Unlike river discharges to the oceans7, the global fluxes of evaporation and transpiration are poorly constrained owing to a lack of methodology to decouple these two water fluxes at the catchment scale. Stable isotope ratios of oxygen (18O/16O) and hydrogen (2H/1H) in water can be used to separate transpiration from evaporation8, because the two processes have different effects on these ratios in water. The physical process of evaporation enriches residual water in the heavy isotopes of oxygen and hydrogen, whereas the biological process of transpiration does not produce an isotopic fractionation, assuming an isotopic steady state over annual timescales8,9,10,11. The pathway water takes after falling as precipitation within a catchment includes mixing, evaporation (fractionation labelled) and transpiration (non-fractionation labelled), until the remaining water accumulates in a downstream lake or river. Each of these catchment processes is ultimately recorded by the isotopic composition of the lake’s water. We have compiled a data set of δ18O and δ2H values of large lake waters and capitalize on dissimilar isotope effects between evaporation and transpiration to decouple and quantify these two freshwater losses from Earth’s surface (isotope content is given by (R sample /R V-SMOW − 1) × 103 ‰, where R is 18O/16O for δ18O and 2H/1H for δ2H, and V-SMOW represents standard mean ocean water).

To proceed with this calculation, we first report on the stable oxygen and hydrogen isotope compositions of Earth’s large lakes (Fig. 1). The isotopic compositions of lake waters show a broad range in δ18O and δ2H values: −23‰ to +15‰ and −180‰ to +80‰, respectively. Well-mixed lakes (for example, the North American Great Lakes and Lake Baikal) have relatively homogenous stable isotope compositions, whereas perennially stratified or shallow lakes tend to have greater isotopic variability (for example Lake Kivu and Lake Chad). Headwater lakes located at high latitudes and altitudes (for example, Kluane Lake) have the lowest δ18O and δ2H values, whereas the closed basin lakes of eastern Africa have the highest δ18O and δ2H values (for example, Lake Afdera and Lake Turkana). Global lakes do not follow one systematic evaporation trend, reflecting the unique climatology and hydrology of each individual lake catchment. The global meteoric water line plotted in Fig. 1 is a regression of δ18O and δ2H values of precipitation samples on a global scale12. This regression produces a δ2H/δ18O slope of eight that can be closely reconciled by liquid–vapour isotope effects at chemical equilibrium13. However, the disequilibrium process of evaporation results in a strong kinetic isotope effect, with the light isotopologues preferentially partitioned into the vapour phase. This results in δ2H/δ18O slopes of less than eight, driving the isotope composition of lake waters ‘below’ the global meteoric water line. Information on the percentage of evaporative losses is retained by the difference between the lake’s isotope composition and that of waters entering a basin (δ input ; Fig. 1, inset schematic graph). Our global data compilation shows that nearly all lakes fall to the right of the global meteoric water line in δ18O–δ2H space as a result of evaporation, except in special cases where waters evaporate upwind and re-precipitate in a downwind lake basin (for example Lake Biwa). In what follows, we develop equations describing a stable isotope mass balance of waters within a lake catchment to estimate the percentage of catchment transpiration, and apply these equations to δ18O and δ2H data for large lake waters.

Figure 1: δ18O and δ2H values of large lakes and semi-enclosed seas. The global meteoric water line12 (GMWL) is shown. The map at top left shows catchment areas covered by the data set. The schematic graph at bottom right shows water inputs to a lake (diamond) and the evaporation trajectory of a lake (percentages refer to evaporation amount). Full size image Download PowerPoint slide

A lake catchment in hydrologic steady state can be described by a balance between water inputs (I, precipitation and inflows from upstream lakes) and water losses such as precipitation intercepted by vegetation (xP, where x is the fraction of intercepted precipitation for the catchment), open-water and soil evaporation (E), transpiration (T) and liquid losses to rivers or groundwater discharge (Q): Similarly, a stable isotope mass balance of a lake catchment is obtained by considering the isotopic composition of each water flux (δ I , δ E and so on): By combining equations (1) and (2), we develop a new equation describing transpiration losses from a catchment: Each parameter in equation (3) can be estimated from gridded data sets or lake-specific studies, except for the isotope composition of evaporate (δ E ). To calculate δ E (ref. 14), we use an evaporation model based on laboratory-derived liquid–vapour fractionation factors13. The isotope composition of soil and open-water evaporate are grouped into one term (δ E ), and the isotopic composition of transpired moisture (δ T ) is calculated using both shallow and deep-water sources under the assumption that the catchment is in steady state at an annual time step11 (on average, water molecules spend multiple years within lakes examined here). Physical, isotopic and hydroclimatic data sets for each lake are compiled from available reanalysis and interpolated data15,16,17,18 and are used in equation (3) to calculate catchment transpiration (Methods).

We find that transpiration accounts for more than two-thirds of total surface water evapotranspiration for 85% of the catchments examined (Fig. 2 and Fig. 3a). Remarkably, transpiration also accounts for the majority of evapotranspiration in desert catchments (average, 75%; range, 35% to 95%). In situ transpiration measurements in deserts range from 7% to 80% of evapotranspiration19, in large part because precipitation rates are highest in the headwaters of desert catchments, thereby increasing the importance of these forested ecosystems to the catchment’s water balance. Transpiration rates range from less than 100 mm yr−1 to approximately 1,300 mm yr−1 (Fig. 2 and Supplementary Information, section 1). Our results show that even though open-water evaporation may locally occur at higher rates than transpiration, the fraction of total evapotranspiration represented by evaporation is severely limited by the small areas of open water on Earth’s continents (approximately 3%, globally20). Therefore, we posit that the biological pump of water into the atmosphere during photosynthetic gas exchange (that is, transpiration), rather than the physical process of evaporation, dominates water losses from the continents. Because plant roots are able to tap into groundwater and soil-water reservoirs, transpiration effectively moves deep sources of water into the atmosphere, whereas evaporation is only effective for water at or near the surface, which explains the very high proportion of transpiration to the overall evapotranspiration flux.

Figure 2: Transpiration water losses for 56 lake catchments grouped by ecoregion (18O/16O-based results). Each coloured bar represents results for a single lake catchment. Extents of bars show 25th and 75th percentiles of Monte Carlo simulations. Median transpiration (T; square) outputs of Monte Carlo simulations and total evapotranspiration losses (solid line) are shown. The median result is close to the total evapotranspiration for most of the lakes, demonstrating the dominant role of transpiration in total evapotranspiration losses. Full size image Download PowerPoint slide

Figure 3: Transpiration and carbon fluxes within 73 lake catchments. a, Transpiration losses as a percentage of total evapotranspiration. b, Transpiration rates. c, Gross primary productivity for 10% of Earth’s continental area. Coloured diamonds are shown for small basins as a visual aid. Inverted triangles represent compiled in situ transpiration measurements (for example sap flow9). Full size image Download PowerPoint slide

Our results are supported by a cross-plot comparison of isotope-based and conventional open-water evaporation rates (R2 = 0.78 (squared correlation coefficient), slope = 0.92; Supplementary Fig. 1); geographic similarity between compiled in situ transpiration measurements9,10 (of, for example, sap flow) at the forest stand level and our estimates (Fig. 3); and agreement between 18O/16O- and 2H/1H-based evaporation rates using equation (3) (R2 = 0.78, slope = 0.94; Supplementary Figs 2 and 3 and Supplementary Information, section 1). We also note that the time step of our calculated transpiration fluxes ranges from 1 to 1,000 years, averaging 40 years, as dictated by the hydrologic residence time of each lake (Supplementary Table 3). To scale up our calculation to Earth’s ice-free land surface, and to provide a fourth check corroborating our catchment transpiration results, we estimate the global transpiration from Earth’s landmasses (excluding Antarctica) from a stable isotope mass balance of Earth’s entire freshwater reservoir. This estimate is based on the deuterium excess parameter, which includes information contained in both 18O/16O and 2H/1H (d = δ2H − 8δ18O). We obtain a similar expression to equation (3) based on the deuterium excess: Terrestrial precipitation (P = 110,000 ± 10,000 km3 yr−1 (ref. 17), d P = 9.5 ± 1‰ (ref. 12)) is the only input of water to the continents, and water is lost through river discharges to the oceans (Q = 37,300 ± 700 km3 yr−1 (ref. 7), d Q = 6.8 ± 3.8‰; Supplementary Information, section 2), terrestrial evaporation (d E = 75 ± 30‰ (this work)), transpiration (d T = 8 ± 3‰) or interception by vegetation (xP = 7,500 ± 1,500 km3 yr−1 (refs 16, 17), d P = 9.5 ± 1‰ (ref. 12)). Solving equation (4) shows that transpiration accounts for 80% to 90% of terrestrial evapotranspiration (respectively the 25th and 75th percentiles of a Monte Carlo sensitivity analysis). Volumetrically, transpiration converts 62,000 ± 8,000 km3 yr−1 of liquid water into atmospheric vapour, requiring 33 ± 4 W m−2 of latent heat, or roughly half of all solar energy absorbed by the continents21 (approximately 70 W m−2). Results show that 90% of precipitation falling on land18 (111,000 km3 yr−1) is already appropriated to ecosystems for either primary production (62,000 ± 8,000 km3 yr−1) or as aquatic habitat in rivers7 (37,000 km3 yr−1), an important consideration for diversion of in-stream flows.

We can use our transpiration fluxes to calculate carbon assimilation by terrestrial vegetation by linking the water and carbon cycles22. The molar ratio of CO 2 assimilated during photosynthesis to transpired H 2 O is known as water-use efficiency (WUE) and is dependent on a variety of factors including the type of photosynthetic pathway used by a particular plant species (C 3 , C 4 or CAM) and atmospheric conditions such as vapour pressure deficit and CO 2 concentration23. We compile WUE data and couple these to atmospheric vapour pressure deficits and C 3 –C 4 vegetation abundances to develop global grids of WUE (Methods). Catchment transpiration fluxes and gridded WUEs are applied to calculate gross primary production within each catchment (Fig. 3c). On the global scale, we weight our global grids of WUE according to vegetation density and calculate the global WUE of the terrestrial biosphere to be 3.2 ± 0.9 mmol CO 2 per mol H 2 O. Applying this ratio to our global transpiration flux (62,000 ± 8,000 km3 yr−1 of H 2 O), we calculate gross primary production to be 129 ± 32 GtC yr−1. This flux is consistent with a recent estimate of 123 ± 8 GtC yr−1 (ref. 6) and provides a fifth line of support for the large transpiration fluxes reported in our work.

Linkages made here between the water and carbon cycles highlight a new stable-isotope-based methodology that can be used to monitor and map ongoing changes to Earth’s water cycle24,25 as well as modifications to carbon assimilation rates under increased atmospheric temperatures and CO 2 concentrations. Our results show that the water and carbon cycles are linked in such a way that transpiration must account for more than 80% of continental evapotranspiration to maintain a mass balance between these two biogeochemical fluxes (plant transzpiration and CO 2 uptake). Given the importance of transpiration, it follows that the physiological response of vegetation to a warmer and CO 2 -enriched atmosphere will have a dominant effect on future changes to evapotranspiration and the terrestrial hydrological cycle. Furthermore, changes in natural ecosystems via land-use modifications or climate changes will have notable effects on river discharges and, consequently, fluvial sediment loads, chemical weathering on continents, and atmospheric latent heat transport.

Climate change is expected to affect global transpiration3. Considering the dominance of transpiration in continental evapotranspiration shown here, future changes in global transpiration will affect land temperatures by altering latent heat fluxes from continents, and will also change the fraction of precipitation entering rivers. Our catchment-scale results can be applied as a calibration tool for climate models, which should shift the prevailing focus on physical climate data towards ecosystem water requirements and so result in better predictions of continental evapotranspiration and water recycling in a warmer future climate.