Coupling remote sensing and in situ measurements to understand evapotranspiration downregulation in a subalpine conifer forest

Understanding how evapotranspiration (ET) is downregulated from its potential maximum (Ep) is one of the major difficulties remaining for consolidating a predictive understanding of evaporative fluxes over space and time. Under moisture-limited conditions, which characterize much of the terrestrial surface for much of the time, actual evaporative flux (Ea) is less than Ep. Both soil water potential (ψm) and vapor pressure deficit (VPD) can contribute to this downregulation, but the relative magnitude of their contributions remains contested and varies widely across landscapes and plant communities. To close this gap in a high-elevation, snow-dominated conifer forest, we quantified the sensitivity of ET downregulation to ψm and VPD during post-snowmelt drydowns using remote sensing estimates constrained by flux measurements from three subalpine eddy covariance sites. We derived measurements of daily Ea and Ep from OpenET, an emerging source of high-resolution global daily ET estimates. We estimated ET downregulation as Ea / Ep and quantified the sensitivity of this ratio to midday observations of VPD and ψm from in situ sensors at 10 conifer forest sites. Intercomparisons between remote sensing and eddy covariance estimates showed reasonably strong agreement between sources of Ea data. The slopes of least-squares regression lines for ensemble estimates ranged from 0.78 to 0.89 across the three flux sites, indicating that OpenET tended to underestimate Ea on average. Across sites, Ea / Ep declined as ψm became more negative. Ea / Ep increased as VPD increased up to a threshold around 1.1 kPa, beyond which the ratio values declined. We also identified interactions between VPD and soil water potential: when soil water potential was high, VPD contributed more to ET downregulation. When water potential was low, it became the more limiting factor, and VPD contributed less to ET downregulation. Our initial results advance the mechanistic understanding of ET downregulation, a particularly acute need in mountain regions, where uncertainty in ET predictions remains high. In identifying both independent and interacting effects of VPD and ψm on forest ET, we emphasize the importance of accounting for both factors in assessing forest responses to limiting hydroclimatological conditions. Our approach also demonstrates the viability of a simple, effective method to characterize some of the dominant constraints on ET in terrain where many of the fundamental assumptions of eddy covariance cannot be met. Our study improves the hydrological modeling community’s capacity to attribute changes in ET to increasing atmospheric demand and declining soil water availability, with further implications for modeling ecosystem dynamics and vegetation-atmosphere feedbacks. Future work will include observations from additional sites and data spanning both high and low snow years.