A Grid-Based Conceptual Hydrological Modelling Framework Using Remotely Sensed Inputs: Preliminary Insights

Conceptual hydrological models play a central role in streamflow estimation, yet their lumped formulations often fail to represent spatial variability in hydrological processes, thereby limiting their performance. With the growing availability of satellite observations, global reanalysis products and high-resolution terrain datasets, grid-based conceptual modelling has become increasingly feasible. However, despite the widespread use of the Génie Rural à 4 Paramètres Journalier (GR4J) model, its grid-to-grid implementations remain limited, even though these frameworks offer clear advantages for capturing spatial heterogeneity and enabling modelling in data-scarce regions. This study presents a grid-based GR4J framework coupled with Muskingum-Cunge routing and driven entirely by remote sensing and reanalysis-based inputs, applied across four Australian catchments representing tropical, semi-arid, temperate, and humid subtropical climates. To implement this framework, the catchments were discretised into 0.1° grids aligned with the spatial resolution of GPM IMERG precipitation and GLEAM potential evapotranspiration, enabling these inputs to be applied at the grid level to generate runoff. Flow routing was done using Muskingum-Cunge method through the channel grids obtained based on flow-direction map derived from a digital elevation model (DEM), enabling sequential upstream-to-downstream runoff transfer across the grid network. Model calibration and validation were carried out using observed daily streamflow at the study catchment outlets for the periods 2005–2018 and 2018-2023 respectively. Model performance was evaluated using the Kling-Gupta Efficiency (KGE) under two configurations: the grid-based framework and the conventional lumped GR4J model. Both achieved calibration KGE values above 0.6; however, the grid-based model consistently showed superior performance during validation. The tropical basin exhibited the greatest improvement, with KGE increasing from 0.09 to 0.51, while the semi-arid and temperate basins showed 21% and 14% gains, respectively. Performance in the humid subtropical basin remained comparable across both configurations. Overall, the grid-based framework shows clear benefits in accounting for spatial variability.