Presentation of two INFO-Crue projects carried out at the HC3 Laboratory of ÉTS

*Webinar in French

This webinar presents the results of two research projects conducted by the HC3 Laboratory group at ÉTS as part of Info-Crue. The projects are: Improved historical reconstruction of daily flows and annual maxima in gauged and ungauged basins and Development of a gridded observational dataset for the post-processing of simulations according to hydrological performance.

In the first case, multi-model weighting techniques were applied to a set of flow series from 144 HYDROTEL calibrations. The various calibrations were generated by modifying: (1) the HYDROTEL parameters, (2) the model input meteorological data sets, and (3) the evapotranspiration module used. 

In the second case, precipitation and temperature grids were generated from a combination of station, reanalysis and satellite data. These grids cover all of Quebec and are available for the period from 1970 to 2019, at a spatial resolution of 0.1° and a temporal resolution of 24 hours.

Improved historical reconstruction of daily flows and annual maxima in gauged and ungauged basins

Presented by Richard Arsenault (École de technologie supérieure)

This study presents a method to improve the reconstruction of flow histories on gauged and ungauged basins. To achieve this, the HYDROTEL distributed hydrological model was implemented on a test bed of 96 gauged watersheds in Quebec. Different multi-model weighting techniques aimed at combining several HYDROTEL calibrations were compared to an optimal interpolation (OI) assimilation method, currently used by the Direction de l'expertise hydrique, to combine one of the HYDROTEL calibrations with observations. Multi-model weighting techniques were applied to a set of 144 HYDROTEL calibrations that were generated by modifying the parameters, meteorological datasets, and evapotranspiration modules.

Our results indicate that the proposed multi-model weighting method (especially the Granger-Ramanathan variant "A") can outperform the IO method without the drawbacks that the latter may introduce. The proposed method can also be further improved by increasing the number and variety of hydrologic simulations, and by calculating its weights over a shorter time period. In contrast, the IO method cannot use this additional information, thus reaching a performance plateau. This study shows that it is possible to improve the simulations of regional hydrological models, both for global and peak flows, as well as over the historical period for gauged and ungauged basins. This can then be used to better estimate the risk of flood frequency analysis and other statistical analyses.

Development of a gridded observational dataset for the post-processing of simulations according to hydrological performance

Presented by Annie Poulin (École de technologie supérieure)

As part of the research work of the INFO-Crue project, the HC3 Laboratory team has been mandated to develop a set of observational grids of precipitation and temperature. These grids cover the period from 1970 to 2019, their spatial coverage extends over all of Quebec at a resolution of 0.1° and the data are available on a daily basis. The objective of this project is to create an optimal combination, for Quebec, of data from various sources. The resulting dataset, named GRNCH-QC, is intended in particular to be used for the post-processing of climate simulations.

The approach used to generate the grids is based on a combination of station, reanalysis and satellite data, and is inspired by the method used by Beck et al. for the generation of MSWEP V2 precipitation grids. The selected products were selected and combined in order to maximize the added value of the GRNCH-QC grids over the territory of the province of Quebec. The analysis of different precipitation and temperature metrics, as well as the calibration and validation of the HMETS and GR4JCN hydrological models in several Quebec watersheds, confirmed the overall advantage of the GRNCH-QC grids over MSWEP V2, the MELCC GCQ grids and all the data sets integrated in GRNCH-QC. Furthermore, the hydrological modeling results show the advantage of GRNCH-QC grids over GCQ grids in northern Quebec and Labrador, where station data are scarce.