Mapping risk of soil phosphorus deficiency using geostatistical approach: A case study of Brahmaputra plains, Assam, India

Analysis and interpolation of spatial variability of soil properties is very important for site specific management. The objective of this study was to determine the spatial distribution and degree of risk of phosphorus (P) deficiency in Goalpara district, Assam using statistics and geostatistics. A total number of 1397 soil samples were collected from 1953 km² area using 1 km×1 km grid at 0–25 cm depth and analyzed for Bray-1 P. Data set was positively skewed, therefore log-transformation was applied in order to achieve normality in the data set. Geostatistical analyses were carried out, including experimental variogram and model fitting and cross-validation. The ordinary kriging estimate map of P showed that the availability of P for plants was low in 78.6% area of the district. Another 6.0% area was ranked medium, while only 3.4% area was mapped with the high in available P. The probability map produced based on indicator kriging showed that the probability of deficiency of available P in the higher class [0.9 -1.0] is 53%, whereas, probability greater than 0.5 accounted 99% of the total area of the district. This means that there is chance of 99% area of the district showing P deficiency in crops. The standard deviation (SD) map generally showing the SD values are high for the points having the true values high. Thus, the study will help in to minimize both yield loss and environmental threats such as eutrophication due to under or over dose of P fertilizer application.