NATIONAL LEVEL SPATIAL MODELING OF AGRICULTURAL PRODUCTIVITY: STUDY OF INDIAN AGROECOSYSTEM

摘要

Traditional decision support systems based on crop simulation models are normally site-specific. In order to address the effects of spatial variability from one place/region to other of soil conditions, and weather variables on crop production, spatial model namely "Spatial-EPIC" using Geographic Information System (GIS) was developed linking with biophysical agricultural management simulation models. With the development of this model any size of agroecosystem starting from a field to a country and even bigger can be modeled. A country level Indian agroecosystem was simulated as an application of model development and have been detailed with validation in this paper. It also helped to predict spatial yield variability on a farm level, region level, state level and so on as a function of soil water conditions under various weather regimes and management practices based on their socio-economic resources they prevail. GIS-based model differing in their resolutions (approx50 km grid size and approx10 km grid size) were applied to two level study respectively at whole India level and then one of the Indian province called Bihar. Results showed that at both resolution level crop yield varied significantly as a function of the data detailed due to their resolution (pixel sizes) as well as function of seasonal climatic variation, soil water holding characteristics and provided crop management time-series information.