Offshore Wind Resource Estimation in Mediterranean Area Using SAR Images.
Calaudi, Rosamaria1; Arena, Felice2; Badger, Merete3; Sempreviva, Anna Maria4
1University Mediterranea Reggio Calabria -Italy, ITALY; 2UNIRC, ITALY; 3DTU-Wind Energy, DENMARK; 4ISAC National Council of Research - CNR, ITALY

Estimation of wind resource is important for planning wind farms. The energy output of a wind farm can be predicted by knowing the local wind climate. Usually, the wind climatology is based on at least one year of accurate wind measurements, but before such data are available at a site, satellite-based wind mapping can be a helpful tool in giving the first estimates of the wind conditions. In some areas of the world where no observations are available, satellite Synthetic Aperture Radar (SAR) and other remote sensing wind measurements can aid in giving the first estimate of the 10 m wind conditions at a site. The Mediterranean is less windy than the North European Seas and coastal waters are deep; however, technological progress in buoyant wind turbines and less harsh meteorological condition might be the winning factors since approximately 94% of resource area lies in water deeper than 60m. In this area joint analysis of data of SAR images crossed with data from the model and experimental data on the wind energy available have not been realized yet, and the used methodology is at the state of the art. Here, we focus on the SAR images that have the advantage of high spatial resolution (down to 100m) allowing to derive information close to the coast but with the disadvantage of low time resolution causing lack of information on regimes with low time scale. We are investigating SAR images from mission ENVISAT sensor ASAR acquired in Wide Swath Mode-WSM-) in the Mediterranean starting 1 March 2002 to 8 April 2012. The scenes are granted free of cost through the ESA 11849 project (Principal Investigator: Rosamaria Calaudi). SAR wind mapping is performed using the Johns Hopkins University, Applied Physics Laboratory (JHU/APL) software APL/NOAA SAR Wind Retrieval System (ANSWRS version 2.0) (Monaldo 2000; Monaldo et al. 2006). The following automated procedure is applied: the pixels are calibrated to obtain the Normalized Radar Cross-Section (NRCS) value, and then pixels are averaged to around 500 m, and finally the geophysical model function CMOD-5.n is used to retrieve the equivalent neutral wind speed using the wind direction a priori from the US Navy Operational Global Atmospheric Prediction System (NOGAPS). The spatial averaging of SAR pixels suppresses noise effects from longer ocean waves and from speckle, an inherent property of imaging radars. The 6-hourly model wind vectors are available at a 1 degree latitude and longitude grid and the wind vectors from the lowest model level around 10 m above the surface are used. To match the satellite data, the wind vectors are interpolated in time and space. The statistical analysis of SAR wind maps will be performed with the Satellite-Wind Atlas Analysis and Application Program (S-WAsP) tool developed by Technical University of Denmark, DTU Wind Energy Department. Here, we focus on a case study in Calabria, a long, narrow and mountainous peninsula in South Italy where breezes play a major role for the local climate causing wind direction variability coast to coast. We considered a 10m mast, measuring hourly wind speed and direction located at the coastline at the harbor of the town Crotone, belonging to the marine network of sensors of ISPRA (Institute for Environmental Protection and Research). For this especial case study a subset of 44 satellite images from ENVISAT-SAR is investigated. The scenes were selected to cover the area of interest; there are 36 images from around 8:50 UTC and 8 images from around 20:45 UTC. Three points were chosen at offshore distances of 4.5, 50 and 200 km and we performed a comparative analysis between wind data from SAR images, the experimental data, and data from the NOGAPS for the whole year 2009. Statistical analysis, shows that the correlation coefficient R2 of wind speed between SAR images and NOGAPS varies from R2=0.7 to R2=0.8 from 4.5 km to 50 km offshore respectively; between SAR and the mast, R2 varies from R2=0.4 to R2=0.6. R2 between NOGAPS data wind direction of two point chosen to 4,5 km offshore of two opposite sides of peninsulas, is 0.85. Around Calabria, from 4,5 to 50 km, R2=0.99, so this confirms, in this case, the limit of the NOGAPS horizontal resolution. Works are in progress to investigate other areas in Mediterranean and to study the accuracy of the ANSWRS 2.0 - SAR wind retrieval system, using as input RAMS mesoscale model at 10 km resolution during July August 2009 as a test. The methodology may be useful in feasibility studies for offshore or coastal sites where no suitable observations are available or as a guide to positioning of an offshore meteorological mast for wind resource estimation. This new technique is seen as a supplement to classical wind sampling and modeling efforts, not as a stand-alone alternative.