Scholarlink

<p>The study of the soil moisture variation is crucial for effective water resource management, drought forecasting, sustainable agriculture, and climate change analysis. Agriculture is the basis of this unique research paper. This study examines how soil moisture varies with the seasons (January, March, May, and August) and years (from 2020 through 2023) in the Bhuj area of Gujarat, India. For the estimation of soil moisture with precision, an artificial neural network (ANN) model is developed by incorporating meteorological and spatial parameters identified through a detailed evaluation. The developed ANN model was calibrated (trained) and validated (tested for prediction of soil moisture) with SMAP Level-4 satellite data. Across all study periods, the model shows strong performance, with an average mean squared error of 8.84 × 10<sup>−7</sup>, mean absolute error of 0.00052, and <i>R</i><sup>2</sup> of 0.99. These metrics varied only slightly across seasons (mean squared error ± 1.45 × 10⁻⁶, mean absolute error ± 0.00029, <i>R</i><sup>2</sup> ± 6.66 × 10⁻<sup>5</sup>), indicating consistent reliability. The results demonstrate the effectiveness of combining remote sensing data with localized modeling for monitoring soil moisture in semi-arid areas, offering valuable insights for environmental and agricultural applications.</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

Soil Moisture Prediction Using Artificial Neural Networks and Remote Sensing Data

  • 期刊论文

  • 发表日期：

    2025年7月30日

查看全文
  • Aryan Singh,
  • Avdhesh Kumar,
  • Ankit Singh,
  • Rohit Kumar Tiwari,
  • Navin Chaurasiya,
  • Adarsh Singh,
  • Shashank Shekhar Singh,
  • Manish Pratap Singh

摘要

The study of the soil moisture variation is crucial for effective water resource management, drought forecasting, sustainable agriculture, and climate change analysis. Agriculture is the basis of this unique research paper. This study examines how soil moisture varies with the seasons (January, March, May, and August) and years (from 2020 through 2023) in the Bhuj area of Gujarat, India. For the estimation of soil moisture with precision, an artificial neural network (ANN) model is developed by incorporating meteorological and spatial parameters identified through a detailed evaluation. The developed ANN model was calibrated (trained) and validated (tested for prediction of soil moisture) with SMAP Level-4 satellite data. Across all study periods, the model shows strong performance, with an average mean squared error of 8.84 × 10−7, mean absolute error of 0.00052, and R2 of 0.99. These metrics varied only slightly across seasons (mean squared error ± 1.45 × 10⁻⁶, mean absolute error ± 0.00029, R2 ± 6.66 × 10⁻5), indicating consistent reliability. The results demonstrate the effectiveness of combining remote sensing data with localized modeling for monitoring soil moisture in semi-arid areas, offering valuable insights for environmental and agricultural applications.