Development of Digital Soil Theme Maps using Soil Health Cards by Geostatistical Methods

Abstract

Soil properties are continuously and spatially variable and this tendency is to be captured for their estimation. Conventional soil resource maps, being expensive, are prepared once in a while with little scope for revision. But that problem could be addressed by using the soil health cards (SHCs) issued to individual farmers by using appropriate geostatistical techniques in generating the updated soil theme maps. As required in the interpolation, soil health cards are the point sources of data including the geolocation of a field from where the soil sample was collected. Altogether 12966 SHCs of West Godavari district of Andhra Pradesh were interpolated for pH, electrical conductivity (EC), organic carbon (OC), available nitrogen (N), phosphorus (P), potassium (K), sulphur (S), iron (Fe), copper (Cu), zinc (Zn), manganese (Mn) and boron (B) using geostatistical tool, Ordinary Kriging (OK). Total of 9046 and 3920 SHCs were used for interpolation and validation, respectively. In another attempt to find utility of interpolated soil maps, initially, 101 rice fields distributed in different mandals (sub-district units) as identified from satellite data were randomly selected. These rice fields were represented as points using geographical information system (GIS) tools. Different soil variables from interpolated maps were extracted to these point vectors from interpolated soil maps along with mean mandal rice productivity as derived from crop cutting experiments for statistical analysis. Data were classified to four groups of N (<200, 201-300, 301-400 and >400 kg ha-1) and three groups of OC (<0.5, 0.5-0.75 and >0.75 %). General linear model with productivity as dependant variable, N and OC groups as fixed factors with EC, P, K, Zn and Fe as covariates yielded good results (R2= 59%) with significant contributions from N and OC groups, EC, P and Fe. These maps being in digital format, spatial analysis is possible in GIS environment when compared to other forms of soil maps with non-spatial data. These digital maps are the important inputs for predicting fertility in unsampled areas and help in decision making with a better resolution of maps.