Contrasting turbulent heat fluxes at two snow covered mountainous sites

The seasonal snowcover is an important water resource in many mountain environments, playing a key role in storing water for delayed release into streams. Accurately representing snow surface energetics is crucial to modelling the response of the snowpack to atmospheric forcing when using coupled atmospheric and hydrological models. Correctly specifying the turbulent sensible and latent heat fluxes is an important aspect of land surface models, affecting both modelling snow accumulation and melt, as well as setting the lower boundary in atmospheric models. In complex mountainous terrain standard MO similarity theory is increasingly shown to break down, and thus new avenues must be explored as to the best methods to represent turbulent fluxes in these environments. Two contrasting sites at Fortress Mountain Snow Laboratory, Alberta, Canada, are used to explore the variability of turbulent fluxes to snow in complex terrain and the potential difficulties in parameterisation of turbulent fluxes from AWS data. The Bonsai Clearing site is a forest-sheltered, very low wind speed environment (mean wind speed < 1 m s^-1) situated in the base of a valley at 2100 m above sea level (a.s.l.), and has a deep (> 1.5 m) and consistent snowpack throughout the winter. In contrast, Fortress Ridge is situated on a very exposed alpine rolling ridge top at 2323 m a.s.l. that experiences high wind speed (mean wind speed > 3 m s^-1) and a shallow and intermittent snowpack. Three dimensional sonic anemometer and fast-response hygrometer measurements at these two contrasting but proximal (< 1 km) sites allow turbulent flow properties and heat fluxes to be quantified through a period of the late winter and spring. Preliminary data show friction velocity at both sites does not scale well on the mean wind speed, instead scaling more effectively on the variance of the horizontal wind speed. Whilst horizontal wind speed variance contributed to shear stress, it did not always contribute proportionately to turbulent heat fluxes. Despite their different settings the turbulent heat flux magnitudes were surprisingly very similar at both sites, and generally modest (< +/- 100 W m^-2). Sensible heat flux was downwards to the relatively cold snowpacks and latent heat flux was indicative of snow sublimation during non-precipitation periods.