Date of Award

1-1-2022

Language

English

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

College/School/Department

Department of Atmospheric and Environmental Sciences

Content Description

1 online resource (xxi, 180 pages) : color illustrations, color maps.

Dissertation/Thesis Chair

Jiping JL Liu

Committee Members

Andrea AL Lang, Mathias MV Vuille, Liming LZ Zhou

Keywords

Arctic sea ice, CMIP6, projection, snow depth, snow mass budget, variability, Sea ice, Snow, Snow surveys, Remote-sensing images

Subject Categories

Atmospheric Sciences | Climate

Abstract

The Arctic has experienced rapid environmental changes in recent decades. Motivated by the important role of snow over sea ice in influencing Arctic climate variability and change, this dissertation aims to improve our knowledge of spatiotemporal variability of snow depth and processes over Arctic sea ice in the observations and model simulations.First, we evaluate snow depth over Arctic sea ice during 1993–2014 simulated by the models from the Coupled Model Intercomparison Project phase 6 (CMIP6) against recent satellite retrievals. The CMIP6 models capture some aspects of the observed snow depth climatology and variability. The observed variability lies in the middle of the models’ simulations. All the models show negative trends in snow depth during 1993–2014. However, substantial spatiotemporal discrepancies are identified. Compared to the observation, most models have late seasonal maximum snow depth (by two months), remarkably thinner snow for the seasonal minimum, an incorrect transition from the growth to decay period, and a greatly underestimated interannual variability and thinning trend of snow depth over multi-year sea ice. Future projections suggest that snow depth in the Arctic will continue to decrease. Under the SSP5-8.5 scenario, the Arctic will be almost snow-free in summer and fall and the accumulation of snow starts from January. Further investigation suggests that model resolution, the inclusion of a high-top atmospheric model, and biogeochemistry processes are important factors for snow depth simulation. Second, for the first time, we intercompare snow mass budget processes in the Arctic simulated by the CMIP6 models using new diagnostics that have not been available for previous models. The multi-model mean shows that snowfall (snow melt) is the dominant process contributing to 100% (70.4%) of the annual snow growth (loss). Snow mass change through sea-ice dynamics, snow-ice conversion, and snow sublimation contribute 10.9%, 9.7%, and 9.0% to the total snow mass loss. The seasonal cycle of all snow processes simulated by most of the CMIP6 models generally follows similar variations, such as minimum snowfall and snow-ice conversion and maximum snow melt in summer, and maximum snow sublimation (snow mass change due to sea-ice dynamics) in late (early) spring. There is reduced Arctic-wide snow mass change due to snow melt and sea-ice dynamics during 1993-2014. However, there are a number of key differences between the CMIP6 models. There is a large spread of snowfall in summer and spring, and a large spread of snow-ice conversion (sublimation) from autumn to spring (late autumn to late spring). Almost half of the models indicate decreasing trends of snowfall during 1993-2014 whereas the other half have no trend. The simulated trends of snow sublimation and snow-ice conversion also diverge in the CMIP6 models. Future projections suggest a significant decrease in snowfall in the Arctic from 2015 to 2099 under the SSP5-8.5 scenario. Snow melt, snow-ice conversion, snow sublimation, and sea-ice dynamics are also projected to be reduced. Third, we identify dominant spatiotemporal variability patterns of the observed daily Arctic snow depth change in September-December during 1993-2018 using self-organizing map (SOM) analysis. Two dominant patterns are identified, one features an anomalous increase of snow depth in the Canadian Arctic (CA) and the other has an anomalous increase of snow depth in the central Arctic north to Greenland (NG). Further analyses of atmospheric dynamic and thermodynamic states linking with the two patterns suggest that the CA pattern is mainly associated with enhanced moisture and temperature from the anticyclonic anomaly in the Pacific sector and an anomalous Greenland high. The NG pattern is mainly associated with a westward shift of storm track in the North Atlantic due to an anticyclonic anomaly over northern Eurasia. We also investigate the possible connection between the two patterns and Arctic sea ice cover and sea surface temperature outside the Arctic. Finally, we validate whether the CMIP6 models can reproduce the dominant daily snow depth change patterns identified above. It is found that most of the models can reproduce the NG pattern associated with the shifted north Atlantic storm track, while about half of the models can reproduce the CA pattern which is related to high-pressure anomalies over the Pacific sector and Greenland. We also extend the atmospheric dynamics and thermodynamics analyses associated with the CA and NG patterns for the CESM2 model family with different configurations. The results show that the time and magnitude of atmospheric responses (e.g., specific humidity and air temperature) to snow events simulated by the CESM2 model family are different from those in the observations. Nevertheless, CESM2-WACCM is relatively better in reproducing the observed atmospheric responses associated with both the CA and NG patterns compared with other CESM2 family models, indicating a potential role of the inclusion of high-top atmosphere and fine atmospheric resolution in simulating the realistic spatiotemporal variability of the Arctic snow depth.

Download

Share

COinS