Evaluation of Wind Flows and Turbulent Fluxes in Complex Terrain of Canadian Rockies

Abstract

In mountains, the role of diurnal wind (i.e. valley, slope winds) due to differential heating, radiation and topography in controlling fluxes of heat and water vapour is not well understood. Since data in high mountain areas are limited, high resolution models can help resolve near-surface processes and their diurnal changes to use as an input to hydrological models for more accurate predictions of evapotranspirartion and future water resources. Improvements over recent years in the resolution of Numerical weather prediction (NWP) models and large-eddy-simulation (LES) have had made great progress on resolving the atmospheric boundary layer (ABL) and boundary layer processes over mountainous terrain. In this work, the Weather Research and Forecasting (WRF) model is used to simulate flow in LES mode over the complex terrain of the Fortress Mountain and Marmot Creek research basins (MCRB and FMRB, respectively), Kananaskis Valley,Canadian Rockies, Alberta in mid- and late summer. The days selected in this study allow for development of thermally-induced wind circulation and ABL processes. However, the use of terrain-following coordinates in most numerical weather prediction models results in errors that propagate through the domain and can result in numerical instability. To avoid this issue when simulating flow over steep terrain a local smoothing approach was used, where smoothing is applied only where slope exceeds some predetermined threshold. The results are compared with global smoothing, which uniformly filters terrain, and is already implemented in WRF. Local smoothing with the cumulus parametrization activated only for the parent domain provides better predictions for surface wind direction, improved predictions for net radiation, and better RMSE for humidity, and was used for the rest of the analysis on turbulence kinetic energy (TKE) and near- surface processes. The model shows that valley flows are impacted by wind gusts and topographic wind originated from higher elevations blowing into the valley. In this study, up-valley flows were stronger in the wide but deeper Kananaskis Valley in MCRB, as compared to the narrower and shallower valley in FMRB. In addition, cold-air pools seem to linger longer in the deeper and wider valley at MCRB, but air temperature was lower in the early morning at the shallower but narrower valley at FMRB. The removal of the cold air pool due to temperature rise happened earlier in the valley in FMRB than in the valley of MCRB due to an elevated inversion layer of the deeper valley.

Boundary layer processes and turbulence in complex terrain are influenced by thermally-induced flows, as well as dynamical or non-local winds. Data from three high-frequency eddy covariance systems at a northwest-facing slope location, and at two ridgetops at the south and north valley side walls of the Fortress Valley were combined with LES to investigate the influence of diurnal mountain flows on TKE. Simulated cross sections showed up-valley flow was inclined toward the northern valley wall at the southeast side of the valley, and the interactions between the up-valley flow and the cross-ridge flows contribute to TKE in the valley. It was found that there is a strong correlation between TKE and wind speed at ridgetops, while TKE in the valley correlated strongly with the wind speed at the northern ridgetop. Furthermore, TKE budget analysis showed that horizontal shear could be an important source of TKE production at the northwest-facing slope station in the Fortress Valley. The variability observed in TKE budget components across different locations within this high mountain basin indicates the significance of both horizontal and vertical exchange processes in the mechanisms governing TKE production.

The final portion of this study evaluated model predictions of sensible and latent heat fluxes versus observations at three eddy-covariance locations in the Fortress Valley. The differences between model predictions and observations illustrates the crucial role of soil moisture, along with net radiation, in controlling the heat and evaporative fluxes in mountainous terrain. The observations over July and August were further used to quantify the variability of the sensible and latent fluxes with soil moisture content and net radiation, as influenced by elevation and vegetation. Observations showed that despite variations in vegetation type and elevation, the latent heat flux exhibited a weak correlation with soil moisture at each site but displayed a strong correlation with net radiation at all sites for both wet and dry days. But when all study sites were compared together for mid- versus late summer sunny days, it was noted that the local topography and soil moisture, radiation, and local flows can all have important impacts on turbulent fluxes. The findings also indicate that longer term data with a wider range of soil moisture, and topographical features (i.e slopes, aspect) will be beneficial for more in depth future studies on exchange processes in mountainous terrains.