Local atmospheric vapor pressure deficit as microclimate index to assess tropical rainforest riparian restoration success.

Characterizing microclimatic variables, such as vapor pressure deficit (VPD), is crucial for monitoring ecological processes and biodiversity dynamics of forests, among other terrestrial ecosystems. Approaches using technologies such as remotely piloted aircraft (RPA) have demonstrated potential for assessing the biophysical interface between forests and the atmosphere by obtaining high-resolution microclimatic metrics in space and time. In the present study, we developed a microclimatic approach based on VPD modeling to quantify the success of forest restoration in a tropical rainforest landscape. We used the photogrammetric technique Structure from Motion (SfM) with RPA to estimate three-dimensional forest structures and evaluated its influence in obtaining metrics for VPD modeling. A total of 30 plots of 314 m were analyzed at five stages of riparian forest development, including areas of early-stage passive restoration (E10, 10 years and E14, 14 years), mid-stage natural forest regeneration (M26, 26 years and M29, 29 years), and an old-growth forest (REF). These plots were used to calibrate and validate the VPD model (∼70 % training data and ∼ 30 % test data, with k = 10). Old-growth forests exhibited an average VPD of 0.19 kPa, lower than younger forests that exceeded the 1.0 kPa threshold. The 50th and 75th percentiles of the height distribution explained 86 % and 83 % of the variance in VPD (RMSE of 0.34 kPa), respectively, and demonstrated the potential use of this metric to predict the effects of forest structure on VPD. Results show that early-stage restoration sites are exposed to higher threshold limits of VPD, which can affect ecosystem functioning. Spatial characterization allows for identifying target areas for interventions, increasing our capacity to support better decisions in forest management.