Downscaling Vapor Pressure Deficit across Complex Forested Terrain

Abstract The forest microclimate shapes many aspects of forest functioning, including plant regeneration and wildfire occurrence. In complex landscapes with varying terrain, the forest microclimate varies at fine spatial scales (10–100 m 2 ). However, accurately mapping this variation remains challenging. Vapor pressure deficit (VPD) is an important microclimatic variable for plant growth and fire activity, yet few studies have specifically focused on downscaling VPD. The aim of this study was to examine the drivers of in-forest VPD in temperate eucalypt forests and develop a model to predict in-forest VPD. We use microclimate data from 37 in-forest weather stations, installed across an aridity gradient in southeastern Australia. We used linear mixed models within an information theoretic approach to develop a predictive model for daily maximum in-forest VPD using open VPD and topographic variables. The highest-ranked model included fundamental topographic drivers of vegetation structure and microclimate in forested landscapes: aspect, elevation, and slope, in addition to open VPD. The model had high accuracy when tested against independent data. We used this model to map fine-scale (30 m 2 , daily) maximum in-forest VPD across a topographically complex case study landscape. Predicted in-forest VPD showed considerable spatial and temporal variations not captured by coarse-scale open VPD. This represents a significant advancement in our ability to model microclimatic conditions in temperate eucalypt forests and has the potential to advance our understanding of how ecosystem processes vary at fine spatial scales.