Surface-based temperature inversions and permafrost dynamics in the mountains of northcentral Yukon: linking observations and downscaled climate models

Abstract

Surface-based temperature inversions (SBIs) exert a dominant control on surface air temperature patterns in high-latitude mountain environments. Long-term climate observations are sparse in these regions, so gridded climate datasets are commonly utilized. However, SBIs are poorly represented in these datasets, introducing systematic bias into permafrost models that rely on gridded or downscaled air temperature fields. In this thesis, I investigate the role of SBIs in shaping surface air temperature variability in subarctic Yukon valleys and evaluate how improved representation of these patterns influence permafrost modelling outcomes. Using a dense network of in-situ air temperature sensors along elevational transects, we quantify the spatial and temporal variability of SBIs in dissimilar valley environments. Results demonstrate that strong SBIs in subarctic valleys with strong gradients of warming with elevation near the valley bottom and much weaker gradients above. These inversion-driven patterns are poorly resolved in coarse-resolution reanalysis datasets, leading to systematic biases in air temperature estimates. To address this limitation, inversion-aware bias reduction approaches for downscaling reanalysis air temperatures are evaluated and developed. A physically informed model (DReaMIT) is introduced to improve transferability. Results show that incorporating SBI processes substantially improves representation of air temperature patterns and reduces model error. Finally, the influence of SBI-aware climate forcing on permafrost and ground thermal regimes is assessed using the Northern Ecosystem Soil (NEST) model. Results demonstrate that SBI-driven temperature adjustments can significantly alter predicted ground temperatures and active layer thickness. While snow redistribution remains the dominant control on local variability, SBI-related air temperature adjustments can exceed its influence at some sites. Overall, this thesis highlights the critical importance of representing SBIs in permafrost modelling and provides new methods for reducing SBI-related bias in climate forcing data in high-latitude complex terrain, improving confidence in predictions of permafrost change in high-latitude mountain environments.