Vol. 12, Special Issue 10 (2023)

Author(s):
Fanesh Kumar, Jitendra Sinha, Vivek Kumar Tripathi and Nilima Jangre

Abstract:
Agriculture is the largest consumer of water through evapotranspiration and percolation. Irrigated agriculture remains the largest user of water globally, which may increase the risk of water scarcity for agricultural sector in the future. So, there is an immense need for enhancing per drop more crop. The present study focuses on assessment of rice water productivity for Pindrawan command area using ANN model and GIS. 20 years meteorological data and crop coefficient (Kc) of rice at different crop growth stages were analyzed to estimate crop evapotranspiration (ETc). Production and productivity data of the study area were obtained from the Agriculture Department Government of Chhattisgarh. The values of Crop Water Productivity (CWP) using ANN models (A1, A2 and A3) ranged A1(0.583- 0.615), A2 (0.582-0.614) and A3 (0.573-0.604) kg-m-3, while Field Water Productivity (FWP) ranged A1(0.329-0.347), A2(0.329-0.347), A3(0.326-0.344) kg-m-3 in study area. The values of CWP as estimated by CROPWAT and ANN models were found to be quite close. However, the ANN models estimated CWP little higher side (3.21%). Three different ANN models have been integrated in this study, depending on the combination of inputs to be given to the network. The first model (A1PC) has six inputs, the second model (A2PC) only has 4 inputs and third model (A3PC) only has 2 inputs. Performance evaluation of the models has been carried out by calculating the mean absolute deviation (MAD), root mean square error (RMSE), Absolute Prediction Error (APE), coefficient of correlation (CC), Nash-Sutcliffe coefficient efficiency (CE), and Index of Agreement (IOA).