New perspectives on the frequency of crop failures in Finland

Data on annual harvest output is needed to make any assumptions on the frequency of medieval and early modern crop failures and their possible connection to hunger.¹²⁸ Yet, officials in Russia and Finland started to collect annual yield data

127 Jutikkala 1994, 11; Lappalainen 2012, 32.

128 Le Roy Ladurie 1972a, 289–292; Pfister 1978, 224.

systematically from the second half of the 19th century onwards.¹²⁹ From Russia, estimations of medieval and early modern crop yields are only available from some sporadic years and locations.¹³⁰ Similarly in Finland, earlier estimates of crop yields with annual resolution only cover some decades continuously and these are limit-ed to the southernmost part of the country.¹³¹ Thus, evidence of annual grain yield variability needs to be gathered from indirect evidence.

Tithes are commonly used as indirect evidence of agricultural output.¹³² How-ever, as discussed in Article IV, grain tithes, which were based on annual harvest output, were collected only from limited regions and periods in Finland. Moreover, due to several missing values, constructing continuous tithe time-series is impossi-ble. Consequently, crop failure events cannot solely be detected from tithe material.

Alternatively, peasant petitions to the Crown documented early modern crop failure and hunger events in Finland. The peasant could appeal for a tax reduction in the year of crop failure and food scarcity.¹³³ Yet, when assessing the spatio-temporal dis-tribution of the peasant petitions mentioning crop failure or hunger (Appendix 2), it is evident that the material is biased to western Finland and to certain decades. Thus, instead of providing relevant and reliable material on the severity and frequency of past crop failure and hunger events, the petitions might simply indicate the temporal and spatial patterns of petitions as a peasant coping strategy. For example, the fact that peasants in 17th century Eastern Finland did not appeal for tax reduction due to crop failure or hunger after the 1640s (Appendix 2) does not imply that the period following would have been a time of abundant harvests and full bellies. Instead, the majority of lands in this part of the country were donated to the nobility at this time.

This meant that peasants fell under the nobility’s authority and were to pay their tax-es and rents to them instead to the Crown.¹³⁴ Consequently, peasants could not ap-peal to the crown for tax reductions, or at least did not see it as worthwhile to do so.

Thus, a yield ratio series reconstructed from tree rings is the only continuous, and perhaps the most reliable, material to study crop failure history in Finland covering the whole study period.

From the period when quantitative data on annual crop yield fluctuations are available, the 1860s onwards, the reconstruction corresponds to the known coun-try-wide Finnish crop failures. All the crop failure events fall on years when the re-constructed values are below the 10th percentile of the corresponding interval (see, chapter 4.1.2 for the definition of the intervals). Thus, this suggests that the recon-struction may be indicative of large-scale crop failure events. Indeed, all the severe¹³⁵ crop failure events examined in Article IV in 17th century Ostrobothnia fall below the 10th percentile of the corresponding interval (Figure 5a). From these, at least 1601 and 1695 are also known to be severe crop failure years in Russia (Article IV).

129 OSF 1869, 2; Kahan 1968, 353. At the time, Finland was an autonomius part of the Russian Empire

130 Slicher van Bath 1963, 14–15.

131 Tornberg 1989, 60, 66, 78.

132 See, e.g., Pfister 1978; Kain 1979; Le Roy Ladurie & Goy 1982; Dodds 2004.

133 Holm 2005, 387.

134 Kujala 2003, 39.

135 Severe crop failure = traceable impact on the livelihood of the peasants. See Article IV for details.

-4 -3 -2 -1 0 1 2

1300 1350 1400 1450 1500 1550 1600 1650 1700 1750 1800 1850 1900 -4

-3 -2 -1 0 1 2

1300 1350 1400 1450 1500 1550 1600 1650 1700 1750 1800 1850 1900

Shifts in the reconstructed values 10th percentile

Severe crop failures (this study)

Crop failures in Finland (Melander & Melander 1924) a)

b)

Reconstructed yield ratio anomalies (z-scores)

Figure 5. Reconstructed climate-mediated yield ratio anomalies (grey solid line) and the sig-nificant (0.05 level) shifts in the reconstructed values (black dashed line). The grey dashed line indicates the 10th percentile limits of the sub-periods. The dots indicate large-scale crop failures a) analysed in this study over the periods 1300–1500 (Novgorod and Ladoga), 1600–

1699 (Ostrobothnia, Finland) and 1866–1913 (whole Finland) and b) in Finland according to Melander & Melander (1924) over the period 1300–1900.

Source: Melander & Melander 1924; Article I, II, III IV.

However, only half of the documented crop failure events in medieval Novgorod from the 14th century onwards fell under the 10th percentile of the corresponding intervals. Thus, the ability to identify severe crop failures declines with a greater distance, when moving to a slightly different agro-climatic regime. Yet, as discussed in the Results, drought and excessive rainfall also caused yield losses in Novgorod.

Because Novgorod is located further south than Finland, crop yield variability may be explained by more a complex relationship between both precipitation and tem-perature. In fact, the descriptive analysis carried out in Article II found that two of these recorded crop failure events (1366 and 1371), which did not fall under the 10th percentile, resulted from summer drought. Thus, the reconstruction may only be in-dicative of crop failure events in those areas in the north-east where temperature is the principal yield-limiting factor.

Previously, Holopainen and Helama have challenged the relevance of the chronol-ogy of Finnish crop failures published by Melander and Melander in 1924.¹³⁶ Later, Voutilainen qualified the criticism by stating that the chronology does not necessarily i) represent crop failures of equal size in terms of a regional drop in food supply, nor ii) represent solely major crop failures.¹³⁷ Here, the crop yield reconstruction provides additional evidence. Unlike this study, where the severity of the crop failures in the 17th and the latter half of the 19th century were validated either from several independent sources (Article IV) or from quantitative and spatial crop yield data (Article I), the crop failure events presented in the chronology hardly correspond with the recon-struction (Figure 5b). This provides further evidence that the established chronology of Finnish crop failures includes local harvest shortfalls and minor failures.

However, it is important to recall that, at best, the reconstruction can be used as an indirect source to track major crop failures. It does not indicate whether these fail-ures may or may not have any impact on the livelihood of the people. Climate almost never monocausally explains the human consequences of harvest failures. Dynamic, local, and society-specific factors, like availability of alternative food sources, trade networks, food prices, political stability and storage capacity, among other things, have likely contributed to every hunger crisis.¹³⁸

4.2 CAUSALITY BETWEEN CLIMATE AND HUNGER