Evaluation of stratification criteria for regional scale assessment of soil fertility parameters in the semi-arid

Abstract

Statistically sound soil sampling strategy is a prerequisite for assessing the soil fertility status of a region. We tested a stratified, random sampling methodology to ascertain whether surface geology, microwatershed, position on the toposequence, and size of farm holding explained the differences in the fertility parameters of soils in the semi-arid tropical region of Karnataka State, India. Descriptive statistics revealed that out of 119 farmer fields sampled, many were deficient in the following nutrients: 50% in available phosphorous (P), 78% in boron (B), 67% in zinc (Zn), and 72% in sulfur (S). Discriminant analysis indicated that fertility studies on a regional scale need to include good representation of soil heterogeneity based on parent material (surface geology), hydrological properties, and positions on toposequence but not the size of the farm holding. The proposed stratified random sampling technique enabled representation of the heterogeneity in the whole population even with a small sample set size.