A common approach to investigating species' niches is to examine relationships between spatial variation in environmental conditions and contemporary species occurrences, using species distribution models (SDM or niche models). The relationships between past species distributions and environmental variation over time are less commonly explored. One way to examine effects on species changes over time is to use paleo-datasets to parameterize niche models, where the use of temporal variation allows for making more direct links between past species and environmental conditions through records of past changes. We examined the impact of five environmental variables (temperature, incidence of external nutrient input, local [within bog] moisture, incidence of regionally dry periods, and fire activity) on temporal variation in peatland species composition, occurrences, and abundances (Sphagnum, Eriophorum, Carex, and Ericaceous dwarf shrubs) using a high-resolution peat macrofossil paleo-record spanning the last ~10,000 years from the Store Mosse bog (south-central Sweden). Our results showed that species composition was affected by external nutrient input, local moisture conditions and incidence of regionally dry conditions. The presence and abundance of different species groups were mainly affected by external nutrient input and the incidence of regionally dry periods. Moreover, hummock Sphagna benefited from external nutrient input and low moisture, and in one species, warmer temperatures. Intermediate Sphagna from cooler temperatures with no external nutrient input, and hollow Sphagna from cooler temperatures and external nutrient input. Lastly, our results showed that environmental effects differed between the successional stages of the peatland in one case. Overall, the observed species' responses imply that peatland carbon dynamics will shift with future changes in climate. By examining links between climate and species responses of the past, this study demonstrates that the paleo-data approach in SDMs can contribute to a better understanding of the environmental effects influencing species distributions on longer time scales, thereby providing a valuable tool to improve predictions of future climate change effects.

Northern peatlands are terrestrial ecosystems that provide specialized habitats in which biomass production exceeds decomposition, resulting in accumulation of organic matter. Understanding what factors drive species changes in future climate conditions in these systems is of high importance since this has the potential to affect ecosystem functioning and biodiversity, and by extension carbon sequestration. In ecology, a common method for investigating species’ relationships with climate variation, linked with spatial information, is species distribution modelling (SDM). This method typically uses information about current climate conditions tied to locations of species occurrences, forecasting the effects of change on future geographic distributions based on the implicit assumption that temporal variation can be substituted by contemporary spatial variation. 

This assumption might not be met for several reasons, namely (1) species changes often occur over much longer time-scales than the ones involved in contemporary ecology, and therefore (2) responses to climatic changes are time-lagged. Incorporating paleo-records of actual (past) changes in species distributions and climate conditions therefore provides a much more direct way to model species responses to climate change. In this project, a combination of methods from the fields of paleoecology and ecology were employed to create a novel approach to explore species distribution changes over time in boreal peatlands. This was done by first reconstructing the vegetation of two proximal peatlands (Store Mosse and Dala Mosse bogs; Paper I and III) in south-central Sweden, followed by statistical modeling of the species data and climatic parameters over time (obtained from independent paleoclimate data; Paper II and III), creating Temporal Paleo-Species Distribution Models (Temporal Paleo-SDMs). 

Paper I identifies factors driving species changes in Store Mosse bog based on internal (successional steps and biotic interactions) and external (climatic) processes. This study tests the assumption that climate has been the main driver of species change by producing a high-resolution postglacial vegetation reconstruction using macrofossil analysis, which is assessed against a set of independent proxy records representing changes in local and regional hydrology, nutrient input, and temperature. Paper II uses the same high-resolution plant macrofossil dataset from Store Mosse and pairs this with independent information about local and regional climate conditions, nutrient input and fire incidence during the same period to create the first Temporal Paleo-SDM and thereby assess the relationships between bog species and climate variability over time (reaching ~10 000 cal yr BP). Paper III tests the repeatability of the Temporal Paleo-SDM method by applying it to a new high-resolution species dataset from Dala Mosse, using the same climate parameters as in Paper II. This thesis bridges across paleoecology and ecology and shows the power of interdisciplinary collaborations and demonstrates the useful contributions they can make in future peatland research. 

Substantial uncertainties exist regarding how future climate change will affect storminess (storm frequency and intensity) in Ireland and the United Kingdom (UK). Knowledge about spatiotemporal variations of past storminess gives us a better understanding of its mechanisms on centennial to millennial time scales, as well as the impact of external forcing on future storminess in climate models. Here, we present the oldest storm record to date from Ireland, covering the last 8000 years, reconstructed from the Roycarter Bog, a coastal blanket bog in north-western Ireland. The sequence was analysed for grain-size, chemical, mineral and organic molecular composition. The chronology was built on 11 AMS radiocarbon dates. The deposit characteristics, location and low inorganic content suggest aeolian transport of particles to the bog throughout the studied period. Cluster analysis of the grain-size frequency curves, along with the coarse to fine sand ratio, allowed the identification of eleven storm periods (cal yr BP): 6150–5500 (1); 4970–4130 (2); 4000 (3); 3490–3290 (4); 3230 (5); 2850–2590 (6); 2170–1920 (7); 1440 (8); 1225–890 (9); 620–470 (10); and 290–230 (11). During the mid-Holocene, the relative sea level was lower and the local beach sources located further away, giving a longer transport distance compared to the late Holocene. In the latter part of the mid-Holocene (6150–4130 cal yr BP), during the Holocene thermal maximum, increased storminess and wind strengths were inferred for north-western Ireland, manifested as two longer storm periods. During the late Holocene the storm frequency increased, and a greater number (9) of shorter storm periods were recorded. Comparison between our results and regional peat palaeostorm records from Scotland, north of our study site, showed an antiphase relationship between storminess in Ireland and Scotland during the latter part of the mid-Holocene, but mostly in-phase storminess over the last 3000 years. Taken together, enhanced wind strength and storminess were recorded during the warmer mid-Holocene, while an increased frequency of storm events occurred in the cooler late Holocene. Mid-Holocene storm periods occurred during locally wet periods, while most of the storm periods during late Holocene occurred during drier phases. Alternatively, the elevated mineral input during late Holocene promoted microbial activity and peat decomposition. The apparent variability in cyclicity and frequency between the mid- and late Holocene indicates that the processes governing storminess in the region shifted. This calls for further studies ahead, including climate modelling, to disentangle the complex processes governing storminess on millennial to centennial time scale.

Northern peatlands store globally-important amounts of carbon in the form of partly decomposed plant detritus. Drying associated with climate and land-use change may lead to increased fire frequency and severity in peatlands and the rapid loss of carbon to the atmosphere. However, our understanding of the patterns and drivers of peatland burning on an appropriate decadal to millennial timescale relies heavily on individual site-based reconstructions. For the first time, we synthesise peatland macrocharcoal records from across North America, Europe, and Patagonia to reveal regional variation in peatland burning during the Holocene. We used an existing database of proximal sedimentary charcoal to represent regional burning trends in the wider landscape for each region. Long-term trends in peatland burning appear to be largely climate driven, with human activities likely having an increasing influence in the late Holocene. Warmer conditions during the Holocene Thermal Maximum (∼9–6 cal. ka BP) were associated with greater peatland burning in North America's Atlantic coast, southern Scandinavia and the Baltics, and Patagonia. Since the Little Ice Age, peatland burning has declined across North America and in some areas of Europe. This decline is mirrored by a decrease in wider landscape burning in some, but not all sub-regions, linked to fire-suppression policies, and landscape fragmentation caused by agricultural expansion. Peatlands demonstrate lower susceptibility to burning than the wider landscape in several instances, probably because of autogenic processes that maintain high levels of near-surface wetness even during drought. Nonetheless, widespread drying and degradation of peatlands, particularly in Europe, has likely increased their vulnerability to burning in recent centuries. Consequently, peatland restoration efforts are important to mitigate the risk of peatland fire under a changing climate. Finally, we make recommendations for future research to improve our understanding of the controls on peatland fires.

Boreal peatlands are facing significant changes in response to a warming climate. Sphagnum mosses are key species in these ecosystems and contribute substantially to carbon sequestration. Understanding the factors driving vegetation changes on longer time scales is therefore of high importance, yet challenging since species changes are typically affected by a range of internal and external processes acting simultaneously within the system. This study presents a high-resolution macrofossil analysis of a peat core from Store Mosse bog (south-central Sweden), dating back to nearly 10 000 cal. a BP. The aim is to identify factors driving species changes on multidecadal to millennial timescales considering internal autogenic, internal biotic and external allogenic processes. A set of independent proxy data was used as a comparison framework to estimate changes in the bog and regional effective humidity, nutrient input and cold periods. We found that Store Mosse largely follows the expected successional pathway for a boreal peatland (i.e. lake -> fen -> bog). However, the system has also been affected by other interlinked factors. Of interest, we note that external nutrient input (originating from dust deposition and climate processes) has had a negative effect on Sphagnum while favouring vascular plants, and increased fire activity (driven by allogenic and autogenic factors) typically caused post-fire, floristic wet shifts. These effects interactively caused a floristic reversal and near disappearance of a once-established Sphagnum community, during which climate acted as an indirect driver. Overall, this study highlights that the factors driving vegetation change within the peatland are multiple and complex. Consideration of the role of interlinked factors on Sphagnum is crucial for an improved understanding of the drivers of species change on short- and long-term scales.

Store Mosse (the ‘Great Bog’ in Swedish) is one of the most extensive bog complexes in southern Sweden (~77 km²), where pioneering palaeoenvironmental research has been carried out since the early 20th century. This includes, for example, vegetation changes, carbon and nitrogen dynamics, peat decomposition, atmospheric metal pollution, mineral dust deposition, dendrochronology, and tephrochronology. Even though organic matter (OM) represents the bulk of the peat mass and its compositional change has the potential to provide crucial ecological information on bog responses to environmental factors, peat OM molecular composition has not been addressed in detail. Here, a 568-cm-deep peat sequence was studied at high resolution, by attenuated reflectance Fourier-transform infrared spectroscopy (FTIR-ATR) in the mid-infrared region (4000–400 cm^–1). Principal components analysis was performed on selected absorbances and change-point modelling was applied to the records to determine the timing of changes. Four components accounted for peat composition: (i) depletion/accumulation of labile (i.e. carbohydrates) and recalcitrant (i.e. lignin and other aromatics, aliphatics, organic acids and some N compounds) compounds, due to peat decomposition; (ii) variations in N compounds and carbohydrates; (iii) residual variation of lignin and organic acids; and (iv) residual variation of aliphatic structures. Peat decomposition showed two main patterns: a long-term trend highly correlated to peat age (r = 0.87), and a short-term trend, which showed five main phases of increased decomposition (at ~8.4–8.1, ~7.0–5.6, ~3.5–3.1, ~2.7–2.1 and ~1.6–1.3 ka) – mostly corresponding to drier climate and its effect on bog hydrology. The high peat accumulation event (~5.6–3.9 ka), described in earlier studies, is characterized by the lowest degree of peat decomposition of the whole record. Given that FTIR-ATR is a quick, non-destructive, cost-effective technique, our results indicate that it can be applied in a systematic way (including multicore studies) to peat research and provide relevant information on the evolution of peatlands.

We present a high-resolution peat paleodust and accumulation rate record spanning the last 8300 years from Draftinge Mosse (400 ha), southern Sweden (57 degrees 06'27.6 '' N 13 degrees 42'54.1 '' E). The record was analysed for peat accumulation rates (PAR), elemental concentrations, mineralogy, and plant macrofossil content. Five periods of increased mineral deposition were recorded. The first event occurred between similar to 6280 and similar to 5570 cal BP, during the fen to bog transition. This is followed by four atmospheric mineral dust events (DE) which were recorded in the ombrotrophic section of the sequence at (cal BP): similar to 2200; similar to 1385-1150; similar to 830-590, and from similar to 420 to the present. Statistical analysis and elemental ratios indicated that both the mineralogy and grain size shifted when the system transitioned from fen into bog, showing that the governing transport process shifted with the peat-land succession stages. This highlights the importance of identifying peatland succession stages within peat paleodust studies. Following all four DE, increases in PAR were observed, implying a coupling to dust deposition. Comparison of DE and PAR with a paleodust record from Store Mosse, a 20 times larger bog located ca 18 km away (Kylander et al. 2016), showed that both PAR and dust deposition are largely represented by single-core reconstructions, indicating that they are driven by a common climate forcing mechanism. However, higher PAR and dust deposition rates were observed in the more moderately sized Draftinge Mosse, suggesting that the size of the bog is important to consider in peat paleodust studies. Furthermore, the smaller bog responded more rapidly to hydrological changes, indicating that the size of the bog affects its' buffering capacity. Authigenic carbonates, observed here during episodes of rapid peat growth, coincide with changes in REE ratios, indicating that authigenic peat processes potentially cause REE fractionation.