Wed, Aug 04, 2021:On Demand
Background/Question/Methods
Tree water use contributes substantially to urban water balance. There is an urgent need for improved assessment of urban tree water use and sensitivity to increased evaporative demand associated with atmospheric droughts. We used a hierarchical Bayesian model testing framework to assess the role of environment, species, and trait variance in driving transpiration rate estimated from sap flux measurements for eight common species of the Los Angeles urban ecosystem. We hypothesized that (1) this modelling approach would accurately predict transpiration rate under dynamic relative humidity (RH), photosynthetically active radiation (PAR), and temperature (T) for trees of different size (diameter at breast height; DBH). We further predicted that (2) the inclusion of species in the model would improve accuracy of water use estimation and that (3) species-specific leaf and tree-scale traits could represent species and further improve our model predictions across atmospheric conditions. We considered a set of hydraulic and morphological traits, including leaf thickness (LT), leaf mass per area (LMA), leaf density (LD), leaf carbon and nitrogen concentrations (δ13C and δ15N), leaf osmotic potential at turgor loss point (πtlp), leaf hydraulic conductance (Kleaf), leaf sensitivity to hydraulic decline (the leaf water potential at 50% loss of Kleaf ; P50), and tree sapwood area to leaf area ratio (SA:LA).
Results/Conclusions Across the eight species included in our study, we found substantial variation in mean annual transpiration rate (8.5 kg d-1 for Quercus agrifolia to 95 kg d-1 for Gleditsia triacanthos) measured throughout the year with daily mean RH, PAR and T ranging from 0.1 - 4.6 kPa, 3.0 - 60.6 μmol m-2 s-1, and 8.2 - 35.5 ºC, respectively. We found novel relationships between urban tree physiological and functional traits, including coordination of πtlp with P50 and a trade-off between maximum Kleaf and P50. Our basic model including only environmental factors predicted tree-level transpiration under the range of observed atmospheric conditions with powerful accuracy. Including species as a nested factor improved predictions, as did featuring traits in a second nested level that included LT, πtlp, SA:LA, and δ13C. Our models including species’ traits also resolved differences across species in mean annual transpiration rates. The inclusion of traits in predictions of tree water use can enable the resolution of differential water use according to species composition and provides a foundation for improved scaling predictions for canopies and ecosystems in urban areas experiencing atmospheric droughts.
Results/Conclusions Across the eight species included in our study, we found substantial variation in mean annual transpiration rate (8.5 kg d-1 for Quercus agrifolia to 95 kg d-1 for Gleditsia triacanthos) measured throughout the year with daily mean RH, PAR and T ranging from 0.1 - 4.6 kPa, 3.0 - 60.6 μmol m-2 s-1, and 8.2 - 35.5 ºC, respectively. We found novel relationships between urban tree physiological and functional traits, including coordination of πtlp with P50 and a trade-off between maximum Kleaf and P50. Our basic model including only environmental factors predicted tree-level transpiration under the range of observed atmospheric conditions with powerful accuracy. Including species as a nested factor improved predictions, as did featuring traits in a second nested level that included LT, πtlp, SA:LA, and δ13C. Our models including species’ traits also resolved differences across species in mean annual transpiration rates. The inclusion of traits in predictions of tree water use can enable the resolution of differential water use according to species composition and provides a foundation for improved scaling predictions for canopies and ecosystems in urban areas experiencing atmospheric droughts.