Résumé
Aim: Global forests face increasing stresses from novel climate states, altered disturbance regimes and the spread of pests and pathogens. Understanding where and when mortality occurs across species' ranges—and identifying associated mortality agents—can improve predictions of forest shifts, inform management and enhance vegetation model carbon cycle components. Location USA. Time Period 2003–2023. Taxa Studied Pinus elliottii, Pinus palustris and Pinus taeda. Methods: Using inventory data on 129,970 trees from 14,517 forest plots censused two to five times, recorded agents of mortality (competition, weather events, fire, animals, insects and pathogens), and applied Bayesian size-dependent survival models, multinomial regressions and spatial eigenvector-based analyses to assess spatial and temporal variation in the survival of three major US pine species. Results Species varied ontogenetically, spatially and temporally in survival. Most interspecific differences occurred in saplings and small trees (< 10 cm diameter), and near species' maximum stature. Pinus taeda showed a marked decline in survival across the natural populations over the last two decades. Spatially, survival patterns were significant for all three species; P. taeda showed consistent spatial structuring at all three spatially examined resolutions, with mortality associated primarily with weather, insects and competition. Pinus elliottii and P. palustris showed spatial structures at one resolution each. As expected, competition dominated in low-mortality areas, consistent with background mortality from succession and gap-phase dynamics. High mortality patches were dominated by disturbance agents such as weather events and insects. Main Conclusions: Exploring range-wide demography while jointly accounting for ontogenetic, spatial and temporal variations is essential in determining species vulnerability to environmental and anthropogenic changes. Models to predict future trends in mortality need to include mechanisms beyond climate envelopes (such as fire, storms, pests and pathogens) in order to properly capture species' whole-range responses to global change.