DECADAL VARIABILITY OF THE DANUBE RIVER FLOW IN THE LOWER BASIN AND ITS RELATION WITH THE NORTH ATLANTIC OSCILLATION

Transcription

1 INTERNATIONAL JOURNAL OF CLIMATOLOGY Int. J. Climatol. 22: (2002) Published online in Wiley InterScience ( DOI: /joc.788 DECADAL VARIABILITY OF THE DANUBE RIVER FLOW IN THE LOWER BASIN AND ITS RELATION WITH THE NORTH ATLANTIC OSCILLATION NOREL RÎMBU, a CONSTANŢA BORONEANŢ, b, *CARMENBUŢĂ b and MIHAI DIMA a a Bucharest University, Faculty of Physics, Department of Atmospherics, Bucharest, Romania b National Institute of Meteorology and Hydrology, Bucharest, Romania Received 20 February 2001 Revised 5 February 2002 Accepted 12 February 2002 ABSTRACT The decadal variability (>5 years) of the Danube river flow in the lower basin and its connection with the North Atlantic Oscillation (NAO) is analysed for the period Associated linkages with precipitation (PP) in the European sector, global sea surface temperature (SST) and atmospheric circulation for the period , and the 500 hpa geopotential heights (G500) over the Northern Hemisphere for the period are also investigated. The results show that there is an out-of-phase relationship between the time series of the Danube river flow anomalies and the NAO index. The time series of a PP index, defined as the average of normalized precipitation anomalies over a large area including the Danube basin, presents a time evolution similar to that of the river flow anomalies. The correlation maps between the river flow anomalies and global SST show coherent large-scale patterns. High values of the Danube river flow are associated with a tripole-like SST structure in the North Atlantic similar to that appearing during the negative phase of the NAO, and with negative SST anomalies in the central North Pacific and positive SST anomalies in the eastern and central tropical Pacific. Physically consistent sea level pressure and 500 hpa geopotential height are obtained. Copyright 2002 Royal Meteorological Society. KEY WORDS: Europe; Danube river lower basin; correlation; composite maps; decadal variability; North Atlantic Oscillation; precipitation 1. INTRODUCTION Research studies show that climate variations influence many components of the climate system. It has been widely recognized that changes in the cycling of water between land, sea and air can potentially have significant impacts on the environment and on many sectors of the economy and society through their effects on water resources and their management. (Arnell, 1995, 1999; Arnell and Reynard, 1996). Evidence from long hydrological records shows that periods with anomalous hydrological behaviour (Arnell et al., 1993) are associated with persistent climatic anomalies. Interannual to decadal variability of the atmosphere over the North Atlantic region is characterized by the North Atlantic Oscillation (NAO) teleconnection pattern (Bjerknes, 1964; Hurrell, 1995). The NAO is a fluctuation in the pressure gradient across the North Atlantic with centres of action being the Icelandic low and the Azores high. It is the dominant mode of atmospheric behaviour in the North Atlantic throughout the year, mostly pronounced during winter and a primary climatic factor orchestrating hemispheric-scale climatic fluctuations centred over the Atlantic. Research studies focused on the influence of the NAO on the variability of various climatic elements at time scales ranging from seasonal to interannual and decadal show that the NAO is responsible for generating systematic, large-amplitude patterns in the anomalies of temperature, precipitation, wind speed, latent and * Correspondence to: Constanţa Boroneanţ, National Institute of Meteorology and Hydrology, Sos. Bucureşti-Ploiesti 97, Bucharest, Romania; boroneant@meteo.inmh.ro Copyright 2002 Royal Meteorological Society

2 1170 N. RÎMBU ET AL. sensible heat fluxes, and hence sea surface temperature (SST), over much of the extra-tropical North Atlantic (van Loon and Rogers, 1978; Kushnir, 1994; Hurrell and van Loon, 1997). The NAO has also been associated with the path and intensity of Atlantic storm tracks and with evaporation and precipitation patterns in the Atlantic-European region (Hurrell, 1995). Because the signature of the NAO is strongly regional, a simple index of the NAO has been defined as the difference of the standardized sea level pressure (SLP) anomaly measured at Lisbon, Portugal, and at Stykkisholmur, Iceland (Hurrell, 1995). Correlations with hydrological data have shown that, when the NAO index is high, river flow (particularly in winter) is above average in northern Europe and below average in southern Europe (Shorthouse and Arnell, 1997; Dettinger and Diaz, 2000). Because river flows depend directly on precipitation, it is evident that there is a linkage between precipitation anomalies associated with extreme phases of the NAO and river flow regimes in Europe. At the decadal time scale (>5 years) the NAO strongly influences the moisture balance (i.e. precipitation minus evaporation) over Europe. During the positive phase of the NAO (deep Icelandic low and strong Azores high), enhanced precipitation over northern Europe associated with less precipitation over central and southern Europe occurs (Hurrell, 1995; Hurrell and van Loon, 1997; Rîmbu et al., 2001). A reverse situation occurs during the negative phase of the NAO. The Danube river has a very large catchment basin extending from central Europe (upper basin) to southeastern Europe (lower basin). Its flow regime and other hydrologic characteristics are subject to significant influences due to climate variability (Bondar and Buţă, 1995). Based on existing observational data in the Danube river basin, many research studies have pointed out the effect of precipitation and temperature changes on the Danube flow regime (Starosolszky and Gauzer, 1998), on the possible climate impacts on the water resources in the Danube river basin (Behr, 1998; Petrovic, 1998), and on changes in hydrological characteristics for selected river basins in the case of climate-change scenarios (Dvorak et al., 1997; Stănescu et al., 1998). Because the decadal variability in the NAO has become especially pronounced since the early 1970s and has determined decade-long regional climatic anomalies (winter dry conditions over southern Europe and Mediterranean, and wet anomalies from Iceland eastward through Scandinavia (Zorita et al., 1992; Wilby et al., 1997; Werrity and Foster, 1998)), we expect that the NAO signal might also be detected in the decadal variations of the Danube river flow in the lower basin. The goal of the present study is to investigate the role of the NAO on decadal variability of the Danube river flow in the lower basin and its connections with atmospheric circulation and SST at a global scale based on observed data. It is important to study the linkage between the climate variability and hydrological regime in order to understand the physical mechanisms that determine it and to help in guiding the development of policies for mitigation or prevention of drastic effects of long-term climate variability on various ecological and socio-economical sectors. The paper is organized as follows. After the Introduction a brief description of the principal characteristics of the Danube river basin follows in Section 2. The data and methods are described briefly in Section 3. The decadal variability of river flow and its relation with the NAO and precipitation is presented in Section 4. Section 5 presents the decadal variations of Danube flow in the lower basin in connection with global SST, SLP and 500 hpa geopotential height (G500). A summary of the results and conclusions are presented in Section SOME CHARACTERISTICS OF THE DANUBE RIVER BASIN The Danube river is the second longest in Europe, cutting across the territories of many countries. It has a total length of 2870 km and a catchment area of about km 2. From the Black Forest Mountains in Germany to the Black Sea in Romania, the Danube flows through 13 countries. It is the largest transboundary river basin in Europe. The river serves multiple functions as a waterway, natural resource and source of energy, and also plays an important role in the ecological balance of the region. The Danube represents a natural resource of water for industry, agriculture, domestic water and groundwater supply.

3 CLIMATIC CONTROL OF RIVER DANUBE FLOW 1171 The Danube river basin can be divided into three sub-regions (Figure 1): the upper, the central and the lower basin. The upper basin extends from the source (Germany) to Bratislava, Slovak Republic. The central basin is the largest and comprises the section from Bratislava to the Iron Gates dam (the former Yugoslavia and Romania). The lower basin is formed by the Romanian Bulgarian lowlands and its upland plateau and mountains. The climate of the Danube basin is very diverse. There is an influence of Atlantic climate in the western part of the upper basin, a Mediterranean influence in the southern part of central and lower basin, and a continental climate elsewhere. The Danube river flow is determined principally by the precipitation and evaporation processes from the Danube catchment basin. The mean quantity of precipitation that falls in a certainareaof the Danube catchment basin is strongly dependent on the orography. Because one-third of the basin is formed by mountains and the remainder consists of hills and plains, the annual precipitation total ranges from about 2000 mm per year in the high regions to only about 500 mm per year in the plains. Evaporation is also important for the water balance in the Danube catchment basin, especially in the lower regions where the mean annual evaporation varies between 450 and 650 mm per year. 3. DATA AND METHODS 3.1. Data The primary quantity analysed in this study is the annual mean of the Danube s discharge. The time series of annual mean flows were calculated from monthly mean flows of the Danube river measured at six hydrological stations located in the lower basin, on the Romanian border (Figure 1). These data were provided by the National Institute for Meteorology and Hydrology (NIMH), Bucharest, Romania. The names of the stations, their locations (latitude and longitude) and some simple statistical characteristics of the river flow for the period are presented in Table I. The decadal variations of the Danube s flow are strongly related to decadal variations of precipitation over the river catchment area. The monthly precipitation data set (PP) for global land areas, with a spatial resolution of 5 latitude 5 longitude constructed by Hulme (1992, 1994), has been used. From the period PP data set, the period has been extracted for the present study. Figure 1. The Danube river catchment basin and the locations of the Romanian hydrological stations from the lower basin used in this study

4 1172 N. RÎMBU ET AL. Table I. The hydrological stations from the Danube river lower basin, their coordinates and some simple statistical characteristics Station Coordinates Annual mean Standard deviation (m 3 s 1 ) Variance a Coefficient of discharge (%) variability Original Decadal b (m 3 s 1 ) Orşova E, N Corabia E, N Turnu Măgurele E, N Giurgiu E, N Călăraşi E, N Ceatal Izmail E, N a Decadal/original. b (Decadal Standard deviation/annual mean) 100. The SST data set is extracted from the global analyses of Kaplan et al. (1997, 1998, 2000) derived from in situ data using a statistical method known as optimal smoothing (OS). The spatial resolution of the SST data set is 5 latitude 5 longitude. SLP over the ocean (resolution 4 latitude 4 longitude) was extracted from the Kaplan data set representing reduced space optimal interpolation of the global SLP anomalies from the Comprehensive Ocean Atmosphere Data Set (COADS). The anomalies were calculated with respect to the climatological annual cycle estimated from the COADS data for the period (Kaplan et al., 2000). From the Kaplan data set for the period , both for SST and SLP, we have selected the period The 500 hpa geopotential heights (G500) were taken from the National Centers for Environmental Production National Center for Atmospheric Research (NCEP/NCAR) reanalyses data set (Kalnay et al., 1996). The horizontal resolution of the G500 field is 2.5 latitude 2.5 longitude and the period analysed is Methods All data were processed in the same way. First, annual means were calculated from monthly means. Then, annual anomalies with respect to the mean and normalized by local standard deviation estimated for the period were produced. The annual normalized anomalies were smoothed with a 5 year running mean filter to obtain the decadal component of the series. To explain the decadal variability of the river s flow we have drawn composite maps. All maps corresponding to the times when the normalized flow anomaly was lower (higher) than one standard deviation were averaged. The map of the difference between high-flow and low-flow averaged maps was used to identify the large-scale precipitation and atmospheric circulation anomalies associated with decadal variability of the Danube s flow in its lower basin. Correlation maps between the time series of decadal anomalies of flow and the global SST and SLP have been produced to investigate possible coherent large-scale connections. 4. RELATION WITH PRECIPITATION AND THE NAO In this section we analyse the relation between decadal variations of the Danube river flow and decadal variations of precipitation in the river catchment basin. To obtain a quantitative measure of the strength of decadal (>5 year period) variations of the flow, the variances for both the original (annual means) and the decadal time series of the annual mean flow measured at selected hydrological stations have been calculated. The ratio between the decadal and original variance

5 CLIMATIC CONTROL OF RIVER DANUBE FLOW 1173 for each station is presented in Table I. Consistent with partition of variance between interannual and decadal time scales of river flows in Europe (Dettinger and Diaz, 2000), more than 30% of the mean annual flow variability is contained in the decadal component at all selected stations in the Danube lower basin. The time series of annual means of the Danube s discharge recorded at selected hydrological stations during the period are presented in Figure 2 (thin line). The solid line represents the decadal component obtained by smoothing the series with a 5 year running mean filter. A simple visual inspection of these time series shows that the decadal variations at all stations are quite similar. A decreasing trend is evident at all stations during the last two decades. The correlation coefficients between the decadal flow time series at Ceatal Izmail and the remaining five stations from the Danube lower basin have been calculated and are presented in Table II. Because the six time Figure 2. The time series of annual means of the Danube river flow in the lower basin (thin line) measured at hydrological stations presented in Table I for the period The solid line represents the decadal component obtained by smoothing the series with a 5 year running mean filter

6 1174 N. RÎMBU ET AL. Table II. The linear correlation coefficients between the time series of annual flow at Ceatal Izmail station and five hydrological stations from the Danube river lower basin Correlation coefficient Orşova Corabia Tr. Măgurele Giurgiu Călăraşi Ceatal Izmail series of the Danube decadal streamflow are highly correlated with each other (the correlation coefficients are greater than 0.91), for simplicity, in the following we consider only the time series of flow at Ceatal Izmail as representative for the Danube river lower basin and have referred it as the Danube flow. The flow behaviour has been analysed in connection with the precipitation field and the NAO. The decadal flow variations have been analysed in connection with decadal variations of precipitation (PP) in the Danube catchment basin. A composite PP map has been constructed to identify the PP patterns associated with decadal flow variations (Figure 3). This map shows that high values of precipitation over the Danube basin catchment occur in association with high values of Danube flow. It is interesting to note that higher precipitation anomalies are located in the upper and central basins, which are the regions with the main tributaries to the total Danube flow in the lower basin. Accordingly, the decadal variability of flow in the lower basin reflects mainly the decadal variability of precipitation from the upper and central catchment basins. Many research studies have shown that decadal precipitation variability over Europe is related to the decadal variability of the NAO (Hurrell, 1995; Rîmbu et al., 2001). Therefore, the NAO signal should be present in the European river flow time series. Our goal was to search for the NAO influence on the decadal variability of the Danube river flow in its lower basin. Based on the PP pattern presented in Figure 3, a PP index has been defined by averaging the normalized PP anomalies from the region (40 55 N; 5 35 E). This domain includes the whole Danube river catchment basin. The correlation coefficient between the time series of normalized river flow anomalies and the time Figure 3. The composite map of precipitation in the European sector based on the decadal component of the Danube river flow. The map was obtained as the difference between the averaged maps of positive and negative composites. The composites were constructed from PP maps for which the decadal normalized anomalies of river flow were higher/lower than one standard deviation. Contour interval is 30 mm

7 CLIMATIC CONTROL OF RIVER DANUBE FLOW 1175 series of PP index is The high value of the correlation coefficient between PP index and flow suggests that decadal PP variability is dominant compared with evaporation variability over this region at the decadal time scale. This is also consistent with the values of coefficient of variability (i.e. standard deviation/mean 100) of river flow (about 10%) and of precipitation over the Danube river catchment basin (about 7%). In Figure 4 we present the decadal time series of the NAO index (thin line), the normalized Danube river flow anomalies (solid line) and the PP index (dotted line). It is evident from Figure 4 that there is an out-of-phase relationship between the Danube flow and the NAO index. The correlation coefficient between these time series is 0.75, consistent with other studies (Shorthouse and Arnell, 1999; Dettinger and Diaz, 2000) that showed that the river flows tend to be lower than normal in central and southern Europe and higher than normal in northernmost Europe when the NAO is in its positive phase. As Figure 4 shows, since 1970 the NAO index has been on an upward trend while the Danube s flow was continuously decreasing. A simple visual inspection shows that the time series of Danube flow and the PP index present similar decadal variations, with precipitation leading the river flow. The cross-correlation functions between the monthly means of the Danube river flow anomalies and the PP and NAO indices (not shown) reveals that precipitation variations are in phase with the NAO index variations and lead the river flow variations by 2 3 months. This time delay between precipitation (or NAO) and streamflow variations is small compared with the time scale considered in our analysis. 5. GLOBAL CONNECTIONS OF RIVER FLOW 5.1. Relations with global SST According to the results presented in the previous sections, the decadal variations of Danube s discharge can be related to the NAO to the extent in which the NAO controls the precipitation regime in the Atlantic European sector (Hurrell and van Loon, 1997; Rîmbu et al., 2001). Although our analysis was focused on the NAO influence on the Danube s flow in the lower basin, we looked at some atmospheric and ocean field associations at the global scale. Several studies have established that large-scale SST fluctuations can be linked to atmospheric circulations that produce precipitation fluctuations (Dai et al., 1997; Latif et al., 2000). One of the most widely studied Figure 4. The time series of normalized anomalies of the Danube s discharge at Ceatal Izmail station (solid line), NAO index (thin line), PP index (dashed line). All time series were normalized and smoothed with a 5 year running mean filter

8 1176 N. RÎMBU ET AL. phenomenon is El Niño Southern Oscillation (ENSO), which generates coherent anomaly patterns of temperature and precipitation in regions all over the globe (Dai and Wigley, 2000). However, the direct impact of ENSO on the North Atlantic and Europe appears to be weak. Rajagopalan et al. (1998) have reported observational evidence of significant coherence at decadal time scales, with no phase lag, between tropical South Atlantic SSTs and the NAO, with warm SSTs associated with the positive phase of the NAO. The correlation map between the time series of the Danube river flow and global SST represented in Figure 5(a) emphasizes coherent large-scale patterns. In the North Atlantic a tripole-like SST pattern similar to the SST pattern associated with the negative phase of the NAO (Hurrell, 1995) appears in connection with positive anomalies of the Danube river flow. The centres of the tripole anomaly SST pattern appear to be displaced towards the eastern coast of North America compared with the SST pattern in the North Atlantic associated with the NAO. Higher than average values of the Danube river flow tend to be associated with cooler than average SST anomalies in the central North Pacific and warmer than average SSTs along the western coast of North America, eastern and central tropical Pacific. Our results are compatible with the negative correlation between North Pacific SSTs and river flows throughout North America, tropical Africa, central and southern Europe reported by Dettinger and Diaz (2000). Although the Pacific SST pattern associated with the Danube flow decadal variability bears some resemblance to the SST pattern characterizing the Pacific Decadal Oscillation (PDO) (Mantua et al., 1997), the shift from positive (negative) to negative (positive) phase of PDO from 1948 (1977) is not clearly evident in the variability of the Danube river flow (Figure 2). This suggests that although the North Pacific SSTs exhibit coherent patterns similar to the PDO, their direct influence on decadal variations of the Danube river flow are smaller than the North Atlantic processes (i.e. the NAO) that control European hydrological behaviour. a) b) Figure 5. (a) The correlation map between the time series of decadal component of the Danube river flow measured at Ceatal Izmail station and SST. (b) As in (a), but for SLP

9 CLIMATIC CONTROL OF RIVER DANUBE FLOW Connection with global SLP The correlation map between the time series of the Danube river flow and SLP over the North Atlantic (Figure 5(b)) emphasizes coherent large-scale SLP patterns consistent with the corresponding SST patterns (Figure 5(a)). Over the North Atlantic, high values of flow are related to a dipole-like pattern of SLP anomalies similar to that corresponding to negative phase of the NAO, and are consistent with the negative correlation between the NAO index and the Danube flow. In agreement with SST anomalies, the central North Pacific is dominated by negative SLP anomalies. It is evident from Figure 5(b) that the SLP anomaly pattern over the central and eastern Pacific does not present an out-of-phase relation with the SLP anomalies over the western tropical Pacific, as in the case of the interannual ENSO phenomenon. This pattern might suggest that the SST and SLP anomalies from the tropical Pacific are not generated through the atmosphere ocean interaction processes that produce the typical interannual ENSO, and that their impact on the decadal variability of the Danube river is small. However, the coherent patterns presented in Figure 5 might be the result of a superposition of different decadal modes of climate variability, such as the NAO (Hurrell, 1995) or PDO (Mantua et al., 1997). The mechanisms by which the correlations are established are still uncertain, because the way in which such decadal modes interact is not yet clarified (Dettinger and Diaz, 2000) Connection with Northern Hemisphere G500 To assess better and confirm the links between the decadal Danube flow and large-scale SSTs and atmospheric circulation patterns described in the previous sections, we have constructed a composite map of annual 500 hpa geopotential heights. The large-scale atmospheric circulation patterns associated with decadal variability of precipitation over Europe, which influence the decadal variability of the Danube river flow are clearly emphasized in this field. The composite map of G500 (Figure 6) emphasizes a dipole-like pattern in the North Atlantic that is characteristic of the negative phase of the NAO, and a large area of negative G500 anomalies in the North Pacific that are consistent with the corresponding SLP patterns presented in Figure 5(b). The G500 variations corresponding to the difference between the high values and low values of the Danube river flow are higher in the North Atlantic than the corresponding variations over the North Pacific. This confirms our supposition that the North Atlantic processes (i.e. the NAO) play the principal role in generating the Danube flow decadal variations. Low values of G500 over central and southern Europe associated with high values of the Danube river flow are consistent with the increasing of precipitation over a large area that includes the entire Danube river catchment basin (Rîmbu et al., 2001). 6. SUMMARY In this study we have analysed the decadal variability of the Danube river flow in its lower basin using the flow records from six hydrological stations on the Romanian border. The decadal variations dominate the year-to-year Danube flow variations at all stations analysed. Because all six time series of the annual mean of the Danube flow are highly correlated and present a similar time evolution, we considered only the river flow at Ceatal Izmail station in our analysis. The decadal flow variability was analysed in connection with land precipitation over Europe and the NAO. The composite PP map showed that high values of PP anomalies throughout the river catchment occur in association with high values of the Danube river flow anomalies. As expected, the decadal variations of river flow are in good agreement with the decadal variations of precipitation in the Danube catchment basin, being largely controlled by North Atlantic processes, particularly by the NAO. Decadal NAO variability is reflected in flow variability as an out-of-phase relationship, in the sense that the Danube river flow in the lower basin tends to be lower than normal when the NAO is in its positive phase, and vice versa. The correlation map between the Danube river flow and global SSTs shows that higher than average values of river flow tend to be associated with a tripole-like SST pattern in the North Atlantic, similar to the SST

10 1178 N. RÎMBU ET AL. Figure 6. The composite map of Northern Hemisphere geopotential height at 500 mbar (G500) based on the time series of the decadal Danube river flow. Solid line corresponds to high flow and dashed line to low flow. The regions where the difference between the composite G500 maps corresponding to high flow and low flow are higher (lower) than 10 gpm are shaded (light shaded). The contour interval is 10 dam pattern associated with the negative phase of the NAO, and with cooler than average SST anomalies in the central North Pacific and warmer than average SSTs along the western coast of North America, eastern and central tropical Pacific. The corresponding SLP correlation map emphasizes a dipole-like pattern of SLP anomalies over the North Atlantic, similar to that characterizing the negative phase of the NAO, and a large area of negative SLP anomalies in the North Pacific. Consistent with the global coherent patterns in the fields of SST and SLP, the composite map of geopotential heights at 500 hpa confirms the dominant role of the NAO in generating large-scale atmospheric circulation patterns and its associated precipitation modes that influence the hydrologic fluxes in Europe. The composite and correlation patterns presented in this paper demonstrate that decadal variations of the Danube s flow respond to climate forcings on nearly global scales, just as does the decadal precipitation variability over Europe. ACKNOWLEDGEMENTS The authors would like to thank Dr Glenn R. McGregor and the anonymous referees for their helpful comments and suggestions. N. Rîmbu and M. Dima thank Dr Gerrit Lohmann and Dr Ute Merkel for fruitful discussions during their visit at the Max Plank Institute in Hamburg that helped to improve this paper. REFERENCES Arnell NW Scenarios for hydrological climate change impact studies. In The Role of Water and the Hydrological Cycle in Global Change, Oliver HR, Oliver S (eds). NATO ASI Series 1, vol. 31. Springer-Verlag: Berlin, Heidelberg; Arnell NW The effect of climate change on hydrological regimes in Europe: a continental perspective. Global Environmental Change 9: Arnell NW, Reynard NS The effects of climate change due to global warming on river flows in Great Britain. Journal of Hydrology 183: Arnell NW, Krasovskaia I, Gottschalk L River flow regimes in Europe. In Flow Regimes from International Experimental and Network Data (FRIEND), vol. 1. Hydrological Studies, Gustard A (ed.). Institute of Hydrology: Wallingford, Oxfordshire, UK; Behr O Possible climate impacts on the water resources of the Danube river basin. In Proceedings of the Second International Conference on Climate and Water, Espoo, Finland, August; Bjerknes J Atlantic air sea interactions. Advances in Geophysics, vol. 10. Academic Press: 1 82.

11 CLIMATIC CONTROL OF RIVER DANUBE FLOW 1179 Bondar C, Buţă C Trends of water discharges, sediment discharges and salinity of Danube in the Romanian sector. Romanian Journal of Hydrology and Water Resources 2: Dai A, Wigley TML Global patterns of ENSO-induced precipitation. Geophysical Research Letters 27(9): Dai AG, Fung IY, DelGenio AD Surface observed global land precipitation variations during Journal of Climate 10: Dettinger MD, Diaz HF Global characteristics of stream flow seasonality and variability. Journal of Hydrometeorology 1: Dvorak V, Hladny J, Kasparek L Climate change hydrology and water resources impact and adaptation for selected river basins in the Czech Republic. Climatic Change 36: Hulme M A global land precipitation climatology for the evaluation of general circulation models. Climate Dynamics 7: Hulme M Validation of large-scale precipitation fields in general circulation models. Global Precipitation and Climate Changes, Desbois M, Desalmand F (eds). Springer-Verlag: Hurrell J Decadal trends in the North Atlantic oscillation: regional temperatures and precipitation. Science 269: Hurrell JW, van Loon H Decadal variations in climate associated with the North Atlantic oscillation. Climatic Change 36: Kalnay EM, Kanamitsu M, Kistler R, Collins W, Deaven D, Gandin L, Iredell M, Saha S, White G, Woollen J, Zhu Y, Chelliah M, Ebisuzaki W, Higgins W, Janowiak J, Mo KC, Ropelewski C, Wang J, Leetmaa A, Reynolds R, Jenne R, Joseph D The NCEP/NCAR 40-year reanalysis project. Bulletin of the American Meteorological Society 77: Kaplan A, Kushnir Y, Cane MA, Blumenthal MB Reduced space optimal analyses for historical data sets: 136 years of Atlantic sea surface temperatures. Journal of Geophysical Research 102: Kaplan A, Cane MA, Kushnir Y, Clement AC, Blumenthal MB, Rajagopolan B Analysis of global sea surface temperature Journal of Geophysical Research 103: Kaplan A, Kushnir Y, Cane MA Reduced space optimal interpolation of historical marine sea level pressure: Journal of Climate 13: Kushnir Y Interdecadal variations in the North Atlantic sea surface temperature and associated atmospheric conditions. Journal of Climate 7: Latif M, Arpe K, Roeckner E Oceanic control of decadal North Atlantic sea level pressure variability in winter. Geophysical Research Letters 27: Mantua NJ, Hare SR, Zhang Y, Wallace JM, Francis RC A Pacific decadal climate oscillation with impact on salmon. Bulletin of the American Meteorological Society 78: Marshall J, Kushnir Y Atlantic climate variability. White paper. Petrovic P Possible climate impacts on water resources of the Danube river basin. Case study: sub-basin of Nitra river. In Proceedings of the Second International Conference on Climate and Water, Espoo, Finland, August; Rajagopalan B, Kushnir Y, Tourre YM Observed decadal midlatitude and tropical Atlantic climate variability. Geophysical Research Letters 25: Rîmbu N, Le Treut H, Janicot S, Boroneanţ C, Laurent C Decadal precipitation variability over Europe and its relation with surface atmospheric circulation and sea surface temperature. Quarterly Journal of the Royal Meteorological Society 127(572B): Shorthouse C, Arnell NW Spatial and temporal variability in European river flows and the North Atlantic oscillation. In FRIEND 97 Regional Hydrology: Concepts and Models for Sustainable Water Resource Management, Gustard A, Blazkova S, Brilly M, Demuth S, Dixon J, van Lanen H, Llasat C, Mkhandi S, Servat E (eds). IAHS Publication No IAHS: Walling Ford; Shorthouse C, Arnell NW The effects of climatic variability on spatial characteristics of European river flows. Physics and Chemistry of the Earth Part B Hydrology Oceans and Atmosphere 24(1 2): Stănescu VA, Corbuş C, Ungureanu V, Simota M Quantification of the hydrological regime modification in the case of climatic changes. In Proceedings of the Second International Conference on Climate and Water, Espoo, Finland, August; Starosolszky O, Gauzer B Effect of precipitation and temperature changes on the flow regime of the Danube river. In Proceedings of the Second International Conference on Climate and Water, Espoo, Finland, August; Van Loon H, Rogers JC The seesaw in winter temperatures between Greenland and Northern Europe. Part I: general description. Monthly Weather Review 106: Werrity A, Foster M Climatic variability and recent changes in rainfall and river flows in Scotland. In Proceedings of the Second International Conference on Climate and Water, Espoo, Finland, August; Wilby RL, O Hare G, Barnsley N The North Atlantic oscillation and British Isles climatic variability, Weather 52: Zorita E, Kharin V, von Storch H The atmospheric circulation and sea surface temperature in the North Atlantic area in winter: their interaction and relevance for Iberian precipitation. Journal of Climate 5: