Narratives of an uncertain future

xtreme weather events have shaped societies throughout history, and their risks are being altered by global warming. To prepare and adapt, society needs relevant climate information. Storylines—physically consistent narratives of historical events—provide a way to connect extreme events with their causes, impacts, and future alternatives under plausible scenarios. While traditionally focused on meteorological processes, storylines must address impacts and adaptation to benefit stakeholders. This thesis explores storyline methods for impact assessment and adaptation planning. Chapter 2 examines crop failures in the United States, using the 2012 extreme event as a reference. We identify two types of future analogues: events with similar meteorological conditions and those with similar impacts. Machine learning is applied to a large climate projection ensemble, showing that while meteorological analogues remain rare, impact analogues increase substantially under global warming. This highlights the added value of impact modeling in estimating crop failures beyond meteorological conditions. In Chapter 3, we extend the analysis to simultaneous soybean failures in the US, Brazil, and Argentina based on the 2012 event. Using a hybrid model, we link meteorology to crop production and explore adaptation scenarios. Results indicate that changes in mean climate drive simultaneous failures, suggesting adaptation to mean climate can reduce risk. However, local increases in extreme weather still elevate failures, underlining the regional importance of extremes. Chapter 4 shifts focus to flood impacts, using Tropical Cyclone Sandy (2012) and its compound coastal flooding impacts in New York City as a case study. The storyline approach explores alternative realizations of Sandy under different scenarios, including climate change and variability in cyclone landfall. We find that 1~m of sea level rise increases flood volumes by a factor of 4.2, while inland precipitation from varied landfall positions increases flood volumes 2.5-fold. Impacts on critical infrastructure depend on whether events are precipitation- or surge-dominated, highlighting the method's potential to assess risks under changing conditions. In Chapter 5, we apply a similar method to Cyclone Idai (2019) in Beira, Mozambique, and evaluate adaptation strategies. Scenarios combining global warming, sea level rise, and spring tides show drastic increases in exposure and damages—up to 37-fold for population exposure and 56-fold for economic losses. Adaptation strategies demonstrate their effectiveness: a combined flood protection and accommodation approach reduces impacts by 80%, while a seawall reduces them by only 10%. Including adaptation in storylines enables the assessment of strategy effectiveness and residual impacts, aiding decision-making for coastal adaptation. This thesis demonstrates that meteorological storylines can quantify impacts, capturing nonlinear relationships between climate and impacts using impact models (statistical, physical, or hybrid). Storylines improve risk assessment by evaluating changes in the frequency and severity of historical extremes under future conditions. Incorporating internal variability and compound events allows for a broader understanding of risks through “what-if” scenarios. Finally, integrating adaptation options into storylines helps quantify their effectiveness and communicate trade-offs, making climate information more actionable. By advancing storyline methods, this work bridges climate science with impact assessment and adaptation planning, offering tools to better inform decision-making under climate change.