Explaining CO2 fluctuations observed in snowpacks

Abstract

Winter soil carbon dioxide (CO2) respiration is a significant and understudied component of the global carbon (C) cycle. Winter soil CO2 fluxes can be surprisingly variable, owing to physical factors such as snowpack properties and wind. This study aimed to quantify the effects of advective transport of CO2 in soil-snow systems on the subdiurnal to diurnal (hours to days) timescale, use an enhanced diffusion model to replicate the effects of CO2 concentration depletions from persistent winds, and use a model-measure pairing to effectively explore what is happening in the field. We took continuous measurements of CO2 concentration gradients and meteorological data at a site in the Cape Breton Highlands of Nova Scotia, Canada, to determine the relationship between wind speeds and CO2 levels in snowpacks. We adapted a soil CO2 diffusion model for the soil-snow system and simulated stepwise changes in transport rate over a broad range of plausible synthetic cases. The goal was to mimic the changes we observed in CO2 snowpack concentration to help elucidate the mechanisms (diffusion, advection) responsible for observed variations. On subdiurnal to diurnal timescales with varying winds and constant snow levels, a strong negative relationship between wind speed and CO2 concentration within the snowpack was often identified. Modelling clearly demonstrated that diffusion alone was unable to replicate the high-frequency CO2 fluctuations, but simulations using above-atmospheric snowpack diffusivities (simulating advective transport within the snowpack) reproduced snow CO2 changes of the observed magnitude and speed. This confirmed that wind-induced ventilation contributed to episodic pulsed emissions from the snow surface and to suppressed snowpack concentrations. This study improves our understanding of winter CO2 dynamics to aid in continued quantification of the annual global C cycle and demonstrates a preference for continuous wintertime CO2 flux measurement systems.