How Joannites’ economy eradicated primeval forest and created anthroecosystems in medieval Central Europe

During European states’ development, various past societies utilized natural resources, but their impact was not uniformly spatially and temporally distributed. Considerable changes resulted in landscape fragmentation, especially during the Middle Ages. Changes in state advances that affected the local economy significantly drove trajectories of ecosystems’ development. The legacy of major changes from pristine forest to farming is visible in natural archives as novel ecosystems. Here, we present a high-resolution densely dated multi-proxy study covering the last 1500 years from a peatland located in CE Europe. The economic activity of medieval societies was highly modified by new rulers—the Joannites (the Order of St. John of Jerusalem, Knights Hospitaller). We studied the record of these directorial changes noted in the peat profile. Our research revealed a rapid critical land-use transition in the late Middle Ages and its consequences on the peatland ecosystem. The shift from the virgin forest with regular local fires to agriculture correlates well with the raising of local economy and deforestations. Along with the emerging openness, the wetland switched from alkaline wet fen state to acidic, drier Sphagnum-dominated peatland. Our data show how closely the ecological state of wetlands relates to forest microclimate. We identified a significant impact of the Joannites who used the novel farming organization. Our results revealed the surprisingly fast rate of how feudal economy eliminated pristine nature from the studied area and created novel anthroecosystems.


Site description. Pawski Ług is an ombrotrophic Sphagnum-dominated peatland located in western Poland
in the Łagów Lakeland (52°19′45″N, 15°16′30″E). The area is covered by morainic hills and a number of small glacial lakes. The peatland covers an area of 3.67 ha, and its vegetation is currently dominated by Sphagnum fallax, S. angustifolium, and other species typical for raised bogs: Drosera rotundifolia, Oxycoccus palustris, Eriophorum vaginatum, and Ledum palustre. The peatland is surrounded by the mixed forest mainly composed of Quercus, Fagus sylvatica, and Pinus sylvestris. Pawski Ług is protected as a nature reserve since 1970.

Results
Regional historical background. Historical settlement structures near the Pawski Ług mire (Fig. 1) developed in the late Middle Ages as elements of a feudal estate with the central administrative unit in Łagów. The oldest traces of permanent settlement in the vicinity of the site (within a 5-km radius) relate to Łagów and the people of the Lusatian culture, which developed in this area between approximately 1300 BCE and 500 cal. BCE 36 . The first traces of Slavic settlement were found between 800 and 950 CE. It may have been the settlement of the Leubuzzi tribe, recorded in the work of the German chronicler Adam of Bremen. At the beginning of the second half of the tenth century, these areas were probably temporarily incorporated into the emerging Polish State, which was associated with the destruction of the network of older settlements of the Middle Oder region (hence perhaps noticeable traces of fire) and the construction of a new type of fortified settlements 37 . In 1124-1125 CE, Polish prince Bolesław Krzywousty founded a bishopric with the capital in Lubusz (German: Lebus) for the areas lying in the fork of the rivers Warta and Oder, which gave the name to the entire Terra Lubucensis region 38 .
In 1249-1252 CE, the area was granted in exchange for military aid by the Silesian Prince Bolesław Rogatka to the Archbishop of Magdeburg and the Magdeburg Margraves 39 . The latter built wooden fortifications (castrum Lagowe) along with a stone edifice (Hus Lagow) in Łagów for guarding the eastern border of Brandenburg 40 , and they started the process of colonization. It entailed the foundation of new villages or bringing new German settlers to villages previously inhabited by the Slavic population. By the mid-fifteenth century, the number of Terra Lubucensis settlements recorded in written sources increased from 61 to 175 41 . In 1347 CE, the Brandenburg Margrave Ludwig V pledged Łagów together with the adjacent property to the Knights of the Order of St. John and failed to repay the loan. In 1350 CE, he conferred on the Order the ownership of the manor house and town in Łagów and 21 other settlements. The Joanittes were an international order. Encouraging European elites to make pilgrimages to Jerusalem and support Christianity in Palestine, they also founded their houses in Central Europe. They also took possession of their estates in Brandenburg, Pomerania, and the Kingdom of Poland 42 . The Brandenburg Joannites established a commandery in Łagów, i.e., the headquarters of the basic unit of the organizational structure of the Order and built a castle there. In the late Middle Ages, there were 11 villages within a 5-km radius of the site, which became the Order's property at the beginning of the fifteenth century 42 . All these formed a relatively stable settlement network in the following centuries. Joannites' estate near Łagów was organised in a typical central European way. The Joannites performed the role of a feudal landlord, to whom Brandenburgian margraves granted land, and whose role was to modernize it in order to maximise www.nature.com/scientificreports/ the profits from agriculture, in the same manner as it was done in the estates of Brandenburgian, Pomeranian, Prussian and Polish members of nobility 43,44 . The Joannites enlarged the brick castle and the bailey with houses of servants and craftsmen, thus creating a kind of district, next to which the old settlement on the slopes of the Falcon Hill functioned. Nearby villages were a part of the "table estate" of the Order, which means that in addition to providing the income in terms of money, they were supposed to supply the brothers with food. The lifestyle of the Joannites was more of a knight's fraternity than of a church order, and in 1538 AD, they converted to Protestantism in Brandenburg (and in Łagów). They became a secular order and kept their property as an organization 42 . After the cessation of the Order of St. John in 1812, their property was transferred to the Prussian state, and from 1819, it was transferred to various aristocratic families, which at first continued to develop agricultural production in the area and the large manorial farms 45 , but environmental sources also indicate some changes in the trait of human economic activity.
Multi-proxy, high-resolution peat archive: changes in forest and peatland ecosystems. By using the high-quality age-depth model (Fig. 2) and several proxies, we reconstructed three main stages of land-use change linked to peatland ecosystem functioning and structure. The synthetic Fig. 3 shows the synopsis of changes in the last 1500 years (Fig. 3). Individual proxy diagrams for pollen, testate amoebae (TA), and plant macrofossils are available as Supplementary Figures S1, S2, and S3, respectively. From ca. 500 to 1350 CE, herbaceous plants occurred together with Sphagnum. Moreover, Nymphaea alba sclereids were present in the entire phase, indicating open water habitats (Suppl. Fig. S2). TA communities changed abruptly (see Suppl. Fig. S3). The population of Assulina muscorum and Amphitrema wrightianum decreased. Other mixotrophs such as Archerella flavum, Hyalosphenia papilio, and Heleopera sphagni increased, and new species appeared in the communities, especially those typical for very wet habitats, for example, Difflugia globulosa and Centropyxis aculeata 46 . The water table, decreased slightly at ca. 500 cal. CE to ca. 15 cm and increased again reaching the surface of the mire. The reconstructed water table was the highest in the entire profile at ca. 900-1200 CE. Local vegetation was dominated by a dense forest with Pinus sylvestris, Betula, Alnus, and Quercus. Corylus avellana retreated   48 . Because of low frequency/lack of human pollen indicators, we assume that this telmatic stage was an effect of climate influence (Fig. 3). An increase in fire activity at the end of the phase coincides with water table lowering and a slight decrease in AP values (Suppl. Fig. S1). This gradual increase in charcoal and a decrease in forest vegetation can be the first stronger human signal in this record associated with the building of the Polish State.   www.nature.com/scientificreports/ and rapid peat growth (220 cm of peat accumulated over ca. 55 years) were probably related to the formation of a floating mat on the surface of the water basin, and Sphagnum growth was driven by autogenic succession. TA record estimates water table lowering as the reconstructed depth-to-water table (DWT) decreased to ca. 10 cm (Fig. 3 Fig. S3). However, the values of pollen human indicators do not suggest intensive pressure on the peatland. The decrease in C. betulus and Alnus and the increase in fire activity (expressed as an increase in microscopic charcoal influx and more dynamic charcoal curve) suggest that forest was exploited for timber and the wood was probably burned for household maintenance 49 .
After 1800 CE, forest regenerated as indicated by the pollen of F. sylvatica (presently surrounding the mire) and Betula. Moreover, the abundance of cereal pollen increased, which with the lack of coprophilous fungi suggest a change in land management from grazing to croplands. The population density around our site was not high in the early phase. We inferred no sign of extensive deforestation. The mire was surrounded by the forest that locally burned, as shown by the macrocharcoal remains. A conspicuous feature is an increasing trend of local fires that shows 6 peaks (Fig. 3) and then decreased at ca 870 CE. The local fires were most possibly caused by the local Slavic societies, whose economy is described as mixed, with agriculture and husbandry supplemented with hunting. In ca. 1050 CE, the area was incorporated into the Polish State; however, the human impact in this phase was still low. Local fire intensity was low until 1150 CE together with the signs of increasing local deforestation. Microcharcoal suggests an increasing trend in regional fires and through the denser settlement. It is assumed that the population of Terra Lubucensis in the Piast period doubled considering the tribal period but remained at a low level. The increase in the level of rural and proto-urban settlement relates to the annexation of Terra Lubucensis by the March of Brandenburg in the middle of the thirteenth century. This growth is mirrored in accelerating deforestation in ca 1300 CE, which is apparent primarily by the decrease in C. betulus. The decline of this species (and all broadleaved trees) indicates the twilight of pristine hornbeam forest and abrupt land-use change associated with growing pressures from woody biomass extraction related to local economy. For ca 300 years (1050-1349 CE), the area was managed first by the Polish State and then by Margraves of Brandenburg. The latter began intensive colonization; however, the most dramatic nature disturbances appeared in the next centuries.
It is also important to mention that the mesotrophic mire possessed shallow open water conditions that ended together with the increasing landscape openness. Thus, a strong link between the forest structure and the wetland ecosystem has never been documented in such detail. A comparable example is provided by 24 that described increased catchment erosion and nutrient loading that are commonly recognized impacts of deforestation on wetlands. The authors suggest that deforestation increases water tables and leads to lower retention; therefore, reforestation may drastically alter this water balance, and in some cases, the protected wetland will cease to exist. However, by using our paleoecological record, it was already shown that many different scenarios are possible 51 , and deforestation does not always mean higher water table in wetlands 27 , but it may also lead to terrestrialization. In the case of this study, open water disappeared together with clearcutting that disturbed hydrology and transformed soils surrounding the mire. Despite this change, the water table was high enough to sustain the rapid peat growth in the Joannites stage; however, the ecological state of the wetland was completely different. After this date, increasing cereal abundance is visibly associated with the gradual deforestation. Moreover, since ca. 1350 CE, there were no local fires, which was most possibly caused by the lower local wood/fuel availability. Simultaneously, regional fires (represented by microcharcoal) had an anthropogenic origin and were related to regular wood usage, while peaks might represent wars and battles. Abruptly increasing openness implied ecological revolution and a tipping point for the vegetation, local people, and wetland ecosystem. A similar loss of the close-to-pristine nature in the late Middle Ages was inferred in many locations of Europe and the rest of world, and it was caused by different societies/nations 52,53 . Details of each of those studies are very complex as intricate drivers relate to historical and economic transitions. A reliable example is the impact of Polish State development 54 and Teutonic knights' expansion 55    www.nature.com/scientificreports/ usually allocated for the upkeep of local churches and parish priests 58 . The peasants paid rent to the Joannites for the land in their possession, and the Order also kept its manors in Granów and probably in Łagów. From 1600 CE, meadows and pasture forests (with oaks) with some croplands were surrounding the peatland. This kind of agriculture pattern was then sustained by the Joannites through the next 400 years. The visual negative correlation between deforestation events and mire acidification is a remarkable feature, and it might be related to humic acid pulses from the surrounding eroded soils. Even a short event associated with C. betulus regeneration ca. 580-520 BP is reflected in Sphagnum decrease. The presence of Sphagnum appears to be closely associated with the catchment processes, which was probably related to flushes of humic acids from the surrounding eroded soils. This is the first record with so close deforestation-mire response pattern. Over the last 500 years, the mire was most possibly surrounded by the open fields and dispersed Pine forest. Increased abundance of coprophilous fungi spores and cereal pollen grains suggests croplands and grazed meadows. Stable population and settlement structures continued in the area over several hundred years until the nineteenth century, as shown on the maps of David Gily from 1802 to 1803 AD. The moments of crisis over this period, however, are visible in historical sources and in charcoal signal. The change in land-use structure is apparent in the pollen data that revealed afforestation by Betula and F. sylvaticaca. 1800 CE. After that time, the land management changed. The monastery's unified economic policy was replaced by different land management strategies by different owners who took over the monastery's property after the dissolution of the Joannites. Consequently, in the first half of the nineteenth century, the production of local agriculture diminished. For Pawski Ług, the date 1350 CE implies the verge of the loss of pristine forests in CE Europe and entering societies from the forest tribe to the state period and feudal economy. Local deforestations cumulatively had global consequences for the soil's transformation and wetland ecosystems that also suffered from these disturbances, which still possess the legacy effects 24 having the origin from the late Middle Ages. It was also recently described that forest community responses to climate change are most closely related to microclimate change 59 . Past woody biomass extraction in medieval around Łagów affected the local microclimate and finally affected not only forest microclimate but also wetland functioning.
The water table in the mire decreased locally; however, we propose two alternative explanations. The water table decreased considerably in the peatland; therefore, ombrotrophic conditions became prevalent, and the wetland was colonized by Sphagnum or the fen was acidified 60 , which allowed Sphagnum to establish a floating mat that was accumulating in the peat in the stable hydrological conditions. This kind of deforestation-generated acidification was also observed in one peat core in Tuchola Pinewoods (N Poland), where it was associated with deforestation and pine planting that subsequently triggered Sphagnum expansion 61 . In the other case, Sphagnum expansion is related to the drainage and forest management (Pinus sylvestris plantations) in W Poland 62 . The only tree species that increased abundance around the mire, probably through the secondary succession, was Pinus sylvestris. Consequently, the soil experienced progressive soil acidification that also affected Pawski Ług and supported Sphagnum development in the basin. Sphagnum prefers acid conditions to grow, and it is also an important ecosystem engineer that acidifies habitat by itself 63 . However, we are sure that the mire has never been minerotrophic again and became bog until presently. The establishment of Sphagnum was subsequently responsible for the rapid peat accumulation that provided an exceptionally high-resolution archive of the past changes in the last 600 years. Considering stable hydrology, we would rather choose the second explanation with the floating mat development and local acidification. If it was the dry shift, we could expect more pronounced dry phases that were not existing in the profile and the peat was growing fast (2.6 mm/year). Several examples of floating mats with fast PAR were reported for Polish and European peatlands, e.g., Dury Lakes (10 mm/year) 64 , Mukrza (4.6 mm/year) 65 as well as in Central Italy (5-6 mm/year) 66 or Jaczno (1.38 mm/year) 30 . The stable and high ground water table in the floating Sphagnum mat conditions explains the rapid and constant peat growth in Pawski Ług mire.
After 1812 and Joannites' dissolution, a secondary succession of the forest with beech occurred, and intensive agriculture in the region was abandoned. Post-Joannite estates were first taken over by the Prussian state and later by some Prussian aristocratic families. The system of large manor farms, producing for sale, was maintained, but the local economy declined due to the exclusion of Łagów from the plans for new transport routes. Industrialization processes were not established there, and depopulation due to migration to the western parts of Germany affected the region 67 . The change came at the end of the nineteenth century, when Łagów-the former seat of the Joannites-became a tourist resort. The park and game reserve, which had been created by the Joannites, were developed, and finally, in 1909, a railway connection was established.
Our study provides a high-quality reconstruction of the impact of economy on the forest and wetlands. We investigated the past transitions from the primeval forest managed by Slavic tribes to the agriculture under Brandenburgians and Joannites. The novel landscape led to the development of the novel Sphagnum-dominated wetland ecosystem. We suggest that the anthropogenic ecosystems are often regarded as close to pristine. Our study provides the arguments to discuss the timing of the Anthropocene period 5 and emergence of the novel ecosystems 68 .

Methodology
Coring. The 5-m core was sampled in spring 2016 by using a Wardenaar sampler to recover a sample from the top 1 m. The remaining part of the profile was extracted using an INSTORF sampler (1 m long and 80 mm diameter). Because the first meter of the peat was too unconsolidated (soft), it was not possible to retrieve it using any peat corer; hence, we decided to perform the analyses only on the section between 100 and 500 cm. The samples were transported to the laboratory and stored frozen. The material was subsampled from the de-frosted core for each proxy and dated and analyzed at a range of resolutions. Pollen, microscopic charcoal, and TA were sampled every 5 cm, whereas plant macrofossils and macroscopic charcoal were analyzed contiguously every 1 cm.  72 . Historical data on the surrounding area was extracted from the available medieval and early modern written sources (chronicles, inventories, old maps) and historical and archeological studies, which were analyzed by traditional historical, spatial, and statistical methods.

Palynological analysis.
A total of 80 samples (2 cm 3 in volume, sampled every 5 cm) were prepared using standard laboratory procedures for palynological analysis following standard procedures 73 and using Lycopodium marker. Pollen, cryptogam spores, and selected nonpollen palynomorphs (NPPs) were counted under a binocular microscope until the total pollen sum (TPS) in each sample reached at least 500. Pollen grains were identified using atlases and keys [74][75][76] . The results of the palynological analysis were expressed as percentages calculated on the basis of the ratio of an individual taxon to the TPS, i.e., the sum of arboreal pollen (AP) and nonarboreal pollen (NAP) excluding aquatic and wetland plants including Cyperaceae, and cryptogams. For better understanding, some pollen taxa that are human impact indicators were grouped into cultivated land indicators and major ruderals 77-79 . Charcoal analysis. Microscopic charcoal particles (size: > 10 μm) were counted from the same slides as pollen and NPPs 80 until the number of charcoal particles and Lycopodium spores, counted together, exceeded 200 81 . The calculations of microscopic charcoal accumulation rates (MIC) follow the formula proposed by 82 , i.e., MIC = Ct × PAR, where Ct is the concentration of charcoal particles (unit: particles/cm 2 /yr) 80 .
For macroscopic charcoal analysis, 400 contiguous samples (1 cm 3 ) were prepared by bleaching to create a more visible contrast between the charcoal and the remaining organic matter and by wet sieving through a 100μm mesh following the method described by 83 . The particles were divided into 4 fractions, and the entire sample was analyzed under 60× magnification. Macroscopic charcoal influx (proxy for local fires 84 ) or accumulation rates (MAC, particles/cm 2 /year) were calculated using the charcoal concentrations and PAR. www.nature.com/scientificreports/ Plant macrofossils analysis. The macrofossil composition of the 400 contiguous peat samples (volume: ≈3 cm 3 ) was determined by sieving each sample through a 125-μm diameter mesh. Plant remains were later scanned using a binocular microscope (10 × -50 × magnifications) and identified using an extensive reference collection of type material 85 . Volume percentages were estimated for all components except for seeds, Betula spp. catkin scales, Eriophorum vaginatum spindles, Carex spp. nutlets, Chara oospores, and Sphagnum spore capsules, which were counted and expressed as the number (n) present in each subsample.
Testate amoebae analysis. Peat for TA analysis (2 cm 3 in volume) was sampled from the same depths as those for pollen and microscopic charcoal analyses. Peat samples were washed under 0.3-mm sieves following the method described by 86 . TA were analyzed under a light microscope between 200 × and 400 × magnification, with a minimum of 100 tests per sample 87 . Several keys and taxonomic monographs [88][89][90][91][92][93] as well as internet resources 94 were used to achieve the highest possible taxonomic resolution. The results of the TA analysis were used for the quantitative DWT reconstructions 46 .
Statistics and diagrams. Diagrams with paleoecological proxy data were plotted using C2 95 or Tilia graph 96 software. To adjust the description of the proxy data, we divided each diagram into zones based on the pollen and NPP spectra, as these proxies represent the broadest spatial set of environmental changes. To draw the synthesis figure, DataGraph was used 97 . The final graphical output was edited using Affinity Designer (affinity.serif.com). Quantitative reconstruction of the TA-based DWT was performed in C2 software 95 by using a training set developed for northern Poland by 46 . The local training set was used to avoid biogeographical and taxonomic bias 98,99 . Nonmetric Multidimensional Scaling (NMDS) on the Bray-Curtis dissimilarity was applied using the vegan package 100 to explore trajectory of changes over time in vegetation in relation to environment. Selected environmental variables (DWT, MIC, MAC, vascular plants, and Sphagnum) were fitted onto the ordination post hoc with the envfit function. Numerical analyses were performed using R version 3.6.2 101 . www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.