Improved understanding of the physical impacts to the climate caused by anthropogenic emissions has led to an increased focus on the role of policy on limiting mitigating the cause and the impacts of climate change. The emissions trajectory that society takes will play a major role in determining a wide variety of climate impacts from health to economic growth. This dissertation explores how idealized modeling can be used to inform the potential outcomes of the policies meant to mitigate the negative impacts of climate change and optimize the benefits of emission reduction. Chapter 1 explores the potential trade-offs between increased temperatures and reduced health impacts of considering aerosol emissions in international climate policy. Using an idealized model, it is possible to demonstrate that both health and temperature benefits can result from considering either health or temperature separately. Chapter 2 interrogates whether climate policy in the United States can be used to reduce existing racial inequities in air pollution exposure. The modeling shows that prioritizing by exclusively income or cost in national policy would result in worsening inequities, but that prioritizing by race explicitly has the opposite effect. Using a set of more realistic air quality simulations with randomized emission reduction, it appears that the transportation has a unique potential for air quality equity improvements. Lastly, Chapter 3 focuses on statistical models that use natural variation in annual weather to predict how climate will influence human outcomes. Using an idealized outcome variable and real data of global gross domestic product, it is possible to demonstrate that imperfect measurement networks and the mean shift in climate both work to underestimate future climate impacts.

Racial and ethnic minorities in the United States often experience higher-than-average exposures to air pollution. However, the relative contribution of embedded institutional biases to these disparities can be difficult to disentangle from physical environmental drivers, socioeconomic status, and cultural or other factors that are correlated with exposures under status quo conditions. Over the spring and summer of 2020, rapid and sweeping COVID-19 shelter-in-place orders around the world created large perturbations to local and regional economic activity that resulted in observable changes in air pollution concentrations, compositions, and distributions. Here, we use the pandemic-related emergency order and subsequent economic slowdown to causally estimate pollution exposure disparities in California. Using both public ground-based sensor data and a citizen-science network of monitors for respirable particulate matter (PM2.5), along with satellite records of nitrogen dioxide (NO2), we show that the initial sheltering-in-place period produced disproportionate air pollution reduction benefits for Asian, Hispanic/Latinx, and low- income communities. By linking these pollution data with weather, geographic, socioeconomic, and mobility data in difference-in-differences models, we demonstrate that these disparate pollution reductions cannot be explained by environmental conditions, geography, income, or local economic activity and are instead driven by non-local activity. This study thus provides causally-identified evidence of systemic racial and ethnic bias in pollution control under business-as-usual conditions.