Examining the Validity of Mechanistic Microclimatic Downscaling in a Montane System

Wilk, Andrew; Peterman, William

ABSTRACT: Accurately predicting localized weather and abiotic conditions under forested cover like temperature, relative humidity, soil moisture, and soil temperature at fine spatial temporal scales at any point on earth would be revolutionary in ecology. Recent advances in mechanistic weather downscaling and the development of an R package, microclimc , have made these tools available to the research community. In this study we apply this new model at the locations of two long-term, high-resolution weather stations run by the USGS to understand how well the model can predict multiple abiotic variables. We found that predicted air temperature and soil temperature at a depth of 15 cm were highly correlated with measured values. while relative humidity and soil moisture were not. Additionally, we saw that seasonal trends in error are present, and the most accurate predictions occurred during the summer (May–June). We believe that a worthy endeavor would examine the effects that different parameter tunings have on the accuracy and stability of the models and that there is great potential for mechanistic weather downscaling.