| 25 Aug 2026
Assessment of historical heavy precipitation in Poland based on documentary evidence
Abstract. Anthropogenic climate change is intensifying the hydrological cycle and increasing extreme precipitation across much of Europe, yet understanding of pre-instrumental precipitation extremes in Central Europe remains limited by short instrumental records. This study presents the first comprehensive, documentary-evidence-based database of heavy and extreme precipitation events in Poland for the 12th–18th centuries. An assessment of a documentary-based database of heavy precipitation identified 477 such events. The extreme precipitation events were identified and classified into three types (long-lasting, heavy, or violent), based on reported impacts, duration, and affected regions. Results showed that the number of documented events increased over time, from a single event in the 12th century to 242 in the 18th century, with a peak of 182 events in the 1701–1750 sub-period. Silesia (37 %) and the Baltic Coast and Pomerania (28 %) were the most frequently affected regions, and summer accounted for the majority of all events. The duration of 273 events could be established, of which just over half lasted a single day. Notably, 316 impact assignments described damage never framed as a flood. These findings confirm that documentary sources capture fine-grained detail on the timing, duration, and consequences of past extreme-precipitation episodes, offering a valuable counterpart to multi-proxy reconstructions of longer-term hydroclimatic variability in Poland and Central Europe. This dataset offers a long-term baseline for evaluating present-day and projected changes in Central European precipitation extremes.
This paper presents a new database of historical extreme rainfall events in Poland for the pre-instrumental period (12th-18th centuries). The database draws on a variety of documentary sources. Events are characterised by the type of precipitation (heavy, long-lasting, and violent) and into 4 impact categories following a method used in the authors’ previous documentary flood database. The authors present summary statistics of the dataset, including seasonal and regional totals for each category of event. They find interesting similarities between historical and recent occurrences of extreme rainfall: for example, the relatively small Silesia region, which recently experienced devastating floods in 2024 following Storm Boris, consistently accounts for a disproportionate number of extreme precipitation events in the database in each century. They conclude that anthropogenic warming is intensifying existing tendencies towards heavy rainfall in this region, rather than it being an entirely new phenomenon.
In my view, the strongest conclusions that are drawn regard the regional partitioning of impacts and occurrence. It is very useful to draw attention to regions which have experienced flood and heavy rainfall events in the past, particularly when meteorological mechanisms can be identified as in Silesia. This is a strong motivation for the work. The biggest limitation is the changes in record keeping over time (and potentially between regions – see my comments below). In the case of event counts, it can introduce spurious trends into the counts over time. For example, though there is a general trend towards increased occurrence in later centuries, the authors acknowledge this may simply reflect more consistent record keeping towards the latter part of their period. They note some similarities with occurrence of heavy rainfall and flooding in other regions (Spain, Sweden) in the early 18th century, although presumably the same issues with record keeping could apply here too. However, I appreciate that they are careful to outline these caveats and avoid drawing conclusions about trends too strongly.
The language and figures are of a very high standard throughout. Results are well presented and organised in a sensible way, and the title and abstract are clear and appropriate to the text. The authors discuss relevant literature at several stages: the introduction sets results in the context of literature regarding recent changes in extreme precipitation over Central Europe, and the Discussion compares the authors’ conclusions to findings from similar databases in other European countries. Overall I enjoyed reading this paper – it is of a high standard and is a valuable contribution to the field.
I have a couple of general (and fairly minor) comments which the authors may wish to consider, and a couple of specific points. Otherwise I am pleased to recommend it for publication in Climate of the Past.
Comment on usefulness of the database
The authors state that the database provides a “long-term baseline against which modern and projected changes in Central European precipitation extremes can be assessed” (L374). However, it is not a quantitative dataset, and there are limitations regarding its use to infer long-term trends due to the changes in record keeping. I wonder if it’s worth commenting specifically on what kind of baseline they mean: for example, can this database help our understanding of future climate change? In my opinion, it seems most useful in the regional and seasonal partitioning of heavy rainfall events in Poland. This could be useful in assessing climate model simulations – for example, do simulations of the preindustrial period reproduce this partitioning or are there existing model biases? And do we see changes in the partitioning in future projections? A couple of sentences of comment from the authors would be interesting.
Comment on changes in number of events recorded in each region
Fig 3b. The authors mention that the 30 Years War may have contributed to the relatively low number of events in the 17th vs 18th centuries (L364). Do they have any intuition as to why the growth in events in the Baltic Coast/Pomerania region is so much higher than other regions in the 18th century? I wondered if it reflected non-meteorological factors, e.g. development along the coast region leading to more events being recorded (just an idea – the authors will be more familiar with this region’s history). If so, this could be an important caveat to discuss regarding the regional partitioning more broadly, i.e. to what extent could the differing event frequencies in different regions reflect non-meteorological differences in population, occupations, etc.?
This could also be relevant for Figs 6 and 7. Silesia dominates the flood categories in Fig 7a., but less so Other Impacts, despite having by far the heaviest share of Violent Rain (Fig 6a). Before seeing the figure, I would have expected that Violent Rain is particularly important for non-flooding damage, e.g. to crops and built structures, although my assumption could be wrong. However, the authors also note that the number of Other Impacts events grows much more than the flood categories in the 18th century (Fig. 7b), and I noticed that the Baltic Coast and Pomerania region has a higher share of Other Impacts (23%) than in the flooding categories. We saw in Fig 3b that the Baltic Coast and Pomerania region recorded many more events in the 18th Century than in previous ones. So, is it possible that the time-averaged results in Fig 7a are being skewed by the growth in recordings from one particular region (Baltic Coast/Pomerania), which may reflect social rather than meteorological changes? I don’t know if this is true – I would be interested to hear the author’s thoughts. They might also consider plotting a version of Fig 7a but excluding the 18th Century, for example, to see if this is the case.
Specific comments
L162: “There were only nine and ten heavy precipitation events in the 13th and 14th centuries, respectively.” – the authors may want to clarify that there were only nine and ten recorded events, as we know that not all events would have been recorded (the authors are careful to make this distinction elsewhere).
Fig 4b: Could also plot the number of events where no month could be determined – or at least state it in the text (477 – 404 = 73, I think? This would be helpful to state or plot explicitly).