Separating local topography from snow effects on momentum roughness in mountain regions

Parametrization of momentum surface roughness length in mountainous regions continues to be an active research topic given its application to improved weather forecasting and sub-grid scale representation of mountainous regions in climate models. A field campaign was conducted in the Val Ferret watershed (Swiss Alps) to assess the role of topographic variability and snow cover on momentum roughness. To this end, turbulence measurements in a mountainous region with and without snow cover have been analyzed. A meteorological mast with four sonic anemometers together with temperature and humidity sensors was installed at an elevation of 2500 m and data were obtained from October 2011 until May 2012. Because of the long-term nature of these experiments, natural variability in mean wind direction allowed a wide range of terrain slopes and snow depths to be sampled. A theoretical framework that accounted only for topographically induced pressure perturbations in the mean momentum balance was used to diagnose the role of topography on the effective momentum roughness height as inferred from the log-law. Surface roughness depended systematically on wind direction but was not significantly influenced by the presence of snow depth variation. Moreover, the wind direction and so the surface roughness influenced the normalized turbulent kinetic energy, which in theory should not depend on these factors in the near-neutral atmospheric surface layer. The implications of those findings to modeling momentum roughness heights and turbulent kinetic energy (e.g. in conventional K-epsilon closure) in complex terrain are briefly discussed.

Turbulent Transport in Complex, Alpine Environments

Holly Jayne Oldroyd

Turbulent exchanges between the atmosphere and the underlying surface transport heat, moisture, momentum and pollutants, and thus understanding them is crucial to water resource management, meteorological predictions, climate modeling, wind energy production and pollution transport modeling and mitigation. The majority of established theories and techniques used to model, predict and observe turbulent exchange assume the underlying surface is flat, uniform and homogeneous. Consequently, in mountainous regions, turbulent exchanges are difficult to model and predict because they are intrinsically linked to the underlying surface conditions and complexity. Few observational studies have focused on the vertical structure of turbulent fluxes over mountainous terrain due to practical challenges. In sum, researchers and practitioners urgently need better theories and techniques to study and model turbulent exchange in alpine environments. In this dissertation, we analyze turbulent exchange measured over a steep, alpine slope in Switzerland to make quantitative comparisons with standard flat-terrain âlawsâ and practices and develop new theories and techniques adapted to mountainous terrain. First, flux-profile relations derived from Monin-Obukhov similarity theory are widely used in a variety of models to relate the surface conditions to atmospheric variables. We show that for katabatic flow, or thermally-driven air drainage that often occurs at night over sloping terrain, these relations break down because they do not account for the strong vertical gradients in the observed turbulent fluxes. However, the measurements exhibit clear functionalities that we use to derive new, empirical flux-profile relations for katabatic flow. Our functions indicate increased turbulent mixing compared to the flat-terrain relations. Misalignment between gravity and the inclined surface helps explain these functional differences because it reduces (and even reverses over steep slopes) the typical nighttime buoyant suppression of turbulent kinetic energy, which acts vertically via gravity. We also revise the governing flow equations to properly account for this misalignment. Second, standard data treatment techniques, such as sensor tilt corrections, do not hold true over complex topography. We develop an optimized sector-wise planar fit (SPF) tilt correction to best select the SPF input parameters for any complex site, and quantify the resulting sensitivities in the momentum fluxes. Finally, we present a mechanistic study that combines measurements and the one-dimensional momentum balance. Results show that frictional and buoyant mechanisms dominate, but that drag forces from the alpine grass canopy and the modeled outer-layer pressure (for conditions with low wind speeds) are also significant. In summary, this one-dimensional model or the newly developed flux-profile relations could replace inappropriate, flat-terrain wall models in larger-scale numerical models to improve atmospheric predictions for katabatic flow regimes. Additionally, terrain appropriate sensor tilt corrections and proper alignment of buoyant forces improve turbulent flux estimates for complex and steep topography, and hence have ramifications for similarity or budget analyses that require accurate quantification of turbulent exchange.

EPFL2015

Atmospheric boundary layers in complex terrain and over ice, snow and vegetated surfaces

Andreas Christen, Marc Diebold, Chad Higgins, Wolf Hendrik Huwald, Michael Lehning, Holly Jayne Oldroyd, Marc Parlange

Recent field campaigns held at the Val Ferret watershed in 2012 provided turbulent measurements in the atmospheric surface layer with and without snow cover. The turbulent kinetic energy (TKE) over the snow was reduced in comparison to the measurements obtained over bare surface. The “smoothing” of the surface by snow probably has a small role to play in the decrease of the TKE but the importance of the snow cover itself still has to be determined. Recent measurements obtained during the Plaine Morte 2013 field campaign using sonic anemometers are analysed. We discuss how the snowpack impacts the atmospheric turbulence under various snowpack conditions.

2013

Characterizing the spatial snow depth distribution in mountainous terrain

Thomas Grünewald, Tobias Jonas, Michael Lehning, Henning Löwe

Spatial snow depth information is an essential component for many applications such as the computation of snow melt rates in hydrologic applications or the energy balance in meteorological or climate models. Due to computational constraints, small-scale distributed modeling is rarely feasible for large regions. This is especially true over mountainous, complex terrain. Sub grid parameterizations for spatial snow depth distributions are therefore particularly valuable in the context of larger-scale applications. Past research has shown that for rather homogeneous landscape units the pre-melt spatial distribution of snow depth can be approximated by a log-normal distribution. Furthermore, seasonally recurring snow accumulation patterns have been reported, mostly shaped by terrain and wind. We focus on heterogeneous landscape units in order to investigate the impact of topographic parameters on sub grid snow depth distribution at peak of winter. Here we analyze a new data set with a particular view on developing a sub grid parameterization of snow depth for operational hydrologic applications and verifying previously developed models. Snow depth data obtained from an airborne laser scanning campaign in the Wannengrat Alpine catchment as well as from a unique opto-electronic scanning data set (Sensor ADS 80, Leica Geosystems) in the larger Dischma catchment (40km2) are used. Both areas are located above Davos, in the eastern Swiss Alps. A sub grid parameterization of snow depths variability at peak of winter will allow to compute snow covered area as a function of spatial melt rates during the ablation period, an issue which is also relevant for the parameterization of sub grid albedos in other large-scale models.

2013

Turbulent Transport in Complex, Alpine Environments

Holly Jayne Oldroyd

EPFL2015