Spatial Interpolation of Meteorological Data in Complex Terrain Using Temporal Statistics

Abstract

Diagnostic wind field numerical models have significant difficulty developing representative wind velocities in complex terrain. A large of this difficulty begins with the initial wind field interpolation. If this interpolated wind field does not closely represent the true winder mass-consistent adjustments cannot retrieve the correct atmospheric flow patterns. Presently, the initial interpolation in diagnostic models is almost exclusively done using a simple one-over-separation-squared (1/r2) interpolation algorithm. This algorithm uses the closest or larger number of measurements of the wind velocity closest to the interpolation location. In this paper, we explore different interpolation algorithms using not only the measurement field at the interpolation time, but also the statistical relationships between stations. These algorithms were tested with data from the 1980 Atmospheric Studies in Complex Terrain (ASCOT) sponsored by the Department of Energy. The results show that, while consistent (though in most cases marginal) improvement in interpolated wind speeds was obtained, little improvement was derived for interpolated wind direction.