Application of the Combination of Remote Sensing and Machine Learning Approaches in Predicting Hydrological Parameters: A Bibliometric Analysis

The integration of remote sensing data with machine learning (ML) techniques has emerged as a robust and effective paradigm for predicting key hydrological parameters, including evapotranspiration, soil moisture content, and land surface temperature. This study presents a comprehensive scientometric analysis of research trends and international collaborative networks within this rapidly evolving field. Data pertinent to this investigation were retrieved from the Web of Science Core Collection database and subsequently analyzed using the Bibliometrix R package and VOSviewer software. These analyses facilitated the identification and visualization of complex interrelationships among scholarly publications, contributing authors, topical keywords, and affiliated countries/institutions. The findings reveal a prominent trend toward the application of advanced ML algorithms, such as Artificial Neural Networks (ANNs) and Random Forest (RF), in conjunction with remotely sensed data acquired from platforms like MODIS, Sentinel, and SMAP, particularly in regions characterized by limited in situ observational data. Furthermore, the utilization of multi-source data fusion and sophisticated ML algorithms for enhanced simulation accuracy of hydrological processes and improved predictive capabilities for climate change impacts and drought events has been identified as a key emerging research direction. Notably, the increasing reliance on satellite-derived datasets, including MODIS, SMAP, and the Normalized Difference Vegetation Index (NDVI), for hydrological parameter estimation in data-scarce environments constitutes another significant observation. Beyond identifying prevailing research trends, this study critically examines existing challenges, knowledge gaps, and potential avenues for future research endeavors in this domain.