Predicting cotton production using Infocrop-cotton simulation model, remote sensing and spatial agro-climatic data

• Published in: Current science (Bangalore) 2008-12, Vol.95 (11), p.1570-1579
• Main Authors: Hebbar, K. B., Venugopalan, M. V., Seshasai, M. V. R., Rao, K. V., Patil, B. C., Prakash, A. H., Kumar, V., Hebbar, K. R., Jeyakumar, P., Bandhopadhyay, K. K., Rao, M. R. K., Khadi, B. M., Aggarwal, P. K.
• Format: Article
• Language: English
• Subjects: Agricultural soils; Agrology; Cotton; Crop management; Crop production; Crops; Energy crops; Modeling; Simulations; Soil depth
• ISSN: 0011-3891

A methodology is described to predict cotton production on a regional basis using the integrated approach of remote sensing (RS), geographic information system (GIS) and a crop simulation model, i.e. Infocrop-cotton model. This model is based on an indigenous crop growth simulator called Infocrop. The Infocrop-cotton model was calibrated and validated to simulate the effect of diverse weather, soil, and agronomic management practices on growth, development and yield of cotton varieties and hybrids using results of several diverse field experiments (60 datasets). These experiments were conducted during 2000–01 to 2004–05 in major cotton-producing states of India across locations spreading from Hisar (29°10′N, 75°46′E) to Coimbatore (11°00′N, 77°00′E) with varying management practices, weather and soil. The model satisfactorily simulated the trends in leaf area, dry matter growth, days to flowering and seed cotton yield. The simulated time to flowering and maturity varied between 54 and 80 days as well as 136 to 193 days, with an RMSE value of 3 and 8.5 days respectively. Total biomass and seed cotton yield showed an accuracy of 86 and 89% respectively. The model also precisely simulated water deficit and N stress, the two important abiotic constraints for dryland cotton production. The Infocrop-cotton model was used in conjunction with RS and GIS techniques for developing an integrated approach for deriving cotton production estimates. Resourcesat-1 LISS III data of October/November months corresponding to peak vegetative stage of cotton crop were used to derive spatial distribution of cotton crop. The study area was classified as polythesian polygons based on pedo-climatic variables, namely soil type, soil depth and rainfall pattern using GIS. Cotton yields for each of these polygons were simulated using the crop model and were aggregated to determine the total production of the district. The prediction of cotton production was more accurate to the partially irrigated or irrigated districts and not for the rainfed districts. The utility of the integrated approach in prediction of cotton production at the regional level has been discussed.