Authors: Anne Lausier*, University of Maine
Topics: Environment, Hazards, Risks, and Disasters, Water Resources and Hydrology
Keywords: Precipitation, Quantile Regression, Agriculture, Risk
Session Type: Poster
Start / End Time: 1:10 PM / 2:50 PM
Room: Lincoln 2, Marriott, Exhibition Level
Presentation File: No File Uploaded
Numerous human and environmental systems are sensitive to the spatial and temporal distribution of precipitation, including agriculture, water supply, and ecosystems. Trends in observed precipitation form an important line of evidence to understand how changes may increase system vulnerabilities. Linear trends reported in US and global climate assessments reflect changes in mean annual precipitation. Mean trends may not reflect changes across other quantiles in the precipitation probability distribution, including the tails (very high and low precipitation levels), leading to the systematic mischaracterization of climate risk. Here we reanalyze global annual precipitation using quantile regression to reveal overlooked trends. We find trends in the tails inconsistent with the mean in 44.4% of land area coinciding with 38.4% of the global population. A case example of agriculture shows that 40.7% of rainfed agricultural regions are exposed to trends undetected by the mean. Previously undetected trends offer a more accurate view of a changing climate, while underscoring how risk may be systematically mischaracterized across a range of regions and sectors. This work enables the reappraisal of risk aggregated over thresholds in human and environmental systems, supporting a revaluation of threats and identification of appropriate adaptation strategies.