An Interpretable Machine Learning Framework for Unraveling the Dynamics of Surface Soil Moisture Drivers
Authors:
- Zahir NikraftarZNZahir Nikraftar
- Esmaeel PariziEsmaeel Parizi
- Mohsen SaberMSMohsen Saber
- Mahboubeh BoueshaghMBMahboubeh Boueshagh
- Mortaza TavakoliMTMortaza Tavakoli
- Abazar Esmaeili MahmoudabadiAMAbazar Esmaeili Mahmoudabadi
- Mohammad Hassan EkradiMEMohammad Hassan Ekradi
- Rendani MbuvhaRMRendani Mbuvha
- Seiyed Mossa HosseiniSeiyed Mossa Hosseini
Open Access
- 18 July 2025
- journal article
- research article
- Published by MDPI AG in Remote Sensing
- Vol. 17 (14) , 2505
Abstract
Understanding the impacts of the spatial non-stationarity of environmental factors on surface soil moisture (SSM) in different seasons is crucial for effective environmental management. Yet, our knowledge of this phenomenon remains limited. This study introduces an interpretable machine learning framework that combines the SHapley Additive exPlanations (SHAP) method with two-step clustering to unravel the spatial drivers of SSM across Iran. Due to the limited availability of in situ SSM data, the performance of three global SSM datasets—SMAP, MERRA-2, and CFSv2—from 2015 to 2023 was evaluated using agrometeorological stations. SMAP outperformed the others, showing the highest median correlation and the lowest Root Mean Square Error (RMSE). Using SMAP, we estimated SSM across 609 catchments employing the Random Forest (RF) algorithm. The RF model yielded R2 values of 0.89, 0.83, 0.70, and 0.75 for winter, spring, summer, and autumn, respectively, with corresponding RMSE values of 0.076, 0.081, 0.098, and 0.061 m3/m3. SHAP analysis revealed that climatic factors primarily drive SSM in winter and autumn, while vegetation and soil characteristics are more influential in spring and summer. The clustering results showed that Iran’s catchments can be grouped into five categories based on the SHAP method coefficients, highlighting regional differences in SSM controls.Keywords
All Related Versions
This publication has 124 references indexed in Scilit:
- Ensemble machine learning approaches for estimating soil texture components in loess soils of Golestan ProvinceEarth Science Informatics, 2025
- Groundwater sustainability assessment in the Middle East using GRACE/GRACE-FO dataHydrogeology Journal, 2023
- Soil moisture mapping for different land-use patterns of lower Bhavani river basin using vegetative index and land surface temperatureEnvironment, Development and Sustainability, 2023
- Investigating relationship between soil moisture, hydro-climatic parameters, vegetation, and climate change impacts in a semi-arid basin in IranArabian Journal of Geosciences, 2021
- Evaluation of gridded soil moisture products over varied land covers, climates, and soil textures using in situ measurements: A case study of Lake Urmia BasinTheoretical and Applied Climatology, 2021
- Spatial and temporal variability of soil moisture in relation with topographic and meteorological factors in south of Ardabil Province, IranEnvironmental Monitoring and Assessment, 2018
- Soil moisture response to rainfall on the Chinese Loess Plateau after a long‐term vegetation rehabilitationHydrological Processes, 2018
- GIS-based groundwater potential mapping using boosted regression tree, classification and regression tree, and random forest machine learning models in IranEnvironmental Monitoring and Assessment, 2015
- Spatial and temporal dynamics of soil moisture in a Mediterranean mountain area (Vallcebre, NE Spain)Ecohydrology, 2012
- Climate and Land Degradation — an Overview2007