Abstract. Cold region hydrology is very sensitive to the impacts of climate warming.
Impacts of warming over recent decades in western Canada include glacier
retreat, permafrost thaw, and changing patterns of precipitation, with an
increased proportion of winter precipitation falling as rainfall and shorter durations of snow cover, as well as consequent changes in flow regimes. Future warming is expected to continue along these lines. Physically realistic and
sophisticated hydrological models driven by reliable climate forcing can
provide the capability to assess hydrological responses to climate change.
However, the provision of reliable forcing data remains problematic,
particularly in data-sparse regions. Hydrological processes in cold regions
involve complex phase changes and so are very sensitive to small biases in
the driving meteorology, particularly in temperature and precipitation,
including precipitation phase. Cold regions often have sparse surface
observations, particularly at high elevations that generate a large amount
of runoff. This paper aims to provide an improved set of forcing data for
large-scale hydrological models for climate change impact assessment. The
best available gridded data in Canada are from the high-resolution forecasts
of the Global Environmental Multiscale (GEM) atmospheric model and outputs
of the Canadian Precipitation Analysis (CaPA), but these datasets have a
short historical record. The EU WATCH ERA-Interim reanalysis (WFDEI) has a
longer historical record but has often been found to be biased relative to
observations over Canada. The aim of this study, therefore, is to blend the
strengths of both datasets (GEM-CaPA and WFDEI) to produce a less-biased
long-record product (WFDEI-GEM-CaPA) for hydrological modelling and climate
change impact assessment over the Mackenzie River Basin. First, a
multivariate generalization of the quantile mapping technique was
implemented to bias-correct WFDEI against GEM-CaPA at 3 h ×0.125∘ resolution during the 2005–2016 overlap period, followed by a
hindcast of WFDEI-GEM-CaPA from 1979. The derived WFDEI-GEM-CaPA data are
validated against station observations as a preliminary step to assess their
added value. This product is then used to bias-correct climate projections
from the Canadian Centre for Climate Modelling and Analysis Canadian
Regional Climate Model (CanRCM4) between 1950 and 2100 under RCP8.5, and an
analysis of the datasets shows that the biases in the original WFDEI product
have been removed and the climate change signals in CanRCM4 are preserved.
The resulting bias-corrected datasets are a consistent set of historical and
climate projection data suitable for large-scale modelling and future
climate scenario analysis. The final historical product (WFDEI-GEM-CaPA,
1979–2016) is freely available at the Federated Research Data Repository at
https://doi.org/10.20383/101.0111 (Asong et al.,
2018), while the original and corrected CanRCM4 data are available at
https://doi.org/10.20383/101.0162 (Asong et al., 2019).
Application of a framework for high resolution climate change impact assessment on grapevine (Vitis vinifera L.) at regional scale: Phenology, yield and biotic stress responses