Design and Development of a Machine Learning and IoT-Based Decision Support System Smart Tool for Climate-Smart Irrigation Water Management

Efficient irrigation is critical for sustaining wheat production in semi-arid lowlands where water scarcity limits yield and water productivity (WP). This study develops an integrated framework combining machine learning (ML), IoT-based soil moisture monitoring, and a decision support system (DSS) tool to model wheat yield and water productivity for two lowland-adapted wheat varieties, Ogolcho (OWV), under different irrigation scheduling scenarios. Seasonal climate variables, soil physical properties, crop growth stage information, and real-time soil moisture data were used to train and evaluate seven ML models for yield and WP prediction within an offline DSS environment. Among the tested models, Random Forest, Gradient Boosting, and Support Vector Regression demonstrated superior performance, achieving high predictive accuracy (R² = 0.95–0.99; RMSE ≤ 0.06). The results indicate that OWV consistently outperformed KWV in both yield and water productivity across irrigation regimes. Optimal performance was observed under 100% and 125% of crop evapotranspiration (ETc), followed by 75% ETc,while deficit irrigation (50% ETc) and excessive irrigation (150% ETc) reduced productivity due to water stress and water surplus effects, respectively. The developed offline DSS prototype integrates ML-based prediction with field-level climate and soil data to provide real-time irrigation scheduling recommendations and productivity estimates without requiring internet connectivity. This enhances its applicability in resource-limited and remote farming areas. The combined ML and IoT-driven DSS framework offers a scalable, data-driven approach for improving irrigation efficiency and supporting climate-resilient wheat production in water-limited environments. Furthermore, in addition to the English language version, it gives a service by local language versions like Amharic (በአማረኛ) and expands other languages like Afan Oromo (አፋን ኦሮሞ፤ ትግረኛ......).