Effects of climate change on potential range shifts and composition of Türkiye's terrestrial ecosystems

Abstract

At the current critical juncture with climate change, centennial projections of potential species distributions in biodiversity hotspots, using dynamic vegetation models with integrated ecological notions such as competition, may provide vital insight into conservation efforts. The central aim of this thesis research is to estimate the current and future potential distribution of Türkiye's forests and their composition in the absence of anthropogenic landcover change, to calibrate the bioclimatic thresholds of certain key taxa to reflect their regional distribution in simulations using a process-based dynamic regional-to-global vegetation model (DGVM) to introduce new taxa to the simulations where necessary and in this manner to contribute to past studies on the distribution of European forest taxa by filling an important spatial gap. For this purpose, potential distribution of 25 forest taxa, and a C3-grass type are simulated by means of LPJ-GUESS v. 4.1 for the Anatolian Peninsula, keeping a spatial window large enough to track temporal changes in the vegetation range and composition in its surrounding areas. The model was forced with ERA5-Land reanalysis data for the historical period (1961-2014), and five different global climate model (GCM) contributions to the 6th phase of the Coupled Model Intercomparison Project (CMIP6) for the scenario period (2015-2100). The GCM datasets were interpolated to ERA5-Land's resolution and bias corrected with the climatology of the reference period 1995-2014 from the ERA5-Land dataset. The reference period for bias correction was selected in accordance with the 6th assessment report of Intergovernmental Panel on Climate Change (IPCC 2021), and in part, to reflect the current state of global warming. Shared Socioeconomic Pathways (SSP) 5-8.5, the high emissions scenario, was selected for all simulations. The bioclimatic tolerance levels of all plant functional types were calibrated using the temperature components of 1 km CHELSA dataset, and 9 km ERA5-Land datasets, calculating climatologies where necessary using Climate data operators (CDO) and R computing language; species distribution maps from European forest genetic resources programme (EUFORGEN); and the forest inventory data from General Directorate of Forestry (GDF) of Türkiye. In situ site visits were also undertaken to critical regions within the study area to further finetune the bioclimatic tolerance levels of certain key taxa, and further analysis of both observation data and the simulation results were made using R and ArcGIS. The simulation results from multiple model runs highlighted a consistent increase in woody species biomass for the entire study region, by the end of the century. The increase in biomass, mainly in the grasslands ecosystems was an important outcome of these series of simulations, which highlighted a consistent encroachment of woody taxa into terrain that is otherwise dominated by grasslands in simulations, in part reigniting the discussion over the past anthropogenic influence on the current vegetation composition of the Peninsula. The mountain systems also came out as important potential refugia for many cold-favoring woody taxa. Throughout the centennial simulations, the tree species prioritized an altitudinal move rather than a latitudinal one, consolidating their ranges around high elevation massifs, occasionally giving up their density dominance but gaining biomass. The composition of forest and woodland ecosystems shifted in large part in favor of the temperate taxa, as these species expanded their distribution ranges throughout the study area, without any particular prioritization for altitude or latitude. Temperate taxa with higher tolerance for drought increased their overall biomass and density, while the drought sensitive species gave up their dominance but continued to exist in smaller groups in areas where resource limitation became a determinant for the outcome of competition. All simulated taxa survived the 140-year simulation period at some capacity (there were no potential regional extinctions), finding some form of a footing along this ancient land bridge – some giving up their largely monotypic/homogeneous forest ranges for mixed forest compositions; others either leaving their dominant presence and becoming a "member of the team"; or exerting their dominance over a larger area. Overall, the results from these series of simulations, sans human presence, suggest a high potential for future forest cover throughout the study region by the end of the century under a high emissions scenario, and important changes in vegetation composition including encroachment of grasslands ecosystems by woody taxa. Our findings also corroborate recent paleo studies that highlight the early forest capacity of Asia Minor, and with the right environmental policies in place to ensure a fair human-nature co-existence, the region may continue to play its role as a cradle.