From small to mega river basins−adaptation of a lumped hydrological model for runoff simulation

• Parsimonious hydrological models may be enabled to use in large-scale water management. • CAWM model was adapted to simulate runoff in perennial river basins. • Model applied to six basins ranging from areas from 1,200 to 4.6 million km 2 . • Model efficiently reproduced flows in semi-arid and humid climatic regions. • Model validation showed good to very good performance indicators in all applications. Conceptual lumped hydrological models are widely used to simulate surface runoff in small catchments, particularly in intermittent or temporary river systems. This study presents an updated version of the CAWM IV (Campus Agreste Watershed Model), originally developed for hydrological simulations in small semi-arid basins, in which the core mathematical structure is preserved while key conceptual extensions are introduced. The main update consists of the inclusion of a groundwater reservoir and an explicit soil–groundwater exchange threshold, enabling the representation of baseflow generation and runoff processes in perennial rivers and under more humid conditions. The applicability of the updated CAWM model is evaluated across six Brazilian river basins with contrasting drainage areas, ranging from 1,316 km 2 to approximately 4.6 million km 2 , including the Amazon, Madeira, Tocantins–Araguaia, Parnaíba, das Velhas, and Sirinhaém basins. Model performance is assessed during calibration and validation periods using standard efficiency metrics and compared with the GR4J and HEC-HMS models. Across the mega basins, CAWM showed robust and stable performance, with Nash–Sutcliffe Efficiency (NSE) values ranging from approximately 0.76 to 0.94 and low volumetric bias. In the smaller basins, CAWM achieved satisfactory to very good performance (NSE typically between 0.74 and 0.85), comparable to or better than HEC-HMS and clearly outperforming GR4J. Overall, the results demonstrate that the updated CAWM model can reliably reproduce river flow dynamics across basin scales.