Integration of remote sensing and hydrological modeling in water resources management: approaches and performance gains

Abstract

Hydrological models support water permit decisions, system operation and flood forecasting, yet they depend on sparse monitoring networks that limit their calibration and reliability. This study aimed to analyze, in recent scientific literature, the approaches for integrating orbital data and hydrological models, the performance gains obtained and the limitations reported. This is an integrative literature review conducted in the Scopus, Web of Science, ScienceDirect, Capes Journal Portal and SciELO databases, covering the period from 2021 to 2026, with reporting guided by the PRISMA 2020 statement. After identification, screening and eligibility assessment, 18 studies composed the final sample. Four integration approaches were identified: use of orbital products as input forcing, multivariable calibration, sequential data assimilation, and statistical or machine learning post-processing. Gains were consistent but conditional: assimilation raised the proportion of gauges showing improvement from 24 to 62 percent depending on the transformation applied to the observation, and peak flow forecasting improved between 5 and 46 percent. Multivariable calibration did not always improve streamflow, but it made the remaining water balance components consistent. It is concluded that the gain depends more on the design of the integration and on the quality of the chosen product than on simply adding orbital data to the model.