Author(s): Palash Dandotia; K. S. Prof. Hariprasad

Linked Author(s): PALASH KRISHNA DANDOTIA

Keywords: Precision irrigation; Irrigation scheduling; Crop water stress index; Artificial neural network; Deep Learning; Climate Change

Abstract: Water scarcity is a critical concern confronting sustainable wheat agriculture, particularly in arid and semi-arid regions. Precision irrigation presents a promising solution to conserve water while fostering agricultural sustainability. As a globally cultivated staple crop, wheat demands substantial water resources, necessitating efficient water management practices. The Crop Water Stress Index (CWSI) serves as a valuable method for determining wheat's irrigation requirements. CWSI provides a quantitative measure of the level of water stress experienced by a crop based on various physiological and environmental factors. This study evaluates the application of machine learning techniques, specifically Feed Forward Back Propagation Artificial Neural Network (FF-BP-ANN), to compute CWSI in wheat crops under varying irrigation levels. FF-BP-ANN is a multilayer perception model with an error backpropagation technique, designed to capture nonlinear relationships between input and output variables. Meteorological variables such as relative humidity, air temperature, and canopy temperature were utilized to calculate CWSI. The FF-BP-ANN model iteratively adjusts weights based on input data to minimize errors, thereby learning the underlying relationships and accurately predicting CWSI. The data is being further used for stimulating the effect of change in air temperature and relative humidity with its effect on CWSI. Thus, we may be able to find suitable shift in time for crop plantation and change in yield in accordance to global warming using Intergovernmental Panel on Climate Change (IPCC), Coupled Model Intercomparison Project CMIP6 dataset projected for 2050.

DOI: https://doi.org/10.64697/978-90-835589-7-4_41WC-P2061-cd

Year: 2025