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.

Koshava is a local wind usually observed in the cold part of the year over the large part of Serbia, parts of Romania, parts of Hungary, and east Croatia. The main characteristics of the Koshava wind are its high wind speed, southeasterly direction, persistence, and gustiness. This paper analyses the synoptic [Mediterranean cyclones (MCs) and Eurasian anticyclones (EAs)] as well as the mesoscale contributors to the Koshava wind. Time and spatial features of the pressure systems have been analysed using an automatic cyclone tracking scheme. The results show that the MCs and EAs that generate Koshava are approximately 1000 km away from the Koshava region. It is demonstrated that the strong anticyclones are the main trigger for the Koshava wind. Less than 3% of the Koshava winds occurred without the occurrence of either an anticyclone or a cyclone. The Synoptic Koshava Index (SKI), developed in this study and defined as the difference between the area-averaged mean sea level pressures in the anticyclone and cyclone regions, has been shown to be a good indicator of the Koshava occurrence. Koshava has also been investigated from the perspective of the gap flow windstorm. The across-mountain pressure difference, ΔP, is the most important mesoscale contributor to the Koshava characteristics. The across-mountain potential temperature difference, Δ , together with ΔP, can be successfully used to predict Koshava's occurrence. A simple (linear) probabilistic model for forecasting the Koshava mean hourly wind speed has been constructed. The results demonstrate that the occurrence of the Koshava wind can be predicted with the significantly higher accuracy than the Koshava mean hourly speed. The correlation between Koshava and winds at higher levels is small.

An extreme Koshava episode (EKE) from 30 January to 4 February 2014 has been studied. Koshava is a local windstorm in Southeast Europe. EKE was characterized by wind gusts above 45 m s −1 and deep snowdrifts. Strong Eurasian anticyclone in combination with large temperature gradient between the anticyclone region and the Mediterranean caused the EKE. The anticyclone had the probability of occurrence of 0.1%. Koshava layer was either statically stable or adiabatic. The event was numerically modelled using two mesoscale models: WRF-NMM and NMMB. Wind directions were forecasted more accurately than the mean speed and gusts.