Josias Gloy’s research while affiliated with Alfred-Wegener-Institut Helmholtz-Zentrum für Polar- und Meeresforschung and other places

Treelines across the Northern Hemisphere are shifting upward and northward in response to global warming, particularly in boreal forests, where climate change progresses more rapidly at high elevations and latitudes. These shifts intensify competition for resources, threaten endemic alpine species, and disrupt established ecological relationships, leading to biodiversity loss. However, significant heterogeneity and regional variation exist in how treelines respond to environmental changes, with many underlying drivers and constraints still poorly understood. This study aims to enhance understanding of alpine treeline dynamics and improve vegetation model predictions under changing climatic conditions. We evaluated the relative impact of key factors influencing treeline migration velocity and examined the effects of varying snow regimes on treeline migration within the alpine treeline ecotone. To achieve this, we incorporated a novel snow module into the vegetation model LAVESI (Larix Vegetation Simulator), enabling the integration of precipitation outside the growing season, snow accumulation, and snowmelt processes. This module allows for explicit modelling of the positive and negative impacts of snow depth on tree growth and treeline migration, while accounting for stochastically occurring extreme events and capturing full weather variability. Our findings reveal site-specific responses to factors driving treeline shifts and forest expansion, with localised conditions playing a critical role in shaping migration dynamics. The Canadian and the Russian sites demonstrate clear insights into primary migration drivers, while the high variability at the Alaskan site indicates more complex local dynamics and greater predictive uncertainty. The study highlights the significant role of snow in modulating migration potential, as snow accumulation creates favourable conditions for seedling germination and growth while also posing risks of increased mortality from snow loads or avalanches. These results underscore the importance of incorporating snow-related processes into vegetation models to improve the accuracy of predictions for boreal forest dynamics. Overall, this study provides valuable insights into tree migration processes, highlighting the varied predictability of treeline responses across regions. These findings carry significant implications for refining vegetation models and guiding conservation strategies to sustain alpine tundra resilience in the face of accelerating climate change.

Background: Wildfires are recognized as an important ecological component of larch-dominated boreal forests in eastern Siberia. However, long-term fire-vegetation dynamics in this unique environment are poorly understood. Recent paleoecological research suggests that intensifying fire regimes may induce millennial-scale shifts in forest structure and composition. This may, in turn, result in positive feedback on intensifying wildfires and permafrost degradation, apart from threatening human livelihoods. Most common fire-vegetation models do not explicitly include detailed individual-based tree population dynamics, but a focus on patterns of forest structure emerging from interactions among individual trees may provide a beneficial perspective on the impacts of changing fire regimes in eastern Siberia. To simulate these impacts on forest structure at millennial timescales, we apply the individual-based, spatially explicit vegetation model LAVESI-FIRE, expanded with a new fire module. Satellite-based fire observations along with fieldwork data were used to inform the implementation of wildfire occurrence and adjust model parameters. Results: Simulations of annual forest development and wildfire activity at a study site in the Republic of Sakha (Yakutia) since the Last Glacial Maximum (c. 20,000 years BP) highlight the variable impacts of fire regimes on forest structure throughout time. Modeled annual fire probability and subsequent burned area in the Holocene compare well with a local reconstruction of charcoal influx in lake sediments. Wildfires can be followed by different forest regeneration pathways, depending on fire frequency and intensity and the pre-fire forest conditions. We find that medium intensity wildfires at fire return intervals of 50 years or more benefit the dominance of fire-resisting Dahurian larch (Larix gmelinii (Rupr.) Rupr.), while stand-replacing fires tend to enable the establishment of evergreen conifers. Apart from post-fire mortality, wildfires modulate forest development mainly through competition effects and a reduction of the model’s litter layer. Conclusion: With its fine-scale population dynamics, LAVESI-FIRE can serve as a highly localized, spatially explicit tool to understand the long-term impacts of boreal wildfires on forest structure and to better constrain interpretations of paleoecological reconstructions of fire activity.

With changing climate, the boreal forest could potentially migrate north and become threatened by droughts in the south. However, whether larches, the dominant tree species in eastern Siberia, can adapt to novel situations is largely unknown but is crucial for predicting future population dynamics. Exploring variable traits and trait adaptation through inheritance in an individual-based model can improve our understanding and help future projections. We updated the individual-based spatially explicit vegetation model LAVESI (Larix Vegetation Simulator), used for forest predictions in Eastern Siberia, with trait value variation and incorporated inheritance of parental values to their offspring. Forcing the model with both past and future climate projections, we simulated two areas - the expanding northern treeline and a southerly area experiencing drought. While the specific trait of 'seed weight' regulates migration, the abstract 'drought resistance' protects stands. We show that trait variation with inheritance leads to an increase in migration rate (∼ 3% area increase until 2100). The drought resistance simulations show that, under increasing stress, including adaptive traits leads to larger surviving populations (17% of threatened under RCP 4.5 (Representative Concentration Pathway)). We show that with the increase expected under the RCP 8.5 scenario vast areas (80% of the extrapolated area) of larch forest are threatened and could disappear due to drought as adaptation plays only a minor role under strong warming. We conclude that variable traits facilitate the availability of variants under environmental changes. Inheritance allows populations to adapt to environments and promote successful traits, which leads to populations that can spread faster and be more resilient, provided the changes are not too drastic in both time and magnitude. We show that trait variation and inheritance contribute to more accurate models that can improve our understanding of responses of boreal forests to global change.

Boreal forests of Siberia play a relevant role in the global carbon cycle. However, global warming threatens the existence of summergreen larch-dominated ecosystems, likely enabling a transition to evergreen tree taxa with deeper active layers. Complex permafrost–vegetation interactions make it uncertain whether these ecosystems could develop into a carbon source rather than continuing atmospheric carbon sequestration under global warming. Consequently, shedding light on the role of current and future active layer dynamics and the feedbacks with the apparent tree species is crucial to predict boreal forest transition dynamics and thus for aboveground forest biomass and carbon stock developments. Hence, we established a coupled model version amalgamating a one-dimensional permafrost multilayer forest land-surface model (CryoGrid) with LAVESI, an individual-based and spatially explicit forest model for larch species (Larix Mill.), extended for this study by including other relevant Siberian forest species and explicit terrain. Following parameterization, we ran simulations with the coupled version to the near future to 2030 with a mild climate-warming scenario. We focus on three regions covering a gradient of summergreen forests in the east at Spasskaya Pad, mixed summergreen–evergreen forests close to Nyurba, and the warmest area at Lake Khamra in the southeast of Yakutia, Russia. Coupled simulations were run with the newly implemented boreal forest species and compared to runs allowing only one species at a time, as well as to simulations using just LAVESI. Results reveal that the coupled version corrects for overestimation of active layer thickness (ALT) and soil moisture, and large differences in established forests are simulated. We conclude that the coupled version can simulate the complex environment of eastern Siberia by reproducing vegetation patterns, making it an excellent tool to disentangle processes driving boreal forest dynamics.

Boreal forests of Siberia play a relevant role in the global carbon cycle. However, global warming threatens the existence of summergreen larch-dominated ecosystems likely enabling a transition to evergreen tree taxa with deeper active layers. Complex permafrost-vegetation interactions make it uncertain whether these ecosystems could develop into a carbon source rather than continuing atmospheric carbon sequestration under global warming. Consequently, shedding light on the role of current and future active-layer dynamics and the feedbacks with the apparent tree species is crucial to predict boreal forest transition dynamics, and thus for aboveground forest biomass and carbon stock developments. Hence, we established a coupled model version amalgamating a one-dimensional permafrost-multilayer forest land-surface model (CryoGrid), with LAVESI, an individual-based and spatially explicit forest model for larch species (Larix Mill.), extended for this study by including other relevant Siberian forest species and explicit terrain. Following parametrization, we ran simulations with the coupled version to the near future to 2030 with a mild climate-warming scenario. We focus on three regions, covering a gradient of summergreen forests in the east at Spasskaya Pad to mixed summergreen-evergreen forests close to Nyurba, and the warmest area at Lake Khamra in the south-east of Yakutia, Russia. Coupled simulations were run with the newly implemented boreal forest species and compared to runs allowing only one species at a time, as well as to simulations using just LAVESI. Results reveal that the coupled version corrects for overestimation of active-layer thickness (ALT) and soil moisture and large differences in established forests are simulated. We conclude that the coupled version can simulate the complex environment of central Siberia reproducing vegetation patterns making it an excellent tool to disentangle processes driving boreal forest dynamics.

Wildfires are an essential ecological process, located at the interface between atmosphere, biosphere, and geosphere. Climate-related changes in their appearance and frequency will shape the boreal forest of tomorrow, the largest terrestrial biome responsible for numerous important ecosystem functions. Changing fire regimes could also increase pressure on fire management and become a threat for humans living in Siberia. However, a lack of long-term fire reconstructions complicates the understanding of the main drivers in the larch-dominated forests of eastern Siberia. At the same time, this lack of long-term understanding also aggravates the validation of fire-vegetation models, and thus predictions of future changes of fire regimes in this vital region. Here, we present a new fire module being built for the individual-based, spatially explicit vegetation model LAVESI (Larix Vegetation Simulator). LAVESI is able to simulate fine-scale interactions in individual tree’s life stages and detailed population dynamics, now expanded by the ability of wildfires igniting and damaging biomass. Fire-vegetation simulations were computed around the catchment of Lake Khamra (SW Yakutia), which experienced forest fires in the years 2007 and 2014 according to remote sensing imagery. From the lake, we previously contributed a new, sedimentary charcoal-based fire reconstruction of the late Holocene. Testing the fire module at a current study site, where modern and historic data has already been collected, allows us to improve it, and look into ways in which the fire reconstruction might help inform the model, before eventually scaling it up to cover larger regions. This represents a first step towards a reliable fire-vegetation model, able to predict future impacts of fires on both the forests of eastern Siberia, as well as the humans living there.