Last-Century Increases in Intrinsic Water-Use Efficiency of Grassland Communities Have Occurred over a Wide Range of Vegetation Composition,Nutrient Inputs,and Soil pH

Last-century climate change has led to variable increases of the intrinsic water-use efficiency (W_i; the ratio of net CO₂ assimilation to stomatal conductance for water vapor) of trees and C₃ grassland ecosystems, but the causes of the variability are not well understood. Here, we address putative drivers underlying variable W_i responses in a wide range of grassland communities. W_i was estimated from carbon isotope discrimination in archived herbage samples from 16 contrasting fertilizer treatments in the Park Grass Experiment, Rothamsted, England, for the 1915 to 1929 and 1995 to 2009 periods. Changes in W_i were analyzed in relation to nitrogen input, soil pH, species richness, and functional group composition. Treatments included liming as well as phosphorus and potassium additions with or without ammonium or nitrate fertilizer applications at three levels. W_i increased between 11% and 25% (P < 0.001) in the different treatments between the two periods. None of the fertilizers had a direct effect on the change of W_i (ΔW_i). However, soil pH (P < 0.05), species richness (P < 0.01), and percentage grass content (P < 0.01) were significantly related to ΔW_i. Grass-dominated, species-poor plots on acidic soils showed the largest ΔW_i (+14.7 μmol mol⁻¹). The ΔW_i response of these acidic plots was probably related to drought effects resulting from aluminum toxicity on root growth. Our results from the Park Grass Experiment show that W_i in grassland communities consistently increased over a wide range of nutrient inputs, soil pH, and plant community compositions during the last century.The intrinsic water-use efficiency (W_i) of plants is controlled by photosynthetic carbon assimilation and stomatal conductance via the leaf-level coupling of CO₂ and water fluxes. A general, but variable, increase of W_i under rising atmospheric CO₂ has been observed in long-term studies (Peñuelas et al., 2011; Franks et al., 2013; Saurer et al., 2014), but little is known about other environmental or ecosystem factors, which may interact with the effect of increasing CO₂ on W_i. An improved understanding of putative interactive mechanisms is important because changes in W_i may have significant effects on the global terrestrial carbon and water cycles (Gedney et al., 2006; Betts et al., 2007). This study explores the interactive effects of the increase in atmospheric CO₂ (observed over the last century), nutrient loading, and soil pH together with other related effects on plant species richness and functional group composition on the coupling of plant CO₂ and water fluxes in a seminatural grassland in southeastern England.W_i is a leaf-level efficiency that has also been termed potential water-use efficiency or physiological water-use efficiency, as it excludes the direct influence of vapor pressure deficit (VPD), a parameter determined by environmental conditions, on leaf-level water-use efficiency (Farquhar et al., 1989; Franks et al., 2013). W_i reports the relationship between net CO₂ assimilation rate (A_n) and stomatal conductance for water vapor (g_H2O):(1)According to the first law of Fick, A_n can be given as the product of the stomatal conductance for CO₂ (g_CO2) and the concentration gradient between the atmosphere (c_a) and the leaf internal gas space (c_i): A_n= g_CO2 (c_a – c_i). Using g_CO2 (c_a – c_i) instead of A_n in Equation 1, replacement of g_H2O/g_CO2 by the numerical value of g_H2O/g_CO2 (1.6) and rearrangement yields the following alternative expression of W_i:(2)Equation 2 reveals that past changes of W_i must have been controlled by two parameters: the change of c_a and the concurrent change of 1 – c_i/c_a, the relative gradient for CO₂ diffusion into the leaf (Franks et al., 2013). A change in the relative gradient is determined by the changes in A_n relative to g_H2O, as leaves respond to changing c_a and other environmental factors. In particular, Equation 2 shows that any variation in the climate change response of W_i is determined by the c_i/c_a response, if the comparison is made for vegetation at the same location and in the same period of time.Studies with C₃ vegetation, including trees/forests and C₃ grasslands, have revealed a general increase of W_i in the last century (Bert et al., 1997; Duquesnay et al., 1998; Feng, 1999; Arneth et al., 2002; Saurer et al., 2004; Barbosa et al., 2010; Köhler et al., 2010; Andreu-Hayles et al., 2011). In many cases, c_i/c_a, estimated by ¹³C discrimination (Farquhar et al., 1989), varied relatively little. Indeed, it has been suggested, based on theoretical grounds and empirical evidence from studies over geological/evolutionary to short time scales, that adaptive feedback responses will tend to maintain c_i/c_a approximately constant (Ehleringer and Cerling, 1995; Franks et al., 2013), as plants optimize carbon gain with respect to water loss (Cowan and Farquhar, 1977). Yet, c_i/c_a-dependent variation in the W_i response to climate change has also been noted (Peñuelas et al., 2011; Köhler et al., 2012) over the last century, indicating that additional factors, perhaps including other global change drivers, can modify the W_i response over this time scale, at least transiently. A meta-analysis by Peñuelas et al. (2011) reports c_i/c_a-dependent increases of W_i for different forests between 6% and 36% from the early 1960s to 2000s. A recent study by Saurer et al. (2014) on European forest trees found increases in W_i ranging from 1% to 53% during the last century. The strongest increase of W_i was recorded in regions where summer soil-water availability decreased in the last century. For different grassland communities, the c_i/c_a-dependent increases of W_i varied between 13% and 28% at one site (Köhler et al., 2012) from 1915 to 2009. Evidently, such variation can have important repercussions for the coupling of terrestrial CO₂ and water fluxes. Yet, little is known about the mechanism(s) underlying the variation.At the Park Grass Experiment (PGE) at Rothamsted, England, Köhler et al. (2012) observed a nitrogen supply-dependent enhancement of the W_i response on plots receiving nitrate fertilizer and maintained at a near-neutral soil pH by liming. However, the actual relationship between nitrogen supply and W_i response did not hold when the unlimed control (soil pH approximately 5.2) was included in the comparison. Remarkably, however, there was a significant positive relationship between the grass content of the community and the W_i response of the experimental plots in the investigation. These results suggested that the effect of nutrient supply on the W_i response of the grassland communities was indirect, perhaps working via effects on soil pH and/or vegetation composition (plant species richness or functional group composition).The PGE provides a unique opportunity to study century-scale variation in the c_i/c_a-dependent variation of W_i for a wide range of diverse grassland communities. Much of the extant ecosystem-scale variability of plant species richness and soil pH in temperate grasslands of Europe (Ceulemans et al., 2014) is included in the range of plot-scale plant species richness and soil pH at the PGE (which is reported in this investigation). The different long-term applications of fertilizer and lime over the past century have resulted in substantial changes in soil pH, species richness, and grass content on the experimental plots, but in most cases, within-plot changes over the study period considered here (1915–2009) were comparatively small (Crawley et al., 2005; Silvertown et al., 2006). All experimental plots are located at the same site and are exposed to the same weather conditions. Consequently, trends in climate as a direct driver for differences in W_i between plots can be ruled out.Here, we explore putative mechanisms underlying eventual c_i/c_a-dependent variation of W_i during the last century at the PGE by, first, quantifying the sustained effect of a wide range of contrasting fertilizer treatments (n = 16) on the change of W_i during the last century and, second, analyzing the relationships between the observed W_i response of treatments and the respective nutrient status, soil pH, plant species richness, and plant functional group composition of the grassland communities.