Parsimonious representation of plant water use strategies via non-dimensional parameter groups

<p>Process-based models are needed to improve estimates of water and carbon cycles in variable climatic conditions. Yet, their utility is often limited by our inability to directly measure plant stomatal and hydraulic traits at scales suitable to quantify characteristics of whole ecosystems. Inferring such parameters from ecosystem-scale data with parsimonious models offers an avenue to address this limitation. To this aim, we use a simple representation of the water flux through the soil-plant-atmosphere continuum (SPAC) and derive a parameterization of Feddes-type soil water-limitation constraints on transpiration (expressed via a soil moisture dependent function β). This parameterization explicitly accounts for community-effective plant eco-physiological traits as encoded in the SPAC model parameters. We express analytically the fractional loss of conductivity in well-watered conditions and the soil saturation thresholds at which transpiration is down-regulated from its well-watered rate and at which transpiration ceases, as a function of non-dimensional parameter groups. These non-dimensional groups combine plant stomatal and hydraulic traits, soil texture and climate. We implement the theoretical β function into a soil water balance and infer distributions of plant traits which best-match FLUXNET observations in a range of biomes. Finally, we analyze the resulting non-dimensional groups to explore patterns in plant water use strategies. Our results indicate that non-dimensional groups reflect combinations of plant traits which are adapted to growing season environmental conditions and these groups may be more meaningful model parameters than individual traits at ecosystem scales. Additionally, using non-dimensional groups instead of focusing on individual parameters reduces risks of equifinality and provides future opportunities to exploit satellite data to quantify robust ecosystem-scale parameters. This analysis provides a parsimonious and functionally accurate alternative to account for ecosystem hydraulic controls and feedbacks and can help overcome limitations of commonly used empirical water-limitation constraints.</p>