Scholarlink

<p>This study evaluates the effectiveness of artificial intelligence (AI) applications, particularly Multilayer Perceptron Neural Networks (MLP-NN) and Gridded Precipitation Products (GPPs), in predicting the Drought Deciles Index (DDeI) across Iraq’s diverse climatic zones. Using 47 years of precipitation data from 22 ground stations (GS) and two GPP datasets, APHRODITE (Asian-Precipitation Highly-Resolved Observational Data Integration Towards Evaluation of Water Resources) and TRMM (Tropical Rainfall Measuring Mission), the study developed and validated MLP-NN models for drought monitoring. The novelty lies in integrating multiple datasets and leveraging advanced machine learning techniques for precise drought predictions across arid and semi-arid regions. The MLP-NN models were calibrated using precipitation data from GS and validated using GPP datasets for three climatic zones. APHRODITE demonstrated superior performance with an R<sup>2</sup> value of 0.922, particularly in arid regions, while TRMM performed well in areas with higher precipitation, achieving an R<sup>2</sup> of 0.895 and a Nash–Sutcliffe Efficiency of 0.96 in Zone 1. However, TRMM showed limitations in Zone 3, characterized by lower precipitation, where its R<sup>2</sup> dropped to 0.63. Overall, the MLP-NN-DDeI models successfully predicted drought indices, indicating them as reliable tools for monitoring and managing drought conditions. The model can be used for accurate and comprehensive drought predictions in Iraq and similar arid regions, supporting sustainable water resource management and policy planning.</p>

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

Integrating gridded precipitation data and machine learning for enhancing drought prediction in Iraq

  • 期刊论文

  • 发表日期：

    2025年4月2日

查看全文
  • Ali H. Ahmed Suliman,
  • Taymoor A. Awchi,
  • Shamsuddin Shahid

摘要

This study evaluates the effectiveness of artificial intelligence (AI) applications, particularly Multilayer Perceptron Neural Networks (MLP-NN) and Gridded Precipitation Products (GPPs), in predicting the Drought Deciles Index (DDeI) across Iraq’s diverse climatic zones. Using 47 years of precipitation data from 22 ground stations (GS) and two GPP datasets, APHRODITE (Asian-Precipitation Highly-Resolved Observational Data Integration Towards Evaluation of Water Resources) and TRMM (Tropical Rainfall Measuring Mission), the study developed and validated MLP-NN models for drought monitoring. The novelty lies in integrating multiple datasets and leveraging advanced machine learning techniques for precise drought predictions across arid and semi-arid regions. The MLP-NN models were calibrated using precipitation data from GS and validated using GPP datasets for three climatic zones. APHRODITE demonstrated superior performance with an R2 value of 0.922, particularly in arid regions, while TRMM performed well in areas with higher precipitation, achieving an R2 of 0.895 and a Nash–Sutcliffe Efficiency of 0.96 in Zone 1. However, TRMM showed limitations in Zone 3, characterized by lower precipitation, where its R2 dropped to 0.63. Overall, the MLP-NN-DDeI models successfully predicted drought indices, indicating them as reliable tools for monitoring and managing drought conditions. The model can be used for accurate and comprehensive drought predictions in Iraq and similar arid regions, supporting sustainable water resource management and policy planning.