Variations in Conifer Tree Surface Temperature across a Hydrologic Gradient in Sagehen Creek Watershed, Sierra Nevada, USA

Forest canopy temperature provides information about longwave radiation and sensible energy exchange, and has applications for understanding forest water content and tree water use. Forest canopy temperature is influenced by many environmental factors, including air temperature, solar zenith angle, cloudiness, and canopy shading. In this research we aim to understand how tree- and hillslope-scale canopy temperature varies through time and how it is related to air temperature. We installed five thermal infrared (TIR) sensors around a single Lodgepole Pine tree in Sagehen Creek watershed, Sierra Nevada, USA. Canopy temperature of the tree varies from 1^oC in March, as the snow melts, to around 41 ^oC in July during the peak of tree water use. Daily variation of canopy temperature around the tree follow the same pattern of air temperature, however, at the top of the tree it is warmer than other aspect of tree. In the shaded part of the tree, canopy temperature is approximately same as air temperature.

We also used unmanned aerial system (UAS) thermal imagery to explore hillslope-scale variation in canopy temperature across a gradient of a wet meadow to a dry hillslope. The UAS-based TIR imagery required numerous calibration efforts to resolve absolute temperature. A constant 0°C reference calibration method using melting snow as a natural black-body radiometric calibration target is developed and applied. Subtracting the snow-based bias correction from temperature maps resulted in about 1.5 °C error for retrieving forest canopy temperature and 0.3 °C bias for retrieving water surface temperature from airborne TIR observations. Canopy temperature across the hillslope averaged 11 ^oC in April and 29 ^oC in July, but varied by 6 ^oC and 9 ^oC, respectively across individual trees. Overall, our results show that canopy temperature varies systematically with air temperature in shaded, but that spatiotemporal deviations between the two may hold information about tree water use and water stress.