Publications of the Department of Geology D9Helsinki 2006

Reconstructions of past climates from documentary and natural sources

in Finland since the 18th century

Jari Holopainen

Academic dissertation

To be presented with the permission of the Faculty of Science of the University of Helsinki for public criticism in Lecture Room E204 of Physicum, Kumpula,

on November 24th, 2006, at 12 o´clock

Ph.D. thesis No. 192 of the Department of Geology, University of Helsinki

Supervised by:Prof. Matti EronenDepartment of GeologyUniversity of Helsinki, Finland

Reviewed by:Prof. Jukka KäyhköDepartment of GeographyUniversity of Turku, Finland

Dr. Heikki TuomenvirtaFinnish Meteorological InstituteHelsinki, Finland

Discussed with:Dr. Ari VenäläinenFinnish Meteorological InstituteHelsinki, Finland

Cover: Satellite image globe courtesy of Planetary Visions

ISSN 1795-3499ISBN 952-10-2610-3 (paperback)ISBN 952-10-2611-1 (PDF)http://ethesis.helsinki.fi /

Helsinki 2006Yliopistopaino

Jari Holopainen: Reconstructions of past climates from documentary and natural sources in Finland since the 18th century, University of Helsinki, 2006, 33 pp., University of Helsinki, Publications of the Department of Geology D 9, ISSN 1795-3499, ISBN 952-10-2610-3, ISBN 952-10-2611-1 (pdf-version).

Abstract

In this paper both documentary and natural proxy data have been used to improve the accuracy of palaeoclimatic knowledge in Finland since the 18th century. Early meteorological observations from Turku (1748-1800) were analyzed fi rst as a potential source of climate variability. The reliability of the calculated mean temperatures was evaluated by comparing them with those of contemporary temperature records from Stockholm, St. Petersburg and Uppsala. The resulting monthly, seasonal and yearly mean temperatures from 1748 to 1800 were compared with the present day mean values (1961-1990): the comparison suggests that the winters of the period 1749-1800 were 0.8 ºC colder than today, while the summers were 0.4 ºC warmer. Over the same period, springs were 0.9 ºC and autumns 0.1 ºC colder than today. Despite their uncertainties when compared with modern meteorological data, early temperature measurements offer direct and daily information about the weather for all months of the year, in contrast with other proxies. Secondly, early meteorological observations from Tornio (1737-1749) and Ylitornio (1792-1838) were used to study the temporal behaviour of the climate-tree growth relationship during the past three centuries in northern Finland. Analyses showed that the correlations between ring widths and mid-summer (July) temperatures did not vary signifi cantly as a function of time. Early (June) and late summer (August) mean tempera-tures were secondary to mid-summer temperatures in controlling the radial growth. According the dataset used, there was no clear signature of temporally reduced sensitivity of Scots pine ring widths to mid-summer temperatures over the periods of early and modern meteorological observations. Thirdly, plant phenological data with tree-rings from south-west Finland since 1750 were examined as a palaeoclimate indicator. The information from the fragmentary, partly overlapping, partly non-systematically biased plant phenological records of 14 different phenomena were combined into one continuous time series of phenological indices. The indices were found to be reliable indicators of the February to June temperature variations. In contrast, there was no correlation between the phenological indices and the precipitation data. Moreover, the correlations between the studied tree-rings and spring temperatures varied as a function of time and hence, their use in palaeoclimate reconstruction is questionable. The use of present tree-ring datasets for palaeoclimate purposes may become possible after the application of more sophisti-cated calibration methods. Climate variability since the 18th century is perhaps best seen in the fourth paper study of the multiproxy spring temperature reconstruction of south-west Finland. With the help of transfer functions, an attempt has been made to utilize both documentary and natural proxies. The reconstruction was verifi ed with statistics showing a high degree of validity between the reconstructed and observed temperatures. According to the proxies and modern meteorological observations from Turku, springs have become warmer and have featured a warming trend since around the 1850s. Over the period of 1750 to around 1850, springs featured larger multidecadal low-frequency variability, as well as a smaller range of annual temperature variations. The coldest springtimes occurred around the 1840s and 1850s and the fi rst decade of the 19th century. Particularly warm periods occurred in the 1760s, 1790s, 1820s, 1930s, 1970s and from 1987 onwards, although in this period cold springs occurred, such as the springs of 1994 and 1996. On the basis of the available material, long-term temperature changes have been related to changes in the atmospheric circulation, such as the North Atlantic Oscillation (February-June).

5

Contents

List of original publications 6

1 Introduction 7

1.1 The need for perspective 7

1.2 Basics of palaeoclimatological reconstruction 8

1.3 Context of early meteorological observations 8

1.4 Objectives of the studies 10

2 Materials and methods 11

2.1 Analyses of the early meteorological observations 11

2.2 Tree-ring data analyses 11

2.3 Method for combining the plant phenological data 12

2.4 Verifi cation of the reconstructions 12

2.5 Multiproxy reconstruction 13

3 Results 13

3.1 The early temperature records from Turku 13

3.2 Climate-tree growth relationship from the 18th to the 20th century in northern Finland 14

3.3 Early plant phenological data and tree-rings from south-west Finland

as a palaeoclimate indicator 15

3.4 A multiproxy reconstruction of spring temperatures in south-west Finland 15

4 Discussion 17

4.1 Early efforts to organize weather observations 17

4.2 The role of scientifi c academies and volunteers 18

4.3 The utilization of early plant phenological data 19

4.4 Time dependency of climate sensitivity in tree-rings 20

4.5 The value of seasonal multiproxy reconstruction 20

4.6 Interpretation of the spring temperature variability since 1750 21

4.7 Prospects for future studies 22

5 Conclusions 24

6 Acknowledgements 25

7 References 26

Papers I-IV 35

6

List of original publications

This is an article dissertation based on the material and results presented originally in the following papers referred to in the text with Roman numerals, as designated below:

I Vesajoki, H. and Holopainen, J. 1998. The early temperature records of Turku (Åbo), south-west Finland 1749-1800. In: Frenzel B. (ed) Documentary climatic evidence for 1750-1850 and the fourteenth century. Paläoklimaforschung – Palaeoclimate Research 23: 151-161.

II Helama, S., Holopainen, J., Timonen, M., Ogurtsov, M. G., Lindholm, M., Meriläinen, J., Eronen, M. 2004. Comparison of living-tree and subfossil ringwidths with summer temperatures from 18th, 19th and 20th centuries in Northern Finland. Dendrochronologia 21/3: 147-154.

III Holopainen, J., Helama, S., Timonen, M. 2006. Plant phenological data and tree-rings as palaeoclimate indicators in south-west Finland since AD 1750. International Journal of Biometeorology 51: 61-72.

IV Holopainen, J., Helama, S., Kajander, J. M., Korhonen, J., Launiainen, J., Nevanlinna, H., Reissell, A., Salonen, V-P., 2006. A multiproxy reconstruction of spring temperatures in south-west Finland since AD 1750. Submitted to Climatic Change.

In paper I, the author was responsible for coding and analyzing the data from Turku. The author also participated in planning and writing the article. In paper II, the author was responsible for the analysis of the early meteorological data from Tornio and Ylitornio and the homogenization of the temperature data. In papers III and IV, the author performed the statistical analysis and wrote the papers. SH performed part of the statistical analysis. Others participated by providing data and ideas.

7

1 Introduction

1.1 The need for perspective

Our increased understanding that the world is undergoing a period of signifi cant warming has stimulated interest in the study of past climates. According to instrumental observations of climate, global mean temperatures have been rising since the beginning of the 20th century (Brohan et al. 2006), and a signifi cant part of this warming is likely due to anthropogenic emissions of greenhouse gases, although natural factors may also have contributed (IPCC 2001). To know how unusual the 20th century has been, we need to be able to place it in the context of longer-term climate variability. Within this context, records from high latitude regions, where seasonal contrasts are relatively large, are of special interest, since, at present, the most notable climate change involving several feedback mechanisms is occurring in the arctic and boreal regions (Betts 2000; Harding et al. 2002; ACIA 2004).

Modern instrumental weather observations are the most accurate measures of climatic fl uctuations, but their length is usually no more than slightly over 150 years. For instance, in Finland the era of modern instrumental meteorological observations began around the middle of the nineteenth century, when routine meteorological observations were started in several localities (Heino 1994; Tuomenvirta 2004). The longest meterological observation series comes from Helsinki, where Prof. G. G. Hällström organized regular observations of meteorological phenomena on October 4, 1828 (Johansson 1906; Heino 1994; Holopainen 2004). Proxy data as indicators of climate variability must therefore be used to reconstruct earlier changes. This research fi eld is called palaeoclimatology (Bradley 1999). The types of proxy data used for the reconstruction of past climates have been described in great detail by Ingram et al. (1978), Lamb (1982), Eronen (1990), Pfi ster (1992), Bradley (1999), and recently Brázdil et al. (2005). The classifi cations are mainly based on dividing evidence into natural and anthropogenic data according to their origin. The natural data may fall into geophysical (e.g. isotopes, sediments, moraines) and biological (e.g. pollen, tree-rings) evidence. Anthropogenic data may be grouped into documentary (e.g. descriptive reports, instrumental observations, phenonological observations, harvest times for various crops) and material (e.g. paintings, archaeological remains) evidence according to their form and to the place where they originated.

A number of previous (multi)proxy studies have focused on global and hemispheric temperature reconstructions over the past few centuries or over millennia, based on both empirical proxy data (Bradley and Jones 1993; Jones et al. 1998; Mann et al. 1998, 1999; Crowley and Lowery 2000; Briffa et al. 1998b, 2001, 2002, 2004; Esper et al. 2002; Cook et al. 2004; Pollack and Smerdon 2004; Moberg et al. 2005; Oerlemans 2005; D’Arrigo et al. 2006; Osborn and Briffa 2006) and model simulations including forcing data (Crowley 2000; Bauer et al. 2003; González-Rouco et al. 2003; von Storch et al. 2004; Zorita et al. 2004; Goosse et al. 2005a, b; Weber 2005; Hegerl et al. 2003, 2006). Large-scale reconstructions, however, mask the details of the spatial patterns and seasonal

8

timing of climate changes (Jones and Kelly 1983; Helama 2004; Wanner 2005), and therefore local reconstructions of different seasons as separate records are needed to understand the interactions in the mixture of natural and anthropogenic forcing agents along with internal climatic variability.

1.2 Basics of palaeoclimatological reconstruction

Palaeoclimatological work examines the relationships between the variations of proxy records and certain climate parameters such as temperature, precipitation and air pressure. Although climate is not the only factor bearing an impact on proxy records, its signal can be extracted and separated from the extraneous noise (Bradley and Jones 1992). This can be done statistically by comparing modern meteorological records with overlapping records of proxy data. Palaeoclimatological examination is therefore an empirical test for detecting which climate parameters bear signifi cant impact for proxy record variability (cf. Helama 2004). Ideally, the strength and the sign of the parameters can be de-termined.

In practice, the proxy-based reconstruction of the past climate is based on a methodology whereby the relationship between the proxy data and the climatic parameter is processed through response and transfer functions (Briffa et al. 1988; Bradley and Jones 1992; Brázdil et al. 2005). Response functions analyze the sensitivity of proxy data to the climate. Transfer functions are statistical equations relating proxy records and climatic series over the calibration period. Calibration shows how, and to what extent, proxy material is climate dependent. It is assumed that the modern relationships observed have operated, unchanged, throughout the period of interest (the principle of uniformitarianism) (Fritts 1976; Bradley 1999). It is commonly stated as: the present is the key to the past.

For those cases where a strong covariation between climate and the proxy record is evident and relatively unchangeable as a function of time, a statistical equation for the relationship can successfully be established and verifi ed with independent data. Verifi cation is processed by visual and statistical means including several tests of association (Fritts 1976; Briffa et al. 1988). The statistical procedures ensure that the maximum amount of climate variance in explained by the proxy parameters over the calibration period. However, in assessing the reconstruction performance, there is no substitute for independent data. Many reconstructions split the climate data into two halves, calibrating on one half and verifying on the other (Bradley and Jones 1992). The process may then be repeated, reversing the calibration and verifi cation periods.

Some observations may not need direct calibration if recent comparable observations are available. This applies to such data as early weather diaries and instrumental measurements, rain and snow frequency, dates of the fi rst and last snowfall, and river freeze and thaw dates (Bradley and Jones 1992).

1.3 Context of early meteorological observations

The direct weather data, as described below, have characteristics which make them an exceptionally valuable source of palaeoclimatic information. The most important of these sources from Finland have been discussed by Vesajoki and Tornberg (1994) and Vesajoki et al. (1996). Early meteorological observations in particular can extend the modern climatic record backwards in time and increase its length suffi ciently to improve the existing statistics on climate variability. Holopainen (2004) has summarized these activities from Finland together with information about which archives the diaries are stored in.

9

Figure 1. Page of the meteorological journal from Turku by Johan Leche, December 1753.

10

These journals include not only temperature data, but also observations of air pressure, wind directions, wind speed, cloud coverage, precipitation type, amount of precipitation and weather phenomena (Figure 1). These data can be used to compute time series of monthly values of measured weather variables, similar to modern meteorological data (Holopainen 1995, 2004; Vesajoki and Holopainen 1995, Vesajoki et. al. 1995; Holopainen and Vesajoki 2001). Early meteorological observations are, however, often fragmentary and discontinuous (in relation to the modern meteorological time series) (Heino 1994; Holopainen 1995; Holopainen 2004), and perhaps that is the reason why they have not been regularly used in palaeoclimatological calibrations in Finland. Despite these uncertainties, usage of the earliest possible meteorological observational data can improve our understanding of the dependency of different proxies on climate over the past centuries. The use of such data also provides a temporal context for this study.

1.4 Objectives of the studies

The aim of this thesis has been to collect and analyze potential and useful proxies to re-construct climate variability in Finland before the era of modern instrumental meteorological observations (papers I, II, III, IV). With such information, an attempt to understand the causes of such climatic variations in spring times is presented in paper IV.

In paper I, the background of the temperature series of Turku, which covers the period 1748-1800, is presented, and the mean temperatures obtained from the series are compared with existing neighbouring temperature series to evaluate the reliability of the Turku series.

In Finland, ring widths of Scots pine have been used to reconstruct summer temperatures (e.g. Lindholm 1996; Eronen et al. 2002; Helama et al. 2002; Helama 2004). At the coniferous forest limit of northernmost Finland, a strong positive correlation between ring widths and the temperature of growing seasons (especially for July) has been reported in several papers (e.g. Sirén 1961; Mikola 1962; Lindholm 1996). Previous climate-growth studies have operated within the time span of modern meteorological data (over the last 120 years at most); therefore early temperature data from Tornio (1737-1749) and Ylitornio (1801-1838) offer an interesting opportunity to study the temporal behaviour of the climate-growth relationship during the past three centuries in northern Finland (paper II).

Plant phenological observations in Finland began in the mid-eighteenth century on the initiative of the Swedish botanist Carl von Linné (Lappalainen and Heikinheimo 1992; Häkkinen 1999). This data, however, has not been regularly used in climate-indicator studies. One reason may be the fragmentary and discontinuous nature of these data compared to modern phenological time series. Another reason may be the lack of regional meteorological data. Fortunately, old time series of climatological measurements are available from Turku for the period 1748-1823 (paper I; Holopainen 2004). Plant phenological data with tree-rings from south-west Finland since 1750 were examined as a potential palaeoclimate indicator (paper III).

No single proxy alone is adequate for reconstructing the full spectrum of past climate variability (e.g. Mann 2002). In addition to the plant phenological observations presented in paper III, the region of south-west Finland provided other potential sources for recon-structing past climate. The idea of reconstructing spring temperature variations in south-west Finland came from two previous studies. In Finland, the annual mean temperature has increased by 0.76°C in the 20th century, and most of the warming has occurred in spring (Tuomenvirta 2004). Interestingly, the early climatological data collected in Turku indicates that spring temperatures in south-west Finland during the period of 1748 to 1823 were 0.68°C colder than during the period of 1961 to 1990 (Holopainen 2004). In paper IV, spring temperatures were reconstructed using a set of varved sediments and documentary records

11

starting from 1750 from south-west Finland. This approach also attempts to demonstrate how to bridge the gap between early (1748-1823) and modern meteorological (1873 onwards) observation series by utilizing different kinds of proxy data in conjuction.

2 Materials and methods

2.1 Analyses of the early meteorological observations

The present dataset includes early meteorological observation series from Turku (1748-1823), Tornio (1737-1749) and Ylitornio (1792-1838). The journals mentioned above are the most complete, cover the longest period, and are probably the most useful for palaeo-climatological purposes. A common feature of these datasets is varying observation times. Observations were usually taken two to three times per day at varying hours. In order to further increase the accuracy of the average temperatures, the given observational hours and the observed diurnal temperature cycle have been taken into account by using the formula (Moberg and Bergström 1996):

td = (t1 + ∆1 + t2 + ∆2 +... tn + ∆n)/n

where td is diurnal mean temperature, t1, t2, …tn denote the recorded temperatures and ∆1, ∆2, …∆n denote the mean deviation from the diurnal mean temperature at the actual observation hour as determined from modern data (1956-1965) by Heino (1973). The number of terms, n, is determined by the number of observations made each day. The monthly mean temperature has been obtained as the average of all daily means. This technique has been used to determined monthly mean temperatures from Turku from 1748 to 1800 (Holopainen 1999), Tornio from 1737 to 1749 (Holopainen and Vesajoki 2001) and Ylitornio from October 1800 until December 1838. Furthermore, the Turku temperature series was later expanded and homogenized to cover the period 1801-1823 by Holopainen (2004). As a consequence, homogenized temperature data was used in the analyses of papers II, III and IV.

2.2 Tree-ring data analyses

Tree-ring material of Scots pine (Pinus sylvestris L.) has been collected from northern Finland (paper II) and south-west Finland (paper III). In paper II, ring-width data were used to develop three independent living-tree chronologies from Peuravuono (68º32’N, 28º00’E), Karhunpesäkivi (68º49’N, 27º18’E) and Riekkovaara (68º25’N, 27º16’E). Due to the low number of subfossil tree-ring series over the 18th and 19th centuries, one regional subfossil pine chronology was constructed from tree trunks from the bottom sediment of small lakes in northern Finland (Figure 1 in paper II). In paper III, cores from living trees along the shore of Lake Kaitajärvi and Lake Vähä-Melkutinjärvi (60º40’N, 23º35’E) were extracted by an increment borer. Subfossil samples were collected from tree trunks (megafossils) from the bottom sediment of the same lake.

Disks were cut by saw after lifting the trunks to the surface, and sampled trunks were returned to the lake. After this, the ring widths were measured to the nearest one-hundredth of a millimeter. Series of ring widths were carefully cross-dated (Fritts 1976) using several numerical procedures (Holmes 1983; Van Deusen 1990) in addition to visual comparisons of the plotted measurement series

12

I

on the computer screen. Individual ring-width series of trees were indexed in a process called tree-ring standardization (Fritts 1976; Cook 1985; Helama 2004) in order to remove the agesize related trend in radial growth. This was done using a negative exponential curve, a linear regression line or a line through the series mean (Fritts et al. 1969). Indices were derived from the modelled curve by division. Due to high serial correlation, indices were prewhitened using Box and Jenkins (1970) methods of autoregressive and moving average time series modelling (Cook 1985; Monserud 1986; Guiot 1986; Biondi and Swetnam 1987). The order of the autoregressive moving average process was determined using Akaike’s (1974) Information Criteria. Prewhitening transforms autocorrelated series into a series of serially independent observations. These indices were used in paper II and III to compute correlations between climate and tree-rings.

In paper II, chronology confi dence was determined using statistics of Expressed Population Signal (EPS) (Wigley et al. 1984). EPS was computed as a function of mean intertree correlation and sample size, and any values greater than 0.85 are indications of reliable chronology.

2.3 Method for combining the plant phenological data

In paper III, the plant phenological data from south-west Finland compiled by Linkola (1924) were examined for a potential proxy as a palaeoclimatological indicator. While several phenological events were observed in the eighteenth and the beginning of nineteenth centuries, many of the time series span only a few years. Hence, the challenge in paper III is to combine separate records of different plant species into a single long continuous series (Table 1 in paper III).

A network of phenological observation sites represents a variety of years, latitudes, altitudes, aspects of continentality (proximity of the Baltic Sea), and microclimates. Therefore, in order to avoid possible biases due to the variability of data availability each year, each phenological series of observation dates were transferred into dimensionless phenological indices. In paper III, all site- and phenomenon-specifi c series were fi rst standardized to present an average of zero and a standard deviation of one. The mean phenomenon-specifi c series were averaged as an arithmetic mean for annually resolved time series representing the variability in the particular plant phenomenon. As a result, each phenomenon-specifi c mean series was based on normalized site-specifi c index se-ries. These series were compared to each other, to tree-ring index series, and to early and modern meteorological time series. Finally, the mean phenological index was computed as an average of all available phenomenon-specifi c index series. The mean series were then calibrated and further verifi ed against the instrumental climate records.

2.4 Verifi cation of the reconstructions

In order to demonstrate the skill of the reconstruction of the data, Pearson correlations, Reduction of Error (RE) and Coeffi cient of Error (CE) statistics were used (Fritts 1976; Briffa et al. 1988). RE and CE are the measures of the variance in common between the observed and reconstructed climatic variables, and any positive value of RE or CE indicates a reasonable skill in the reconstruction.

13

II

2.5 Multiproxy reconstruction

Guiot (1992), Mann et al. (1999) and Mann (2002) have demonstrated that a combined approach using complementary proxy and documentary data increases reconstruction reliability. Spring temperatures were reconstructed with a multiproxy database for south-west Finland covering the period since 1750 (paper IV). The proxy records used here were ice break-up in the Aurajoki River (Hällström 1839; Johansson 1932), the Baltic Sea ice extent (Seinä and Palosuo 1993, 1996; Seinä et al. 2001), the plant phenological index (paper III) and the annual varve thickness in Lake Pyhäjärvi (Itkonen and Salonen 1994). Due to their different characteristics, the records were integrated into one palaeo-climate model using time-scale dependent calibration techniques. Short- and long-period time scales were separated in the case of each series using cubic smoothing splines (Cook and Peters 1981). Then high and low frequency variabilities were reconstructed separately and combined into the fi nal model. Details of the methods used in paper IV are given by Guiot (1985) and Briffa et al. (1988).

3 Results

The main results and conclusions drawn in separate papers of this volume are presented in the sections below (3.1-3.4). These results are further discussed and summarized in the context of previous palaeoclimatological literature in the subsequent section (4).

3.1 The early temperature records from Turku

The monthly temperatures of the early meteorological observations from Turku covering the latter part of the eighteenth century were studied fi rst as a potential proxy record before the routine meteorological observation series from Helsinki (paper I). The reliability of the calculated mean temperatures has been evaluated by comparing them with those of the contemporary temperature records of Stockholm, St. Petersburg and Uppsala. The resulting monthly, seasonal and yearly mean temperatures of 1748 to 1800 were compared with present day mean values (1961-1990): the comparison suggests that the winters of the period 1749-1800 were 0.8ºC colder than current winters, while the summers during the period 1749-1800 were 0.4ºC warmer than current summers. Over the same period, springs were 0.9ºC and autumns 0.1ºC colder than today.

From the palaeoclimatological point of view, one clear advantage that early meteorological data has over other proxy data is that observational data provide daily information for all months of the year, something that other proxies never do. As in all early meteorological observations, there are uncertainties in these series. Information about the measuring instruments and their position is insuffi cient, except for the observation period of 1748 to 1763 conducted by Johan Leche, who gives a detailed picture of the construction and calibration of the instruments. Observation times changed considerably from time to time, and during the period of 1764-1785 in most cases only two measurements a day were taken. After the meeting of the Societas Meteorologica Palatina in Mannheim in 1785, three observations a day were set at 6 a.m., 2 p.m. and 10 p.m. In the series, there is a gap of three months in 1781, and the observations for the years 1795 and 1796 are missing altogether. Furthermore, the original data from the period of 1801-1826, some extreme values and summaries excluded, seem to have disappeared in the great fi re of Turku in 1827.

14

Table 1. Mean temperatures (Tm) and mean temperature anomalies (∆Tm) for 30-year periods. Anomalies with respect to 1961-1990 are denoted by ∆Tm. Modifi ed from Holopainen (2004).

1751-1780 1781-1810 1901-1930 1931-1960 1961-1990

Winter Tm -5.99 -7.29 -4.79 -4.93 -5.24

∆Tm -0.76 -2.05 0.45 0.31

Spring Tm 2.94 1.87 2.79 2.94 3.47

∆Tm -0.53 -1.60 -0.68 -0.53

Summer Tm 16.04 16.53 14.91 15.88 15.61

∆Tm 0.43 0.92 -0.70 0.27

Autumn Tm 5.41 5.29 5.02 5.79 5.64

∆Tm -0.23 -0.35 -0.62 0.15

Annual Tm 4.63 4.11 4.49 4.92 4.87

∆Tm -0.24 -0.76 -0.38 0.05

In order to further increase the accuracy of the average temperatures covering the period 1748-1800, a technique which was described in section 2.1 was used (Holopainen 1999; Holopainen 2004). Table 1 summarizes the variations of seasonal and annual temperatures for fi ve consecutive 30-year periods in Turku, south-west Finland. Climate in south-west Finland has fl uctuated between warmer and colder conditions: on an annual level the coldest consecutive 30-year period was 1781-1810 and the warmest 1931-1960. On a seasonal level, the summers from the periods 1751-80 and 1781-1810 were warmer than the present standard normal period, while the winters were evidently colder. The last 30-year spring period was the warmest on record, by 1.60 °C warmer than the period 1781-1810. In the period 1931-1960 autumns were warmer, by 0.15 to 0.77 °C, than other periods.

3.2 Climate-tree growth relationship from the 18th to the 20th century in northern Finland

The analysis of temperature records from Tornio (1737-1749) and Ylitornio (1800-1838) was a continuation of the homogenization procedure of early meteorological observations. In paper II, these data were used for the analysis of the importance of summer temperatures on the tree-ring growth of Scots pine (Pinus sylvestris L.) over the past three centuries. Three living-tree chronologies, one subfossil pine chronology and one composite tree-ring chronology were constructed from latitudinal and altitudinal forest limits of pine in northern Finland and compared with early and modern climatological data. Response functions were derived by means of using fi ve subperiods as follows: 1738-1748, 1802-1822, 1825-1835, 1861-1926 and 1927-1992. It was shown that correlations between ring widths and mid-summer (July) mean temperatures did not vary signifi cantly as a function of time. Early summer (June) and late summer (August) mean temperatures were secondary in relation to mid-summer temperatures in controlling the radial growth. Early summer temperatures governed pine radial growth most clearly during the 19th century, whereas late summer temperatures had the strongest infl uence on ring widths during the 18th century and the latter part of the 20th century. There was no clear signature of temporally reduced sensitivity of Scots pine ring widths to mid-summer temperatures over the periods of the early meteorological observations. The subfossil pine chronology, constructed using pines recovered from small lakes along the forest limit zone, showed a consistent pattern of response to summer temperatures in relation to living-tree chronologies.

15

3.3 Early plant phenological data and tree-rings from south-west Finland as a palaeoclimate indicator

Plant phenological data and tree-rings have been tested for their palaeoclimatic value in south-west Finland since 1750 (paper III). The information from fragmentary, partly overlapping, partly non-systematically biased plant phenological records of 14 different phenomena (a total of 3144 observations) was combined into one continuous time series of phenological indices. The method is described in detail in section 2.3. These series were compared to each other, to living tree and subfossil tree-rings, and to early and modern meteorological time series. Phenological indices showed a strong positive correlation with February to June temperatures indicating that they can be used for palaeoclimate interpretations. In contrast, the correlations between phenological indices and pre-cipitation data were around zero. Furthermore, the correlations between the studied tree-rings and spring temperatures varied as a function of time, and hence, their use in palaeo-climate reconstruction is questionable. The use of present tree-ring datasets for palaeo-climate purposes may become possible after the application of more sophisticated cali-bration methods.

3.4 A multiproxy reconstruction of spring temperatures in south-west Finland

In order to study the spring temperature variation in south-west Finland from 1750 until the present, a reconstruction of a multiproxy database was analysed (paper IV). The proxy records used here were the ice break-up in the Aurajoki River, the Baltic Sea ice extent, the plant phenological index and the annual varve thickness in Lake Pyhäjärvi. This study was in many respects based on the use of the oldest phenological observations from south-west Finland. Each record has its limitations since the instructions on how the data were collected are patchy, and the conditions in which the data were collected are not always known. Despite this, these old observation series can be utilized by integrating them into one palaeoclimate model using time-scale dependent calibration techniques.

1750 1800 1850 1900 1950 2000

-4

-2

0

2

4

Tem

pera

ture

ano

mal

y (d

eg)

Calendar year

Figure 2. Reconstruction of the spring temperatures (February-June) through multiproxy data (1750-1906) and instrumentally observed (1907-2004 in Turku) temperatures for south-west Finland from 1750 to the present (reference period 1961-90). Individual years are represented as a thin line in the diagram. The values are fi ltered with a low-pass fi lter as described in Fritts (1976). In practice, the fi ltration method which was used fi lters out fl uctuations of less than 8 years.

16

The reconstruction was verifi ed with statistics showing a high degree of validity between the reconstructed and observed temperatures. The reconstruction demonstrates that springs have featured a warming trend since around the 1850s (Figure 2.). This is in contrast with the period of 1750 to around 1850 when springs featured larger low-frequency variability, as well as a smaller range of annual temperature variation. The coldest spring-times occurred around the 1840s and 1850s and the fi rst decade of the 19th century. Particularly warm springtimes occurred in the 1760s, 1790s, 1820s, 1930s, 1970s and from 1987 onwards, although in this period cold springs occurred, such as the springs of 1994 and 1996. Long-term temperature changes have been related to changes in atmos-pheric circulation, such as the North Atlantic Oscillation (NAO) (February-June).

17

4 Discussion

4.1 Early efforts to organize weather observations

The rise of experimental science in the 17th century changed the whole image of physics as well as meteorology. The development of meteorology from an almost completely theoretical discipline to an experimental science par excellence at the University of Turku has been described in detail by Kallinen (1995). The development of instruments, especially the invention of the barometer and thermometer, played a key role in this change. Both instruments were inventions of the 17th century (Middleton 1969; Camuffo 2002).

Soon, individuals and organizations began to establish networks of meteorological instruments to help quantify and record the weather. The earliest network was organized in 1653-54 by the Grand Duke of Tuscany, Ferdinand II de’ Medici. This network included a number of Italian stations, i.e. Florence, Vallombrosa, Cutigliano, Bologna, Parma, Milan and, in addition, Paris (France), Innsbruck (Austria), Osnabruck (Germany) and Warsaw (Poland). The network’s activities ended in 1667. This international initiative was followed by attempts to establish more local network-based undertakings, beginning with James Jurin and the Royal Society of London (1724-1735), Luis Cotte and the Societé Royale de Médicine, Paris (1776-1786) and Karl Theodor von Pfalz and John Jacob Hemmer and the Societas Meteorologica Palatina, Mannheim (1781-1792) (Camuffo 2002; see also Kington 1988). Other well-known, though more recently derived, long meteorological series are those prepared from De Bilt (from 1706) of the Netherlands and the Central England temperature series (Jones and Bradley 1992). The latter is a monthly temperature series from 1659 and a daily one from 1772 (Manley 1974; Parker et al. 1992). Recently, long-period daily series of temperature and pressure have been produced by the IMPROVE project (Camuffo and Jones 2002) using data from Padova (from 1725), Uppsala (1722), St. Petersburg (1743), Stockholm (1756), Milan (1763), Central Belgium (1767) and San Fernando/Cadiz (1786).

The earliest organized meteorological observations in Finland were initiated around 1730, when H. D. Spöring, Professor of Medicine, started observations (Johansson 1913; Seppinen 1988; Heino 1994; paper I; Holopainen 2004). Temperature measurements, however, were restricted to the year 1731. In addition, there are observations from one year before or after the period 1730-31, but the date of the records is missing (Holopainen 2004). More systematic and carefully documented measurements were started on October 7, 1748 by Spöring’s successor, Professor Johan Leche, and the observations were published in the Proceedings of the Royal Swedish Academy of Sciences (1762a, 1762b; 1763a, 1763b, 1763c, 1763d, 1763e). After Leche’s death, the observations were continued using the same instruments by Pehr Kalm, Professor of Economics, until October 11, 1779. However, the identity of the observer is uncertain in the years 1764-1771. Kalm was succeeded by Chief Gardener Blomberg, who was responsible for the observations from October 6 until November 3, 1786. According to Johansson (1913), a teacher named Nordberg was responsible for the observations for the next seven years, and he was followed by Professors Mether and Hellenius. From 1801 through 1825 the observations were performed by Gustaf Hällström, Professor of Physics. The original measurement records, however, excluding some extreme values and summaries and the observations from 1827, seem to have disappeared in the great fi re of Turku on September 4, 1827 (paper I, Holopainen 2004). After this event Hällström initiated meteorological observations in Helsinki, on October 4, 1828 (Johansson 1906; Heino 1994).

The most interesting source of historical information is the data from Tornio, where observations were made from 1737 to 1749 (Vesajoki et al. 1995; Holopainen and Vesajoki 2001 and references therein). In addition to Turku and Tornio, climatological observations were also initiated at several localities in Finland before the year 1828. Holopainen (2004) has summarized these activities

18

together with information about which archives the diaries are stored in. The establishment of stations since 1785 had been infl uenced by the Meteorological Society of Mannheim (set up in 1780), which arranged one of the fi rst international networks of meteorological observations (Johansson 1913, Cappel 1980; Kington 1988; Heino 1994; Moberg 1996; paper I; Holopainen 2004). In May 1785, King Gustav III promulgated a circular to the boards (konsistorierna) of the gymnasiums in Sweden. The senior masters and priests were obliged to make observations three times a day, and after each year send their observation journals to the Academy (Moberg 1996). Some of the observers were very enthusiastic and performed the observations for several decades. An example of this is the organ player Johan Portin, who made meteorological observations in Ylitornio from 1792 to 1838 (paper II).

In Finland, several occasional meteorological observations covered time spans of a few years (Holopainen 2004), although it is questionable if these observations should be considered as being incorporated in the national programme. It is also possible that several journals of observations are still hiding in the Finnish archives to be analyzed, since they may be fi led in categories other than documents of climatology in the Finnish National Archive or the library of Finnish Meteorological Institute. Examples of such are climatological time series from Hailuoto (Carlö in Swedish) (1817-1836) and Vöyri (Wöro in Swedish) (1800-1824), which are described by Hällström (1843a, 1843b). Still, the earliest climatological records are likely from Turku and Tornio.

4.2 The role of scientifi c academies and volunteers

The turning point in the development of the climatological network in Finland can be seen as the year 1881, when the Finnish Meteorological Institute was established (Heino 1994). In those days, phenological observations were also closely connected with offi cial meteorological activities (Lappalainen and Heikinheimo 1992). The present work has reviewed and studied hundreds of thousands of single observations representing the work of dozens, probably hundreds of people over the past 300 years. Their work is greatly acknowledged here. That work has been largely voluntary, especially before the 1900s, and was based on people’s interest in their surrounding environment (Heino 1994).

Both meteorological and phenological observation activities have been established and maintained by various scientifi c academies in Finland since the 1750s. The fi rst recorded initiatives to systematically gather phenological observations in Finland and Sweden were by the Swedish astronomer Anders Celsius and the Swedish botanist Karl von Linné (Johansson 1945; Lappalainen and Heikinheimo 1992; Häkkinen 1999). The fi rst published Finnish observations were made by Professor Johan Leche in Turku (Leche 1763e; see also Leche 1889). The next steps in the development of meteorological and phenological observation activities were made during the years 1792-1795 by the society “Pro Natura” and after that by the Finnish Economic Society (Suomen Talousseura, Finska Hushållningssällskapet), which undertook as one of their tasks the establishment of meteorological stations (Johansson 1945; Heino 1994). A more comprehensive climatological-phenological monitoring program was initiated after the foundation of the Finnish Society of Sciences and Letters (Suomen Tiedeseura, Finska Vetenskaps-Societen) in 1838 in Helsinki (Lappalainen and Heikinheimo 1992; Häkkinen 1999). Its observations started in 1846 and continued until 1965. Since then, the collection of phenological observations has been managed jointly by the Finnish Society of Sciences and Letters, the Finnish Museum of Natural History and various biological research stations (Häkkinen 1999). Since 1961, the Finnish Youth Association for Environmental Protection (Luonto-liitto) has also organized the collection of phenological observations (Lappalainen and Heikinheimo 1992) followed by the Finnish Forest Research Institute in 1997 (http://www.metla.fi /metinfo/fenologia/index-en.htm).

19

4.3 The utilization of early plant phenological data

The records of the foundation of the Finnish Society of Sciences and Letters were used in many studies on the relationship between temperature and the timing of plant phenological events, especially after 1896 (e.g. Lappalainen 1994; Heikinheimo and Lappalainen 1997; Häkkinen et al. 1995, 1998; Häkkinen 1999; Linkosalo 2000; Linkosalo et al. 1996). The fi rst studies on climate change in Finland, however, had already been carried out in the 19th century when Moberg (1865; 1868) studied possible trends in the timing of phenological events over the years by means of linear regression (Häkkinen 1999). Moberg (1894) also compiled and published phenological observations from Finland from the 1750s until 1845. In addition to Moberg’s work (1857a; 1857b), the variation in the timing of events between years was studied, and tables with mean times of the occurrences of phenological events at different geographical locations were constructed by Johansson (1911; 1945) and Linkola (1924).

In paper III, the data compiled by Linkola (1924) were examined for a potential proxy as a palaeoclimatological indicator (Table 1 in paper III). Methods for combining phenological time series have been applied in several studies (Häkkinen et al. 1995; Linkosalo et al. 1996; Linkosalo 2000; Schaber and Badeck 2002; Chuine et al. 2004). It is important to note that the preceding studies are based on the combination of time series for a single species of plant. While several phenological events were observed in the eighteenth century and at the beginning of the nineteenth century, many of the time series span only a few years. An attempt to combine the information from fragmentary, partly overlapping, partly non-systematically biased plant phenological records into one continuous time series of phenological indices was done in paper III.

The indices were found to be reliable indicators of the February to June temperature variations. Based on the correlations over the calibration and verifi cation period, as well as on RE and CE statistics, the most reliable reconstructions were derived for the spring temperature time window from February to May or June, or from March to May or June (Figure 5 in paper III). When attempting to reconstruct the spring climate for a narrower time window, for example from March to April or without March, temperatures generally failed due to the negative values of the CE statistics. Consistency can be observed not only at interannual but also at longer, multidecadal time scales. The two records, the observed and reconstructed temperature records, match better over the calibration than the verifi cation period, a feature which is common to nearly all palaeoclimate reconstruction models due to regression statistics. Apart from statistical issues, dissimilarities may also occur for a variety of other reasons. Early parts of the meteorological and phenological series contain fewer observations (compared to modern data) made by more or less untrained observers, and therefore the information is usually insuffi cient or does not exist at all. Despite these uncertainties, the 18th and 19th century phenological observations (at least in the form of regionally normalized phenological mean indices) provide statistically reliable palaeoclimate information for selected spring months. Since it is not necessary to have the same species observed at all stations, this approach could also serve as a suitable method for larger scale studies for which it is often diffi cult to fi nd a continuous dataset of common observed phases.

20

4.4 Time dependency of climate sensitivity in tree-rings

In this study process, the relationship between climate and ring width was studied at the northern forest limit (paper II) and in south-west Finland (paper III).

The dendroecological implication of the uniformitarianism principle is that the relationships which govern environmental and tree growth conditions must have all been the same in the past as in the present (Fritts and Swetnam 1989). Interestingly, Briffa et al. (1998a) reported reduced sensitivity in high-latitude tree-rings of several species to growing season temperatures during the latter part of the 20th century. Possible contributors to dissociation could increase the infl uence of a late snow delaying the onset of season tree growth or enhanced ultraviolet radiation on the photosynthetic process, or be a possible consequence of falling concentrations of ozone in the stratosphere (Vaganov et al.1999; Briffa et al. 2004; Helama 2004). This premise was the motivation for the study of the relationship between summer temperatures and Scots pine ring widths in Northern Finland in paper II, since early meteorological observations from Tornio and Ylitornio have been available since the 1730s. According to paper II, it was found that there was no clear signature of temporally reduced sensitivity of Scots pine ring widths to mid-summer temperatures over the periods of early meteorological observations from Tornio and Ylitornio. Furthermore, the response of living-tree and subfossil pine chronologies to summer temperature did not show signifi cantly dissimilar patterns of response to the temperatures of the different summer months. These fi ndings suggest that ring widths from subfossil wood could, at least in optimal forest-limit conditions as described in paper II, be used as indicators of past mid-summer temperatures.

On the other hand, the correlations between the studied tree-rings and the spring temperatures in south-west Finland varied as a function of time, and thus their use in palaeoclimate reconstruction is questionable. It should be noted that the relationship between tree-rings, or any other similar natural time series, may be time-scale dependent (Guiot 1985; Timm et al. 2004) and that the use of the present tree-ring dataset for palaeoclimate purposes may become possible after more sophisticated calibration methods than those applied in paper III. Furthermore, the tree-ring data from south-west Finland may correlate in a more straightforward fashion with some other variables, such as soil moisture and large-scale atmospheric indices e.g. North Atlantic Oscillation (Helama 2004; Helama and Timonen 2004). For example, Scots pine tree-rings are known to correlate signifi cantly with early summer precipitation in south-east Finland (Lindholm et al. 1997; Helama and Lindholm 2003).

4.5 The value of seasonal multiproxy reconstruction

The proxy-based reconstruction of past climates is based on a methodology whereby a modern calibration dataset (proxy and instrumental climate data covering the same period) is used to derive a statistical relationship between the two from which a transfer function is derived to describe an estimation of the instrumental data from the proxy data. This transfer function is then used to estimate the climate back in time from the entire length of the proxy series.

Both documentary and geophysical proxy records were integrated in paper IV into one palaeoclimate model. Although each record has its drawbacks since the instructions on how to collect the data have been insuffi cient and the conditions in which data have been collected are not always known, these old observation series can be utilized by integrating them into one palaeoclimate model using time-scale dependent calibration techniques. The reconstruction was verifi ed with statistics showing a high degree of validity between the reconstructed and observed temperatures (Table 3 in paper IV).

21

Seasonal multiproxy reconstructions are important since the reconstruction of annual temperature is sometimes biased in favour of a specifi c season. Jones et al. (2003) have shown that winters have warmed relative to summers during the last two centuries compared with earlier centuries and that usually the winter season in the Northern Hemisphere dominates the annual average (see also Guiot et al. 2005). Therefore it is crucial to reconstruct a seasonal signal of temperature variability by the proxies. This is not always easy since each proxy responds to temperature in its own way. Using seasonal multiproxy reconstruction it could be expected that the potentially non-climatic variations in some of the records will be largely cancelled out.

4.6 Interpretation of the spring temperature variability since 1750

Climate is known to vary on different spatiotemporal scales, from interannual climatic variability to very long-period variations related to the evolution of the atmosphere and changes in the lithosphere (Bradley 1999). Climate variability can arise from a number of factors, which has given rise to discussion in several studies (e.g. Crowley 2000; Bauer et al. 2003; González-Rouco et al. 2003; von Storch et al. 2004; Zorita et al. 2004; Goosse et al. 2005a, 2005b; Weber 2005; Hegerl et al. 2003, 2006). On time scales of centuries, there is variability caused by external forcing, including natural (e.g. volcanic and solar radiative) and anthropogenic (greenhouse gas and sulphate aerosol) infl uences. There are also internal modes of climate system variability (e.g. North Atlantic Oscillation (NAO) and Arctic Oscillation), which may also produce signifi cant changes in regional climate in which certain modes may be favoured or amplifi ed by external forcing (Bradley 1999).

The position of Finland between the relatively warm North Atlantic and the large Eurasian land mass makes it very sensitive to any changes in circulation (Heino 1994). The spring temperature changes in paper IV have been closely associated with changes in atmospheric circulation, as indicated by the NAO (February-June) as it acts in the whole of Scandinavia and western Europe (Hurrell and van Loon 1997; Chen and Hellström 1999; Yoo and D’Odorico 2002). Westerly winds bringing moist air from the Atlantic result in mild wintertime and cool summertime temperatures, while continental air fl owing from the east has opposite effects. The frequent westerly winds of the 1990s have contributed to the mild winter months in Finland (Tuomenvirta 2004).

The weak response of the global forcings on a local scale does not at all mean that there is no impact on the temperature. The question of spatial and temporal scales of climate variability is crucial in this matter. Regional or local climate is generally much more variable than climate on a hemispheric or global scale because regional or local variations in one region are compensated to some extent by opposite variations elsewhere. An illustrative experiment in this respect is given by Goosse et al. (2005b), who concluded that on hemispheric and nearly hemispheric scales, the impact of the main natural and anthropogenic forcing is clearly stronger than on regional or local scales.

The question of spatial and temporal scales of climate variability relate closely to the re-lationship of the 20th century climate to the past centuries and millennia. For these periods the terms commonly used are the Medieval Climatic Anomaly (MCA; AD 900-1300), the Little Ice Age (LIA; AD 1300-1860), and the Recent Global Warming phase (RGW; 1860-present). Their spatial scales of action extend from the regional to the hemispheric and global. The timing of these cold and warm periods has been demonstrated to vary geographically and, in particular, on regional and European scales (Bradley and Jones 1993; Hughes and Diaz 1994; Bradley et al. 2003; Goosse et al. 2005b).

The survey of the multiproxy data of spring temperatures covered only a latter part of the LIA period, thus only a few remarks can be made here. Before 1860, a monotonously cold period had not been experienced in south-west Finland: certain intervals were colder than others. The coldest

22

springtimes in the spring temperature record of the 1840s and 1850s are known from historic records as periods of starvation and high mortality, especially in the area of Helsinki (Solantie 1999). Starvation and high mortality continued in Finland, especially during the years 1867-68. In the Famine Year of 1867 cold weather conditions dominated the whole of May and continued during the fi rst half of June (Jantunen and Ruosteenoja 2000). The growing period remained short, and a severe crop failure resulted. In the following year about 8% of the population of Finland perished from hunger, typhoid fever and other infectious diseases (Turpeinen 1986). This was about three times higher than normal.

Recently Goosse et al. (2005b) concluded that the concepts of the Medieval Warm Period and Little Ice Age were both hemispheric-scale phenomena since the temperatures averaged over the Northern Hemisphere were respectively generally higher/lower during those periods because of stronger/weaker external forcing at that time. On the local scale, the sign of the internal temperature variations determines to what extent the forced response will be actually visible or even masked by internal noise. Moreover, Matthews and Briffa (2005) have re-evaluated the concept of the Little Ice Age in detail. They concluded that the concept of the Little Ice Age will remain useful only by continuing to incorporate the temporal and spatial complexities of glacier and climatic variations as they become better known and by refl ecting an improved understanding of the Earth-atmosphere-ocean system and its forcing factors through the interaction of palaeoclimatic reconstruction with climate modelling.

4.7 Prospects for future studies

The principal aim of any palaeoclimatic study is to improve our understanding of the climate of the past on different spatial scales. Not only temperature reconstructions, but also the use of the data of different climate state variables such as air pressure and precipitation can improve palaeoclimate knowledge. This was not beyond the scope of this thesis, and there is a defi nite need for such work. This approach allows the resulting climate fi eld reconstruction to be analyzed and related synoptically in order to gain more knowledge of the dynamics of past climate variability in the region. Examples of this kind of approach are presented by Wanner and Luterbacher (2002) and Casty (2005).

An analysis of early plant phenological data since 1750 from south-west Finland suggests that this dataset can be used for palaeoclimatological purposes. It was shown that the information from fragmentary, partly overlapping, partly non-systematically biased plant phenological data of different phenomena can be utilized by combining them into one continuous time series of phenological indices. This method made possible for regional estimates of temperature variations in south-west Finland. Similarly, work with the network of early plant phenological observations from elsewhere in Finland since 1750 will produce information about the spatio-temporal climate variability.

In the present work, only ring widths have been measured and analyzed. However, densitometric measurements of tree-rings yield time series which are known to correlate with climate better than ring widths (Briffa et al. 1988a). It is probable that densitometric measurements using the same tree-ring data that have been presented in paper II and paper III will yield improved results or even different types of climatic information if density variables are used in combination with ring-width variables.

Additional proxies for palaeoclimate studies not used in the present work could include the fl uctuations of grain harvest times in the 16th to 19th centuries (Tornberg 1989; Vesajoki et al. 1996). Tornberg (1989) has already described the fl uctuations in the harvests in south-west Finland from the 1550s to the 1860s. The use of these data may help to discover information on year-to-year variability in climatic fl uctuations in southern Finland as well as climatic impacts on society in the past.

23

Utility was a concept that combined the natural sciences and economic policy from the early decades of the 18th century until the 1770s. In this period began the early stages of climatological and phenological observations in Sweden and Finland. According to Niemelä (1998), people of the 18th century subscribed to a deep faith in the endless natural resources of nature, largely based on the theories proposed by Olof Rudbeck in the late 17th century. Physico-theology further suggested that all things in nature had been created for the benefi t and use of man. The immense resources of nature were to be found and utilized, which in turn was the main task of the natural sciences. In relation to climate, utilism manifested an interest in a warming climate by clearing fi elds and draining peatland. If there was to be impending signifi cant warming, even silk culture could be considered a potential fi eld of agrology in Finland (Niemelä 1998). In seeking a more precise focus for future studies, it could be asked from where the impulses for agro-climatological experiments came and what the role is of early meteorological observations in the debate over the warming climate. As was presented in section 3.1, the measurements support the idea of warm summers in the period 1751-1780.

If we review the orientation of society of the past, one could note with irony that the main goal has been reached. Warming has been going on for more than a century on a global scale and also in Finland. To better understand humans’ long-term infl uences on the changing climate system, it is imperative to identify and diagnose the interdecadal variations of the separate, long-term proxy records.

24

5 Conclusions

The objective of this investigation has been to collect and analyze potential and useful proxies to reconstruct climate variability in Finland before the era of modern instrumental meteorological observations. The following conclusions can be drawn based on the results and their interpretation.

● Spring temperatures in south-west Finland since 1750 have fluctuated between warmer and colder conditions on inter-decadal and inter-annual timescales. The multi-proxy reconstruction demonstrates that springs have become warmer and have featured a warming trend since around the 1850s. In the period from 1750 to around 1850, springs have been characterized by larger multidecadal low-frequency variability, as well as by a smaller range of annual temperature variations.

● Reconstructed spring temperature changes have been related to changes in atmospheric circulation, as indicated by the North Atlantic Oscillation (NAO) (February-June).

● The early temperature records provide the most exact information on climate variability before the era of modern meteorological observations, although there are uncertainties about the metadata. One clear advantage that early meteorological data has over other proxy data is that observational data provide daily information for all months of the year, something that other proxies never do.

● Plant phenological records are valid proxy data for reconstructing spring temperatures. The information from fragmentary, partly overlapping, partly non-systematically biased plant phenological records can be utilized by combing them into one continuous time series of phenological indices.

● Tree-ring based palaeoclimate reconstructions provide evidence for summer temperatures. At the coniferous forest limit of northernmost Finland there is a strong positive correlation between the ring widths of Scots pine and growing season temperatures (especially for July). Meanwhile, correlations between the studied tree-rings and spring tem-peratures in south-west Finland varied as a function of time, and thus their use in palaeo-climate reconstruction is questionable. This might be a consequence of the relationship between tree-rings or any other similar natural time series which may be time-scale dependent, and the use of the present tree-ring dataset for palaeoclimate purposes may become possible after more sophisticated calibration methods are applied than those in paper III.

● Multiproxy reconstructions offer a possibility to utilize geological, biological and documentary proxy data in combination. Paper IV discusses the possibility of producing a reliable climate reconstruction through the integration of the oldest documentary and geological proxy records into one model. Verification of the reconstruction over the independent time period shows that the reconstruction is statistically reliable. All the verification statistics show a high degree of validity between the reconstructed and observed temperatures in Turku.

25

6 Acknowledgements

I am most grateful to my supervisor Professor Matti Eronen, who gave me an opportunity to fi nish my Ph.D. and encouraged and supported me during the study process in many ways. Dr. Samuli Helama is greatly thanked for all the help, teamwork and fruitful discussions, as well as for reading and making comments on my writings. Professor Veli-Pekka Salonen supported this work in an important way with encouragements and ideas in the latter part of the work. Dr. Raino Heino, Dr. Samuli Helle, Dr. Risto Häkkinen, Dr. Kirsti Jylhä, Lic. Phil. Juha M. Kajander, M. Sc. Johanna Korhonen, Prof. Markku Kulmala, Prof. Jouko Launiainen, Dr. Tapio Linkosalo, Dr. Anders Moberg, Dr. Heikki Nevanlinna, Dr. Jari Niemelä, Dr. Anni Reissell and M. Sc. Mauri Timonen are thanked for their co-operation and valuable advice during the study process. The careful measurement of ring widths by several assistants at the Saima Centre for Environmental Sciences, Savonlinna, and at the Finnish Forest Research Institute, Rovaniemi Forest Research Station is acknowledged. Prof. Jukka Käyhkö and Dr. Heikki Tuomenvirta reviewed this synopsis, and their work is greatly acknowledged. I owe special thanks to Dr. Heikki Vesajoki, who encouraged me to study documentary proxy records and supported me in the beginning of my studies.

This study was carried out at the Department of Geology, University of Helsinki, and the helpfulness of all my coworkers there and the facilities provided by the institute are all greatly appreciated. The present research was funded by the Academy of Finland (grant to Dr. Heikki Vesajoki in 1992-93), The Jenny and Antti Wihuri Foundation, The Maj and Tor Nessling Foundation, The Integrated Land Ecosystem – Atmosphere Processes Study (iLEAPS), the University of Helsinki (a grant towards the completion of the doctoral dissertation) and the Finnish Cultural Foundation/Satakunta Foundation. All sponsors are gratefully acknowledged. The Graduate School in Geology is acknowledged for their support for the language revision of this thesis.

I warmly thank my parents, Antero and Liisa, for their support and interest during this study. Also my past and present colleagues and friends who have supported me and helped me during this study, in many ways, are greatly acknowledged. My daughter Aino-Maria has participated in researching the theory and practice of the everyday, although I began my work long before she was born. Without her being, however, the study may have never been completed.

Finally, I would like to thank Ulla, for her patience, support and love.

Helsinki, 15 September 2006 J.H.

26

7 References

ACIA 2004. Impacts of a Warming Arctic: Arctic Climate Impact Assessment. Cambridge University Press. http://www.acia.uaf.edu/

Akaike, H. 1974. A new look at the statistical model identifi cation. IEEE Transactions on Automatic Control, AC-19: 716-723.

Bauer, E., Claussen, M., Brovkin, V. and Huenerbein, A. 2003. Assessing climate forcings of the Earth system for the past millennium. Geophysical Research Letters Vol. 30, No. 6, 1276, doi:10.1029/2002GL016639.

Betts, R.A. 2000. Offset of the potential carbon sink from boreal forestation by decreases in surface albedo. Nature 408: 187-190.

Biondi, F. and Swetnam T.W. 1987. Box-Jenkins models of forest interior tree-ring chronologies. Tree-Ring Bulletin 47: 71-96.

Box, G.E.P. and Jenkins, G.M. 1970. Time-series analysis, forecasting and control. Holden-Day, San Fransisco, 575 p.

Bradley R.S. 1999. Paleoclimatology: Reconstructing climates of the Quaternary. Academic Press, San Diego, 610 p.

Bradley, R.S. and Jones, P.D. 1992. Climate since A.D. 1500: Introduction. In: Bradley, R.S. and Jones, P.D. (eds.) Climate since A.D. 1500. Routledge, London, pp. 1-16.

Bradley R.S. and Jones P.D. 1993. “Little Ice Age” summer temperature variations: their nature and relevance to recent global warming trends. The Holocene 3: 367–376.

Bradley, R.S., Briffa, K.R., Cole, J., Hughes, M.K. and Osborn, T.J. 2003. The climate of the last millennium. In: Alverson, K., Bradley, R. S. and Pedersen, T. F. (eds.) Paleoclimate, global change and the future. Springer Verlag, Berlin, pp. 105-141.

Brázdil, R., Pfi ster, C., Wanner, H. von Storch, H. and Luterbacher, J. 2005. Historical climatology in Europe – The state of the art. Climatic Change 70: 363-430.

Briffa, K.R., Jones, P.D., Pilcher, J.R. and Hughes, M.K. 1988: Reconstructing summer temperatures in northern Fennoscandia back to AD 1700 using tree-ring data from Scots pine. Arctic and Al-pine Research 20: 385-394.

Briffa, K.R., Schweingruber, F.H., Jones, P.D., Osborn, T.J., Shiyatov S.G. and Vaganov E.A. 1998a: Reduced sensitivity to recent tree-growth to temperature at high northern latitudes. Nature 391: 678–682.

Briffa, K.R., Jones, P.D., Schweingruber, F.H. and Osborn, T.J. 1998b. Infl uence of volcanic erup-tions on Northern Hemisphere summer temperatures over 600 years. Nature 393: 450–455.

Briffa, K.R., Osborn, T.J., Schweingruber, F.H., Harris, I.C., Jones, P.D., Shiyatov, S.G. and Vaga-nov, E.A. 2001. Low-frequency temperature variations from a northern tree ring density network. Journal of Geophysical Research 106, D3, 2929–2941.

Briffa, K.R., Osborn, T.J., Schweingruber, F.H., Jones, P.D., Shiyatov, S.G. and Vaganov, E.A. 2002. Tree-ring width and density data around the Northern Hemisphere: part 1, local and regional cli-mate signals. Holocene 12: 737–758.

Briffa, K.R., Osborn, T.J. and Schweingruber, F.H. 2004. Large-scale temperature inferences from tree rings: a review. Global and Planetary Change 40: 11–26.

Brohan, P., Kennedy, J.J., Haris, I., Tett, S.F.B. and Jones, P.D. 2006. Uncertainty estimates in region-al and global observed temperature changes: a new dataset from 1850. Journal of Geophysical Research 111, D12106, doi:10.1029/2005JD006548.

27

Camuffo, D. 2002. History of the long series of daily air temperature in Padova (1725-1998). Climatic Change 53: 7-75.

Camuffo, D. and Jones, P.D. (eds.) 2002. Improved understanding of past climate variability from early daily European instrumental sources. Kluwer Academic Publishers, Dordrecht, Boston, London, 392 p.

Cappel, A. 1980. Societas Meteorologica Palatina (1780-1795). Annalen der Meteorologie 16: 10-27.

Casty, C. 2005. Climate variability over the North Atlantic/European and Alpine regions since 1500. Ph. D. Dissertation, University of Bern, 143 p.

Chen, D. and Hellström, C. 1999. The infl uence of the North Atlantic Oscillation on the regional temperature variability in Sweden: spatial and temporal variations. Tellus 51 A: 505-516.

Chuine, I., Yiou, P., Viovy, N., Seguin, B., Daux, V. and Le Roy Ladurie, E. 2004: Historical phenology: Grape ripening as a past climate indicator. Nature 432: 289-290.

Cook, E.R. 1985. A time series analysis approach to tree-ring standardization. Ph. D. Dissertation, University of Arizona, Tucson, 171 p.

Cook, E.R. and Peters, K. 1981. The smoothing spline: A new approach to standardizing forest interior tree-ring width series for dendroclimatic studies. Tree-Ring Bulletin 41: 45-53.

Cook, E.R., Esper, J. and D’Arrigo, R.D. 2004. Extra-tropical Northern Hemisphere land temperature variability over the past 1000 years. Quaternary Science Reviews 23: 2063–2074.

Crowley, T.J. 2000. Causes of climate change over the past 1000 years. Science 289: 270–277.Crowley, T.J. and Lowery, T.S. 2000. How warm was the Medieval warm period? Ambio 29: 51-54.D’Arrigo, R., Wilson, R. and Jacoby, G. 2006. On the long-term context for late twentieth century

warming. Journal of Geophysical Research 111, D03103, doi:10.1029/2005JD006352.Eronen, M. 1990. Muuttuva ilmasto. Terra 102:4: 220-238.

Eronen, M., Zetterberg, P., Briffa, K.R., Lindholm, M., Meriläinen, J. and Timonen, M. 2002. The supra-long Scots pine tree-ring record for Finnish Lapland: Part 1, chronology construction and initial inferences. The Holocene 12, 6: 673-680.

Esper, J., Cook, E.R. and Schweingruber F.H. 2002. Low-frequency signals in long tree-ring chronologies for reconstructing past temperature variability. Science 295: 2250-2253.

Fritts, H.C. 1976. Tree Rings and Climate. Academic Press, London, 567 p.Fritts, H.C., Mosimann, J.E. and Bottorff, C.P. 1969. A revised computer program for standardizing

tree-ring series. Tree-Ring Bulletin 29: 15-20.Fritts, H.C. and Swetnam T.W. 1989. Dendroecology: a tool for evaluating variations in past and

present forest environments. Advances in Ecological Research 19: 111-188.Gonzáles-Rouco, F., von Storch, H. and Zorita, E. 2003. Deep soil temperatures as proxy for surface

air-temperature in a coupled model simulation of the last thousand years. Geophysical Research Letters Vol. 30, No. 21, 2116, doi:10.1029/2003GL018264.

Goosse, H., Crowley, T.J., Zorita, E., Ammann, C.M., Renssen, H. and Driesschaert, E. 2005a. Modelling the climate of the last millennium: what causes the differences between simulations? Geophysical Research Letters Vol. 32, L06710, doi:10.1029/2003GL018264.

Goosse, H., Renssen, H., Timmermann, A. and Bradley, R.S. 2005b. Internal and forced climate variability during the last millennium: a modeldata comparison using ensemble simulations. Quaternary Science Reviews 24: 1345–1360.

Guiot, J. 1985. The extrapolation of recent climatological series with spectral canonical regression. Journal of Climatology 5: 325-335.

28

Guiot, J. 1986. ARMA techniques for modelling tree-ring response to climate and for re-constructing variations of paleoclimates. Ecological Modelling 33: 149-171.

Guiot, J. 1992. The combination of historical documents and biological data in the reconstruction of climate variations in space and time. Paläoklimaforschung – Palaeoclimate Research 7: 93-104.

Guiot, J, Nicault, A., Rathgeber, C., Edouard, J.L., Guibal, F., Pichard, G., and Till, C. 2005. Last-millenium summer-temperature variations in western Europe based on proxy data. The Holocene 15, 4: 489-500.

Häkkinen, R. 1999. Analysis of bud-development theories based on long-term phenological and air temperature series: application to Betula sp. leaves. Ph. D. Dissertation, Finnish Forest Research Institute, Research Papers 754, 59 p.

Häkkinen, R., Linkosalo, T. and Hari, P. 1995. Methods for combining phenological time series: application to bud burst in birch (Betula pendula) in Central Finland for the period 1896-1955. Tree Physiology 15: 721-726.

Häkkinen, R., Linkosalo, T. and Hari, P. 1998. Effects of dormacy and environmental factors on timing of bud burst in Betula pendula. Tree Physiology 18: 707-712.

Hällström, G.G. 1839, ‘Specimina mutanti currente seculo temporis, quo glacies fl uminum annuae dissolutae sunt’, Acta Societatis Scientiarum Fennicae 1, 129-148.

Hällström, G.G. 1843a. Bidrag till kännedom om Finska orters klimatförhållande: om luftvarmen på Carlö. Acta Societatis Scientiarum Fennicae: 119-129.

Hällström, G.G. 1843b. Bidrag till kännedom om Finska orters klimatförhållande: om luftvarmen i Wöro. Acta Societatis Scientiarum Fennicae: 131-143.

Harding, R., Kuhry, P., Christensen, T.R., Sykes, M.T., Dankers, R. and Linden van der S. 2002. Climate feedbacks at the tundra-taiga interface. Ambio 12: 47-55.

Hegerl, G.C., Crowley, T.J., Baum, S.K., Kim, K-Y. and Hyde, W.T. 2003. Detection of volcanic, solar and greenhouse gas signals in paleo-reconstructions of Northern Hemispheric temperature. Geophysical Research Letters 30, 1242, doi:10.1029/2002GL016635.

Hegerl, G.C., Crowley, T.J., Hyde, W.T. and Frame, D.J. 2006. Climate sensitivity constrained by temperature reconstructions over the past seven centuries. Nature 440: 1029-1032.

Heikinheimo, M. and Lappalainen, H. 1997. Dependence of the fl ower bud burst of some plant taxa in Finland on effective temperature sum: implications for climate warming. Annales Botanici Fennici 34: 229-243.

Heino, R. 1973. Lämpötilan vuorokausivaihtelusta ja siihen vaikuttavista tekijöistä. Tutkimusseloste 46, Ilmatieteen laitos, 58 p.

Heino, R. 1994. Climate in Finland during the period of meteorological observations. Finnish Meteorological Institute Contributions 12, Helsinki, 209 p.

Helama, S. 2004. Millennia-long tree-ring chronologies as records of climate variability in Finland. Ph. D. Dissertation, University of Helsinki, Yliopistopaino, 54 p.

Helama, S. and Lindholm, M. 2003. Droughts and rainfall in south-eastern Finland since AD 874, inferred from Scots pine ring-widths. Boreal Environmental Research 8 (2): 171-183.

Helama, S. and Timonen, M. 2004. Meteorologisia aikasarjoja puiden vuosilustoista. Ilmastokatsaus 9(12):1 5-6.

Helama S., Lindholm M., Timonen M., Meriläinen, J. and Eronen M. 2002. The supra-long Scots pine tree-ring record for Finnish Lapland: Part 2, interannual to centennial variability in summer temperatures for 7500 years. The Holocene 12, 6: 681-687.

Holmes, R.L. 1983. Computer-assisted quality control in tree-ring dating and measurement. Tree-Ring Bulletin 43: 69-75.

29

Holopainen, J. 1995. Lounais-Suomen ilmasto 1700-luvun jälkimmäisellä puoliskolla. Licenciate dissertation, University of Joensuu, Department of Geography and Regional Planning, unpub-lished. 66 p.

Holopainen, J. 1999. Turun vuosien 1748-1800 lämpötilamittaussarjan homogenisointi. Terra 111:1: 37-39.

Holopainen, J. 2004. Turun varhainen ilmastollinen havaintosarja. (English abstract: The early climatological records of Turku). Finnish Meteorological Institute Reports 8, Edita, Helsinki, 59 p.

Holopainen, J. and Vesajoki, H. 2001. Varhainen lämpötilahavaintosarja Torniosta vuosilta 1737-1749. Terra 113:3: 196 - 201.

Hughes, M.K. and Diaz, H.F. 1994. Was there a ‘Medieval Warm Period’ and if so, where and when? Climatic Change 26: 109-142.

Hurrell, J.W. and Van Loon, H. 1997. Decadal variations in climate associated with the North Atlan-tic Oscillation. Climatic Change 36: 301–326.

Ingram, M.J., Underhill, D.J. and Wigley, T.M.L. 1978. Historical climatology. Nature 276: 329-334.

IPCC 2001. Climate Change 2001. The Scientifi c Basis. In: Houghton, J.T., Ding, Y., Griggs, D.J., Noguer, M., Linden van der, P.J., Dai, X., Maskell, K., and Johnson, C. A. (ed) Cambridge Uni-versity Press, Cambridge, UK, 881 p.

Itkonen, A. and Salonen, V-P. 1994. The response of sedimentation in three varved lacustrine sequences to air temperature, precipitation and human impact. Journal of Paleolimnology 11:323-332.

Jantunen, J. and Ruosteenoja, K. 2000. Weather conditions in northern Europe in the exceptionally cold spring season of the Famine Year 1867. Geophysica 36(1-2): 69-84.

Johansson, O.V. 1906. Über die Bestimmung der Lufttemperatur am Meteorologischen Observatori-um in Helsingfors. Eine Kritische Studie.Wissenschaftliche Beilage, Meteorologisches Jahrbuch für Finnland Bd. I 1901, 110 p.

Johansson, O.V. 1911. Aus G.G. Hällströms hinterlassenen papieren, Berechnungen von Daten über phänologische Erscheiningen und dem Auf- und Zugang der Gewässer. Bidrag till kännedom af Finlands natur och folk 72:2.

Johansson, O.V. 1913. Några drag ur meteorologins historia i Finland före 1800. Terra 25: 185-209.

Johansson, O.V. 1932. Isförhållandena vid Uleåborg och I Torne älv. Bidrag till kännedom af Fin-lands natur och folk 84 (3): 3-45.

Johansson, O.V. 1945. Det fenologiska observation materialet i Finland och provstudierna av samma. Bidrag till kännedom af Finlands natur och folk 88 (8): 5-118.

Jones, P.D., Kelly, P.M., 1983. The spatial and temporal characteristics of Northern Hemisphere sur-face air temperature variations. Journal of Climatology 3: 243-252.

Jones, P.D. and Bradley, R.S. 1992. Climatic variations in the longest instrumental records. In: Bradley, R.S. and Jones, P.D. (eds.) Climate since A.D 1500. Routledge, London, pp. 118-142.

Jones, P.D., Briffa, K.R., Barnett, T.P. and Tett, S.F.B. 1998. High-resolution palaeoclimatic records for the last millennium: interpretation, integration and comparison with General Circulation Model control-run temperatures. The Holocene 8: 455-471.

Jones, P.D. and Moberg, A. 2003. Hemispheric and large-scale surface air temperature variations: an extensive revision and an update to 2001. Journal of Climate 16:206-223.

Jones, P.D., Briffa, K.R. and Osborn, T.J. 2003. Changes in the Northern Hemisphere annual cycle: implications for paleoclimatology. Journal of Geophysical Research 108: 4588-4591.

30

Kallinen, M. 1995. Change and stability. Natural philosophy at the Academy of Turku (1640-1713). Studia Historica 51, Hakapaino Oy, Helsinki, 439 p.

Kington, J.A. 1988. The weather of the 1780s over Europe. Cambridge University Press, Cambridge, 166 p.

Klingbjer, P. and Moberg, A. 2003. A composite monthly temperature record from Tornedalen in northern Sweden, 1802-2002. International Journal of Climatology 23:1465-1494.

Lamb, H.H. 1982. Climate, History and the Modern World. Methuen, London, 387 p.

Lappalainen, H. 1994. Examples of plant phenological events in Finland and their relation to temperature. Societas Memoranda Societatis pro Fauna et Flora Fennica 70: 105-121.

Lappalainen, H. and Heikinheimo, M. 1992. Relations between climatological and plant phenological observations. Vol. 1. Survey of plant phenological observations in Finland from 1896 to 1965. Meteorological Publications 20, 74 p.

Leche, I. 1889. Förteckning på tiden, då de Allmænnaste træ och buskar kring Åbo utslagit blad och blommor åhren 1750, 51 och 52, enligen Kongl. Vet. Acad. begæran utrönt af I. Leche. Bidrag till kännedom af Finlands nature och Folk. Finska Vetenskaps-Societen, Fyationdeåttonde Häftet: 471-482.

Leche, J. 1762a. Utdrag af väderleks journalen, som blifvit hållen i Åbo, ifrån och med år 1750, til och med år 1761. Första stycket, om blås-vädren. Kongliga Vetenskaps Academiens Handlingar: 179-192.

Leche, J. 1762b. Utdrag af väderleks journalen, som blifvit hållen i Åbo, ifrån och med år 1750, til och med år 1761. Andra stycket. Kongliga Vetenskaps Academiens Handlingar: 303-313.

Leche, J. 1763a. Utdrag af väderleks journalen, som blifvit hållen i Åbo, ifrån och med år 1750, til och med år 1761. Tredje stycket. Om nederbörds-vattnet. Kongliga Vetenskaps Academiens Handlingar: 15-27.

Leche, J. 1763b. Undervisning om sättet, at förfärdiga barometrar. Kongliga Vetenskaps Academiens Handlingar: 81-99.

Leche, J. 1763c. Utdrag af väderleks journalen, som blifvit hållen i Åbo, ifrån och med år 1750, til och med år 1761. Fjärde stycket, om barometrens förändringar. Kongliga Vetenskaps Academiens Handlingar: 100-107.

Leche, J. 1763d. Utdrag af 12 års thermometer-observationer, gjorda i Åbo. Kongliga Vetenskaps Academiens Handlingar: 177-190.

Leche, J. 1763e. Utdrag af 12 års meteorologiska observationer, gjorda i Åbo: sjätte och sista stycket. Kongliga Vetenskaps Academiens Handlingar: 257-268.

Lindholm, M. 1996. Reconstruction of past climate from ring-width chronologies of Scots pine (Pi-nus sylvestris L.) at the northern forest limit in Fennoscandia. Ph. D. Dissertation, University of Joensuu, Publications in Sciences 40, 169 p.

Lindholm, M., Timonen, M., Meriläinen, J., Vanninen, P., and Eronen, M. 1997. Effects of climate on the growth of Scots pine in the Saimaa lake district, south-eastern Finland, in the southern part of the boreal forest belt. Dendrochronologia 15: 151-168.

Linkola, K. 1924. Kesien aikaisuudesta Lounais-Suomessa vuosina 1750-1923. Talonpoika I, Turku, pp. 41-74.

Linkosalo, T. 2000. Analyses of spring phenology of boreal trees and its response to climate change. Ph. D. Dissertation, University of Helsinki, Department of Forest Ecology, Publications 22, 55 p.

Linkosalo T., Häkkinen, R., and Hari, P. 1996. Improving the reliability of a combined phenological time series by analysing observation quality. Tree Physiology 16:661-664.

31

Manley, G. 1974. Central England temperatures: monthly means 1659 to 1973. Quarterly Journal of the Royal Meteorological Society 100: 389-405.

Mann, M.E. 2002. The value of multiple proxies. Science 297: 1481-1482.

Mann, M.E., Bradley, R.S. and Hughes, M.K. 1998. Global-scale temperature patterns and climate forcing over the past six centuries. Nature 392: 779-787.

Mann, M.E., Bradley, R.S. and Hughes, M.K. 1999. Northern hemisphere temperatures during the past millennium: inferences, uncertainties, and limitations. Geophysical Research Letters 26: 759-762.

Matthews, J.A. and Briffa, K.R. 2005. The ‘Little Ice Age’: re-evaluation of an evolving concept. Geografi ska Annaler 87 A (1): 17-36.

Middleton, W.E.K. 1969. Invention of the meteorological instruments. The John Hopkins Press, Baltimore, 349 p.

Mikola, P. 1962. Temperature and tree growth near the northern timber-line. In: Kozlowski T.T. (ed) Tree growth. The Ronald Press Company, New York, 265-274.

Moberg, A. 1857a. Naturhistoriska dag-anteckningar gjorda i Finland, Åren 1750-1845. Notiser ur Sällskapets pro Fauna & Flora fennica förhandlingar. Bihang till Acta Societatis Scientarium Fennicæ. Trädje Häftet, p. 95-250.

Moberg, A. 1857b. Om de ifrån år 1750 till 1850 in Finland gjorda naturalhistoriska dag-anteckning-ar och deras betydelse I klimatologisk hänseende. Biddrag till Finlands naturkännedom 2.

Moberg, A. 1865. Botaniska bidrag till Finlands klimatologi. Inbjudningsskrift (S.O. Lindbergs pro-fessur installation). Helsingfors.

Moberg, A. 1868. Ornithologiska bidrag till Finlands klimatologi. Inbjudningsskrift (F. W. Mäklins professur installation). Helsingfors.

Moberg, A. 1894. Fenologiska iakttagelser i Finland åren 1750-1845. Biddrag till kännedom of Finlands natur och folk 55, 165 p.

Moberg, A. 1996. Meteorological Observations in Sweden Made Before A.D. 1860. In: Temperature Variations in Sweden since the 18th Century. The Department of Physical Geography, Stockholm University, Dissertation Series No. 5, 98 s.

Moberg, A. and Bergström, H. 1996. Homogenization of Swedish temperature Data. Part III: The long temperature Records from Uppsala and Stockholm. In: Temperature Variations in Sweden since the 18th Century. The Department of Physical Geography, Stockholm University, Disserta-tion Series No. 5, 98 s.

Moberg, A., Sonechkin, D.M., Holmberg, K., Datsenko N.M. and Karlén, W. 2005. Highly variable Northern Hemisphere temperatures reconstructed from low- and high-resolution proxy data. Na-ture 433: 613-617.

Monserud, R.A. 1986. Time series analysis of tree-ring chronologies. Forest Science 32: 349–372.

Niemelä, J. 1998. Vain hyödynkö tähden? Valistuksen ajan hyötyajattelun, luonnontieteen ja talous-politiikan suhde Pehr Adrian Gaddin elämäntyön kautta tarkasteltuna. Suomen Historiallinen Seura. Historiallisia tutkimuksia 199, Tammer-Paino Oy, Tampere, 358 p.

Oerlemans, J.H. 2005. Extracting a climate signal from 169 glacier records. Science 308: 675–677.

Osborn, T.J. and Briffa, K.R. 2006. The spatial extent of 20th-century warmth in the context of the past 1200 years. Science 311: 841-844.

Parker, D.E., Legg, T.P. and Folland, C.K. 1992. A new daily central England temperature series, 1772-1991. Journal of Climatology 12: 317-342.

Pfi ster, C. 1992. Monthly temperature and precipitation in central Europe 1525-1979: quantifying

32

documentary evidence on weather and its effects. In: Bradley and Jones (eds.) Climate since A.D. 1500. Routledge, London, pp. 118-142.

Pollack, H. N. and Smerdon, J. E. 2004. Borehole climate reconstructions: spatial structure and hemispheric averages. Journal of Geophysical Research Vol. 109, D11106, doi:10.1029/2003JD004163.

Seinä, A. and Palosuo, E. 1993. Itämeren suurimpien vuotuisten jääpeitteen laajuuksien luokittelu 1720-1992. Meri 20: 5-20.

Seinä, A. and Palosuo, E. 1996. The Classifi cation of the Maximum Annual Extent of Ice Cover in the Baltic Sea 1720-1995 - Based on the material collected by Rirto Jurva (winters 1720-1940) and the material of the Ice Service of the Finnish Institute of Marine Research (winters 1941-1995). Meri 27: 79-91.

Seinä, A., Gronvall, H., Kalliosaari, S. and Vainio, J. 2001. Ice seasons 1996-2000 in Finnish sea ar-eas. Meri 43: 3-27.

Seppinen, I. 1988. Ilmatieteen laitos 1838-1988. Ilmatieteen laitos, Helsinki, 290 p.

Schaber, J. and Badeck, F-W. 2002. Evaluation of methods for the combination of phenological time series and outlier detection. Tree Physiology 22: 973-982.

Sirén, G. 1961. Skogsgränstallen som indicator for klimatfl uktuationerna i norra fennoskandien under historisk tid. Communicationes Instituti Forestatis Fennie 54: 1-66.

Solantie, R. 1999. Helsingin väestökriisi 1852-56: ilmasto, imperialismi, nälkä ja kolera. Historiallinen Aikakauskirja 97: 37-42.

von Storch, H., Zorita, E., Jones, J.M., Dimitriev, Y., Gonzáles-Rouco, F. and Tett, S.F.B. 2004. Reconstructing past climate from noisy data. Science 306: 679-682.

Timm, O., Ruprecht, E. and Kleppek, S. 2004. Scale-Dependent Reconstruction of the NAO Index. Journal of Climate 17: 2157-2169.

Tornberg, M. 1989. Ilmaston- ja sadonvaihtelut Lounais-Suomessa 1550-luvulta 1860-luvulle. Turun Historiallinen Arkisto 44: 58-87.

Tuomenvirta, H. 2004. Reliable estimation of climatic variations in Finland. Finnish Meteorological Institute Contributions 43, Yliopistopaino, Helsinki, 158 pp.

Turpeinen, O. 1986. Nälkä vai tauti tappoi. Kauhuvuodet 1866-1868. Historiallisia tutkimuksia 136. Jyväskylä, 307 pp.

Vaganov, E.A., Hughes, M.K., Kirdyanov, A.V., Schweingruber, F.H. and Silkin, P.P. 1999. Infl uence of snowfall and melt timing on tree growth in subarctic Eurasia. Nature 400: 149-151.

Van Deusen, P.C. 1990. Popular formulations for modeling tree rings from climate: A unifying approach. Journal of Environmental Quality 20: 823-827.

Vesajoki, H. and Holopainen J. 1995. Keskilämpötilojen vaihtelut Lounais-Suomessa 1700-luvun jälkimmäisellä puoliskolla. Terra 107:3: 136-144.

Vesajoki, H. and Tornberg, M. 1994. Outlining the climate in Finland during the pre-instrumental period on the basis of documentary sources In: Frenzel B. (ed) Climatic trends and anomalies 1675-1715. Paläoklimaforschung – Palaeoclimate Research 13: 51-60.

Vesajoki, H., Tornberg, M. and Kajander, J.M. 1996. Documentary sources for reconstructing past weather and climate in Finland. Ams-Varia 25: 39-42.

Vesajoki, H., Narinen, M. and Holopainen, J. 1995. The early temperature records from Tornio, northern Finland 1737-1749. In: Heikinheimo, P. (ed) International conference of the past, present and future climate. Publications of the Academy of Finland 6, pp.183-186.

Wanner, H., 2005. Call for PAGES initiative on “Past Regional Climate Variability”. PAGES News

33

13 (1): 19-21.Wanner, H. and Luterbacher, J. 2002. The LOTRED approach - a fi rst step towards a “paleoreanaly-

sis” for Europe. PAGES News 10: 9–11.Weber, S.L. 2005. A timescale analysis of the Northern Hemisphere temperature response to volcanic

and solar forcing. Climate of Past 1: 9-17.Wigley, T.M.L., Briffa, K.R. and Jones, P.D. 1984. On the average value of correlated time series,

with applications in dendroclimatology and hydrometeorology. Journal of Climate and Applied Meteorology 23: 201-213.

Yoo, J. and D’Odorico, P. 2002. Trends and fl uctuations in the dates of ice break-up of lakes and rivers in Northern Europe: the effect of the North Atlantic Oscillation. Journal of Hydrology 268: 100-112.

Zorita, E., von Storch, H., Gonzales-Rouco, F.J., Cubasch, U., Luterbacher, J., Legutke, S., Fischer-Bruns, I. and Schlese, U. 2004. Climate evolution in the last fi ve centuries simulated by an at-mosphere-ocean model: global temperatures, the North Atlantic Oscillation and the Late Maun-der Minimum. Meteorologische Zeitschrift 13, 4, 271-289.