OOS 21-1 - Phenology-driven variation in evapotranspiration: implications for early drought monitoring and detection

Background/Question/Methods
Spring phenological timing has important consequences for ecosystem carbon uptake, with earlier leaf out associated with greater carbon uptake. However, the consequences of an earlier start to spring for ecosystem water cycling have been relatively unexplored. Some evidence suggests a connection between enhanced spring evapotranspiration (ET) during years with an early start of spring, and subsequent soil moisture deficits during the growing season -- raising the possibility that springtime phenological timing could function as an early drought indicator. However, this strategy depends on the extent to which drought metrics can disentangle phenology-driven increases to ET from drought-driven decreases to leaf-level water use. The relationship between spring phenology and summer soil moisture will also depend on the pace of canopy development; if leaf area expands more rapidly in years when spring starts later, the relationship between spring phenology and summer soil moisture will be weakened. Here, we leverage ecosystem scale flux data, remote sensing observations, and tree-level phenology data to understand: 1) How well popular drought indices capture drought onset in years when it coincides with leaf development, and 2) how the pace of canopy development interacts with the timing of leaf emergence to determine the likelihood of summer soil drought.

Results/Conclusions
Across multiple sites that experienced an early spring in the same year as a summer drought, a plant drought index – the Evaporative Stress Index (or ESI) – was unable to capture the early stages of drought due to increased springtime ET driven by advanced phenology. However, a variant of the ESI that corrects for the relationship between leaf area and surface conductance is able to detect the emerging drought 7-10 weeks other than the conventional ESI and other drought metrics. Next, we focus our attention on the links between springtime ET and summer water use. Using information on ET and greenness from flux towers and satellite remote sensing, we demonstrate a relationship between springtime phenology and summer soil moisture in many places; however, we demonstrate an inverse relationship between the start of phenological spring and the rate of leaf area development at both the ecosystem and tree level that complicates a more straightforward relationship between spring phenological timing and summer drought risk. Overall, springtime ET and phenology are still promising targets for early drought detection, but they must be evaluated together with information about climate conditions that influence both water supply and the pace of canopy development.