Process-Based Estimates of Seasonal Streamflow for Ungauged Catchments

The seasonal water balance is at the core of overall catchment responses, but current methods to simulate seasonal water availability at ungauged locations are unreliable. Here we enhance two simple models from Woods (2003 and 2009, Advances in Water Resources) to estimate actual evaporation, changes in storage, and streamflow, using summary statistics of precipitation, temperature, and potential evaporation from the CAMELS-US dataset. Specifically, we first use sinusoidal functions to simulate observed precipitation, temperature, and potential evaporation to quantify several parameters (e.g., climate dryness index). We then use these variables and parameters to force our simple model representing the main hydrological processes to generate estimated streamflow. We then assess the model’s ability to simulate the seasonal flow regime across many catchments worldwide. Finally, we identify the dominant variables and processes controlling the seasonal water balance and discuss the limitations of our model. This allows finding in which situations we can reliably estimate seasonal variation in catchment streamflow without flow measurements, and other cases where model refinement is needed. Our study is important to improve our understanding of seasonal catchment hydrology and for predicting the seasonal hydrological differences between various hydro-climatic conditions or catchments, especially in locations with sparse measurements.