@COMMENT This file was generated by bib2html_grace.pl <https://sourceforge.net/projects/bib2html/> version 0.94
@COMMENT written by Patrick Riley <https://sourceforge.net/users/patstg/>
@COMMENT This file was prepared using the NASA Astrophysics Data System (ADS)
@COMMENT https://ui.adsabs.harvard.edu/
@ARTICLE{2026AdSpR..78.4529F,
       author = {{Foroodi}, Zahra and {Amerian}, Yazdan and {Mahbuby}, Hany},
        title = "{Weighted downscaling of groundwater storage anomalies using random forests in Tehran region}",
      journal = {Advances in Space Research},
     keywords = {Groundwater storage, GRACE, Downscaling, Random forests},
         year = 2026,
        month = sep,
       volume = {78},
       number = {5},
        pages = {4529-4547},
     abstract = "{Gravity Recovery and Climate Experiment (GRACE) data have been
        extensively used to monitor terrestrial water storage (TWS) and
        groundwater storage (GWS). Estimating accessible groundwater is
        a crucial for managing water consumption in arid and semi-arid
        regions such as Tehran, while the coarse resolution of GRACE
        satellites poses a limitation for estimating groundwater storage
        anomalies (GWSA) at regional scale. Therefore, downscaling GWSA
        is of significant importance. In this study, we estimated three
        parameters{\textemdash}time lag, scale, and
        shift{\textemdash}for each observation well and converted
        groundwater level anomalies (GWLA) time series into GWSA time
        series. We propose a novel weighted Random Forest (RF)-based
        downscaling method that leverages an autoregressive (AR) model.
        Since the decline in GWSA in Tehran, the capital of Iran, has
        intensified over time, temporal factors such as year exert a
        stronger influence than climatic variations such as reduced
        precipitation, with precipitation ranking second in importance.
        Therefore, both temporal and climate-related features were
        simultaneously incorporated to effectively train the Random
        Forest (RF) model. The RF model was trained in both weighted and
        unweighted forms. In the unweighted case, the root mean square
        error (RMSE) and R$^{2}$ were 2.4 cm and 0.94, respectively,
        whereas in the weighted-case, they were 2.18 cm and 0.95,
        respectively. The results demonstrate the high performance and
        accuracy of the proposed method. Additionally, we estimated the
        effective resolution using the variogram method and the mean
        distance between wells, achieving an improved resolution of the
        downscaled GWSA to 0.05{\textdegree}.}",
          doi = {10.1016/j.asr.2026.06.077},
       adsurl = {https://ui.adsabs.harvard.edu/abs/2026AdSpR..78.4529F},
      adsnote = {Provided by the SAO/NASA Astrophysics Data System}
}