Smart Irrigation through IoT and Sensor Fusion: A Multiscale Review for Precision Agriculture

Agriculture accounts for approximately 70% of global freshwater withdrawals, making efficient irrigation management critical for sustainable food production. Smart irrigation systems based on the Internet of Things (IoT) and sensor fusion have emerged as promising tools; however, prior reviews often emphasize single sensing scales or specific technologies and rarely provide task-aware, cross-scale performance synthesis with implementation-oriented guidance. This article presents a systematic recent (2020–2025) multiscale synthesis of smart irrigation monitoring across soil, plant, canopy, and atmospheric drivers, reporting quantitative evidence from 41 empirical studies spanning diverse crops and climates and contextual insights from 7 systematic reviews (48 records in total). To ensure comparability, we separate regression, classification/detection, and early-warning tasks and report their metrics accordingly. Across comparable regression settings, cross-scale fusion is associated with consistent performance gains over single-source monitoring, typically reducing prediction error (RMSE/MAE) by 15–35% and improving model fit (e.g., Δ R 2 in the range of +0.10 to +0.25, with larger gains reported in specific downscaling cases). For classification/detection tasks, studies report improvements in discrete-task metrics under study-specific definitions. Several studies also report earlier stress detection, with lead time ranging up to 84 h relative to the baseline defined in each study.