Keywords

1 Introduction

This chapter addresses climate variability in the Baltic Sea basin over the past millennium (1000 years). Climate change at the millennial time scale was not discussed in the previous assessment of climate change in the Baltic Sea region (BACC Author Team 2008).

Understanding of millennial climate variability is mainly based on proxy data (see Sect. 3.2). Historical notes in the form of chronicles containing information about extreme weather events, and weather-related disasters are important sources of data about climatic conditions in the past millennium. Together with historical documents, dendroclimatology has provided much of the information on the climatic conditions of the past millennium. Tree-ring width and wood density are the main sources of dendroclimatological data . Other proxy data are of less importance for the past millennium and are more relevant for reconstructing the climatic conditions of the Holocene as a whole. Temperature and precipitation data for the Baltic Sea region, as reconstructed from different proxy data sets were also compared against the results of a regional climate model simulation at the Swedish Meteorological and Hydrological Institute (Schimanke et al. 2012). The chapter also describes the main climatic drivers, of which solar radiation factors, atmospheric circulation patterns and volcanic eruptions appear the most important. The most unusual weather conditions in the Baltic Sea area in each century have been described in several publications (e.g. Rojecki 1965; Borisenkov and Pasetsky 1988, 2002; Pfister 1992, 1999; Glaser 2008).

2 Data Sources and Methodology

Research on millennial climate variability is mainly based on proxy data. This is because long-term instrumental measurements of meteorological elements (usually limited to atmospheric pressure , air temperature and wind) were recorded at only a few stations and over a relatively short period.

The first non-regular measurements of atmospheric pressure by means of a barometer constructed by Torricelli in 1643 were made in the period 1649–1658 in Clermont-Ferrand, Florence, Paris and Stockholm (von Rudloff 1967). The first regular meteorological measurements were launched in Greenwich (1774). The oldest continuous series of atmospheric pressure data for the Baltic Sea basin and surrounding areas date back to the nineteenth century: Prague (1800), Oslo (1816), Warsaw (1826), St Petersburg (1837), Copenhagen (1842), Stockholm (1844), Berlin (1848), Uppsala (1855), Haparanda (1859) and Helsinki (1882).

The first non-regular attempts at measuring air temperature began between 1654 and 1670 in Florence and Pisa in Italy (von Rudloff 1967); however, the longest homogenised and uninterrupted series of temperature data began in 1659 in central England (Manley 1974). In the Netherlands, direct temperature measurements started in 1705 in De Bilt (Cowie 2007). For the Baltic Sea basin, temperature measurements were initiated around the mid-1700s, with a denser coverage of the southern region in the late eighteenth century. The oldest continuous observational records of temperature are from Sweden: Uppsala (1739—non-regular measurements started in 1722), Lund (1741) and Stockholm (1756); Russia: St. Petersburg (1743); and Denmark: Copenhagen (1798) (von Rudloff 1967). A few temperature series began in the eighteenth century in central Europe: Berlin (1719), Jena (1770), Prague (1775), Warsaw (1779), Vilnius (1781), Wroclaw (1791) and Kraków (1792).

The oldest measurements of precipitation in Europe started in 1715 in Hoofddorp-Zwanenburg in the Netherlands and in 1725 in Padova in Italy. In the Baltic Sea area, the oldest precipitation station is Uppsala (1739) (von Rudloff 1967). In the first half of the nineteenth century, precipitation was also measured in a few other places: Warsaw (1803), Prague (1804), Copenhagen (1805), Jena (1827), Dresden (1828), Helsinki (1844) and Berlin (1847).

Historical notes in the form of chronicles containing information about extreme weather events and weather-related disasters are important sources of data about climatic conditions of the past millennium. In many cases, such notations are very precise as they locate events in space and time, sometimes even with an accuracy of a day. A systematic daily weather diary carried from 1502 to 1540 in Kraków and surroundings by Marcin Biem, a professor of Kraków Academy (Bokwa et al. 2001; Limanówka 2001) is unique on the European scale. A similar weather diary for the north-eastern part of Poland was kept by Jan Antoni Chrapowicki (Nowosad et al. 2007; Przybylak and Marciniak 2010) for the period 1656–1685. Borisenkov and Pasetsky (1988, 2002) compiled information about climate extremes and natural disasters from Russian chronicles. In Switzerland, Pfister (1992) established the European Centre of Historical Climate with the European Climate Historical database—EURO-CLIMHIST (Pfister 1992; Brázdil et al. 2010). Later, Glaser (2008) published the complete climate history for central Europe and Germany, covering the past 1200 years. Brázdil et al. (2005) discussed the state of European historical-climatological research with special attention to data sources, methods and significant results.

Together with historical documents, dendroclimatology has provided a large part of information on climatic conditions of the past millennium. Tree-ring width and wood density (Ljungqvist 2010) are the main sources of dendroclimatological data . Several recent multi-proxy reconstructions were made for the northern hemisphere (Jones et al. 1998, 2001a, b, 2009; Mann et al. 1998, 1999, 2008; Bertrand et al. 2002; Mann and Jones 2003; Jones and Mann 2004; Moberg et al. 2005; Ljungqvist 2010; Ljungqvist et al. 2012). On a regional scale, the most important reconstructions based on tree-ring or multi-proxy data are for Fennoscandia (Briffa et al. 1992; Gouirand et al. 2008; Lindholm et al. 2009, 2011; Esper et al. 2012), Finland (Helama et al. 2002, 2005, 2009b; Ogurtsov et al. 2008; Luoto and Helama 2010), central and northern Sweden (Gunnarson and Linderholm 2002; Linderholm and Gunnarson 2005; Moberg et al. 2006; Grudd 2008), eastern Norway (Kalela-Brundin 1999), Germany (Glaser 2008; Glaser and Riemann 2009), the north-western Baltic Sea (Klimanov et al. 1985), Russia (Klimenko and Solomina 2010), and Poland (Przybylak et al. 2005, 2010; Przybylak 2007, 2011; Szychowska-Krąpiec 2010; Koprowski et al. 2012). Although many of the data sets concern the Alps (Büntgen et al. 2005, 2006), they are well correlated with the central European mountains (Bednarz 1984, 1996; Bednarz et al. 1999; Niedźwiedź 2004; Büntgen et al. 2007, 2012).

Other proxy data were also applied to the millennium temperature reconstructions, for example, peat-bog deposits (Lamentowicz et al. 2008, 2009), lake fossils and sediments from Tsuolbmajavri Lake in northernmost Finland (Korhola et al. 2000), laminated lake sediments in Gościąż Lake in central Poland (Starkel et al. 1996; Ralska-Jasiewiczowa et al. 1998) and borehole temperatures (Majorowicz et al. 2004; Majorowicz 2010). The reliability of the reconstructions is discussed by Holmström (2011). For northern Sweden, a summer-temperature reconstruction for the past 2000 years was achieved using pollen-stratigraphical data (Bjune et al. 2009). Among other biological proxies, a few reconstructions were based on diatoms (Korhola et al. 2000; Weckström et al. 2006) and chironomids (Korhola et al. 2002). These data are of less importance for reconstructing the climatic conditions of the past millennium and better used for reconstructing the climate of the Holocene as a whole.

Proxy records are clearly useful for helping understand the spatial and temporal variability of climate change, especially over periods shorter than the millennial time frame and which fall outside the instrumental period. Other data sets, such as long-term variability in Baltic Sea ice cover (Koslowski and Glaser 1999), runoff or oxygen conditions (Hansson and Omstedt 2008; Hansson et al. 2011; Hansson and Gustafsson 2011) can be used in combination with climate models to increase understanding.

3 General Features of the Millennial Climate

According to the scientific literature, four climatic periods have occurred over the past millennium (Lamb 1977, 1982; Grove 1988; Folland et al. 1990; Brázdil 1996; Crowley 2000; Crowley and Lowry 2000; Bradley et al. 2003; Brázdil et al. 2005; Xoplaki et al. 2005; NCR 2006; Esper and Frank 2009; Jones et al. 2009; Brázdil and Dobrovolný 2010; Büntgen and Tegel 2011; Büntgen et al. 2011b; Ogurtsov et al. 2011; Przybylak 2011; Ljungqvist et al. 2012):

  • Medieval Warm Period (MWP 900–1350) or Medieval Climate Anomaly (MCA)

  • Transitional Period (TP 1350–1550)

  • Little Ice Age (LIA 1550–1850)

  • Contemporary Warm Period (CW after 1850).

The dates for the four climatic periods are approximate and may differ slightly from one geographical region to another (Ljungqvist et al. 2012). Some shorter intervals are mainly related to the changes in solar activity or large volcanic eruptions. The Contemporary Warm Period is addressed in detail in Chaps. 4–9.

A comparison of late spring (April–June) precipitation and summer (June–August) air temperature across Europe between the southern Baltic Sea and the Alps (Fig. 3.1) was reported by Büntgen et al. (2011b) for the last 2500 years, together with possible impacts on civilisation.

Fig. 3.1
figure 1

Reconstructed April–June (AMJ) precipitation totals (top) and summer (June–August) temperature anomalies (bottom) for central Europe with respect to 1901–2000. Error bars are ±1 RMSE (Root-Mean Square Error) for the calibration periods. Black lines show independent precipitation and temperature reconstructions from Germany (Büntgen et al. 2010) and Switzerland (Büntgen et al. 2006). Bold lines are 60-year low-pass filters. Periods of demographic expansion, economic prosperity and societal stability are noted, as are periods of political turmoil, cultural change and population instability. Büntgen et al. (2011b)

Full size image

Gouirand et al. (2008) reported variability in summer temperature for the whole of Fennoscandia (Fig. 3.2) over the past 1500 years. All four periods of the past millennium (MWP, TP, LIA and CW) are visible. The curves for central Europe (Fig. 3.1) and Fennoscandia (Fig. 3.2) show similarities but the deepest cooling in Fennoscandia is observed at the beginning of eighteenth century, while in central Europe summer temperature, negative anomalies are higher in the first decades of the nineteenth century. Also, a cool episode during the MWP in the first half of the twelfth century in Fennoscandia is sharper than in central Europe.

Fig. 3.2
figure 2

Regional-average summer (June–August) temperatures AD 442–1970 for Fennoscandia created by merging seven reconstructions based on seven networks (Gouirand et al. 2008). The time series is extended to year 2000 with instrumental data. Data are shown as smoothed (Gaussian filtered) temperatures, highlighting variability on timescales longer than 10 years (thin black), 30 years (thick dark grey) and 100 years (thick light grey). The dashed horizontal line is the average for the entire period. The uncertainty in reconstructed temperatures (based on the calibration period statistics) is illustrated by ±2 standard errors with grey shading (for the 10-year smoothing only)

Full size image

Mean January–April air temperature for the period 1170–1994 (Fig. 3.3) and 10-year anomalies of winter and summer air temperature from 1401 to 1800 (Fig. 3.4) have been reconstructed for Poland. The climatic history for Poland over the past millennium was reconstructed by Przybylak (2011) and more comprehensively by Przybylak et al. (2010). Tree-ring reconstruction of January–April air temperature indicates three relatively cool periods: 1475–1500, 1600–1660 and 1725–1830 (Fig. 3.3). The peak temperature of an exceptionally warm episode occurring 1661–1675 is slightly lower than indicated by other reconstructions for central Europe (compare Fig. 3.9). Cold anomalies for winter temperature in Poland (Fig. 3.4) suggest an increased annual temperature range during the LIA. A cool period at the final phase of the LIA in the first half of the nineteenth century is also found in a reconstruction based on the full depth of ground temperature in boreholes (Majorowicz et al. 2004).

Fig. 3.3
figure 3

Reconstructed mean January–April air temperature in Poland for the period 1170–1994 using a standardised chronology of Scots pine (Pinus sylvestris L.) tree-ring widths (modified after Przybylak et al. 2005; Przybylak 2011). Rek-11 and Rek-50 represent 11- and 50-year running means; reconstruction using areally averaged air temperature from Warsaw, Bydgoszcz and Gdańsk for calibration. Obs: measured mean January–April area-averaged air temperature from Warsaw, Bydgoszcz and Gdańsk (Przybylak et al. 2005; Przybylak 2007, 2011)

Full size image
Fig. 3.4
figure 4

Reconstructed mean 10-year air temperature in Poland from 1401 to 1840 based on historical sources winter (DJF) and summer (JJA). 1 and 2 anomalies with respect to 1901–1960 and 1789–1850 means, respectively. Przybylak (2011)

Full size image

Climate change during the past millennium over the Baltic Sea region was simulated by the Swedish Meteorological and Hydrological Institute using a regional climate model (Schimanke et al. 2012). The authors used the Rossby Centre Regional model (RCA3) with boundary conditions from the general circulation model ECHO-G. RCA3 includes radiative forcing, changes in orbital parameters, changes in greenhouse gas concentration and atmospheric circulation. The model covers the whole Baltic Sea area and its surroundings and has a horizontal resolution of about 50 × 50 km. Results were presented as 50-year running means; air temperature is largely underestimated (Schimanke et al. 2012). Biases in annual precipitation are about 20 % in the Baltic Sea region and during winter and autumn exceed 50 %.

The 50-year running means for the annual temperature anomalies over Sweden varied from −0.5 °C in the seventeenth century during the LIA to about +0.4 °C in twelfth to thirteenth centuries representing the MWP (Fig. 3.5). The TP is clearly expressed by anomalies varying from about 0 °C through the fourteenth century to about −0.2 °C at 1550.

Fig. 3.5
figure 5

The 2-m temperature anomaly with regard to the preindustrial mean (950–1900) for the winter (DJF), summer (JJA) and annual mean averaged over the Baltic Sea region. The coloured sections highlight the periods that are defined as MWP (red) and LIA (blue). The darker colours reflect the 50-year periods which are considered for the analysis of the Baltic Sea. After Schimanke et al. (2012)

Full size image

Winter temperature anomalies are larger ranging from about −0.9 °C in 1700 to +0.8 °C at the middle of the twelfth century (Fig. 3.5). Variability in summer temperature anomalies is less than for the annual anomalies, ranging from −0.3 °C in the early eighteenth century to +0.3 °C during the MWP near the mid-twelfth century.

The variability in annual and winter temperature anomalies for the Baltic Sea region presented in Fig. 3.5 (Schimanke et al. 2012) differs from other reconstructions. Winter temperature anomalies are lowest in the latter half of the seventeenth century (−0.9 °C). This shows some agreement with the reconstructed winter temperatures for the greater Baltic Sea area (Eriksson et al. 2007) discussed in Sect. 3.5. A large discrepancy is visible in the first half of the eighteenth century when the reconstructed winter temperature anomaly is greatest in the LIA, while the model simulation indicates a warm episode near 1800, which Eriksson et al. (2007) suggested was the coolest in the LIA (based on 15-year running means for temperature).

Reconstructing past precipitation using proxy measurements is more difficult than for temperature and is only possible for parts of the year. There is no possibility of reconstructing annual precipitation data from proxy data. Büntgen et al. (2011b) reconstructed total precipitation from April to June for central Europe between the southern Baltic Sea and the Alps. Warm and dry summers are typical during the MWP (900–1350). The onset of the TP in the latter half of the thirteenth century is indicated by cooler and wetter summers. Total precipitation between April and June over the past millennium in central Europe is presented in Figs. 3.1 (Büntgen et al. 2011b) and 3.6 (Büntgen et al. 2011c). Generally, the MWP was drier than the first half of the TP and the LIA in spring and early summer seasons.

Fig. 3.6
figure 6

a Central European and regional fir TRW (Tree Ring Width) extremes, and b their centennial changes over the past millennium (network extremes were double weighted), compared to c annual-resolve and 40-year low-passed Central European April–June precipitation variability (Büntgen et al. 2011c)

Full size image

Annual precipitation simulated in the Baltic Sea catchment area by a regional climate model (Schimanke et al. 2012) and shown in Fig. 3.7 indicates that the annual totals are larger during the MWP than the TP and the LIA. The annual precipitation data cannot be compared with the reconstructed data for April–June. However, the very dry period in the latter half of the fifteenth century is marked in both precipitation series. The variability of river runoff to the Baltic Sea follows changes in precipitation. However, the simulation of river runoff for the period 1500–1995 by Hansson et al. (2011) presented in Chap. 5, Fig. 5.3 indicates no significant long-term change.

Fig. 3.7
figure 7

Annual precipitation in the Baltic Sea catchment area (green) and statistical estimated runoff (blue) as 50-year running means. After Schimanke et al. (2012)

Full size image

4 The Medieval Warm Period (MWP 900–1350)

At the turn of the tenth and eleventh centuries, relatively stable climate conditions with few extremes prevailed in the Baltic Sea basin and the surrounding parts of Europe. Hot and dry summers were noted in 993 and 994, respectively. For example, in Russian chronicles, eight droughts (1000, 1025–1028, 1035, 1037, 1092), four wet summers with rain and floods (1002, 1031, 1034, 1043) and seven severe winters (1034/35, 1043/44, 1047/48, 1056/57, 1058/59, 1066/67, 1076/77) were noted in the eleventh century (Borisenkov and Pasetsky 1988, 2002).

In Europe, the warmest conditions occurred between 1200 and 1250, and the MWP ended about 1350 (Borisenkov and Pasetsky 1988, 2002). Chernavskaya (1996) reconstructed the June temperature changes in European Russia over the past two millennia based on pollen analysis in peat bogs. Data from Polistovo (56.8°N, 38.1°E) suggest an earlier occurrence of the MWP (ninth to tenth centuries) in European Russia than in central Russia (tenth to eleventh centuries). Two periods of strong cooling occurred in the middle of the twelfth century and at the end of the fourteenth century. On the East European Plain, summer temperatures during the MWP were found to be above the long-term average between 900 and 1200 (Klimenko and Solomina 2010). For Fennoscandia (Fig. 3.2), summer temperatures were elevated in the tenth and eleventh centuries (Gouirand et al. 2008). Even warmer summer periods were noted in the late twelfth century, succeeding the extreme cool summers in the mid-twelfth century.

Recent investigations of Fennoscandia by Ljungqvist (2010) showed that the MWP occurred between 800 and 1300. At that time, warm-season (May–September) temperatures exceeded the contemporary warming of the end of twentieth century by about +0.5 °C. The start of the warming was noted between the ninth and tenth centuries, and the peak temperature appeared at the beginning of the second half of the twelfth century. In a winter temperature simulation over the Baltic Sea region (Schimanke et al. 2012) during that time anomalies reached their highest value of +0.8 °C for the MWP (Fig. 3.5). Lower temperatures occurred at the end of the MWP, about 1350. A diatom-based July temperature reconstruction for the past 800 years in northern Scandinavia (Weckström et al. 2006; Holmström 2011) indicates that temperature was about 0.2 °C higher in the latter half of the fifteenth century than in 1970–2000. An exceptionally warm period occurred in 1220–1250 and in the latter half of the fifteenth century (1470–1500) in the TP. The frequency of extreme temperature events in Russia increased in the twelfth century (Borisenkov and Pasetsky 1988, 2002). Winter-simulated temperature indicates the second warm episode of the MWP in the latter half of the thirteenth century (Schimanke et al. 2012) for the Baltic Sea region (Fig. 3.5). At the beginning of the fourteenth century, climatic conditions cooled rapidly. In 1315, a serious famine in northern Europe resulted from a series of very cold winters (1302/03 and 1305/06) and cool and wet summers (1314–1317) across the whole of Europe (Cowie 2007).

There is less information available on precipitation in the MWP (Gouirand et al. 2008; Büntgen et al. 2011b, c). Nevertheless, a regional dendroclimatic precipitation reconstruction from southern Finland showed a uniquely prolonged rainfall deficit coinciding with the MWP (Helama et al. 2009a). The drought was particularly severe between 1000 and 1200. The simulation of annual precipitation for the Baltic Sea catchment using a regional climate model (Schimanke et al. 2012) shows that the driest period was the mid-eleventh century and that two wet periods occurred in the first half of the twelfth century and the latter half of the thirteenth century (Fig. 3.7). As a generalisation, relatively dry periods occurred in Europe in the years: 1272–1291, and 1300–1309, while the wettest conditions were noted in 1312–1322 (Borisenkov and Pasetsky 1988, 2002). In April–June in central Europe (Fig. 3.6), wet conditions were observed in 1052, 1187 and 1316, and the driest in 1167 (Büntgen et al. 2011c). Distinct wet periods with frequent floods were recognised in this region at 1020–1030 and 1075–1100 (Starkel 2001).

5 The Transitional Period (TP 1350–1550)

The increase in the intra-seasonal variability of climate at the end of the MWP in the period 1270–1350 is considered to be the beginning of the LIA; however, Brázdil et al. (2005) suggested that the following 200-year period should be treated as transitional between the MWP and the LIA. This period was characterised by a great variability of climatic conditions. At that time, temperature decreased by about 1.2 °C, but cooling occurred until the latter half of the sixteenth century (Borisenkov and Pasetsky 1988, 2002). During this period, the decreasing tendency of mean annual and seasonal temperatures simulated for Baltic Sea region (Fig. 3.5) is clearly visible.

Over the majority of Europe and Russia, very unfavourable conditions for agriculture in the period 1400–1480 were linked to large fluctuations in temperature and precipitation. For example, the summers of 1428, 1434, 1436 and 1438 were hot and dry and the summer of 1435 was cool and dry, whereas the summers of 1432, 1437 and 1439 were extremely wet with flooding (Borisenkov and Pasetsky 1988, 2002). In central Europe (Fig. 3.6), the first part of the TP up to about 1430 was very wet according to the reconstructed April–June precipitation curve (Büntgen et al. 2011c) and followed by very dry conditions with an extremely dry and hot spring and summer in the year 1504 (Glaser 2008).

In Poland (Fig. 3.4), severe winters were detected in four decades: 1401–1410, 1451–1460, 1511–1520 and 1541–1550 (Przybylak 2011). The warmest were winters during the years 1521–1530. Similar thermal conditions in winter based on historical sources were found in Latvia (Jevrejeva 2001) and Estonia (Tarand and Nordli 2001). The summers were relatively warm in two decades: 1471–1480 and 1491–1500.

The first halves of the fifteenth and sixteenth centuries appear relatively warm periods, but there was large variability (Helama et al. 2009b). Climatic variability may be reflected in the proxy instability. According to a diatom-based reconstruction, the warmest 30-year non-overlapping period in northern Scandinavia occurred in 1470–1500 (Weckström et al. 2006); however, according to a dendroclimatic reconstruction in this region, the summers of the late fifteenth century were anomalously cold (Helama et al. 2009b). There were very warm conditions in Fennoscandia in summer at the end of the TP (Gouirand et al. 2008; Fig. 3.2).

6 The Little Ice Age (LIA 1550–1850)

In the latter half of the sixteenth century, the temperature dropped. This tendency was particularly clear in the period 1569–1579. Another sequence of extremely wet and cool summers was identified at the end of the sixteenth century (Borisenkov and Pasetsky 1988, 2002). The longest consecutive cold period occurred from the late sixteenth century to the mid-eighteenth century (Gouirand et al. 2008), which is in very good agreement with a regional climate model simulation for temperature in the Baltic Sea region (Schimanke et al. 2012; Fig. 3.5). But a short sequence of very warm summers was observed in the latter part of eighteenth century, just before the prolonged cooling at the end of the eighteenth and during the nineteenth century (Fig. 3.2).

Eriksson et al. (2007) analysed the complex description of winter climate conditions during the LIA for the greater Baltic Sea region (50–70°N, 0–30°E). Their study is based on well-documented time series of ice cover , sea-level pressure and winter surface-air temperatures. Using winter temperature in connection with atmospheric circulation and ice conditions, they found four cold and three warm periods during the LIA (Fig. 3.8). In the latter half of the sixteenth century, a cool phase (1562–1576) passed to a relatively mild period (1577–1591). In the seventeenth century, two phases of cold winters, 1597–1629 and 1663–1706, were divided by the mild period of 1630–1662. The first half of the eighteenth century (1707–1750) which includes the warmest winter (1723/24) is considered to be the major warm period. At this time, the maximum ice extent in the Baltic Sea was similar to present-day conditions, albeit there are many uncertainties related to the observations at that time (Seinä and Palosuo 1996; Hansson and Omstedt 2008). The period 1730–1745 has also been described as particularly variable interannually, swinging from extremely mild to extremely cold winters (Jones and Briffa 2006). The occurrence of cold winters is related to the Late Maunder Minimum (1675–1715) which has also been discussed by Luterbacher et al. (2001) and in Chap. 4, Sect. 4.2.3. The longest cool period in the final phase of the LIA in the Baltic Sea region (1750–1877) coincided with the Dalton Minimum (1790–1840) in solar activity (Eriksson et al. 2007; see also Chap. 4). During the entire LIA, no downward trend in ice break-up date of the river Daugava was detected (see also Chap. 5).

Fig. 3.8
figure 8

Winter climate conditions in the greater Baltic Sea region since 1500. The grey colour shows seasonal winter data from two gridded data sets: (top to bottom) Baltic Sea mean winter air temperature. Black line in all panels is a 15-year running mean. Blue and red fields cover periods classified as cold and mild, respectively (Eriksson et al. 2007)

Full size image

A new reconstruction of the Baltic Sea region climate for the past 500 years was prepared by Brázdil et al. (2010) on the basis of instrumental data and documentary evidence under the MILLENNIUM project. January–April mean temperatures were reconstructed for Stockholm (1502–2008) and central Europe (1500–2007). In central Europe, the coldest conditions were observed in the sixteenth century, while in central Sweden the end of the seventeenth century was cooler (Fig. 3.9). In central Europe, the warmest period was the first half of the eighteenth century, while in Stockholm such conditions occurred at the end of the eighteenth century. In the recently modelled and reconstructed northern Scandinavian summer temperatures since 1500 (Fig. 3.10), the longest cool period covered the first half of the seventeenth century (Büntgen et al. 2011a) and the beginning of the eighteenth century as well as in the early nineteenth century. The warmest conditions occurred just after the middle of the eighteenth century.

Fig. 3.9
figure 9

Comparison of reconstructed JFMA (January–April) temperatures for Stockholm (1502–2008) and CEuT (Central European Temperature) (1500–2007) (anomalies from the 1961–1990 mean). Original series of CEuT (a) and Stockholm (b) are smoothed with a 30-year Gaussian filter (c) and compared using 31-year running correlations between unfiltered data (d). The horizontal solid line in d denotes the critical value of correlation coefficients for α = 0.05 for one-tailed t test (Brázdil et al. 2010)

Full size image
Fig. 3.10
figure 10

Reconstructed and modelled northern Scandinavian summer temperature variations over the past five centuries (Büntgen et al. 2011a). a Actual (black) and reconstructed (red) JJA (June–August) temperature anomalies (°C) with error estimates (grey) and the ten warmest and coldest decades superimposed (colour boxes). b Comparison of the actual and reconstructed temperatures with five existing (green) reconstructions. c Comparison of the actual and reconstructed temperatures with CCSM3 surface-air temperature (pink) and sea surface temperature (blue) model simulations . Mean and variance of the data are scaled against JJA temperature (1860–2006), expressed as anomalies (relative to 1961–1990) and 20-year low-pass filtered

Full size image

There are several climate reconstructions based on the documentary and instrumental data for the period since 1500. Koslowski and Glaser (1999) constructed an ice winter severity index and temperature for the western Baltic Sea region, covering the period of 1500–1759. Temperature series were reconstructed for Stockholm (Moberg et al. 2002; Leijonhufvud et al. 2010), Tallin (Tarand and Nordli 2001), Poland as a whole (Dobrovolný et al. 2010; Luterbacher et al. 2010; Przybylak et al. 2010) and the Tatra Mountains (Bednarz 1984; Niedźwiedź 2004; Büntgen et al. 2007), and several parts of central Europe (Pfister 1992; Brázdil 1996; Luterbacher et al. 2004; Dobrovolný et al. 2010).

Glaser et al. (2010) reconstructed the variability of floods in Europe since 1500. The precipitation conditions are presented in Figs. 3.1 and 3.6. The wettest conditions during the LIA were observed in 1675 (Glaser 2008) and the first half of the eighteenth century, and the driest were the years 1504 and 1762 (Büntgen et al. 2011c). In southern Finland, conditions were markedly wetter during the LIA than the MWP, as inferred from tree rings (Helama et al. 2009a). An exception was a period of transient drought which, as the same dendroclimatic reconstruction indicates, occurred during the first decades of the nineteenth century. A regional model simulation of annual precipitation (Schimanke et al. 2012) shows that the driest period of the LIA was the latter half of the seventeenth century (Fig. 3.7). For central-east Sweden the low May–June precipitation was found by Jönsson and Nilsson (2009) to occur in the periods 1560–1590 and 1694–1751. The clustering of heavy rainfall years is typical for the LIA, causing the floods in central Europe in 1590–1610, 1705–1715 and 1800–1815 (Starkel 2001). They coincide with the relatively cool periods of Spörer, Maunder and Dalton solar activity minima.

In the seventeenth century, the winter of 1657/58 was exceptionally cool. Modelled Baltic Sea ice extent indicates that it was one of the coldest winters for the Baltic Sea since 1500 (Hansson and Omstedt 2008). The sequence of cool summers and winters occurred over the period 1690–1699. In Sweden, climatic conditions favourable for agriculture occurred in 1604–1620 (Borisenkov and Pasetsky 1988, 2002). According to a diatom-based July temperature reconstruction in Finnish Lapland (Weckström et al. 2006), particularly cold 30-year periods were detected between 1640–1670 and 1750–1780. In several reconstructions, the low-frequency variability of LIA temperature is underestimated (e.g. von Storch et al. 2004; Zorita et al. 2007; Christiansen and Ljungqvist 2011). A new method of temperature reconstruction (Christiansen and Ljungqvist 2011) showed that in the seventeenth century, the lowest temperature anomaly in the cooling northern hemisphere reached a 50-year smoothing value 1.1 °C below the contemporary level (1880–1960). The period 1630–1700 was the coolest consecutive period of the entire past millennium.

At the beginning of the eighteenth century, the winter of 1708/09 was perhaps the coldest winter of the past 500 years (Luterbacher et al. 2004). In Poland, the coolest winters were recorded in the decades 1701–1710 and 1741–1750, while the coldest summers were found in the decade 1731–1740 (Przybylak 2011). Very cold winters were also observed at the end of eighteenth century in 1783/84, 1788/89, 1794/95 and 1798/99 (Borisenkov and Pasetsky 1988, 2002). This cold period continued to the end of the first half of the nineteenth century. In 1815, the Tambora volcanic eruption in Indonesia discharged large amounts of ash into the upper atmosphere, resulting in the famous ‘year without a summer’ in 1816. This particular year, the summer in western Europe was unusually cold. However, that was not the case in the Baltic Sea basin (Briffa and Jones 1992). In eighteenth century in Poland, precipitation showed large variability. The period 1731–1750 was wetter than normal (Przybylak 2011). But between 1751 and 1766, during a generally dry period with 13 dry years, the wettest was 1755 and the driest 1762.

Dendroclimatological studies have identified several cool and rainy summers in the Carpathian Mountains (southern Poland) in the latter part of the LIA: 1650–1660, 1670–1685, 1690–1719 and 1735–1745 (Bednarz 1984, 1996). The final phase of the LIA in the first half of the nineteenth century was also marked by a sequence of exceptionally cold years between 1812 and 1824. In that period, average winter temperature in Russia and a large part of Europe was lower than normal by as much as 10–12 °C (Borisenkov and Pasetsky 1988, 2002). In central Europe, the winter of 1829/30 was extremely cold, as well as the winters of 1822/23 and 1837/38. In Norway, based on farmer diaries (Nordli et al. 2003), the severely cold spring/summer (April–August) decadal temperatures were found around 1740, and the 1800s and 1830s.

The cooling during the LIA had an important influence on human society . At the turn of the thirteenth and fourteenth centuries, the number of farms in northern Norway decreased due to a drop in temperature (Cowie 2007). In Finland, abandoning of farms in the sixteenth and seventeenth centuries coincided with a long-term summer temperature cooling, implying that the desertion may have resulted from a change in climatic conditions that significantly limited agriculture as a means of subsistence (Holopainen and Helama 2009). In the same region, harvest records show that during the poorest years the amount of grain harvested was less than had been sown (Holopainen and Helama 2009). Very unfavourable weather conditions in 1697 resulted in a failed crop . This caused widespread famine, followed by over a third of the Finnish and a fifth of the Estonian population dying in just a few years (Cowie 2007). Also, the cold winter of 1657/58 permitted the Swedish King Charles X to walk across the frozen Belts and Sound with his army and occupy all of Denmark, except for Copenhagen, with a large loss of land for the Danish (Neumann 1978; Hansson and Omstedt 2008).

There are discrepancies in the dates of both the end of the LIA and the beginning of the CW. The majority of scientists agree on 1850 as being crucial (e.g. Grove 1988). However, some climatologists claim that the LIA did not finish until the last decades of the nineteenth century (e.g. Lamb 1977). In southern Poland (Tatra Mountains) according to summer temperature data, the final episode of the LIA lasted 103 years: 1793–1895 (Niedźwiedź 2004).

7 Conclusion

According to the scientific literature, there are four climatic periods of the past millennium: the Medieval Warm Period (MWP 900–1350), the Transitional Period (TP 1350–1550), the Little Ice Age (LIA 1550–1850), and the Contemporary Warm Period (CW after 1850). The MWP started at the beginning of the tenth century with relatively stable climate conditions, and few extremes prevailed in the Baltic Sea basin and the surrounding parts of Europe. The warmest conditions occurred between 1200 and 1250. Two periods of strong cooling were detected in the middle of the eleventh and at the beginning of the fourteenth century. During the MWP in Fennoscandia, warm-season (May–September) temperatures exceeded the contemporary warming of the end of twentieth century by about +0.5 °C. The following 200-year period should be treated as a transitional period between the MWP and LIA. This period was characterised by a great variability in climatic conditions; at that time, temperature decreased by about 1.2 °C. In the latter half of the sixteenth century, the temperature dropped, initiating the LIA. Winter temperatures in combination with atmospheric circulation and ice conditions indicate four cold and three warm periods during the LIA. In the recently modelled and reconstructed northern Scandinavian summer temperatures since 1500, the longest cool period prevailed during the first half of the seventeenth century and at the beginning of the eighteenth century, as well as during the first years of the nineteenth century. During these main historical climatic periods, climatic conditions were not uniform. Shorter warm/cool and wet/dry fluctuations were observed depending on regional factors.