Lack of documentation on past harvest fluctuations limits the exploration of long-term trends in crop production and agricultural adaptation strategies. A long-term perspective is needed, however, to understand the wide spectrum of potential human responses to environment and climate change. Therefore, we used tree-ring density series as proxy data to reconstruct climate-mediated yield ratio (harvested grain in relation to sown) in central and northern Finland over the period ad 760–2000. The reconstruction explains 50% of the variance in recorded yield ratio variability over the calibration period (ad 1866–1921). The reconstruction illustrated several intervals of increased and reduced yield ratio over the past 13 centuries. The long-term development of the agricultural prerequisites is characterized by distinct intervals defined statistically as ad 760–1106 (highest yield ratios), 1107–1451, 1452–1694, 1695–1911 (lowest yield ratios) and 1912 onwards. The results provide insight into the establishment and development of crop cultivation in the agricultural margin. The reconstruction suggests that continuous crop cultivation was established in the study region during a favourable period of climatic conditions supporting high yields. Thereafter, the climate-mediated yield ratio declined in the long run until the turn of the 20th century. Periods of agricultural transformations, those previously demonstrated in pollen data and historical documents, followed the onsets of the low yield ratio phases indicated by our reconstruction. Thus, we suggest that ever since the establishment of crop cultivation, climate can be considered as an important factor contributing to the development of the agricultural history in the north.
Prior to the industrial period, most societies were largely dependent on local agricultural production. Although crop cultivation has been the main source of livelihood for thousands of years among different societies, high-resolution evidence on past harvest fluctuations is available only from the time documentary records began and have survived. Exploring long-term trends in crop production would, however, provide important insight into the wide spectrum of potential human responses to environment and climate change (Bocinsky and Kohler, 2014; Pfister, 2010). The location of Finland in northernmost Europe overlaps with the northern limits of crop cultivation (Holopainen and Helama, 2009), where the historical crop yields were highly sensitive to temperature fluctuations (Huhtamaa et al., 2015) and where the link between climate and social changes can be particularly strong (Holopainen et al., 2012; Klimenko, 2016; Parry, 1975). In fact, the historical documents from Finland demonstrate that during the coldest years the amount of harvested grain may well have been less than that sown (Holopainen and Helama, 2009). In this paper, our aim is to explore the fluctuations of climate-mediated crop yield ratio in this region (c. north of 62°N) over the past 1200 years.
At the northern margin of crop cultivation, the seasonal course of temperatures and the length of the growing season are the main limiting factors for agriculture (Peltonen-Sainio et al., 2008). In historical Finland, the spring and summer (April–August) temperatures were particularly crucial for the crop yield. Late onset of the growing season and cool summers delayed the ripening of the crops to early autumn, when the risk of crop damage because of night frost increased (Huhtamaa et al., 2015). This effect has been observed with 17th-century phenology and harvest data indicating that early ripening resulted in higher crop yield (Holopainen and Helama, 2009). Most recently, crop losses owed to night frost have occurred less frequently in the second half of the 20th century (Mukula and Rantanen, 1987), synchronous with the warming trend (Mikkonen et al., 2015).
Historical crop cultivation in Finland was limited mainly to two principal grains tolerant of the short growing season, that is, winter rye (Secale cereale L.) and spring-sown barley (Hordeum vulgare L.). The agrarian population aimed to minimize the agricultural risks caused by a cool climate by combining different cultivation methods, like slash-and-burn and arable cultivation, and by balancing the cultivation of the two grains (Solantie, 2012; Taavitsainen et al., 1998). Nevertheless, harvest success followed temperature fluctuations. Years of cool and frosty summers ended in poor harvest and crop failure (Huhtamaa et al., 2015; Jutikkala, 2003).
Although pollen evidence on Finnish cultivation history is widely available (Alenius et al., 2013; Josefsson et al., 2014; Lahtinen and Rowley-Conwy, 2013; Taavitsainen et al., 1998), the annual fluctuations in agricultural success or crop failure events cannot be detected from the pollen analyses. For most parts of Finland, the first written documentation on crop yields is available from the 16th or 17th centuries onwards. Yet these records are commonly indirect sources of such information, like tithe records, from which annual fluctuations and yield quantities are difficult to estimate (Leijonhufvud, 2001; Seppälä, 2009). Other written sources, like registers of sown seed in bailiffs’ accounts, are not feasible for studies on long-term yield variability as these are available only for some sporadic years. Previous estimates of pre-modern Finnish crop yields with an annual resolution cover only some decades continuously and these are limited to south-western Finland (Tornberg, 1989), to areas where the adverse effects of climatic fluctuations probably posed the smallest conceivable risk to crop cultivation in the country (Mukula and Rantanen, 1987). Furthermore, the existing Finnish crop failure chronologies (Manninen, 1860; Melander and Melander, 1924; Myllyntaus, 2009) are spatially biased towards southern and western Finland and misleadingly include Swedish and Russian events. In addition, the accuracy of the earliest crop failure events (14–16th centuries) listed in the chronologies is difficult to prove, as no references to contemporary sources are given.
Country-wide national statistics on crop yield have been recorded continuously since the mid-19th century. Therefore, any reliable information on past crop production is available only for the most recent centuries, although crop cultivation has been the main source of livelihood for a much longer time. A similar lack of written sources limits exploration of the pre-19th-century crop production in many other regions, for example, most parts of North-America (Lemon, 1967).
An alternative approach to unveiling the past crop yield variability involves the exploration of proxy data. It was recently evidenced that tree-ring proxies (maximum latewood density (MXD) in particular) may provide novel material for studies of past crop yield fluctuations when harvest data are not available (Huhtamaa et al., 2015). In that study, it was shown that most of the provincial yield ratio series correlated in a similar way with the temperature and MXD proxies. The common limiting factor, growing season temperature, largely explains the correlation between the MXD proxy and the yield ratio series. Tree-ring records have been successfully used to reconstruct drought-sensitive rain-fed maize yields in Mexico and south-west United States (Bocinsky and Kohler, 2014; Burns, 1983; Therrell et al., 2006) and agricultural production in Florida (Maxwell and Knapp, 2012; Maxwell et al., 2013). Similarly, the reconstruction of agricultural reserves was derived from a tree-ring reconstruction of growing season rainfall in South Carolina, US (Anderson et al., 1995). Up to now, to our knowledge, temperature-sensitive crop yields have not been reconstructed from tree-ring data. Here, we present a climate-mediated yield ratio reconstruction over the period ad 760–2000 from MXD data.
Pollen analyses show that the cultivation of cereals started in Finland at the beginning of the Iron Age, in c. 500 bc (Lahtinen and Rowley-Conwy, 2013) although signs of cultivation can also be found from the Neolithic period onwards (Alenius et al., 2009, 2013; Josefsson et al., 2014). Yet, in different locations of our study area, the earliest evidence of continuous crop cultivation is largely found between the 6th and 10th centuries ad (Alenius and Laakso, 2006; Alenius et al., 2004, 2008, 2009, 2013; Augustsson et al., 2013; Baudou et al., 1991; Ojala, 2001; Taavitsainen et al., 1998; Tiljander, 2005; Wallin and Segerström, 1994). Thus, our new reconstruction effectively covers the era of continuous crop cultivation in the study region. In addition, our reconstruction covers two climatologically interesting periods, the so-called Medieval Climate Anomaly (MCA) and the ‘Little Ice Age’ (LIA). The onset and the termination of these periods vary spatially in the Northern Hemisphere and the dating depends on the climate variable and the season in question. In Finland, a period of summer season warmth, when the long-term average temperatures were close to or above the 1961–1990 mean, spanned from the early-9th to the beginning of the 12th centuries (Helama et al., 2014; Matskovsky and Helama, 2014). Multi-centennial phases of summer drought and mild winters, both also considered typical characteristics of the MCA, prevailed in the studied area c. ad 900–1200 (Haltia-Hovi et al., 2007; Helama et al., 2009a; Tiljander et al., 2003). Somewhat converse of the MCA, cool and wet summers and severe winters have been seen as characteristic of the ‘LIA’ in Finland. Generally, wet summers took place c. 1220–1650 (Helama et al., 2009a) and the cooling trend of winters started in the late 13th century and lasted until the 20th century (Haltia-Hovi et al., 2007). The summer temperatures were variable between the 12th and 15th centuries including warmer and colder decades, but in the latter half of the 15th century, summer temperatures fell significantly and remained below their 1961–1990 mean until the early 20th century (Helama et al., 2014; Matskovsky and Helama, 2014). By combining our results with palaeoecological and historical evidence, we hope to demonstrate the significance of climate variability on the establishment and development of crop cultivation in this region of agricultural margin.
Materials and methods
Crop yield data
It may not be feasible to derive relationships between crop yield and climate-related proxy records using crop statistics from the era of industrial agriculture when aiming to reconstruct crop variability over the historical and even prehistorical era. Such comparisons could be difficult because of several technical and methodological developments which have taken place in agricultural practices during the recent past as well as changes in the genetic plant material (Hietala-Koivu, 2002; Holopainen and Helama, 2009). For example, breeding has brought new crop varieties that are more resistant to weather stress and (weather-related) plant pests and diseases. Thus, industrialization of agriculture has not only increased the yield but may have also altered the crop responses to climate variability. However, changes in agricultural practices and crop species were rather minor throughout the historical period (from around the Middle Ages to the 19th century) compared with the leap from the ‘old’ agriculture of the 19th century to the ‘industrial’ agriculture of today (Rasila, 2004). Moreover, the industrialization of agriculture did not arrive in Finland until the beginning of the 20th century, much later than in western Europe (Haapala, 2009). Thus, we used crop yield statistics from ad 1866 to 1921 to reconstruct past climate-mediated crop yield fluctuations. Annualized data from the northern administrative provinces (Kuopio, Oulu and Vaasa) of 19th-century Finland were used to represent temperature-sensitive crop yield variations. The responses of the two principal grains at the time, winter rye and barley, were investigated.
Altogether, these data included 12 time series comprising sown and harvested grains of rye and barley in the provinces of Kuopio, Oulu and Vaasa (Figure 1). Here, we made use of annual yield ratio, the amount of harvested grain in relation to sown (i.e. the quantity of harvested seed per quantity of sown seed, which is a common way to record harvest over historical times in the Nordic countries). These six time series entered the following analyses.
Area of study. Three historical provinces of Finland: Oulu (O, excluding the areas above the northernmost limit of cereal cultivation), Kuopio (K) and Vaasa (V), and the approximate area of sampling sites of the MXD series (dashed circles).
The values for the year ad 1876 were missing from all of the three provinces. This is probably because of the new regulations concerning the compilation of statistics enacted in January 1877 (Tilastollinen Toimisto, 1884). The missing data were infilled by adopting an arithmetical mean of the annual yield ratio over the five subsequent (1871–1875) and five following (1877–1881) years of the corresponding grain and province.
To calculate a time series of the mean yield ratio for the study region, the annual values of province and grain-specific yield ratios were statistically weighted according to the total amount of grain harvested in each province. First, the proportion () between the amount of harvested grains (rye and barley) from a province () and the sum of harvested grains (rye and barley) in all of the three provinces () was determined as
This proportion was determined separately for the provinces of Kuopio, Oulu, and Vaasa. Second, the proportion () between the harvested grains (rye or barley) in each of the three provinces () and the sum of both harvested grains (rye and barley) () in that province was determined as
This proportion was determined separately for rye and barley in the three provinces. Finally, a single mean yield estimate was calculated for the study region. The weight for each of the six time series () was expressed as
This showed a sum of the six weights needed to calculate a time series of the mean yield estimate for the study region, equal to 1.0 (Table 1). These proportions were first calculated by using the total quantity of harvested grains as documented over the ad 1866–1921 period. In order to investigate whether the relative amounts of different species of harvested grains had changed substantially over the past centuries and whether such changes could influence the calculation of the weighted mean yield ratio estimate, additional estimates of past harvests were collected from tithe records. In historical Finland, a tithe was approximately 5–10% of the harvest that was paid to the Church as a tax. Although estimating the total quantity of crop production from tithe records may not be possible, these records can indicate some general trends in crop production, like the proportion of different grain species (Leijonhufvud, 2001). Tithe records were available for the provinces of Oulu and Vaasa during the 1560s, 1600s and 1680s. No documentary evidence on past harvests was found for the province of Kuopio over these early periods. Experimentally weighted mean yield estimates were calculated over the period ad 1866–1921 using the obtained weight representative of the 1560s, 1600s, 1680s and 1866–1921 harvests (because of the lack of documentary evidence, the proportions for Kuopio province were time-invariant in this calculation) (Table 1). It was found that the four estimates remained virtually inseparable (Figure 2). Therefore, the proportions obtained over the ad 1866–1921 period were adopted and the weighted mean time series, the mean yield estimate, was used in all subsequent analyses as a variable to be explained by the MXD proxy data.
Proportions used to calculate the time series of weighted mean yield ratio. These weights represent the 1560s (Period I), 1600s (Period II), 1680s (Period III) and ad 1866–1921 (Period IV) harvests.
Rye
Barley
Vaasa
Oulu
Kuopio
Vaasa
Oulu
Kuopio
Period I
0.184
0.072
0.168
0.276
0.168
0.132
Period II
0.212
0.079
0.168
0.248
0.161
0.132
Period III
0.239
0.089
0.168
0.221
0.151
0.132
Period IV
0.267
0.074
0.168
0.193
0.166
0.132
Comparison of the four estimates of the weighted mean yield ratio. These estimates were calculated over the period 1866–1921 using the obtained weight representative of the 1560s (Period I), 1600s (Period II), 1680s (Period III) and ad 1866–1921 (Period IV) harvests. The four estimates appear nearly inseparable.
Dendroclimatic data
Tree-ring data were used to reconstruct the crop yield variability. The maximum density (MXD) of the latewood portion of each ring was used as a proxy for past crop variability. Our MXD data originated from subfossil and living Scots pine (Pinus sylvestris L.) tree-ring samples from Lapland and northern Finland (Matskovsky and Helama, 2014) and southern Finland (Helama et al., 2014). The northern part of this dataset includes 430 series from Lapland and northern Finland and covers the period from 216 bc to ad 2010, the southern part of the dataset including 161 series and covering the period between the years ad 673 and 2000. These datasets were previously constructed for mean chronologies and reconstructions of past climate variability in the northern and southern parts of the study region. Non-climatic trends were removed from the individual MXD series using the frequently applied routine of regional curve standardization (RCS; Briffa et al., 1992). Refinements of the original RCS method included the signal-free approach (Melvin et al., 2013), its correction procedure (Matskovsky, 2011) and the use of multiple curves in detrending the data (Melvin et al., 2013) (for details, see Helama et al., 2014; Matskovsky and Helama, 2014). Tree-ring data were subsequently averaged into mean MXD chronologies and the resulting chronologies transformed into proxy estimates. The MXD data of Lapland and northern Finland were averaged into one regional mean (northern) chronology and the MXD data of southern Finland into one regional (southern) chronology. The northern MXD proxy explained more than 60% of the observed summer (June–August) temperature variance (Matskovsky and Helama, 2014), whereas the southern MXD proxy explained up to 60% of the observed growing season (May–September) temperature variance (Helama et al., 2014). These time series of the corresponding MXD proxies were adopted here from the original publications (Helama et al., 2014; Matskovsky and Helama, 2014). The common period of the two MXD records was ad 760–2000. Previous comparisons have demonstrated high correlativity of these tree-ring proxies with the Finnish crop data (Huhtamaa et al., 2015). Here, the records of MXD proxies were combined by normalizing the records into z-scores (over the ad 1866–1921 period) and by averaging the normalized data into a mean MXD record. This time series was used in all subsequent analyses as a variable explaining the variations in past crop variability.
Huhtamaa et al. (2015) found that the time series of yield ratios were not normally distributed. Thus, the mean yield ratio data (1866–1921) was tested for normality using the Shapiro–Wilk test. The test statistic (0.875) showed that the data were not normally distributed (p < 0.001). Also, the test statistic (0.9518) for the data of our MXD proxy showed deviation from normality (p < 0.05) over the common period 1866–1921. Consequently, the records were Box–Cox transformed (Box and Cox, 1964) into normally distributed data for comparisons.
Reconstructing the crop figure
The transformed MXD record was scaled to the mean and variance of the yield ratio record. Calibration of this type was first done over the early period (ad 1866–1893) and quantified using the squared Pearson correlation (r2). The late period (1894–1921) was withheld from the calibration and was used for verification of the reconstruction. Consequently, Pearson correlation was calculated over the late period and the reconstruction was further verified using statistical tests of the reduction of error (RE; Fritts, 1976) and coefficient of efficiency (CE; Briffa et al., 1988). Test statistics of RE and CE above zero were taken to indicate an acceptable level of skill in the reconstructions (Briffa et al., 1988; Fritts, 1976). This early calibration–late verification examination was followed by late calibration–early verification where the transformed MXD record was scaled to the mean and variance of the yield ratio record over the late period 1894–1921 and further verified (using r2, RE and CE) over the early period 1866–1893 withheld from calibration.
Finally, the scaling was done using the full yield ratio data over the common period (ad 1866–1921) and applied over the full MXD period ad 760–2000. The final reconstruction of climate-mediated yield ratio was obtained by back-transforming the reconstructed values. The long-term changes in the final reconstruction were quantified using a method based on a sequential t-test analysis. This method was used to characterize the changes in the yield ratios as differences between mean values of two subsequent regimes (Rodionov, 2004) as recorded in our reconstruction.
Results
Proxy data were successfully used for reconstructing the mean yield ratio variability in the study region over the period ad 1866–1921 (Table 2). Early and late calibrations explained between 40% and 50% of the observed variance in the mean yield ratio. Similar estimates were evident for the early and late verifications. Both of the verification statistics, the RE and CE, did remain positive over the early and late verification periods. We explored whether changing the number of years over which the missing value (ad 1876) was averaged influenced the reconstruction statistics (see Table S1, available online). However, the calibration and verification tests remained highly stable for any corresponding change in the length of the interval used to estimate the missing value. Over the full calibration period, the proxy data explained roughly 50% of the variance in the mean yield ratio. The connection between documented and reconstructed yield ratio variability became less obvious in the latter half of the 1910s (Figure 3), which perhaps can be explained by the increased socio-political instability, which culminated in 1918 with the Finnish Civil War (Arosalo, 1998; Haapala, 1995; Lamberg and Pajunen, 2010). The emerging concerns which contributed to the increasing instability, like the availability and wages of the agricultural workers, landownership and land use rights, probably had more profound impact on the annual yield fluctuations than during the preceding more peaceful decades.
Calibration and verification statistics of the reconstruction of the early (ad 1866–1893) and late (ad 1894–1921) periods.
Calibration period
1866–1893
1894–1921
1866–1921
Verification period
1894–1921
1866–1893
Calibration
Variance explained
0.497
0.427
0.496
Verification
Variance explained
0.427
0.497
Reduction of error
0.134
0.333
Coefficient of efficiency
0.122
0.326
Documented and reconstructed climate-mediated yield ratio. The dashed grey bars indicate the documented crop failures of the study region.
The reconstruction corresponds with the documented crop failure events in northern and central Finland over the calibration period by indicating low yield ratio values in years of crop failure (Figure 3). These events occurred in ad 1867, 1877, 1878, 1881, 1892, 1899 and 1902. We acknowledge that not all crop shortages were caused by climatic downturns. Our study area is located on the European periphery where long winters blocked the trading routes over the Baltic Sea for several months and land roads were few in number and poorly maintained. Grain markets in pre-modern Finland were marginal, especially when compared with western Europe or the Baltics. Therefore, the farmer’s yield was determined not only by climate and weather but also by how much seed grain he had managed to store in the previous years. As only part (if any) of the fields could be sown after a year of crop failure because of empty stores and lack of seed grain, the next harvest was likely poor as well (Jutikkala, 2003). For example, whereas a short and cool growing season and night frosts did cause a crop failure in ad 1877, the failure in the following year was largely owed to the poor quality and quantity of the seed grain (Savonlinna, 1878; Uusi Suometar, 1877). Thus, the crop failure of ad 1878 is not apparent in our reconstruction.
The full reconstruction (Figure 4) indicates the intervals of potentially low and high yield ratios. It is worth noting that the reconstruction does not illustrate the total quantities of the crop production but that the reconstruction has the unit of relative crop yield, that is, the amount of harvested grain in relation to sown. This means that had there been a storage reserve of any importance determining the following year’s harvest success, this effect is not considered in the reconstruction. Moreover, the non-climatic factors leading to gradual growth of the yield ratios in many parts of Europe over the last millennium (Slicher van Bath, 1963), the selection of seed grain, increasing use of fertilization and/or technical advances, inter alia, are not considered in the reconstruction. These limitations notwithstanding, the reconstruction is indicative of yield ratio responses to climate variability, mainly to growing season temperature (Huhtamaa et al., 2015), which was favourable or unfavourable for the crop yields. Five distinct phases of such intervals are discernible throughout the full reconstruction period. These periods are also evident when the reconstruction is analysed using the sequential algorithm (Rodionov, 2004). A period of highest yield ratios was found from ad 760 to 1106. This was followed by a period of lower yield ratios and increased yield variability for 1107–1451. The yield ratios further decreased in 1452–1694. The reconstruction indicates a period of lowest yield ratios for 1695–1911. After that, the yield ratio rose, especially from the 1920s onwards, to values comparable to the period of high yield ratios of ad 760–1106 (Figure 4).
Reconstructed climate-mediated yield ratio anomalies ad 760–2000. The black line shows the significant (0.05 level) shifts in the reconstructed values.
By comparison, the 50-year periods of highest and lowest yield ratios were found for ad 1048–1097 and 1704–1753, respectively. It is notable that the timing of these periods was consistent with the overall long-term development of agricultural success in the region. That is, the period ad 1048–1097 was coeval with the early period of ameliorated yields between ad 760 and 1106, whereas the period ad 1704–1753 overlapped with the periods of deteriorated yields between ad 1695 and 1911.
Discussion
Climate variability, yield ratios and human responses
Throughout the written history of Finland, delayed onset of summer and night frost have been named as the main reasons for crop failure and famine. Supposedly, harvest was poor or totally lost once every second to fourth year (Gadd, 1785; Myllyntaus, 2009). This framework raises a key question: why did people choose to earn their livelihoods from temperature-sensitive crop cultivation in ‘the land of the killer-frost’ (Jutikkala, 2003; Lappalainen, 2012; Solantie, 2006)? Our reconstruction suggests that in the 8th–10th centuries AD, when continuous crop cultivation was established in Finland, the risk of temperature-driven crop failure was notably lower and the crops were generally higher than during the historical period (c. 13th century ad onwards). Climatic conditions during the latter half of the 1st millennium did not pose such a threat to crop cultivation as they were to do during the 2nd millennium AD. We admit, however, that we do not have information on possible historical plant pests and diseases related to these climatic conditions, which may have affected the crop variability in a range of ways the reconstruction does not portray.
The continuous period of high crop yields coincides with an episode of multi-centennial summer season warmth, associated with the MCA in the region and around north-west Europe (Goosse et al., 2012; Luoto and Helama, 2010; Ogilvie et al., 2000; Sundqvist et al., 2010). The warm climatic regime of the MCA was interrupted by a period of distinctly cold winter and summer temperatures c. ad 1110–1150 (Helama et al., 2009b; Linderholm et al., 2015; Tiljander et al., 2003), a period also evident in our reconstruction (Figure 4). Following this phase of sudden drops in temperature and crop yield, pollen evidence suggests a period of agricultural intensification and innovations in the eastern and southern parts of the study area, respectively (Alenius and Laakso, 2006; Alenius et al., 2004, 2008, 2013). Thus, people might have coped with the increased risk of poor harvest by expanding cultivated areas and by introducing new high-yielding methods, like slash-and-burn cultivation. In slash-and-burn agriculture, woods are burnt and the crop is sown in the ashes, whereas in arable cultivation permanent fields are re-sown year after year. Slash-and-burn agriculture provided higher yields than arable cultivation, as the burning of wood and humus released valuable nutrients for the crops. However, the rotation speed was slow (the regrowth of the forest and the renewal of the nutrition levels took several decades) and the method required extensive virgin forest areas as only two to four consecutive harvests could be obtained after the clearance (Soininen, 1975). Therefore, the pre-modern farmers commonly practised arable and slash-and-burn cultivation in parallel, to gain profit from both the land-demanding yet high-yielding slash-and-burn fields and the steady but low-yielding arable fields (Solantie, 2012).
The yield ratios rose in the early 15th century to values comparable to the 8th–10th centuries, but suddenly dropped in the latter half of the century (Figure 4). The rapid mid-15th century cooling, which followed a major atmospheric circulation change over the North Atlantic (Dawson et al., 2007; Meeker and Mayewski, 2002) and coincided with the culmination of the Spörer solar minimum (Miyahara et al., 2006), has been evidenced in various summer and winter season reconstructions of the region (Haltia-Hovi et al., 2007; Helama et al., 2009b; Klimenko and Solomina, 2010; Luoto and Helama, 2010; Zhang et al., 2015). As in the 12th century, a transformation in agricultural practices followed the deteriorating climate and lowering yield ratios of the second half of the 15th century. Pollen evidence indicates a rapid intensification of agricultural activities (Augustsson et al., 2013; Baudou et al., 1991; Ojala, 2001; Taavitsainen et al., 1998; Tiljander, 2005) and changes in cultivation practices (Alenius and Laakso, 2006) in the 16th and 17th centuries AD. The beginning of the 16th century ad was, also according to written sources, marked by a rapid and notable expansion of the cultivated area (Korpela, 2012). During this period, a new rye variety (korpiruis in Finnish) which was suitable for slash-and-burn fields in the coniferous forests was introduced to the area (Korpela, 2012; Taavitsainen et al., 1998), which might partly explain the agricultural expansion despite the unfavourable climate.
The culmination of the ‘LIA’ in Finland has been commonly dated to the late 17th and early 18th centuries ad (Luoto, 2013; Luoto and Helama, 2010; Tiljander et al., 2003), which is synchronous with the onset of the phase of the lowest yield ratios in our reconstruction. The Maunder solar minima (c. 1645–1715) and several volcanic eruptions preceded the culmination (Shindell et al., 2003). Similarly to the earlier phases of lowering yield ratios, the following century was marked by an extensive increase of slash-and-burn cultivation in the eastern parts of the study area (Pitkänen and Huttunen, 1999; Taavitsainen et al., 1998). In addition, based on the tithe and statistical records, the main cultivated crop in the study area (except in the northernmost province) changed from barley to rye during the 18th century ad. These changes coincide with the 50-year period of the lowest ratios (ad 1704–1753) indicated by our reconstruction. Winter rye provided steadier yield and decreased the risk of crop failure owed to early-autumn night frost, as the autumn-sown rye ripened earlier than the spring-sown barley (Solantie, 2012; Tornberg, 1989). Only in the northernmost province of the study area (see Figure 1) was the cultivation of rye hardly worthwhile during this period because of long winters and related wintering damage (Solantie, 2012).
Overall, the periods of agricultural transformation since ad 760, evidenced in the pollen analyses and historical documents, appear to coincide with deteriorating climate and increased hardships in Finnish agriculture as indicated by the onsets of further declining yield ratios suggested by our reconstruction.
Expansion and decline of the crop cultivation in the northern agricultural margins
Whether crop cultivation was introduced in Finland prior to the Iron Age (prior to c. 500 bc) or during the Iron Age is still under debate (Lahtinen and Rowley-Conwy, 2013). Nevertheless, during this early stage of crop cultivation, agriculture served merely as a supplementary livelihood option (Korpela, 2012). Instead of agriculture, the majority of the population in the study area relied on the so-called wilderness economy (erätalous in Finnish), mainly in terms of fishing and hunting. Continuous crop cultivation began in Finland only when the climatic conditions were favourable for a long period of time for high yields, as indicated by the period of highest yield ratios shown by our reconstruction. In the north-easternmost parts of our study region, the crop cultivation never replaced wilderness economy as the main livelihood strategy (Korpela, 2012; Solantie, 2012). We assume that the climatic conditions may not have been sufficiently favourable, possibly not even in the 8–11th centuries AD, for steady high yields at these latitudes. Our reconstruction suggests that the adoption of a new livelihood strategy towards agriculture was perhaps reinforced by the rather hospitable climate and generous crop yields, and even more so in central Finland in the late 1st millennium AD, when the risk of crop failure was much lower than during the next millennium. Nevertheless, the establishment of crop cultivation in the agricultural margin was a gradual and slow process, requiring new cultivation techniques and genetic adaptation of the cultivated varieties (Taavitsainen et al., 1998). In the course of time, crop cultivation became more and more important, and farmers increased the cultivated area and invested in innovations in order to maximize their yields.
Also on the eastern side of the study area, in North-West Russia, c. ad 950–1100 was marked by a warmer climate and intensive agricultural expansion to the north (Klimenko, 2016). In addition, it has been suggested that c. ad 800–1100 North Atlantic temperatures were warm, possibly even comparable to the 20th-century conditions (Ogilvie et al., 2000). Over this time period, Iceland and Greenland were settled and agriculture (mainly animal husbandry but also crop cultivation) was introduced to the islands (Patterson et al., 2010). Thus, evidence of climate-induced expansion of agriculture at the turn of the millennium can be found across northernmost Europe.
The sudden cooling in the second half of the 15th century challenged a society relying on crop cultivation, which by that time had become the primary livelihood strategy in central Finland. It is assumed that the population had two options: either to abandon crop cultivation and return to a wilderness economy or to cope with the adverse conditions. However, the resources available in the forests and lakes might not have been sufficient to feed the whole population. The population in Finland grew steadily throughout the Middle Ages, as the European population crisis of the 14th century did not reach it (Orrman, 2003). In addition, the Swedish crown strengthened its control in the studied area from the 14th century ad (Korpela, 2012). The new authorities set taxes which were mainly to be paid in agricultural products. The crown was not benevolent to hunter-gatherers, as the administrative control (i.e. taxing) of these wandering people was difficult and ineffective. Instead, establishing authority over farmers who cultivated the same lands, year after year, generation after generation, confined to their fields, was far more manageable. Thus, the authorities aimed actively to settle people to practise agriculture during the 15th and 16th centuries ad (Korpela, 2012). Therefore, the administration by the crown may have had a decisive role in maintaining regional agriculture, which possibly explains why the crop cultivation did not end in the study area at the turn of the 15th and 16th centuries ad, despite the deteriorating climatic conditions, which were no longer supportive of high crop yields.
In North-West Russia, temperatures also dropped abruptly from the mid-15th century onwards. The population in the far north responded to the climatic degradation and related yield decline by gradually switching from agriculture to maritime trade (Klimenko, 2016). The biggest northern European city at the time, Veliky Novgorod in North-West Russia, suffered a decade-long period of food scarcity in the mid-15th century ad (Huhtamaa, 2015). The intense colonization of the Russian north stopped in the latter half of the 15th century (Klimenko, 2016). Moreover, many farms were deserted in Sweden c. ad 1460–1500, which probably indicates agricultural crisis (Vahtola, 2003). The human consequences of the mid-15th century climate change in the North Atlantic were even more dramatic, as the Greenland settlements disappeared and the crop cultivation in Iceland was abandoned altogether (Dugmore et al., 2007; Patterson et al., 2010). Thus, the multi-decadal period of low yield ratios in c. ad 1450–1540 suggested by our reconstruction was probably a period of agricultural hardship across northernmost Europe. Interestingly, Finland seems to be a rare place where the climatic deterioration did not result in human crisis in the late 15th century.
Agriculture was introduced and established rather rapidly in Iceland and Greenland when climate supported the agricultural economy. In Finland, the shift from a wilderness economy to crop cultivation was a slow and gradual process that took place over centuries (Rasila, 2004; Taavitsainen et al., 1998). Thus, its society had the generations-long memory necessary to understand and adapt to the local environment. Therefore, it appears that local adaptation strategies and innovations, like slash-and-burn cultivation in Finland, to cope with lowering crop yields, was crucial for the existence of agricultural societies on the northern limits of crop cultivation. In addition, Finland had a low population density and extensive land area. Thus, continuous clearing of new fields and high-yielding cultivation methods that required large land areas, like slash-and-burn cultivation, were possible there (Taavitsainen et al., 1998). For example, between the mid-16th and late 17th centuries, the number of farmsteads remained rather unchanged (c. 30,000 farms) but the average area of cultivated fields per farm doubled (from c. 2.6 to 5.5 ha; Nummela, 2003). Had the crown not supported crop cultivation or had there been no such extensive virgin forest areas, Finland might have faced agricultural and/or demographic crisis like the Norse colonies in the second half the 15th century. Moreover, it has been suggested that instead of the severity of cooling the increased interannual climate variability affected the North Atlantic agriculture most adversely in the 15th century (Dugmore et al., 2012; McGovern et al., 2007). Consequently, the ability to understand and predict the behaviour of environment, climate and future yields decreased (McGovern et al., 2007). Although cooling significantly lowered the yield ratios in Finland in the mid-15th century, the annual variability of the crop yields did not increase (Figure 4). The yield ratios remained steadily low year after year. Thus, perhaps the Finnish farmers could anticipate the future conditions far better than their North Atlantic fellows.
Expanding the area of cultivation was the main response to the deteriorating climate and lowering yield ratios in Finland. This was the only way to maintain a sufficient level of crop production as more efficient methods of crop-raising were unknown in pre-industrial Finland (Soininen, 1975). Also population growth occasionally created an additional pressure to increase the crop production over the ‘LIA’. Yet clearing of new fields was determined by the current land use policies and the availability of virgin land areas suitable for cultivation. Moreover, crop cultivation remained sensitive to climate fluctuations although the field area (and thus the quantity of crop production) may have increased. Consequently, during the period of lowest yield ratios indicated by our reconstruction (ad 1695–1911), devastating famines broke out in Finland in 1696 and 1867, when 25–30% and 8% of the population, respectively, died from malnutrition and related diseases. Both of these famines were triggered by large-scale crop failures caused by low temperature extremes (Jantunen and Ruosteenoja, 2000; Neumann and Lindgrén, 1979).
Conclusion
Tree-ring density (MXD) data were used to reconstruct climate-mediated yield ratio variations in central and northern Finland over the period ad 760–2000. Our reconstruction suggests that permanent agriculture was not introduced to ‘the land of the killer-frost’. Instead, crop cultivation was established in a land where climate supported high yields. Thereafter, historical agriculture was able to adapt to the changing environmental conditions. Periods of agricultural transformations, which are evident in the pollen analyses and historical documents, coincide with long-term periods of altered yield ratios suggested by our reconstruction. The agricultural population responded to the changing climatic conditions, which lowered the crop yields, by increasing the area of cultivation and adopting high-yielding crop varieties and cultivation methods. However, agriculture may not have remained an important source of subsistence during the coolest periods of the ‘LIA’ without the control exercised by the Swedish crown that aimed to persuade people to practise agriculture.
The reconstruction presented here provides novel material for understanding the long-term trends in crop production and the expansion and establishment of permanent crop cultivation in central and northern Finland. We anticipate that a similar approach, where temperature-sensitive records such as MXD data are used to reconstruct past crop fluctuations, may help to detail and enhance the understanding of agricultural history in other areas, particularly in areas where crop cultivation has been practised under strong climatic influence, but where long-term documentation of historical harvests is not available. Similar to the study region (northern Europe), such conditions have been evidenced to prevail in the corresponding areas of the New World. Were such reconstructions to be accomplished, valuable information for historical research on, for example, pre-colonial and colonial northern North-America could well be gained.
Footnotes
Acknowledgements
We wish to kindly thank Jari Holopainen, Jukka Korpela, Christian Rohr, Philip Slavin and the two anonymous referees for their comments and suggestions on the manuscript.
Funding
This study was partially supported by the Academy of Finland (funding decision no. 288267).
References
1.
AleniusTLaaksoV (2006) Palaeoecology and archaeology of the village of Uukuniemi, Eastern Finland. Acta Borealia23(2): 145–165.
2.
AleniusTGrönlundESimolaH. (2004) Land-use history of Riekkalansaari Island in the northern archipelago of Lake Ladoga, Karelian Republic, Russia. Vegetation History and Archaeobotany13(1): 23–31.
3.
AleniusTLaventoMSaarnistoM (2009) Pollen-analytical results from Lake Katajajärvi – Aspects of the history of settlement in the Finnish inland regions. Acta Borealia26(2): 136–155.
4.
AleniusTMikkolaEOjalaAE (2008) History of agriculture in Mikkeli Orijärvi, eastern Finland as reflected by palynological and archaeological data. Vegetation History and Archaeobotany17(2): 171–183.
5.
AleniusTMökkönenTLahelmaA (2013) Early farming in the northern boreal zone: Reassessing the history of land use in southeastern Finland through high-resolution pollen analysis. Geoarchaeology28(1): 1–24.
6.
AndersonDGStahleDWCleavelandMK (1995) Paleoclimate and the potential food reserves of Mississippian societies: A case study from the Savannah River Valley. American Antiquity60(2): 258–296.
7.
ArosaloS (1998) Social conditions for political violence: Red and white terror in the Finnish Civil War 1918. Journal of Peace Research35(2): 147–166.
8.
AugustssonAGaillardMJPeltolaP. (2013) Effects of land use and climate change on erosion intensity and sediment geochemistry at Lake Lehmilampi, Finland. The Holocene23(9): 1247–1259.
9.
BaudouEEngelmarkRLiedgrenL. (1991) Järnåldersbygd i Österbotten: en ekologisk-arkeologisk studie av bosättningskontinuitet och resursutnyttjande. Vaasa: Scriptum.
10.
BocinskyRKKohlerTA (2014) A 2,000-year reconstruction of the rain-fed maize agricultural niche in the US Southwest. Nature Communications5: 5618.
11.
BoxGECoxDR (1964) An analysis of transformations. Journal of the Royal Statistical Society: Series B: 211–252.
12.
BriffaKRJonesPDBartholinTS. (1992) Fennoscandian summers from AD 500: Temperature changes on short and long timescales. Climate Dynamics7(3): 111–119.
13.
BriffaKRJonesPDPilcherJR. (1988) Reconstructing summer temperatures in northern Fennoscandinavia back to AD 1700 using tree-ring data from Scots pine. Arctic and Alpine Research20: 385–394.
14.
BurnsBT (1983) Simulated Anasazi Behaviour Using Crop Yields Reconstructed from Tree Rings: AD 652–1968. Tucson: University of Arizona.
15.
DawsonAGHickeyKMayewskiPA. (2007) Greenland (GISP2) ice core and historical indicators of complex North Atlantic climate changes during the fourteenth century. The Holocene17(4): 427–434.
16.
DugmoreAJKellerCMcGovernTH (2007) Norse Greenland settlement: Reflections on climate change, trade, and the contrasting fates of human settlements in the North Atlantic islands. Arctic Anthropology44(1): 12–36.
17.
DugmoreAJMcGovernTHVésteinssonO. (2012) Cultural adaptation, compounding vulnerabilities and conjunctures in Norse Greenland. Proceedings of the National Academy of Sciences109(10): 3658–3663.
18.
FrittsHC (1976) Tree Rings and Climate. London: Academic Press.
19.
GaddPA (1785) Afhandling, om medel til allmogens bärgning under säd- och foderbrist; jämte upgift af enskilta och allmänna anstalter, som därtil fordras. Turku: Tryckt hos K. Acad. boktr. J. C. Frenckells enka.
20.
GoosseHGuiotJMannME. (2012) The medieval climate anomaly in Europe: Comparison of the summer and annual mean signals in two reconstructions and in simulations with data assimilation. Global and Planetary Change84: 35–47.
21.
HaapalaP (1995) Kun yhteiskunta hajosi: Suomi 1914–1920. Helsinki: Painatuskeskus.
22.
HaapalaP (2009) Modernisation of Finland 1800–2000. In: PerkiöM (ed.) Perspectives to Global Social Development. Tampere: Tampere University Press, pp. 48–66.
23.
Haltia-HoviESaarinenTKukkonenM (2007) A 2000-year record of solar forcing on varved lake sediment in eastern Finland. Quaternary Science Reviews26(5): 678–689.
24.
HelamaSMeriläinenJTuomenvirtaH (2009a) Multicentennial megadrought in northern Europe coincided with a global El Niño–Southern Oscillation drought pattern during the Medieval Climate Anomaly. Geology37(2): 175–178.
25.
HelamaSTimonenMHolopainenJ. (2009b) Summer temperature variations in Lapland during the Medieval Warm Period and the Little Ice Age relative to natural instability of thermohaline circulation on multi-decadal and multi-centennial scales. Journal of Quaternary Science24(5): 450–456.
26.
HelamaSVartiainenMHolopainenJ. (2014) A palaeotemperature record for the Finnish Lakeland based on microdensitometric variations in tree rings. Geochronometria41(3): 265–277.
27.
Hietala-KoivuR (2002) Landscape and modernizing agriculture: A case study of three areas in Finland in 1954–1998. Agriculture, Ecosystems & Environment91(1): 273–281.
28.
HolopainenJHelamaS (2009) Little Ice Age farming in Finland: Preindustrial agriculture on the edge of the Grim Reaper’s scythe. Human Ecology37(2): 213–225.
29.
HolopainenJRickardIJHelamaS (2012) Climatic signatures in crops and grain prices in 19th-century Sweden. The Holocene22(8): 939–945.
30.
HuhtamaaH (2015) Climatic anomalies, food systems and subsistence crises in medieval Novgorod and Ladoga. Scandinavian Journal of History40(4): 572–590.
31.
HuhtamaaHHelamaSHolopainenJ. (2015) Crop yield responses to temperature fluctuations in 19th century Finland: Provincial variation in relation to climate and tree-rings. Boreal Environment Research20: 707–723.
32.
JantunenJRuosteenojaK (2000) Weather conditions in northern Europe in the exceptionally cold spring season of the famine year 1867. Geophysica36(1–2): 69–84.
33.
JosefssonTRamqvistPHHörnbergG (2014) The history of early cereal cultivation in northernmost Fennoscandia as indicated by palynological research. Vegetation History and Archaeobotany23(6): 821–840.
34.
JutikkalaE (2003) Halla aina uhkana. In: RasilaVJutikkalaEMäkelä-AlitaloA (eds) Suomen maatalouden historia. Osa 1: Perinteisen maatalouden aika: esihistoriasta 1870-luvulle. Helsinki: Suomalaisen Kirjallisuuden Seura, pp. 292–299.
35.
KlimenkoV (2016) Thousand-year history of northeastern Europe exploration in the context of climatic change: Medieval to early modern times. The Holocene26(3): 365–379.
36.
KlimenkoVSolominaO (2010) Climatic variations in the East European Plain during the last millennium: State of the art. In: PrzybylakRMajorowiczJBrázdilR. (eds) The Polish Climate in the European Context: An Historical Overview. Dordrecht: Springer, pp. 71–101.
37.
KorpelaJ (2012) Sisä-Suomen asuttaminen ja väestön kasvu myöhäiskeskiajalla ja uuden ajan alussa. Historiallinen Aikakauskirja110(3): 275–290.
38.
LahtinenMRowley-ConwyP (2013) Early farming in Finland: Was there cultivation before the Iron Age (500 BC)?European Journal of Archaeology16(4): 660–684.
39.
LambergJAPajunenK (2010) Agency, institutional change, and continuity: The case of the Finnish Civil War. Journal of Management Studies47(5): 814–836.
40.
LappalainenM (2012) Jumalan vihan ruoska: Suuri nälänhätä Suomessa 1695–1697. Helsinki: Siltala.
41.
LeijonhufvudL (2001) Grain Tithes and Manorial Yields in Early Modern Sweden: Trends and Patterns of Production and Productivity c. 1540–1680. Uppsala: Swedish University of Agricultural Sciences.
42.
LemonJT (1967) Household consumption in eighteenth-century America and its relationship to production and trade: The situation among farmers in southeastern Pennsylvania. Agricultural History41(1): 59–70.
43.
LinderholmHWBjörklundJSeftigenK. (2015) Fennoscandia revisited: A spatially improved tree-ring reconstruction of summer temperatures for the last 900 years. Climate Dynamics45(3–4): 933–947.
44.
LuotoTP (2013) How cold was the Little Ice Age? A proxy-based reconstruction from Finland applying modern analogues of fossil midge assemblages. Environmental Earth Sciences68(5): 1321–1329.
45.
LuotoTPHelamaS (2010) Palaeoclimatological and palaeolimnological records from fossil midges and tree-rings: The role of the North Atlantic Oscillation in eastern Finland through the medieval climate anomaly and Little Ice Age. Quaternary Science Reviews29(17–18): 2411–2423.
46.
McGovernTHVésteinssonOFridrikssonA. (2007) Landscapes of settlement in northern Iceland: Historical ecology of human impact and climate fluctuation on the millennial scale. American Anthropologist109: 27–51.
47.
ManninenA (1860) Mietteitä katovuosista Suomessa. Kuopio: Aschan.
48.
MatskovskyVV (2011) Estimation of biases in RCS chronologies of tree rings. Journal of Siberian Federal University4(4): 389–404.
49.
MatskovskyVVHelamaS (2014) Testing long-term summer temperature reconstruction based on maximum density chronologies obtained by reanalysis of tree-ring data sets from northernmost Sweden and Finland. Climate of the Past10(4): 1473–1487.
50.
MaxwellJTKnappPA (2012) Reconstructed tupelo-honey yield in northwest Florida inferred from Nyssa ogeche tree-ring data: 1850–2009. Agriculture, Ecosystems & Environment149: 100–108.
51.
MaxwellJTKnappPAOrtegrenJT (2013) Influence of the Atlantic multidecadal oscillation on tupelo honey production from AD 1800 to 2010. Agricultural and Forest Meteorology174, 129–134.
52.
MeekerLDMayewskiPA (2002) A 1400-year high-resolution record of atmospheric circulation over the North Atlantic and Asia. The Holocene12(3): 257–266.
53.
MelanderKRMelanderG (1924) Katovuosista suomessa. In: PalménEG (ed.) Oma maa V. Porvoo: WSOY, pp. 350–359.
54.
MelvinTMGruddHBriffaKR (2013) Potential bias in ‘updating’ tree-ring chronologies using regional curve standardisation: Re-processing 1500 years of Torneträsk density and ring-width data. The Holocene23(3): 364–373.
55.
MikkonenSLaineMMäkeläHM. (2015) Trends in the average temperature in Finland, 1847–2013. Stochastic Environmental Research and Risk Assessment29(6): 1521–1529.
56.
MiyaharaHMasudaKMurakiY. (2006) Variation of solar cyclicity during the Spoerer Minimum. Journal of Geophysical Research: Space Physics111: A03103.
57.
MukulaJRantanenO (1987) Climatic risks to the yield and quality of field crops in Finland: I. Basic facts about Finnish field crops production. Annales Agriculturae Fenniae26(1): 1–18.
58.
MyllyntausT (2009) Summer frost. A natural hazard with fatal consequences in pre-industrial Finland. In: MauchCPfisterC (eds) Natural Disasters and Cultural Responses: Case Studies toward a Global Environmental History. Lanham, MD: Lexington Books, pp. 77–102.
59.
NeumannJLindgrénS (1979) Great historical events that were significantly affected by the weather: 4, The great famines in Finland and Estonia, 1695–97. Bulletin American Meteorological Society60(7): 775–787.
60.
NummelaI (2003) Asutus, pelto ja karja. In: RasilaVJutikkalaEMäkelä-AlitaloA (eds) Suomen maatalouden historia. Osa 1: Perinteisen maatalouden aika: esihistoriasta 1870-luvulle. Helsinki: Suomalaisen Kirjallisuuden Seura, pp. 133–156.
61.
OgilvieAEBarlowLKJenningsAE (2000) North Atlantic climate c. AD 1000: Millennial reflections on the Viking discoveries of Iceland, Greenland and North America. Weather55(2): 34–45.
62.
OjalaA (2001) Varved Lake Sediments in Southern and Central Finland: Long Varve Chronologies as a Basis for Holocene Palaeoenvironmental Reconstructions. Espoo: Geological Survey of Finland.
63.
OrrmanE (2003) Suomen keskiajan asutus. In: RasilaVJutikkalaEMäkelä-AlitaloA (eds) Suomen maatalouden historia. Osa 1: Perinteisen maatalouden aika: esihistoriasta 1870-luvulle. Helsinki: Suomalaisen Kirjallisuuden Seura, pp. 67–86.
64.
ParryML (1975) Secular climatic change and marginal agriculture. Transactions of the Institute of British Geographers64: 1–13.
65.
PattersonWPDietrichKAHolmdenC. (2010) Two millennia of North Atlantic seasonality and implications for Norse colonies. Proceedings of the National Academy of Sciences107(12): 5306–5310.
66.
Peltonen-SainioPJauhiainenLHakalaK (2008) Climate change and prolongation of growing season: Changes in regional potential for field crop production in Finland. Agricultural and Food Science18(3–4): 171–190.
67.
PfisterC (2010) The vulnerability of past societies to climatic variation: A new focus for historical climatology in the twenty-first century. Climatic Change100(1): 25–31.
68.
PitkänenAHuttunenP (1999) A 1300-year forest-fire history at a site in eastern Finland based on charcoal and pollen records in laminated lake sediment. The Holocene9(3): 311–320.
69.
RasilaV (2004) Overview of the history of Finnish agriculture: From prehistory to the 21st century. In: MarkkolaP (ed.) Suomen maatalouden historia. Osa 3: Suurten muutosten aika: Jälleenrakennuskaudesta EU-Suomeen. Helsinki: Suomalaisen Kirjallisuuden Seura, pp. 490–507.
70.
RodionovSN (2004) A sequential algorithm for testing climate regime shifts. Geophysical Research Letters31: GL019448.
SeppäläS (2009) Viljana, nahkoina, kapakalana: Talonpoikien maksamat kruununverot Suomessa vuosina 1539–1609 [Grain, Skins, Dried Fish: Crown Taxes Paid by Peasants in Finland during the Years 1539–1609]. Helsinki: Suomalaisen Kirjallisuuden Seura (in Finnish with English abstract).
73.
ShindellDTSchmidtGAMillerRL. (2003) Volcanic and solar forcing of climate change during the preindustrial era. Journal of Climate16(24): 4094–4107.
SoininenAM (1975) Vanha maataloutemme. Helsinki: Suomen Historiallinen Seura.
76.
SolantieR (2006) Maataloutta pakkaspirun maassa. Historiallinen Aikakauskirja104(2): 185–190.
77.
SolantieR (2012) Ilmasto ja sen määräämät luonnonolot Suomen asutuksen ja maatalouden historiassa [The Role of the Climate and Related Nature Conditions in the History of the Finnish Settlement and Agriculture]. Jyväskylä: University of Jyväskylä (in Finnish with English abstract).
78.
SundqvistHSHolmgrenKMobergA. (2010) Stable isotopes in a stalagmite from NW Sweden document environmental changes over the past 4000 years. Boreas39(1): 77–86.
79.
TaavitsainenJPSimolaHGrönlundE (1998) Cultivation history beyond the periphery: Early agriculture in the North European boreal forest. Journal of World Prehistory12(2): 199–253.
80.
TherrellMDStahleDWDiazJV. (2006) Tree-ring reconstructed maize yield in central Mexico: 1474–2001. Climatic Change74(4): 493–504.
TiljanderM (2005) Holocene Sedimentary History of Annual Laminations of Lake Korttajärvi, Central Finland. Helsinki: University of Helsinki.
83.
TiljanderMSaarnistoMOjalaAEK. (2003) A 3000-year palaeoenvironmental record from annually laminated sediment of Lake Korttajärvi, central Finland. Boreas32(4): 566–577.
84.
TornbergM (1989) Ilmaston- ja sadonvaihtelut Lounais-Suomessa 1550-luvulta 1860-luvulle [Fluctuations in climate and harvest in South West Finland from the 1550’s to the 1860’s]. Turun Historiallinen Arkisto44: 58–87 (in Finnish with English abstract).
VahtolaJ (2003) Population and settlement. In: HelleK (ed.) The Cambridge History of Scandinavia Volume 1: Prehistory to 1520. Cambridge: Cambridge University Press, pp. 559–580.
87.
WallinJESegerströmU (1994) Natural resources and agriculture during the Iron Age in Ostrobothnia, western Finland, investigated by pollen analysis. Vegetation History and Archaeobotany3(2): 89–105.
88.
ZhangPLinderholmHWGunnarsonBE. (2015) 1200 years of warm-season temperature variability in central Fennoscandia inferred from tree-ring density. Climate of the Past Discussions11(1): 489–519.
