Impact of measured and simulated tundra snowpack properties on heat transfer
Abstract
Variability in vertical profiles of snowpack microstructure controls transfer of heat through a snowpack and the temperature of underlying soils. In situ measurements of snow and soil properties from four field campaigns during two different winters (March and November 2018, January and March 2019) were compared to an ensemble of CLM5.0 (Community Land Model) simulations, at Trail Valley Creek, Northwest Territories, Canada. Snow MicroPenetrometer profiles (n = 1050) allowed snowpack density and thermal conductivity to be derived at higher vertical resolution (1.25 mm) than with traditional snowpit observations (3 cm resolution, n = 115). Comparing measurements with simulations shows model overestimation of snow thermal conductivity by a factor of 3, leading to a cold bias in wintertime soil temperatures (RMSE = 5.8 oC). Bias-correction of the simulated thermal conductivity (relative to field measurements) improved simulated soil temperatures (RMSE = 2.1 oC). Multiple linear regression shows the required correction factor is strongly related to snow depth (R2 = 0.77, RMSE = 0.066). The correction factor is dependent on differences between measured and simulated snow depth values, particularly early in the winter as simulations do not adequately represent the high proportions of depth hoar in shallow arctic snowpacks. This is important, as wintertime soil temperatures act as a control on subnivean soil respiration, and thus improving their simulation helps to reduce uncertainties surrounding Arctic winter carbon fluxes and budgets.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2021
- Bibcode:
- 2021AGUFM.C13A..08D