A non-stationary downscaling and gap-filling approach for GRACE/GRACE-FO data under climatic and anthropogenic influences

Abstract Gravity Recovery and Climate Experiment (GRACE) and GRACE Follow-On (GRACE-FO) are being increasingly used as valuable data sources for hydrological monitoring. However, their coarse spatial resolution is considered as a limitation for regional studies, especially in areas with remarkable h...

Full description

Saved in:
Bibliographic Details
Main Authors: Seyed Mojtaba Mousavimehr, Mohammad Reza Kavianpour
Format: Article
Language:English
Published: SpringerOpen 2025-04-01
Series:Applied Water Science
Subjects:
Online Access:https://doi.org/10.1007/s13201-025-02427-z
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849730992516890624
author Seyed Mojtaba Mousavimehr
Mohammad Reza Kavianpour
author_facet Seyed Mojtaba Mousavimehr
Mohammad Reza Kavianpour
author_sort Seyed Mojtaba Mousavimehr
collection DOAJ
description Abstract Gravity Recovery and Climate Experiment (GRACE) and GRACE Follow-On (GRACE-FO) are being increasingly used as valuable data sources for hydrological monitoring. However, their coarse spatial resolution is considered as a limitation for regional studies, especially in areas with remarkable hydroclimate variability. In this study, a novel approach is presented for downscaling, and gap filling of terrestrial water storage (TWS) in Tehran province, Iran. Non-stationarity in the GRACE/GRACE-FO derived TWS is a significant challenge for predictive models. In this regard, the Hodrick–Prescott filter was adopted to detrend the TWS data. Afterward, several machine learning and deep learning techniques are employed for TWS prediction using Global Land Data Assimilation System and the fifth-generation ECMWF reanalysis (ERA5) datasets. The methodology is employed for bridging the gap between GRACE and GRACE-FO as well. Subsequently, the models are trained with different combinations of input variables and their performance is evaluated against the actual values. In parallel, a separate regression model based on the temporal index of the sample is developed for trend estimation and highlighting the role of anthropogenic activities. The proposed methodology is employed for bridging the gap between GRACE and GRACE-FO as well. The models with the highest accuracy are fed by input data with a spatial resolution of 0.25° × 0.25° to obtain fine-resolution TWS. Finally, the downscaled TWS derived from the predictive model is applied to calculate groundwater storage (GWS). The monthly TWS prediction results exhibit a strong correlation (CC = 0.93) and a low error (RMSE = 4.75 cm), underscoring the effectiveness of the proposed approach. TWS and GWS computations reveal rapid declines in groundwater-level prevailing by anthropogenic factors which exacerbate water crisis issues and environmental problems in the study area.
format Article
id doaj-art-6f662ad77fde4964af56eccdabff0b9f
institution DOAJ
issn 2190-5487
2190-5495
language English
publishDate 2025-04-01
publisher SpringerOpen
record_format Article
series Applied Water Science
spelling doaj-art-6f662ad77fde4964af56eccdabff0b9f2025-08-20T03:08:43ZengSpringerOpenApplied Water Science2190-54872190-54952025-04-0115511510.1007/s13201-025-02427-zA non-stationary downscaling and gap-filling approach for GRACE/GRACE-FO data under climatic and anthropogenic influencesSeyed Mojtaba Mousavimehr0Mohammad Reza Kavianpour1Faculty of Civil Engineering, K.N. Toosi University of TechnologyFaculty of Civil Engineering, K.N. Toosi University of TechnologyAbstract Gravity Recovery and Climate Experiment (GRACE) and GRACE Follow-On (GRACE-FO) are being increasingly used as valuable data sources for hydrological monitoring. However, their coarse spatial resolution is considered as a limitation for regional studies, especially in areas with remarkable hydroclimate variability. In this study, a novel approach is presented for downscaling, and gap filling of terrestrial water storage (TWS) in Tehran province, Iran. Non-stationarity in the GRACE/GRACE-FO derived TWS is a significant challenge for predictive models. In this regard, the Hodrick–Prescott filter was adopted to detrend the TWS data. Afterward, several machine learning and deep learning techniques are employed for TWS prediction using Global Land Data Assimilation System and the fifth-generation ECMWF reanalysis (ERA5) datasets. The methodology is employed for bridging the gap between GRACE and GRACE-FO as well. Subsequently, the models are trained with different combinations of input variables and their performance is evaluated against the actual values. In parallel, a separate regression model based on the temporal index of the sample is developed for trend estimation and highlighting the role of anthropogenic activities. The proposed methodology is employed for bridging the gap between GRACE and GRACE-FO as well. The models with the highest accuracy are fed by input data with a spatial resolution of 0.25° × 0.25° to obtain fine-resolution TWS. Finally, the downscaled TWS derived from the predictive model is applied to calculate groundwater storage (GWS). The monthly TWS prediction results exhibit a strong correlation (CC = 0.93) and a low error (RMSE = 4.75 cm), underscoring the effectiveness of the proposed approach. TWS and GWS computations reveal rapid declines in groundwater-level prevailing by anthropogenic factors which exacerbate water crisis issues and environmental problems in the study area.https://doi.org/10.1007/s13201-025-02427-zGroundwater levelGRACEMachine learningNon-stationary time seriesDownscalingHodrick–Prescott filter
spellingShingle Seyed Mojtaba Mousavimehr
Mohammad Reza Kavianpour
A non-stationary downscaling and gap-filling approach for GRACE/GRACE-FO data under climatic and anthropogenic influences
Applied Water Science
Groundwater level
GRACE
Machine learning
Non-stationary time series
Downscaling
Hodrick–Prescott filter
title A non-stationary downscaling and gap-filling approach for GRACE/GRACE-FO data under climatic and anthropogenic influences
title_full A non-stationary downscaling and gap-filling approach for GRACE/GRACE-FO data under climatic and anthropogenic influences
title_fullStr A non-stationary downscaling and gap-filling approach for GRACE/GRACE-FO data under climatic and anthropogenic influences
title_full_unstemmed A non-stationary downscaling and gap-filling approach for GRACE/GRACE-FO data under climatic and anthropogenic influences
title_short A non-stationary downscaling and gap-filling approach for GRACE/GRACE-FO data under climatic and anthropogenic influences
title_sort non stationary downscaling and gap filling approach for grace grace fo data under climatic and anthropogenic influences
topic Groundwater level
GRACE
Machine learning
Non-stationary time series
Downscaling
Hodrick–Prescott filter
url https://doi.org/10.1007/s13201-025-02427-z
work_keys_str_mv AT seyedmojtabamousavimehr anonstationarydownscalingandgapfillingapproachforgracegracefodataunderclimaticandanthropogenicinfluences
AT mohammadrezakavianpour anonstationarydownscalingandgapfillingapproachforgracegracefodataunderclimaticandanthropogenicinfluences
AT seyedmojtabamousavimehr nonstationarydownscalingandgapfillingapproachforgracegracefodataunderclimaticandanthropogenicinfluences
AT mohammadrezakavianpour nonstationarydownscalingandgapfillingapproachforgracegracefodataunderclimaticandanthropogenicinfluences