Abstract
Climate change increasingly affects tree growth and carbon allocation in subalpine forests across the Western United States, necessitating new approaches to better monitor its impacts. Although the use of thermal data in monitoring tree growth and carbon allocation remains in its infancy, recent research has identified leaf temperature (TL) as a strong predictor of stem radial growth (SRG). The goal of this study was to evaluate the potential of satellite-based estimates of canopy temperature (TC) and soil moisture (SM) to provide insights into daily and annual tree growth dynamics. We utilized generalized additive mixed models and linear mixed models with spaceborne estimates of TC (ECOSTRESS) and SM (SMAP) as the primary predictors of daily resolved SRG (SRGD) across five growing seasons (2020 – 2024) in a subalpine mixed-conifer forest of Central Idaho. Models predicted micrometer-level SRGD with moderate precision and accuracy (R2 = 0.56; RMSE = 21.45 µm), while the binomial representation of daily growth/no growth (SRGD-B) was predicted with 87 % accuracy. When multiplying the annual mean predicted SRGD by the annual mean difference in days between satellite data acquisitions (5.58 ± 5.53 days), we were able to predict annual growth performance (%; i.e., how much a tree grew relative to its average growth across years) with high precision (R² = 0.76) and accuracy (RMSE = 7.63 %). These findings highlight the efficacy of thermal satellite remote sensing data for tracking intra- and inter-annual tree growth dynamics, and the potential for scaling satellite-based tree growth estimates across broad regions.