COS 284-3 - Climate sensitivity and memory differ between trees that survive and trees that die during drought

Abstract: Increases in drought frequency and severity are causing large-scale forest dieback across the Southwest and globally. Climatic conditions of past months or years may be just as important as current conditions for tree growth under repetitive droughts. Therefore, the primary goal of this study was to understand how drought stress can alter the timescale and magnitude of tree growth responses to climate, possibly revealing early warning signs of drought-related mortality. We leveraged existing data on ring-widths for three widespread tree species in northern Arizona (quaking aspen, Populus tremuloides; Douglas fir, Pseudotsuga menziesii; white fir, Abies concolor) to compare climate sensitivity and memory of drought-surviving trees with those that died. We focused on the effects of temperature and precipitation during different seasons (growing vs dormant seasons). We fit a hierarchical Bayesian mixed model using 577 trees across 22 sites that incorporated the effects of these different seasonal climate variables and their interactions, and we estimated the relative importance of these variables at different annual lags (current year up to 4 years prior to growth). We found site- and species-specific differences in climate sensitivity and memory between trees that died and those that survived. However, across all three species, we found that drought-surviving trees generally had weaker growth responses to the interaction between winter precipitation and winter temperature compared with trees that died; this 2-way interaction effect for white fir was -0.14±0.017 (live) vs -0.1±0.019 (dead). The magnitude and direction of the effect of winter temperature on tree growth depended on winter precipitation levels. Trees that died had higher growth with higher winter temperatures, even when winter precipitation was low. Yet, drought-surviving trees had higher growth with increased winter temperatures, but only during wet winters. Furthermore, we discovered differences in the duration of climate memory of winter precipitation, with drought-surviving aspen and Douglas fir trees having longer memory compared with trees that died. Winter temperature was most important one and three years prior to ring formation for living aspen and Douglas fir trees, respectively. Aspen trees that succumbed to drought, however, showed relatively short memory of winter precipitation, such that precipitation received during the winter immediately prior to ring formation was most important. This study suggests that identifying regional or species-specific patterns of the lagged climate effects (e.g., drought), and potential warning signs of tree mortality could help improve forest productivity predictions under warming climates with increasingly variable precipitation.