Post on 04-Jun-2020
Data Normalized Difference Vegetation Index (NDVI) is a commonly used numerical indicator for the presence of live green vegetation. Higher values indicate higher amounts and productivity of vegetation. The study used NDVI data from AVHRR and MODIS, participation data from TRMM, land surface temperature data from MODIS, and El NINO index of Sea Surface Temperature NINO3 and NINO 3.4.
Study Area
a. The study area is comprised of the three easternmost provinces of
Mongolia: Khentii, Dorod, and Sukhbaatar, and the area borders China and Russia.
b. It contains a population of approximately 800,000 to 900,000 Mongolian Gazelle .
c. The steppe ecosystem is characterized by rolling hills, flat plains, and scattered ponds and springs. Regional vegetation is largely grasses and forbs, with some shrubs and few trees, all found within a predominantly sandy, loamy soil.
d. The climate is defined by long, cold winters and short, warm summers with most precipitation occurring in July and August (200-300 mm).
Introduction
The Mongolian Gazelle has experienced a population decline of as high 17 million and a range reduction of 665,000 km 2 from historic levels (declines from peak of 95% and 60%, respectively) due to pressures from poaching, mining, and expansion of livestock cultivation. Current conservation lands cover only a small portion of potential gazelle habitat, and landscape-level conservation strategies are needed to retain intact grasslands and promote protection of migrating gazelle. Objectives 1) Provide a comprehensive view of where and when forage
conditions are most suitable for Gazelle in the Eastern Mongolian Steppe.
2) Explore the local and global climate drivers of grassland dynamics.
Implications of Grassland Trends and Climate Linkages (1982-2011) for Mongolian Gazelle Habitat Conservation
Methodology
Results
Nick Cuba*, ncuba@clarku.edu David Eitelberg+, deitelberg@clarku.edu Qiqi Jiang+, qjiang@clarku.edu Mike Towle+, mtowle@clarku.edu * Graduate School of Geography, Clark University + GIS for Development and Environment, Dept. of IDCE, Clark University
References: Buermann, W., B. Anderson, C.J. Tucker, R.E. Dickinson, W. Lucht, C.S. Potter, R.B. Myneni. (2003). Interannual covariability in Northern Hemisphere air temperatures and greenness associated with El Nino-Southern Oscillation and the Arctic Oscillation. Journal of Geophysical Research. 108(D13), 4396 Wang, J., K.P. Price, P.M. Rich. (2001). Spatial patterns of NDVI in response to precipitation and temperature in the central Great Plains. International Journal of Remote Sensing. 22(18), 3827-3844 Milner-Gulland, E.J. & Lhagvasuren, B. (1998) Population dynamics of the Mongolian gazelle Procapra gutturosa: an historical analysis. Journal of Applied Ecology, 35, 240–251. Mueller, T., K.A. Olson, T.K. Fuller, G.B. Schaller, MG Murray, P. Leimgruber. (2007). In search of forage: predicting dynamic habitats of Mongolian gazelles using satellite-based estimates of vegetation productivity. Journal of Applied Ecology, 10.1111/j.1365-2664.2007.01371.x Yu. F., K.P. Price, J. Ellis, J.J. Feddema, P. Shi. (2004). Interannual variations of the grassland boundaries bordering the eastern edges of the Gobi Desert in central Asia. International Journal of Remote Sensing. 25(2), 327-346
Regression Analysis. Once per metric.
Extract mean R2 values for each habitat suitability class Independent:
Climate Metric*
Dependent: Grassland
R2 maps for study area
Select highest
performing R2 output
(mean)
Avg R2 -High Suitable area
Avg R2 -Mod(H) suitable area
Avg R2 -Mod(L) suitable area
Avg R2 - Low suitable area
NDVI (1982-2007)
Mean # Suitable Days/Year
Rate of change in # Suitable Days/Year
Standard deviation # Suitable Days/Year
Long term trends
Linear trends
Interannual variability
Threshold?
High/Low
High/Low
High/low
High/low
High/low
reclass
High suitable area
Mod (High) suitable area
Mod (Low) suitable area
Low suitable area
Regression Analysis. Once per metric.
Extract mean R2 values for each habitat suitability class
Climate Metric*
R2 maps for study area
Select highest
performing R2 output
(mean)
Avg R2 -High Suitable area
Avg R2 -Mod(H) suitable area
Avg R2 -Mod(L) suitable area
Avg R2 - Low suitable area
NINO index
Signif. +/Other # SD/y trend
High/Low SD of # SD/y High/Low # SD/y
LONG-TERM MEAN LINEAR TRENDS INTERANNUAL VARIABILITY
Mean # Suitable Days/Year (SD/y) 1982-2007
Rate of change in # SD/y St Dev. Of # SD/y
Gazelle Habitat Suitability
Habitat Suitability
Cat. Mean R2
High 0.55
Mod (H) 0.58
Mod (L) 0.48
Low 0.51
Suitability analysis Linear Regression
1. Local climate regression on grassland
NDVI & 2 months rolling precipitation
NDVI & 1 lag land surface temperature
Cat. Mean R2
High 0.7
Mod (H) 0.67
Mod (L) 0.58
Low 0.55
2. Teleconnections
Precipitation & NIÑO 3
NDVI
lag-1
Feb_2 March_1 March_2 April_1 April_2
Precip
Lag-5
Jan Feb March April May Jun July Aug
Cat. Mean R2
High 0.058
Mod (H) 0.061
Mod (L) 0.059
Low 0.058
“Suitable” Days were defined using Mueller et al. (2007) relative NDVI thresholds based on August map for each year 1982-2007: (31% and 74% of range of observed NDVI values. From this information we derived three metrics.
0
50
100
150
200
250
300
350
400
450
500
550
600
650
-0.4
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
1982
1987
1992
1997
2002
2007
# Suitab
le
Days
IDEAL THRESHOLD
RANGE
1. Suitable days
2. Suitability Variables
Conclusions
The overall Gazelle habitat suitability map produced in this project indicate that locations in the northwest and the east of the study area are the most suitable for Gazelle grazing. Protected areas in these locations are seen to provide highly suitable grazing conditions for Gazelle, and an effective future strategy may focus on expanding the borders of current protected areas, or on establishing new protected areas in close proximity.
The results of linear regression analysis measured the strength of the relationships between precipitation and temperature, and indicated the temporal relationships at which these relationships are strongest. The best-performing metrics: one month lagged surface temperature and the three-month rolling average of accumulated precipitation, are both seen to correlate more strongly with NDVI in the areas of highest suitability for Gazelle. This finding suggests that regional climate forecasts will have important implications for the success of Mongolian Gazelle conservation.
Although analysis of the linkages between ENSO and climate conditions in the study area failed to discover strong connections, further testing for relationships at additional time-steps (e.g. seasons) or over a longer time interval may show significant correlation between these variables.