摘要

The steady-state, Reynolds-averaged momentum equations, with a simple eddy viscosity closure, are solved numerically to compute the spatial variation in surface-layer mean windspeed over irregular, gently rolling terrain. Simulations, with both this non-linear model "ASL3D" and (for comparison) with the pre-existing linear "Mixed Spectral Finite-Difference" or MSFD model, are compared with observed winds from anemometers aligned on a 140 m transect in a rolling field near Lacombe, Alberta. Recorded wind speeds, normalized and aggregated by wind direction sector, characterize local wind variation over terrain whose elevation varied by roughly 10 m over a radius of about half a kilometer from the instrumented transect.
For northeast and southwest winds particularly, both models agree well with the observations. In southeast winds, observed spatial variation of the wind was weak, except that an anemometer close to fences and gates recorded distinctly lower speeds: provided those obstructions are represented by adding a localized sink in the momentum equations, the ASL3D model transect is (again) in quite good agreement with the observations. For northwest winds, however, agreement of modelled and measured transects is poor, presumably because a steep, wooded slope lay upwind from the anemometer array. Overall the linear correlation coefficient between modelled and observed fractional deviations of wind speed from the reference value is 0.72.
Other than as regards the flexibility to represent such complications as fences, plant canopies (etc.), computed wind fields over for the present terrain do not suggest any compelling advantage of the more laborious non-linear model (ASL3D) over the semi-analytical MSFD treatment. It is concluded that, when applied over gentle terrain, the skill intrinsic to even such a simple paradigm as ASL3D (and MSFD) represents a meaningful and potentially useful alternative to the neglect of lateral inhomogeneity.

  • 出版日期2018-2-15