Long‐term trends of the ‘Koshava’ wind during the period 1949–2010

Abstract

In this study, a comprehensive analysis of long-term trends of the Koshava wind during the period between 1949 and 2010 is carried out. Koshava is a strong wind that blows from southeast quadrant over Serbia, Bulgaria and Romania. The Siberian high and West-Mediterranean cyclones, together with the orography of the eastern Balkan, are the main drivers of the Koshava wind. The trend analyses are performed on wind data sets from five synoptic weather stations, all situated in the region where the Koshava wind is fully developed. Koshava wind speeds are divided into two categories: (1) all wind speeds and (2) wind speeds above 5 ms −1 . Two homogeneity tests are used to inspect the quality of wind speed and wind direction time series. The Mann-Kendall test and Sen's slope estimator are used to analyze trends of the Koshava wind speeds and the annual number of days with the Koshava wind. Statistically significant negative trends of the Koshava wind speeds and wind activity are observed at all five weather stations and are more pronounced for wind speeds above 5 ms −1 . The negative trends of the Koshava wind are mostly related to the changes in the synoptic circulation, temperature and weakening of the Siberian high and West-Mediterranean cyclones. It is shown that the observed decline of the Koshava wind has a significant impact on reducing the wind energy potential in the region.

Homogeneity analysis of wind data from 213 m high Cabauw tower

“…In practice, only significant magnitudes of RDI indicate a break point. The RDI threshold value is 0.100, while pronounced break points are indicated with RDI values of 0.150 and higher (Petrović, ; Romanić et al, ). The inhomogeneities with RDI being very close to 0.100 need to be interpreted cautiously.…”

“…Inhomogeneities due to the instrumentation change are typically very pronounced (Petrović and Curley, ; Romanić et al, ). Second, the inhomogeneities due to the instrumentation are most distinct around the near‐calm wind speeds (Romanić et al, ). For that reason, this inhomogeneity should have appeared and be even stronger at the lower heights since the frequency of occurrence of small wind speeds increases with moving closer to the surface (Figure ).…”

“…This dependency is due to the cubic relationship between wind speed and wind energy. On an annual basis, the mean wind power density can be estimated as (Romanić et al, ): where P is the wind power density in W m −2 , ρ is the air density (1.22 kg m −3 ), A and k are the Weibull‐A (m s −1 ) and Weibull‐k parameters, respectively, and Γ is the gamma function defined as a definite integral: …”

“…In the situations when RDI is around 0.100, the random noise cannot be a priori excluded as a cause of ‘false inhomogeneity,’ but, at the same time, the earlier papers by Petrović (, ), Petrović and Curley () and Romanić et al . (), as well as the Monte‐Carlo analysis in Section 3 showed that the noise level in the RDI series is below 0.100. The changes of instrumentation should affect the data series regardless of instrument's position in space, which includes the wind sensor heights.…”

“…Analysing homogeneity of wind data requires separate investigations of wind speed and wind direction series. Inhomogeneities in wind speeds do not necessarily have to coincide with inhomogeneities in wind directions and vice versa (Romanić et al, ). For example, switching from 16 direction wind rose to 8 or 32 direction roses results in wind direction inhomogeneities, but it would not influence wind speeds.…”