Rainfall Variability, Extreme Events, and Climate Information in the Central American Dry Corridor

Abstract

Central America is consistently identified as a global hotspot for future anthropogenic climate change and is already exposed to considerable natural climate variability and frequent extreme events including tropical storms, heavy flooding, and drought. Pervasive inequalities and systemic vulnerabilities exacerbate the impacts of climate variability and change, especially for the millions of families who rely on rainfed agriculture throughout the region. This is particularly true for the Central American Dry Corridor, a region marked by less rainfall and more frequent dry spells in comparison to the rest of Central America. While projections for 21st century drying due to anthropogenic climate change are robust throughout Central America, internal climate variability is expected to dominate precipitation anomalies for the next several decades and very few locations have had significant or consistent changes in precipitation. However, smallholder farmers in the Guatemalan Dry Corridor have reported recent changes in the distribution and intensity of rainfall, making it harder to predict and contributing to crop losses and decreased production of food staples in recent years. Reducing uncertainty in near-term and future climate variability and change is therefore essential for mitigating climate risk and for improving hazard reduction, resource management, and agricultural decision-making throughout the region. This in part hinges on the production and dissemination of accurate and usable climate information. By combining instrumental climate data, smallholder farmers' observations of climate variability, and precipitation forecasts from climate models, my dissertation seeks to reduce some of these uncertainties. Expanding on this objective, I integrate these sources of information to (1) systematically assess how rainfall variability and extreme precipitation events in the Dry Corridor are changing, (2) determine the relevance of climate variability and change for agricultural decision-making within heterogeneous socio-environmental systems, and (3) evaluate the quality, use of, and access to climate information within these systems. I conclude that natural climate variability continues to be a predominant driver of Central American rainfall patterns, but that analyses at smaller scales reveal regional differences and the importance of evaluating an array of climate metrics that capture distinct features of precipitation variability. I also highlight the role of farmers' observations of rainfall variability in agricultural decision-making and how this influence is mediated by other factors. Finally, I show that the technical quality of climate information related to seasonal climate forecasts provides evidence of skill, depending on the region and metric, but that the use and value of this information remains low among smallholder farmers throughout the region. This dissertation underscores the need for multi-scalar, interdisciplinary studies to develop a better understanding of current and future climate patterns and risks in order to support more informed decision-making and adaptation to climate change.