Premium
An efficient method for distributing wind speeds over heterogeneous terrain
Author(s) -
Winstral Adam,
Marks Danny,
Gurney Robert
Publication year - 2009
Publication title -
hydrological processes
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.222
H-Index - 161
eISSN - 1099-1085
pISSN - 0885-6087
DOI - 10.1002/hyp.7141
Subject(s) - terrain , wind speed , wind direction , environmental science , meteorology , wind power , spatial variability , snow , scale (ratio) , range (aeronautics) , geology , geography , mathematics , statistics , cartography , materials science , engineering , electrical engineering , composite material
High spatial variability of wind over mountain landscapes can create strong gradients in mass and in energy fluxes at the scale of tens of metres. Variable winds are often cited as the cause of high heterogeneity in snow distribution in non‐forested mountain locations. Distributed models capable of capturing the variability in these fluxes require a time series of distributed wind data at a comparably fine spatial scale. Application of atmospheric and surface wind flow models in these regions has been limited by our ability to represent this complex process in a computationally efficient manner. Simplified models based on terrain and vegetation parameters are not as explicit as more complex, fluid‐flow models, but are computationally efficient for real‐time operational use. We developed and applied a simplified wind model based on analysis of upwind terrain to predict wind speeds across diverse topographies at three mountainous research locations. Each site was instrumented with a network of wind sensors to capture the full range of wind variability present. Differences in upwind topography were significantly related ( p < 0·0001) to wind‐speed differences between sites. Wind speeds at each sensor location were modelled from each of the other intra‐site locations as if data from only one sensor were available. The wind model explained 69% of the observed variance with a mean absolute prediction error of 0·8 m/s, 19% of the observed wind mean. These results were very encouraging given the inherent complexity and profound variability of processes determining wind patterns in these systems. Copyright © 2008 John Wiley & Sons, Ltd.