Variability of geostrophic airflow over Poland, 1951-2014

Abstract The paper presents the analysis of the anemological conditions variability over Poland with the usage of geostrophic wind vector as an objective (and homogenous) information concerning the airflow over the area of research. The geostrophic wind vector components are calculated using SLP and air temperature (at sigma 995 level) at selected gridpoints which were subsequently interpolated to a central point thus describing the average flow over the research area. The data originated from NCEP/NCAR Reanalysis and its temporal range was 1951-2014. The analysis covers statistical characteristics of the overall annual cycle as well as trend analysis of the airflow features over Poland: geostrophic wind vector module (V), and its zonal (u) and meridional (v) components. Aside from general statistical characteristics for averages and extremes (quantiles 10% and 90%) GEV distribution was fitted to maximum annual/monthly geostrophic wind speed values which allowed the estimation of return levels for selected return periods. For the period 1951-2014 average geostrophic wind velocity over Poland equals 7.4 ms−1 and the 99% quantile exceeds 21 ms−1. Maximum speed ever recorded equalled 37.6 ms−1. Geostrophic wind vector module (V) and its components (u, v) exhibit clear annual cycle with the highest V values in winter. Positive (westerly) u values dominate in the colder part of the year. In spring the dominance of eastern advection appears and in summer the prevalence of westerly flow is only minimal. There exists a distinctive variability of decadal directional structure and this is clearly visible in the substantial increase in the share of western sector frequencies in 1981-1990 and following decade. Monthly V averages do not exhibit (except October) statistically significant trends whereas in spring and summer months as well as for annual averages of u component trend is significant. There are virtually no significant changes in the v values. GEV analysis allowed the year to be divided into two parts. Warm one with relatively low return levels – for many months not exceeding 20 ms−1 even for 50y return period. On the other hand winter months return level values exceed 30 ms−1 even for relatively short return periods (20y) with upper estimates for 100y return period closing to 40 ms−1.