Sampling of SAR imagery for wind resource assessment

Merete Badger⁽¹⁾, Jake Badger⁽¹⁾, Charlotte Hasager⁽¹⁾ and Morten Nielsen⁽¹⁾

⁽¹⁾ Risoe DTU, P.O. Box 49, DK-4000 Roskilde, Denmark

Abstract

Wind resources over the sea can be assessed from a series of wind fields retrieved from Envisat ASAR imagery, or other SAR data. Previous wind resource maps have been produced through random sampling of 70 or more satellite scenes over a given area of interest followed by fitting of a Weibull function to the data. Here we introduce a more intelligent sampling strategy based on wind class methodology that is normally applied in numerical modeling of wind resources. The aim is to obtain a more representative data set using fewer satellite SAR scenes.

Thirty years of NCEP/NCAR re-analysis data are used to define a number of geostrophic wind classes showing climatologically representative large-scale meteorological conditions for the area of interest. A screening of the Envisat ASAR archive is then made to find imagery to match each individual wind class. SAR samples entering the statistical analysis are weighted according to the frequency of occurrence for each wind class. The new sampling strategy has been applied within a wind and solar resource assessment study for the United Arab Emirates funded by Masdar and coordinated by UNEP. Wind statistics computed from the SAR data agreed well with numerical modeling results at the 10-m level.

Workshop presentation