An experimental 392-year documentary-based multi-proxy (vine and grain) reconstruction of May-July temperatures for Kőszeg, West-Hungary

Abstract

In this paper, we present a 392-year-long preliminary temperature reconstruction for western Hungary. The reconstructed series is based on five vine- and grain-related historical phenological series from the town of Kőszeg. We apply dendrochronological methods for both signal assessment of the phenological series and the resultant temperature reconstruction. As a proof of concept, the present reconstruction explains 57% of the temperature variance of May–July Budapest mean temperatures and is well verified with coefficient of efficiency values in excess of 0.45. The developed temperature reconstruction portrays warm conditions during the late seventeenth and early eighteenth centuries with a period of cooling until the coldest reconstructed period centred around 1815, which was followed by a period of warming until the 1860s. The phenological evidence analysed here represent an important data source from which non-biased estimates of past climate can be derived that may provide information at all possible time-scales.

Appendix: background information

Main characteristics of the phenology data series

Information, collected mainly from the Town Council Protocols of Kőszeg, on the beginning of vine harvest and election day of vine shepherds, and the beginning of grape ripening, is available sporadically from the late sixteenth century, and more continuously from the late 1610s onwards. Dates of collecting and summarising wine and grain tithes were derived from tithe accounts which were similarly kept in the town archives of Kőszeg. These data are sporadically available in documented form from the late sixteeenth century, and, due to the fact that Catholic church gave up its feudal rights and thus, this form of taxation ceased to exist in Hungary from 1848 onwards, are more continous from the late 1610s until 1847. In the case of all four series, each value was calculated from Calendar dates. Quality indices (1–5) were extracted from the descriptive data collected in the Kőszeg Book of Vinesprouts (n.d.) from 1740 onwards, although sporadic information is available for more than a century before.

Beginning of grape ripening: nomination and oath of vine shepherds

The so-called vine shepherds, nominated by the actual hill-masters, were elected in each year around the day of Saint Laurence (10 August) by the town council. Moreover, Saint Laurence’ day was one of the dates when prediction/forecast was made about the wine quality and quantity of the forthcoming harvest (Bél 1984). The election or nomination of vine shepherds was mentioned in the town council protocols sporadically from the 1580s and more continuously from the late 1610s onwards with only few years missing. Vine shepherds were chosen and paid only from the days when soft grapes were already found and reported back to the town council (by hill-masters) in the vineyards. While this is essentially true before the late 1680s and after the 1780s, the date of shepherd-nominations (starting day of their contract) was predominantly fixed to 10 August regardless of the maturity of grapes between the late 1680s and late 1780s, even if short reports on stage of ripening were provided instead. Although, similarly to other series, other circumstances, such as disease, rainy weather, human factors etc. sometimes could also have some effect on this date (e.g. 1675: election on time, but very late grain and vine harvests), it is still one of the best indicators of actual early summer temperature conditions (see “Results and discussion”).

The master series: beginning of grapevine harvest

As indicated in the Introduction, the beginning of grape harvesting is one of the most common biophysical indicators applied in phenology-based temperature reconstructions. Similar to the election of vine shepherds, the beginning of the vine harvest was discussed in the town council usually at least some days before the event, but the final decision was made by the mayor and a broader council of 24 elected men (see KTCP, Szövényi 1966-1970). Except for a few years, dates of vine harvest are almost continuously available in the town council protocols from the late 1610s onwards. Possible reasons for the gaps are related to (1) the document (actual entire volume) being missing (e.g. KTCP 1693) or (2) extreme years when ripening was either very early (e.g. KTCP 1728, 1729), late (KTCP 1715, 1721) or did not occur at all due to poor weather conditions (i.e. very cold/wet conditions). Although vine harvest date series show a strong correlation with May–July temperature (Fig. 3), for single years, the beginning of grape harvest can in some cases also be influenced by other factors. The influences of these factors are usually minimal however (see Appendix Table 3, and details below). Due to the detailed descriptions of the decision-making process concerning the starting date of annual harvests, in many cases, it is possible to provide information about the influencing factors other than May–July temperature, and therefore provide additional uncertainties for certain years (see Appendix Table 3).

Table 3 Factors affecting vine harvest dates reported

Among other influencing natural factors, September–October weather conditions (temperature, precipitation, frost and hail) might have some additional impact in determining the first day of harvest. However these impacts usually only affect the starting dates by 2–3 days (Appendix Table 3). Nevertheless, even in cases of unfavourable (social, economic) conditions, the council did not necessarily provide permission for an earlier harvest: good/normal quality received priority over quantity (e.g. KTCP 1722, 1725). Other natural reasons, only in exceptional cases, had an impact which resulted an alteration of the ’real’ starting date by more than 2–3 days. Among human and economic influencing factors, the delay of harvest because of sunny autumn weather, for gaining higher quality but risking actual quantity, was probably important due to the fact that the town primarily lived on wine production and wine export and thus, quality was at least as important as quantity.

Moreover, no harvest started on the day (and preceding 1–2 days) of the country Fair of Kőszeg (Saint Orsolya’s day), namely on 21 October or 1–2 days before (see, e.g., Szövényi 1966–1970). This rule lost its importance after the 1870s (e.g. vintage on 20–21 October: 1873, 1877, 1879). In general, society proved strongly flexible. A good example of flexible social reaction is that when the harvest was in danger (e.g. bad weather, birds, wasps, etc.), the church allowed Sunday/feast work in order to avoid any delay. Even in the case of the most severe non-climatic situations, for example war and plague, the town council seemingly managed to find the way (paying soldiers for protection, hiring people from the countryside for work) to ensure the correct timing of the harvest with respect to the ripening of the vines (KTCP: e.g. 1709, 1710). Only one case is known from the town council protocols (KTCP: 1705) when, besides early ripening, the beginning of harvesting was allowed earlier (only in really exceptional cases), clearly to stop a sudden decrease of harvested grapes, caused by war and surrounding armies.

Consequently, although human factors should also be considered, ripening and weather conditions dominated the decision of the official beginning of the vine harvest (also emphasised in the town protocols: KTCP: 25 June 1672). As such, in general the reasons of changing or postponing the harvest date were mainly related to natural factors—primarily to weather and partly to other natural conditions. Even if human impact played some role (e.g. country fair, feast days etc), its rare influence did not affect the harvesting date by more than a couple of days, and thus, compared to natural factors, its significance was somewhat marginal. These conditions were demonstrated by the instruction of the town council, as a strong tradition which was also generalised and fixed for the future (KTCP: 27 Nov. 1716): ’As far as vine is concerned, similarly to the practice in the neighbourhood, we always have to keep ourselves to the weather conditions.’

Wine quality

Since in our study period the town primarily lived on the income derived from wine export towards such areas as Silesia, Prussia, Saxony and Austria (e.g. Szövényi 1965, 1966–1970), wine quality had special importance, and thus it was documented several times in the town council protocols and by contemporary authors such as Matthias Bél from the late sixteenth century onwards. However, a systematic set of data is available only from 1740 in the Book of Vinesprouts. In this book, apart from other phenological parameters, information is provided for each year, describing how good or bad (sweet or sour) the vine/wine was in the actual year. Based on this descriptive qualitative information, we used the following 5-point index system: very bad/sour (1), worse than average, not good, bad (2), average (3), good, better than average (4) and very good/sweet/strong (5). Although other scaling systems (e.g. 3-point: Pfister 1984; 4-point: Brázdil et al. 2008; 7-point: Strömmer 2003) are also utilised in the scientific literature, in our case the (higher) number of years with very bad wine, and the very exceptional cases of years with extreme qualities in the descriptions, did not provide enough reason for applying scaling methods other than the 5-point one. In the case of vine quality, apart from May–July, the importance of August–October temperatures is much more pronounced than that of vine harvest date series (also emphasised by Péczely 1982).

Although the same information is provided concerning the harvested amount in each year, similarly to quantitative tithe information, this type of information had to be excluded from the present reconstruction due to low correlation values with measured temperature.

Dates of collecting vine and grain tithes

Town citizens paid their wine and grain taxes to the bishop of Győr (see, e.g., Bél 1984), and the town authority was responsible to collect it each year until 1848. Information on the circumstances and the way of tithe collection procedures and their documentation are quite well-known due to the fact that, especially in 1672, very detailed descriptions of the precisely defined, long-term (’ex antiquo usu’) unchanged and ’obligatory’ tradition of wine and grain tithe collection was described to the Royal Chamber. These detailed descriptions and instructions were, for example, included in the town council protocols (25 June and 14 July 1672). Grain tithes were collected in the field, while vine tithes were taken after people pressed out the juice. Moreover, almost every year, mainly again in the town council protocols and partly in the town master’s accounts, information was provided about circumstances of tax-collection, lunch payments of tax-collectors, some tithe-auctions, etc.

In general, this type of information is rather similar to dates of grain tithe auctions, first described in Switzerland (Pfister 1979). Even if collection dates of vine and grain tithes are mainly dependent on ripening and harvest dates, other parameters such as the end of grain harvest or vintage together with the end of pressing out of the juice also had to be accounted for. Autumn weather conditions unfavourable for transportation (rain, bad roads, etc.) and sometimes other uncertainties (e.g. disease, feasts, security questions) also played some role in defining the date for conscription of individual taxes. Nevertheless, due to the fact that the presently available two series are based on the actual dates of defining annual tax which was carried out in the town, transportation and other problems will have caused less significant delays.

Homogeneity issues: vine and grain types and management throughout the study period

Predominant vine types and conservative management practices: the study period and beyond

On the basis of the data available in the 1451-urbarium of Kőszeg (G-Güns), in the mid-fifteenth century the area covered by the vineyards was estimated to be ca. 390 ha, in 1552 ca. 357 ha (Bariska 1998), around 354 ha (1738) and then 372 ha in 1745 (Promontorial conscriptions 1738; Szövényi 1966–1970; Bariska 2001). According to the cadaster survey carried out in 1845, the extent of vineyards was still 356 ha. At the same time, the extent of arable lands was 631 ha (Fényes 1851). The area of vineyards was still similar (348 ha) in the mid-1870s (Chernel 1877), when a rapid reduction started, continuing into the twentieth century (Bariska 2001). Thus, the extension (and location) of vineyards did not change significantly throughout the study period. Wine-related information always played an important role in administrative documentation, especially because the town itself also possessed a large number of vineyards in various locations. The vineyards, owned by the town, were usually the places where the vine harvest was initially started by the hillmaster and vine shepherds, while ’common’ vine harvests usually started one or two days later.

When, in the early eighteenth century, Bél (1984) described the Kőszeg wine region, he not only provided information concerning his own period but, as he himself mentioned, described the long-term tradition. Thus, a rather detailed documentation of vine management traditions and techniques are available, among others, related to methods of wine production, soils, location of vineyards, risks of wine production with special emphasis on weather, vine types, circumstances concerning the harvest (timing, way of harvesting, location and extension of vineyards, etc.). Many of his descriptions are supported by the detailed documentation, available in the local administrative documentation, namely in the town council protocols, the Book of Vinesprouts and the accounts of the town. All these details provide us with evidence concerning a rather uniform management practice which had changed little over space and time.

Changes in vine type affect the climatic interpretation of the data. Therefore, determination of the main vine types and their biological response to climate is important. Twentieth century vine types of the Kőszeg region, for example, have vine ripening and grapevine harvest significantly (weeks) earlier than vine types of the ‘traditional’ (pre-late nineteenth century) period. The main vine types, mentioned in the (seventeenth–)eighteenth and nineteenth century sources of Kőszeg wine production (Table 1), are traditionally wide-spread and provided the basis of white wines, either in the western parts or in the whole Carpathian Basin (Kozma 2000; Hajdu 2003). Similarly to earlier periods, in the 1830s and 1870s, Furmint and Zirfandel were still mentioned as the most common vine types of the region (Schams 1833; Chernel 1877), but apart from that, the growing importance of other (also red) wines can be detected in the second half of the nineteenth century (e.g. Great Burgunder, Riesling, Tramini). Although no detailed information is available about the vine types of the area before the early eighteenth century, traditionally in vine plantations, the sticks from local vinestocks were used (Bél 1984), which means that the same vine types were also predominant in at least several decades prior to the description. Thus, concerning prevailing vine varieties, the area was quite conservative up to the late nineteenth century, and external impacts such as new varieties clearly played a marginal role (see also Fényes 1851) and therefore phenophases, quality and quantity of vines were dependent on the same vine types (Appendix Table 4).

Table 4 Reported main and rare vine types in Kőszeg: eighteenth and nineteenth centuries

The (biological) full ripening of these, ‘traditional’ vine types occur approximately at the same time: in early to mid-October (Hajdu 2003). This phenological information is in good agreement with the information kept in the town council protocols of Kőszeg, where hillmasters and the mayor usually reported the full ripening of grapes around early or mid-October, when decisions about the actual vine harvest (starting date, preparations, etc.) were made. When first ripening is mentioned in connection with the election of vine shepherds (around 10 August), Zirfandel with its somewhat earlier ripening appeared in the town protocols (KTCP 17 Aug. 1721: ’cirfandli’, 10 Aug. 1739: ’Czirfandel’). Among the grapes harvested for producing wine, Zirfandel might have played an important role in defining the starting date of grape ripening. In traditional wine production, unlike today, Kőszeg wines were predominantly white ones. This situation remained approximately the same until the late nineteenth century (see, e.g., Bél 1984; Schams 1833; Fényes 1851; Chernel 1877). Kőszeg wine as a ’trade-mark’ often meant a mixture or marriage of the juice or wine of these different types (Bél 1984), which is important in understanding quality descriptions and judgements, similarly documented in the Book of Vinesprouts from 1740 onwards. A gradually growing importance of red wines can be followed especially from the mid- and late nineteenth century (e.g. Chernel 1877).

What do we know about grain: predominant cereal types

Rather precise information is available concerning harvested grain types due to the fact that in almost each case, while defining the share of harvesters, the harvested grain types were listed in the town council protocols. In the tithe accounts, only the word ’grain’ (H-’gabona’, G-’traidte’) appeared to denote taxation. Nevertheless, more information on the harvested grain types was described in the town council protocols while ordering the annual share (in crop percent) of harvesters: winter crops and spring crops were harvested each year, one after the other, respectively. Winter crops such as wheat and rye were detailed in the protocols every second or third year from 1650 until 1760. Spring crops were mainly oat and barley (KTCP, e.g. 1650, 1653–1654, 1660, 1663–1664, 1709, 1725–1726, 1744, 1760), but sometimes also buckwheat (KTCP, e.g. 1726, 1731–1732, 1739–1740) and millet (KTCP, e.g. 1732) were mentioned. From the late eighteenth century, maize also appeared as part of the crops harvested (KTCP, e.g. 1794). Although no direct information is available concerning the quantity sown and harvested proportions of different grain types, wheat had usually the greatest importance in the area, followed by rye and barley and then oats together with other spring crops (e.g. SzTCP, 1790, 1840). Even if there is not always detailed information available concerning management and cereal types, the town protocols and town management accounts similar to the case of vine, show rather conservative and predominantly unchanged agricultural practices through the study period.