Climate Projections Drive C1W
Author: Dr. Andre R. Erler
Weather and climate are the primary drivers of hydrology and are of critical importance to the mission of Canada 1 Water. Consequently, projections of climate change impacts on water resources can only be trusted if the major climate processes controlling hydrology across Canada are represented in the climate models that are used for projection. This requires high-resolution regional climate modeling with particular attention to prominent features of the Canadian landscape and climate, such as the Rocky Mountains, the Great Lakes and extensive snow and permafrost cover. Furthermore, this year CMIP6, a new repository of climate model ensembles, has been made available in support of the most recent IPCC Assessment Report (AR6).
The Canada 1 Water climate modeling team will take advantage of this latest set of global climate projections and use projections from several leading international climate modeling centers as boundary conditions for high-resolution regional climate projections, specifically for Canada. These regional climate projections will resolve major land features and mountain ranges at 12 km resolution (the highest resolution currently used at this scale), and a new representation of large lakes will be employed to better capture lake-effect precipitation. Furthermore, the land surface model will be upgraded to provide the best possible simulation of snow cover, a major driver of hydrologic change, as well as permafrost. This is especially important, as most efforts to-date have failed to achieve an adequate representation of the arctic land surface.
Due to the high computational cost of such simulations, only a small ensemble will be generated, using different process representations as well as global climate scenarios (from different modeling centers). The focus will first be on a single moderate to high emission scenario (SSP3-7.0), as this is deemed most valuable for adaptation planning.
Despite the technical sophistication of the planned regional climate simulations, it is anticipated that some biases (errors) will remain and are unavoidable. However, errors in surface temperature in particular tend to cause even larger errors in ice and snow cover. For this reason, a bias-correction step will be applied, after which the land-surface component, which includes snow and permafrost, will be rerun, to correct errors on snow and permafrost extent. This step is considered critical for modeling of hydrological impacts of climate change that is often omitted when climate projections are bias-corrected or downscaled.
Once completed (approximately two years from now), the regional climate projections will be contributed to the NA-CORDEX project, which is a repository of regional climate projections for North America. If everything goes well, the C1W team may be the first to contribute projections based on the latest CMIP6 global projections.
Canada 1 Water is now nearing the end of the first year and the climate modeling team, which is based at the University of Toronto and Aquanty, has already set up the regional climate model (the Weather Research and Forecasting model, WRF) at the Niagara super-computing facility and conducted multi-year test runs with different model configurations. It is expected that several multi-decadal historical simulations will be completed before the end of the year, so that work on the CMIP6 boundary conditions can begin, and climate projection runs can commence mid-next year. At the same time, work has begun on testing the land surface model in stand-alone mode, to optimize the representation of snow and permafrost. To this end, observational datasets have been acquired and we have initiated collaborations with domain experts at the Geological Survey of Canada, the Canadian Forest Service and the University of Waterloo.