20th century changes in carbon isotopes and water-use efficiency: tree-ring-based evaluation of the CLM4.5 and LPX-Bern models

Abstract

Abstract. Measurements of the stable carbon isotope ratio (δ 13 C) on annual tree rings offer new opportunities to evaluate mechanisms of variations in photosynthesis and stomatal conductance under changing CO 2 and climate conditions, especially in conjunction with process-based biogeochemical model simulations. The isotopic discrimination is indicative of the ratio between the CO 2 partial pressure in the intercellular cavities and the atmosphere (c i /c a ) and of the ratio of assimilation to stomatal conductance, termed intrinsic wateruse efficiency (iWUE). We performed isotope-enabled simulations over the industrial period with the land biosphere module (CLM4.5) of the Community Earth System Model and the Land Surface Processes and Exchanges (LPX-Bern) dynamic global vegetation model. Results for C3 tree species show good agreement with a global compilation of δ 13 C measurements on leaves, though modeled 13 C discrimination by C3 trees is smaller in arid regions than measured. A compilation of 76 tree-ring records, mainly from Europe, boreal Asia, and western North America, suggests on average small 20th century changes in isotopic discrimination and in c i /c a and an increase in iWUE of about 27 % since 1900. LPXBern results match these century-scale reconstructions, supporting the idea that the physiology of stomata has evolved to optimize trade-offs between carbon gain by assimilation and water loss by transpiration. In contrast, CLM4.5 simulates an increase in discrimination and in turn a change in iWUE that is almost twice as large as that revealed by the tree-ring data. Factorial simulations show that these changes are mainly in response to rising atmospheric CO 2 . The results suggest that the downregulation of c i /c a and of photosynthesis by nitrogen limitation is possibly too strong in the standard setup of CLM4.5 or that there may be problems associated with the implementation of conductance, assimilation, and related adjustment processes on long-term environmental changes.

Limited evidence forCO2‐related growth enhancement in northern Rocky Mountain lodgepole pine populations across climate gradients

“…Two exceptions to the relatively weak increase in iWUE were the low and mid elevation sites of the west transect, where increases were nearly double those of other sites (Figures c and a). The increases at these sites were far more in line with observations of temporal changes in iWUE globally (Keller et al., ; Peñuelas et al., ; Silva & Anand, ).…”

Observed and modelled historical trends in the water‐use efficiency of plants and ecosystems

“…With one exception (Battipaglia et al, ), free‐air carbon dioxide enrichment (FACE) experiments studying the effects of elevated c a on plants and ecosystems under natural, open‐air conditions have shown responses broadly consistent with the third strategy, that is constant c i / c a (Ainsworth & Long, ; Gimeno et al, ; De Kauwe et al, ), giving rise to the expectation that plants growing in normal conditions should behave accordingly. Most tree‐ring studies in natural environments have yielded results consistent with this strategy (Frank et al, ; Keller et al, ; Peñuelas et al, ; Saurer et al, ). However, there is also empirical support for plants following scenario (i) with constant c i (Dorado Liñán et al, ) or scenario (ii) where c i increases at the same rate as c a (McCarroll et al, ; Treydte et al, ).…”

“…Leaf-level studies using stable carbon isotopes in tree rings, employing Equation (4), have indicated an increase of iWUE leaf of 0.2 ± 0.1% year -1 over the 20th century (e.g. Frank et al, 2015;Keller et al, 2017;Peñuelas, Canadell, & Ogaya, 2011;Saurer et al, 2014;Silva & Anand, 2013), approximately proportional to the recent increase in c a (Figure 1b). Some of these studies have shown a weakening iWUE leaf response to c a towards the end of the 20th century in regions where increases both in temperature and in D may have offset the increase in iWUE leaf due to increasing c a (Gagen et al, 2011;Waterhouse et al, 2004).…”

“…CMIP phase 6 (CMIP6) has strongly recommended the implementation of carbon isotope formulations in models (Jones et al, 2016). Carbon isotopes have been included in a few models, including LPJ (Scholze, Ciais, & Heimann, 2008), LPX-Bern (Keller et al, 2017;Saurer et al, 2014), and CLM4.5 (Duarte et al, 2017;Keller et al, 2017;Raczka et al, 2016). In general, however, vegetation models calculate ∆ 13 C following the simple model of Farquhar, O'Leary, et al (1982), that is Equation (3), and the simulations are then compared to ∆ 13 C directly inferred from tree rings using Equation (2) between results obtained from the different data sources, and (c) different modelling approaches and assumptions for WUE.…”

“…As a result, data‐based estimates of WUE are not easily comparable and often appear to disagree with one another (e.g. Frank et al, ; Keenan et al, ; Keller et al, ), obscuring the interpretation of the results. Even though inherent differences between plant functional types (PFTs) have been reported (Brodribb, McAdam, Jordan, & Feild, ; Lin et al, ), and though varying site conditions may induce different physiological responses, the large variation in WUE trends between tree‐ring‐based estimates and those from eddy‐covariance measurements reported in the literature is unexpected (Medlyn & De Kauwe, ) and at odds with current understanding of scaling.…”