Processing and Interpretation

Data processing

For their climate reconstructions historical climatologists need continuous, long and homogenous time series with information that can be converted into quantitative entities. Researchers on climate impacts, however, require information about temperature and precipitation with a high resolution for their studies on societal mitigation and adaptation strategies.

Bio-physical proxy data allow an estimation of temperatures for periods of several months. Those estimation periods, though, may be based on quite diverse monthly temperatures and precipitation patterns: For example, an early rye harvest might be caused by an extremely warm May and an average June or by a rather cool May and a very warm June or by an extremely warm April, an average May and a rather warm June. Descriptions of weather preserved in chronicles and weather diaries enable us to reconstruct the general conditions of single months. Those individual reports, however, are non-uniform and cannot be classified in all cases.

These so-called Pfister indices are divided into seven different classes (cf. Mauelshagen 2010). These contain information about the order, but not about the magnitude of the differences.

  • -3: extremely cold/extremely dry
  • -2: very cold/very dry
  • -1: rather cold/rather dry
  • 0: average without any significant positive or negative trend
  • +1: rather warm/rather wet
  • +2: very warm/very wet
  • +3: extremely warm/extremely wet

For example, a month with the temperature index -3 is colder than one with the index -2, but nothing can be said about the magnitude of the difference. A basis is always needed for such comparisons. For the Pfister indices contained within Euro-Climhist the basis is always the period 1901 to 1960. On the one hand, it is no more part of the Little Ice Ace Age. On the other hand, it is not yet much affected by rapid Global Warming. Accordingly, a month with a temperature index of -3 was very cold according to the period 1961 to 1960.

Historical climatologists deduce monthly and seasonal temperature and precipitation indices from their documentary evidence using a specific methodology, which has been developed during the last years. Indices are based upon a great variety of personal sources of different spatio-temporal resolution. Most of them are not quantified, they differ in form and content and they are often incomplete. Nevertheless, some part of the evidence may be quantified. In particular this holds for days with precipitation mentioned in daily weather reports, which can be aggregated by month (series 2-7). Moreover, descriptions of extreme temperatures frequently include quasi-objective observations which are comparable over time, such as hints to the – advanced or delayed – stage of vegetation, the duration or absence of a snow-cover, the freezing of rivers and lakes or the appearance of spring vegetation in the cold season. Descriptions of floods or low waters are mentioned as indication for heavy rainfall or long droughts.

In addition, climate-sensitive evidence is available in the files of institutions. For example, the time of grain or grape harvest is registered year by year in some files, often for several centuries (series 14 and 15). Temperatures over several consecutive months can be estimated from such time series using appropriate statistical methods and similar to the evidence from natural archives such as tree-rings.

In order to get Pfister temperature and precipitation indices, the entirety of data available for a month or a season needs to be integrated in a data field. The different data should corroborate and complete each other. Moreover, a meteorologically plausible tendency should emerge. Due to the increase of evidence for Switzerland after 1550 a monthly temperature and precipitation index is obtained.

The well documented extremely cold April 1731 provides a suitable example how to deduce a monthly temperature and precipitation index:

Euro-Climhist: Query for April, 1731
(excluding weather impacts) (access 19th August, 2015)
1731-April 1-10 / snow melts completely / Ct. Nidwalden
1731-April 11-20 / enduring snow-cover / Ct. Appenzell-Innerrhoden
1731-April 21-25 / cold / Ct. Nidwalden
1731-April 21-30 / huge snow-masses / Winterthur (ZH)
1731-April 21-30 / snow cover lasting some days / Ct. Nidwalden
1731-April / cold / Ct. Geneva
1731-April / easterly wind / Bätterkinden (BE)
1731-April / very long duration of snow-cover / Winterthur (ZH)
1731-May 7 / sweet cherry trees bloom (127 days after New Year), extremely late / Mittelland
1731-April / Pfister temperature index: -3 extremely cold / Mittelland
1731-April / aggregate precipitation 102 mm: wet / Zurich (ZH)
1731-April / 11 precipitation days: rather dry / Winterthur (ZH)
1731-April / Pfister precipitation index: 0 average / Mittelland

On a whole, this data-field is meteorologically plausible corresponding to a very cold month of April. The extreme character of the month can be concluded from several indictors: The extremely long snow-cover in Winterthur and in Nidwalden (Stans), the prevailing chilly north-easterly wind and finally the extreme delay of the sweet cherry blossom in the Swiss Mittelland, which is a good indicator of April temperatures (Rutishauser, Studer 2007).

The Czech (historical) climatologist Petr Dobrovolný and his co-authors (Dobrovolný  et al. 2010) have assessed monthly temperatures from 1501 to 2000 on the basis of temperature indices for Germany, the Czech Republic and Switzerland and the subsequent instrumental series (eighteenth century onwards). Their approach allowed assessing the standard error of estimate.

The connection between climate reconstruction and climate impact research is made possible by monthly or seasonal indices for temperature and precipitation that can be deduced from historical documentary sources. The evidence available for a time interval is sorted chronologically and then interpreted in an integrated manner month by month or season by season. The thermic and hygric conditions are assessed with reference to a monthly or seasonal temperature and a precipitation index.



Societies take up the topic of climate change primarily in the wake of impressive extremes and their consequences; therefore people asked and ask for explanation. At present, they are often attributed to the enhanced Greenhouse Effect without adequate reflection. Analog historical cases are, on the other hand, often quoted by climate sceptics to demonstrate that such extremes "did always happen". Isolated extreme events may not be used as evidence neither "in favor" nor "against" the occurrence of Global Warming, because they cannot be statistically analyzed. A single "Siberian" winter does not indicate a new Ice Age, and extremely warm summers such as 2003 have been exceeded by the eleven-month lasting hot and dry weather in 1540 (Wetter et al. 2014).

The history of weather and climate does not end with the establishment of the measurement network of MeteoSwiss in 1864. Vice versa, many questions related to the history of weather and climate rise from current weather events. For this reason Euro-Climhist includes some selected long series of instrumental and proxy data up to present times.

On the one hand the climate-historical interpretation draws on the graphical representation of results from documentary evidence, and on the other hand on the comparison with results from data out of the archive of nature, primarily tree-rings as well as reconstructed historical variations of Alpine glaciers.

Caution is needed in the interpretation of certain proxy data. The unfolding of leaves of a chestnut tree in the centre of Geneva was systematically registered since 1808 (series 17). The curve displays a strong trend towards an earlier blossoming since the late nineteenth century. This feature needs to be mainly credited to Urban Warming, as the comparison with a series of the first blossom of chestnut trees in the rural environment of Hallau (Canton Schaffhausen) indicates (Wetter, Pfister 2014).