ИСТИНА

The study objective was to test a hypothesis that a hydrological model subjected to comprehensive evaluation procedure performs better and is more robust comparing with the roughly tested model i.e. the first model is, most likely, more suitable for climate impact assessment. The hypothesis testing was done on an example of the ECOMAG physically-based hydrological model that was setup to two great Arctic basins: the Lena and the Mackenzie rivers. Two versions of the ECOMAG model were compared: (1) the model with the a priori assessed parameters (without any calibration); (2) the model calibrated/evaluated against the streamflow observations for long-term period (split-sample test was used). The comprehensive evaluation procedure, which includes enhanced tests of model performance and model robustness, was applied for every model’s version. The model performance was assessed at multiple sites within the catchments and for multiple hydrological indicators of interest (high flow, low flow, FDC, multi-year trends). The model robustness was assessed through statistical significance of the differences in the model performance criteria for climatically contrasting periods composed from the historical meteorological data. The experiments demonstrated that the uncalibrated model was not able to pass the evaluation test for both basins. Calibrated ECOMAG performed well and occurred robust enough to be used for climate impact study in the Lena River basin as a whole and its sub-basins. For the Mackenzie River basin, the model has partly passed the comprehensive test. Additional study should be carried out to improve simulation of the observed trends.