Modelling Infiltration and quantifying Spatial Soil Variability in a Wasteland of Kharagpur, India

Infiltration is vital for both irrigated and rainfed agriculture. The knowledge of infiltration characteristics of a soil is the basic information required for designing an efficient irrigation system. The most difficult task in the field, however, is to determine the spatial variability of infiltration process, which is of great importance for precision agriculture. The objective of the present study was to model infiltration and analyse the spatial variability of infiltration characteristics in a wasteland of Kharagpur, West Bengal, India. A total of 24 infiltration tests were conducted on a systematic squared grid pattern over the study area using double-ring infiltrometers. The observed infiltration data from all the test sites were fitted to four selected infiltration models and a best-fit model for individual sites was identified. Infiltration characteristics of the area generally indicated low basic infiltration rates (0·1–1·1 cm h−1) for most sites. The Philip two-term model was found to be the best model for describing the infiltration process at most test sites. The parameters of this model (i.e. sorptivity S and transmissivity factor A) showed a wide variation among the test sites. The sorptivity-based scaling factor αS and the transmissivity-based scaling factor αA were computed and the observed infiltration data were scaled. Scaling achieved through αA was found better than through αS. Optimum scaling factors αopt were then obtained by the least-squares method and the scaling of infiltration data were repeated. Scaling factors based on the arithmetic, geometric and harmonic means of αS and αA were also computed for scaling. It was found that αopt and the scaling factor based on harmonic mean of αS and αA scaled the infiltration data more effectively (lowest sum of squares error) than other scaling factors. Moreover, the best-fit infiltration model parameters (S and A), and all the scaling factors were subjected to frequency analysis and a best-fit statistical distribution was identified. Based on the results of this study, it is concluded that the study area has wide spatial variability.