Investigating runoff sensitivity in the land-surface model matsiro to reduce low runoff bias

Abstract

The Minimal Advanced Treatments of Surface Interaction and RunOff (MATSIRO), which has been used as a land-surface scheme in the global climate model, the Model for Interdisciplinary Research on Climate (MIROC), calculates Dunne runoff and base runoff using the TOPography-based MODEL (TOPMODEL). In past experiments that used MATSIRO, the runoff and its response to precipitation were too low compared to observation. We conjectured that those biases could be attributed to the water table’s excessive depth. Its depth was diagnosed based on grid-mean soil moisture using a saturation threshold that was originally set to almost equal the porosity. In this study, sensitivity experiments, in which the threshold was decreased to 75, 50, 25, and less than 13 % of the porosity, were conducted, and the subsequent effects on river flow were investigated in the Chao Phraya River basin, Thailand, as a case study. As a result, both Dunne and base runoff increased along with the response of river flow to precipitation. The simulated river flow matched observations most closely with a threshold of 50 % saturation. In addition, soil moisture and the Bowen ratio changed significantly with the runoff changes induced by the threshold changes. These results suggested the importance of the relationship between grid-mean soil moisture and groundwater level for the TOPMODEL. Preliminary global experiments indicate that runoff sensitivity might be dependent on climate zone.