To improve our knowledge concerning climate change in Poland from 1001–1500, we constructed three new late-winter–early-spring temperature reconstructions covering the period since the 12th century based on tree-ring widths (Figs. 9 and 10a–c). In addition, winter and summer temperatures for the 15th century were reconstructed based on documentary evidence (Fig. 5). Unfortunately, the scarcity of historical sources (see Fig. 2) for the area of present-day Poland does not allow such reconstructions for the earlier centuries. For the complete synthesis of Poland's climate in the study period and for comparison purposes, we used all available reconstructions in the literature of cold-season temperatures based on tree-rings widths (mainly P. sylvestris) covering medieval times: January–April 1170–1994 (Przybylak et al., 2001), December–March 1091–2006 and December–March 1109–2004 (A. alba; Szychowska-Krąpiec, 2010), February–March 1168–2000 (Koprowski et al., 2012), and November–April 1200–2010 (Balanzategui et al., 2017). See also the recently published inventory of all available dendrochronologies for Poland (Opała-Owczarek et al., 2021, Table 5.2 therein). Information about climate conditions occurring in Poland in the 11th century is available only from analysis of biological proxies taken from laminated (varved) sediments in Lake Żabińskie (NE Poland). Quantitative reconstructions of the cold-season (Hernández-Almeida et al., 2015) and August temperatures (Hernández-Almeida et al., 2017) are available for the entire millennium (Fig. 10f and g).

The 11th century in Poland was the coldest in the entire study period in light of both chrysophyte-based reconstruction of a number of days below 4 ^∘C (DB4 ^∘C) representing the severity of winter and chironomid-based reconstruction of August temperatures (Fig. 10f and g). Particularly cold was the first half of the 11th century. This century was also colder in comparison to the contemporary period (1951–2000). For the 12th century, we also have data from old (Szychowska-Krąpiec, 2010, Fig. 56 therein) and new (see Figs. 9 and 10b, c) reconstructions of air temperature from Lesser Poland based on fir and pine chronologies. The 12th century was markedly warmer than the 11th century and also warmer than at present. In particular, the summer seasons in the first half of this century were very warm (August temperature of about 18 ^∘C; Fig. 10g), while in the second half they were close to the 1951–2000 norm (16.9 ^∘C). As shown in Fig. 10f and g, the 12th century was the warmest century in the entire study period in terms of summer temperature, while the cold-half-year temperature was slightly colder (except the last 2–3 decades) than in the next 3 centuries. The reconstructions of winter temperature for Lesser Poland based on dendrochronological data (Fig. 10b and c) are characterized by strong fluctuations oscillating from about 0 to −2 ^∘C relative to present means. But on average, winter conditions were warmer than in the 13th, 14th, and 15th centuries.

Proxies represented the entire cold season (see Fig. 10d and f), revealing the existence of warm temperatures in the 13th century that were particularly high in its first half. The Scots pine reconstructions of late winter and early spring (Feb–Mar) for both Lesser Poland and the Kuyavian-Pomeranian region (Figs. 9 and 10a, b, d) generally also confirm this finding. The average temperature of this season was only slightly colder (by 0.2 ^∘C) than at present. Summer temperatures in the 13th century oscillated below and above the long-term contemporary norm. Similarly to winter temperature, summers were also on average slightly colder than at present by the same value. However, this century was significantly colder (by 0.3–0.6 ^∘C) than the 12th, 14th, and 15th centuries.

Good agreement exists for the cold-season temperature dendrochronological reconstructions for Poland in the 14th century. They indicate the existence of colder conditions in the 1330s and the last 3 decades of the century and warmer conditions in the middle of the century (see Fig. 10). On the other hand, values of the reconstructed number of DB4 ^∘C are slightly lower than the norm for the period 1001–2000 (100.3 d) and significantly lower than the present norm (102.7). That is, winters were mild and thermically stable in this time (see Fig. 10f). Summer temperature was relatively high in almost the entire century and very high in the last decade, when it was comparable to that in the first half of the 12th century (see Fig. 10g). On average, the temperature was only slightly lower than in the 12th century (by 0.1 ^∘C), but it was higher than in the latter half of the 20th century (by 0.4 ^∘C).

For the 15th century, we have many excerpts describing the weather. Based on these sources, winter and summer temperatures were reconstructed (see Fig. 5). They document the existence in that century of very cold winters and very warm summers, which means that climate continentality was very high. Sadowski (1991) estimated that Poland's climate continentality was greater in the 15th century than at any other time in the period 1201–1980 that he had analysed. The chironomid-based reconstruction of August temperatures (Fig. 10g) shows a similar time pattern to the summer temperature changes presented in Fig. 5. In both reconstructions, higher temperatures were observed in the second half of the century, and particularly in the 1470s. In the case of winter temperatures, all dendrochronological reconstructions generally confirm that they were usually lower than at present (Fig. 10), although large oscillations are seen. Some contradictions, however, are noted regarding the occurrence of the 2 coldest decades in the reconstruction based on documentary evidence (the 1430s and 1450s). The first cold decade was noted in Lesser Poland, but not in northern and central Poland. In the latter areas, very warm winters occurred in this time. Similar results are shown by the reconstruction of the number of DB4 ^∘C. At this time, they were the fewest (only 96 d) of the entire study period (Fig. 10f) and even the entire millennium (Hernández-Almeida et al., 2015). On the other hand, there is correspondence of the thermal character of winter between the reconstructions based on documentary evidence and the biological proxy (the number of DB4 ^∘C) for the second half of the 15th century. The coldest winters (in the 1450s) were significantly better registered in trees growing in northern and central Poland, while in southern Poland a warming was even noted (Fig. 10). The decrease in winter temperature in Poland in the 1430s and 1450s was confirmed by the reconstruction of the ice winter index in the western Baltic Sea for the 15th century (Koslowski and Schmelzer, 2007).

We observed that, for Scots pine from Lesser Poland and northern Poland, negative pointer years outnumbered positive pointer years. This was probably caused by the influence of unfavourable climatic conditions prevailing in the study period and long and severe winters. For the period 1361–1500, 38 pointer years were found for which information about weather conditions was available in the historical sources. For as many as 71 % of pointer years, descriptions of extreme weather conditions were found that favoured the annual tree growth being either low (e.g. severe and long winter, drought) or high (e.g. warm winter, heavy rainfall).

The interpretation of climate based on pointer years in oaks is more difficult because they depend simultaneously on temperature and precipitation. Krąpiec (1998) found that positive pointer years for oaks in Poland are associated mainly with heavy rainfall in the growing season following a mild winter, while negative ones are linked to dry years with cold and snowless winters and to spring frosts damaging the cambium. For example, the negative pointer years of 1314 and 1317 in Lesser Poland were caused by severe cold weather that caused a famine in Bohemia and neighbouring areas (Brázdil and Kotyza, 1997). The analysis of moon rings (MRs) of oaks is easier to interpret, allowing long and severe winters to be determined. In the period 1361–1500, 19 such winters were recorded, and, for almost half (47.4 %), the presence of moon rings was found.

As results from the above summary for Poland in medieval times, most of the proxies allow only for the temperature reconstruction of the entire cold half-year (November–April) or some of its sub-periods. Information about the weather in summer for the whole of the study period is limited to only one reconstruction of August temperature based on chironomid assembles (Fig. 10g). For late medieval times, some information is also available from documentary evidence. As a result, it is difficult to compare our results presented in the paper with those from the Northern Hemisphere and the European territory since proxy series of data (tree rings in particular) for the mentioned areas are biased towards the warmer seasons of the year (e.g. Table 1 in Ljungqvist, 2010). Fortunately, some comparisons are possible using existing reconstructions for European regions based on documentary evidence. However, due to the scarcity of available historical sources, the comparison should be limited mainly to late medieval times.

Comparison of August temperature from NE Poland (Fig. 10g) against summer temperature (June–August) reconstructed for Europe by Luterbacher et al. (2016) shows quite good correspondence. In both reconstructions, the coldest temperatures in the period 1001–1500 occurred in the 11th century and the mid-15th century. Cold summers were also noted from the mid-13th century to the mid-14th century. On the other hand, in both reconstructions, the markedly warmest summers occurred in the 12th century. However, in Poland they ended earlier than in Europe (in the middle part of this century) but started earlier, at the end of the 11th century. Smaller summer warmings in both areas were also noted at the turn of the 15th century and in the second part of the 15th century.

In Europe, however, the scale of warming in these times was significantly smaller than in the 12th century, while in Poland it was comparable across the periods, but the duration of the warming was shorter in both areas. The comparison of the results is most reliable for the 15th century, for which we reconstructed summer temperature (Fig. 5). The second half of the century was clearly warmer than the first, not only in Poland but also in the Czech lands (Brázdil, 1994, 1996), Switzerland, and in the entirety of central Europe (Camenisch et al., 2016), as well as in Scandinavia (Gouirand et al., 2008). The decade 1471–1480 was the warmest in the entire 15th century in Poland (Fig. 5) and also in the Low Countries and central Europe (Camenisch, 2015; Riemann et al., 2015; Camenisch et al., 2016), while the warmest summers occurred in the 1480s in the Czech lands (Brázdil, 1994, 1996) and in the 1490s in Switzerland (Trachsel et al., 2010; Camenisch et al., 2016). For the 15th century, there exists a correspondence between the run of the August temperature in Poland (Fig. 10g) and reconstructions of summer temperature for some European areas based on documentary evidence and tree rings (Fig. 2; see also reconstructions 2–5 in Camenisch et al., 2016). The summer temperature was colder in the first part of the century than in the second.

Luterbacher et al. (2010) found that the correspondence of temperatures between Poland and Europe is lower for summer than for winter. In the latter season, there exists a very high correlation (0.96, at interannual and multidecadal timescales) between reconstructed Polish and (non-Polish) European mean temperatures (e.g. excluding the grid points representing Poland) over the period 1500–2000. However, for Europe there are only some reconstructions based on documentary evidence (e.g. Brázdil, 1994, 1996; Pfister et al., 1996, 1998; van Engelen et al., 2001; Shabalova and Van Engelen, 2003; Glaser and Riemann, 2009; Camenisch, 2015; Camenisch et al., 2016; Retsö and Söderberg, 2019) and modelling works (e.g. Goosse et al., 2006; Schimanke et al., 2012). The scarcity of historical sources for the study period, particularly before the 15th century (e.g. Brázdil, 1996; Glaser and Riemann, 2009; Oliński, 2022; see also Fig. 2), means that the reconstruction uncertainties can be significant. Nevertheless, a correlation exists for some periods between Poland's winter temperature and reconstructions for European areas based on documentary evidence.

For example, such agreement was observed in the 11th century, when winters were lower than average in western–central Europe as well (Pfister et al., 1998). Similar results were also found for Europe north of the Alps (Alexandre, 1987, after Brázdil, 1996) and the Czech lands (Brázdil, 1996), while the opposite was found in the Low Countries (van Engelen et al., 2001). In the 12th century, there is still good agreement with the reconstruction of the DB4 ^∘C, but less agreement with the reconstructions of February–March temperatures (compare Fig. 10 in Pfister et al., 1998, and Fig. 9 here). On the other hand, all reconstructions reveal a significant winter warming at the end of the 12th century that has also been noted in Germany (Glaser and Riemann, 2009) and the Low Countries (van Engelen et al., 2001). In the 13th century, changes in the reconstructed temperature in Poland and western–central Europe differed much, particularly in the first half of the century. However, the warming of winters in this time also occurred in England, Ukraine, and Russia, but not in Europe north of the Alps, the Czech lands (Fig. 4 in Brázdil, 1996), Germany (Glaser and Riemann, 2009), or the Baltic Sea basin (Schimanke et al., 2012). On the other hand, a good coherence was usually stated for the 14th century, where in the two compared areas, colder conditions prevailed, except in the middle of the century (compare Fig. 2 in Pfister et al., 1996), with available reconstructions from Poland, including new ones presented in Fig. 9. Agreement is also seen for Europe north of Alps, but not for the Czech lands, Ukraine, or Russia (Brázdil, 1996) and the Low Countries (van Engelen et al., 2001). The reconstructed winter (DJF) temperature in Poland for the 15th century based on documentary evidence (Fig. 5) matches well with the data presented for the Czech lands, Russia, England (Brázdil, 1994, 1996), the Baltic Sea basin (Schimanke et al., 2012)., Sweden (Retsö and Söderberg, 2019), the Low Countries (Camenisch, 2015), the Alps (Mangini et al., 2005), and the whole of Europe (Goosse et al., 2006), but only partly with data for Germany (Glaser and Riemann, 2009).

Ljungqvist (2010) found that the Northern Hemisphere proxy collection is biased toward the year's warmer seasons. The reason for this is that predominant among the proxies in the reconstruction process are tree rings taken mainly from mountain regions (see PAGES 2k Consortium, 2013, 2017; Tardif et al., 2019). Nevertheless, the reconstructed summer temperatures are frequently used to represent annual values (e.g. Cook et al., 2004; Moberg et al., 2005; Ljungqvist, 2010) due to the high correlation between seasons on decadal and longer timescales within the instrumental period. However, for smaller areas that are dominated by lowlands – like Poland – this is not the case. There is lack of correlation between mean summer and winter temperatures for Poland in 1901–2000. The correlation of summer temperature with the annual temperature is 0.4 and is 2 times smaller than the correlation of winter temperature (0.79). Therefore in Poland, winter temperature represents the annual temperature better than summer temperature. It is thus justified to use winter temperatures as annual proxies and thus to use them to distinguish the occurrence or not of the MWP (MCA) in Poland.

Due to the coverage of the entire millennium and also of the whole cold season, the best proxy that can presently be used to delimit the MWP in Poland is DB4 ^∘C (Fig. 10f). These data locate this period in the time period 1180–1440 (excluding the period 1260–1270, partly a consequence of a volcanic eruption in 1268), when the winters were generally shorter and warmer than today (1951–2000). The other proxies usually do not allow the beginning of the MWP to be distinguished (due to the reconstructed data series starting too late) but only its end. There exist, however, some essential differences between proxy results. The MWP is also suggested to have ended in the first half of the 15th century (e.g. in line with DB4 ^∘C data) by Sadowski (1991). Some other authors concluded that the MWP in Poland ended at the beginning of the 14th century (Maruszczak, 1991; Kotarba, 2004; Balanzategui et al., 2017; Szychowska-Krąpiec, 2010, only fir chronology) or even the beginning of the 13th century (Koprowski et al., 2012). Based on new February–March reconstructions (Fig. 9), it is possible to distinguish the MWP (from the end of the 13th century to the mid-15th century) only in series representing the Kuyavia-Pomerania region. These are thus in correlation with DB4 ^∘C data, except the first 100 years. All the mentioned quantitative reconstructions suggest that the MWP in Poland was comparable to or warmer than the current temperature (1951–2000). For Europe, various time frames are given for the MWP, but most suggest that this period ended in the 14th century (e.g. Brázdil, 1996, for the Czech lands; Glaser and Riemann, 2009, for Germany; Millet et al., 2009, for the northern French Alps; Niemann et al., 2012, for the Swiss Alps; Niedźwiedź et al., 2015, for the Baltic Sea basin). The simulation of mean winter European temperatures presented by Goose et al. (2006, their Fig. 3) locates the MWP as extending until the first half of the 15th century; however, when we take mean summer temperature, the MWP ended in the beginning of the 13th century. The lack of spatial coherence in timing of the MWP across the entire globe was also recently confirmed by Neukom et al. (2019) based on proxy data published by PAGES 2k Consortium  (2017). They locate the MWP from 751 to 1350, and according to their Fig. 3, within this time span the peak warming in Poland occurred between 800 and 1000. However, we need to add that the PAGES 2k databases do not contain a single proxy data series taken from Poland.

The warm/cold periods observed in Poland during the study period are an effect of natural variability in the climatic system (see the course of the set of forcings presented in Fig. 1 in Goose et al., 2006). For example, the low temperatures in the 11th century can be connected with negative anomalies of total solar irradiance (TSI), while in the 15th century they can also be associated with increased volcanic activity and the negative phase of the North Atlantic Oscillation (NAO) (Trouet et al., 2009). For example, the very cold decade of the 1450s observed both in Europe (Goose et al., 2006) and globally (e.g. Tardif et al., 2019) is most often explained by the Kuwae eruption of 1452 or 1453 (Gao et al., 2006). According to those authors “the Kuwae eruption was the largest stratospheric sulfate event of that period [since 1300, authors], probably surpassing the total sulfate deposition of the Tambora eruption of 1815, which produced 59 kg SO₄ km⁻² in Antarctica and 50 kg SO₄ km⁻² in Greenland.” The eruption probably changed atmospheric circulation in the North Atlantic, leading to the occurrence of the negative phase of the NAO (see Fig. 2 in Trouet et al., 2009), which is important in shaping the climate in Poland, as we show in the next paragraph.

Any extreme thermal condition in Poland in a season or entire year is most often caused by anomalous behaviour of atmospheric circulation and only infrequently by volcanic activity. According to investigations made by Przybylak et al. (2003) for the period 1500–1990, the main cause of cold winters in Poland in this time was the negative phase of the NAO (in which advection of cold air from the north and east dominates), while warm winters were observable when the NAO was in the positive phase (advection from the west dominates). The occurrence of summers that are extremely warm or very warm in Poland is also connected with changes in atmospheric circulations. However, cold summers are caused by advection from the west and north and warm summers with southerly and easterly advection (Bartoszek, 2017; see Table 3 therein). It is reasonable to assume that the relationships observed in the contemporary summer climate of Poland are also obtained in the study period.

The tree-ring chronologies supporting the conclusion of this study are available from Elżbieta Szychowska-Krąpiec, Marcin Koprowski, and Marek Krąpiec upon request.

The supplement related to this article is available online at: https://doi.org/10.5194/cp-19-2389-2023-supplement.

Conceptualization: RPr, PO, MKo; methodology: RPr, PO, MKo, MKr, ESK; validation: RPr, PO, MKo, MKr, ESK; formal analysis: RPr, PO, MKo, MKr; investigation: RPr, PO, MKo, MKr, ESK, RPu, AP; resources: RPr, PO, MKo, MKr, ESK, AP; data curation: RPr, PO, MKo, MKr, ESK, AP; writing – original draft: RPr, PO, MKo, MKr, ESK; writing – review and editing: RPr, PO, MKo, MKr, ESK, RPu, AP; visualization: RPr, PO, MKo, MKr, RPu; project administration: RPr; funding acquisition: RPr, ESK, MKr.

The contact author has declared that none of the authors has any competing interests.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors.

The research work of Piotr Oliński and Rajmund Przybylak was supported by grants funded by the National Science Centre, Poland (grant nos. 2013/11/b/HS3/01458 and 2020/37/B/ST10/00710). Marcin Koprowski and Radosław Puchałka were supported by a grant funded by the National Science Centre, Poland (grant no. 2020/37/B/ST10/00710). Elżbieta Szychowska-Krąpiec and Marek Krąpiec were supported by the grant of the Faculty of Geology, Geophysics and Environmental Protection of AGH University of Science and Technology in Kraków (grant no. 16.16.140.315).

This paper was edited by Natalia Piotrowska and reviewed by two anonymous referees.