ALBERT

Description: Most northern peatlands developed during the Holocene, sequestering large amounts of carbon in terrestrial ecosystems. However, recent syntheses have highlighted the gaps in our understanding of peatland carbon accumulation. Assessments of the long-term carbon accumulation rate and possible warming driven changes in these accumulation rates can therefore benefit from process-based modelling studies. We employed an individual- and patch-based dynamic global ecosystem model with dynamic peatland and permafrost functionality and vegetation dynamics to quantify long-term carbon accumulation rates and to assess the effects of historical and projected climate change on peatland carbon balances across the pan-Arctic. Our results are broadly consistent with published regional and global carbon accumulation estimates. A majority of modelled peatland sites in Scandinavia, Europe, Russia and Central and eastern Canada change from carbon sinks through the Holocene to potential carbon sources in the coming century. In contrast, the carbon sink capacity of modelled sites in Siberia, Far East Russia, Alaska and western and northern Canada was predicted to increase in the coming century. The greatest changes were evident in eastern Siberia, northwest Canada and in Alaska, where peat production, from being hampered by permafrost and low productivity due the cold climate in these regions in the past, was simulated to increase greatly due to warming, wetter climate and greater CO2 levels by the year 2100. In contrast, our model predicts that sites that are expected to experience reduced precipitation rates and are currently permafrost free will lose more carbon in the future.

Description: Dynamic global vegetation models (DGVMs) are designed for the study of past, present and future vegetation patterns together with associated biogeochemical cycles and climate feedbacks. However, current DGVMs lack functionality for the representation of peatlands, an important store of carbon at high latitudes. We demonstrate a new implementation of peatland dynamics in a customised "Arctic" version of the dynamic vegetation model LPJ-GUESS, simulating the long-term evolution of selected northern peatland ecosystems and assessing the effect of changing climate on peatland carbon balance. Our approach employs a dynamic multi-layer soil with representation of freeze-thaw processes and litter inputs from a dynamically-varying mixture of the main peatland plant functional types; mosses, dwarf shrubs and graminoids. The model was calibrated and tested for a sub-arctic mire in Stordalen, Sweden, and validated at a temperate bog site in Mer Bleue, Canada. A regional evaluation of simulated carbon fluxes, hydrology and vegetation dynamics encompassed additional locations spread across Scandinavia. Simulated peat accumulation was found to be generally consistent with published data and the model was able to capture reported long-term vegetation dynamics, water table position and carbon fluxes. A series of sensitivity experiments were carried out to investigate the vulnerability of high latitude peatlands to climate change. We found that the Stordalen mire may be expected to sequester more carbon in the first half of the 21st century due to milder and wetter climate conditions, a longer growing season, and CO2 fertilization effect, turning into a carbon source after mid-century because of higher decomposition rates in response to warming soils.

Description: This dataset corresponds to the pollen-based REVEALS estimates of 25 plant taxa for Europe and associated standard errors, published in Marquer et al. (2017). This is part of the results from the Swedish project LandClim I (Gaillard et al., 2010; Trondman et al., 2015; Marquer et al., 2014, 2017). The study area includes a large part of northern and Central Europe, i.e. Ireland, Great Britain and a latitudinal transect from the Alps in the south to northernmost Norway. These REVEALS estimates are based on 151 pollen records (small/large, lakes/bogs/mires) that were selected from the European Pollen Database (Fyfe et al., 2009; Giesecke et al., 2014), the Alpine Palynological Data-Base (University of Bern, Switzerland), or were provided directly by individual data contributors. The selected pollen records are grouped into 36 1° x 1° grid-cells. Twenty-five consecutive time windows over the last 11,700 years BP are used: 0-100, 100-350, 350-700 BP for the three first time windows, and 500 calendar years each from 700 to 11,700 BP. For details about the REVEALS model, see Sugita (2007). In the excel file, the folder "Metadata" contains the explanation of abbreviations in the data folders and information about the pollen records used for the REVEALS reconstructions. All REVEALS estimates and their SEs are given in proportions of the grid cell (the total of all REVEALS estimates sum up to 1). The codes of the 25 consecutive time windows are given in the folder "Code time windows". The results of the 36 grid cells are in the folder "REVEALS 36GCs" and the related standard errors in the folder "SE_REVEALS 36GCs". Note that in the folder "Metadata", the GPS coordinates correspond to the upper left (NW) corners of each grid cell.

Description: Matlab script file of a two-dimensional (2-D) peat microtopographical model together with other supplementary files that are required to run the model.