Ecosystem-scale Plant Isohydricity Controls Grassland Productivity Sensitivity to Aridity

Droughts are expected to become more frequent and more intense under climate change. Plant mortality rates and biomass declines in response to drought depend on stomatal and xylem flow regulation. Plants operate on a continuum of xylem and stomatal regulation strategies from very isohydric (strict regulation) to very anisohydric. Co-existing species may display a variety of isohydricity behaviors. As such, it can be difficult to predict how to model the degree of isohydricity at the ecosystem scale by aggregating studies of individual species. This is nonetheless essential for accurate prediction of ecosystem drought resilience and for prediction of changes in leaf area (which in turn affect evapotranspiration fluxes and the hydrologic balance). In this presentation, we define and introduce a metric for the degree of isohydricity at the ecosystem scale in analogy with a recent metric introduced at the species-level. Using data from the AMSR-E satellite, this metric is evaluated globally based on diurnal variations in microwave vegetation optical depth (VOD), which is directly related to leaf water potential. Areas with low annual-mean radiation are found to be more anisohydric. Except for evergreen broadleaf forests in the tropics, which are very isohydric, and croplands, which are very anisohydric, land cover type is a poor predictor of ecosystem isohydricity, in accordance with previous species-scale observations. It is therefore also a poor basis for parameterizing water stress response in land-surface models. We further use the new dataset to show that within United States (US) grasslands, the productivity (as estimated by summer Normalized Difference Vegetation Index, NDVI) of anisohydric ecosystems is more than three times as sensitive to aridity – as assessed by the vapor pressure deficit (VPD) and precipitation - than that of isohydric ecosystems. Grasslands - especially anisohydric ones - are also far more sensitive to variations in VPD than to variations in precipitation. The precipitation sensitivity of summer NDVI increases only for the most anisohydric ecosystems. The effect of well-constrained increases in VPD under climate change may dominate the effect of more uncertain precipitation changes on future grassland water and carbon fluxes.