B106-09
Using Vegetation Greenness to Predict Seasonal Variation in Evaporative Fraction at AmeriFlux Sites
Tuesday, 15 December 2020: 21:02
Virtual
Adam M Young1, Mark A Friedl2, Steve E Frolking3, Carlos Carillo4, Toby Ault4, Eric Beamesderfer5, Xiaolu Li6, Thomas E Milliman7, Minkyu Moon8, Bijan Seyednasrollah9, Andrew D Richardson10 and AmeriFlux collaborators, (1)Northern Arizona University, School of Informatics, Computing and Cyber Systems, Flagstaff, AZ, United States, (2)Boston University, Earth and Environment, Boston, MA, United States, (3)University of New Hampshire, Institute for the Study of Earth, Oceans, and Space, Durham, NH, United States, (4)Cornell University, Department of Earth and Atmospheric Science, Ithaca, NY, United States, (5)Northern Arizona University, School of Informatics, Computing and Cyber Systems, Flagstaff, United States, (6)Cornell University, Ithaca, NY, United States, (7)Univ. of New Hampshire, Institute for the Study of Earth, Oceans and Space (EOS), Durham, NH, United States, (8)Boston University, Boston, MA, United States, (9)Northern Arizona University, School of Informatics, Computing, and Cyber Systems, Flagstaff, AZ, United States, (10)Northern Arizona University, School of Informatics, Computing & Cyber Systems, Flagstaff, AZ, United States
Abstract:
Vegetation phenology is tightly linked to significant seasonal shifts in the Bowen ratio, with latent heat flux (LE) becoming dominant relative to sensible heat flux (H) after leaf emergence. However, the overall importance of phenology as a control of seasonal variation in LE relative to other factors remains unclear. Here, we investigated how phenology acts as a control of daily LE using AmeriFlux data. Phenology was quantified using vegetation greenness from PhenoCam imagery. We focused on characterizing if and how evaporative fraction (EF = LE/(LE + H)) and bulk surface conductance (Gs) responded to phenology. First, we quantified the timing of seasonal transitions in EF and Gs, and then compared these timings to phenological transition dates from PhenoCam. Second, we constructed an a priori conceptual model linking pathways among vegetation greenness and other environmental variables to Gs and EF. We used structural equation modeling (SEM) to quantify the relative strength of these pathways, allowing us to identify the primary controls of EF. We analyzed ~170 site-years of data spanning a precipitation gradient across the contiguous U.S.
From our results, vegetation greenness emerged as a key predictor of seasonal shifts in the surface-energy balance. We found clear evidence that changes in EF and Gs were directly tied to leaf emergence after green-up; phenological transition dates were strongly correlated to the timing of seasonal shifts in EF and Gs (r2 = 0.68). The SEM pathway linking vegetation greenness ➝ Gs ➝ EF had the strongest influence over EF. For example, the average path values (PV; higher PV = higher importance) linking these variables were 0.54 and 0.91, respectively. Comparatively, the PV linking soil water content ➝ Gs was 0.18. The mechanism underlying this linkage from vegetation greenness ➝ Gs is related to increased transpiration after leaf emergence, which leads to higher EF through increases in LE. These relationships indicate that seasonal changes in vegetation are a dominant control over seasonal evapotranspiration and ecosystem water balance. Understanding the strength of these linkages will become increasingly important for anticipating climate change impacts such as drought, wildfire hazard, boundary-layer dynamics, and micro-climate surface-warming.
