Ecosystem water relations

Abstract

‘Water potential’ is the biophysically relevant measure of water status in vegetation relating to stomatal conductance, hydraulic conductance, and mortality thresholds; yet this cannot be directly related to fluxes of water at plot- to landscape-scale without understanding its relationship with ‘water content’.  The relationship between water content and water potential has long been of interest to plant and soil scientists and is typically determined at small scale on excised plant parts or soil samples.  But how can water potential be scaled from leaf to canopy, branch to stem, tree to forest?  How does water content vary throughout a plant, and across individuals and species? In this talk I will outline some practical considerations for deriving representative values for ecosystem water potential and content: the ecosystem pressure-volume curve.  I will discuss how ecosystem water status is influenced by the boundaries we apply to the system, which differ depending on whether we are interested in interpreting remote sensing data, models, or field measurements.  And I will describe the concept of the ‘state-based’ model which relates to steady-state vegetation types that emerge predictably in response to a given climate or hydraulic environment. To support this discussion, I will present ecosystem pressure-volume curves from nine sites including tropical rainforest, savanna, temperate forest, and a long-term Amazonian rainforest drought experiment.  The results from these preliminary analyses show that the relationship between the water stored in biomass consistently scales with biomass across systems, as does the vegetation-level hydraulic capacitance; while the relative measures of water storage and hydraulic capacitance show no trend across ecosystems.  Such cross-biome relationships in water relations hold promise for improving our understanding of vegetation-climate feedbacks over large spatial and temporal scales, and enhancing our capacity to interpret remote sensing data.