ContextThe southeastern U.S. experiences tornadoes and severe thunderstorms that can cause economic and ecological damage to forest stands resulting in loss of timber, reduction in short-term carbon sequestration, and increasing forest pests and pathogens. ObjectivesThis project sought to determine landscape-scale patterns of recurring wind damages and their relationships to topographic attributes, overall climatic patterns and soil characteristics in southeastern forests. MethodsWe assembled post-damage assessment data collected since 2012 by the National Oceanic and Atmospheric Administration (NOAA). We utilized a regularized Generalized Additive Model (GAM) framework to identify and select influencing topographic, soil and climate variables and to discriminate between damage levels (broken branches, uprooting, or trunk breakage). Further, we applied a multinomial GAM utilizing the identified variables to generate predictions and interpolated the results to create predictive maps for tree damage. ResultsTerrain characteristics of slope and valley depth, soil characteristics including erodibility factor and bedrock depth, and climatic variables including temperatures and precipitation levels contributed to damage severity for pine trees. In contrast, valley depth and soil pH, along with climactic variables of isothermality and temperature contributed to damage severity for hardwood trees. Areas in the mid-south from Mississippi to Alabama, and portions of central Arkansas and Oklahoma showed increased probabilities of more severe levels of tree damage. ConclusionsOur project identified important soil and climatic predictors of tree damage levels, and areas in the southeastern U.S. that are at greater risk of severe wind damage, with management implications under continuing climate change.