Does a successful comprehensive evaluation increase confidence in a hydrological model intended for climate impact assessment?
- 27 November 2020
- journal article
- research article
- Published by Springer Science and Business Media LLC in Climatic Change
- Vol. 163 (3), 1165-1185
- https://doi.org/10.1007/s10584-020-02930-z
Abstract
The objective of the study is to verify a hypothesis that a hydrological model, which successfully passed a comprehensive evaluation test (CE-test), is more suitable for climate impact study than that which failed the test. In our study, the CE-test is a specially designed model evaluation procedure, including a set of enhanced tests of model performance and robustness. The hypothesis verification is carried out with two models, ECOMAG and SWAP, which are applied for the Lena and Mackenzie River basins. The following three versions of every model are compared: (1) version A with a priori assigned parameters (without any calibration); (2) version B calibrated against streamflow observations at the basin outlets only, and (3) version C calibrated against streamflow observations at several gauges within the basins. We found that the B and C versions were successful in passing the CE-test, while the A versions failed the test. The C versions performed better than the B versions, especially at the monthly time scale. Then, all model versions were forced by global climate model (GCM) ensemble data to simulate flow projections for the twenty-first century and assess the projection uncertainty. Summarizing the results, we found that the differences in projections (in terms of mean annual changes in discharge and their uncertainties) between A version and two other versions were nearly three times larger than the differences between the B and C versions. Thus, the CE-test results together with the estimated differences in projections give us reason to conclude that the successful comprehensive evaluation of a model increases its confidence and suitability for impact assessment.Keywords
Funding Information
- Russian Science Support Foundation (19-17-00215, 16-17-10039)
This publication has 46 references indexed in Scilit:
- Time stability of catchment model parameters: Implications for climate impact analysesWater Resources Research, 2011
- Climate non-stationarity – Validity of calibrated rainfall–runoff models for use in climate change studiesJournal of Hydrology, 2010
- Decomposition of the mean squared error and NSE performance criteria: Implications for improving hydrological modellingJournal of Hydrology, 2009
- The simulation of runoff from small catchments in the permafrost zone by the SWAP modelWater Resources, 2006
- Cold region river discharge uncertainty—estimates from large Russian riversJournal of Hydrology, 2005
- Model Parameter Estimation Experiment (MOPEX): An overview of science strategy and major results from the second and third workshopsJournal of Hydrology, 2005
- The distributed model intercomparison project (DMIP): motivation and experiment designJournal of Hydrology, 2004
- The simulation of heat and water exchange in the boreal spruce forest by the land-surface model SWAPJournal of Hydrology, 2003
- The hydrology of tomorrowHydrological Sciences Journal, 1986
- A Statistical Exploration of the Relationships of Soil Moisture Characteristics to the Physical Properties of SoilsWater Resources Research, 1984