Long-term changes in persistence of atmospheric circulation were examined for a large number of classifications of daily circulation patterns, both objective and subjective, collected within the COST 733 action. Long-term changes in the lifetime of circulation types are mostly insignificant, the types with positive and negative trends are approximately the same for most classifications. One of the few exceptions is the subjective catalogue of Hess and Brezowsky, in which the lifetime gets abruptly longer in the mid-1980s (Fig. 1). This feature of the Hess-Brezowsky catalogue, discussed recently in several papers, thus appears to be an artifact of the catalogue, likely resulting from its subjective nature (Cahynová and Huth, 2009a).

Fig. 1. Changes in the mean annual persistence (residence time) over all circulation types in selected subjective and objective circulation classifications. HBGWL - German subjective Hess-Brezowsky catalogue, PECZELY - Hungarian subjective catalogue, OGWL - "objectivized" Hess-Brezowsky, OGWL-3d+ - "objectivized" Hess-Brezowsky with a constraint of at least 3-day duration of synoptic situations. Years start in December in order to encompass the whole winter season.

The influence of long-term circulation changes on trends in surface climate elements was examined using multiple catalogues of circulation types for 11 climate elements in the Czech Republic. Two independent methods revealed that the circulation changes affect surface climatic trends in a considerable manner only for temperature in winter and autumn when the observed temperature trends can be explained by circulation changes from 20% in autumn and 30-50% in winter. Trends in other seasons and in other climate elements (e.g., precipitation, cloudiness, sunshine duration, wind components) are thus caused by changing climatic properties of individual circulation types (Cahynová and Huth, 2009b).

Relative drought indices (Standardized Precipitation Index [SPI], Palmer Drought Severity Index [PDSI]) were used to assess the present and future drought conditions in the Czech Republic (Dubrovský et al., 2009; Trnka et al., 2009). In the climate-change impact experiments, the future climate at 45 Czech stations is represented by modifying the station specific observed series according to climate change scenarios based on 5 Global Climate Models from the IPCC-TAR database. Changes in the SPI-based drought risk closely follow the modelled changes in precipitation, which is predicted to decrease in summer and increase in both winter and spring. Changes in the PDSI (which is considered to be superior to SPI in climate change impact studies as it accounts for the effects of both temperature and precipitation) indicate an increased drought risk at all stations under all climate-change scenarios (Fig. 2).

Fig. 2. Relative PDSI (Palmer drought severity index calibrated with weather data pooled from 45 Czech stations and the 1961-2000 period) in the present and future climates. The future climate projection is based on 5 GCMs (CSIRO-Mk2, CGCM2, GFDL-R30, HadCM3, CCSR/NIES) and relates to 2060-99 period assuming SRES-A2 emission scenario. The left panel shows the spectrum of drought conditions in the set of 45 stations in terms of average values of rPDSI (the negative values of PDSI indicate drier conditions than the present-climate all-station average). The right panel shows a percentage of rPDSI-based drought months in the 40-year series (Dubrovský et al., 2009).

Scenarios of changes in extreme precipitation were examined in 24 future climate runs of 10 regional climate models, focusing on the area of the Czech Republic (Kyselý and Beranová, 2009). The peaks-over-threshold analysis with increasing threshold censoring was applied to estimate multi-year return levels of daily rainfall amounts. The results showed that heavy precipitation events are likely to increase in severity in winter (Fig. 3) and (with less agreement among models) also in summer. The inter-model and intra-model variability and related uncertainties in the pattern and magnitude of the change are large, but the scenarios tend to agree with precipitation trends recently observed in the area, which may strengthen their credibility. In most scenario runs, the projected change in extreme precipitation in summer is of the opposite sign than a change in mean seasonal totals, the latter pointing towards generally drier conditions. A combination of enhanced heavy precipitation amounts and reduced water infiltration capabilities of a dry soil may severely increase peak river discharges and flood-related risks.

Fig. 3. Relative changes (in %) of 50-yr return values of daily precipitation amounts between the late 21st century scenarios and control climate (1961-1990) in winter (DJF) in 10 regional climate models. A2/B2 denotes the SRES emission scenarios; a driving global model is given in the heading of panels if different from the HadAM/HadCM model. Larger (smaller) crosses indicate gridboxes in which the estimated 90% (80%) confidence intervals of the 50-yr return values do not overlap, i.e. the change is statistically significant approximately at the 0.01 (0.07) level.