Climate Change in Latvia
CLIMATE CHANGE IN LATVIA
Editor Māris Kļaviņš UDK 551(474.3) Cl 620
The publication of the book has been supported by the INTERREG III B project ASTRA, Latvian National Research Programme: “Climate Change and Waters”, and Latvian Science Council cooperation project “Effects of Climate Change on the Nature of Latvia, 2006-2009”
Co-editors: M. sc. Valērijs Rodinovs Dr. biol. Guntis Brūmelis Dr. geol. Ervīns Lukševičs Dr. biol. Viesturs Melecis Dr. chem. Linda Eglīte Dr. geogr. Agrita Briede Dr. geogr. Elga Apsīte
Editor-in-chief: Professor Māris Kļaviņš English language editor: Māra Antenišķe Cover photo: Ivars Druvietis
Lay-out and Cover design: Arnis Čakstiņš
The authors acknowledge Latvian Environment, Geology and Meteorology Agency for providing hydrological and meteorological data for climate change analysis
All the papers published in the present volume have been reviewed. No part of the volume may be reproduced in any form without the written permission of the publisher.
ISBN 9984-802-70-1 © Latvijas Universitāte, 2007 5
CONTENTS
Foreword ...... 7 THE CHARACTER OF CLIMATE CHANGE ...... 9 Lita Lizuma, Māris Kļaviņš, Agrita Briede, Valērijs Rodinovs Long-term Changes of Air Temperature in Latvia ...... 11 Māris Kļaviņš, Valērijs Rodinovs Long-term Changes of River Discharge Regime in Latvia ...... 21 Agrita Briede, Lita Lizuma Long-term Variability of Precipitation in the Territory of Latvia ...... 35 Māris Kļaviņš, Valērijs Rodinovs, Anita Draveniece Large-scale Atmospheric Circulation Processes as the Driving Force in the Climatic Turning Points and Regime Shifts in the Baltic Region ...... 45 Māris Kļaviņš, Agrita Briede, Valērijs Rodinovs Ice Regime of Rivers in Latvia in Relation to Climatic Variability and North Atlantic Oscillation ...... 58 Anita Draveniece, Agrita Briede, Valērijs Rodinovs, Māris Kļaviņš Long-term Changes of Snow Cover in Latvia as an Indicator of Climate Variability ...... 73 Tatjana Koļcova, Lita Lizuma, Svetlana Rogozova, Marta Smith Climate Change Impacts on Hydrological Processes in Latvia ...... 86 Uldis Bethers, Juris Seņņikovs Mathematical Modelling of the Hydrological Processes in the Aiviekste River Basin ..... 96 THE IMPACTS OF CLIMATE CHANGE ...... 121 Gunta Spriņģe, Māris Kļaviņs, Jānis Birzaks, Agrita Briede, Ivars Druvietis, Linda Eglīte, Laura Grīnberga, Agnija Skuja Climate Change and Its Impacts in Inland Surface Waters ...... 123 Gunta Grišule, Agrita Briede Phenological Time Series in Latvia as Climate Change Indicators ...... 144 Māris Laiviņš, Solvita Rūsiņa The Dynamics of Pine Forest Vegetation as an Indicator of Climate Change and Eutrophication in the Integrated Monitoring Stations in Latvia ...... 154 Ivars Druvietis, Agrita Briede, Laura Grīnberga, Elga Parele, Valērijs Rodinovs, Gunta Spriņģe Long-term Assessment of Hydroecosystem of the River Salaca, North Vidzeme Biosphere Reserve, Latvia ...... 173 Didzis Elferts The Influence of Climatic Factors on the Radial Growth of the Scots Pine on a Dune Island of Raganas Mire ...... 187 Gunta Spriņģe, Agrita Briede, Ivars Druvietis, Elga Parele, Valērijs Rodinovs Changes of the Hydroecosystem of Lagoon Lake Engure, Latvia (1995-2006) ...... 194 Māris Baltiņš, Māris Kļaviņš, Valērijs Rodinovs Climate Related Mortality Changes in Latvia, 2000-2004 ...... 210 6 Climate Change in Latvia
CLIMATE POLICY AND TECHNOLOGIES ...... 217 Ieva Bruņeniece, Valdis Bisters, Māris Kļaviņš Climate Change Policy Instruments in Latvia ...... 219 Dagnija Blumberga, Marika Rochas, Anna Vološčuka, Ivars Veidenbergs Benchmarking for Energy Climate Technologies in Latvia ...... 252 Sylvestre Njakou Djomo, Dagnija Blumberga Combining the “Well to Gate” and Scenarios Analyses to Assess Hydrogen Transition Pathway in Latvia ...... 261 7
FOREWORD
Climate change is one of the most topical issues in the present society, considering its relevance for each and every human being in the world. However, many of the discussed questions are still the subject of academic studies. Climate change is undoubtedly global, nevertheless, its local manifestation very much depends on regional processes and features. Thus, taking into account the diversity of local (regional) processes, regional studies of climate change impacts are highly relevant. The aim of this book is to analyse climate change patterns and their possible influences on environmental processes and nature of Latvia thus filling in the gaps of knowledge about the character of climate change.
Editor Māris Kļaviņš
The Character of Climate Change
Climate Change in Latvia 10 Climate Change in Latvia The Character of Climate Change 11
Long-term Changes of Air Temperature in Latvia
Lita Lizuma*, Māris Kļaviņš, Agrita Briede, Valērijs Rodinovs University of Latvia, Faculty of Geographical and Earth Sciences, Raiņa blvd. 19, LV 1586, Rīga, Latvia *Latvian Environment, Geology and Meteorology Agency, Maskavas str. 165, LV 1019, Rīga, Latvia
Long-term changes of air temperature can be considered one of the most important indicators of climate change. The long term changes of air temperature in Latvia were investigated. A significant increase in air temperature can be observed for the whole period of observations, however, it has been more expressed during the last decades. The most significant increase of air temperature can be observed in cities, especially, Rīga. Air temperatures have increased, 24-hour amplitudes in air temperature have decreased. The character of changes stresses the importance of local changes in urban environment. Key words: long-term variability, temperature, trends, Latvia
INTRODUCTION
Climate change and global warming are among the most important environmental problems today. Continuously increasing levels of atmospheric carbon dioxide have been recorded for over five decades, resulting in global warming. This phenomenon has been reported to have increased the global average surface temperature by 0.6±0.2 ◦C (IPCC 2001), whose increase may be responsible for climate change (including, for example, the changes in the amount of precipitation and storm patterns). Qualitative evaluation of these changes is based on an analysis of long-term meteorological data series. There are numerous studies in which the general trends of air temperature changes in northern Europe have been demonstrated (Jaagus 1996; Jaagus 1998; Keevallik 2003; Jaagus 2006). Similar studies have also demonstrated changes in Latvia’s climate over the last century (Treilība 1995; Lizuma 2000, 1999). The aim of this study is 12 Climate Change in Latvia to analyze the most important changes in the air temperature in Latvia since the start of the regular observations, especially considering the availability of unique data series of meteorological data collected at the country’s oldest meteorological station, known as Rīga-University. The station was established in 1795 and, along with similar facilities in Vilnius and Tallinn, is among the oldest in the Baltic region. Inasmuch as the Rīga-University meteorological station is located in the very center of the city, its data can be used to study the conditions of an urban environment and changes therein.
MATERIALS AND METHODS
Data used in the study were obtained from the Latvian Environment, Geology and Meteorology Agency, from 22 stations. To characterize the climate and its changes, data on temperature were used. To investigate the links to wide-scale climatic forcing, we used the extended North Atlantic oscillation (NAO) index (Luterbacker et al. 2002). The sum of normalized deviations from the reference climate indicator values has been expressed as Σ(K-1). The multivariate Mann-Kendall test (as described by Hirsch et al. 1982; Hirsch and Slack 1984) for monotone trends in time series of data grouped by sites was chosen for the determination of trends, as it is a relatively robust method concerning missing data, and it lacks strict requirements regarding data heteroscedasticity. The Mann-Kendall test was applied separately to each variable at each site, at a significance level of p<0.5. The trend was considered as statistically significant at a 5 % level if the test statistic was greater than 2 or less than -2 (Hirsch and Slack 1984).
RESULTS AND DISCUSSION
Climatic conditions and air temperature in Latvia are influenced by large scale atmospheric circulation processes, intensity of Solar irradiation, and also by local factors such as vicinity to the Baltic Sea. Mean annual air temperature in Latvia (Fig. 1) (from 22 meteorological stations) is + 5.8 oC with a variation of ± 2.2 oC. The lowest temperatures are observed in the most northern and uphill located stations (northern part of Latvia). The highest temperatures are characteristic of Rīga and the regions directly influenced by the Baltic Sea. Regular observations of air temperature in the territory of Latvia started in 1795 and the trend of temperature changes clearly indicates an increase in the air temperature (Fig. 2) and is in well agreement with the pattern of temperature changes in Uppsala (Sweden). The Character of Climate Change 13
Fig. 1. Mean annual temperature in Latvia and the observed maximal and minimal values of temperature
It was first noted in the 1960s that Rīga’s air temperature was slowly warming (Temņikova 1969). Later research showed that over the course of the last century the average air temperature has increased by 0.5 °C in Latvia as a whole, and by 1 °C in Rīga (Treilība 1995). Analysis of the series of data concerning average air temperature over the course of 155 years (1851-2006) demonstrates an increase in the average temperature of 1.4 °C (Fig. 3a). The significance level of the calculated trend is 99.9%. In comparison with neighbouring countries, similarity of climate change pattern is evident. For example, in Lithuania, air temperature significantly varied during the last 200 years. The late 18th century witnessed it rise by 0.4-0.5 oC. The greatest changes in temperature have been observed in winter and spring seasons. The period 1940 to the present could be characterized as more extreme (Bukantis and Rimkus 2004). The common features for the Baltic Sea region is temperature increases in the 20th century (Førland et al. 2002). It is found that the variability of annual mean temperature of the Baltic Sea region is about five times larger than the variability of global mean temperatures. The warming trend characteristic of the Baltic area has been observed to extend further to the Arctic (Førland et al. 2002). 14 Climate Change in Latvia
Fig. 2. Long-term changes of temperature in Rīga and Uppsala.
Mean annual maximum and minimum temperatures have also increased over the course of time – the mean minimal temperature has increased by 1.9 °C between 1913 and 2006, while the mean maximum temperature has increased by 1.7 °C (Fig. 3b and c). The indicator does differ by season. The mean maximum temperature has increased more rapidly in the latter half of spring (April and May) while the minimum temperature has increased more rapidly in winter season. The uneven shifts in minimal and maximal temperature are the cause of changes in the amplitude of temperatures over a 24-hour period. In the city of Rīga we no longer see as intensive changes in air temperature over the course of day and night as was the case at the beginning of the 20th century. The character of the changes of air temperature in general is similar to other meteorological observation stations in Latvia (Fig. 5). Trend analysis of seasonal temperature changes according to the Mann-Kendall test criteria for period (1950-2003) demonstrates a statistically significant increase of temperature in all the 22 meteorological stations (at p < 0.001, test statistics > 2). The greatest increases in average air temperature have been recorded in spring (March, April, and May) and early winter (November and December) (Fig. 4). In the latter half of winter and in summer the trend has been less pronounced. The Character of Climate Change 15
Fig. 3. Long-term changes of : a) annual mean air temperature in Rīga; b) annual mean maximal temperature in Rīga; c) annual mean minimal temperature in Rīga. 16 Climate Change in Latvia
Fig. 4. Increase of mean monthly temperature (°C) (1851-2006) in Rīga.
Table 1. Characteristics of periods with higher and lower annual temperature
Length of Mean Length of Mean Periods with lower Periods with higher periods, temperature, periods, temperature, annual temperature annual temperature years oC years oC Jelgava (1832-2004) 1838-1856 19 5.8 1832-1837 6 6.6 1875-1895 21 5.8 1934-1939 6 6.9 1951-1956 6 5.5 1971-1975 5 6.6 1961-1970 10 5.6 1981-1984 4 6.7 1976-1980 5 5.3 1988-2003 16 6.8 Liepāja (1895-2003) 1900-1909 10 6.4 1909-1914 6 7.2 1915-1933 19 6.3 1934-1939 6 7.6 1940-1947 8 5.9 1948-1951 4 7.4 1952-1956 5 6.3 1971-1975 5 7.6 1961-1970 10 6.3 1988-2003 16 7.7 1976-1981 6 6.1 Mērsrags (1896-2003) 1898-1909 12 5.7 1934-1939 6 6.8 1925-1933 9 5.7 1943-1951 9 6.5 1961-1970 10 5.7 1971-1975 5 6.9 1976-1981 6 5.5 1988-2003 16 6.8 The Character of Climate Change 17
Fig. 5. Long-term (1950-2003) trend statistics (according to Mann-Kendal test) for meteorological stations in Latvia. 18 Climate Change in Latvia
For summer and winter seasons the increase of air temperature is evident unless its statistical significance is lower, however, during autumn seasons there are no changes in air temperature in Latvia, as the test statistics is not significant for all meteorological stations. Unless for all observation stations well expressed linear trends can be observed, periodic changes of years with increased (or decreased) temperatures in reference years can be observed (Table 1). The length of the periods with higher and lower annual temperature ranges from 4 to 21 years, but usually it is ~ 6 years and the difference between higher and lower annual temperature within these periods is > 1 oC. The importance of periodic changes of air temperature is yet more evident if integral curves of annual temperature (Fig. 6) are analysed for 3 stations – Jelgava, Liepāja, and Mērsrags.
Fig. 6. Integral curves of annual temperature.
The periodicity of interannual variability of temperature is especially evident for years 1944-1993, however, application of integral curves of annual temperature for the analysis of long term changes of air temperature indicates deregulation of climatic processes during the last decades. The Character of Climate Change 19
CONCLUSIONS
This analysis of long-term changes of air temperature in Latvia has showed that the climate of Latvia has changed during the last centuries. Air temperatures have increased, 24-hour amplitudes in air temperature have declined. The air temperature changes are influenced by the overlapping of two processes: a) global changes; b) changes in the urban environment. The process of global warming and related changes in atmospheric circulation have led to higher air temperatures and greater cloudiness, which in turn leads to lower sunshine durability and greater precipitation (Lizuma 2000). The process of urbanization is also of importance, inasmuch as the data were obtained in meteorological observations in the city centre; it could be said that the changes have also been influenced by the urban environment. A city is both a producer and a consumer of energy, and it is a major source of pollution. Both of these factors lead to a more rapid increase in mean temperatures than in the surrounding territories. Still, it is most likely that the increase of air temperature is more due to global changes. 20 Climate Change in Latvia
REFERENCES
Bukantis A., Rimkus E. (2004) Lithuanian Climate in the 18th-21th Centuries, Long term ecological research Baltic conference, Vilnius, 14. Førland E.J., Hanssen-Bauer T., Jónsson C., Kern-Hansen Ø., Nordli O.E., Tveito E., Vaarby L. (2002) Twentieth-century Variations in Temperature and Precipitation in the Nordic, Arctic. Polar Record, 38, 203-210. Hirsch R.M., Slack J.R. (1984) A Nonparametric Trend Test for Seasonal Data with Serial Dependence, Water Resources Res., 20(6), 727-732. Hirsch R.M., Slack J.R., Smith R.A. (1982) Techniques of Trend Analysis for Monthly Water Quality Data, Water Resources Res., 18 (1), 107-121. Intergovernmental Panel on Climate Change, IPCC Third Assessment Report. (2001) Available from http://www.ipcc.ch/. Jaagus J. (1996) Climatic Trends in Estonia during the Period of Instrumental Observations and Climate Change Scenarios. Estonia in the System of the Global Climate Change. Institute of Ecology. Publication, 4, (Ed. J.-M. Punning), 35-48. Jaagus J. (1998) Climatic Fluctuations and Trends in Estonia in the 20th Century and Possible Climate Change Scenarios. Climate Change Studies in Estonia (Eds. T. Kallaste, P. Kuldna). Tallinn, Stockholm Environment Institute Tallinn Centre, 7-12. Jaagus J. (2006) Climatic Changes in Estonia during the Second Half of the 20th Century in Relationship with Changes in Large-scale Atmospheric Circulation. Theor. Appl. Climatol., 83(1-4), 77-88. Keevallik S. (2003) Changes in Spring Weather Conditions and Atmospheric Circulation in Estonia (1955-1995). Int. J. Climatol., 23, 263-270. Lizuma L. (2000) An Analysis of a Long-Term Meteorological Data Series in Riga. Folia Geographica VIII Living with Diversity in Latvia, 53-60. Luterbacker I., Huplaki E., Dietrich D., Jones P.D., Davies T.D., Portis D., Gonzalez-Rouco I.F., von Storch H., Gyalistras D., Casty C., Waner H. (2002) Extending North Atlantic Oscillation Reconstructions Back to 1500. Royal Meteorological Society Atm. Sc. Let., 2; 114-124. (http://www.cru.uea.ac.uk/cru/data/naojurg.htm) Treilība M. (1995). Analysis of Meteorological Observation Series in Latvia. International Conference on Past, Present and Future Climate. Publications of the Academy of Finland 6/95, 310-311. Темникова Н.С. (1969). Климат Риги и Рижского взморья. Гидрометеоиздат, Ленинград. 138 стр. The Character of Climate Change 21
Long-term Changes of River Discharge Regime in Latvia*
Māris Kļaviņš, Valērijs Rodinovs University of Latvia, Faculty of Geographical and Earth Sciences, Raiņa blvd. 19, LV 1586, Rīga, Latvia
The study of changes in river discharge is important for the development of efficient water resource management system, as well as for the development and validation of climate change impact models. The hydrological regimes of rivers and their long-term changes in Latvia were investigated. Four major types of hydrological regimes of rivers, which depend on climatic and physico-geographic factors, were characterized. These factors are linked to the changes observed in river discharge. Periodic oscillations of discharge intensity and low- and high-water flow years are common to the major rivers of Latvia. A frequency of river discharge regime changes of about 20 and 13 years was estimated for the studied rivers. A significant dependence of the river discharge regime changes on the climate change has been found. Key words: long-term variability, discharge, trends, Latvia
INTRODUCTION
A thorough investigation of the increasing human impact on the environment and studies of environmental change are of utmost importance. Long term observations of hydrologic systems provide time series of evapotranspiration, precipitation, and river discharge. These data series can be analysed from different points of view. For example, the study of the hydrological cycle is important in investigation of climatic variation and in hydrological applications (Arnell 1992). Considerable attention has been paid to the study of global climate change, to the relations between global processes of atmospheric circulation (NAO, ENSO), and to the hydrological cycle (Perry et al. 1996; Amarasekera et al. 1997; Simpson and Colodner 1999), as well as to the regional impacts of global climate changes (Gleick 1986). Future climate changes may have substantial impacts on river discharge patterns, as well as on extreme events, their magnitude 22 Climate Change in Latvia and probability of occurrence (Krasovskaia and Gottschalk 1993). River discharge data can also be used to validate hydrological cycle calculations in climate models (Zeng 1999). River discharge time series have been extensively studied worldwide (Molenat et al. 1999; Costa and Foley 1999; Lins and Slack 1999). The relevant trends regarding global climate changes have been identified in Nordic countries (Kite 1993; Rosenberg et al. 1999; Vehviläinen and Huttunen 1997). In Finland, climate change may result in the increase of mean discharge by 20-50 % (Vehviläinen and Lohvansuu 1991). Extensive study of river discharge trends in the USA have discovered that the USA is becoming wetter, with less extreme events (Lins and Slack 1999). Similarly, river discharge patterns have been studied in terms of linear trend analysis, even though they can be much more complex (Pekarova et al. 2003). Analysis of river discharge patterns is important for the Baltic countries, which are located in a climatic region directly influenced both by atmospheric processes in the Northern Atlantic and by continental impacts from Eurasia. The earliest observations of river discharge in Latvia can be dated back to the 19th century for the Daugava River, and long series of data have been accumulated. Studies conducted on river discharge trends in Estonia confirm the importance of such analysis (Jaagus et al. 1998). Long-term streamflow analysis is essential for effective water resource management and therefore has immense socio-economic significance. Discharge analysis in respect to global climatic changes is also very important at present, considering the predicted changes in this region. The aim of this study is to analyse the hydrological regime and long- term changes of river discharge in Latvia.
MATERIALS AND METHODS
The study area covered the whole territory of Latvia (Fig. 1), and also reference sites of rivers in neighbouring areas were used. In Latvia, there is a dense net of rivers flowing through Quaternary sediments. The total number of rivers is 12 500, of which only 17 are longer than 100 km. The total length of rivers is ~ 37950 km and the mean density of the river network is 588 m per 1 km2. The average annual runoff of rivers is about 35 km3, of which more than 50 % forms in neighbouring countries. The hydrological regime in rivers is influenced not only by the climate (precipitation and air temperature), but also by factors such as geomorphology, geological structure, soil composition, and land-use patterns (Table 1). The coverage of lakes and wetlands in river basins also affects the river streamflow. More than 90 % of the total runoff in Latvia is comprised by the five largest rivers. The Character of Climate Change 23
In general, the dominance of natural habitats indicates a rather low level of anthropogenic impact.
Fig. 1. Hydrologic regions of Latvia (I-IV parentheses), water discharge (▲), and air temperature and precipitation (●) study sites.
Table 1. The characteristics of the studied rivers
Basin size, Length, Water runoff, Forest area, Bog area, Agricultural area, River km2 km km3/year % % % Daugava 87900 1005 20.4 43 5 50 Lielupe 17600 119 3.6 22 3 71 Venta 11800 346 2.9 32 5 62 Gauja 8900 452 2.2 47 5 48 Salaca 3420 95 0.95 34 15 45 Bārta 2020 98 0.63 - - - Irbe 2000 32 0.44 - - - Tulija 57 15 0.018 - - -
The chemical composition of the studied rivers can be considered to be representative of the inland waters of Latvia (Table 2). Commonly, watercourses have not been subjected to major anthropogenic pollution, with the exception of the lowest reaches of the rivers and selected sites on the Daugava River below large cities such as Daugavpils, Līvāni, and other (Kļaviņš et al. 1999). 24 Climate Change in Latvia
Table 2. Mean water chemical composition of the studied rivers (1990-2001 seasonal samples)
HCO - SO -2 Ca+2 Mg+2 Cl- COD N-NO - P Pb Zn Cu River 3 4 3 tot mg/l mg/l mg/l mg/l mg/l mg/l mg/l mg/l µg/l µg/l µg/l Daugava 140.9 26.4 40.8 10.2 11.2 31 1.10 0.065 0.1 4.4 1.2 Lielupe 275.3 114.8 99.7 23.9 28.5 39 2.20 0.084 0.3 4.1 2.5 Venta 253.3 44.5 52.5 17.4 16.5 27 1.52 0.039 0.3 2.7 0.9 Gauja 207.9 34.0 52.6 13.6 14.5 30 0.96 0.039 0.7 4.8 1.9 Salaca 175.3 29.4 50.9 12.5 8.5 34 0.98 0.029 0.6 3.3 2.3 Bārta 236.9 32.0 69.2 13.7 10.8 23 1.12 0.045 Irbe 147.1 26.6 44.4 9.6 8.9 37 0.56 0.036 Tulija 186.5 27.6 53.9 12.1 7.0 29 0.88 0.030
Fig. 2. Monthly changes of temperature and amount of precipitation in observation stations (Rūjiena and Daugavpils).
The climatic conditions of Latvia are dominated by the transport of cyclonic air masses from the Atlantic Ocean, leading to comparatively high humidity, uneven distribution of atmospheric precipitation through the year, mild winters, and moist summers. In general, the spatial heterogeneity of the climate of Latvia is determined by physiogeographical features, such as upland relief, distance to the Baltic Sea, and the cover of forests and mires. More precipitation is common to uplands (> 200 m above sea level), and differences between regions can reach up to 250 mm annually. For climate characterization, monthly temperature and precipitation of Daugavpils (in SSE of Latvia) and Rūjiena’s (in NE of Latvia) meteorological stations The Character of Climate Change 25 have been represented (Fig. 2). For centenary trend estimation of the air temperature and precipitation, data from the Meteorological Station Rīga- University were used (Fig. 3). Data used in this study were obtained from the Latvian Environment, Geology and Meteorology Agency.
Fig. 3. Long-term changes of temperature and precipitation for the Meteorological Station Rīga-University.
Discharge measurements covered the last 65 years for the Gauja River and 125 years for the Daugava River. For trend analysis, mean annual discharge values calculated as arithmetic means from monthly records were used. The streamflow data have been tested by Fisher test for data homogeneity before the analyses of variability (Table 3). Table 3. The results of Fisher test statistics
Number of Standard River F F p, % observations deviation empirical theoretical Salaca, 1927-1959 33 10.51 1.10 1.76 0.05 Salaca, 1960-2004 45 9.68 Gauja, 1940-1959 20 14.55 1.87 1.99 0.05 Gauja, 1960-2004 45 18.99 Daugava, 1920-1959 40 115.35 1.02 1.69 0.05 Daugava, 1960-2004 45 113.11 Lielupe, 1921-1959 39 20.95 1.60 1.71 0.05 Lielupe, 1960-2004 45 17.03 Venta, 1920-1959 40 17.33 1.45 1.70 0.05 Venta, 1960-2004 45 20.47 26 Climate Change in Latvia
The length of observation has been divided into two periods that differs in intensity of agricultural activities. Obtained results indicated that time series of river flow are homogenous (Femp where: Qi – discharge in year i; Q0 – mean discharge for the entire period of observation. Using this approach, the integral curve is produced by summing these deviations . By integration of the deviations, the amplitude of the oscillations increases proportionally to the length of the period, with one- sign deviations in the row. The analyses of integral curves allow to identify precisely the significant changing points of low-water and high-water discharge periods. High-water discharge periods are considered to be the years for which K > 1, and low-water flow periods are indicated by a K < 1. The multivariate Mann-Kendall test (as described by Hirsch et al. 1982; Hirsch and Slack 1984) for monotone trends in time series of data grouped by sites was chosen for the determination of trends, as it is a relatively robust method concerning missing data, and it lacks strict requirements regarding data heteroscedasticity. The Mann-Kendall test was applied separately to each variable at each site, at a significance level of p≤0.05. The trend was considered as statistically significant at a 5 % level if the test statistic was greater than 2 or less than -2 (Hirsch and Slack 1984). RESULTS AND DISCUSSION Depending on the hydrological regime, the river basins in Latvia can be grouped into 4 hydrological regions showed in Figure 1. The hydrological regions differ in the seasonal river discharge variability in spring and autumn, in the relative proportion between spring and autumn floods (Fig. 4), and also in other factors (precipitation, evapotranspiration, runoff, temperature): Type I. The Venta River and small rivers along the coast of the Baltic Sea. The rivers in this region have two main discharge peaks – during the spring snow melt and in the late autumn during intensive rainfall; Type II. The Lielupe River and small rivers in the central part of Latvia. This group of rivers receive a major part of their discharge from direct surface The Character of Climate Change 27 runoff, spring floods dominate, and the role of permanent water discharge during the year is comparatively low (~ 40 %); Type III. Basins of the Salaca River, Gauja River, and small rivers along the Gulf of Rīga coast. This group of rivers are characterised by substantial snowmelt floods and comparatively smaller (than type I) rain floods in autumn. 50-60 % of the total runoff takes place in spring; Type IV. Small and medium sized rivers in the basin of Daugava River (as it can be seen on the example of the Rēzekne River). More than half of the river discharge takes place during spring floods, and the water discharge pattern is characterised by steep fluctuations of water discharge intensity. Differences in annual precipitation in Latvia range from 63 % to 150 % in comparison with mean values. More precipitation occurs in the warm period (IV-X) of the year, reaching 63-70 % of the annual total. Mean air temperature decreases in the direction from the West to the East. Interannual temperature variability and interannual variability have a comparatively small significance (Lizuma et al. 2001). Changes in river discharge were determined using linear trend analysis, as it is a commonly used approach in the study of river discharge. Figure 5 and Table 4 shows that the discharge trends in the rivers of Latvia and the north-eastern part of the Baltic Sea are evident: the discharge has significantly increased for the Venta, Gauja, Bārta, Irbe, and Tulija rivers, but the changes are significant and increasing for all of the other studied rivers. Fig. 4. Patterns of seasonal changes of river discharge for major rivers in Latvia. 28 Climate Change in Latvia It is also observed that river discharge is characterized by a stronger increase if the period of trend analyses for the last 50 years is taken. It should be mentioned that discharge trends and trends for precipitation and temperature are similar for the II, III, and IV hydrological region. Regarding the Venta River, located in the type I hydrological region, a positive trend of discharge is more expressed. An observation period of more than 150 years at the Meteorological Station Rīga-University (Fig. 3) shows that during the last century, the mean annual temperature has increased by about 0.8-1.4 oC, and the total annual precipitation by about 7.5 mm every year (Lizuma 2000). Using moving average values (in this case with a step of 6 years), a good coherence is seen between the changes in annual precipitation at the Meteorological Station Rīga-University and the discharges of the largest and mid-sized rivers in the Baltic region (the Daugava, Nemunas, Narva, and Pärnu rivers), which flow into the eastern coast of the Baltic Sea (Fig. 6). Table 4. Significance test for temporal changes of water discharge for rivers in Latvia.* River, sampling Period of Normalised Period of Normalised test station observation test statistic observation statistic Daugava, Daugavpils 1905-2004 -1.16 1961-2004 2.41 Venta, Kuldīga 1905-2004 2.39 1961-2004 1.09 Lielupe, Mežotne 1920-2004 -0.91 1961-2004 1.94 Gauja, Sigulda 1939-2004 1.82 1961-2004 2.50 Salaca, Lagaste 1926-2004 1.07 1961-2004 2.79 Aiviekste, Lubāna 1959-1999 1.65 1961-2003 2.25 Dubna, Sīļi 1948-1998 1.57 1961-1999 3.00 Bārta, Dūkupji 1950-1999 2.35 1961-1999 2.53 Irbe, Vičaki 1955-1999 2.19 1961-1999 2.67 Tulija, Oļi 1961-2004 2.85 * - The trend can be considered as statistically significant at a 5 % level if the test statistics is greater than 2 or less than -2. Figure 6 also indicates the periods with low and high water levels, and the presence of regular cyclic processes. A close relationship between meteorological data and discharge can be found when studied for periods longer than 60 years. The Character of Climate Change 29 Fig. 5. Long-term changes of river discharge in Latvia. 30 Climate Change in Latvia Fig. 6. Long-term changes of mean annual discharge of the rivers in the Baltic region and precipitation. 1 – precipitation (station Rīga-University); 2 – the Nemunas River; 3 – the Daugava River; 4 – the Narva River; 5 – the Pärnu River. Data were levelled to a 6-year moving average The use of integral curves allows to identify oscillation patterns better. Figure 7 shows integral curves for water discharge in the five largest rivers in Latvia. Differences are seen between the Lielupe and the other four large rivers of Latvia, and in all rivers there is an apparent difference between observations before and after 1920. For example, in the Lielupe River, water discharge decreased from 1986 to 2000, in contrast to the other rivers that showed a stable increasing tendency. As it can be seen in Figure 7, in the year 1996 the water discharge reached the lowest value during the last ten years. The difference in flow patterns between the Lielupe River and other rivers in Latvia can be explained bearing in mind that the sampling station in Lielupe, is situated quite upstream (110 km) and thus can reflect more than 50 % of the total river discharge. The Lielupe River basin is moderately affected by melioration and by the construction of various hydrotechnical constructions (dams, ponds, etc) (Kraukle 1987). Agricultural activities also influence the water flow regime in this river. The Character of Climate Change 31 Fig. 7. Normalized integral curves for coefficients of the annual runoff of rivers in Latvia. The general patterns of the periodicity of water flow regime in several major rivers in Latvia are summarised in Table 5. For the last 100-125 years low discharge periods of rivers of Latvia are longer than high discharge periods and they last from a minimum of 10 years up to a maximum of 21-27 years. High discharge periods used to last from 10 years (6-8 years), however, during the last 30 years for the biggest rivers (except Lielupe) their prolongation can reach 20 to 27 years. Goudie (1992) described sinusoidal changes of river discharge in eastern Europe. Short- term fluctuations with mean duration 4-6 years have been previously found in Estonia and Finland (Hiltunen 1994; Jaagus 1995). An approximately 20-year periodicity has been suggested in the earlier studies of rivers in the Baltic region and eastern Europe (Glazacheva 1988), along with a period of about 20 to 50 years of monthly mean precipitation and water level which may be the result of interference of the precipitation and temperature regimes. In previous studies, a 26 year periodicity of the Daugava River flow was considered the main period, which includes 2-, 6- and 13-year cycles (Glazacheva 1988). However, there is no well-defined explicitness of the physical meaning of river discharge regime. 32 Climate Change in Latvia Table 5. Changes of low and high discharge periods for the largest rivers in Latvia Low discharge Q High discharge Years mean, K Years Q m3/s K period m3/s period mean, Daugava (1881-2004) 1881-1901 21 401 0.87 1902-1908 7 595 1.29 1909-1921 13 442 0.96 1922-1936 15 549 1.19 1937-1952 16 419 0.90 1953-1958 6 555 1.20 1959-1985 27 401 0.87 1986-2004 19 490 1.06 Total, mean 77 416 0.90 47 547 1.18 Venta (1897-2004) 1900-1923 24 60.2 0.92 1924-1930 7 72.1 1.10 1931-1949 19 57.0 0.87 1950-1959 10 69.9 1.07 1960-1977 18 57.1 0.88 1978-2002 25 79.1 1.21 Total, mean 61 58.1 0.89 42 73.7 1.13 Salaca (1927-2004) 1933-1952 20 25.6 0.84 1927-1932 6 44.9 1.48 1963-1976 14 22.4 0.74 1953-1962 10 34.6 1.14 1977-2004 28 33.9 1.11 Total, mean 34 24.0 0.79 44 37.8 1.24 Gauja (1940-2004) 1940-1952 13 62.5 0.89 1953-1962 10 84.5 1.21 1963-1977 15 55.8 0.80 1978-2004 27 77.4 1.10 Total, mean 28 59.2 0.84 37 81.0 1.15 Lielupe (1921-2004) 1933-1942 10 49.4 0.89 1921-1932 12 71.9 1.29 1963-1977 15 39.8 0.72 1943-1962 20 61.8 1.11 1984-1997 14 48.9 0.88 1978-1983 6 66.3 1.19 1998-2004 7 66.8 1.18 Total, mean 39 46.0 0.83 45 66.7 1.20 It is important to recognize that the assessment of factors driving the changes of river discharge is far beyond the aims of this article, for basically these questions are part of global climate change problems. Long term changes of river discharge patterns can be directly related to changes in North Atlantic Oscillation (Fig. 8). One can only guess the factors determining the oscillatory pattern of river discharge, nevertheless, it can urge to reconsider the conclusions based on short-term observations and also the conclusions drawn when analyzing river discharge changes only as a linear process. The Character of Climate Change 33 Fig. 8. Long-term changes of the Venta River and the Pärnu River discharge and North Atlantic oscillation index (data were levelled up to a 10-year moving average). CONCLUSIONS River discharge regime during the last century has been subjected to major changes, highly possible in relation to climate change. Well expressed regular changes of high-water and low-water periods are evident. REFERENCES Amarasekera K.N., Lee R.L., Williams E.R., Eltahir E.E.B. (1997) ENSO and the Natural Variability in the Flow of Tropical Rivers. J. Hydrol., 200, 24-39. Arnell N.W. (1992) Factors Controlling the Effects of Climate Change on River Flow Regimes in a Humid Temperate Environment. J.Hydrol., 132, 321-342. Costa M.H., Foley J.A. (1999) Trends in the Hydrologic Cycle of the Amazon Basin. J. Geophys. Res., 104, (D12), 14,189-14, 198. Gleick P. (1986) Methods for Evaluating the Regional Hydrologic Impacts of Global Climatic Changes. J. Hydrol., 88, 97-116. Glazacheva L. Long-term Trends of the River Run-off, Air Temperature in the Baltic Region and Atmospheric Circulation in the Euro-Atlantic Sector. In: The Factors of Regime Formation, Hydrometeorological Conditions and Hydrochemical Processes in the Seas of USSR. Leningrad, Hydrometeorological Agency. 227-241. (in Russian) 34 Climate Change in Latvia Goudie A. (1992) Environmental Change, Clarendon Press, Oxford. Hirsch R.M., Slack J.R. (1984) A Nonparametric Trend Test for Seasonal Data with Serial Dependence. Water Resources Res., 20(6), 727-732. Hirsch R.M., Slack J.R., Smith R.A. (1982) Techniques of Trend Analysis for Monthly Water Quality Data. Water Resources Res., 18 (1), 107-121. Jaagus J., Järvet A., Roosaare J. (1998) Modelling the Climate Change Impact on River Runoff in Estonia. In: Climate change Studies in Estonia (Eds. T.Kallaste, P.Kuldna), Stockholm Environment Institute Tallinn Centre, Tallinn, 117-127. Kite G. (1993) Analysing Hydrometeorological Time Series for Evidence of Climatic Change. Nordic. Hydrol., 24, 135-150. Kļaviņš M., Rodinov V., Kokorīte I., Kļaviņa I. (1999) Chemical Composition of Surface Waters of Latvia and Runoff of Dissolved Substances from the Territory of Latvia. Vatten, 55, 97-108. Krasovskaia I., Gottschalk L. (1993) Frequency of Extremes and Its Relation to Climate Fluctuations. Nordic Hydrol., 24, 1-12. Kraukle L. (1987) The Changes of the River Runoff in Latvia Due to the Melioration Activities. Proceedings of Latvian Hydrometeorological Agency, 1(21), 57-81 (in Russian). Lins F.H. Slack J.R. (1999) Streamflow Trends in the United States, Geophys. Res. Lett., 26, (2), 227-230. Lizuma L. (2000) An Analysis of Long-term Meteorological Data Series in Rīga. Folia Geogr., 7, 53-61. Lizuma L., Briede, A. (2001) The Long-term Variations of Temperature and Precipitation in Latvia. In: Proceedings of 2nd World Congress of Latvian Scientists, Rīga, 273. Molenat J., Davy P., Gascuel-Odoux C., Durand P. (1999) Study of Subsurface Hydrologic Systems Based on Spectral and Cross-spectral Analysis of Time Series. J.Hydrol., 223(1- 4), 152-164. Pekarova P., Miklanek P., Pekar J. (2003) Spatial and Temporal Runoff Oscillation Analysis of the Main Rivers of the World during the 19th-20th Centuries. J.Hydrol., 274, 62-79. Perry G.D., Duffy P.B., Miller N.L. (1996) An Extended Data Set of River Discharges for Validation of General Circulation Models. J. Geophys. Res., 101, D16, 21339-21349. Rosenberg N.L., Epstein D.J., Wang D., Vail L., Srinivasan R., Arnold J.G. (1999) Possible Impacts of Global Warming on the Hydrology of the Ogallala Aquifer Region. Climatic Change, 42, 677-692. Simpson H.J., Colodner D.C. (1999) Arizona Precipitation Response to the Southern Oscillation: a Potential Water Management Tool. Water Resources Res., 35(12), 3761- 3769. Vehvilainen B., Lohvansuu J. (1991) The Effects of Climate Change on Discharges and Snow Cover in Finland. J. Hydrol. Sci., 36 (2), 109-121. Vehviläinen B., Huttunen M. (1997) Climate Change and Water Resources in Finland, Boreal Env. Res., 2(1), 3–18. Zeng N. (1999) Seasonal Cycle and Interannual Variability in the Amazon Hydrologic Cycle. J.Geophys. Res., 104 (D8), 9097-9106. The Character of Climate Change 35 Long-term Variability of Precipitation in the Territory of Latvia Agrita Briede, Lita Lizuma* University of Latvia, Faculty of Geographical and Earth Sciences, Raiņa blvd. 19, LV 1586 Rīga, Latvia *Latvian Environment, Geology and Meteorology Agency, Maskavas str. 165, LV 1019, Rīga, Latvia Long-term trends in Latvia’s annual and monthly precipitation for the period 1922 to 2003 are investigated in this study. The statistical significance of a trend at a study site is evaluated with the help of the Mann-Kendall test. Results of the test indicate that trends of precipitation differ from station to station. Even the modest uplands influence the spatial distribution of precipitation in Latvia, which in the most cases forms barriers for the west flow. Therefore spatial distribution of precipitation is very much a local phenomenon and there are no clearly expressed trends for time series in the similar physiogeographical district neither for annual nor for a single month values (monthly mean values and maximal diurnal monthly values). On the whole, the most significant increase in precipitation has taken place in January and March. However, the Mann-Kendall test indicates a decreasing tendency for September. Key words: precipitation, long-term variability, seasonal distribution, Latvia INTRODUCTION Precipitation globally has increased by about 2% during the 20th century (IPCC 2001). Nevertheless, the changes are not globally uniform. The increases in precipitation in the latter part of the 20th century have been marked over northern Europe (Schönwiese and Rapp 1997), but a general decrease is typical for southward to the Mediterranean (Piervitali et al. 1998; Romero et al. 1998). The dry wintertime conditions over southern Europe and the Mediterranean and wetter-than-normal conditions over many parts of northern Europe and Scandinavia have been explained by the strong positive 36 Climate Change in Latvia values of the North Atlantic Oscillation, with more anticyclonic conditions over southern Europe and stronger westerlies over northern Europe (Hanssen- Bauer and Førland 2000). A number of studies have been investigating whether trends exist in precipitation records (Groisman and Easterling 1994; Karl and Knight 1998; Zhang et al. 2000; Jaagus 2006). During the last century the amount of precipitation has increased by 12 % in the territory which lies between 55 oN and 85 oN (Folland et al. 2001). Long-term changes in atmospheric precipitation are influenced by changes in atmospheric circulation processes, mostly in the cold season. The aim of the study is to determine long-term trends in the time series of precipitation and to analyse changes in seasonal and spatial distribution of precipitation in Latvia. MATERIAL AND METHODS This study is based on mean monthly series of 24 meteorological and gauging stations obtained from Latvian Environment, Geology and Meteorology Agency (Anonymous 1922-2003). The stations are distri- buted all over the territory of Latvia (Fig. 1). Monthly maximal diurnal precipitation has been analysed for 9 gauging stations, located in different physiogeographical regions. Precipitation measurements in Latvia have been made since 1850, and the oldest station is Rīga-University station. However, it should be mentioned that during the war periods observations were interrupted in most of the stations. Changes of precipitation were analysed for different time periods: 1950- 2003, 1961-1990 and for the whole observation period at each site. The links between large-scale atmospheric processes were analysed using North Atlantic Oscillation (NAO) index (http://www.cru.uea.ac.uk/cru/data/nao. htm). For the trend analysis annual, monthly mean, and seasonal time series were used. The seasonal means were calculated as arithmetic means of the monthly values. Trends in the time series of precipitation at 24 stations in Latvia are analysed using the non-parametrical Mann-Kendall test (Libiseller and Grimvall 2002). The Mann-Kendall test was applied separately to each variable at each site, at a significance level of p≤0.01. The trend was considered as statistically significant at a 5 % level if the test statistic was greater than 2 or less than -2. The slope of linear regression was obtained multiplying the slope with the number of observation years (80). The Character of Climate Change 37 Fig. 1. The location of meteorological/gauging stations selected for the study. RESULTS AND DISCUSSION Climate in Latvia is determined by the country’s location in the northwest of the Eurasian continent and by the closeness to the Atlantic Ocean. A rather high cyclonic activity in Latvia annually results in about 170-200 days of precipitation. A typical feature of winter weather is the interchange of cold periods and thaws for several days. The annual amount of precipitation is 703 mm, which exceeds evapotranspiration by 245 mm each year. Table 1. The seasonal distribution of precipitation across selected stations. Station/ Parameter Kolka Ainaži Zosēni Zosēni Jelgava Liepāja Rēzekne Rīga LU Rīga Ventspils 1945-2003 1922-2003 1922-2003 1922-2003 1922-2003 1922-2003 1922-2003 1922-2003 Annual precipitation, mm 630 635 601 687 728 603 649 686 Mean winter precipitation 122 112 120 155 143 104 118 163 (DJF), mm Mean spring precipitation 108 123 101 112 139 119 117 117 (MAM), mm Mean summer precipitation 199 223 188 188 242 217 220 184 (JJA), mm Mean autumn precipitation 201 172 192 231 204 163 196 222 (SON), mm 38 Climate Change in Latvia The higher sums of annual precipitation (>700 mm) for the selected periods (1950-2003, 1961-1990, 1993-2003) are typical of stations located in Vidzeme (central part) and Western Kurzeme’s uplands, but the change pattern is not similar for all stations in the same periods. About 62 % of precipitation falls during summer and autumn in most of the studied stations. Generally the greatest amount of precipitation falls in July and August, but in the west of the country, in September and October (Table 1). A high temporal variability of the trends can be observed for annual precipitation (Fig. 2). The total amount of annual precipitation in Latvia has a tendency to increase for the last 50 years, and higher values are observed in the western part. Fig. 2. Mann-Kendall test statistics of annual precipitation (1950-2003). The investigation of changes in monthly precipitation series can provide a more detailed overview of the timing of significant changes in annual precipitation. Precipitation have a tendency to increase in January, March, and June, while the decrease takes place only in September (Table 2). The results of trend analyses show that the long-term changes of annual precipitation are not uniform over the territory of Latvia. Generally it is possible to see The Character of Climate Change 39 the same tendency for most of the meteorological stations in a particular month, but there are no clearly expressed trends for time series in the similar physiogeographical districts, neither annual nor a single month; it could be due to the fact that spatial distribution of precipitation is very much a local phenomenon. Table 2. Mann-Kendall test values for monthly precipitation (1950-2003). Statistically significant values (p ≤ 0.01) are in bold. Meteorological/ I II III IV V VI VII VIII IX X XI XII gauging station Ainaži 1.81 1.44 2.19 -0.07 -0.43 1.89 -0.98 -0.19 -2.39 0.43 0.52 -0.68 Alūksne 3.58 1.99 1.82 -0.57 0.33 2.02 -0.29 -0.39 -1.28 0.75 0.76 0.21 Bauska 2.12 0.98 0.58 -0.94 0.15 0.18 0.22 -1.13 -1.73 0.60 -0.19 0.45 Dagda 2.68 2.10 2.03 -0.87 -0.05 1.05 -1.04 -2.40 -0.14 -0.31 -0.54 1.81 Daugavpils 1.52 1.82 1.07 -0.36 -0.56 1.49 0.29 0.34 -0.98 0.79 0.16 1.85 Dobele 1.88 1.16 1.78 -0.47 -0.01 1.72 1.36 -0.82 -0.37 1.09 0.26 -0.07 Gulbene 3.37 2.22 2.46 -0.90 0.31 2.22 0.35 -0.79 -0.88 1.52 1.32 1.68 Jelgava 2.09 1.04 1.77 -0.68 0.90 2.29 0.92 -0.03 -0.47 0.04 0.24 1.14 Kolka 0.69 1.06 1.63 0.40 -0.92 1.94 0.13 -0.07 -1.08 0.93 -0.87 -1.53 Kuldīga 1.19 1.29 2.70 0.23 0.43 2.03 -0.47 -0.15 -1.18 1.60 0.15 0.09 Liepāja 0.66 1.01 2.53 -1.05 -0.10 1.38 -0.80 0.10 -1.99 0.93 0.92 -0.91 Mērsrags 2.17 2.11 1.78 0.74 0.70 1.65 0.19 0.46 -1.49 1.40 0.46 -0.30 Pāvilosta 0.42 1.80 2.46 -0.73 -0.93 1.34 -0.63 -0.02 -2.51 0.66 1.49 -0.37 Priekuļi 1.93 0.99 1.04 -0.70 -0.09 0.79 0.28 -1.27 -2.12 0.25 -0.25 -0.44 Rēzekne 2.76 2.03 2.39 -1.19 0.28 0.57 0.72 -0.75 -1.03 -0.31 -0.81 1.23 Rīga-LU -0.60 -0.78 -0.07 -1.50 -0.11 0.82 0.39 0.80 -2.14 0.19 -0.54 -1.02 Rūjiena 2.48 1.25 2.42 -0.36 0.19 1.51 -1.02 -0.46 -2.04 0.73 -0.70 -0.15 Saldus 1.52 1.94 1.89 0.01 -1.19 1.03 0.58 -0.36 -1.04 0.96 0.28 -0.07 Skrīveri 2.42 2.09 1.17 -1.07 -0.49 1.62 0.01 -0.88 -2.23 0.99 0.28 1.09 Skulte 1.63 1.52 1.36 -0.27 0.56 1.85 0.11 -0.13 -2.15 0.29 0.23 0.28 Stende 1.67 2.44 1.97 0.31 -0.04 2.19 0.19 -0.63 -1.25 1.98 -0.01 -0.25 Ventspils 1.43 1.81 2.17 -0.12 0.03 2.09 0.18 -0.02 -1.65 0.69 0.28 -0.21 Zosēni 2.13 1.31 1.16 -1.08 0.29 1.63 -0.51 -1.31 -1.40 1.13 -0.46 0.23 Zīlāni 1.89 0.93 0.66 -0.70 -0.65 0.90 0.08 -1.28 -1.57 0.56 -0.10 -0.08 During the cold half of the year (November – March), a statistically significant increase (p≤0.01, test statistic ≥ 2) of precipitation has been observed for 13 of the 24 stations for the period 1950-2003. However, there are no clearly expressed changes for the warm period (April – October) (Fig. 3). 40 Climate Change in Latvia Fig. 3. Mann-Kendall test values of precipitation in the warm (IV-X) and cold (XI-III) periods (1950-2003). The total changes of the precipitation are discovered with the help of a slope of linear regression for 8 stations under 80 years of observation (Fig. 4). It is evident that precipitation increases in a different amplitude in winter period. It is typical of all selected stations. Relatively similar decreasing patterns (in 7 of 8 cases) have been observed for summer and autumn seasons, contributing to the decreasing summer and autumn precipitation in the annual total. In the previous studies it is proven that annual cold-period precipitation in Latvia has become more abundant and is more evident in those parts where the prevailing winds and relief fosters the ascending of air masses (Treilība 1995). In Europe an apparent precipitation increase was typical of the first half of the 20 century. The Character of Climate Change 41 Fig. 4. The slope of seasonal precipitation (mm) for period 1922-2003. Investigations also showed a larger increase in winter and autumn seasons, more pronounced in higher latitudes (Folland et al. 2001). The studies done in Estonia indicated precipitation increase by 80-180 mm (or 10-25 %) during the 20 century for 9 stations (Jaagus 1998, 2006). The maximal diurnal sums of monthly precipitation have been analysed considering 9 gauging stations (Atašiene, Carnikava, Dundaga, Dzērbene, Kuldīga, Lejasciems, Ludza, Melturi, and Nereta) during the period of 55 years (1948-2003). The statistically significant increases (p≤0.01, test statistic ≥ 2) of maximal diurnal precipitation in most cases (7 of 9) have been in February (Fig. 5). For March and January there is a strong tendency of increase for most of the cases, but statistically significant are the changes for 4 and 3 gauging stations, respectively. Generally, for the maximal diurnal sum of the monthly precipitation, the change pattern is similar to that of the monthly mean values. The possible connection between large-scale atmospheric circulation and air monthly precipitation has been investigated using NAO monthly and seasonal indexes. The obtained results indicate that these connections have different seasonal influence on precipitation in Latvia (Table 3, Fig. 6). Monthly correlation coefficients of precipitation and NAO indexes are mostly significant for January, February, March, and December. The positive NAO indexes in winters are related to western atmospheric transfer and have resulted in precipitation and temperature increase. That could be explained by a rather strong connection between NAO and climatic variability in Latvia, especially in winter. 42 Climate Change in Latvia Fig. 5. Monthly maximal diurnal temperature changes (1948-2003) for two gauging stations. The long-term fluctuations of NAO indexes and winter precipitation anomalies are showed in Figure 6. There is a positive relationship between those two parameters, which indicates the increase of precipitation in the territory of Latvia in the case of a positive NAO index. The correlation coefficient of the winter precipitation and NAO indexes is more significant considering the last thirty years because a positive phase of NAO indexes dominates. Fig. 6. Relationship between yearly winter precipitation anomalies (in comparison with 1961-1990) and NAO indexes (1925-2003). Mean values of winter precipitation are summed up for 10 meteorological stations. The Character of Climate Change 43 Table 3. Correlation coefficients between NAO indexes and precipitation (1950-2003). ± Statistically significant correlation coefficients (rα0.05; n54= 0,26) are marked in bold. Meteoro- logical/ I II III IV V VI VII VIII IX X XI XII gauging station Ainaži 0.47 0.40 0.47 0.09 -0.18 -0.22 0.17 0.11 0.28 0.03 0.14 0.31 Alūksne 0.40 0.34 0.33 0.09 0.01 -0.20 0.02 -0.10 0.07 0.12 0.21 0.11 Bauska 0.31 0.16 0.30 0.08 -0.20 -0.07 -0.22 -0.10 0.14 -0.02 0.23 0.17 Daugavpils 0.37 0.33 0.28 0.18 0.01 -0.24 -0.12 -0.12 0.03 -0.07 0.03 0.06 Dobele 0.34 0.25 0.39 -0.07 -0.33 -0.20 0.01 -0.20 0.10 -0.08 0.15 0.12 Gulbene 0.35 0.29 0.28 0.11 -0.03 -0.18 -0.11 -0.03 0.11 0.02 0.22 0.08 Jelgava 0.29 0.08 0.29 0.02 -0.26 -0.25 0.11 -0.13 0.14 -0.03 0.07 0.09 Kolka 0.26 0.25 0.38 0.08 -0.06 0.12 0.19 -0.08 0.12 0.07 0.08 0.06 Liepāja 0.41 0.22 0.35 -0.01 -0.04 -0.17 0.08 -0.13 -0.02 -0.12 0.13 0.12 Mērsrags 0.28 0.18 0.25 0.12 -0.24 -0.09 0.25 0.06 0.12 -0.10 0.07 -0.02 Pāvilosta 0.32 0.31 0.33 0.06 -0.09 -0.21 0.12 -0.03 0.01 -0.11 0.07 0.06 Priekuļi 0.47 0.43 0.43 0.13 -0.17 0.04 0.01 -0.06 0.22 0.15 0.20 0.27 Rēzekne 0.34 0.17 0.25 0.09 -0.02 -0.31 0.12 -0.12 0.06 0.05 0.09 0.17 Rīga-LU 0.35 0.17 0.21 -0.06 -0.28 0.00 0.07 -0.01 0.11 -0.11 0.09 0.20 Rūjiena 0.47 0.45 0.46 0.15 -0.16 -0.06 0.09 0.02 0.25 0.07 0.21 0.30 Saldus 0.51 0.43 0.45 0.18 -0.26 -0.03 0.07 -0.08 0.14 0.00 0.15 0.19 Skrīveri 0.43 0.28 0.40 0.07 -0.11 0.05 -0.08 -0.07 0.23 0.07 0.18 0.26 Skulte 0.36 0.32 0.42 0.12 -0.20 -0.25 0.04 0.03 0.15 0.00 0.07 0.04 Stende 0.50 0.41 0.43 0.21 -0.25 -0.15 0.06 -0.07 0.02 0.03 0.13 0.14 Ventspils 0.44 0.37 0.41 0.02 -0.09 -0.14 0.02 -0.06 0.02 -0.06 0.06 0.15 Zīlāni 0.37 0.28 0.28 0.06 0.01 -0.30 -0.21 -0.14 0.12 0.01 0.14 0.26 Zosēni 0.51 0.39 0.41 0.06 -0.11 -0.21 -0.15 0.02 0.20 0.10 0.23 0.25 CONCLUSIONS In this study the trends of annual and seasonal precipitation in Latvia for the period of 1950-2003 and for the whole observation period were investigated with the use of the Mann-Kendall test and the slope of linear regression. Analyses of annual and seasonal precipitation show significant variations in trends between stations. For individual stations (13 of 24), an overall increasing trend is evident in precipitation series for the cold period. On a monthly scale, strong increasing trends for most of the stations were obtained for January, March, and June. A corresponding statistically significant decreasing monthly trend was typical only of September. The maximal diurnal sum of the monthly precipitation has also increased in the cold period of the year. North Atlantic Oscillation seems to play an important 44 Climate Change in Latvia role in the precipitation regime in Latvia. Particularly, it explains significant proportion of winter precipitation variability in the territory. REFERENCES Anonymus (1922-2003) Meteoroloģisko novērojumu tabulas [Meteorological tables]. LVĢMA datu fonds, Rīga. Briede A., Lizuma L. (2004) The Long-term Variability of Climate in Latvia. LTER Baltic conference. Abstract. Vilnius, Lithuania, 13. Folland C.K., Karl T.R., Christy J.R., Clarke R.A., Gruza G.V., Jouzel J., Mann M.E., Oerlemans J., Salinger M.J., Wang S.-W. (2001) Observed Climate Variability and Change. In: J. T. Houghton et al. (eds.), Climate Change 2001: The Scientific Basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, 99-181. Hanssen-Bauer I., Førland E.J. (2000) Temperature and Precipitation Variations in Norway 1900-1994 and Their Links to Atmospheric Circulation. Int. J. Climatol., 20, 1693- 1708. Jaagus J. (1998) Climatic Fluctuations and Trends in Estonia in the 20th Century and Possible Climate Change Scenarios. Climate Change Studies in Estonia (Eds. T. Kallaste, P. Kuldna). Tallinn, Stockholm Environment Institute Tallinn Centre, pp. 7-12. Jaagus J. (2006) Climatic Changes in Estonia during the Second Half of the 20th Century in Relationship to Changes in Large-scale Atmospheric Circulation. Theor.Appl.Climatol. 83, 77-88. Libiseller C., Grimvall A. (2002) Performance of Partial Mann-Kendall Test for Trend Detection in the Presence of Covariates. Environmetrics, 13, 71–84. Lizuma L., Briede A. (2001) Gaisa temperatūras un nokrišņu svārstību ilgtermiņa izmaiņas Latvijā. Pasaules Latviešu zinātnieku kongress, Rīga, lpp. 273. IPCC (Intergovernmental Panel on Climate Change). (2001) Climate Change 2001: The Scientific Basis. (www.ipcc.ch/pub/tar/wg1/index.htm) Piervitali E., Colacino M., Conte M. (1998) Rainfall over the Central-Western Mediterranean Basin in the Period 1951-1995. Part I: Precipitation Trends. Geophys. Space. Phys., 21C(3), 331-344. Romero R., Guijarro J.A., Ramis C., Alonso S. (1998) A 30-year (1964-1993) Daily Rainfall Data Base for the Spanish Mediterranean Regions: First Exploratory Study. Int. J. Climatol., 18, 541-560. Schönwiese C.D., Rapp J. (1997) Climate Trend Atlas of Europe Based on Observations 1891-1990. Kluwer Academic Publishers, Dordrecht, 228 pp. Treilība M. (1995) Analysis of Meteorological Observation Series in Latvia. International Conference on Past, Present and Future Climate. Publications of the Academy of Finland 6/95, 310-311. The Character of Climate Change 45 Large-scale Atmospheric Circulation Processes as a Driving Force in the Climatic Turning Points and Regime Shifts in the Baltic Region Māris Kļaviņš, Valērijs Rodinovs, Anita Draveniece University of Latvia, Faculty of Geographical and Earth Sciences, Raiņa blvd. 19, LV 1586, Rīga, Latvia Large scale atmospheric circulation processes can be considered to be the most important factor influencing climate in the Baltic region. Analysis of the long-term trends of main atmospheric circulation types (westerly circulation, southerly and easterly circulation, and meridional or northerly airflow) demonstrates significant changes of the main circulation patterns which are associated with significant changes in the climate of Latvia, and also of the Baltic region. Climate warming can be associated with increased temperature and amount of precipitation, at first in winter-seasons, and increased intensity of zonal circulation. Key words: atmospheric circulation; climatic turning points; long-term changes INTRODUCTION Amongst major factors influencing climate there are: large-scale atmospheric circulation arising from uneven distribution of solar radiation on the Earth’s surface; the Earth’s rotation and interaction between atmosphere, hydrosphere, and lithosphere. Large-scale atmospheric circulation is the most important factor influencing climate in northern Europe (Chen 2000; Jaagus 2005). As it has been previously shown (Girs 1971), the atmospheric circulation processes over central and eastern Europe can be described using three most common circulation patterns: W – westerly circulation, E – southerly and easterly, and C – meridional or the northerly airflow. This classification is in well agreement with the other suggested classification of large scale atmospheric circulation processes over the central Europe – the 46 Climate Change in Latvia Hess and Brezovsky classification system (Hess and Brezovsky 1969), also known as the Grosswetterlagen. The combination of these atmospheric circulation types allows to describe long term climatic processes. Amongst factors influencing large scale atmospheric circulation there are solar activity, changes of Earth rotation, intensity of major oceanic streams (at first the Gulf stream), and other processes. Several studies have demonstrated an increase in the westerly circulation in winter season associated with the North Atlantic oscillation (NAO) which causes higher temperature and precipitation in Europe (Chen and Hellstrom 1999) and is possibly related to global climate change. In several studies the close correlation between NAO index and different indicators of climate change, such as temperature, amount of precipitation, snow regime, and others, has been demonstrated (Loewe and Koslowski 1998; Chen and Hellstrom 1999; Paeth et al. 1999; Uvo 2003). A significant part of the overall climate system is based on the close interplay of hydrological regime and other processes in the climate system. However, the indicators of hydrological regime such as river discharge have been comparatively little studied, especially in respect to large scale atmospheric circulation processes. The aim of this study is to analyse the changes of large-scale atmospheric circulation processes and their impact on the changes of the main climate indicators in the Baltic region. MATERIALS AND METHODS Data used in the study were obtained from the Latvian Environment, Geology and Meteorology Agency. To characterize the climate and its changes, data on temperature, amount of precipitation, ice regime on the rivers of Latvia, snow cover, and river discharge were used. Large-scale atmospheric circulation processes have been studied using the catalogues of the indexes of atmospheric circulation. The Arctic and Antarctic Research Institute (AARI), St. Petersburg, Russia, has produced the catalogue of the daily forms of the Atlantic-Eurasian atmospheric circulation processes since 1891 and has been successful in providing long-term forecasts for the W, E, and C patterns. This catalogue has been used to study decadal- scale fluctuations in atmospheric circulation, which have been linked to changes in temperature and precipitation (Girs 1971), as well as changes in climate and the rotation of the solid earth (Sidorendkov and Svirenko 1991). Decadal-scale fluctuations in the Atlantic-Eurasian atmospheric circulation are represented by the atmospheric circulation index (ACI). ACI is calculated as the integral curve of negative meridional (C) anomalies (Beamish et al. 1998). To investigate the links to wide-scale climatic forcing, we used the extended North Atlantic oscillation (NAO) index (Luterbacker et al. The Character of Climate Change 47 2002) and the recently suggested Baltic winter index (WIBIX) (Hagen and Feistel 2005). The values of the climatic processes as well as atmospheric circulation indexes have been separated from the indicators of the warm season (from November till March and further indicated in subscript). The sum of normalized deviations from reference climate indicator values has been expressed as Σ(K-1). RESULTS AND DISCUSSION The analysis of the nature of atmospheric circulation processes over the last century (1900-2000) reveals (Fig. 1, 2) that the dominant types of circulation processes do display longitudinal atmospheric circulation processes: either westerly (W) or northerly and easterly (E), and this type of circulation dominates in autumn-winter seasons, forming up to 80 % of the total (Fig. 3). Fig. 1. Types of atmospheric circulation processes. W – westerly; C - northerly and E – easterly. Meridional (C – northerly) circulation pattern gains significance only during spring-summer seasons, but the share of this circulation pattern does not exceed 1/3 of the overall. Nevertheless, long-term analysis of circulation processes (Fig. 3) shows that it is possible to find years (1902, 1947) or even periods (1900-1902, 1940-1948) when meridional circulation dominates the overall atmospheric circulation processes. Starting from the 50ties, zonal circulation (E + W) dominates the atmospheric circulation process, at first due to the increase of the dominance of southerly and easterly (E) airflow. In the mid 80ties it ceased and has been substituted with increased westerly (W) circulation, accompanied by higher temperatures and increased precipitation in winter. Unless the atmospheric circulation processes do have evident oscillatory pattern, it is possible to 48 Climate Change in Latvia establish a centennial increasing trend of easterly and southerly (continental) atmospheric circulation and decreasing westerly airflow influence (Fig. 4). Fig. 2. Centennial seasonal distribution of atmospheric circulation processes in the Baltic Region (1900-2000). Fig. 3. Long-term changes of large-scale atmospheric circulation processes in the Baltic region. Meridional circulation C does not have a well expressed pattern of circulation, however, it is possible to observe a well expressed periodicity of 15-20 years. The Character of Climate Change 49 Fig. 4. Trends of long-term changes of large-scale atmospheric circulation processes (3-year moving average). These trends are more evident for the warm season of the year (April – October) than for the cold season (November – March) (Fig. 5). Fig. 5. Trends of seasonal long-term changes of large-scale atmospheric circulation processes in the Baltic region. The changes of atmospheric circulation processes do have a well expressed periodicity of 6, 10-12, and 15-20 years, and possibly this periodicity can be 50 Climate Change in Latvia related to solar activity and the processes of movement of the Earth in the space. The periodicity of air mass circulation processes allows identifying the periods with the dominance of a respective atmospheric circulation type during the last century (Table 1). In each of these periods one atmospheric circulation type is dominant, except for the period of 1940-1950 when all three airflow patterns have had equal impacts. A comparative analysis of zonal (W+E) and meridional (C) atmospheric circulation types summing up yearly anomalies for corresponding circulation patterns allows obtaining integral curves which demonstrate centennial changes of air mass transport in 3 periods (30-33 years) when zonal airflow was influenced by meridional circulation patterns (Fig. 6). Table 1. The character of changes in dominance of atmospheric circulation processes Prolongation of period, ACI-W ACI-E ACI-C Dominant Periods years % % % circulation type 29 1900-1928 40 36 24 W 11 1929-1939 33 45 22 E 11 1940-1950 30 36 34 W+E+C 12 1951-1962 29 45 26 E 11 1963-1973 20 52 28 E 20 1974-1993 26 53 21 E 6 1994-2000 35 39 26 E+W Changes of indicators (NAO index, WIBIX, and ACI) describing large- scale atmospheric processes have similar patterns (Fig. 7). Atmospheric circulation indexes according to Vangengeim-Girs classification (ACI- W) also have the same pattern of changes, as far as they describe or are developed considering the processes over the North Atlantic and their impact on continental climate. The analysis of atmospheric circulation processes allows explaining the variability of meteorological conditions over the Baltic region (temperature, precipitations, prolongation and intensity of snow cover, ice regime on rivers, lake level fluctuations). Often closer correlations are evident for winter period when zonal circulation processes dominate (Fig. 8). Less evident are the correlations between atmospheric circulation processes and climate indicators in warm seasons (spring-summer-autumn), as it has also been proven in our previous studies. Cumulative curves of deviations from reference values (Fig. 8) both for the indicators (WIBIX index, ACI) of atmospheric circulation processes and meteorological parameters (temperature) allows identifying the years which in a centennial perspective can be considered the climate turning points (1902, 1939, 1972, 1987). The Character of Climate Change 51 Fig. 6. Integral curves of zonal (W+E) and meridional (C) atmospheric circulation anomalies. Fig. 7. Long-term changes of atmospheric circulation processes as described by their indices (1- winter NAO index; 2- WIBIX index; 3- ACI-W annual cold period 6 year moving average). 52 Climate Change in Latvia Fig. 8. Cumulative curves of normalized values of WIBIX index, deviations from reference winter temperature (Rīga-University station), and the number of days in winter season with a dominant zonal circulation for the last 100 years. The breaking points indicating changes in the character of circulation during the last century (the beginning of the century, year 1940, the beginning of the 70 ties) can be associated with significant changes of climate indicators (winter temperatures, amounts of precipitations, etc). The long-term changes of atmospheric circulation patterns are coherent with the long-term changes of river discharge, as it can be seen on the example of the Venta River. The same pattern is common to the largest rivers of the eastern coast of the Baltic region (the Daugava, Narva, Nemunas) as well as for mid-sized rivers (the Lielupe, Gauja, Pärnu, Salaca), where a positive correlation can be found with the zonal circulation (W) indexes. However, for the Venta River discharges a negative correlation was found with the meridional circulation type C (Table 2, Fig. 9). For the characterization of climatic processes, both direct measurements of climatic indicators and the analysis of processes influencing atmospheric circulation can be used. Atmospheric circulation can be characterized both by circulation indices and monthly frequency of circulation types. In studies of climate change in Europe (Hurrell and van Loon 1997; Loewe and Koslowski 1998; Chen and Hellstrom 1999), the authors have most broadly used North Atlantic oscillation (NAO) index, characterizing westerly airflow or the intensity of zonal circulation. NAO index is defined as the difference in normalized sea-level pressure between the Azores high and Icelandic low. The classification of the NAO index data is based on the definition of 3 categories: high (NAO >1) – strong westerly, normal (NAO ~ 1) and low (NAO < -1) – weak westerly. The Character of Climate Change 53 Table 2. The correlation of Baltic region river discharge with the regime of large-scale atmospheric circulation processes. River ACI-E ACI-W ACI-C Neman -0,17 0,26* -0,09 Narva -0,25* 0,42** -0,19 Daugava -0,35** 0,36** 0,04 Venta 0,04 0,14 -0,25* Pärnu -0,12 0,44** -0,39** Lielupe -0,27* 0,38** -0,08 Gauja -0,16 0,40** -0,21 Salaca -0,16 0,43** -0,26* ** Correlation is significant at a 0.01 level (2-tailed). * Correlation is significant at a 0.05 level (2-tailed). Table 3. Correlation between atmospheric circulation indices, snow cover parameters, and winter temperature in Latvia. Days with snow cover Depth of snow cover To winter Ainaži Alūksne Rūjiena Ainaži Alūksne Rūjiena Rūjiena ACI-Ecold 0,46 0,32 0,40 0,33 0,47 0,33 -0,22 ACI-Wcold -0,56 -0,28 -0,49 -0,45 -0,54 -0,38 0,45 ACI-Ccold -0,08 -0,17 -0,06 -0,01 -0,12 -0,08 -0,18 In order to describe atmospheric circulation patterns over Arctic, they are classified into three basic circulation forms: 1) W – westerly circulation; 2) E – southerly and easterly; 3) C – meridional or the northerly airflow, as suggested (Girs 1971) mainly according to the thermobaric wave position in the upper troposphere. Vangengeim and Girs’s atmospheric circulation classification scheme (Fig. 1) shows the dominant directions of air mass transfer and sea-level atmospheric pressure fields. Recently it has been demonstrated that Vangengeim and Girs’s atmospheric circulation classification scheme allows to analyse and explain climatic processes not only in northern and eastern Europe (Aasa et al. 2004; Jaagus 2005), but also in numerous parts of Siberia (Berezovskaya et al. 2005). In several studies it has been indicated that Vangengeim and Girs’s atmospheric circulation classification scheme is in well agreement with the other suggested classification schemes of large scale atmospheric circulation processes over the central Europe – the Hess and Brezovsky classification system (Gerstengarbe and Werner 1999), also known as the Grosswetterlagen. This atmospheric circulation classification scheme considers: 1) the character of circulation over North Atlantic; 2) the positions of frontal zones; 3) the positions of dominant cyclonic and anticyclonic air 54 Climate Change in Latvia masses. Although it focuses on central Europe, it is actually valid over the whole territory of Europe. It has been shown (Keevallik and Loitjärv 1999) that this classification, which consists of 29 typical circulation patterns and one undetermined case, is valid for Estonia, however, temperature and precipitation data display that some patterns bring different weather to central Europe and Estonia (ibid). In addition, Post and Tuulik (1999) have shown that for the circulation patterns of zonal and mixed circulation groups the upper air flow and air mass types are similar in central Europe and Estonia, but for the meridional circulation group there are different air masses over Estonia and over CE. Fig. 9. Long-term changes of the Daugava River and the Venta River discharge and the character of atmospheric circulation processes. The Character of Climate Change 55 It is suggested that the flows are similar but Estonia’s North-eastern position is the reason for having more transformed and cold air masses (ibid). Because of the neighbouring position of Latvia and Estonia and the small size of the territories of both countries, it might be supposed that the same processes occur in Latvia. To study climate change processes in the winter season, the Baltic winter index has recently been suggested (WIBIX) (Hagen and Feistel 2005). This derived climate index is based on monthly values of the first principal component of 1) winter anomalies of air pressure difference between Gibraltar and Reykjavik to describe North Atlantic oscillation; 2) sea level anomalies of Landsort (Sweden) to characterize the filling level in the Baltic; 3) maximum Baltic ice cover to include the influence of continentally dominated alignments of atmospheric centres of action. The results of this study allow to suggest that the impact of the airflow from North Atlantic in spring-summer-autumn seasons is of less importance, thus a greater role is played by the meridional processes of atmospheric circulation. The time from April to September covers the spring period of transition from winter to summer atmospheric circulation in the northern hemisphere (end of March to early June), which is more dependent on solar radiation, the summer period, and the transition back to winter circulation. It has been shown that during summer season, westerly circulation is at the weakest, and during spring/autumn seasons, when it gradually weakens/ intensifies, meridional circulation develops more often, whereby either arctic air masses or warm midlatitude or even subtropical air arrives in Latvia (Krauklis and Draveniece 2004). However, if westerly airflow can be associated with an increased transport of warmer air masses and increased amount of precipitations, then in the case of a dominant meridional circulation the transfer of air masses from North Atlantic could be either blocked or could be a source of increased precipitation. In general, the meridional circulation is mostly the reason for atmosphere blocking. CONCLUSIONS The analysis of large-scale atmospheric circulation processes allows to identify the years which in a centennial perspective can be considered as the climate turning points (1902, 1939, 1972, 1987) associated with significant changes of climate indicators (winter temperatures, amounts of precipitation, etc). They are evident if the regional and local meteorological processes are studied. Nevertheless, in the studies of climate change it is important to consider not only atmospheric circulation over North Atlantic, but also other large-scale circulation processes. 56 Climate Change in Latvia REFERENCES Aasa A., Jaagus J., Ahas R., Sepp M. (2004) The Influence of Atmospheric Circulation on Plant Phenological Phases in Central and Eastern Europe. Int. J.Climatol., 24, 1551-1564. Bardossy A., Filiz F. (2005) Identification of Flood Producing Atmospheric Circulation Patterns. J.Hydrol., 313, 48-57. Berezovskaya S., Yang D., Hinzman L. (2005) Long-term Annual Water Balance of the Lena River. Global Planetary Change, 48, 84-95. Chen D., Hellström C. (1999) The Influence of the North Atlantic Oscillation on the Regional Temperature Variability in Sweden: Spatial and Temporal Variations. Tellus, 51A, 505- 516. Chen D. (2000) A Monthly Circulation Climatology for Sweden and Its Application to a Winter Temperature Case Study. Int. J. Climatol., 20, 1067 – 1076. Girs A.A. (1971) Long-term Fluctuations of Atmospheric Circulation Patterns and Its Forecasts. Gidrometeoizdat: Leningrad, 480 p. (in Russian). Hagen E., Feistel R. (2005) Climatic Turning Points and Regime Shifts in the Baltic Sea Region: the Baltic Winter Index (WIBIX) 19659 – 2002. Boreal Environ. Res., 10, 211- 224. Hess P., Brezovsky H. (1969) Katalog der Grosswetterlagen Europas. Bericht des Deutschen Wetterdienstes., Deutscher Wetterdienst, Offenbach am Main 2nd ed. Nr 113, Bd. 15. Hurrell J.W., van Loon H. (1997) Decadal Variations in Climate Associated with the North Atlantic Oscillation. Clim. Change, 36, 301-326. Jaagus J. (2005) Climatic Changes in Estonia during the Second Half of the 20th Century in Relationship With Changes in Large-scale Atmospheric Circulation. Theor. Appl. Climatol., 83(1-4), 77-88. Keevallik S., Loitjärv K. (1999) European Circulation Patterns and Synoptic Situation in Estonia. Publicationes Instituti Geographici Universitatis Tartuensis, 85, 123-132. Krauklis Ā., Draveniece A. (2004) Landscape Seasons and Air Mass Dynamics in Latvia. Folia Geographica, 12, 16-48. Loewe P., Koslowski G. (1998) The Western Baltic Sea Ice Season in Terms of a Mass-related Severity Index 1879-1992 (II) Spectral Characteristics and Association with the NAO, QBO, and Solar Cycle. Tellus, 50(A), 219-241. Luterbacker I., Huplaki E., Dietrich D., Jones P.D., Davies T.D., Portis D., Gonzalez-Rouco I.F., von Storch H., Gyalistras D., Casty C., Waner H. (2002) Extending North Atlantic Oscillation Reconstructions Back to 1500. Royal Meteorological Society Atm. Sc. Let., 2; 114-124 http://www.cru.uea.ac.uk/cru/data/naojurg.htm). Paeth H., Hense A., Glowienka-Hense R., Voss R., Cubash U. (1999) The North Atlantic Oscillation as an Indicator for Greenhouse-gas Induced Regional Climate Change. Clim. Dynamics, 15, 953-960. Post P., Tuulik J. (1999) Estonian Weather Elements and European Circulation Patterns. Phys. Chem. Earth (B), 24 (1-2), 97-102. Sidorendkov I. S., Svirenko P.I. (1991) Long-term Changes in Atmospheric Circulation and Fluctuations in the Primary Natural Synoptic Region. pp. 93-104. In Atmospheric Processes of the Earth, the Studies of the Hydrometeorological Scientific Research Centre, USSR, Vol. 316, 152 p. Uvo C.B. (2003) Analyses and Regionalization of Northern Europe Winter Precipitation Based on Its Relationship with North Atlantic Oscillations. Int. J. Climatol., 23, 1185-1194. Yoo J.C., D’Odorico P. (2002) Trends and Fluctuations in the Dates of Ice Break-up of Lakes and Rivers in Northern Europe: the Effect of the North Atlantic Oscillation. J.Hydrol., 268, 100-112. The Character of Climate Change 57 FOOTNOTES ∗ Modified from publication in Proceedings of Latvian Academy of Sciences 58 Climate Change in Latvia Ice Regime of Rivers in Latvia in Relation to Climatic Variability and North Atlantic Oscillation* Māris Klaviņš, Agrita Briede, Valērijs Rodinovs University of Latvia, Faculty of Geographical and Earth Sciences, Raiņa blvd. 19, LV 1586, Rīga, Latvia Different elements of the hydrological regime of rivers in Latvia were studied to evaluate the regional impacts of climate change. River discharge in Latvia is directly influenced by the distribution of atmospheric precipitation and generally follows an oscillating pattern. The discharge in winter-spring season is very much related to ice regime. Ice cover period for all rivers in Latvia has become shorter during the last decades. Nevertheless, a very long period of records of ice-breakup in the Daugava River (from 1529) indicates oscillation around mean values. The ice regime and thus the seasonal river discharge are very much influenced by North Atlantic oscillation patterns. Key words: ice regime of rivers, long-term changes, climate change, NAO indexes INTRODUCTION Multitude of different factors have to be taken into account in the evaluation of climate change and the patterns of processes in the environment. Records of the exact time of the ice freezing and breakup in rivers allow the assessment of long-term and seasonal variability of climate, especially in relation to climate change (Livingstone 1997; Singh and Kumar 1997; Magnuson et al. 2000). There are three major reasons why ice regime studies are important: a) as the exact dates of freezing and breakup of lakes and rivers have been recorded for many rivers (also in Latvia) well before scientific observation began; the data cover a longer period than other hydrological factors; b) the ice regime of waters models the hydrological regime during the period of maximal discharge after the melting of accumulated precipitation; c) the condition of ice is a sensitive and reliable indicator of climate. The Character of Climate Change 59 It has been observed that the river discharge and ice regimes during winter are related to the North Atlantic oscillation (NAO) pattern (Hurell 1995; Osborn et al. 1999) of large-scale anomalies in the North Atlantic atmospheric circulation. It is also proven that Southern oscillation may influence the ice regime of lakes and rivers in the Northern Hemisphere (Robertson et al.2000). The so-called positive phases of NAO (associated with strong westerly winds and an increased flow of warm and moist air in western Europe) cause late and warmer winters and early springs (Hurrell 1995; Paeth et al. 1999; Chen and Hellström 1999). The air temperature regime and the occurrence of warm rainfall, influenced by the air flows from the North Atlantic (as indicated by NAO), significantly affect the ice regime (Loewe and Koslowski 1998), hence the river discharge pattern (Hurrell and van Loon 1997). The next major factor which possibly influences the ice regime is global warming processes (Ruosteenoja 1986; Singh et al. 2000). The breakup dates on the rivers in the Northern Hemisphere for the last two centuries provide a consistent evidence of later freezing and earlier breakup (Magnuson et al. 2000). In several studies the ice regime trends of inland waters have been analyzed, since easily identifiable parameters describing ice breakup have been recorded for a long period of time (Arai 2000; Benson et al. 2000; Beltaos 1997; Likens 2000). These studies have clearly shown long-term changes of climate, and also the dependence of environmental processes and the ice regime of surface waters in Northern Europe on the NAO (Yoo and D’Odorico 2002). The ice regime of the Baltic Sea has been previously analyzed with the aid of historical time series of ice breakup at the port of Rīga (Jevrejeva 2001) with an aim to reconstruct the climate (Tarand and Nordli 2001). The aim of this study is to assess long-term changes of the ice regime of rivers in Latvia in relation to long-term climate change (temperature and precipitation), river discharge, and NAO oscillation patterns. MATERIALS AND METHODS P. Stakle (1931) published the first data on the time series of ice breakup of the River Daugava. Data on river discharge and ice regime were obtained from the Latvian Environment, Geology and Meteorology Agency, and temperature records for the period 1795 to 2002 from the Meteorological Station Rīga–University. The standard homogeneity test was applied on the data before analyses (Lizuma and Briede 2003). During the study period, sampling and observation methods followed standard procedures and historical observations were re-evaluated to adjust them for the present time-counting system (Stakle 1931). Only original data were used in this study and there were no replacements of the data. Locations of sampling sites and regular monitoring stations are shown in Figure 1. 60 Climate Change in Latvia To investigate links to wide-scale climatic forcing, the extended NAO index was used (Luterbacker et al. 2002). The multivariate Mann-Kendall test (Hirsch et al. 1982, Hirsch and Slack 1984) for monotone trends in time series of data grouped by sites, plots, and seasons was chosen for the determination of trends, as it is a relatively robust method concerning missing data and it lacks strict requirements regarding data heteroscedasticity. Fig. 1. The location of streamflow stations (•) on the studied rivers, and precipitation stations (*) The Mann-Kendall test was applied separately to each variable at each site, at a significance level of p<0.5. A trend was considered as statistically significant at a 5 % level if the test statistic was greater than 2 or less than -2. In this study the seasonal NAO index form year 1709 was used (Luterbacker et al. 2002). Classification of the NAO index data is based on 3 categories: high (NAO >1) – strong westerly, normal (NAO ~ 1) and low (NAO < -1) – weak westerly. RESULTS AND DISCUSSION The climate, hydrological processes, and ice regime of the inland waters of Latvia are determined by the country’s physico-geographic location: a flat surface topography (57 % of Latvia’s territory is below 100 m above sea level), the dominance of Quaternary glacial and ancient sea sediments The Character of Climate Change 61 (moraine loam and sand are parent soil materials), and dominance of humid podsol soil. The climate is humid (mean precipitation ranges from 600 to 850 mm per year) and comparatively cold, and the mean density of the river network is 588 m per 1 km2. The monitoring network of hydrological and ice regime observation covers a significant number of rivers, and long rows of observations have been accumulated (Fig. 1). Long-term data on river discharge do not show significant trends, for example, for the Daugava, Venta, Salaca, and Dubna Rivers (Fig. 2); the process can rather be described as periodic oscillation around mean values. However, for a short term, the changes are significant. For all the period of hydrological observations (more than 100 years for the Rivers Daugava and Venta and more than 60 years for other major rivers), in the last twenty years (1982-2002) the water discharge of rivers in central and eastern Latvia has increased by ~ 4 % (except the River Lielupe), and in the western part of the country by about 6 %. Seasonal discharge regime has also changed significantly, as it will be shown further. Regardless of the long-term patterns of water discharge, the present water flow regime has comparatively increased in comparison to centennial mean values of the rivers of Latvia. The long-term trends of annual river discharge are insignificant. The trends of particular seasons appear to show different patterns (Fig. 3). For the Daugava, Venta, and Lielupe Rivers there is a trend of increase in winter (December – February), but not in other seasons. A very significant increase of river discharge in winter can be observed in the last two decades. The long-term variability of seasonal air temperature is shown in Figure 4. Seasonal air temperatures, according to the records of the meteorological station Rīga-University, have changed substantially during the last 200 years (1795-2002). Air temperature in winter has increased by 1.9 ºC, in spring by 1.3 ºC, and in autumn by 0.7 ºC. Mean annual temperature has increased by 1.0 ºC. In comparison to the reference level (1961-1990), the lowest mean temperature of annual and seasonal temperatures (autumn, spring, and summer) occurred during the period from 1830 to 1930. The temperatures in winter have been increasing gradually since the 19th century, but during the 1830-1930 period the long-term minimum was not reached. Relevant increases of air temperature in winter and spring have been observed since the 1970ies. 62 Climate Change in Latvia Fig. 2. Long-term changes of mean annual discharge of the rivers in Latvia. Discharges were levelled to a 6-year moving average. The Character of Climate Change 63 Fig. 3. Long-term changes of seasonal (winter, spring, and summer) discharge ratios in relation to mean annual discharge in the Daugava, Venta, and Lielupe rivers. 64 Climate Change in Latvia There are direct links between temperature, ice regime on rivers, and their discharge pattern. The time series of dates of ice breakup in the Daugava River at Daugavpils (Table 1) indicates a mean date of April 3. The breakup time can differ by more than a month, depending on the distance from the Baltic Sea and Gulf of Rīga and the river catchment characteristics. Table 1 The basic characteristics and ice regime of the studied rivers of Latvia 2 3 River streamflow station 0,17 (95%) 0,17 breakup Decrease, = day/10year Duration of of Duration Runoff, km Mean date of Mean date of freezing over over freezing p Distance from Basin area, km the Baltic Sea or Gulf of Rīga, km Rīga, of Gulf Average number of of number Average observations, years days with ice cover ice with days Daugava-Daugavpils 370 64500 14,38 1925-2000 24. Nov. 03. Apr. 101 2,8 Lielupe-Mežotne 100 9390 1,76 1921-2000 26. Nov. 27. Mar. 86 3,0 Venta-Kuldīga 60 8320 2,09 1926-2000 02. Dec. 22. Mar. 65 3,2 Gauja-Sigulda 40 8510 2,23 1939-2000 01. Dec. 30. Mar. 78 4,1 Salaca-Lagaste 20 3220 0,97 1927-2000 26. Nov. 29. Mar. 77 5,1 Pededze-Litene 350 978 0,26 1959-2000 17. Nov. 02. Apr. 108 - Bērze-Baloži 50 904 0,16 1960-2000 05. Dec. 14. Mar. 80 - A linear decreasing trend indicates the change of the date of ice breakup to an earlier one (Table 1, Fig. 5, 6). The calculated regression equation estimated that the time of ice cover during the 20th century (observation period of 77-60 years) has changed to an earlier of 2.8 to 5.1 days every 10 years. In general, the change in the river breakup toward earlier dates, indicating an earlier start of flooding, can explain the increase of winter runoff of rivers in Latvia, which is associated with climatic variability, obvious from temperature charts. However, there are clear differences between the studied rivers, and the changes have not been consistent for different time periods. For example, a change toward an earlier ice breakup has not been a typical feature for the entire period of observations for the Daugava (Fig. 5) – the longest observation series in Europe. The downward trend was more expressed during the last 150 years, especially for the last 30 years. No downward trend was detected for the initial period, which includes the Little Ice Age. Lengths of the periods are not equal and mild winters can be followed by hard winters. However, the periodicity cannot be considered a fixed cycle, it is more like a quasi-periodic process (Fig. 5). The Character of Climate Change 65 Fig. 4. Mean annual temperature observed in Rīga in winter (December–February), spring (March–May) and autumn (September–November). The reference level is the mean of the period 1961-1990. 66 Climate Change in Latvia Fig. 5. Time series of ice breakup data in the Daugava River (dashed line shows the trend from 1860 to 2003, and continuous line from 1530 to 1859). Similar trends of ice breakup were noticed examining the data from the Lielupe, Salaca, Venta, and Gauja Rivers (Fig. 6). The applied Man-Kendall test verified that the number of days during which a river is covered with ice has been significantly decreasing. A downward trend was noticed for all the seven selected rivers, which are located in different parts of Latvia (Table 2, Fig. 7). The data were statistically significant (less than –2) for the Salaca, Gauja, and Bērze. A relevant correlation between days with ice cover and NAO was also found for the Daugava and Pededze Rivers. The length of ice cover on rivers was negatively correlated with NAO winter indexes (December – March) since the middle of the 20th century (Table 2). However, there were some periods when these correlations were insignificant (for example, in the middle of the 19th century, and the transition period from the 19th to the 20th century). The relation between the time series of data of ice breakup in the Daugava River and the NAO indexes was analysed, using integral curves (Figure 8). The integral curves for ice breakup and NAO indexes calculated for the last 300 years allowed to identify a long-term variability (40-90 years). In addition, the directions of these fluctuations are opposite – the increase of one indicator is related to the decrease of the other. The Character of Climate Change 67 Fig. 6. Time series of ice break-up dates in the Lielupe, Salaca, and Venta Rivers. Fig. 7. Length of ice cover (days) in the rivers of Latvia: l- Daugava; 2-Lielupe; 3-Gauja; 4- Salaca; 5- Venta. Time series are levelled to a 6-year moving average. The processes over the North Atlantic appear to have a significant influence on the climate of Latvia, especially in winter (December, January, and February) or the cold season (October – March (Fig. 8-10). 68 Climate Change in Latvia Table 2. The trend statistics for days with ice cover in rivers and the correlation of ice-cover period with NAO indexes (December – March). Bērze Venta Gauja Salaca Lielupe Pededze Daugava 1959-2000 1939-2000 1925-2000 1926-2000 1926-2000 1920-2000 1960-2000 Correlation with NAO - index -0,52 -0,54 -0,44 -0,62 -0,57 -0,60 -0,70 p-value (one-sided test) 0,001 <0,001 <0,001 <0,001 <0,001 <0,001 <0,001 Mann-Kendall test criteria* -1,533 -1,849 -2,547 -1,737 -2,146 -1,831 -2,467 p-value (one-sided test) 0,063 0,032 0,005 0,041 0,016 0,034 0,007 *- The trend can be considered as statistically significant at a 5 % level if the test statistics is greater than 2 or less than -2 Fig. 8. Curves of NAO time series (October to April) and days with a negative sum of temperatures. Figure 8 shows the winter period series of NAO index in the 20th century and the sum of negative temperatures in Rīga. In Figure 8, winter refers to the values from December to February. A strong negative correlation between the NAO index and the sum of negative temperatures and monthly mean temperatures shows that the processes over North Atlantic are the driving force for climate in winter in the territory of Latvia. According to our data, spring and summer temperatures do not appear to be related to the NAO The Character of Climate Change 69 index (Fig. 9). Strong correlations with NAO indexes are typical only of the cold period: winter (r = 0.81) and autumn (r = 0.44). Fig. 9. Correlation coefficient between NAO and mean monthly temperatures. Fig. 10. Correlation coefficient between NAO and the precipitation sum in the warm and cold periods for seven sampling stations in Latvia. 70 Climate Change in Latvia We assume that in the warm period the amount of precipitation in Latvia is linked with a different atmospheric circulation pattern, which is at present unclear (Fig. 10). The NAO has been observed to influence winter precipitation with varying intensity along the Norwegian coast, in Northern Sweden, and in Southern Finland where mountain relief plays an important role (Uvo 2003). The stable relationship found in our study highlights the strong linkage between large-scale NAO forcing and regional scale climate processes in Latvia. Moreover, the negative correlation between winter temperatures and NAO indexes has become stronger during the last 100 years. CONCLUSIONS Several conclusions can be drawn concerning climatic and hydrological parameters and ice breakup dates in the rivers of Latvia: • River discharge has undergone periodical fluctuations and the observed long-term trends of river discharge are significant for winter season, especially in the last 30 years. • The study confirms an increase of winter runoff in relation to the total runoff of the studied rivers. • A significant temperature increase trend was observed for the winter period (1.9 oC since 1795). A linear change is pronounced, overlapping with periodical oscillations. • The ice cover period in the selected rivers has been decreasing. The reduction of the ice-cover period for the last 30 years is 2.8 up to 5.1 days every 10 years. The time of ice breakup depends not only on the meteorological conditions in the particular year and the distance from the Baltic Sea, but also on global climate change. The trends are not consistent between periods, and mild and hard winters clearly alternate. The periodicity cannot be considered as a fixed cycle, it is more like a quasi-periodic process. REFERENCES Arai T. (2000) The Hydro-climatological Significance of Long-term Ice records of Lake Suwa, Japan. Verh. Internat. Verein Limnol., 27, 2757-2760 Beltaos S. (1997) Onset of River Ice Breakup. Cold Region. Sci. Technol., 25, 183-196 Benson B.J., Magnuon J.J., Jacob R.L., Fuenger S.L. (2000) Response of Lake Ice Breakup in the Northern Hemisphere to the 1976 Interdecadal Shift in the North Pacific. Verh. Internat. Verein Limnol., 27, 2770-2774 Chen D., Hellström C. (1999) The Influence of the North Atlantic Oscillation on the Regional Temperature Variability in Sweden: Spatial and Temporal variations. Tellus, 51A, 505-516 The Character of Climate Change 71 Hurrell J.W., van Loon H. (1997) Decadal Variations in Climate Associated with the North Atlantic Oscillation. Clim. Change, 36, 301-326 Jevrejeva S. (2001) Severity of Winter Seasons in the Northern Baltic Sea Between 1529 – 1990: Reconstruction and Analysis. Clim. Res., 17, 55-62 Likens G.E. (2000) A Long-term Record of Ice Cover for Mirror Lake, New Hampshire: Effects of Global Warming? Verh. Internat. Verein Limnol., 27, 2765-2769 Livingstone D.M. (1997) Break-up Dates of Alpine Lakes as Proxy Data for Local and Regional Mean Surface Air Temperature. Clim. Change, 37, 407 – 439 Lizuma L., Briede A. (2003) Ilggadīgo (1795-1996) gaisa temperatūras novērojumu rindu viendabības izvērtējums. LU 61. Zinātniskā konference. Ģeogrāfija. Ģeoloģija. Vides zinātne. Referātu tēzes. Rīga, 84.-87. lpp Loewe P., Koslowski G. (1998) The Western Baltic Sea Ice Season in Terms of a Mass-related Severity Index 1879-1992 (II) Spectral Characteristics and Association with the NAO, QBO, and Solar Cycle. Tellus, 50(A), 219-241 Luterbacker I., Huplaki E., Dietrich D., Jones P.D., Davies T.D., Portis D., Gonzalez-Rouco I.F., von Storch H., Gyalistras D., Casty C., Waner H. (2002) Extending North Atlantic Oscillation Reconstructions Back to 1500. Royal Meteorological Society Atm. Sc. Let., 2; 114-124 (http://www.cru.uea.ac.uk/cru/data/naojurg.htm) Magnuson J.J., Robertson D.M., Benson B.J., Wynne R.H., Livingstone D.M., Arai T., Assel R.A., Barry R.G., Card V., Kuusisto E., Granin N.G., Prowse T.D., Stewart T.D., Vuglinski V.S. (2000) Historical Trends in Lake and River Ice Cover in the Northern Hemisphere. Science 289, 1743-1746 Osborn T.J., Briffa K.R., Tett S.F.B., Jones P.D., Trigo R.M. (1999) Evaluation of the North Atlantic Oscillation as Simulated by a Coupled Climate Model. Clim. Dynamics, 15, 685-702 Paeth H., Hense A., Glowienka-Hense R., Voss R., Cubash U. (1999) The North Atlantic Oscillation as an Indicator for Greenhouse-gas Induced Regional Climate Change. Clim. Dynamics, 15, 953-960 Robertson D.M., Wynne R.H., Chang W.Y.B. (2000) Influence of El Niño on Lake and River Ice Cover in the Northern Hemisphere from 1900 to 1995. Verh. Internat. Verein Limnol., 27, 2784-2788 Ruosteenoja K. (1986) The Date of Break-up of Lake Ice as a Climatic Index. Geophysica, 22, 89-99 Singh P., Kumar N. (1997) Impact Assessment of Climate Change on the Hydrological Response of a Snow and Glacier Melt Runoff Dominated Himalayan River. J. Hydrol., 193, 316-350 Singh P., Kumar N., Arora M. (2000) Degree-day Factors for Snow and Ice for Dokriani Glacier, Garhwal Himalayas. J. Hydrol., 235, 1-11 Stakle P. (1931) Hidrometriskie novērojumi Latvijā līdz 31.X1929. Finansu ministrijas Jūrniecības departaments. Rīga, 374 lpp. Tarand A., Nordli P.O. (2001) The Tallinn Temperature Series Reconstructed Back Half a Millenium by Use of Proxy Data. Clim. Change, 48, 189-199 Yearly bulletins of surface waters in Latvia. 1920-2002. Latvian Hydrometeorological Agency (in Russian, Latvia) Uvo C.B. (2003) Analyses and Regionalization of Northern Europe Winter Precipitation Based on Its Relationship with North Atlantic Oscillations. Int. J. Climatol., 23: 1185-1194. Yoo J.C., D’Odorico P. (2002) Trends and Fluctuations in the dates of Ice Break-up of Lakes and Rivers in Northern Europe: the Effect of the North Atlantic Oscillation. J.Hydrol., 268, 100-112 72 Climate Change in Latvia FOOTNOTES * Modified from publication in Proceedings of Latvian Academy of Sciences The Character of Climate Change 73 Long-term Changes of Snow Cover in Latvia as an Indicator of Climate Variability* Anita Draveniece, Agrita Briede, Valērijs Rodinovs, Māris Kļaviņš University of Latvia, Faculty of Geographical and Earth Sciences, Raiņa blvd. 19, LV 1586, Rīga, Latvia Long-term (1945-2004) changes of snow cover in Latvia have been studied using data of 28 observation sites. Over the territory, there are remarkable differences in snow cover duration: from 70 days in the western part up to 136 days in the uplands of eastern Latvia. A general decrease in snow cover has been observed in Latvia, including the decrease of the duration of snow cover. Moreover, both duration and snow cover depth exhibit not only regional variability, but also a marked oscillatory pattern. The snow cover, the temperature regime, and the amount of precipitation is very much influenced by large scale atmospheric oscillation over North Atlantic. Key words: snow cover, long-term changes, Latvia, NAO index, Baltic winter climate index INTRODUCTION Multitude of different factors have to be taken into account to evaluate climate change and different kinds of processes in the environment. Records of the dates of snow cover appearance and parameters allow to assess the long-term and seasonal variability of climate, especially in relation to climate change (Easterling et al. 2000; Dye 2002). There are several reasons why snow regime studies are important, and these are as follows: a) the exact dates of the beginning of snow season have been recorded well before the beginning of regular scientific observations, as a result, the data cover long period of observations; b) the snow cover parameters are of importance for hydrological regime during the period of maximal discharge of accumulated atmospheric precipitation; c) snow cover conditions are a sensitive and reliable indicator of climate (Gutzler et al. 1992; Groisman et al. 1994; Huntington 74 Climate Change in Latvia et al. 2004; Falarz 2004; Ye 2000). Snow cover affects the Earth’s surface energy and hydrological budgets, influences terrestrial ecosystems (soil and air temperatures, soil water recharge, the availability of radiation to ground- layer vegetation) and biogeochemical cycles of elements (Gerland et al. 2000; Arora and Boer 2001; Dye and Tucker 2003). Snow cover season can directly influence the prolongation of the growing season of vegetation (Menzel and Fabian 1999). Satellite and surface observations have revealed that the spatial extent of annual snow cover in Northern Hemisphere land areas has decreased over the last decades, at first in the spring season (Brown 2000; Ye 2000; Falarz 2004). Both anthropogenically forced atmospheric warming and natural climate variability can serve as the source of snow cover changes (Brown 2000; Dye 2002). The studies of the annual timing of snow cover disappearance and onset, and the duration of the snow-free period can provide important information about climatic change, ecosystem changes, and the regional importance of these processes. While many regionally oriented studies have been performed worldwide (Ye 2000; Schwartz and Reiter 2000; Falarz 2004), in Latvia snow cover has not been much studied (Barloti 1932; Темникова 1958; Zirnītis 1962; Draveniece 1998). The aim of this study is to assess the long-term changes of the snow cover in Latvia in relation to long-term climate change (temperature and precipitation) and large scale atmospheric oscillation processes over North Atlantic. MATERIALS AND METHODS Data on snow cover were obtained from the Latvian Environment, Geology and Meteorology Agency (Anonymous 1945-2004). Only original data were used in this study, and there were no replacements of the data. The location of meteorological and gauging stations is shown in Figure 1. To investigate the links to wide-scale climatic forcing, we used the extended North Atlantic oscillation (NAO) index (Luterbacker et al. 2002) and the recently suggested Baltic winter index (WIBIX) (Hagen and Feistel 2005). The classification of the NAO index data is based on 3 categories: high (NAO >1) – strong westerly, normal (NAO ~ 1) and low (NAO < -1) – weak westerly. The multivariate Mann-Kendall (Hirsch et al. 1982; Hirsch and Slack 1984) test for monotone trends in time series of data grouped by sites, plots and seasons was chosen for the detection of trends, as it is a relatively robust method concerning missing data and it does not require strict requirements regarding data heteroscedasticity. The Mann-Kendall test was applied separately to each variable at each site, at a significance level of p<0.05. A trend was considered as statistically significant at a 5 % level if the test statistic was greater than 2.0 or less than -2.0. The Character of Climate Change 75 Fig. 1. The location of meteorological (■) and gauging stations (▲) used for the study. 1-Pāvilosta, 2-Liepāja, 3-Kolka, 4-Dobele, 5-Ventspils, 6-Saldus, 7-Mērsrags, 8-Ainaži, 9-Jelgava, 10-Bauska, 11- Stende, 12-Skulte, 13-Rīga, 14- Daugavpils, 15- Priekuļi, 16-Zīlāni, 17- Rūjiena, 18-Rēzekne, 19—Dagda, 20-Skrīveri, 21-Gulbene, 22- Zosēni, 23-Alūksne, 24- Rucava, 25- Vičaki, 26- Praviņi, 27-Višķi, 28-Dzērbene. RESULTS AND DISCUSSION The regional distribution of snow cover is closely related to the air temperature distribution over the territory of Latvia. Moving inland from the coastline of the Baltic Sea and the Gulf of Rīga, the duration of snow cover and snow depth increases, particularly in the regions where the surface topography, steepness, and the orientation of upland slopes towards the prevailing winds force an upward movement of air masses (Krauklis and Draveniece 2004). The large body of water of the Baltic Sea exerts a strong influence on the extent and variability of the snow cover over Latvia, reaching along the coastal belt, about 30-100 km wide. It shows, for example, in the mean number of days with snow cover (Fig. 2), as well as in the mean and minimum number of days with snow cover (Fig. 3-6). Snow cover is unevenly distributed over the territory of Latvia. 76 Climate Change in Latvia Fig. 2. Mean duration of snow cover (days) in Latvia over the period 1945-2004. Fig. 3. Mean minimal duration of snow cover (days) in Latvia over the period 1945-2004. The Character of Climate Change 77 Fig. 4. Duration of snow cover (days) in Latvia in 1995/1996 winter (severe winter). The variability of snow cover duration (number of days with snow cover) in the northern part of Latvia is less pronounced than in the western and south- western part (Fig. 5). The timing of snow cover onset and disappearance, as well as the duration of continuous snow cover, are very dynamic in Latvia. The onset of snow cover in autumn over the territory of Latvia takes place in November 1-23: along the Baltic Sea coastline it is first established in the 3rd decade of November (20-23 November), along the Gulf of Rīga, the Zemgale Plain, and Kurzeme Upland – in the 2nd decade of November. The earliest autumn snow cover appears in Vidzeme Upland and in the regions to the north of it, the East Latvia Lowland and the Latgale Upland, in the 1st decade of November. According to long-term records, the mean date of snow cover onset in Alūksne is November 1, in Dagda and Zosēni – November 5. A continuous snow cover in Latvia begins 30-44 days after the onset of the first snow cover in autumn. Besides, to the east of an imaginary line connecting the towns Skrīveri, Priekuļi, and Rūjiena, pre-winter lasts for 30- 39 days, but over the rest of Latvia it is a week longer. A continuous snow cover develops from December 5 to January 6 (Alūksne and Ventspils); over the most part of the territory it develops within the 2nd and 3rd decade of December. In some winters the process of the development of continuous snow cover lasts for almost 3 months. For instance, in the winter of 1987/88, a continuous snow cover formed in Dagda only on February 3, in Zosēni – on February 9, but the western regions did not see a continuous snow cover at all. 78 Climate Change in Latvia Fig. 5. Mean number of days (average, maximum and minimum) with snow cover over the period 1945-2004. Fig. 6. Mean depth (cm) of snow cover in winter 1995/1996 (severe winter). The geographical location of a site also affects the snow depth (both seasonal average snow depth and maximum decade snow depth) and has to be considered the leading factor affecting the snow cover (Fig. 6). For The Character of Climate Change 79 example, in a very snowy winter (1995/1996), a permanent snow cover stood all over Latvia throughout March reaching the depth of 64 cm in Vidzeme Upland. It should be mentioned that there is a wide variation of maximum snow depth: almost half of all winters are warm, nevertheless, the others are colder than average or even severe. The average number of days with snow cover grows towards the eastern part of Latvia: from 69 days in the coastal territories of Kurzeme (Rucava) to 134 days in Alūksne, and the number of days with continuous snow cover comprise accordingly 67 and 116 days. The investigation of mean duration of snow cover showed that snow covers the ground for more than 100 days in the territories to the east of an imaginary meridional line connecting the towns of Skrīveri, Priekuļi and Rūjiena. The highest values of the total number of days with snow cover, the number of days with continuous snow cover, and the snow depth have been recorded at Alūksne Upland. To the west from the above-mentioned imaginary line mean snow depth comprises 6-12 cm, but to the east of it – 17-29 cm (Priekuļi and Alūksne, respectively). The eastern part of Latvia differs from the western part not only in considerably greater snow depth, but also in much wider variability. Fig. 7. The statistics of the Mann-Kendall test for a trend in the number of snow cover days over Latvia (1945-2004). 80 Climate Change in Latvia The duration of snow cover shows a decreasing trend. The decrease is expressed by the number of days. The linear regression analysis shows that within a 50-year period the duration of snow cover in Latvia has decreased by 3-27 days (Fig. 8). A considerable decrease was detected in Zosēni, Rūjiena, Ainaži. Along the coast of the Baltic Sea and in Kurzeme, the snow cover duration has decreased on average by 16 days. The least changes have taken place in Rīga, Alūksne, Gulbene, Jelgava, and Daugavpils. Three of these monitoring sites (Alūksne, Gulbene, Daugavpils) are located in the eastern part of the country, characterized by a larger degree of continentality. A very small change in Rīga and Jelgava might be explained by the specific character of city microclimate and the growing urbanisation. Fig. 8. The decrease in the duration of snow cover (days) in Latvia for the period 1945-2004. The variability of snow cover duration shows a linear trend, however, a marked periodicity can be observed (Fig. 9, 10). The periodicity is displayed in the changes of snow parameters and it is apparent all over Latvia irrespective of geographical factors (the location of the observation site). The character of long-term variability in snow cover duration and depth is similar for the most distant observation sites (Fig. 10). There has been a negative correlation between the length of a period with snow cover and NAO winter indexes (December – March) as well as between the length of a period with snow cover and Baltic winter climate index since the middle of the 20th century (Table 1). The Character of Climate Change 81 Table 1. The correlation of days with snow cover and NAO indexes (December – March), and Baltic winter climate (WIBIX) index. Bauska Rūjiena Pāvilosta Priekuļi Alūksne Saldus NAO winter index -,665** -,174 -,215 -,180 -,440** -,466** WIBIX index -,697** -,300* -,160 -,285* -,524** -,470** ** Correlation is significant at a 0.01 level (2-tailed) * Correlation is significant at a 0.05 level (2-tailed). Fig. 9. Fluctuations of snow cover depth and the number of snow cover days during the last 50 years (5-year moving average). Regular snow cover observations in Latvia began already in 1891 (Jelgava, Kolka), as a result, 10 of the existing 23 meteorological stations have long data series (> 75 years). Besides, snow cover is observed at gauging stations, too. However, there are few studies on snow cover in Latvia (Barloti 1932; Zirnītis 1963; Teмникова 1958), while today most studies on climate change recognize the major role of snow cover variability (Brown 2000; Huntington et al. 2004; Falarz 2004; Ye 2000). 82 Climate Change in Latvia Fig. 10. Variations of the number of days with snow cover (1945-2004). ▬▬ - average in Latvia and 1-Alūksne; 2- Zosēni; 3- Bauska; 4- Liepāja (5-year moving average). Snow cover is considered to be a sensitive indicator of climate change because the snow depth, its extent and duration directly depend on near- surface air temperature and the amount of precipitation, and provides a good feedback on climate. Moreover, the snow cover duration and its depth have a pronounced effect on the growth of natural and cultivated plants. The seasonal snow cover stores large quantity of water, which afterwards greatly influences the surface water and groundwater cycle, hence the snow cover characteristics are directly related to the development of hydro-electric power generation or assessment of the flood risk. Finally, a good snow cover is an essential factor for the development of recreation and winter sports. In winter season, when western circulation is more vigorous than in any other season and it moves extra-tropical cyclones and air masses rapidly through the mid-latitudes, Latvia’s weather is highly inconsistent. As a result, almost half of all winters are mild and the others are colder than average or even severe. Each year there is a considerable deviation from long-term mean of the snow cover duration and depth. In December – January the incoming short-wave radiation reaches a minimum and durable cloud cover resulting from frequent frontal passages and establishing air mass cloudiness cuts off The Character of Climate Change 83 direct solar radiation thus fostering the forming of permanent snow cover. The latter forms 4-6 weeks after the formation of the first snow cover; on average in December 8-29, except the coastal territories adjacent to the Baltic Sea, where it forms in the first decade of January (Krauklis and Draveniece 2004). The long-term variability of snow cover can be considered an important climate change indicator. The snow cover duration was tested with the help of the nonparametric Mann-Kendall test, which is often used to investigate trends in hydro-meteorological time series which have high seasonal variability and contain outliers. A trend was considered statistically significant if the Mann-Kendall test statistic was greater than 2.0 (positive or negative). The calculations of trends in long-term snow cover duration (Fig. 7) showed that over a 50-year period (1946-1996) the length of snow cover in Latvia has decreased, except for some observation sites (Liepāja and Dzērbene). Statistically significant trends or pronounced negative tendency in snow cover duration have been found at the following observation sites: Dagda, Rēzekne, Ainaži, Dobele, Vičaki, Rūjiena, Zosēni. The variability of snow cover duration and depth demonstrates pronounced periodicity. This phenomenon has been discovered in a number of countries of the Northern Hemisphere (Brown 2000; Huntington et al. 2004; Falarz 2004; Ye 2000), and it is being related both to changes in solar activity and changes in atmosphere circulation processes over the Atlantic Ocean; the latter directly attributed to factors affecting climate change. Although the variability of snow cover is of an oscillating character, a decreasing trend can be detected using linear regression analysis. The periodically recurring processes also show a tendency to vary. These variations manifest themselves through less frequent occurrence of severe winters and the decrease of snow depth during severe winters, and also through more frequent mild winters with short snow cover duration and thinner snow cover. Another aspect of the periodicity of snow cover change is that even though there is a general increase in average air temperature, there may be periods with durable and thick snow cover during winter. In terms of snow cover duration and depth variations the present period is distinctive because of a decreasing trend in the duration of continuous snow cover, not only in long-term perspective but also in long-term periodicity. The processes over the North Atlantic appear to have a significant influence on the climate of Latvia, especially in winter (December, January, and February) or the cold season (October – March, Fig. 9-10). Changes in snow cover can also be directly related to the recently derived (Hagen and Feistel 2005) Baltic winter climate index (WIBIX), which is based on the monthly values of the principal component of: a) winter anomalies (January – March) of air pressure difference between Gibraltar and Reykjavik to describe 84 Climate Change in Latvia the North Atlantic oscillation; b) sea level anomalies of Landsort (Sweden) to characterize the filling level in the Baltic proper; c) maximum Baltic ice cover to include the influence of continentally dominated alignments of atmospheric centers of action. A strong negative correlation between the NAO index, WIBIX index, and the number of days with the snow cover shows that the processes over North Atlantic are the driving forces for climate in winter over the territory of Latvia. While in winter an intense westerly circulation moves fronts and air masses through the mid-latitudes, in the warm period it weakens considerably and the amount of precipitation in Latvia is related to a different atmospheric circulation pattern. A study of precipitation over the Baltic sea area based on radar data showed that in the cold season the greatest part (55-70 %) of precipitation was frontal, while in summer months non-frontal precipitation comprised 60-80 % (Walther et al. 2003). The NAO has been observed to influence winter precipitation and thus the snow cover with varying intensity along the Norwegian coast, in northern Sweden, and in southern Finland where mountain relief plays an important role (Uvo 2003). The strong relationship found in our study highlights the stable linkage between large-scale NAO forcing and regional scale climate processes in Latvia. REFERENCES Anonymous (1945 – 2004) Meteoroloģiskās pamattabulas, LVĢMA datu fonds Arora V.K., Boer G.J. (2001) Effects of Simulated Climate Change on the Hydrology of Major River Basins. J.Geophys. Res., 106, 3335 - 3348 Barloti J. (1932) Sniega sega Latvijā 1921-1931 Rīga, Zemkopības ministrija. 39 lpp. Brown R.D. (2000) Northern Hemisphere Snow Cover Variability and Change 1915- 97. J.Clim., 13 (13), 2339 – 2355 Draveniece A. (1998). Sniega segas raksturojums Latvijā. Maģistra darbs. Rīga: Latvijas Universitāte. 107 lpp. Dye D.G. (2002) Variability and Trends in the Annual Snow-cover Cycle in Northern Hemisphere Land Areas, 1972-2000. Hydrol. Proc., 16, 3065-3077 Dye D.G., Tucker, C.J. (2003) Seasonality and Trends of Snow Cover, Vegetation Index, and Temperature in Northern Eurasia. Geophys. Res. Lett., 30(7), 581 - 584 Easterling D.R., Karl, T.F., Gallo, K.P., Robinson, D.A., Trenberth, K.E., Dai, A. (2000) Observed Climate Variability and Change of Relevance to the Biosphere. J. Geophys. Res., 105, 20101 – 20114 Falarz M. (2004) Variability and Trends in the Duration and Depth of Snow Cover in Poland in the 20th Century. Int. J.Climatol., 24(13), 1713-1727 Gerland S., Liston G. E., Winther J.G., Orbaek J. B., Ivanov B. V. (2000) Attenuation of Solar Radiation in Arctic Snow: Field Observations and Modelling. Ann. Glaciology, 31, 364 - 368 Groisman P.Ya., Karl T.R., Knight R.W. (1994) Observed Impact of Snow Cover on the Heat Balance and the Rise of Continental Spring Temperatures. Science, 263 (14), 198-200. The Character of Climate Change 85 Gutzler D.S., Rosen R.D. (1992) Interannual Variability of Wintertime Snow Cover across the Northern Hemisphere. J. Clim., 5 (12), 1441-1447. Hagen E., Feistel R. (2005) Climatic Turning Points and Regime Shifts in the Baltic Sea Region: the Baltic Winter Index (WIBIX) 1659-2002. Boreal Environ. Res., 10, 211 - 224 Huntington T.G., Hodgkins G.A., Keim B.D., Dudley R.W. (2004) Changes in the Proportion of Precipitation Occurring as Snow in New England (1949-2000). J.Clim., 17 (13), 2626 – 2636 Falarz M. (2004) Variability and Trends in the Duration and Depth of Snow Cover in Poland in the 20 th century. Int. J. Climatol., 24(3), 1713 – 1727 Krauklis A., Draveniece A. (2004). Landscape Seasons and Air Mass Dynamics in Latvia. Ģeogrāfiski raksti/Folia Geographica, 12, 16-48. Luterbacker I., Huplaki E., Dietrich D., Jones P.D., Davies T.D., Portis D., Gonzalez-Rouco I.F., von Storch H., Gyalistras D., Casty C., Waner H. (2002) Extending North Atlantic Oscillation Reconstructions Back to 1500. Atm. Sc. Let., 2; 114-124 (http://www.cru.uea. ac.uk/cru/data/naojurg.htm) Menzel A., Fabian P. (1999) Growing Season Extended in Europe. Nature, 397, 659 Schwartz M.D., Reiter B.E. (2000) Changes in North American Spring. Int. J. Climatol., 20, 929 - 932 Zirnītis A. (1963). Latvijas PSR klimats. Rīga, Latvijas Valsts izdevniecība, 92 lpp. Uvo C.B. (2003) Analyses and Regionalization of Northern Europe Winter Precipitation Based on Its Relationship with North Atlantic Oscillations. Int. J. Climatol., 23, 1185-1194. Walther A., Bennartz R., Fischer J. (2003). Radar-based Precipitation Classification in the Baltic Sea Area. http://www.fu-berlin.de/iss/extended_p3c.3_walther.pdf Ye H. (2000) Decadal Variability of Russian Winter Snow Accumulation and Its Associations with Atlantic Sea Surface Temperature Anomalies. Int. J.Climatol., 20(14), 1709 - 1728 Темникова Н.С. (1958). Климат Латвийской ССР. Рига, Изд-во АН, 232 с. FOOTNOTES * Modified from publication in Proceedings of Latvian Academy of Sciences 86 Climate Change in Latvia Climate Change Impacts on Hydrological Processes in Latvia Tatjana Koļcova, Lita Lizuma, Svetlana Rogozova, Marta Smith Latvian Environment, Geology and Meteorology Agency, Maskavas str. 165, LV-1019, Rīga, Latvia Climate change and changing meteorological conditions are closely related to changes in hydrological processes and have a direct impact on the production of hydroenergy. Long-term flow analysis is essential for effective water resource management; therefore it has an immense socio- economic significance. In this study the trend analysis of meteorological and hydrological data of rivers in Latvia was carried out. The Mann-Kendall test with a 5 % significance level was applied to the data series for trend analysis. To show the interrelationship between climate and the water resources system, the changes in the annual and seasonal temperature and precipitation as well as snow characteristics were analysed with an emphasis on the trends. For water flow data analysis, three periods were considered: 1921- 2003, 1941-2003, 1961-2003. Due to the interruptions in the observation, air temperature and precipitation were analysed in periods: 1925-2003, 1945-2003, 1961-2003. Snow cover data series were analysed in period 1951-2003. The temperature data series and the amount of precipitation show a significant positive trend for annual, winter, and spring periods. The maximum of the water equivalent of snow is decreasing. The significant changes in the magnitude and timing of floods and the seasonal and annual water runoff in the period 1961-2003 due to the rising air temperature and precipitation have caused an increase in hydropower production. Key words: climate change, river discharge, Latvia INTRODUCTION Climate change leads to significant changes in water resources and further affects hydropower production, so it is of national importance for Latvia inasmuch as hydroenergy is the most important source of renewable energy. Depending on hydrological regime and taking into account hydrological The Character of Climate Change 87 regionalisation (Glazacheva 1990), rivers in Latvia could be grouped into three major types – “marine”, “transitional”, and “continental”. The main source of feeding of the “marine” type rivers is precipitation, which exceeds 50 %; snow melt and groundwater sources form correspondingly 40 % and 10 % of water runoff. Due to the frequent thaws during wintertime, “marine” rivers have frequent winter floods, some of which are higher than spring floods. For the rivers of “continental” type the snow melt water rate is almost equal to that of groundwater. The “continental” rivers have a typical hydrological regime like most of the East-European rivers, with the maximum flood from snow melt in the spring. More than 40 % of the annual flow occurs during this period. The type of feeding of the “transitional” rivers is mixed. The snow melt and rain contribution vary from 45 % to 50 % of the whole runoff; groundwater source comprises about 5 %. The hydrological regime is characterised by an irregular runoff distribution over the year and a small portion of groundwater. The aim of this study is to elaborate the statistical analysis of the variability of hydrological and climatic data (streamflow, precipitation, temperature) under different climate change scenarios. MATERIALS AND METHODS In this study air temperature and precipitation data were analysed for periods 1925-2003, 1945-2003, and 1961-2003. Snow cover is a very important element in water balance. The data of maximum snow water equivalent, number of days with snow cover, and the date of the disappearance of stable snow cover for the observation period 1951-2003 were also employed in the analysis. 43 precipitation data series from hydrological stations were also analysed for the period 1945-2003. At present the Latvian Environment, Geology and Meteorology Agency’s (LEGMA) observation network includes 35 hydrological stations in order to obtain regular water discharge data in rivers. 3 periods were selected for the data series analysis: 1922-2003, 1941-2003, and 1961-2003. Only natural variability was studied, so only the data series unaffected by human activities were selected. The Standard Normal Homogeneity Test (SNHT) (Alexandersson and Moberg 1997) was used for homogeneous annual and seasonal data series of temperature and precipitation. The double-mass plot technique was used to assess the streamflow data series homogeneity. The Mann-Kendall test with a 5 % significance level was applied to the data series for trend analysis. Frequency of annual water flow was determined using integral functions of data deviations from the average runoff. 88 Climate Change in Latvia RESULTS AND DISCUSSIONS The study of trends in streamflow for the periods 1922-2003, 1941-2003, and 1961-2003 revealed that for large parts of Latvia and other Baltic and Nordic countries the streamflow has increased in the periods 1941-2003 and 1961-2003, but not for the longest period 1922-2004 due to some cold and wet years in the 1920s (Hisdal 2006). In general, annual average temperature has been increasing during the three studied periods. The winter and spring seasons show significant positive trends for the whole territory of Latvia. For the summer season, positive trends at a confidence level of 70 % for the period 1945-2003 and at a confidence level of 95 % for the period 1961-2003 were found. The average autumn temperature shows no trend for all periods. The annual precipitation of 1925-2003 shows a positive and also a negative trend in the western part and a negative trend in the most of the meteorological stations in the north-east of the country (Fig. 2). Significant positive trends of annual precipitation were found for the periods 1945-2003 and 1961-2003. Fig. 1. Annual temperature and precipitation data series trends accordingly to Mann-Kendal test criteria. Figure 1 shows the statistical significance of annual temperature and precipitation trends. The horizontal black and grey lines show the calculated trends and significance restrictions. Winter precipitation has increased in all periods. Spring precipitation was increasing in 1945-2003 and 1961-2003, mainly in the western part of the country. For the summer season a negative trend appears in the northern part The Character of Climate Change 89 in the period 1925-2003 and in the eastern part in 1945-2003. In autumn, a positive trend was evident only in the western part of the country in the period 1945-2003. Negative trends in autumn precipitation were found for all periods. The changes in snow conditions occur as the result of the changes in air temperature and precipitation during the cold season. The rise of air temperature during winter and spring cause the snow storage to decline even if winter precipitation has increased. The maximum water equivalent, the number of days with snow, and the length of the period with a stable snow cover have been decreasing during the period 1951-2003. All the stations showed negative trends at a 95 % significance level. The earlier date of the disappearance of snow cover was the result of the increasing temperature in spring. Fig. 2. Trends in annual precipitation, air temperature, and runoff data series. The water runoff reflects the tendencies in precipitation series (Fig. 2). Long-term annual runoff has been slightly decreasing in the period 1922- 2003. Only the data series from the stations in the coastal zone of the Baltic Sea had a weak positive trend in accordance with the increased amount of precipitation. Significant positive changes are present in the annual runoff of the period 1961-2003, in 83 % of the total number of analysed data series. The rest of the series also have a positive trend (70 % confidence level). In the spring flood data series a significant negative trend is present in all the observation periods. There is no trend at all in the coastal region where the spring floods are formed by rain and snowmelt water in a different percentage. The “rearranging” of the feeding in the coastal zone should be noted taking into account the significant increase in winter precipitation and the rising temperature during winter that leads to the decreasing of water equivalent of snow (Koļcova and Rogozova 2006). 90 Climate Change in Latvia Fig. 3. Annual flow, spring flood, and drought data series trends. The trend analysis of the date of the spring flood culmination shows an earlier occurrence of the flood peak. In summer flood, the trend is negative for the longest series of peak value. There is no systematic pattern in these series for a short period of 1961-2003 and in the series of the summer flood’s date of culmination. The trends in the series of the floods’ peak value are shown in Fig. 4. Fig. 4. Spatial distribution of spring (right) and summer (left) flood trends for the periods 1922-2003 (above) and 1961-2003 (below). The Character of Climate Change 91 The number of drought data series is comparatively smaller due to great anthropogenic impact on the base flow: melioration and water adjustment with the dams since the 1960s (Reihan et al. 2006). 30-day minimum discharge averages were used for the minimum flow regime analysis in order to avoid occasional errors and to obtain more reliable calculation results. Fig. 5. Trends in low flow series for the period 1922-2003 (left) and 1961-2003 (right). The low flow (Fig. 5) has been slightly increasing in the coastal zone of Latvia with weak negative changes or no changes over the rest of the territory of Latvia for all analysed periods. As to the seasonal water runoff (Fig. 6), a significant positive trend was found in winter for all periods and in all regions. Fig. 6. Seasonal water runoff data series trends. For the longest period, there is no trend in the spring streamflow in the coastal regions. The coastal and transitional regions of Latvia have both weak 92 Climate Change in Latvia and significant negative trends. For the period of 1961-2003 a positive trend appears on the rivers with lake adjustments (the Salaca River – Lagaste, the Rēzekne River – Griškāni, the Irbe River – Vičaki). There are no systematic patterns in the summer streamflow for the long periods of 1922-2003 and 1941-2003; both negative and positive trends could be observed, and even no trends at all. For the short period of 1961-2003 positive trends (weak and strong) were found in large parts of the territory of Latvia. For autumn during the long period of 1922-2003 there is the tendency of the reduction of flow in the continental region of Latvia. In the rest of the territory the trend is not found. The periods of 1941-2003 and 1961-2003 are characterised by a positive trend in some continental parts of Latvia. In general, the autumn stream flow has no trend at all. The frequency analysis of the annual runoff data series (Fig. 7) shows a cyclic process with 44-46-year semi-periods (Koļcova et al. 2006). J. Barkāns has studied the process and verified on a global scale the characteristics that were observed for the Daugava River (Barkāns and Zicmane 2004). The runoff data of 45 rivers worldwide were collected for the research. J. Barkāns noticed the same process with 44-year semi-periods for all rivers with identical harmonic and differing phases. From 1880, three cycles in annual flow are observed: 1881-1925, 1926-1971, and 1972-around 2016. Each cycle has two wet and two dry periods. In 2005 the second wet period started. Keeping in mind the previous cycles, it will last till 2020. Fig. 7. Annual runoff deviation fluctuations of the Daugava River nearby Daugavpils. The Character of Climate Change 93 The hydropower energy production at the Daugava HPP, which generates more than 90 % of the total hydropower in Latvia, strongly depends on the annual runoff of the Daugava River. The cyclic variations of water flow (annual and maximum) seem to form the basis of the planning of energy generation. Fig. 8. The Daugava River’s monthly water runoff distribution (observed and simulated by SWECLIM) (Koļcova et al. 2006). According to the simulation of the river runoff to the Baltic Sea made with the aid of the Swedish Regional Climate Modelling Program (SWECLIM) for the period 2071-2100, the decrease of annual streamflow and the increase of winter flow are expected (Graham 2004). Figure 8 shows the changes in water runoff of the Daugava River on the basis of four different climate change scenarios (Koļcova and Rogozova 2006). Climate change leads to significant changes in water resources and further affects hydropower production (Fig. 9). Fig. 9. The relationship between the energy generated by the Daugava HPP (black spots) and the annual discharge (bars) of the Daugava River (Koļcova et al. 2006). 94 Climate Change in Latvia Positive runoff trends in summer season have produced a positive effect on the work of the HPP and energy production as a whole. The negative trends of maximum discharges in spring are conducive to a safe work of the HPP during extreme floods (Kriauciuniene et al. 2006). CONCLUSIONS The changes in temperature and precipitation detected in this study influence the annual, seasonal, and extreme water flow. First of all, this effect is reflected by the strong decrease of the spring flood brought about by the decrease of water content of snow in snow cover period. The water runoff reflects the tendency in precipitation series, therefore the winter runoff has a significant positive trend for all the periods analysed. The frequency analysis of annual runoff data series shows a cyclic process with a 90-year period. Taking into account that the Daugava HPPs supplies about 50 % of the demand for electric energy in Latvia, the periodical variations of the Daugava’s water flow must be the basis for planning of energy generation in Latvia. In addition, the minimum flow increase in Kurzeme Upland affects the discharges of rivers in the region where about 30 % of the small HPP are situated. ACKNOWLEDGEMENTS This research was supported by the Climate and Energy Project, funded by the Nordic Energy Research and the Nordic Energy Sector. REFERENCES Alexandersson H., Moberg A. (1997) Homogenization of Swedish Temperature Data. Part A: Homogeneity Test for Linear Trends Int. J. Climatol., 17, 25-34 Barkāns J., Zicmane I. (2004) Electricity Production by Hydropower Plants: Possibilities of Forecasting. Latv. J. Phys. Techn. Sci., 1, 32-38 Glazacheva L. (1990) Hydrological Regionalization in the Set of Natural Regionalizations on Latvian Territory. Latv. PSR ZA Vēstis, 11, 520-532 Graham L.P. (2004) Climate Change Effects on River Flow to the Baltic Sea. Ambio 33, 235- 241 Hisdal H. (2006) ‘Statistical Analysis’ Group: Summary of Selected Studies. CE Abstracts, 151 Koļcova T., Lizuma L., Smith M. (2006) Climate Change and Hydropower Production Perspectives in Latvia. CE Abstract volume, 191 The Character of Climate Change 95 Koļcova T., Rogozova S. (2006) Trends in Flooding of Latvian Rivers. NHP Report 49, 658 Kriauciuniene J., Kovalenkoviene M., Gailiusis B. (2006) Climate Change and Possible Impact on Hydrological Regime in Lithuania. CE Abstract Volume, 163 Reihan A., Koltsova T., Kriauciuniene J., Meilutyte-Barauskiene D., Jarvet A. (2006) Changes in River Runoff in the Baltic States in the 20th century. NHP Report 49, 601 96 Climate Change in Latvia Mathematical Modelling of the Hydrological Processes in the Aiviekste River Basin Uldis Bethers, Juris Seņņikovs University of Latvia, Faculty of Physics and Mathematics, Laboratory of Mathematical Modelling of Environmental and Technological processes, Zeļļu Str. 8, LV 1002, Rīga, Latvia A spatially and temporally distributed mathematical model of the hydrological processes was developed. It was calibrated and applied to the Aiviekste River catchment. The model includes a dynamic flow routing in the stream network. The calibrated model was applied to the modelling of the runoff from the catchment for typical conditions and climate change scenarios. Key words: river discharge, modelling, the Aiviekste River basin INTRODUCTION The Aiviekste River basin (Fig. 1) encompasses a rather large territory (9293 km2) in eastern Latvia, and it is part of the Daugava River basin (Daugavas Projekts 2003). The Aiviekste River basin consists of the rivers Aiviekste, Pededze, and Rēzekne and two major lakes – Lubāns and Rāzna. To the east the Aiviekste River basin borders the Veļikaja River basin and to the north and nortwest the Gauja River basin. The Aiviekste River is located in the central part of the Daugava River basin, in Vidzeme Upland and East Latvia Lowland. The Aiviekste River basin encompasses Aizkraukle, Jēkabpils, Madona, Gulbene, Alūksne, Balvi, Rēzekne, Ludza, Preiļi, and Krāslava districts. The basin of the Aiviekste River includes the basins of the Pededze River (1523 km2) in the north and of the Rēzekne River (2022 km2) in the east. An insignificant part of the Pededze River basin is located in Estonia. The Rēzekne River basin is located in the northern part of the Latgale Upland. The Aiviekste River basin also contains two large lakes: Lubāns and Rāzna. Lubāns Lake is in the Lubāna Lowland, approximately in the middle of the basin system. The Rēzekne River springs from the Rāzna Lake. The Character of Climate Change 97 Fig. 1. The basin of the Aiviekste River. Digital terrain map with the main rivers, lakes and towns. Geology. The Aiviekste River basin is located mostly on the southern slope of Vidzeme Upland, in Vestiena hilly area, and in East Latvia Lowland. To the west it verges the Middle Latvia Lowland. The Pededze River flows in the northwestern part of East Latvia Lowland. It comes from the southeastern part of Alūksne Upland. Its geological setting is characterized by Upper-Devonian sandstone, siltstone, clay, dolomite, and dolomitic marl, which are covered by up to 15 m thick Quaternary deposits, e.g., glacigenic or glaciolacustrine sediments. The network of the incised valleys shapes the geomorphology of the modern river valleys and the placement of other negative landforms. Hydrology. The length of the Aiviekste River is 118 km. The width of the Aiviekste River is 70-90 m, and it reaches average 3.5-4 m depth at the downstream stretch (upstream of the Rīga – Daugavpils railway, i.e., above the influence of the Pļaviņas HPP). In the middle stretches, nearby the Kuja River junction, the Aiviekste River is 40-50 m wide and 2 m deep. Aiviekste is 30-35 m wide and 1.5-2 m deep in the upstream stretches nearby the 98 Climate Change in Latvia Iča River junction. Annual mean flowrate of Aiviekste River at the river’s source from the Lubāns Lake is 12.8 m3/s. Downstream, it reaches 30.3 m3/s at Lubāna town, 57.8 m3/s at the Aiviekste HPP, and 60.2 m3/s at the entry into the Daugava River. The flow of the Aiviekste River is rather slow from the source in the Lubāns Lake down to the entry of Meirānu Canal. Current velocity during the summer periods may be as low as 0,15-0,2 m/s. The right bank tributaries, the Iča and Balupe rivers, are also slow, being lowland type rivers, especially in their lower reaches. The velocity of the river current increases further downstream near Ļaudona town, reaching 0.25-0.3 m/s. It slows down due to the Aiviekste HPP dam, but after the dam up to the railway bridge the current velocity increases to 0.4-0.5 m/s. The source of the Aiviekste River is at 92.5 m a.s.l, its entry in the Daugava River nearby the town Pļaviņas is 72.0 m a.s.l. The right bank tributaries of the Aiviekste River (the Pededze, Liede, Kuja, Arona, Veseta) flow mostly from Vidzeme Upland; they may be characterised as fast flowing rivers. The artificial Meirānu Canal is considered the biggest left bank tributary. It collects waters from most of the southern part of Lubāns Lowland (the Kažauka, Lisiņa, Malmute rivers), and it drains part of the Lubāns Lake. A big part of the Lubāns Lowland rivers were diverted directly into the Aiviekste River by the help of this canal to regulate the level of the Lubāns Lake. As a result, these rivers were straightened and deepened. Such rivers include the Teicija, Malmute, Lisiņa, Kažauka, Komorsta, and others. Urbanisation. There are aproximately 1500 inhabited localities in the Aiviekste River basin. A total of approx. 172 000 inhabitants live in the catchment area and ~30 % of these inhabit the biggest towns: Rēzekne, Madona, Balvi, Alūksne, and Gulbene (CSP 2000). Territories at the risk of flooding. The Lubāns Lake territory is marked to be one of such territories in the National Flooding Risk Map. It encircles the upper reaches of the Aiviekste River, which, together with the Lubāns Lowland, occupy a natural depression. The flooding was rather frequent before the construction of the dams across the Lubāns Lake (annual snow melt floods). Hydropower plants. Only one HPP operates on the Aiviekste River, approximately 15 km upstream its entry into the Daugava River. A total of 8 small HPP can be found on the Aiviekste River’s minor tributaries. There are two HPP operating on the Pededze River. Two small HPP are being renovated on the Rēzekne River and three on Malta River in the Rēzekne River basin. Land use. Let us consider the three main land types – forests and bushes, rural areas, and other (including urban) areas. In this case their division in subbasins is as follows: the main Aiviekste River basin (48%, 47%, 5%), the Pededze River basin (58%, 39%, 3%), the Rēzekne River basin (30%, 60%, 10%). The Character of Climate Change 99 The size of the catchment and the variability of the hydraulic and land- use characteristics is a challenge for the hydrological modelling of the Aiviekste River. MATHEMATICAL MODEL The physically based spatially and temporally distributed dynamic modelling approach was used, following the ideas of Ven Te Chow (1988) (USACE 1994). The modelling system consists of combining of 5 different models: • model of the surface water content; • model of the snow accumulation and melting; • model of the groundwater flow; • dynamic flow routing model; • lake model. The finite element method is used for all models. The scheme of the finite element mesh is shown in Figure 1. The formulations of the particular mathematical models may refer to Figure 1. Surface water and snow balance is calculated for the finite elements; the elevation, lake, and groundwater levels are defined in the finite element nodes, and the rivers (streams) always coincide with the edges of the finite elements. The surface water model solves the equation (1) for the intercepted + ponded water content w. (1) is defined for the surface water finite elements (Fig. 2). The right-hand side of (1) contains the following elements of the surface water balance: 1. Precipitation in a form of rain P; 2. Evapotranspiration E. We used the following parameterisation of the evapotranspiration: (2) Here e and esat are water vapour partial pressure and its saturation value, whilst Ke is a coefficient dependent on the land use, vegetation, and the presence of snow. 3. Velocity of the infiltration Vinfiltr depends on the surface water content, groundwater level, and the presence of snow. 3.1. The infiltration Vinfiltr=0 if the groundwater level reaches the surface (h≥z) or the surface water content is below the ≤ intercepted water content (w wintercepted). 100 Climate Change in Latvia 3.2. The infiltration velocity is at the minimum value (Vinfiltr=Vimin, Vimin is the calibration parameter) during the presence of snow (Vsnow>0). 3.3. The infiltration velocity is at the maximum value (Vinfiltr=Vimax, Vimax is the calibration parameter, both Vimin and Vimax depend on the land use) if snow is absent. Fig. 2. Finite element mesh and the elements of the modelling system. 4. The groundwater saturation excess flow Vsurface is given by equation (3) Here K is the hydraulic conductivity of the soil, ∆lsurface is the empiric parameter. The outflow to the surface water occurs only if the groundwater level exceeds the surface elevation (h>z). The part 0<αs<1 of the groundwater outflow is assumed to freeze (i.e., add-up to snow volume) during winter time (i.e., negative air temperature). 5. The overland flow to/from the neighbouring (ith) elements is given by equation (4) Here n is the land-use-dependent Manning coefficient for the overland flow, S0 is the surface slope (of the finite element), Aelem is the element area, and lout is the length of the considered edge of the finite element (Fig. 2). The overland flow direction is determined by the surface slope. The Character of Climate Change 101 6. Vsnow is the snow-melting rate. The model for the snow-melting and accumulation is adapted from (Zīverts and Jauja 1999). It solves the equation of the balance of the equivalent water content S accumulated in a form of snow: (5) The right hand terms in (5) correspond to: 1. Ps is precipitation in a form of snow; 2. Part of the groundwater excess flow (if any, see (3) above); 3. Vsnow is a degree-day dependent snow melting rate which accounts for the incoming solar radiation I (Zīverts and Jauja 1999) (6) Here T2>0 is a reference temperature, CmeltB and Amelt are the model parameters. The 2D groundwater flow model considers the piezometric head (or groundwater level) h of the upper aquifer (USACE 1998). The groundwater level is defined at the nodes of FE mesh (Fig. 2), and the balance equation stands as (7) Here zg is the level of the aquitard, and Ss is the soil storability. The right hand terms of (7) correspond to, respectively: 1. Groundwater filtration according to Darcy law; 2. Infiltration from the surface water, see p.3 above; 3. Outflow to rivers Vriver at the river nodes (Fig. 2); (8) ∆ Here hriver is the water level in the river node, and lriver is the calibration parameter. 4. Groundwater outflow to surface water Vsurface is given in (3) above. The groundwater model is closed by external no-flow conditions on outer boundaries, whilst h=hlake is set for the lake nodes (Fig. 2). The dynamic flow routing model solves the temporal development of the water level hriver and discharge Q on the staggered finite difference grid along 102 Climate Change in Latvia the rivers. Water level is defined at the model nodes, whilst discharge – at the river segments (Chadwick and Morfett 1994), see also Figure 2. The full St. Venant equations are solved: (9) Here A is the river cross-section area (we used the assumption of parabolic river profiles providing the river bed level from digitised topographic maps), x is the distance along the river, and q is the source parameter: (10) accounting for the surface runoff and groundwater discharge into rivers. The bed friction parameter in (9) is given by equation: (11) Here, P is wetted perimeter. The lake model solves the dynamic balance of the water level hlake in the lakes: (12) Here, Alake is the lake surface area, and the right hand side terms in (12) correspond to river inflow and outflow, surface runoff, groundwater inflow/ outflow, precipitation, and evaporation. INPUT DATA Digital terrain map. The surface topography is created as a digital terrain map in a form of finite element mesh. The base elevation was taken from the US Geological Survey GTOPO-30 (USGS 1996). GTOPO30 is a global digital elevation model (DEM) with a horizontal grid spacing of 30 arc seconds, i.e., approximately 1 kilometre. GTOPO30 was derived from several raster and vector sources of topographic information. GTOPO30, completed in late 1996, was developed over a three-year period through collaborative effort led by the staff of the U.S. Geological Survey’s EROS The Character of Climate Change 103 Data Center (EDC). We checked the elevations for the consistency using topographic maps in a scale 1:50000. River network. Hierarchical and re-current river network was digitised from the topographic maps (The Headquarters of Soviet Army, 1981-82). Main rivers and lakes are shown in Figure 1, while the zoom of the detailed hydraulic system is shown in Figure 3. The hydraulic network consists of linear and connected finite elements. The combination of the detailed stream network (Fig. 3) with the topography permits the construction of the boundaries of the subbasins at any hierarchical level. Fig. 3. Zoom of detailed hydraulic network in the Aiviekste River basin. Land use. The land use in the Aiviekste River basin was assimilated in the electronic map from the CORINE database (EC 2000). The mesh of he land use may be interpolated on the mesh of the DTM, and vice versa, thus obtaining the land use distribution in every subbasin (down to the finite element level of DTM). The map of the land use is shown in Figure 4. For modelling purposes, we distinguished 6 land use types: agricultural, forests, bushes/grasslands, swamps/wetlands, artificial and natural waterbodies. 104 Climate Change in Latvia The forests are a rather representative land-use type in the basin of the Aiviekste itself and on the Pededze River (northern part of the basin), while the agricultural lands prevail in the basins of the Malta and Rēzekne rivers (southeastern part of the catchment area). River discharge observations (Zīverts 2000) were used for the selection of the calibration period. We used 7 hydrometric stations (Fig. 5) for the data analysis. Aiviekste at the Aiviekste HPP is the most downstream station which represents almost whole of the 8660 km2 catchment area. The subbasins with different types of land use, river slope, and meteorological conditions are represented by hydrometric stations: Kuja at Aizkuja – a rather fast right tributary representing Vidzeme Highland, a catchment area of 268 km2; Pededze at Litene represents the Pededze River and the northern part of the catchment, rich in forests, catchment area 978 km2; Iča at Kuderi represents the eastern part of Lubāns depression, catchment area 674 km2; Malta at Viļāni and Rēzekne represents the agricultural southeastern part of the basin, catchment areas 782 km2 and 545 km2, respectively; Aiviekste at Lubāna is located in the central part of the basin (catchment area 7200 km2), and it characterizes the discharge downstream at the confluence of the Pededze River and upstream at the entry to the Meirānu Canal. Fig. 4. The map of land use (codes form CORINE) in the Aiviekste River basin. Zoom into area near Lubāns Lake. River network included. The Character of Climate Change 105 Fig. 5. The location of the hydrometric observation stations. The discharge measurements at the hydrometric stations reported by (Zīverts 2000) were used to (1) select the characteristic time period, (2) calibrate the model. We decided to use three different years (typical, wet, and dry) for the model calibration, assuming the start of the year on 1-Jul (the beginning of summer) and the end of the year on 30-Jun (when the spring snow-melting flood has ended). The discharge data were analysed to discover 3 consecutive years which fulfill these requirements. Such three years were selected in the period from 1-Jul-1976 to 30-Jun-1979. Meteorological data. The data of precipitation and air temperature of these three years were used, provided by Latvian Environment, Geology and Meteorology Agency from three meteorological observation stations (Zīlāni, Gulbene and Rēzekne). We assumed that the interpolation of the meteorological data between these 3 locations represents the variability of the meteorological conditions over the catchment area of the Aiviekste River. The summary of the meteorological data for 3 selected years is presented in Table 1, and in Table 2, there are the hydrometric observations for all the three years and stations. 106 Climate Change in Latvia Table 1. The summary of the meteorological observations Station Gulbene Rēzekne Zīlāni T average, oC4.284.424.87 P average, mm 632 608 647 T, average, dry year 76/77 4.49 4.76 5.09 P, average, dr y year 76/77 477 451 518 T, average, typical year 77/78 4.43 4.50 5.06 P, average, typical year 77/78 635 626 660 T, average, wet year 78/79 3.92 4.00 4.45 P, average, wet year 78/79 785 747 762 The selected years have the following variability (in Zīlāni station): (1) Jul-1976 to Jun-1977 is a dry year (518 mm, deviation -22 %); (2) Jul- 1977 to Jun-1978 is an average year (660 mm); (3) Jul-1978 to Jun-1979 is a wet year (762 mm, deviation +15 %). One may conclude that there is a reasonable variation of both the annual precipitation (up to 15%, southeastern Rēzekne being the driest) and the annual mean air temperature (up to 0.6 ºC, northern Gulbene being the coldest) in the catchment area of the Aiviekste River. The characteristics of the observation data. Let us consider the variability of the runoff at the most downstream station, the Aiviekste HPP (basin area 8660 km2). The runoff modulus at this station is 127 mm, 213 mm, and 334 mm. Discharge at this station is 25 %, 32 %, and 44 % of the precipitation for the three respective years. The runoff modulus is lower during the dry years due to a relatively higher evaporation (high T, low E) and higher infiltraton (low soil moisture, low groundwater level). The runoff modulus is higher for the wet years due to a lower evaporation (low T, high E) and lower infitration (high soil moisture, high groundwater level, distinct spring snow-melting flood). Table 2. Summary of the annual mean discharge (m3/s) at the selected stations Catchment Dry year Typical year Wet year Station area, km2 1976/77 1977/78 1978/79 Aiviekste HPP 8660.00 34.85 58.60 91.75 Aiviekste Lubāna 7200.00 24.18 42.57 62.81 Kuja 268.00 1.47 2.82 3.65 Pededze 978.00 4.67 7.24 12.95 Iča 674.00 2.37 5.76 N/A Rēzekne 505.00 1.49 3.37 5.06 Malta 782.00 2.71 6.14 8.29 The Character of Climate Change 107 Discharge measurements at the Aiviekste stations and the precipitation of the dry year are shown in Figure 6. Dry years are characteristic with a single snow-melting flood. No travel time of the flood signal may be distinguished. There is no immediate response of the precipitation signal in the discharge time-series. The measurements and precipitation for the wet year are shown in Figure 7. Wet years are characterised by multiple rain events during summer and prolonged rainfalls during autumn. The latter results in a distinct autumn high water. Still neither the travel time of the flood signal may be distinguished nor there is an immediate response of the precipitation signal in the discharge time-series. Fig. 6. Daily discharge of the Aiviekste at HPP, the Aiviekste at Lubāna, and precipitation for the dry year. There are certain differences in the discharge time-series for different subbasins. The Kuja and Rēzekne rivers are compared in Figure 8. The Kuja subbasin has a higher slope, whilst the Rēzekne is a typical lowland river. Therefore the response of the discharge on the precipitation events is much more distinct for the Kuja. The comparison of the snow-melting events in both rivers also indicates some climatic variability in the Aiviekste basin. 108 Climate Change in Latvia Fig. 7. Daily discharge of the Aiviekste at HPP, the Aiviekste at Lubāna, and precipitation for the wet year. Fig. 8. Daily discharge of the Kuja at Aizkuja and the Rēzekne at Griškāni for the typical year. The measurements in upstream (Pededze) and downstream (Aiviekste HPP) stations are compared in Figure 9. There is no delay in precipitation or the snow-melt signal. However, the water retention is much higher in the downstream station due to the effect of the water storage in the Lubāns Lake. The Character of Climate Change 109 Fig. 9. Daily discharge of the Aiviekste at HPP and the Pededze at Litene for a typical year. The climate change scenarios suggested by Pirkanmaa Regional Environmental Centre were used and extrapolated over a 50-year period. The seasonal change of the average temperature and precipitation for three (Low, Central and High) scenarios for 50 years is summarized in Table 3. We added the values of these data changes to the representative meteorological observations (1-Jul-1976 to 30-Jun-1978) to calculate the future run- off scenarios. All scenarios are related to the increase of temperature and precipitation, especially during winters. Table 3. The changes of mean meteorological parameters over a 50-year period. Low, Central and High scenarios Spring Summer Autumn Winter Parameter/scenario Annual (III-V) (VI-IX) (X-XII) (I-II) P / Low, %+0.625+1.25+1.25+2.1+1.25 T / Low, ºC +0.5 +0.4 +0.5 +0.65 +0.5 P / Central, % +2.5 +5.0 +5.0 +10.0 +5 T / Central, ºC +2.0 +1.5 +2.0 +3.0 +2.0 P / High, % +3.75 +7.5 +7.5 +12.5 +7.5 T / High, ºC +3.0 +2.25 +3.0 +3.75 +3.0 110 Climate Change in Latvia MODEL CALIBRATION Model calibration was performed for the 3 year period (from 1-Jul-1976 to 30-Jun-1979). The 2D interpolation of the daily meteorological parameters (precipitation, air temperature) from 3 observation stations was used, until we assumed the climatologic mean for air humidity. Fig. 10. Calculated vs. observed daily discharge of the Pededze at Litene. The goal of the calibration was to find a set of model parameters (groundwater model parameters and the land-use dependent surface water run-off, evaporation, melting, and infiltration parameters) which minimizes the deviation from the observed daily discharges in 7 hydrometric stations (Fig. 5). Each three-year model calibration run was preceded by a 90 year- long “warm-up” period of stabilising the grounwater level at a quasi-periodic state to avoid initialisation effects. The calibration results as a set of the model parameters are given in Table 4, and the calculated vs. observed discharges at the selected hydrometric stations are shown in Figures 10 and 11. The summary of the observed and calculated runoff for the subbasins and the whole Aiviekste catchment area is given in Table 5. The Character of Climate Change 111 The general features of the differences in the land-use-dependent calibration parameters may be found in Table 4. Thus, forests have lower evaporation and snow-melt rate values, higher friction of overland flow and volume of intercepted water storage. Table 4. Calibration results – model parameters Parameter Land use Value Hydraulic conductivity of soil K All 10 m/day Soil storativity Ss All 0.25 Manning coefficient Forest 4.5 Manning coefficient Agricultural, swamps 4 Manning coefficient Bushes 4.25 Manning coefficient Artificial 1 wintercepted Forest 40 mm wintercepted Agricultural 30 mm wintercepted Bushes 35 mm wintercepted Artificial, swamps 10 mm Vimin All 0.5 mm/day Vimax All but swamps 2.5 mm/day Vimax Swamps 1.5 mm/day Evapotranspiration coeff. Ke Agricultural, Swamps, Artificial 0.55 mm/(day*mbar) Evapotranspiration coeff. Ke Forests 0.50 mm/(day*mbar) Evapotranspiration coeff. Ke Bushes 0.52 mm/(day*mbar) Ke-snow Agricultural, Swamps, Artificial 0.14 mm/(day*mbar) Ke-snow Forests 0.10 mm/(day*mbar) Ke-snow Bushes 0.12 mm/(day*mbar) Snow melt coefficient CmeltB Agricultural, Swamps 4 mm/(day*K) Snow melt coefficient CmeltB Forests 2.5 mm/(day*K) Snow melt coefficient CmeltB Bushes 3 mm/(day*K) Snow melt coefficient CmeltB Artificial 5 mm/(day*K) Reference temperature T2 Agricultural, Swamps 0.2 ºC Reference temperature T2 Forests 1.0 ºC Reference temperature T2 Bushes 0.5 ºC Reference temperature T2 Artificial 0.0 ºC ∆lsurface All 100 m The agreement between the observed and modelled discharge time- series is reasonable for the small rivers (Fig. 10). Generally, the high-water situations are underestimated in the model. However, both the observation quality and the accuracy of the catchment area may be questioned for the small hydrometric stations located rather far upstream from the rivers’ entry to larger water bodies. Another source of disagreement in model predictions is the possible different land-use ratio in the calibration period 112 Climate Change in Latvia and the disagreement of the actual precipitation in the small catchment to the meteorological observations in a distant observation station. Fig. 11. Calculated vs. observed daily discharge of the Aiviekste at HPP. The agreement between the calculated and the observed discharge values of the Aiviekste River may be assessed as excellent (Fig. 11). The model represents both the high- and low-water situations well, qualitatively and quantitatively. The transition (i.e., the decrease of discharge after rain or snow-melt events) is also rather accurate. The model slightly overestimates the high-water events at the Aiviekste HPP. This shows in the reduction of these peaks when the HPP’s reservoir level is regulated. The underestimation of average runoff of small rivers (except the Malta) and a very good agreement for both the Aiviekste’s stations can be seen in the figures in Table 5. We have concluded that the model calibration was performed successfully and we have used the obtained model parameters for the calculations of the response of the Aiviekste River discharge to the climate change patterns. The Character of Climate Change 113 Table 5. The comparison of mean observed and the calculated discharge of different hydrometric stations. No. River Discharge, observed m3/s Discharge, modelled m3/s 1 Kuja at Aizkuja 2.64 1.82 2Pededze at Litene 8.29 5.86 3Iča at Kuderi4.07 2.76 4 Malta at Viļāni 5.71 5.77 5 Rēzekne at Griškāni 3.30 2.48 6Aiviekste at Lubāna 43.19 43.25 7Aiviekste at HPP 61.73 62.60 RESULTS AND DISCUSSION The distribution of groundwater level is very similar to the surface elevation (Fig. 12), and it has no distinct seasonal variations. The distribution of snow cover indicates the climatic variability in the Aiviekste River basin. The correlation between air temperature (ºC) and the thickness of snow cover (m) at the beginning of winter is shown in Figure 13. Fig. 12. Surface elevation (left) and groundwater level (right). The time-series of the discharge at the selected hydrometric stations are presented in Figures 10 and 11. An interesting feature was found at the source of the Aiviekste River from the Lubāna Lake (Fig. 14). During the beginning of snow-melt flood (as well as after extreme rain events), the flow direction of the uppermost stretch of the Aiviekste River changes, and water (coming mainly from the northern part of the catchment area, i.e., the Pededze River) enters the Lubāns Lake. This leads to a rather fast raise of 114 Climate Change in Latvia the water level in the lake. This situation is not possible nowadays, after the engineering solutions regulating the outflow of water from the Lubāns Lake are implemented. Fig. 13. Air temperature (left) and the thickness of snow cover (right) at the beginning of winter (11-Dec-1976). Fig. 14. The time series of water level of the Lubāns Lake and the discharge of the Aiviekste River from it. The Character of Climate Change 115 The calculation results for the scenarios of climate change are compared at the station Aiviekste at HPP as the discharge time graphs of Reference vs. Low scenario (Fig. 15), Low vs. Central scenario (Fig. 16), and Central vs. High scenario (Fig. 17). The discharges for different seasons and all scenarios are summarised in Table 6. The implementation of the Low climatic change scenario causes a lower accumulation of snow and slightly reduces the maximum of the snow-melt floods (Fig. 15). Winter run-off increases by 17.5 %, while spring run-off decreases by 3.5 %. The yearly run-off decreases by 2.5 % (Table 6) due to slightly higher air temperatures and less concentrated spring floods. Fig. 15. The time series of the discharge of the Aiviekste at HPP for the Reference and Low scenarios. The main qualitative change in the hydrology of the future scenarios occurs in the transition between the Low and Central scenarios (Fig. 16). The increase of winter air temperatures is a reason for multiple winter rainfalls (instead of snowfalls) causing the increase of winter runoff by more than two times, and the decrease of spring runoff by one third. Summer runoff is generally lower due to higher air temperature which increases evapotranspiration. Annual runoff in comparison with the reference situation decreases by 11 %. 116 Climate Change in Latvia Fig. 16. The time series of the discharge of the Aiviekste at HPP for the Low and Central scenarios. Fig. 17. The time series of the discharge of the Aiviekste at HPP for the Central and High scenarios. The Character of Climate Change 117 Table 6. Mean seasonal discharge (m3/s) of the Aiviekste River at HPP for 3 different climatic scenarios Scenario I-II III-V VI-IX X-XII AVERAGE Reference 24.5 123.3 33.4 65.1 62.6 Measured 17.9 120.5 35.7 65.5 61.7 Low 28.8 119.0 30.8 63.7 61.0 Central 69.0 87.5 23.1 59.2 55.8 High 77.9 74.1 20.3 55.5 52.0 The comparison between the Central and High scenarios (Fig. 17) mainly indicates quantitative changes (equalization of winter and spring runoffs, the decrease of runoff during summers and autumns). The annual runoff in comparison with the reference situation decreases by 17 %. CONCLUSIONS The model calibration revealed a high application potential of spatially and temporary distributed mathematical models for river basin hydrology. It was discovered that the set of physically based model parameters predicts the response of river discharge to seasonal forcing by air temperature and precipitation data series. However, the modelling indicated that the scale problem should be resolved to achieve the same calibration accuracy for the whole catchment and particular subbasins. One of the most interesting model applications was illustrated with the calculation of the response of the Aiviekste runoff to the climate change scenarios. We found that even simple and quantitative changes in air temperature and precipitation cause a non-linear response expressed as qualitative change in annual runoff cycle. Some of the results, e.g., the distinct decrease of annual runoff despite the increase of precipitation are far from obvious or are even unexpected. We are sure that a more accurate prediction of the future changes in the runoff pattern may be achieved if forcing data of regional climate models are used because they account for various scenarios of global climate change. 118 Climate Change in Latvia ACKNOWLEDGEMENTS The study was performed with the financial support of Maj and Tor Nessling Foundation. We acknowledge the input of professors Tom Frisk and Māris Kļaviņš throughout the research period. The software used for the modelling was presented by “The Center of Processes Analysis and Research, SIA”. The Character of Climate Change 119 REFERENCES Chadwick A., Morfett J. (1994) Hydraulics in Civil and Environmental Engineering. E & FN Spon. 557 pp. CSP (2000) The Statistical Yearbook. Regions of Latvia in Numbers. Central Agency of Statistics, Riga. Electronic edition (CD). Daugava Project (2003) The Management Plan for the Daugava Basin. 1. Surface Water Body Aiviekste. [In Latvian] EC (2000) National CORINE Land Cover 2000 in Latvia – Project I&CLC2000, Brussels. Headquarters of Soviet Army (1981-82). Military Topographic Maps. Scale 1:50000 (various sheets). U.S. Army Corps of Engineers (1993) River Hydraulics. Engineer Manual. 176 pp. U.S. Army Corps of Engineers (1994) Flood-runoff Analysis. Engineer Manual. 214 pp. U.S. Army Corps of Engineers (1998) Groundwater Hydrology. Engineer Manual. 122 pp. USGS (1996) http://edcdaac.usgs.gov/gtopo30/gtopo30.asp. Ven Te Chow, Maidment D. R., Mays L.W. (1988) Applied Hydrology. McGraw-Hill Book Company. 572 pp. Zīverts A., Jauja I. (1999) Mathematical Model of Hydrological Processes METQ98 and Its Applications. Nordic Hydrol., 30 (2), 109-128. Zīverts A. (2000) Hydrological Calculations in Latvia, Report. Jelgava, University of Agriculture, 2000. Electronic edition (CD).