RESEARCH HIGHLIGHT – Future snow changes over the Columbia Mountains, Canada, using a distributed snow model
Mortezapour, M., Menounos, B., Jackson, P. L., & Erler, A. R. (2022). Future Snow Changes over the Columbia Mountains, Canada, using a Distributed Snow Model. In Climatic Change (Vol. 172, Issues 1–2). Springer Science and Business Media LLC. https://doi.org/10.1007/s10584-022-03360-9
Click here to read the article.
This paper, co-authored by Andre Erler and researchers from the University of Northern British Columbia, investigates climate change impacts on snow depth using a distributed snow model called SnowModel. Snowmelt is an essential water source for communities, and seasonal snow accumulation in many regions is decreasing with each passing year. Water managers, communities, and policymakers can benefit from improved snow modelling forecasts to inform their decision making and understand vulnerabilities to their water supply systems. This study was conducted in the Columbia River basin in British Columbia and focused on projected changes in seasonal snow cover using SnowModel and different long-term climate forcing data inputs. (i.e. dynamically downscaled regional climate forcing data using the Weather Research and Forecasting (WRF) model, and statistically downscaled forcing data provided by the Pacific Climate Impacts Consortium (PCIC). As this study was based in BC, the complex, mountainous topography was an important factor to consider. The data produced by SnowModel was cross-checked with past studies to ensure accuracy and were consistent with earlier research.
The results of the study indicate that elevation and season are the most impactful variables driving snowpack loss, along with temperature and precipitation. Mid-elevation locations (1000–2000 m above sea level) will see the largest loss of snowpack. Locations below 2000m above sea level may see a 60% reduction in snow depth and snow water equivalent. The authors also found that dynamically downscaled forcing data (WRF) result in larger forecasted loss in snowpack compared to the statistically downscaled forcing data (PCIC).
Regardless of the model and forcing data uses, it’s clear that some areas of BC are warming at twice the global average, resulting in decreased snow depth and snow water equivalents. Decreasing snow depth in the winter means that the snow will melt faster in the spring and autumn. While statistically downscaled data and dynamically downscaled data yielding different projected changes in snow depth, they both have their advantages. For example, the dynamically downscaled input data was able to encompass a wider range of interactions across different geographic conditions, while statistically downscaled data had a lower computational demand.
Snowmelt and accumulation are important drivers of hydrology across many regions of Canada, which is why the Canada1Water project will include snowmelt analysis for scenarios just like this.
Want to learn more about how snow depth is incorporated into C1W? Click here!
Abstract
In western North America, many communities rely on runoff from mountain snowpacks. Projections of how future climate change will affect the seasonal snowpack are thus of interest to water managers, communities and policy makers. We investigate projected changes in seasonal snow cover for the twenty-first century for the Canadian portion of the Columbia River Basin using a physically based snow distribution model (SnowModel) at 500 m horizontal resolution. Forcing data for the reference (1979–1994) and future (2045–2059, 2085–2099) periods originate from a 4-member initial condition ensemble of global Community Earth System Model (CESM1) simulations based on the Representative Concentration Pathway (RCP) 8.5 scenario. The ensemble was dynamically downscaled (DD) to 10 km resolution using the Weather Research and Forecasting model (WRF). We also evaluate the performance of SnowModel using publicly available, statistically down-scaled (SD) temperature and precipitation. We project a 38%/28% and 30%/15% decrease in WRF/SD-simulated snow depth and SWE, respectively, by the end of this century relative to the reference period over the entire domain. Our results indicate that the projected loss of snowpack depends largely on elevation and season. Snow depth and snow water equivalent (SWE) are most affected for elevations below 2000m asl, with a reduction of more than 60%. While both simulations show SWE losses in most areas by the end of the century, a stronger projected thinning of the snowpack occurs for the DD-forced simulations compared to the SD-forced simulations.