Simple physics-based models of compensatory plant water uptake: concepts and eco-hydrological consequences

Abstract. Many land surface schemes and simulation models of plant growth designed for practical use employ simple empirical sub-models of root water uptake that cannot adequately reflect the critical role water uptake from sparsely rooted deep subsoil plays in meeting atmospheric transpiration demand in water-limited environments, especially in the presence of shallow groundwater. A failure to account for this so-called "compensatory" water uptake may have serious consequences for both local and global modeling of water and energy fluxes, carbon balances and climate. Some purely empirical compensatory root water uptake models have been proposed, but they are of limited use in global modeling exercises since their parameters cannot be related to measurable soil and vegetation properties. A parsimonious physics-based model of uptake compensation has been developed that requires no more parameters than empirical approaches. This model is described and some aspects of its behavior are illustrated with the help of example simulations. These analyses demonstrate that hydraulic lift can be considered as an extreme form of compensation and that the degree of compensation is principally a function of soil capillarity and the ratio of total effective root length to potential transpiration. Thus, uptake compensation increases as root to leaf area ratios increase, since potential transpiration depends on leaf area. Results of "scenario" simulations for two case studies, one at the local scale (riparian vegetation growing above shallow water tables in seasonally dry or arid climates) and one at a global scale (water balances across an aridity gradient in the continental USA), are presented to illustrate biases in model predictions that arise when water uptake compensation is neglected. In the first case, it is shown that only a compensated model can match the strong relationships between water table depth and leaf area and transpiration observed in riparian forest ecosystems, where sparse roots in the capillary fringe contribute a significant proportion of the water uptake during extended dry periods. The results of the second case study suggest that uncompensated models may give biased estimates of long-term evapotranspiration at the continental scale. In the example presented here, the uncompensated model underestimated total evapotranspiration by 5–7% in climates of intermediate aridity, while the ratio of transpiration to evaporation was also smaller than for the compensated model, especially in arid climates. It is concluded that the parsimonious physics-based model concepts described here may be useful in the context of eco-hydrological modeling at local, regional and global scales.