Detecting tree stress fingerprints using sub-daily sapflow data

Plants are subject to stress conditions at multiple time-scales, from minutes and hours (e.g., radiation stress) to years or decades (e.g., prolonged drought). The processes controlling how plants respond to such stressors are also time-scale dependent, from rapid physiologic and structural responses such as stomatal regulation or leaf movement, to slow responses such as pigment changes or adjustments of growth and allocation. How these different processes evolve and interact under diverse stressors influences tree health and long-term functioning and, depending on plants ability to recover, might lead to tree health decline and mortality. Observations of tree stress from space typically rely on reflectance indices, which are associated with changes or declines in leaf pigment content, leaf area, and/or fractional of vegetation cover. These changes are driven by slow or delayed reactions to environmental stress (leaf discoloration, defoliation, reduced growth, mortality and compositional changes). Microwave measurements, on the contrary, allow to more directly track vegetation water content, but they are typically available at coarse spatiotemporal scales. Signs of plant health decline or onset of mortality trajectories can, thus, take a long time to detect based on currently available remote-sensing information, limiting our ability for early detection of stress hotspots (e.g., stands at risk of drought-induced mortality). Here, we aim to explore the potential to use sub-daily microwave observations for early detection of plant stress, in the context of SLAINTE, a mission idea recently submitted in response to ESAs 12th call for Earth Explorers. To do this, we analyze sapflow measurements covering over a decade in an evergreen broadleaf forest at the Puchabon study site (FRA-Pue, southern France) to evaluate how sub-daily information of vegetation water fluxes might be used to identify onset and development of plant stress. We define a set of sub-daily metrics (timing of peak sapflow, sensitivity to meteorological drivers, hysteretic behaviour) and evaluate how these vary within the growing season, across years and during extreme events for multiple trees. These derived metrics could, in principle, be derived from sub-daily satellite-based observations, facilitating therefore timely assessments of plant health declines.