When evapotranspiration masks drought: transpiration-based water-use efficiency reveals ecosystem-specific thresholds for drought across U.S. coastal zones

Coastal ecosystems store large amounts of carbon relative to upland terrestrial ecosystems, yet their carbon-water relations remain poorly understood. Standard monitoring metrics – meteorological drought indices and evapotranspiration-based, ecosystem-level water-use efficiency (WUEET = GPP/ET) – were designed for water-limited ecosystems and perform poorly in coastal systems. In these systems, tidal flooding, groundwater connectivity, salinity stress, and evaporative losses from standing water and wet soils may decouple total evapotranspiration (ET) from plant physiological status and gross primary productivity (GPP). Using eddy-covariance observations from 64 sites within 80 km of the Atlantic, Gulf, Pacific, and Alaska coasts of the United States, with transpiration (T) estimated through ET partitioning, we test whether transpiration-based WUE (WUET = GPP/T) provides a more reliable indicator of ecological drought in coastal ecosystems. We identify the Standardized Precipitation-Evapotranspiration Index (SPEI) thresholds at which meteorological drought produces physiological carbon-water decoupling. We show that WUEET and WUET diverge most strongly at low T:ET, and converge at high T:ET (p < 0.001). This divergence shows that, in coastal systems, evapotranspiration-based monitoring can obscure vegetation responses, supporting transpiration-based metrics. WUET responded significantly to drought across all seven SPEI timescales ranging from 1 (SPEI-1) to 48 (SPEI-48) months in upland coastal ecosystems, strongest under short-term moisture-anomalies (p < 0.001), while lower elevation freshwater and saline responses were nonsignificant. The SPEI level associated with a ≥5% change in WUET differed among regions, with WUET increasing under short-term dry conditions along the Pacific and Atlantic, declining under prolonged dry conditions in Alaska, and showing no consistent response to dry conditions along the Gulf Coast. These results provide a tractable framework for coastal drought monitoring as follows: assess T:ET to determine metric reliability, stratify vulnerability by ecosystem class, and match SPEI accumulation windows to the characteristic drought-integration timescale of each coastal region.