An Experimental Investigation of Interstitial Air Movement in Seasonal Snow

Graduation date: 2016Access restricted to the OSU Community, at author's request, from June 11, 2016 - December 11, 2016Atmospheric pressure changes do not stop at the permeable snow surface but rather propagate into it. These pressure changes range from high-amplitude, low-frequency events caused by seasonal cycles and synoptic weather systems to small-amplitude, high-frequency events caused by topographic features and turbulence. The effect of pressure changes on interstitial air movement is locally weak but geographically pervasive and temporally persistent so the cumulative impact may be significant over seasonal timescales. Near the snow surface, pressure changes in the high-frequency range caused by turbulence and windflow over topographic features can enhance fluxes of chemically and radiatively active trace species between the snow and atmosphere. Deeper, in multi-year snow that overlays continental ice sheets, low-frequency pressure changes can stimulate air movement. This ventilation process adds complexity for paleoclimate analysis of ice cores to the extent that air trapped in ice has a different age structure than the ice matrix. For decades, investigators have recognized ventilation as a mechanism that enhances mass flux and interstitial air mixing but its effects are indeterminate because the relationship between wind forcing and interstitial air response is poorly constrained.\ud This dissertation addresses the experimental question: what is the effect of wind on interstitial air movement in snow? To address this question, in situ field experiments were designed and performed to measure the depth in seasonal snow affected by wind-generated pressure changes as a function of frequency of the pressure changes. One experimental result was that high-frequency pressure changes have greater amplitude than theory predicted. At first, this result may seem to indicate that high-frequency pressure changes affect interstitial air motion more than predicted but for another finding that high-frequency pressure changes (perturbations) also attenuate more with depth than current theory predicts. Therefore, strong high-frequency attenuation relegates the effect of high-frequency pressure changes to a very thin air layer near the snow surface. Enhanced perturbation pressure attenuation at high frequencies does not directly address the question of the degree to which low frequencies attenuate with depth. However, the high-frequency mismatch between experimental evidence and theory underscores the necessity for future endeavors to test anticipated low-frequency perturbation pressure attenuation with deep snow pressure measurements.\ud More accurate measures of the relationship between wind forcing and the spectral response of pressure changes acquired in this series of field experiments enabled characterization of the distribution of perturbation pressure amplitude as a function of frequency and wind forcing. It was found that kinetic energy of the wind as given by the horizontal and vertical components of the wind is a better diagnostic for perturbation pressure than vertical velocity variance. This finding is relevant when parameterizing perturbation pressure forcing using wind characteristics. A simplified model that convolves perturbation frequency with the spectral distribution of amplitude was used to diagnose frequencies for which water vapor flux (sublimation) is maximized for hydrostatic pressure changes. Applying the meteorological conditions measured in the case studies for this experiment sublimation enhancement was maximized for pressure oscillations with period ranging from 5 to 20 minutes. For shorter time periods the amplitude was too small to achieve a threshold (taken as the roughness length) and for longer time periods amplitude was sufficient but the frequency of the oscillation was insufficient to drive much air exchange between the snow and atmosphere.\ud Finally, interstitial air movement was calculated under various wind conditions by measuring the evolution of a trace gas plume (carbon monoxide) as detected by a network of thin film sensors. Near surface data revealed an advection signature oriented with the prevailing wind. The plume centroid propagated downwind but upwind dispersion was greater than crosswind dispersion. This dispersion signature is consistent with turbulent eddies that propagate downwind. When the measurement network was oriented in a vertical plane, the center of mass of the plume propagated upwards indicating that upward vertical dispersion was enhanced relative to downward dispersion. This finding indicates that the residence time of a neutrally buoyant gas in the upper portion of the snow column is significantly shorter than the same gas located lower in the snowpack