Scholarlink

<p>Surface water salt intrusion affects livelihoods and the environment by degrading soil and water quality and reducing agriculture production. Remote sensing and machine learning are widely used to monitor water quality in coastal and inland river networks, but surface water salinity is little studied. This paper presents a novel approach for predicting surface water salinity using remote sensing data and machine learning algorithms. The input variables comprising hourly salinity monitoring data on surface water, coordinate and altitude data of salinity monitoring stations, and indicators extracted from Sentinel-1B were gathered, analyzed, and standardized to ensure their correlation and compatibility. Bayesian Model Averaging technique was used to assess and select the optimal input variables. Advanced machine learning algorithms were applied to build the surface water salinity predictive map. Various statistical indicators were utilized to evaluate and verify the accuracy and reliability of these predictive models. The findings showed that a hybrid model of Particle Swarm Optimization and Random Forest was the most accurate forecasting model, and it was selected to generate the surface water saline prediction map in this research. The river mouths with the greatest intrusion of surface water salinity are Vam Co, Cua Tieu, Cua Dai, Ba Lai, Ham Luong, and Co Chien rivers. The Hau River experienced lower surface water salinity incursion. The surface water salinity forecasting map provides crucial information for effective water resource management by enabling local authorities to implement targeted interventions, environment monitoring, and livelihood conservation.</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

Processing of Sentinel-1B imaginary to extract water surface salinity incursion using observed data and hybrid models of particle swarm optimization

  • 期刊论文

  • 发表日期：

    2025年1月14日

查看全文
  • Quynh Duy Bui,
  • Hang Ha,
  • Truong Xuan Tran,
  • Minh Tuan Vu,
  • Lam Phuong Nguyen,
  • Chinh Luu

摘要

Surface water salt intrusion affects livelihoods and the environment by degrading soil and water quality and reducing agriculture production. Remote sensing and machine learning are widely used to monitor water quality in coastal and inland river networks, but surface water salinity is little studied. This paper presents a novel approach for predicting surface water salinity using remote sensing data and machine learning algorithms. The input variables comprising hourly salinity monitoring data on surface water, coordinate and altitude data of salinity monitoring stations, and indicators extracted from Sentinel-1B were gathered, analyzed, and standardized to ensure their correlation and compatibility. Bayesian Model Averaging technique was used to assess and select the optimal input variables. Advanced machine learning algorithms were applied to build the surface water salinity predictive map. Various statistical indicators were utilized to evaluate and verify the accuracy and reliability of these predictive models. The findings showed that a hybrid model of Particle Swarm Optimization and Random Forest was the most accurate forecasting model, and it was selected to generate the surface water saline prediction map in this research. The river mouths with the greatest intrusion of surface water salinity are Vam Co, Cua Tieu, Cua Dai, Ba Lai, Ham Luong, and Co Chien rivers. The Hau River experienced lower surface water salinity incursion. The surface water salinity forecasting map provides crucial information for effective water resource management by enabling local authorities to implement targeted interventions, environment monitoring, and livelihood conservation.