Vegetation Distribution and Biophysics, Permafrost, and Feedback Mechanisms to Arctic Change

Abstract

The Arctic terrestrial ecosystems have been changing in response to global warming over the past decades. Studies reported shrub and tree expansion into tundra ecosystems. This process will have feedback on climate change through changes in albedo, evapotranspiration, and the carbon cycle. However, vegetation is not shifting uniformly towards the tundra area. Factors that control this encroachment need to be identified and the underlying mechanism of their interaction with vegetation distribution patterns needs to be explained at a regional and local scale. Meanwhile, historical field observations of ground soil temperatures have shown a general warming trend across the Arctic permafrost. The ongoing warming process of the permafrost strongly affects the regional ecosystems, infrastructure, and indigenous communities in the Arctic. However, the knowledge about the surface ground heat flux, a key factor that controls the subsurface thermal dynamics, is still limited in the cold region. Projections of future climate and subsurface thermal state are conducted based on model simulations with significant uncertainties yet to be addressed before such results can be used to make any conclusion. It is important to solve these issues first to understand the potential Arctic change in the future and infer its feedback to the global climate system.

In this dissertation, these processes have been investigated using field measurements and model results, with a focus on study sites located in Northwestern Siberia. The relationships between spatiotemporal patterns of the encroaching tall vegetation to the microtopography and snow cover are demonstrated. Preference for tree growth locations has been shown for surfaces with divergence characteristics (well-drained). The change of distribution patterns of tree locations is related to the snow distribution shifts by applying clustering analysis.

Data sets for surface ground heat flux GS are sparse in the Arctic region from observation due to challenges in field measurements, and not available from the Global Climate Model (GCM) direct outputs. An uncertainty-informed framework is developed by combining the analytical model and the state-of-art uncertainty quantification (UQ) machinery to reconstruct GS from field measurements and GCM outputs of shallow soil temperatures. The reconstructed GS and its probability distribution are valuable to be used as the boundary forcing for subsurface thermal regime simulations.

Climate projections from GCMs are retrieved to reconstruct the GS for the future in the study sites. A Bayesian Weighted Averaging (BWA) stochastic downscaling method is applied to reduce the bias and uncertainty carried in the GCMs with historical borehole observations. The long-term ground soil temperature is simulated by a physical numerical model calibrated by historical observations until the end of this century using the downscaled annual mean GS as the top boundary control. The projected surface soil temperature is higher than the global average at two of the borehole sites for the selected future climate scenarios. Almost one-seventh of the current permafrost will disappear at one borehole site under the worst carbon emission scenario, with other cases also showing some level of permafrost loss. These will in turn have a huge impact on global climate through carbon release, destabilize the infrastructure foundation, and change the surface and subsurface hydrology and thermal regime. The results stress the urgent need to control the global warming trend. The study of the above-ground vegetation dynamics and below-ground permafrost cryospheric processes all together provides insights to better understand the Arctic change in the future.