Wind resource assessment through WRF mesoscale model and FUROW: Madrid region´s wind map

This work aims to create a wind map at 80 m height of the Community of Madrid using FUROW, the most complete wind software of the market developed by SOLUTE Ingenieros. For this purpose, the wind speed and direction data at mesoscale size has been simulated using the Weather Research and Forecasting (WRF) model, and FUROW has been used to model the microscale wind field as the model is able to use mesoscale/reanalysis information to generate wind maps at a much finer resolution.


Madrid region´s wind map at 80 metres height

In regards to the WRF model set up, it is made up by 3 domains with horizontal resolutions of 72, 24 and 8 kilometers respectively and related between them with a nesting. The outermost domain is centered on Madrid and has a size of approximately 2016 x 2016 km2. The topographic data are obtained from the U.S. Geological Survey (USGS) global elevation dataset using 1 arc-min for the outermost domain and 30 arc-s for the others. The configuration has been set to have 50 terrain-following hydrostatic pressure levels in the vertical for all domains, with the top level being located at 50 hPa and with a bigger concentration of levels in the range of interest between 0-200 meters of height. The longwave and shortwave radiation schemes are based on Dudhia (1989). Moreover, a new Simplified Arakawa-Schubert (SAS) scheme based on Han and Pan (2011) is employed for the cumulus parameterization of the outermost domain. With regard to the planetary boundary layer, the Yonsei University (YSU) parametrization developed by Hong et al. (2006) has been used. In relation with microphysics, it has been chosen a scheme based on Rogers et al (2001) named as Ferrier microphysics. Finally, a simple five-layer land surface model based on Dudhia (1996) is used.

Initial and boundary conditions were obtained from the NCEP FNL (Final) Operational Global Analysis data, which are on 1-degree by 1-degree grids and provided every six hours. This product is from the Global Data Assimilation System (GDAS), which continuously collects observational data from the Global Telecommunications System (GTS). The time period used to carry out the simulations with the WRF model was the year 2013. Output time step of wind speed and direction is 1 hour.

Furthermore, once wind speed and direction data have been simulated and the frequencies distributions have been calculated, a long term adjustment has been performed. In this work, the 2013 wind data has been corrected using 20 years (01/01/1996 – 01/01/2016) of the global model MERRA-2 provided by NASA. In this case, this function becomes crucial due to the fact that the average wind speed for the year 2013 may differ from the average for the period of the 20 years considered.

As it was previously mentioned, the final objective is to create a wind map at 80m high of the Community of Madrid, consequently it has been delimited by the next coordinates: -4.625º West, -2.875º East, 41.2º North and 39.8º South. What is more, this global wind map is divided in 225 grids of 15 km of horizontal resolution approximately, in order to create 225 wind maps which were rejoined at the end to compose the global wind map.

FUROW has a wind resource calculator that allows the user generating wind maps of a given area at microscale resolution (for this case a horizontal resolution of 200m has been used) based on a site climatology and topographic maps (for this case orography was derived from SRTM dataset and roughness from CLC dataset). Due to the limitations of microscale models and the need of capturing wind spatial variations, it is essential to use the WRF model in order to have wind data with a resolution similar to the size of each grid. The use of global models (with much coarser resolution) such as GFS, ERA-Interim, MERRA, etc. is not very appropriate as they do not consider mesoscale phenomena. Consequently, wind speed and direction were obtained for the center of each grid from the simulation outputs of WRF. Finally, after the wind resource calculation for all the grids was performed, an interpolation has been made for the overlapping areas in order to moderate the transitions between the different grids.

For the sake of verification, wind data from several weather stations situated in Rivas (Madrid), Guadalajara at 10m height and Ledanca (Guadalajara) at 50m height have been used. The observed data was corrected using the 20 years of MERRA-2 as commented before, obtaining 2.35m/s, 1.90m/s and 5.14m/s in Rivas, Guadalajara and Ledanca respectively. The wind map at 10m high obtained by FUROW and based on WRF simulations shows values of 2.65m/s and 2.10m/s for Rivas and Guadalajara (being the difference less than 0.3m/s), whereas the wind map at 50m high shows 5.58m/s for Ledanca, making the difference increase up to 0.44m/s. Differences of the first two sites can be explained because measurements are at peri-urban areas and at rooftops, where building effects may be significant. For the third site, it is located near a very steep cliff, so possible effects due to topography may not being grasped accurately.

As a conclusion of this work it can be stated that the procedure developed by SOLUTE Ingenieros can be used to estimate the wind resource at a given area when no measurements are available. Further improvements such as increasing WRF resolution and considering more years of simulation may help obtaining a more accurate map.


Screenshots of the most windy areas in the region
 

Blog tags: