Increasing water use and water use efficiency in dryland wheat. Agronomy Journal 93, 290-8. Van Herwaarden, A.F., Farquhar, G.D., Angus, J.F., Richards, R.A. & Howe, G.N. 1998a. ‘Haying-off’, the negative grain yield

response of dryland wheat to nitrogen fertilizer. I Biomass, grain yield and water use. Australian Journal of Agricultural Research. 49, 1067-81. Van Herwaarden, A.F., Angus, J.F., Farquhar, G.D. & Richards, R.A. 1998b. II Carbohydrate and protein dynamics. Australian Journal of Agricultural Research. 49, 1083-93. Van Herwaarden. A.F., Richards, R.A., Farquhar, G.D. & Angus, J.F. 1998c. III The influence of heat shock and water stress. Australian Journal of Agricultural Research. 49, 1095-110. Wetselaar, R., Smith, G.D. & Angus, J.F. 1998. Field measurement of soil nitrate concentrations. Communications in Soil Science and Plant Analysis 29, 729- 39.

83

Cropping Sequencing If a farm has been in continuous wheat for many seasons and yields are poor or declining, the introduction of alternative species (break crops) and fallows into the cropping sequence might boost yield considerably. Interactions between the wheat and break crops are complex because there are variable roles of water, weeds, nitrogen, disease and weather. This chapter examines these complexities. It explores how and why they might influence the performance of wheat in different sequences and how they might impact on the economics of the farming system. It then outlines trials to test these interactions. These two-to three-year trials when complete should provide the basis for crop sequence management packages for the local region. To complete the story you will need to refer to the chapters “Constraints to cereal-based rainfed cropping in North Africa and methods to measure and minimize their effects”, “Optimizing plant population, crop emergence, establishment and sowing rate”, “Optimizing variety x sowing date for the farm”, and “Optimizing nitrogen use on the farm”.

Which farms could benefit from these trials? ☺ Those farms that have cropped wheat

continuously for many seasons; ☺ those farms with inexplicably declining

yields; ☺ those with ‘whiteheads’ during grain filling; ☺ those farms that are weedy, have root

diseases, nematode problems and generally poor growth;

☺ those consistently producing small, pinched

and shrivelled grain not attributed to drought; ☺ those exposed to soil degradation; ☺ those using clean or weedy fallow; ☺ those with livestock that need forage; ☺ those located near markets for alternative

crops and managed by farmers who are able to diversify.

Terminology: sequences versus rotations This chapter deals with crop sequences not rotations. In rotations different crops are grown in a fixed order year after year but in sequences the order is not necessarily repeated. The term ‘sequence’ captures the flexibility that is a feature of aware farm management, continuously responding to changing market demands and changing on-farm constraints. Here the focus is on short- to medium-term crop sequence trials rather than the fully phased long-term rotations traditionally conducted on research stations.

Background to cropping sequencing Generally, the yields of crops grown in continuous monoculture decline as a result of a build-up of soil or stubble-borne disease that is specific to that crop species. This problem can be overcome in some cases by the use of varieties that are resistant to the disease, but for some soil-borne diseases there is no current genetic resistance. Chemical control is usually not economical so that the only control strategy for these diseases is to grow a non-host or break crop that is not infected by the disease, so that the levels of disease in the

Impact of a break crop on the incidence of ‘whiteheads’ in a following wheat crop in Australia

soil are reduced prior to the next wheat crop. Crop rotation of this sort has been practised for thousands of years. Some break crops have their own chemical or biofumigant effects that can reduce soil pests. As well as reducing the levels of cereal disease, legume break crops, such as peas, lupins and lentils, also fix nitrogen from the air. This may provide a beneficial N contribution to following cereals that may persist for longer than one year. Break crops can also help to reduce weeds and pests that have particular characteristics or life

cycles that are assisted by the host crop.

Finding an appropriate sequence crop The challenge is to find break crops that farmers are willing to include in their crop sequence. They must be crops that will have a high chance of success in most seasons, wet or dry. Farmer adoption will usually depend on the following: • Can farmers readily grow the crop? Do they have the equipment, knowledge and skills? • Is there an accessible market for the product? • Is the product of use to the farm in the event that there might be no market? For example, might it be used as forage for farm animals? • Is the crop as profitable as the wheat crop it replaces? • Will there be other benefits to productivity of the whole farm in the longer-term? Break crops must be adapted to the soils and climate of the area and be accompanied by a good agronomic management package if they are to be adopted successfully by farmers. A correct break crop or sequence in a damp area may be a disaster in an arid area.

What influences a crop’s response to preceding break crops?

Disease and weed control The most significant benefit of break crops to following cereals is the reduction of weeds and soil-borne disease. In North Africa the yield of wheat after food legumes, sunflower, forages and

barley crops has averaged more than 20 percent higher than for wheat after wheat largely because of reduction of weeds and disease. Specifically, in Tunisia and Morocco, bromus weed species were reduced as were foliar and crown root diseases. These diseases are seen in drought years in depressions in fields in spring.

Nitrogen benefits In addition to providing a disease break, grain legume break crops such as faba bean, lentils, chickpeas, peas and lupins may provide residual N to following wheat crops. In some regions these benefits have boosted yields of following wheat crops by up to 50 percent, similar to the effects of applying adequate N fertilizer out of a bag. However, the magnitude of the N benefit depends on effective N fixation by the legume, the amount of biomass produced and how much N is removed in legume seeds or stubble. In North Africa the continuing benefits may be small or even negative if the legume crop is heavily grazed or completely removed. Some legume break crops, such as lentil, may also leave any deep residual soil N untapped because they have a shallow root system and use soil N inefficiently. Since residual N influences both yield and protein content of following cereals, it is important to consider the N balance of the whole system when attempting to understand crop sequence effects (see chapter on “Optimizing nitrogen use on the farm”). Benefits in N nutrition may also arise from non-legume break crops simply because they effect a healthier wheat root system, enabling the wheat crop to use soil N and applied N fertilizer more efficiently and thereby yield better.

Water use by preceding crops In drier environments or in dry seasons the amount of water removed from soils by preceding crops can have a major effect on the growth of following crops. In annual cropping cycles the inclusion of deep-rooted rotation crops such as sunflower) that remove significantly more water than wheat may cause yield loss in some seasons. Similarly, if there is no crop and the land is left as a clean fallow (free of weeds), any water harvested by the soil during that year will contribute to the yield of the following crop.

Impacts on nutrients other than N Be aware of the nutrient requirements and likely removal by prospective break crops. Brassica crops such as canola, remove large amounts of elements from the soil, particularly S and Zn.

The dominant pathogens within a crop sequence differ between environments, soils and seasons. You must identify the pathogens in your system and understand what influences their activity if you want to realistically assess the chance of success of a crop sequence.

These crops may induce deficiency in following crops if adequate amounts are not supplied. Arbuscular mycorrhizal fungi (AMF) colonize the roots of most crop plants and symbiotically help acquire immobile elements such as P and Zn. In general, tap-rooted crop species are dependent on AMF for P and Zn nutrition in low P soils while fibrous rooted cereals are less dependent. Canola and some lupins are among the few species that are not hosts of AMF so that AMF inoculum levels in soil are usually low following these crops. If following crops in the sequence are AMF dependent or if your soils are low in P or Zn, you may need to apply the required nutrients.

Stubble allelopathy and herbicide history Other interactions between crops in the sequence can arise from chemical interference caused by leached compounds from retained residues (allelopathy) or from residues of herbicides applied to the previous crop.

Seasonal interactions All the above-mentioned factors influenced by crop sequence also interact strongly with seasonal conditions. The complexity of the many possible interactions can lead to outcomes that may be difficult to understand or explain. For example, dry periods between seasons increase carry over of soil-borne diseases (dry conditions maintain inoculum levels). Root infection is increased by wet, cool conditions during early growth while dry conditions during grain fill, allow expression of the diseases as “whiteheads”. These heads have small pinched grains or no grains. Alternatively, wet periods prior to sowing may negate any differences in disease inoculum. This may be by reducing inoculum directly. However, it may also be via the indirect route of improving the following crop by refilling the profile with water and mineralizing large amounts of soil N, so making that crop less susceptible to disease. Such seasonal ‘weather’ interactions are largely responsible for variations in response to preceding crops. As there are so many complex variations and interactions, the only way to get a reliable picture across your region is to actually conduct crop sequence trials at several sites over several seasons. The human resource needed will clearly be considerable, ideally being interested, collaborating farmers working on their farms.

Each farmer will benefit by defining together with you the appropriate sequence for their location while you, as coordinator, will learn the driving forces for the region. The overview information should give you confidence to design sequence packages for any site within the region that will have a high chance of working.

Sequences used in North Africa

Wheat:Fallow:Wheat The most common sequence after continuous wheat is wheat-:fallow-wheat. This may be a fallow that grows weeds and the weeds are grazed by farm animals, or it may be a clean fallow, so called because the weeds are removed. Removal on progressive farms is by chemical means but on many farms it is by pulling or minimal tillage. In very arid regions rainfall is the main limitation to crop production and may even be inadequate to produce a harvestable crop each year. If the land is fallowed the rainfall that falls in the break year is stored in the soil for use by the following crop. After a clean fallow this may result in a wheat crop that has twice the yield of the single-year–crop because it has twice the annual water to grow on. Additionally, it has two year’s accumulation of naturally mineralised nitrogen rather than one. This could amount to 80 kgN/ha equivalents (see the chapter on optimising nitrogen use). This effect of fallow is shown in Table 1 where wheat after fallow produced 2.33 times the yield of continuous wheat in a dry year and 1.64 times the yield in a less dry year.

Table 1 Yield of wheat crops (t/ha) after a fallow year in a very dry or less dry season and the interaction with different tillage methods (Adapted from Bouzza 1990)

tillage mean

zero till

min till

conv till

trad- itional

dry year Wheat:wheat 0.55 0.74 0.58 0.55 0.31 Fallow:wheat 1.05 1.72 1.56 1.29 0.61 Fallow effect (multiplier) 2.33x 2.32 2.69 2.35 1.97 less dry year Wheat:wheat 2.26 2.30 2.41 2.11 2.21 Fallow:wheat 3.71 4.66 3.91 3.82 2.44 Fallow effect (multiplier) 1.64x 2.03 1.62 1.81 1.10

Interaction with tillage method Where water is the primary constraint to yield, any tillage method that wastes water will waste yield. So if the soil is extensively turned over and pulverised, that will aerate and dry the soil. The deeper this is done, the more water is wasted. Consequently, the best tillage method is that which aims to disturb the soil minimally; only during planting and only in a narrow path the width of the seed and to the depth needed to cover it. That is zero tillage. Table 1 shows that yield is invariably highest using zero tillage in this arid region whether wheat is grown continuously or after fallow. And, as argued, the benefits of prior fallow more than doubled yield by saving that year’s rainfall (2.32 and 2.03 times). Even minimum tillage caused sufficient stored water to be lost to reduce yield to a lower plateau than zero till but the benefits of the fallow remained proportionately similar.

Wheat:Forage:Wheat Mazhar (1987) evaluated the effect of rotation on herbage yield and wheat grain yield at Sidi El Aydi (Chaouia-Settat) and Jemaa Shaim (Safi-Abda). He assessed rotations including wheat followed by vetch-barley, medics, weedy fallow, clean fallow and concluded that despite large variations in the ranking of the herbage producers among years, medics and vetch-barley always produced more than weeds. Unlike weedy fallow, they also did not depress the yield of wheat in the following season. So if a farmer has stock, these rotations should be considered.

Wheat:Other Crop:Wheat Kacémi (1992) considered yield of wheat following fallow, corn, and the legumes lentil, faba bean and chickpea, and like Bouzza (1990), had zero and minimum tillage treatments throughout. He used two sites over several years. Unlike Bouzza’s finding zero till was not consistently better or worse than minimum tillage. The benefits of a clean fallow to the following wheat crop were clear (Table 2) and better than the effects of any of the legumes or corn. However, in economic terms the benefits of the fallow would have been smaller if the values of the alternate year crops are considered. Furthermore, the inputs associated with the alternate-year crops should be costed into the economic analysis of the two-year sum. Nutrient inputs required by corn can be high, though if not used completely by the corn crop they can benefit the following wheat crop.

Table 2 Effects of five break ‘crops’ on yield (t/ha) of the following wheat crop (data adapted from Kacémi 1992, quoted by El Mejahed, 1997)

sequence Wheat yield rank

fallow-wheat 3.09 1 corn-wheat 2.74 2 lentil-wheat 2.56 3 Faba bean-wheat 2.34 4 Cowpea-wheat 2.23 5

Quantifying the advantages of sequences using a gross margins analysis A good place to start assessing a proposed crop sequence is to do a rough cost and benefits analysis (gross margins). Decide on the crops you might use in the sequence then sum the likely income from each crop in the several-year sequence. Next, subtract the variable costs from that sum (seed, fertilizer, labour, fuel and machinery wear). Compare the result with a sum calculated for continuous wheat crops. How much would the yield of the wheat crop in the sequence have to improve to make an overall improved profit compared with continuous wheat? In some sequences, reduced direct income from the break crop might be offset by an increased income from a higher yielding wheat crop and/or by a reduced requirement for N fertilizer in the following year. Other costs and benefits, which are more difficult to quantify, are those associated with risk. On the negative side, the risk of crop failure may be increased if the break crops are not well adapted to the region. On the positive side, having product diversification in years of cereal failure or low cereal prices may lessen risk. The costs and benefits of spreading sowing and harvest workloads of the different crops and on longer-term improvements in soil fertility are also difficult to assess.

In dry years and in arid areas break crops may be a poor economic option for farmers because of lower yields and risk of crop failure of the break crops themselves, but in wetter years and wetter areas benefits of break crops on reducing disease and supplying N to following wheat crops can be very positive.

How break crops can alter wheat yield: an example study from Australia in dry and wet years (Kirkegaard 2002) The case study is from two trials (I and II) investigating how break crops, in an area in Australia normally in continuous wheat, affected yield of the following wheat crop. The key difference between the studies was that in trial I the post break crop wheat was grown in a dry season (180 mm) while in trial II the season was wet (300 mm).

Field cropping history and methods Wheat was grown on all plots in year 1. In year 2, field pea, canola (a Brassica) or wheat were grown in specific plots. Finally, in year 3, wheat was again grown in all plots. Yields of wheat crops from the final year are shown in the figure. The effective N contribution of the legume to the following wheat crop was examined in a subtrial by comparing its effects on the wheat yield with the effects on yield of three fertilizer N treatments, 0, 50 and 100 kg N ha-1.

Effects on crop performance The wheat crops grown in the dry year (trial I) had a limited yield potential (note that all yields in the left hand graph of the figure fell between 2 and 4 t ha-1). Despite this, they still had a sizable response to fertilizer N (compare the black points at 0 and 50 kg ha-1).

In this dry year the wheat crop following canola performed little better than that following wheat (compare the yellow and black lines). This is because in a dry year, disease barely develops, so the impact of a disease-cleansing crop such as canola cannot be demonstrated. A positive impact of the preceding legume break crop on wheat yield of 1.5 t ha-1 can be seen clearly when no N was supplied to the wheat. This equated with fertilizer applied at 50 kg N ha-1. The preceding pea crop had a grossly negative impact on wheat yield when the large amount of 100 kg N ha-1 was applied. The residual N together with the applied N would have amounted to 150 kg N ha-1 equivalents being available to the wheat. This caused excessive vegetative growth, rapid use of available stored water and finally severe “haying-off” (see chapter on “Optimizing nitrogen use”) when the crop ran out of water in the dry spring of that year. This is of particular importance in arid zones. For the wheat grown in the wetter year (the right hand graph) the level of soil disease was very high and as a result, wheat grown after wheat had low yield, and when N fertilizer was applied, there was an even lower yield than in the dry year. The positive impacts of the preceding break crops in this season of high disease were large, amounting to the more than doubling of wheat yield when fertilizer was applied.

Yield of wheat grown in three crop sequences, after wheat, after a Brassica (canola) and a legume, in two contrasting seasons, dry (180 mm rain) and wet (300 mm) with impact of applying N to the wheat crop shown

Your on-farm trial How dry is your region? It is useful to start by making an estimate of yield potential at your site using local rainfall data. Use the following rough guide. For each mm of growing season rainfall above 100 mm expect 15 (to 20) kg grain ha-1. Thus for 300 mm growing season rainfall, yield potential is (300-100) x 15 kg ha-1 = 3 (to 4) t ha-1. Similarly 200 mm should give 1.5 (to 2) t ha-1. This relationship errs on the conservative side and varies somewhat between regions, but try it anyway to indicate roughly how much yield the area should produce. You will notice that it considerably underestimates production in the Australian case study where management was optimized. If your area is yielding well below the calculation, it is likely that there is an underlying cause of yield limitation other than rainfall such as saline subsoils or soil compaction or water-wasting tillage practices. If the estimate is under 1.5 t ha-1 the amount nitrogen needed by the crop will be fully supplied by mineralized soil nitrogen (see chapter on optimizing nitrogen) so there will be little benefit in adding N or using a crop sequence including legumes that will potentially increase soil N. The primary aim then will be to introduce sequences

and tillage practices that increase water availability to the crop. Consider introducing fallows to harvest water (and N) or more water efficient species like barley into the sequence. If your area has sufficient rainfall to produce 3 t/ha or more, your options for sequences are much broader. But then you will need to consider fertilizer requirements and timing and interactions with the break crops contemplated. After deciding on your general options for sequences as determined by rainfall, next decide on the specific sequences that are most likely to increase farm productivity in the long term. A good starting point will be discussion with the farmers using a cost benefit analysis of the different options.

Arid areas In the arid areas where your sequence options will be limited to continuous wheat, wheat, clean fallow, wheat and wheat, weedy fallow, wheat, with possibly barley introduced at some point, you need to consider costs and outputs associated with the crops but also those associated with livestock on the farm. Is the weedy fallow the best solution for grazing purposes considering the weeds will likely reduce the yield and value of the following wheat crop? Are there alternatives for the animals? If your prospective farmers are not using zero tillage you should be attempting to show the costs

Part of a crop sequence study similar to the case study described (photo in Australia, J Kirkegaard)

and benefits of that technology within your trials. And don’t forget to encourage prospective farmer participants to use good seed by showing the benefits far outweigh the costs of certified seed. After discussions with the farmers you will learn the weaknesses of their approaches and their requirements and only then will you be able to design a trial that will convince them to change their technology to your optimised approach. Be careful though in your design that you are not unduly adventurous with the farmers’ time and resources. Spread treatments across farms If you would like to propose several sequence options then involve several neighbouring farms in the trial. If farms are very small, make a farm a treatment and get two farms to do that same treatment. It may only be a plot on each farm, say 2.8 m wide and 20 m long. Having treatments spread across farms generates a sense of competition between farmers and particularly competition between farmers with the same treatment. This leads to discussion and much faster communal learning than when farmers are isolated. The farms that are not interested to be involved can become the control farms if they grow continuous wheat. They will be prepared to tell you their methods and let you assess their yields. The problem with this approach is that it needs organization and very good communications with farmers and between farmers so they know what they are doing. However, quite junior researchers can often run these trials very well if they are committed and in some cases their inexperience can be beneficial as the farmers will not feel overwhelmed. On the contrary, farmers should not feel they are interacting with someone who is inadequate for the job.

Less dry areas Here your options are broader. However, from the outset make sure that collaborating farmers really understand the potential economic benefits of using the break crops and when and why the benefits might occur as well as the risks involved. Explain if necessary, that the break crops must grow well to generate income in year 1 and maximize benefits to following crops. They are not a simple fallow replacement. Your control treatment will remain continuous wheat but choose break crops that can be handled with current equipment; crops or approaches that will produce a positive effect on following crops;

crops that interest the farmer; and preferably crops that can be sold in the region without large costs for cartage. Propose an alternative cereal such as oats or triticale or barley, an oilseed like sunflower or canola, a grain legume (pea, lentil, chickpea, faba bean) or a fallow for inclusion. These will provide the most information on mechanisms of crop response that can be explored with the farmers as a joint learning exercise. For the break crop year of the sequence choose varieties of the break crops that are reputed to be well-adapted to environments like at the site. You will need to remind the farmers that legume crops must be inoculated with the correct Rhizobium strains to optimize N fixation and seeders must be used that handle the large legume seeds carefully during sowing. A poor start will lead to a poor crop. Furthermore, oilseed crops such as canola and sunflower, and legume crops such as lentils and chickpea may require sprays to control pests and weeds during establishment and may require windrowing. Discuss all the components of your agronomy package and again, make sure the farmers understand why they are doing things and their cost and benefits. Write everything down for your group and copy it for all involved farmers. This will include the names and addresses of all farms and farmers and the treatments they are attempting. As suggested for arid zones, try to spread treatments across farms so that farmers approach your improved technology as a community. If you can get enough farmers interested and committed to your trial for possibly three years, ask one batch of farmers to run the trials out of phase by one year. These may be farmers who are interested with the results of the first year but did not join the trials at the outset. This approach adds season effects to the data set you are collecting.

A field layout and general methods to use Strip plots Whether your trials are on one farm or involve many farms each with a treatment, use strip plots. Do not attempt complex research designs. Do not replicate plots in the research sense. You are attempting to demonstrate a direction of response to farmers, not an exact number. You want the farmers to see how the treatments rank. If you need statistics to show effects, then the chances are that farmers will not be interested to adopt the technology you are proposing.

A strip plot is one or two adjacent runs of a seed drill width (each approximately 2.8m wide) maybe 20 m long with the treatment applied. The strip plot(s) is always within the farmer’s normal crop which serves as an enclosing border. If there are several treatments they can be placed lengthwise as a continuing strip or as adjacent strips in the farmer’s crop. You will find examples of their use in the papers on optimising nitrogen, sowing, crop establishment and tillage options. Replications. Yes or No? There is no requirement for replications, but if you like replication for publication purposes you can nominally split the adjacent strips into three sections lengthwise (6 m each) and treat them as three blocks for harvesting purposes. It is easiest to harvest each whole strip in one run of your harvester, bag the grain then do the next strip and so on. But if you have to harvest by hand with a sickle, 2 m² replication plots within each strip become essential. How to do this is described in the chapter on crop establishment. Locations of treatments: an example The actual positions of the plots will depend on what you are trying to demonstrate. Let’s assume you are making a design for an arid area and demonstrating the advantages of zero tillage over multi-pass conventional tillage and in parallel the use of clean fallow over weedy fallow to harvest water for a following wheat crop. Because the results are dependent on water you

must avoid prejudicing the results by locating the comparison plots some in water-collecting depressions and some on water draining slopes. Make sure the perceived constraint is going to be similar across comparisons. Also as you are expecting differential water storage across treatments, the strip plots must be wide enough to avoid the results being dominated by movement of water between plots either because of root exploration (maybe 1 m laterally) or lateral drainage. In this example, this error could be more than 1 m on either side of the strip plot. In that case you would run two drill widths (2.8 x 2 =5.6 m) as your strip plot and run the harvester down the centre of this area for your sample. Alternatively, if no harvester is available, you could harvest 3 of the 2m² quadrat samples in the centre of each strip as mentioned above. You could locate the strips for a three year sequence as in the diagram below where each box represents a strip plot with its successive crops over the three years. If you have only two years, plant only the bottom rows of each tillage method and exclude the third year wheat crops. The top row of each tillage method is to enable you to duplicate treatments over two seasons thereby checking the interactions between the sequence and weather as in the case study. If you are in an area with higher rainfall and checking more complex sequences, use the same type of layout. Simply include the additional sequences as additional strip plots or allocate

Layout of strip plots looking at the impacts of conventional and zero tillage and clean and weedy fallow on wheat production in a three year crop sequence (the 3- crop sequence is listed in each box). If only two years are available include only the second row (strip) of each tillage method and exclude all third year wheat crops

them to other nearby farms. If including nearby farms for treatments, remember you always need to include there the control plots of continuous wheat. Having the same control at all farms allows you to scale the responses of all treatments at all farms into one matrix. Marking out Prior to the break crop season install permanent pegs near the trial plots that can be used to precisely relocate the plots from year to year. These can be placed on nearby fence-lines or buried in the soil at the edges of the plots below cultivation depth. Plots can be located square from a fence line using a tape measure for distance. Plot pegs can then be re-located from the permanent pegs on the fence without the need to leave them in the field in the off-season. Marking permits any crop residues that have been blown off plots or moved by animals over time to be relocated back onto plots before sowing the following crops in the sequence. It is also critical to control weeds during the off season in clean fallow plots as weeds can host diseases and alter the water and N profiles which were left by the previous crops.

Observations to make and data to collect After the wheat crop preceding the trial has been harvested and you and collaborators have some time, peg out the trial areas in the fields. Pre-sowing soil samples can also be used to assess disease inoculum using glasshouse bioassays, and to determine soil water and N profiles.

Information on how to measure soil water simply and assess water use efficiency is in the chapter “Constraints to cereal-based rainfed cropping”. Rainfall data will be required throughout to interpret plant responses and max/min temperatures are very useful to gauge the speed of the season and for linking crop performance to extreme temperature events. If such data are not available you may consider at least

installing a rain gauge in the trial area. Check in the chapter on variety x sowing date for guidelines on preparation of tables for recording rain and other data.

Data from first year plots and crops From the first year you will need data on final biomass and harvest yield and, if possible, residual levels in the soil of water and nutrients. The soil measurements can be taken following the harvest of the crops (post-harvest data) and immediately prior to sowing the following wheat crops (pre-sowing data). If only one soil measurement is possible the pre-sowing one is preferred as this describes the conditions for the forthcoming wheat crop. It will also show how much water and nitrogen has been harvested by a clean or weedy fallow. Water content and soil N can be measured gravimetrically from soil cores taken from the plots (see the chapter “Constraints to cereal-based rainfed cropping” for details of methods and calculations). Within a plot at least two cores are required to the previous crop’s estimated final root depth. To allow for the high variability of soil N in the surface layer, collect up to ten subsamples at random from the surface 0-10 cm layer and bulk them. Because you will be doing a cost:benefit analysis for the two or three year sequence, you will need to know the costs and values of any activities in the break crop year as well as the wheat year. So do not forget to include data for your legume or oilseed crop, the animal benefits from grazing the weedy fallow, the value of the potatoes grown on the plot and eaten on the farm and so on.

A rain gauge. The inner cylinder contains 26 mm when full then spills over and collects in the outer cylinder. The funnel/cap reduces evaporation

During the second year wheat crop The pre-sowing measurements of soil water and N you will already have taken will be essential for interpreting crop response during the season. Recommended wheat crop measurements during the season are plant numbers per m² at the two to three leaf stage and final biomass, yield and yield components at maturity (check out the methods using grab samples and quadrats described in the chapter “Optimizing plant population……”). Observations of disease development are also required at these times. Each root disease (e.g. Rhizoctonia, nematodes, Fusarium, Eyespot) has different symptoms and different systems for scoring severity. Check the literature. Be aware that some symptoms of root disease can be confused with those of drought. Disease development can be assessed using visual ratings of symptoms or by more quantitative assessments. For example, severity of take-all infection can be assessed either by scoring seminal root black lesions at the four leaf stage as in the accompanying picture, or by the proportion of ‘whiteheads’ to total spikes in the crop during the grain filling stage.

How to interpret the data Make an estimate of yield potential at the each site using the rainfall data collected there. Use the rough guide explained earlier in the chapter. For each mm of growing season rainfall above 100 mm expect 15 (to 20) kg grain ha-1. If treatments are yielding well below the calculation, discuss with collaborating farmers why this might be, particularly if it doesn’t occur on all farms. Decide what special thing might be limiting yield apart from rainfall. Now expand the session to decide exactly what the data say. Many of the things that should be

discussed will be obvious, like how did the break crops fare in their own right; which was the best break crop; were there any real problems associated with the break crops that could be overcome in any repeat study? Before getting into debating and explaining in detail why particular things happened, a good starting point will be discussion of the cost benefit analysis or gross margins analysis. In a nutshell, this will show whether the sequences were worthwhile and whether further optimization of the sequences is required. An example follows.

Working out gross margins The calculations in the following table use actual relative prices (with a nominal currency) and actual data collected from an Australian study like that described earlier. It considers three sequences, wheat-wheat, canola-wheat and peas-wheat, i.e. wheat in the second year of all sequences. Add your own data to a similar table but use local

An example of scoring disease. As infection increases from slight (score 1) to severe (score 3) fewer healthy roots (yellow) are produced and at the same time the proportion of black infected roots increases

Gross margins calculations For break crops in year 1, for the wheat crops of the following year (year 2) and for the combined crop sequences, wheat-wheat, canola-wheat and pea-wheat (last column) Year 2 examples are for a dry year, when no N was applied to the wheat crop, and for a wet year. In the wet year 50 kg N ha-1 was applied after wheat and canola, but not after peas. Price for wheat varies with protein content in this example taken from Australia

1 2 3 4(2x3) 5(4 -1) 6

crop cost$/ha

yieldt/ha

value$/t

value$/ha

margin$ yr 1

margin$ yrs 1+2

Year 1

wheat 200 4 220 880 680

canola 300 2.8 380 1064 764

pea 150 2 200 400 250

Year 2: eg of a dry year after year 1 (0kgN/ha)

wheat 200 2 220 440 240 920

wheat 200 2 220 440 240 1004

wheat 100 4 250 1000 900 1150

Year 2: eg of a wet year after year 1 (50kgN/ha)

wheat 250 2.2 220 484 234 914

wheat 250 5.6 250 1400 1150 1914

wheat 100 6 250 1500 1400 1650

currency and adjust costs and values of crops to reflect local markets. You might decide to rework the example with local prices The example table shows the second year either as a dry season or a wet season. Only the zero N treatment is presented for the dry season as it is often impractical to add N under dry conditions such as are common in North Africa. For the wet season, a 50 kg N ha-1 treatment is shown after wheat and after canola, as a top dressing would be appropriate in such a season. The added N cost is approximately US$50. It is unnecessary to add N after peas. Check the following conclusions using the data from the table and look for parallels and differences in your own data. • Look at data for year 1 crops in the column labelled 5. The gross margin for peas (green row) was poor (US$250), much less than half that for wheat (US$680, grey) and canola (US$764, yellow). Taken on that year alone, canola seemed to be a suitable economic alternative for wheat. • Look lower down the same column for year 2, for the dry season. These crops are all wheat but follow the break crops wheat, canola or peas. Wheat after peas at zero N does very well in economic terms (US$900 versus US$240 for wheat after wheat). Due to carry over of N from the pea crop it yields well (4 t/ha) and has a high value protein content (US$250 versus US$220 per tonne grain). Despite a poor economic performance in year 1, the two-year gross margin for the pea-wheat sequence in column 6 is good (US$1 150 versus US$920 for wheat-wheat) or 25 percent better than wheat: wheat. In your case, this benefit would be reduced if you receive no price advantage for higher protein content grain (12 percent better than wheat: wheat.) • When year 2 is wet and when 50 N is applied, wheat is susceptible to root disease. Under these conditions the effect of the foregoing ‘break’ crops canola and peas was to ‘break’ the disease cycle. Consequently, yield of wheat after canola or peas was almost three times that after wheat and gross margins over four times greater in that year. Considering gross margins for the two-year sequence, using break crops virtually doubled farm income.

How do you explain your results? There are many possible outcomes to your trials because of the potential interactions between break crops, diseases, soil nutrition and weather. You and the farmer could have difficulty explaining the results. To help you, several scenarios follow. Work through them with collaborating farmers and see what applies to your situation. They might explain why there were benefits from the sequences or why they failed. They will also help determine what further treatments or subtreatments might be required to optimize a crop sequence for the farm.

If break crops improve yield by reducing weeds you would expect: • To see more weeds in wheat after wheat than

wheat after break crops. Check the area occupied by weeds in each crop

• To see less weeds in the break crop itself or the clean fallow. Estimate the area of weeds to crop during the break year

• To measure higher levels of soil moisture at the start of the wheat season where a clean fallow was used as the break. Check the records

If there are high levels of weeds even after break crops, it may be because: • the weeds were not effectively controlled in

the break crops either by the crop itself or by chemical sprays;

• the break crop supported similar types of weeds to wheat

If the degree to which crops conserved soil water was important you would expect: • better growth linked to treatments with higher

levels of residual water; • this better growth to be more pronounced in

drier seasons. If break crops improve yield by reducing disease you would expect: • to see higher levels of disease in wheat after

wheat than wheat after break crops. Check records of disease ratings on the wheat crop;

• to see better growth and yield of wheat after any break crop compared with wheat;

Most outcomes from break crop trials can be explained in terms of levels of water, disease, or nitrogen.

• root diseases often lead to small, pinched and shrivelled grains so you would see more of this symptom if there is no break crop.

If there are high levels of disease even after break crops, it may be because: • the disease promulgates on the break crops as

well as on the wheat (e.g. Rhizoctonia solani); • the disease can last longer than one year in the

soil without a host (e.g. cereal cyst nematode); • the first year was very dry and the disease

inoculum was preserved; • grass weeds were present in the break crops or

fallow and these carried the disease over; • the disease was stubble-borne and spores blew

in from nearby fields. If there is no disease in the wheat-wheat crop sequence it may be because: • seasonal conditions did not allow expression

of the disease; it may have been too dry; • though there were high inoculum levels after

wheat in year 1, they declined in the off-season perhaps because of wet conditions;

• the site had low levels of disease initially; • the wheat variety chosen has some level of

tolerance to the disease. If soil N analyses showed N benefits following the break crop you would expect: • more tillers, higher yield and higher protein

after the legume than canola at 0 kg N ha-1; • higher N in the soil after the legume break

crop than after a non-legume break crop; • smaller differences in wheat yields after

legume and canola at 50 than at 0 kg N ha-1; • differences in the amount of N removed by

different crops in year 1. If there are no obvious benefits of N to growth such as no increase in tillering or biomass it may be because: • the site had high initial N. This reduced N

treatment effects and legume N fixation; • it was too dry for N to become a limitation to

growth as water was the dominant constraint; • weeds growing in the off season used up

available N and nullified differences in N between treatments;

• mineralization in the off-season negated differences which existed between treatments;

• the legume did not fix N due to poor nodulation and growth or high initial N levels.

If there are N benefits to growth but not to yield it may be because: • crops lodged after wet, windy conditions; • crops grew vigorously in response to N but

encountered water-stress late in the season causing haying-off (see “Optimizing N Use”.

For further reading Bouzza, A. 1990. Water conservation in wheat rotations under several managements and tillage systems in semi-arid areas. Ph.D. dissertation, University of Lincoln Nebraska, USA.

Clarkson, J.D.S. & Polley, R.W. 1981. Diagnosis, assessment, crop-loss appraisal and forecasting. In M.J.C. Asher & P.J. Shipton, eds. Biology and Control of Take-all, p. 251-269. Academic Press, London, UK.

EL Mejahed K. 1997. Rotation tillage and N fertilizer effects on wheat in semiarid Morocco. In The challenge of production system sustainability: Long term studies in agronomic research in dry areas. ICARDA, Aleppo, Syria.

Kacémi, M. 1992. Water conservation, crop rotations, and tillage systems in semiarid Morocco. Ph.D. dissertation, Colorado State University, Fort Collins, Colorado, USA.

Kirkegaard, J.A., Hocking, P.J., Angus, J.F., Howe, G.N. & Gardner, P.A. 1997. Comparison of canola, Indian mustard and linola in two contrasting environments II. Break crop and nitrogen effects on subsequent wheat yields. Field Crops Research 52, 179-96.

Kirkegaard, J.A 2002. What is the best cropping sequence for the farm? In Explore on Farm: Adapting and adopting good farming practices. FAO, Rome. Italy

Heenan, D.P. 1995. Effects of broad-leaf crops and their sowing time on subsequent wheat production. Field Crops Research 43, 19-29.

Mazhar, M. 1991. Influence des rotations culturales sur la production de blé et la conservation de l'eau (synthese de 6 années d'expérimentation). Mémoire présenté pour le grade ingenieur en chef. Centre Aridoculture, INRA, Settat, Morocco.

Peoples, M.B., Ladha, J.K. & Herridge, D.F. 1995. Enhancing legume N fixation through plant and soil management. Plant and Soil 174, 83-101.