Soil water that evaporates or is tapped by plants is largely separate from that which runs into streams and recharges groundwater. This finding has big implications for our understanding of water cycling. See Letter p.91
Soils can be viewed as the investment managers of the terrestrial water cycle: they accept precipitation capital from the atmosphere and allocate it to sustain and grow various biological and hydrological stocks. These water investments influence plant productivity, run-off to streams and groundwater, and atmospheric humidity, so deciphering how soils partition water is vital if we are to understand and predict the function of these systems. On page 91 of this issue, Evaristo et al.1 suggest that the allocation of water in most soils worldwide follows a conservative, diversified 'strategy', in which new resources are invested as they are obtained, and transfer of capital between accounts is limited.
Water researchers have considered two contrasting scenarios for soil-water allocation. The 'commingled' scenario, which is the one most widely adopted in hydrological models (see refs 2 and 3, for example), assumes that all water is held in a common pool and is withdrawn only as needed. Residual water from past precipitation is tapped by plants, evaporates or hosts biogeochemical reactions until fresh precipitation displaces some or all of it into groundwater reservoirs or streams. This situation has been referred to as hydrological connectivity, because water leaving the soil in any form is drawn from a common pool and is connected to all other flows.
The contrasting scenario could be said to be 'diversified', because new water is allocated to one of several pools as it enters the soil, and transfers between these pools are limited. Previous hydrological research has provided hints of a diversified approach to soil-water investment. For example, in many soils a substantial fraction of infiltrating water moves rapidly through large pores to recharge groundwater and produce stream run-off4. There is surprising evidence that this separation of soil-water pools extends to water withdrawal by plants, with trees and shrubs in two ecological systems5,6 drawing from a soil-water pool that is apparently distinct from that feeding recharge and run-off. The generality and importance of such diversification have been unclear, however, because relatively few studies have been conducted.
Evaristo and colleagues provide compelling evidence that the diversified mode is widespread, if not ubiquitous. The authors adapt previously reported methods5,6 that capitalize on a distinctive shift in the ratios of hydrogen isotopes and oxygen isotopes in soil water as it evaporates from soils. If the water in soil, plants, groundwater and streams all showed a common evaporation shift, this would strongly suggest a commingled situation (Fig. 1a). But in a meta-analysis of data from 47 studies spanning multiple environments and biomes, the authors instead find a similar evaporation shift in soil and plant water, but little or no shift in streams and groundwater (Fig. 1b). This implies diversification: plants and soil evaporation across ecosystems seem to be tapping a pool of water that is largely separate from the pools that generate run-off and recharge.
This finding has major implications for — and raises many questions about — our understanding of water cycling. At a fundamental level, the diversified mode is conservative in that water is allocated to support multiple uses, with less water available than in the commingled model to support run-off or plant growth individually. The mechanisms that maintain the segregation of these pools of water as they move through the soil matrix remain unresolved, but understanding them is crucial for developing accurate models of soil-water partitioning. In particular, the relative roles of physical and temporal segregation remain unclear. Do plants draw water from different parts of the soil matrix from groundwater recharge, or do plant withdrawals happen at a different time from groundwater recharge?
The suggestion that plants that span many biomes have developed strategies of water use that focus on one pool of soil water largely to the exclusion of another is intriguing, given that drought stress is a major driver of plant mortality7. Perhaps the explanation lies in the transient nature of the soil-water stocks that contribute to recharge and run-off. Evaristo and co-workers' observation of the pervasive pattern of association between plant and soil water, but not with run-off-generating water, calls for further research. However, the isotopic methods used by the authors are less useful in the study of non-woody species such as grasses, so extending this work to some fast-growing and potentially less-discriminative water users will require new approaches.
The lack of water exchange between soil pools also calls into question some previous analyses of water-cycle processes (see ref. 8, for example), because it implies that methods for studying water partitioning that use measurements of chemical or isotope tracers in streams may be blind to the part of the soil-water balance sheet that involves plants and soil evaporation. Indeed, a study9 published earlier this year found that balancing the global water-isotope budget requires widespread hydrological separation in soils, consistent with Evaristo and colleagues' results, and called for a revision of previous global flux estimates from studies that did not consider hydrologic separation.
Finally, water allocation is inextricably coupled with soil biogeochemical reactions, from rock weathering to nutrient cycling, and the effect of diversified soil-water allocation on these processes may be enormous. Many reactions occur in thin films of water surrounding mineral grains. If these films are part of a long-lived soil-water pool, and there is little physical exchange of this water with through-flowing, run-off-generating water, what are the implications for the transfer of the reaction products to groundwater and streams? In this sense, the 'trickle-down' effects of soil-water economics may structure the entire soil chemical system. Better understanding of the processes governing soil-water partitioning may ultimately help to resolve long-standing problems in geochemistry — such as discrepancies between field- and laboratory-based mineral-weathering rates10, which are central to our understanding of the global carbon cycle and to proposed geoengineering strategies for coping with anthropogenic emissions of carbon dioxide.Footnote 1
Evaristo, J., Jasechko, S. & McDonnell, J. J. Nature 525, 91–94 (2015).
Liang, X., Lettenmaier, D. P., Wood, E. F. & Burges, S. J. J. Geophys. Res. 99, 14415–14428 (1994).
Neitsch, S. L., Arnold, J. G., Kiniry, J. R. & Williams, J. R. Soil and Water Assessment Tool Theor. Document. Version 2009 (Texas Water Resources Inst., 2011).
Beven, K. & Germann, P. Water Resour. Res. 18, 1311–1325 (1982).
Brooks, J. R., Barnard, H. R., Coulombe, R. & McDonnell, J. J. Nature Geosci. 3, 100–104 (2010).
Goldsmith, G. R. et al. Ecohydrology 5, 779–790 (2012).
Anderegg, W. R. L. et al. Proc. Natl Acad. Sci. USA 109, 233–237 (2012).
Jasechko, S. et al. Nature 496, 347–350 (2013).
Good, S. P., Noone, D. & Bowen, G. Science 349, 175–177 (2015).
Moore, J., Lichtner, P. C., White, A. F. & Brantley, S. L. Geochim. Cosmochim. Acta 93, 235–261 (2012).
Related links in Nature Research
About this article
Calibration method affects the measured δ 2 H and δ 18 O in soil water by direct H 2 O liquid –H 2 O vapour equilibration with laser spectroscopy
Hydrological Processes (2020)
Hydrological Processes (2019)