Scholarlink

<p>Regional modeling of groundwater storage dynamic requires extensive parameterization and conceptualization, often relying on in-situ data collected from monitoring networks. However, the sparse spatial coverage and uneven distribution of these networks can limit the representativeness of such models. The use of remote sensing has greatly enhanced groundwater modeling by providing consistent and spatially comprehensive datasets. Satellite-based Earth observation systems, such as the Gravity Recovery and Climate Experiment (GRACE) and its follow-on mission (GRACE-FO), have significantly improved the ability to monitor water resources across broad spatial and temporal domains, thereby overcoming many of the limitations of traditional monitoring. In this study, groundwater storage (GWS) changes across Hungary, which is underlain predominantly by transboundary aquifers, were quantified by combining GRACE-derived Terrestrial Water Storage (TWS) anomalies with land surface components from the Global Land Data Assimilation System (GLDAS). To address data gaps in the GRACE time series, a Random Forest (RF) machine learning model was applied to impute missing TWS values. Subsequently, GWS anomalies were estimated by subtracting the GLDAS-based surface water components from the completed TWS records. Future GWS dynamics were further forecasted using a bootstrapped RF ensemble, which produced both point predictions and 95% confidence intervals. A Seasonal-Trend decomposition (STL) analysis indicated continued groundwater depletion in Hungary (-0.0375 to -0.0516&#xa0;mm/year) with the long-term trend accounted for the majority of GWS variability, while seasonal fluctuations contributed approximately 10–20% of the total variance. The study concluded that under current climatic and usage patterns, several aquifers are at risk of reaching critical depletion levels. This underscores the urgent need for region-specific groundwater management strategies aimed at regulating extraction and promoting sustainable use to safeguard future water security.</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

Modeling groundwater depletion in Hungary through GRACE and GLDAS observations analyzed with ensemble machine learning models

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月29日

查看全文
  • Musaab A. A. Mohammed,
  • Norbert P. Szabó,
  • Joseph O. Alao,
  • Péter Szűcs

摘要

Regional modeling of groundwater storage dynamic requires extensive parameterization and conceptualization, often relying on in-situ data collected from monitoring networks. However, the sparse spatial coverage and uneven distribution of these networks can limit the representativeness of such models. The use of remote sensing has greatly enhanced groundwater modeling by providing consistent and spatially comprehensive datasets. Satellite-based Earth observation systems, such as the Gravity Recovery and Climate Experiment (GRACE) and its follow-on mission (GRACE-FO), have significantly improved the ability to monitor water resources across broad spatial and temporal domains, thereby overcoming many of the limitations of traditional monitoring. In this study, groundwater storage (GWS) changes across Hungary, which is underlain predominantly by transboundary aquifers, were quantified by combining GRACE-derived Terrestrial Water Storage (TWS) anomalies with land surface components from the Global Land Data Assimilation System (GLDAS). To address data gaps in the GRACE time series, a Random Forest (RF) machine learning model was applied to impute missing TWS values. Subsequently, GWS anomalies were estimated by subtracting the GLDAS-based surface water components from the completed TWS records. Future GWS dynamics were further forecasted using a bootstrapped RF ensemble, which produced both point predictions and 95% confidence intervals. A Seasonal-Trend decomposition (STL) analysis indicated continued groundwater depletion in Hungary (-0.0375 to -0.0516 mm/year) with the long-term trend accounted for the majority of GWS variability, while seasonal fluctuations contributed approximately 10–20% of the total variance. The study concluded that under current climatic and usage patterns, several aquifers are at risk of reaching critical depletion levels. This underscores the urgent need for region-specific groundwater management strategies aimed at regulating extraction and promoting sustainable use to safeguard future water security.