2 RESEARCH FRAMEWORK
2.3 Sources
2.3.1 Tree-ring, crop yield and meteorological data
This thesis uses several published tree-ring width (TRW) and maximum density (MXD) series.64 Paleoclimate reconstructions based on these data can explain c. 60%
of the observed summer temperature and 40% of precipitation variance.65 If possible, MXD series instead of TRW are selected, as the TRW parameter is known to con-tain considerable biological memory.66 This might translate into an overestimation of the persistence of climate anomalies. In this thesis, two mean MXD chronologies, one from southern Finland and one from northern Finland, are used as principal source material (Figure 1).67 These data were also used in Article III to reconstruct cli-mate-mediated yield ratios. In addition, tree-ring based precipitation reconstruction from western Europe and temperature reconstructions from central Sweden and Po-lar Siberia were used for comparison in Article V.68 All the reconstructions have been corrected for the biological influence of tree ageing and site-related noise.69
Data on annual provincial rye and barley yield fluctuations in Finland are collect-ed from statistical reports that are available from the 1860s onwards.70 The yield data were used in assessing crop yield responses to climate variability (Article I) and in reconstructing a climate-mediated yield ratio (Article III). For Article I, the yield data were linearly detrended to avoid spurious correlation coefficients that could arise because of the long-term increasing development of the grain yields. For Article III, the missing values for the year 1876 were infilled by adopting an arithmetic mean of the annual yield ratio over the five preceding and five following years of both grains (rye and barley) in each province. The mean yield ratio used in the reconstruction was calculated from the provincial series with a weighted arithmetic mean, where the weights dependend on the share of the grain and the provincial yield of the total harvest.
Monthly mean temperature and precipitation sum measurements from Finland are available from six weather stations (temperature) and two stations (precipitation) from the mid- to late-19th century.71 The data have been corrected for
inhomogeni-64 Helama et al. 2005, 2009, 2014a, 2014b, 2017; Matskovsky & Helama 2014.
65 Helama et al. 2009, 176, 2014b, 272; Matskovsky & Helama 2014, 1480.
66 Frank et al. 2007, 3302.
67 Helama et al. 2014a; Matskovsky & Helama 2014, respectively.
68 Büntgen et al. 2011; Gunnarson et al 2011; Briffa et al. 2013, respectively.
69 Briffa et al. 1992, 114; Helama et al. 2017, 172.
70 Official Statistics of Finland (OSF); Statistical Yearbook of Finland (SYF).
71 Finnish Meteorological Institute.
ties.72 The meteorological data were used in Article I for exploring the yield response to climate variability.
2.3.2 17th century administrative sources
Prior to the keeping of official yield statistics, tithe payments which were dependent on the amount of harvested grain can be used to estimate agricultural output.73 How-ever, two tithe systems, ‘fixed tithes’ and ‘grain tithes’, were in simultaneous use in Finland. Only in the province of Ostrobothnia was the amount of collected tithes based on the quantity of annual harvest throughout the early modern period.74 In ad-dition to indicating year-to-year harvest fluctuations, the Finnish tithes may indicate also storage reserves as the tithes were commonly collected on the following spring.
The tithe data in Article IV was first collected from each parish of the province and adjusted to correspond to the harvest year.75 Next, the parish series were normalised into z-scores. Last, the normalised parish data were averaged into a mean provincial tithe index series and the variance of the mean record was stabilized (see Article IV for details).
To study the possible human consequences of adverse climate and crop failures, the numbers of farms and deserted farms were collected from the same province of Southern Ostrobothnia.76 In the Nordic countries, registers of deserted farms are common source material to assess societal hardship over the early modern period.
If a farmstead failed to pay taxes for three consecutive years, the farm was marked as deserted in the land registers. Also, farms that were abandoned, or whose inhab-itants had perished, were marked as deserted. Thus, the increase in the number of deserted farms can be expected to indicate increased impoverishment, perhaps even elevated mortality.77 The number of farms is given here as the farm taxation unit at the time, as mantal.
In addition to the Crown keeping records of peasants’ tax paying ability, the peas-ant could appeal for a tax reduction from the Crown. In these petitions the peaspeas-ants give a reason for the wished reduction, which was often crop failure or hunger. The catalogue of early modern peasant petitions until 1720 has been published by Ludvig Mårtensson (1952). The 17th century administrative sources were used alongside the tree-ring data to examine climate-driven agricultural crises and their possible socie-tal aftermaths in Article IV.
72 Klingbjer & Moberg 2003; Tuomenvira 2004.
73 See, e.g., Pfister 1978; Kain 1979; Le Roy Ladurie & Goy 1982; Dodds 2004.
74 Muroma 1991, 15; Seppälä 2009, 102.
75 Finnish national Archives (NAF): Bailiff’s Accounts, Tithe Registers and Verification Books 1598–1634;
NAF: Provincial Accounts, Tithe Registers and Verification Books 1635–1704.
76 NAF: Bailiff’s Accounts, Land and Verification Books 1598–1634; NAF: Provincial Accounts, Land and Verification Books 1635–1704.
77 Kujala 2003, 39.
2.3.3 Medieval narrative sources
For the medieval period, north-west Russian chronicles and annals provide the best written source material to study crop failure and hunger in the north-east. Two chron-icle compilations are used here, the 1914 edition of ‘The Chronchron-icle of Novgorod’ and the ‘Utdrag ur ryska annaler’ which is a selected collection of Russian chronicle text focusing on the north-eastern Baltic region. ‘The Chronicle of Novgorod’ is based on the annals of the city of Velikii Novgorod, which were kept year by year from the early twelfth to the fifteenth century.78 The ‘Utdrag ur ryska annaler’ was compiled by Matthias Akiander (1849) from tens of Russian contemporary and non-contemporary chronicles. Akiander paid special attention to the remarks on food crises, diseases, weather, and various natural phenomena when collecting the records from the Rus-sian sources.79 The compilation can be criticised as fragmentary, however, since the records were not collected systemically.80
The chronicle records are commonly of both descriptive and explanatory charac-ter, i.e. the event itself and its cause is stated. Nevertheless, these narrative sources usually lack information on the extent and severity of the events. Annalists and chron-iclers could also exaggerate the horrors of the famine. Alternatively, all crop failure or hunger events might not have been considered worth writing down. Chronicles served contemporary religious and political agendas and do not necessarily provide accurate descriptions of bygone events. Some annals and chronicles were written from first-hand knowledge, whereas others were compiled decades, or even centu-ries, after the events had taken place. As such, sources may include misdatings. Thus, entries of crop failure, food shortage and famine from contemporary sources cannot be taken as comprehensive lists of past events.81 The medieval narrative sources were analysed qualitatively in Articles II and V.
For the various methods used to analyse the statistical data and the administra-tive and narraadministra-tive sources, see the details in the original articles.
78 Guimon 2010, 1158–1159.
79 Akiander 1849, 7.
80 Korpela 2014a, 20–21.
81 Helbling 2007, 431; Rohr 2007a, 223–226, 2007b, 100; Camenisch 2015, 40–49.
0 200 400 600 800 km
Tornio Artic Circle
N
Ladoga Helsinki
Novgorod S
Figure 1. Map of the study area and places mentioned in the text. The shading indicates the approximate study area of the thesis. The triangles indicate the average sampling sites of the southern (S) and the northern (N) MXD chronologies used in this study. The dark grey area indicates area of 17th century Southern Ostrobothnia.