Signal Processing-Based Assessment of Climate Variability and Agricultural Water Demand

Climate variability has emerged as a major challenge for sustainable agriculture by influencing irrigation water demand and food production through changes in rainfall, temperature, and evapotranspiration patterns. Conventional statistical methods often fail to capture the nonlinear and non-stationary characteristics of climatic time series, limiting accurate assessment of agricultural water requirements. This study proposes a signal processing-based framework to investigate the impact of varying climatic conditions on irrigation water demand and food availability. Historical meteorological data, including precipitation, air temperature, and reference evapotranspiration, are analyzed using discrete wavelet transform (DWT) to decompose climate signals into multiple temporal scales while reducing noise. Time-frequency features, including wavelet energy, spectral entropy, and power spectral density, are extracted to characterize seasonal and long-term climatic variability. These features are integrated with irrigation water requirement estimates and crop production records to evaluate the relationship between climate fluctuations, agricultural water demand, and food availability. The framework was evaluated over Khasi Hills, Meghalaya, India, using climate observations from 2000–2022, including rainfall, air temperature, and MODIS-derived evapotranspiration. Irrigation water demand was estimated using the FAO-56 Penman–Monteith method for rice and maize cultivation. The proposed framework enables improved identification of climate-induced anomalies that influence irrigation scheduling and crop productivity. Results show that the proposed framework explained 89% of the variability in irrigation water demand (R² = 0.89), reduced estimation RMSE by 17.8%, and identified dominant climatic fluctuations at different temporal scales. Wavelet-based features exhibited strong correlations (r = 0.82) with crop productivity and successfully detected drought periods associated with increased irrigation requirements. Unlike conventional statistical analyses, the proposed framework jointly integrates multiscale wavelet decomposition with irrigation-demand estimation and crop productivity assessment, enabling simultaneous characterization of climatic variability and agricultural water requirements. Although demonstrated for a single agricultural region and selected crops, the framework is readily extendable to other climatic zones and cropping systems.