Differences in water balance between grassland and forest watersheds using long-term data, derived using the CoupModel

Abstract

Abstract To quantify the role of land cover during a period of climate change, the runoff response is studied for Plynlimon in Wales, UK. The main objective was two-fold: (i) to create a protocol for modeling water balance on a daily basis; and (ii) to describe the extent to which the impact of land-use changes can be identified and supported by the long-term monitoring data of runoff from two neighboring watersheds with different land covers. The process-oriented CoupModel platform was used to set up the model with a well-defined uncertainty for selected parameters. The behavioral ensembles were applied to simulated daily discharge data for the period of 1992–2010 using subjective criteria to reduce the prior 35,000 candidates with a random uniform distribution of 40 parameters. The accepted ensemble was reduced to 100 candidates by accepting the best root-mean-square error (RMSE) on the accumulated residuals during the simulation period. Similar good performance for the entire period and both watersheds was obtained. The differences in interception evaporation accounted for the most important differences between forest and grassland. The obtained residual demonstrated that changes in the forest cover had an impact on the water balance during the first part of the simulation period.