Published Paper Details:

Climate variability and inefficient irrigation contribute to water stress and reduced crop productivity worldwide. This paper presents a Smart Irrigation Advisory System that integrates a hybrid rainfall forecasting model with an IoT-based soil moisture monitoring network to provide timely, actionable irrigation recommendations. The forecasting module fuses statistical and machine-learning (deep learning) approaches to improve short- and medium-term rainfall prediction. The sensing module comprises distributed low-cost soil moisture sensors (multi-depth), connected via low-power wireless links to a cloud backend. A decision-support engine fuses forecasted rainfall, real-time soil moisture, crop-specific crop water requirements and evapotranspiration estimates to produce irrigation advisories delivered through a mobile/web interface. A pilot deployment in a semi-arid region evaluated the system for one crop season: the hybrid forecasting reduced rainfall prediction MAE by ~14% versus single-model baselines, and advisory-driven irrigation scheduling reduced water usage by ~22% while maintaining yield within 3% of conventional scheduling. The results demonstrate that combining hybrid forecasting and IoT sensing enables more efficient, climate-resilient irrigation.