Modeling Resilience of the Earth System Under Climate Change Mitigation

Abstract

Human livelihoods rely on the ability of the Earth system to buffer greenhouse gas emissions and recover from externally forced and internal disturbances to a dynamic equilibrium. Anthropogenic climate change and human action challenge this resilience of climate, carbon cycle, and ecosystems to disturbances. Therefore, assessing how they affect the future trajectory of the Earth system is critical for adaptation and mitigation planning. Comprehensive Earth system projections are foundational for these assessments. However, they mostly neglect the explicit effects of anthropogenic interventions for stabilizing global climate like artificial carbon dioxide (CO2) removal and deliberate stratospheric aerosol injections. The projections also rarely account for impacts of varying volcanic aerosol levels on climate and carbon cycle. Their accuracy in simulating repercussions of climate extremes on vegetation productivity, an essential carbon sink, also remains unclear. These shortcomings compromise the reliability of long-term risk assessments. This dissertation, based on three manuscripts, explores ways to better represent vegetation resilience, disturbances from volcanic eruptions, and human interventions in projections of climate change mitigation scenarios. To this end, the work includes model development, statistical emulation, and comparisons of models and observations. As part of this work, Earth system model simulations with an explicitly represented, renewable-powered approach for CO2 removal and conversion corroborate the absence of unintended side effects that would substantially offset its removal potential. However, the results also reveal fundamental land use constraints that could cause target conflicts between large-scale CO2 removal, agriculture, and ecosystem services. Further, model emulations with varying volcanic aerosol levels demonstrate that representing this intermittent forcing realistically increases global and regional temperature uncertainty relative to a constant approach commonly applied in climate change assessments. The results suggest a substantial share of this uncertainty propagates to natural carbon sinks. This significantly increases uncertainty about when planetary temperatures may stabilize, and net-zero emissions may be reached, making it vital to better account for varying aerosol levels in future risk assessments. Finally, comparing statistical relationships between temperature--moisture and vegetation productivity extremes in observations and two Earth system models provides confidence in modeled vegetation resilience at regional to global scales. Both models perform worse locally and struggle to match the observed timing between hydroclimatic drivers and vegetation responses. Recalibrating and improving process models could reduce these discrepancies which may be caused by high biases in soil moisture rates of change. Findings from the studies are limited by uncertain and ambiguous scenarios of socioeconomic development and future reliance on deliberate interventions, as well as model shortcomings. Future work could expand explicit modeling to more CO2 removal approaches and explore their interactions within a portfolio of measures. Additional research on simulated soil moisture and carbon cycling is needed to address the observed biases. Overall, the insights from this work advance modeling capabilities for understanding impacts of human action and natural variability on our living environment. These advances expand the range of future scenarios that can be examined, and support the development of next-generation Earth system models and climate change mitigation strategies.