Calibration of hydrological models for medium-term streamflow prediction in a changing climate

摘要

Prediction of medium-term (1-10 years) streamflow has become more challenging under climate change conditions. This paper presents results from a modelling study carried out to evaluate the performance of conceptual rainfall-runoff models in predicting medium-term streamflow using parameter values from model calibration against different periods of observed historical data. Four rainfall-runoff models are calibrated using the past 1 year, 2 years, 5 years, 10 years, 20 years and the entire historical daily climate and streamflow data for 20 catchments in southeastern Australia. The calibrated models are then used to predict the next 1 year, 2 years, 5 years and 10 years of streamflow. Calibrating hydrological models against the more recent data may give better medium-term streamflow predictions because the more recent data are likely to be of better quality, and they represent the current hydroclimate state and current land use and development. The results indicate that at least 10 years of data are needed for the model calibration to properly capture the range of hydroclimate variability. However, the modelling experiments here show little difference in the streamflow prediction results using parameter values from model calibrations against the past 10 or 20 years of data or the entire historical record.