Description of the Probabilistic Wind Atlas Methodology, Deliverable D3.1

Andrea N. Hahmann, Björn Witha, Daran L. Rife, Nikolaos Frouzakis , Constantin Junk, Tija Sile, Magnus Baltscheffsky, Martin Dörenkämper, Yasemin Ezber, Elena Garcia Bustamante, Fidel Gonzalez-Rouco, Sibel Mentes, Jorge Navarro, Stefan Söderberg, Yurdanur Unal

Research output: Book/ReportReportResearchpeer-review

652 Downloads (Pure)


A new ensemble method is explored for estimating the uncertainty of the wind resource within Weather Research and Forecasting (WRF) model simulations. The output of the ensemble simulations is processed to create a "map" showing the uncertainty in the wind resource estimate at each geographic location. This new method is demonstrated by performing a collection of 9 different WRF model simulations using combinations of 3 planetary boundary layer schemes, 2
simulation re-initialization strategies, and 2 methods for initializing the land surface state. The results of the simulations are validated against data from 10 meteorological masts in South Africa, part of the Wind Atlas of South Africa (WASA) project, where a long-term set of high-quality observations exist. The results of the ensemble simulations are encouraging, but further analysis is needed to quantify their utility. A key disadvantage of the ensemble simulation strategy employed
herein, is that some members may tend to be highly similar to others, leading to overconfidence in the mean and spread of the simulations. Such overconfidence yields misleading estimates of the accuracy, value, and uncertainty of the wind resource.
Original languageEnglish
PublisherNEWA - New European Wind Atlas
Number of pages30
Publication statusPublished - 2017


Dive into the research topics of 'Description of the Probabilistic Wind Atlas Methodology, Deliverable D3.1'. Together they form a unique fingerprint.

Cite this