NUMERICAL EXPERIMENTS TO CALIBRATE A HYDROLOGICAL MODEL BASED ON STREAMFLOW SIGNATURES AND HYDROLOGICAL TIME SERIES DATA

Abstract

 One of methods to estimate hydrograph in ungauged basins is to calibrate a hydrological model through setting streamflow signatures or hydrological time series data as the objective variables. To detect the streamflow signatures or hydrological time series data which have ability to reproduce hydrograph as the objective variables in the calibration, numerical experiments were conducted in 23 basins in Japan. Note that we employed Baseflow index (BFI), Flashiness index (FI), annual runoff height (Q_MEAN), and 3 indices of flow duration curve (Q₁₀, Q₅₀, Q90) for the streamflow signatures, and terrestrial water storage anomaly (∆TWS) and soil water anomaly (∆SM) for the hydrological time series data. We found that ∆TWS has ability to reproduce hydrograph as the objective variables in the calibration. The reason is that ∆TWS can relate to all parameters within the hydrological model and have enough information of the model error due to time series data.