The Cryospheric Science Group of the Interdisciplinary Centre on Climate Change (IC3) is pleased to announce the monthly Cryosphere Research Group Seminar Series for December given by PhD candidate Chad Thackeray, entitled:
“Characterizing the uncertainty in projections of Northern Hemisphere spring snow cover”
Chad W. Thackeray1, Christopher G. Fletcher1, Chris Derksen2, Lawrence R. Mudryk3
1 Department of Geography and Environmental Management, University of Waterloo, Waterloo, Canada
2 Climate Research Division, Environment Canada, Downsview, Toronto, Canada
3 Department of Physics, University of Toronto, Toronto, Canada
Several observational studies have shown that Northern Hemisphere (NH) spring snow cover extent (SCE) is decreasing at an alarming rate, not well captured by general circulation models (GCMs). The Intergovernmental Panel on Climate Change’s Fifth Assessment Report states that early spring SCE is likely to decrease by 7-25% by 2100. However, this projection was only assigned a medium confidence level because of a large intermodel spread, and a lack of sophistication in how many models represent snow processes. This study uses two model ensembles, the Coupled Model Intercomparison Project phase 5 (CMIP5) ensemble and the Canadian Earth System Model version 2 large ensemble (CanESM2-LE), to investigate the spread in projected 21st century spring SCE. An observational ensemble (consisting of several satellite-derived and reanalysis products) is also utilized to evaluate historical (1981-2010) trends. A majority of CMIP5 models underestimate the mean observed spring SCE loss over recent decades, with the greatest bias during March and June. These model deficiencies are closely tied to weaker than observed snowmelt sensitivity (snowmelt per degree of warming). In the near-future (2011-2040), the CMIP5 models exhibit a large spread in projected SCE trends (i.e., May: 0 to -1.25 million km2/decade). This partially stems from biases in the seasonal cycle of SCE, whereby variability in present day late-spring SCE (i.e., May: 5.0 - 16.8 million km2) influences melt rates. A similar spread exists within the CanESM2-LE projections during spring (i.e., May: -0.45 to -1.65 million km2/decade). Much of this internal variability (R2 = 0.46) can be explained by differences in NH spring land warming amongst the 50 ensemble members, while changes in snowfall have a lesser influence (R2 = 0.20). This analysis highlights the intermodel and internal uncertainty associated with projected spring SCE, along with their respective driving forces. Further research is ongoing to better understand the potential linkages between SCE trends and parameterizations of snow processes.