[A-HW22] Material transportation and cycling at the land-sea interface: from headwaters to the ocean

[AHW22-07] Enhancing SWAT model for sediment simulation and parameter representativeness in diverse conditions

★Invited Papers

*Li-Chi Chiang¹, Pin-Chih Shih² (1.Associate professor, Department of Bioenvironmental Systems Engineering, National Taiwan University, Taiwan, 2.Research assistant, Department of Bioenvironmental Systems Engineering, National Taiwan University, Taiwan)

Keywords:SWAT, multiple stations, calibration and validation, sediment rating curve, typhoons

Sediment yields within a watershed are influenced by a range of natural disturbances and human activities, and hydrological models are commonly employed for their simulation. However, challenges arise in model calibration and validation due to the nature and availability of sediment data, whether continuous or discrete. This can significantly impact the representativeness of calibrated parameters for the area. This study addresses this issue by introducing sediment rating curves (SRCs) to enhance sediment measurements for model calibration. The investigation explores five strategies for incorporating both measured and estimated sediment data to enhance the calibration and validation of a hydrological model, specifically the Soil and Water Assessment Tool (SWAT), across multiple sediment stations. The strategies considered are as follows: (S1) using solely measured sediment data, (S2-S4) incorporating measured data along with varying proportions of estimated sediment during typhoon events, and (S5) utilizing estimated sediment data throughout the entire simulation period. Results indicate that while the model generally performs better with exclusively measured sediment data (S1), strategies S4 and S5 also prove effective in improving sediment simulation at downstream stations (STN1, 2, 3) and the upstream station (STN4), respectively. Furthermore, the impact of integrating different portions of typhoon-induced sediment data (S2, S3, S4) on model performance varies across stations. In conclusion, the proposed analytical approach demonstrates utility in calibrating sediment parameters for hydrological models when faced with limited or unevenly distributed measured sediment data.