Agroclimatic analysis for mainland East Asia by a GIS approach

Abstract

Climate has been long recognised as an important constraint to crop production. Many agroclimatic analyses have been developed in Mainland East Asian countries to assist their agricultural development and resource management. These analyses were all restricted to using a limited number of data points and static summations of climatic variables. The complex climatic patterns and the non-linear responses of crops to climate cannot be captured by such agroclimatic analyses. This thesis addressed the agroclimatic environment and its impact on crop production using a different philosophy and methodology in order to overcome these shortcomings. There are four related components in this study. They are: 1). regular grid climatic data sets; 2). crop responses to the environmental elements in Mainland East Asia; 3). agroclimatic classification; and 4). crop modeling at selected representative stations from various agroclimatic zones. The regular grid climatic data sets consist of climatic surfaces and a digital elevation model (DEM). In this study they were developed at a resolution of l/20th degree. While this study focused on agroclimatic analysis these data sets can be applied to any other fields that relate to climate such as forestry, ecology and conservation. These climatic surfaces express climatic variables as functions of multi-dimensional thin plate smoothing splines in term of longitude, latitude and elevation. They were developed using the ANUSPLIN package, and are based on a network of up to 3800 stations across Mainland East Asia. Estimates for climatic variables at any location in the Mainland East Asian countries can be calculated from these surfaces with input of the appropriate independent variables. A DEM at a resolution of l/20th degree, calculated using the ANUDEM package and based on terrain data digitised from topographic maps, was used to construct data sets in this study. These data sets consist of 434,484 grid cells across the studying area. Based on such data sets, crop responses to the climatic environment were simulated using a general plant growth model GROWEST. This model transforms the non-linear responses of key plant groups to linear dimensionless scalars. These include a light index (LI), a thermal index (Tl), a moisture index (MI) and an integrated multi-factor growth index (GI). The spatial and seasonal variations of these indices were analysed for each of the 434,484 grid cells across Mainland East Asia. With 39 selected GROWEST attributes, Mainland East Asia was classified into 66 groups and further aggregated to 14 agroclimatic zones using the ALOC and FUSE modules of a numerical taxonomic package PATN. These agroclimatic zones have been given descriptive labels, thus; 1.Cold high plateau zone 2.Hot dry desert 3.Grassland zone 4.Single-crop 5.Double crop/wheat and 6.Double crop/rice 7.Warm hills 8.Warm highlands 9.Tropical mountain tops 10.Tropical forest 11.Triple-crop 12.Humid tropical lowlands 13.Perhumid tropical highlands 14.Perhumid tropical lowlands Finally, 14 representative stations were selected from the major cropping zones of Mainland East Asia for more detailed crop modeling using the DSSAT v3 package. The Seasonal Analysis module was used to model wheat, maize and rice production for for a period of 15 years, and has further demonstrated the major climatic constraints on crop production for various agroclimatic zones.