Delineating Environmental Stresses to Primary Production of U.S. Forests From Tree Rings: Effects of Climate Seasonality, Soil, and Topography

Abstract

Primary production is the entry point of energy and carbon into ecosystems, but modeling responses of primary production to “environmental stress” (i.e., reductions of primary production from nonoptimal environmental conditions) remains a key challenge and source of uncertainty in our understanding of Earth's carbon cycle. Here we develop an approach for estimating annual “environmental stress” from tree rings based on the proportion of the optimal diameter growth rate (from species‐specific allometric equations) that is realized in a given year. We assessed climatic, topographic, and soil drivers of environmental stress, as well as their interactions, using both empirical model experiments and linear mixed effect models. Climate gradients and interannual climate variability dominated spatial and temporal variability of environmental stress in much of the western United States, where the tree‐ring environmental stress index was positively correlated with antecedent climatic water balance (precipitation minus potential evapotranspiration) and negatively correlated with temperature and vapor pressure deficit. Excluding topographic and soil information from empirical models reduced their ability to capture spatial gradients in environmental stress, particularly in the eastern United States, where growth was not as strongly limited by climate. Mean climate conditions and topographic characteristics had significant interaction effects with the climatic water balance, indicating an increasing importance of winter moisture for warmer and drier sites and as elevation and topographic wetness index increased. These results suggest that including effects of antecedent climate (particularly in dry regions) and site topographic and soil characteristics could improve parameterization of environmental stress effects in primary production models.

Frequencies of Multivariate Air Masses Drive Tree Growth

“…This research suggests that, compared to just using recent temperature and precipitation records, air masses offer the potential to improve such predictions. Further, the strength of the TRW/AM relationship suggests that a machine learning approach like that used here or in previous research (e.g., Bodesheim et al, 2022;Dannenberg et al, 2020) could be used to predict growth in locations where it was not actually measured, thus allowing for a climate-informed spatial upscaling of growth anomalies between discrete site locations.…”

“…Despite these limitations, this research shows that not only is tree growth impacted by multivariate, synoptic-scale atmospheric conditions, but moreover, for many species and locations, multivariate air masses are better indicators of tree growth than local temperature and precipitation alone. There has been a keen interest recently in ecological forecasting (Dietze et al, 2018;Luo et al, 2011) and the impacts of antecedent abiotic factors on plant stress (Dannenberg et al, 2019(Dannenberg et al, , 2020Dannenberg & Wise, 2016;Earles et al, 2018;Jiao et al, 2022;Kannenberg, Novick, et al, 2019;Ogle et al, 2014;Peltier et al, 2018). In particular, knowing the potential for tree growth or tree stress in the upcoming months based upon the conditions up to 12 months prior could aid in forecasts for wildfire vulnerability and widespread tree mortality.…”

“…Recent work has shown that tree growth anomalies can be successfully predicted and upscaled using machine learning based on surface meteorological variables (Bodesheim et al., 2022; Dannenberg et al., 2020). To quantify the variability in tree‐ring width that can be explained by all AMs collectively, we trained cascade‐forward artificial neural network (ANN) models for each of the 904 records.…”

“…Tree rings provide a means both to evaluate the impact of climate on forest growth and ecosystem productivity (“climate response,” e.g., Dannenberg et al., 2019, 2020; Kannenberg, Maxwell, et al., 2019; Peltier et al., 2018; Williams et al., 2013) and, given that many tree species can live for hundreds or even thousands of years, to reconstruct past climates (“climate reconstruction,” e.g., Cook et al., 1999, 2015; Wilson et al., 2016; Wise, 2010, 2016). While much work in dendroclimatology has focused on temperature or precipitation/drought, recent work in “synoptic dendroclimatology” (Hirschboeck et al., 1996) examines growth responses to synoptic‐scale atmospheric circulation (Dannenberg & Wise, 2017; Hudson et al., 2019) and reconstructs those circulation patterns in the pre‐instrumental period (Trouet et al., 2018; Wise & Dannenberg, 2014, 2017).…”

Comprehensive Quantification of the Responses of Ecosystem Production and Respiration to Drought Time Scale, Intensity and Timing in Humid Environments: A FLUXNET Synthesis

“…A drought that occurs during a physiologically important period, for example, earlywood growth in trees, will likely be more damaging than a drought occurring during the dormant season. The lagged impact of drought is another important factor, which has been documented in tree rings (Dannenberg et al., 2019, 2020; Kannenberg et al., 2019), eddy covariance data (Knowles et al., 2020), remote sensing data (Berkelhammer et al., 2017), physiological measurements (Earles et al., 2018), and drought monitoring studies (Chang et al., 2018; Li et al., 2020; Waseem et al., 2016). Thus, only considering one or two drought dimensions may not objectively reflect the actual drought impacts.…”