Coastal socio-ecological systems face unprecedented challenges due to climate change, with impacts encompassing long-term, chronic changes and short-term extreme events. These events will impact society in many ways and prompt human responses that are extremely challenging to predict. This dissertation employs complex systems methods of agent-based modeling and machine learning to simulate the interactions between climatic stressors such as increased flooding and extreme weather and socio-economic aspects of coastal human systems. Escalating sea-level rise and intensified flooding has the potential to prompt relocation from flood-prone coastal areas. This can reduce flood exposure but also disconnect people from their homes and communities, sever longstanding social ties, and lower the tax base leading to difficulties in providing government services. Chapter 2 demonstrates a stochastic agent-based model to simulate human relocation influenced by flooding events, particularly focusing on the responses of rural and urban communities in coastal Virginia and Maryland. The findings indicate that a stochastic, bottom-up social system simulator is able to replicate top-down population projections and provide a baseline for assessing the impact of increasingly intense flooding. Chapter 3 leverages this model to assess how incorporating heterogeneity in relocation decisions across socio-economic groups impacts flood-induced relocation patterns. The results demonstrate how this heterogeneity leads to a decrease in low-income households, yet a rise in the proportion of elderly individuals in flood-prone regions by the end of the simulation period. Flood-prone areas also exhibit distinct income clusters at the end of simulation time horizon compared to simulations with a homogenous relocation likelihood. Lastly, Chapter 4 explores relationships between extreme weather and agricultural losses in the Delmarva Peninsula. Existing research on climatic impacts to agriculture largely focuses on changes to major crop yields, providing limited insights into impacts on diverse regional agricultural systems where human management and adaptation play a large role. By comparing various multistep modeling configurations and machine learning techniques, this work demonstrates that machine learning methods can accurately simulate and predict agricultural losses across the complex agricultural landscape that exists on the Delmarva peninsula. The multistep configurations developed in this work are able to address data imbalance and improve models' capacity to classify and estimate damage occurrence, which depends on multiple geographical, seasonal, and climatic factors. Collectively, this work demonstrates the potential for advanced modeling techniques to accurately replicate and simulate the impacts of climate on complex socio-ecological systems, providing insights that can ultimately support coastal adaptation. / Doctor of Philosophy / Coastal areas are facing increasing challenges from climate change, including rising sea levels and extreme weather conditions. This dissertation explores socio-economic consequences of these adverse environmental changes for coastal communities. Disruptive repetitive flooding due to exacerbated rise in sea levels is one of these consequences that may eventually leave some highly exposed coastal communities no alternative but migrating from their residences. Focusing on coastal Virginia and Maryland, Chapter 2 develops a data-informed model that can simulate individual relocation decisions and assess how they impact population changes and migration patterns. Chapter 3 employs this model to investigate how future changes in sea levels affect diverse socio-economic groups, their relocation decisions, and the resulting collective migration flows in flood-prone areas. We found that considering demographic differences leaves highly flood-prone areas with less low-income households, higher elderly individuals, and more economic clusters compared to simulations where these differences are not accounted for. Chapter 4 uses machine learning models to simulate the economic impact of extreme weather events as another manifestation of climate change on the agriculture in the Delmarva Peninsula. Through data-based modeling techniques, we identify the climatic conditions most responsible for agricultural losses and recognize modeling choices that enhance our predictive ability. Collectively, this dissertation demonstrates how sophisticated modeling techniques can be used to better understand the complex ways in which climate change will impact human society, with the ultimate goal of supporting adaptation strategies that can better address these impacts.