Scholarlink

Irrigation faces challenges like water scarcity due to climate change, population surge, and droughts. Precision agriculture (PA) proposes a solution, using evapotranspiration (ET) to determine irrigation demand. ML and DL models are crucial for ET forecasting due to unpredictable weather. Prior research used ML models (LR, MLP, SVR) with observed and forecasted data, while another used TCN and SARIMA models relying on historical ET data. This study compares ML models (LR, MLP, SVR) trained with weather forecasts and observed ET to TCN and SARIMA models using only past ET data. Results highlight TCN’s prowess, suitable for a smart irrigation system. A comparison reveals ET-focused irrigation’s 19.93% water savings, endorsing ML’s role in resource optimization. Soil moisture data suggests refining ML/DL models for precise local water management.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Irrigation System Based on Evapotranspiration Forecasting Using Machine Learning

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ian Poveda,
  • Diego Fuentealba,
  • Nicolás Ruminot-Ahumada,
  • Samuel Montejo-Sánchez

摘要

Irrigation faces challenges like water scarcity due to climate change, population surge, and droughts. Precision agriculture (PA) proposes a solution, using evapotranspiration (ET) to determine irrigation demand. ML and DL models are crucial for ET forecasting due to unpredictable weather. Prior research used ML models (LR, MLP, SVR) with observed and forecasted data, while another used TCN and SARIMA models relying on historical ET data. This study compares ML models (LR, MLP, SVR) trained with weather forecasts and observed ET to TCN and SARIMA models using only past ET data. Results highlight TCN’s prowess, suitable for a smart irrigation system. A comparison reveals ET-focused irrigation’s 19.93% water savings, endorsing ML’s role in resource optimization. Soil moisture data suggests refining ML/DL models for precise local water management.