To achieve this, wind data from local wind measurements at meteorological stations in operation over the past 6 years and installed in different sites of the region were analyzed.
In a second stage, the NCEP/NCAR reanalysis data base for that region was validated by comparison with the local measurements and the last 20 years were used for the characterization of the inter-annual variability of the mean wind velocity.
As expected, the data from local wind measurements indicate that the maximum deviation of the mean wind speed decreases continually with the increasing of the period used in their determination, although at a decreasing rate.
The use of reanalysis data for the period of the measurements made in the different sites showed that data from the NCEP/NCAR represent in an acceptable way the mean inter-annual variability for the measurement sites tested.
The findings point out that the industry standard figures for annual wind variability can be conservative for the Dobrogea region. This will have a direct impact on pre-construction estimates of wind farm production and on financing and debt rising for the projects.
For the Dobrogea region, the studied data pointed out that the inter-annual variability of mean annual wind speed could be much lower than typical wind industry standard figures.
Moreover, the evolution of inter-annual variability with the number of consecutive years of data appears slightly less conservative than the evolution typically assumed by the industry.
The test case presented highlighted the significant drop in the overall uncertainty on wind farm energy yield estimates between 1 and 2 years of observed wind data.
This study should be repeated for other regions with wind farm developments, quite specially those more distant from Central Europe/UK (where the industry standard practices were primarily derived). NCEP/NCAR reanalysis can be a very useful tool for this kind of studies (being freely available and global) but its use for inter-annual wind variability should be tested in each case.