Advertisement

Oecologia

, Volume 187, Issue 4, pp 995–1007 | Cite as

Assessing the interplay between canopy energy balance and photosynthesis with cellulose δ18O: large-scale patterns and independent ground-truthing

  • Brent R. Helliker
  • Xin Song
  • Michael L. Goulden
  • Kenneth Clark
  • Paul Bolstad
  • J. William Munger
  • Jiquan Chen
  • Asko Noormets
  • David Hollinger
  • Steve Wofsy
  • Timothy Martin
  • Dennis Baldocchi
  • Eugenie Euskirchenn
  • Ankur Desai
  • Sean P. Burns
Special Topic

Abstract

There are few whole-canopy or ecosystem scale assessments of the interplay between canopy temperature and photosynthesis across both spatial and temporal scales. The stable oxygen isotope ratio (δ18O) of plant cellulose can be used to resolve a photosynthesis-weighted estimate of canopy temperature, but the method requires independent confirmation. We compare isotope-resolved canopy temperatures derived from multi-year homogenization of tree cellulose δ18O to canopy-air temperatures weighted by gross primary productivity (GPP) at multiple sites, ranging from warm temperate to boreal and subalpine forests. We also perform a sensitivity analysis for isotope-resolved canopy temperatures that showed errors in plant source water δ18O lead to the largest errors in canopy temperature estimation. The relationship between isotope-resolved canopy temperatures and GPP-weighted air temperatures was highly significant across sites (p < 0.0001, R2 = 0.82), thus offering confirmation of the isotope approach. The previously observed temperature invariance from temperate to boreal biomes was confirmed, but the greater elevation of canopy temperature above air temperature in the boreal forest was not. Based on the current analysis, we conclude that canopy temperatures in the boreal forest are as warm as those in temperate systems because day-time-growing-season air temperatures are similarly warm.

Keywords

Stable oxygen isotope Energy balance Canopy temperature Leaf temperature Gross primary production δ18

Introduction

The effect of leaf temperature on photosynthesis is of fundamental importance to plant productivity, distribution, and ecosystem-level carbon and water exchange (Schimper 1903; Walter et al. 1975; Long and Woodward 1988; Larcher 1995), yet because of measurement difficulties, there are few whole-canopy or ecosystem scale assessments of the interplay between temperature and photosynthesis across both spatial and temporal scales. A combination of abiotic and biotic factors control leaf temperature which can deviate from ambient temperatures through variation in absorbed radiation, transpiration and convective heat loss (Raschke 1960; Gates 1962, 1965; Ehleringer 1989). Variation in biotic factors can cause leaf temperatures to be as much as 10 °C above ambient temperature by leaf clumping, which limits convective heat transfer by increasing the branch-level boundary layer (Smith and Carter 1988) or 18 °C below ambient temperature through evaporative cooling and reduction of absorbed radiation via reflective leaf hairs (Ehleringer et al. 1976; Smith 1978). It is also well-established that optimal leaf temperatures for photosynthetic uptake can vary by species, and shifts with seasonal air temperature (Berry and Bjorkman 1980; Long and Woodward 1988; Michaletz et al. 2016). Simultaneous measurements of whole-canopy or ecosystem scale temperature and photosynthesis has been hindered by measurement difficulties (Miller 1971). Whole-plant photosynthesis systems do exist (Barton et al. 2010), but they necessarily prevent the assessment of canopy energy balance in situ. Thermal-imaging cameras allow for remote, season-long canopy-scale temperature measurements in situ (Leuzinger and Körner 2007; Aubrecht et al. 2016), but to gain information on the relationship between canopy temperatures and carbon uptake, these instruments must be used in concert with eddy-covariance measurements of canopy carbon exchange.

The analysis of the stable oxygen isotope ratio (δ18O) of plant cellulose represents an integration of canopy energy balance and photosynthesis. The stable isotope ratio of plant cellulose is weighted more heavily by periods of maximal carbon uptake (Farquhar et al. 1989; Cernusak et al. 2005; Gessler et al. 2007), therefore, whatever information that is contained in cellulose δ18O is also photosynthesis-weighted. As much as 60% of the δ18O of plant cellulose emanates from the δ18O of leaf water (Barbour et al. 2000). The enrichment of 18O in leaf water during plant transpiration is controlled in large part by the ratio of ambient water vapor pressure (ea) to the saturated water vapor pressure at leaf temperature (ei). With the δ18O of water entering a leaf and atmospheric water vapor constant, there is an approximate 0.2‰ change in leaf water δ18O for every 0.01 change in ea/ei, so the ea/ei signal is particularly large. If canopy temperature is equal to air temperature (such that ei is equal to ambient saturated water vapor pressure), then leaf water δ18O, and subsequently cellulose δ18O, can be used to reconstruct relative humidity during a growing season (Saurer et al. 1997, 2000; Anderson et al. 1998; Roden et al. 2000; Robertson et al. 2001; Wright and Leavitt 2006; Porter et al. 2009). If ea during the period of growth is known, then cellulose δ18O can be further deconstructed to obtain a record of ei. Once ei is obtained, solving for photosynthesis-weighted canopy temperature is relatively straight forward, because saturated water vapor pressure has a well-established relationship with temperature (Buck 1981).

It has been known for decades that the oxygen isotope ratio of tree cellulose correlates with some component of ambient temperature (Gray and Thompson 1976; Epstein et al. 1977; Yakir 1992), and Helliker and Richter (2008) (Hereafter H&R) further recognized that the δ18O of cellulose contains an isotope-resolved, photosynthesis-weighted canopy temperature (hereafter Tcanδ). They found a relatively constant value (approximately 21 °C) across tree species of boreal, temperate and subtropical forested biomes, where mean annual temperature (MAT) ranged from − 9 to 24 °C. Song et al. (2011) extended the approach to a larger dataset of tree-ring δ18O and confirmed the results of H&R showing a narrow range of Tcanδ across boreal, temperate and subtropical biomes, but the Tcanδ of tropical trees were clearly warmer, ranging from 25 to 28 °C, and subalpine trees much cooler, around 10 °C. Additionally, Flanagan and Farquhar (2014) applied the approach to grasses and found that Tcanδ estimates were similar to those observed via infrared thermometry. These independent data sets confirm that the isotope approach is consistent, but there are still many questions concerning isotopic fractionation factors (Sternberg and Ellsworth 2011), the proportion of the leaf water isotopic signal that is retained in cellulose within a season and across species (Gessler et al. 2009; Song et al. 2014b) and, perhaps most importantly, there is a lack of an independent test on how well Tcanδ matches other measures of canopy photosynthesis and temperature. In short, independent ground-truthing is needed.

Our goal here is to establish an independent comparison of Tcanδ using eddy-covariance data and canopy air temperature that will both test the efficacy of the isotope approach, and coarsely answer some of the hypotheses that emanate from H&R. The relatively small envelope of Tcanδ observed by H&R can have two explanations that are not mutually exclusive: (1) the biophysical and physiological components of tree canopies allow for some degree of homeothermy and (2) canopy temperatures tend to match canopy air temperature, and seasonal averages of canopy temperature weighted by photosynthesis simply do not vary that much across biomes. Ground-truthing of the Tcanδ method to assess these two hypotheses is somewhat problematic because, as detailed above, there are few comparable metrics. We can, however, use eddy-covariance measurements to assess relationships between estimated gross primary productivity (GPP) and canopy-air temperature. This comparison is less than ideal because eddy-covariance provides an ecosystem-scale measurement and isotopes provide measurements on the scale of the individual, but we can gain some measure of efficacy for the isotope approach if there is general agreement between GPP-weighted canopy temperatures and Tcanδ.

Methods

Site selection and sampling scheme

From 2009 to 2010 we sampled tree cores from 15 forested eddy-covariance sites; for a subset of these sites, water from stems was extracted for isotope analysis. 13 of the sites approximated the biome distribution used by H&R, and two of the sites were subalpine forests that were chosen as low-temperature growing season sites (Table 1). The sites spanned 34° in latitude and more than 3100 m in elevation. The mean annual temperature (MAT) for the period of study ranged from − 3.5 to 20.3 °C. The number of data-years used at each site ranged from 2 to 9, and was determined by the availability of gap-filled, quality-controlled flux data at the time of tree-core sampling.
Table 1

Sites and species within sites used in this study

Ameriflux site

Location

Coordinates

Elevation

Species examined

Years for tree-ring/flux

Austin cary

FL, USA

29.7381, − 82.2188

44

Slash pine, Pinus elliottii

2002–2003

Donaldson

FL, USA

29.7547, − 82.1633

36

Slash pine, Pinus elliottii

2002–2004

NC loblolly pine

NC, USA

35.8031, − 76.6679

12

Loblolly pine, Pinus taeda

2005–2006

Walker branch

TN, USA

35.9588, − 84.2874

343

Chestnut oak, Quercus michauxii

White oak, Quercus alba

1995–1999

Silas little

NJ, USA

39.9712, − 74.4345

48

Pitch pine, Pinus rigida

Black oak, Quercus velutina

2005–2006

Niwot ridge

CO, USA

40.0329, − 105.5464

3050

Subalpine fir, Abies lasiocarpa engelmann spruce, Picea engelmannii lodgepole pine, Pinus contorta Aspen, Populus tremuloides

1999–2007

GLEES

WY, USA

41.3644, − 106.2394

3190

Subalpine fir, Abies lasiocarpa

Engelmann spruce, Picea engelmannii

2002–2003

Oak openings

OH, USA

41.5545, − 83.8438

230

Black oak, Quercus velutina

2004–2005

Harvard forest

MA, USA

42.5378, − 72.1715

340

Red maple, Acer rubrum

2002–2006

Bartlett forest

NH, USA

44.0646, − 71.2881

272

Sugar maple, Acer saccharum

2004–2006

Howland forest

ME, USA

45.2041, − 68.7402

61

Sugar maple, Acer saccharum

Hemlock, Tsuga canadensis

2000–2004

Willow creek

WI, USA

45.806, − 90.0798

515

Sugar maple, Acer saccharum

1999–2004

Thompson 1930

MB, Canada

55.9058, − 98.5247

257

Black spruce, Picea mariana

2002–2004

Thompson 1964

MB, Canada

55.9117, − 98.3822

258

Black spruce, Picea mariana

2002–2004

Fairbanks

AK, USA

63.9198, − 145.3781

518

Black spruce, Picea mariana

2010–2011

GPP-weighted canopy air temperature

To develop gross primary productivity (GPP)-weighted estimates of canopy air temperature, we used air temperature, relative humidity (rH) and GPP from the MDS product. GPP is a robust, data-derived quantity (Baldocchi and Sturtevant 2015) and while different methods to resolve GPP result in an approximate 10–15% variance in seasonal to annual GPP, comparisons across sites are robust when a consistent method is used (Desai et al. 2008). The relationship between nocturnal turbulence data and nighttime net ecosystem exchange (NEE) has been analyzed extensively at each site to achieve ecosystem respiration estimates that reflect well-coupled periods. Most sites use a cut-off value for friction velocity (u*; m s−1) that results in a near-linear relationship between nighttime NEE and air or canopy temperature, binned by temperature increments at progressively greater values of u*. For example, the flat, relatively open stands of the Silas Little, NJ site achieve adequate coupling at a friction velocity values > 0.2 m s−1, while Niwot Ridge, CO has used a values > 0.4 m s−1.

Sites have then gap-filled nighttime and daytime data using standardized procedures, and summed measured and modeled daytime NEE, and gap-filled nighttime and estimated daytime ecosystem respiration to estimate GPP (Falge et al. 2001; Moffat et al. 2007). Each year of flux and meteorological data from each site was initially analyzed to ensure a complete growing season of data existed. We first determined the growing season start and end days, which consisted of removing the nighttime periods when PAR = 0, then averaging the GPP for the remaining light periods of each day. The first day of the growing season was determined to be the first day of seven continuous days with an average GPP greater than 1 g m−2 s−1. The last day of the growing season was determined to be the last day above 1 g m−2 s−1 before seven continuous days below 1 g m−2 s−1.

To independently test Tcanδ we developed an ecosystem-based approach that provides a GPP-weighted canopy air temperature (hereafter Ta-GPP). To arrive at Ta-GPP, we first filtered eddy-covariance data for daytime periods only for the entire study period (2–9 years, depending on the site). At each half-hour time step, GPP was multiplied by ambient air temperature measured at or near canopy height. The sum of these products was then divided by total growing-season GPP to arrive at Ta-GPP (see Eq. 1 below for more details). For three sites, infrared (IR) thermometers (Apogee, Logan, UT, USA) were placed above the canopy to obtain direct canopy temperatures, and these measurements were used to develop GPP-weighted canopy temperatures (TIR-GPP; see Eq. 1 below for more details).
$$T_{\text{GPP}} = \frac{{\mathop \sum \nolimits_{{i = + {\text{GPP}}}}^{N} \left( {{\text{GPP}}_{i} \times T_{x - i} } \right)}}{{\mathop \sum \nolimits_{{i = + {\text{GPP}}}}^{N} {\text{GPP}}_{i} }},$$
(1)
Here i is defined by the first period of positive GPP, or the start of the growing season, N equals all half-hour periods during the growing season with a positive GPP and Tx equals half-hour measures of either ambient air temperatures at canopy height (Ta-GPP) or IR-measured canopy temperature (TIR-GPP). We examined the effect of growing season length on our calculated values of Ta-GPP for individual years at boreal, subalpine and warm temperate sites by extending and contracting the growing season by 6 days. The largest difference in Ta-GPP from our chosen filtering approach was 0.2 °C at the boreal site.

Isotope model description and parameterization

The record of photosynthesis-weighted canopy temperature contained in the stable oxygen isotope ratio (δ18O) of cellulose begins with the enrichment of 18O in leaf water during transpiration. During transpiration, the water molecules containing light isotopes evaporate preferentially, resulting in an enrichment of heavy isotopes in the leaf mesophyll above the isotope ratio of water entering the roots. This enrichment process can be described by the following model developed initially by Craig and Gordon (1965)and modified by Flanagan et al. (1991) and Farquhar and Lloyd (1993),
$$\Delta_{e} = \varepsilon^{*} + \varepsilon_{\text{K}} + \left( {\Delta_{\text{V}} - \varepsilon_{\text{K}} } \right)\frac{{e_{a} }}{{e_{i} }},$$
(2)
where ∆e is the 18O enrichment of water at the evaporative sites above plant source water (approximated by δ18Oe–δ18Osource). ∆V is the δ18O of atmospheric water vapor relative to source water and ε* is the temperature-dependent equilibrium fractionation factor for the evaporation of water and εK is the cumulative kinetic fractionation factor of water vapor diffusing through leaf stomata and the leaf boundary layer (Farquhar and Lloyd 1993). The relative exchange of isotopes between atmospheric water vapor and xylem water is represented by ea/ei (Helliker and Griffiths 2007), which is the ambient vapor pressure divided by the saturation vapor pressure at leaf temperature. ea/ei can be also viewed as leaf-based relative humidity.
The general model to describe cellulose δ18O is (Barbour and Farquhar 2000; Roden et al. 2000):
$$\Delta_{\text{C}} = \left( {\frac{{\Delta_{e} \left( {1 - e^{ - \wp } } \right)}}{\wp }} \right)\left( {1 - p_{\text{ex}} \times p_{x} } \right) + \varepsilon_{\text{O}} ,$$
(3)
where cellulose δ18O is expressed as an enrichment above source water (∆C). The term [(1 − e−℘)/℘] is the Péclet correction and describes the species-specific effect that leaf architecture and water loss have on the balance of enriched evaporative-site water and unenriched vein water in the leaf (Farquhar and Lloyd 1993). The variable px represents the proportion of cellulose-synthesis water that is comprised of unenriched water at the site of cellulose-synthesis. The proportion of oxygen atoms in a sucrose molecule that exchange with this synthesis water is represented by pex. The equilibrium fractionation factor associated with the exchange of oxygen atoms between carbonyl group and the tissue water is represented by εO (Sternberg 1989; Yakir and DeNiro 1990).
To solve for stable isotope-resolved, photosynthesis-weighted canopy temperature (Tcanδ) we insert Eq. 2 into Eq. 3 and solve for the ei that satisfies observed ∆C:
$$e_{i} = \frac{{\left( {\Delta_{\text{V}} - \varepsilon_{\text{K}} } \right)e_{a} }}{{\left( {\frac{{\left( {\Delta_{\text{C}} - \varepsilon_{\text{O}} } \right)\wp }}{{\left( {1 - p_{\text{ex}} p_{x} } \right)\left( {1 - e^{ - \wp } } \right)}}} \right) - \varepsilon^{*} - \varepsilon_{\text{K}} }}.$$
(4)
Saturated water vapor pressure has a well-quantified relationship with temperature and, therefore, a given saturated vapor pressure yields a unique temperature for ei (Buck 1981):
$$T_{{{\text{can}}\delta }} = \frac{{240.97\left( {\ln \frac{{e_{i} }}{0.61365}} \right)}}{{17.502 - \left( {\ln \frac{{e_{i} }}{0.61365}} \right)}}.$$
(5)

To populate model predictions for ei and Tcanδ we used px = 1(Roden and Ehleringer 1999; Cernusak et al. 2005), pex = 0.4 for angiosperms and pex = 0.26 for gymnosperms based on the work of Song et al. (2014b). It has been reported that pex may change dynamically within and across species (Gessler et al. 2009; Song et al. 2014a; Cheesman and Cernusak 2017). However, the majority of pex values have been found to be around 0.4 (Gessler et al. 2014), the value we use for angiosperms. We feel that the work of Song et al. (2014b) justifies the use of pex = 0.26 for gymnosperms, but we did examine the Tcanδ in gymnosperms with both pex values.

V was assumed to be in equilibrium with source water at the mean growing season temperature (Helliker 2014).We assumed a constant \(\wp\) of 0.105, yielding a 5% offset from ∆LW (following H&R) such that Eq. 3, the Péclet model, collapses to a two pool model with 5% unenriched water. At each site, ea was obtained from Ta-GPP and from GPP-weighted ambient relative humidity (rH) which is calculated in a manner similar to Ta-GPP. We used two approaches for εO, we assumed a constant εO of 27.2‰ and a temperature-dependent εO. The temperature-dependent εO was determined from the empirical equation of Sternberg and Ellsworth (2011):
$$\varepsilon_{O} = \, 0.0084T^{2} - \, 0.51T \, + 33.172,$$
(6)
where T represents the temperature at which cellulose is synthesized. For this study, we used mean 24 h growing-season air temperature at each site as the value for T.

Sample processing and mass spectrometry measurements

Tree cores were obtained from radially symmetric, dominant tree species (Ramesh et al. 1985) using a 5 mm diameter increment borer (Haglof, Langsele, Sweden). Cores were collected from five trees per species. The cores were sanded lightly to make annual ring boundaries clearly visible. Rings that corresponded to years for which complete flux data were available were removed with an exacto knife under a light microscope. The samples were then homogenized with a Wiley mill and α-cellulose was extracted following the Brendel procedure modified with an addition of 17% NAOH step to remove hemicellulose (Brendel et al. 2000; Gaudinski et al. 2005). 90–100 μg cellulose samples were weighed into silver capsules. For isotope analysis, cellulose samples were pyrolysed at 1100 °C in a Costech Elemental analyzer. Isotopic composition of the evolved CO gas was determined on a Thermo-Finnigan Delta Plus isotope ratio mass spectrometer, which had a measurement precision of less than 0.23‰ on a standard reference cellulose powder. All cellulose samples were run in triplicate and data were reported on the Standard Mean Ocean Water (SMOW) scale. Tissue water from stem samples was extracted by cryogenic vacuum distillation (West et al. 2006). Water samples (0.5 ml) were analyzed by equilibration for 48 h in 3 ml Exetainer® vials (Labco Limited, UK) with 10/90 mixture CO2/He. Four replicates of 100 ml of the headspace gas was injected into a gas chromatograph and carried in a helium air stream to a Delta Plus isotope ratio mass spectrometer (Thermo-Finnigan, Germany).

Sensitivity analysis

To examine how errors in measured model inputs affect the calculation of Tcanδ, we performed a sensitivity analysis using initial values of ∆C = 30‰, Tair = 20 °C, rH = 50% (assuming ambient rH and leaf-based rH are equal), δ18Osource = − 10‰, δ18O water vapor = − 19.5‰. Offsets for each of these inputs were developed as follows: for the derivation of air temperature and rH errors, we took the mean difference between Ta-GPP and mean daytime air temperature at each site. A similar difference was derived from GPP-weighted rH and mean daytime rH. We chose this method recognizing that most field sites will not be fully equipped with eddy-covariance data to derive Ta-GPP and GPP-weighted rH, but may have nearby weather station data to drive isotope models. This resulted in an air temperature error term of ± 2.6 °C (e.g., air temperature inputs for the sensitivity analysis ranged from 17.4 to 22.6 °C) and rH of ± 5.2%. For δ18Osource and water vapor δ18Oerrors, we took the mean difference between measured and modeled stem δ18O and applied this offset (2.7‰) to both stem and vapor δ18O. This method was chosen to assess the potential error in Tcanδ that arises from having model-only values and/or being highly incorrect on the parameterization of plant source water δ18O.

Results

Using either a constant εO of 27.2‰ or the temperature-dependent εO led to very similar site averages for Tcanδ (Fig. 1), with the exception of the subalpine sites where growing-season-air temperatures were comparatively low. Across sites, the average difference between Tcanδ using temperature-dependent versus constant εO was 0.3 °C, and this small difference was driven primarily by the fact that subalpine Tcanδ was 2.8 °C higher when using a constant εO. The relationship of Tcanδ versus MAT was not significant for a constant εO, and was significant for temperature-dependent εO (F1,14 = 4.98, p < 0.05). After removing the subalpine sites, there was no significant relationship with Tcanδ and MAT from temperate to boreal biomes.
Fig. 1

Site averages for stable oxygen isotope reconstruction of photosynthesis-weighted canopy temperatures (Tcanδ). Tcanδ was calculated using both a constant water-to-carbonyl-group fractionation (εO) of 27.2‰ (open circles) and a temperature-dependent εO (closed circles). Error bars represent standard error and are occasionally smaller than the symbol, n = 5 for each species at each site

To examine the relationship between Tcanδ and Ta-GPP we used the temperature-dependent εO only (Fig. 2). The relationship was highly significant across sites (F1,23 = 99.2, p < 0.0001, R2 = 0.82, Tcanδ = − 3.96 + 1.15Ta-GPP). Canopy over temperature, the difference between Tcanδ and Ta-GPP, had no significant relationships with mean growing season temperatures or MAT across sites, but after removing data from the Niwot Ridge site there was a significant relationship with MAT (F1,19 = 5.3, p < 0.05, R2 = 0.22). Separating results into broad tree-functional types showed that mean gymnosperm Tcanδ was 1.6 °C below Ta-GPP, and mean angiosperm Tcanδ was 0.04 °C below Ta-GPP. Both of these were biased by the relatively low Tcanδ values at the Niwot Ridge site. When Niwot Ridge was excluded, the gymnosperm Tcanδ was 0.6 °C below Ta-GPP and angiosperm Tcanδ was 0.7 above Ta-GPP, but these differences were not significant in either case. When using a pex = 0.4 for gymnosperms, average gymnosperm Tcanδ was 3.3 °C higher than when using pex = 0.26, and the relationship to Ta-GPP changed. The mean gymnosperm Tcanδ shifted from 1.6 °C below Ta-GPP to 1.5 °C above Ta-GPP with Niwot Ridge included, and the temperature differences shifted from 0.6 °C below Ta-GPP to 2.9 °C above Ta-GPP with Niwot Ridge excluded.
Fig. 2

Stable oxygen isotope reconstruction of photosynthesis-weighted canopy temperatures (Tcanδ, closed circles) for individual species at each site and GPP-weighted, infrared-derived canopy temperature (TIR-GPP, open squares) vs. GPP-weighted air temperature (Ta-GPP). Tcanδ was calculated using a temperature-dependent εO. Error bars represent standard error and are occasionally smaller than the symbol, n = 5 for each species at each site

The reconstruction of Tcanδ was most sensitive to the measured input of δ18Osource, where errors were as high as 6.5 °C with a 2.7‰ error in δ18Osource (Fig. 3). Errors in air temperature of 2.6 °C yield about a 2.8 °C error in calculated Tcanδ, and a 5.2% error in rH gives a Tcanδ error of about 1.8 °C. The temperature and rH errors are approximately additive, so using both temperature and rH errors that are + 2.6 °C and + 5.2% yields a Tcanδ error of 4.6 °C (data not shown). Fortunately, temperature and RH tend to be negatively correlated, so that if the temperature used is lower than actual Ta-GPP (as mean daytime growing season temperatures tend to be), then the corresponding rH will be higher than GPP-weighted rH, and the errors tend to offset. This is demonstrated in the final two columns of Fig. 3 where using Tair + 2.6 °C and rH − 5.2% yields a relatively small error of about 1 °C. Using all errors together does not result in errors much larger than those associated with incorrect δ18Osource.
Fig. 3

Sensitivity analysis to examine how errors in model inputs affect the calculation of Tcanδ. Initial model inputs were ∆C = 30‰, Tair = 20 °C, rH = 50%, δ18Osource = − 10‰, δ18O water vapor = − 19.5‰. Offsets for each of these inputs were derived from difference in mean growing season data vs. GPP-weighted data, or the difference between observed and modeled isotopic inputs

We further examined the sensitivity of Tcanδ predictions by focusing on the two similar subalpine sites, Niwot Ridge andt GLEES (Table 2). At the GLEES site, model and observed δ18Osource differed by 0.6‰, the standard error of δ18Osource was ± 1.1‰, and there was little difference in observed cellulose δ18O between tree species (Table 3). For this site, only a temperature-dependent εO was needed for Tcanδ to match Ta-GPP relatively well. For Niwot Ridge, there was a large range in Tcanδ among species that emanated from a 4‰ range in the δ18O of cellulose and δ18Osource. There was also much greater variance in δ18Osource at this site (δ18Osource standard error as high as ± 7.4‰ in subalpine fir). For the site-level average of Tcanδ, modeled δ18Osource and a temperature-dependent εO led to the best match with both Ta-GPP and TIR-GPP.
Table 2

The sensitivity of Tcanδ predictions to model parameterization at two similar subalpine sites

Site

Niwot Ridge, CO

GLEES, WY

 

Model δ18Osource, constant εO (°C)

Model δ18Osource temp-dependent εO (°C)

Observed δ18Osource, constant εO (°C)

Observed δ18Osource, temp-dependent εO (°C)

Tir-gpp (°C)

Ta-GPP (°C)

Model δ18Osource, constant εO (°C)

Model δ18Osource, temp-dependent εO (°C)

Ta-GPP (°C)

Lodgepole pine

17.0 ± 1.2

13.4 ± 0.9

8.9 ± 0.6

6.9 ± 0.5

     

Subalpine fir

13.3 ± 1.0

10.5 ± 0.8

10.5 ± 0.8

8.2 ± 0.7

  

13.9 ± 0.4

10.4 ± 0.3

 

Englemann’s spruce

11.1 ± 1.0

8.7 ± 0.8

5.8 ± 0.7

4.2 ± 0.6

  

14.6 ± 0.8

10.9 ± 0.6

 

Site mean

13.8 ± 1.0

10.8 ± 0.8

8.4 ± 0.7

6.4 ± 0.6

11.7 ± 0.3

11.8

14.2 ± 0.6

10.6 ± 0.4

10.5

Table 3

Abiotic and isotopic model inputs along with resolved Tcanδ by species for each site

Site

MAT (°C)

Daytime growing season (°C)

24-h growing season temp (°C)

Ta-GPP (°C)

GPP-weighted rH (%)

δ18OC (‰)

δ18Osource (‰)

δ18Osource modeled

Temp-dependent εO (‰)

Tcanδ temp-dependent εO (°C)

Tcanδ, constant εO (°C)

Austin cary pine

20.3

23.2

20.3

24.8

61.3

31.6 ± 0.3

− 3.2 ± 1.7

− 4.1

26.3

22.5 ± 0.3

21.7 ± 0.3

Bartlett maple

7.7

17.5

16.0

18.5

55.8

27.6 ± 0.2

− 7.5 ± 0.8

− 10.7

27.2

16.0 ± 0.2

16.0 ± 0.2

Donaldson pine

20.3

23.1

20.3

24.7

60.9

29.1 ± 1.5

 

− 4.1

26.3

20.4 ± 1.0

19.7 ± 1.0

Fairbanks b. spruce

− 1.9

16.3

14.2

16.9

51.3

21.2 ± 0.7

 

− 20.5

27.6

18.1 ± 0.8

18.7 ± 0.9

Florida pine

20.3

23.1

20.3

24.7

60.9

30.7 ± 1.2

 

− 4.1

26.3

21.7 ± 1.0

20.9 ± 1.0

GLEES fir

− 0.6

8.7

7.6

10.5

46.6

28.0 ± 0.4

 

− 16.3

29.8

10.4 ± 0.3

13.9 ± 0.4

GLEES spruce

− 0.6

8.7

7.6

10.5

46.6

27.6 ± 0.3

− 16.9 ± 1.1

− 16.3

29.8

10.9 ± 0.6

14.6 ± 0.8

Harvard maple

8.0

14.8

13.7

19.6

64.8

27.9 ± 0.3

− 6.4 ± 0.7

− 10.1

27.8

18.0 ± 0.3

18.5 ± 0.3

Howland hemlock

6.3

13.7

11.7

17.4

58.3

27.1 ± 0.5

  

28.4

14.3 ± 0.4

15.3 ± 0.4

Howland maple

6.3

13.7

11.7

17.4

58.3

27.6 ± 0.7

− 9.5 ± 2.3

− 10.5

28.4

17.0 ± 0.9

18.6 ± 1.0

NC clearcut pine

17.6

17.3

15.9

21.8

60.5

30.3 ± 0.6

 

− 6.3

27.2

20.3 ± 0.6

20.3 ± 0.6

NC pine

17.6

17.6

15.9

21.8

60.5

31.1 ± 0.3

 

− 6.3

27.2

21.0 ± 0.3

21.0 ± 0.3

Niwot Aspen

2.5

10.2

8.6

11.8

45.0

26.5 ± 0.5

− 13.7 ± 1.3

− 15.3

29.4

6.9 ± 0.4

9.0 ± 0.5

Niwot Fir

2.5

10.2

8.6

11.8

45.0

28.2 ± 0.7

− 13.3 ± 7.4

− 15.3

29.4

8.2 ± 0.7

10.5 ± 0.8

Niwot pine

2.5

10.2

8.6

11.8

45.0

30.5 ± 0.6

− 9.7 ± 3.2

− 15.3

29.4

6.9 ± 0.5

8.9 ± 0.6

Niwot Spruce

2.5

10.2

8.6

11.8

45.0

26.5 ± 0.9

− 10.1 ± 3.5

− 15.3

29.4

4.2 ± 0.6

5.8 ± 0.7

Ohio oak

10.2

18.7

17.7

22.1

58.5

26.0 ± 0.5

 

− 8.3

26.8

20.2 ± 0.6

19.7 ± 0.6

Pine Barrens Oak

12.6

19.9

18.8

24.9

55.4

28.2 ± 1.1

− 7.9 ± 1.1

− 8.1

26.6

25.1 ± 1.5

24.2 ± 1.4

Pine Barrens pine

12.6

19.9

18.8

24.9

55.4

31.8 ± 0.5

− 6.6 ± 0.5

− 8.1

26.6

24.3 ± 0.5

23.6 ± 0.5

Thompson 1930 spruce

1.0

17.1

15.6

18.9

56.1

26.0 ± 0.3

 

− 16.3

27.3

22.7 ± 0.4

22.8 ± 0.4

Thompson 1964 spruce

− 3.5

13.6

12.3

15.1

60.1

25.2 ± 0.2

 

− 16.3

28.2

17.6 ± 0.2

18.8 ± 0.3

Walker Br.oak

14.8

21.7

18.4

23.9

59.5

27.8 ± 1.0

− 9.7 ± 0.7

− 5.9

26.6

27.4 ± 1.6

26.4 ± 1.5

Walker Br. w. oak

14.8

21.7

18.4

23.9

59.5

28.8 ± 0.5

− 8.3 ± 0.4

− 5.9

26.6

26.4 ± 0.8

25.5 ± 0.7

WillowCreek maple

5.9

16.8

15.5

19.3

62.3

27.2 ± 0.3

 

− 10.5

27.3

22.0 ± 0.5

22.1 ± 0.5

Discussion

Plant cellulose δ18O can be used to reconstruct photosynthesis-weighed canopy temperatures, and we found independent confirmation of the isotope approach with the general agreement between Tcanδ, TIR-GPP, and Ta-GPP. This agreement further demonstrates a reasonable amount of confidence in our understanding of controls on cellulose δ18O across several sites and species. We can provide further confirmation of the isotope approach by examining the match between Tcanδ and modeled values of photosynthesis-weighted canopy temperatures at the Walker Branch site, where we have 19 years of hourly computations using the multi-layered, bio-meteorological model Canoak (Baldocchi 1997). By being a multi-layer model, the Canoak outputs are analogous to the Tcanδ approach in that both allow for canopy-wide integrations of photosynthesis and temperature. The mean photosynthesis-weighted temperature from Canoak for all 19 years was 22.2 ± 1.6 °C, and a histogram of these photosynthesis-weighted temperatures (presented as a probability density function, Fig. 4) shows a clear peak between 25 and 27 °C. At this site, the Ta-GPP was 23.9 °C and the site mean (tree cores from 1995 to 1999) for Tcanδ was 26.0 ± 1.1 °C.
Fig. 4

Histogram of photosynthesis-weighted canopy temperatures (presented as a probability density function) from the multi-layered Canoak model at the Walker Branch site for 19 years of hourly computations. The mean photosynthesis-weighted temperature for all 19 years was 22.2 ± 1.6 °C. The site average Tcanδ for Walker Branch is plotted for comparison

There are, however, caveats to the agreement between Tcanδ and the other estimated of photosynthesis-weighted canopy temperatures that range from scale comparisons to our knowledge of isotopic inputs and fractionation factors. Ta-GPP and TIR-GPP are derived in part from eddy-covariance measurements, and therefore, represent an ecosystem-level measure, whereas Tcanδ is derived primarily from measurements on individuals. It is possible that the specific individuals that we measured differed markedly from the aggregate forest in terms of the relationship between canopy temperature, air temperature and photosynthesis. This is unlikely, but it is still a clear shortcoming of the comparison. The fact that GPP-based temperatures were calculated over many years, and Tcanδ was derived from homogenizing tree rings from multiple years, is both a strength and a weakness. It is a weakness largely because we lose any ability to parse out effects that interannual variation in weather may have had on canopy processes. Conversely, it is a strength because by homogenizing tree rings, we smoothed over small-scale variations that would affect isotopic exchange within molecular precursors to cellulose that would affect the value of pex. These variations can include seasonal or interannual changes in the use of stored versus recently assimilated carbon (Gessler et al. 2009, 2014; Song et al. 2014a, b).

Temperature sensitivity on εO appears to have a muted impact on interpreting cellulose δ18O and resolving Tcanδ in most systems. The temperature range over which growth likely occurs in almost all terrestrial plants is well above the temperature range for which Sternberg and Ellsworth (2011) found the greatest change in εO. In applying temperature-based changes in εO to their interpretation of cellulose δ18O across biomes, they assumed that growth occurred at MAT in all systems, and therefore, assumed that growth occurred in the dormant seasons as well as the growing season. Further, they fixed the temperature for εO at 5 °C for any site with MAT lower than 5 °C, a temperature at which little or no plant growth occurs (Körner 2008). The mean growing-season temperature for each site is the preferable control variable for εO, because this temperature best reflects seasonal patterns of cambial cell growth and wall thickening (Moser et al. 2010). Excluding the subalpine sites, which will be discussed further below, mean growing season temperature ranged between 11.7 and 20.3 °C at our sites, and the mean temperature-dependent εO across these sites was 27.3 ± 0.2‰. Reanalysis of the Sternberg and Ellsworth data at temperatures between 10 and 25 °C show that there is no significant relationship with temperature, and the mean εO was 27.8‰. Consistent with this, Roden and Ehleringer (2000) found no temperature-induced changes in εO in Populus angustifolia along a transect where mean growing season temperature differences exceeded 5 °C. Lastly, as was pointed out recently (Zech et al. 2014), Sternberg and Ellsworth found a temperature effect on εO largely because of the a priori assumption that pex was constant and unaffected by temperature, yet variation in turnover time of the carbohydrate pool—quite possibly associated with temperature—can lead to changes in pex (Song et al. 2014b). It is worth noting that a temperature-sensitive εO or pex would have the same ultimate affect in the final Tcanδ calculation.

Several papers have reported pex as a variable that may change dynamically in response to changes in environmental or physiological conditions (Gessler et al. 2009; Song et al. 2014a; Cheesman and Cernusak 2017), but the majority of pex values have been found to be around the value we used here for angiosperms, 0.4 (Cernusak et al. 2005; Gessler et al. 2014). For gymnosperms, we decided to use the pex = 0.26 for our analyses based on the work of Song et al. (2014b), who compared late wood δ18OC of two oak species and one pine species for 2 years, measuring all isotopic inputs as well as TIR-GPP for each species. They found that pex was 0.4 for the angiosperms, but that using a pex of 0.4 for the pine led to model predictions that were approximately 2.5‰ less enriched than observations. Like several other studies (Szczepanek et al. 2006; Reynolds-Henne et al. 2007; Richter et al. 2008; Roden and Farquhar 2012), Song et al. (2014b) found that pines were more enriched in 18O than co-occurring angiosperms. Direct measurements allowed them to rule out higher canopy temperatures, differences in source-water or leaf water δ18O, and/or a larger εO as explanations for the greater enrichment in pine. They concluded that the more enriched δ18O of pine was likely due to a lower pex. A pex = 0.26 provided the best model-data fit, and was within the range observed for a different pine species by Gessler et al. (2009). When using pex = 0.4 for gymnosperms in the current study, we found that resolved Tcanδ for gymnosperms were higher than Ta-GPP (in excess of 3 °C when excluding the Niwot Ridge sites), yet the TIR-GPP estimates of both this study and those of Song et al. (2014b) suggest that, at least with the pines, canopy temperatures matched or were lower than ambient air temperatures. Thus, based on the higher observed δ18OC of gymnosperms that co-occur with angiosperms, the lack of elevated canopy temperatures by direct measurement, and the general energy balance argument that needle-leaved trees should be better coupled to the atmosphere (Jarvis and Mcnaughton 1986), we feel that using a lower pex for gymnosperms is justified. Whether pex = 0.26 is the correct number for all gymnosperms, let alone all pine species in all environments, certainly needs further study.

A particularly interesting example of δ18Osource error and temperature effects on Tcanδ can be seen by examining the two subalpine sites more closely. For this discussion, we assume that Ta-GPP is the correct target value for Tcanδ. GLEES is the simpler scenario because model and observed δ18Osource were similar and there was little difference in observed δ18OC between species. For this site, only a temperature-dependent εO was needed for Tcanδ to match Ta-GPP relatively well. Sternberg and Ellsworth (2011) found the greatest effect on εO between 5 and 10 °C, and the GLEES site has a cool growing season, with average 24 h growing season temperature of 7.6 °C, and Ta-GPP was 10.5 °C. Niwot Ridge has a similarly cool growing season and similar species to GLEES, yet there was poor agreement in Tcanδ among species because both δ18Osource and δ18OC were different. Both Ta-GPP and TIR-GPP were similar at Niwot Ridge, thus suggesting that resolving Tcanδ is highly susceptible to errors in areas where δ18Osource is highly variable. At sites where the δ18O of precipitation inputs vary greatly through a year, there should be an added focus on sampling source water δ18O frequently to avoid large errors in Tcanδ.

The agreement between Tcanδ and Ta-GPP confirms a similar envelope of photosynthesis-weighted canopy temperatures across warm temperate to boreal biomes shown by H&R, but there are observations in this study that contradict some of the conclusions of H&R. These contradictions revolve around the suggestion that (1) Tcanδ invariance can be generalized to forests outside of the range of temperate to boreal forests, and (2) the relationship between canopy temperature to air temperature, or canopy homeothermy. Song et al. (2011) first showed that Tcanδ for tropical and subalpine forests was above and below 20 °C, respectively, so the similar finding here is not surprising. The lack of homeothermy suggested here, however, requires greater discussion.

While the same temperature invariance in temperate to boreal biomes observed by H&R was confirmed in this study, the relationship between air temperature and canopy temperature in the boreal forest was not. The large excursion of Tcanδ from air temperature in the boreal forests is what drove the discussion of homeothermy in H&R. While we do find a significant relationship between MAT and canopy over temperature (Tcanδ − Ta-GPP), the use of MAT as an independent variable is misleading. It is more appropriate to compare over temperature to mean-growing-season temperature across sites, and for this we find no significant relationship. The over temperature of boreal trees was not systematically greater than other systems. The fact that H&R found that Tcanδ was much larger than day-time-growing-season air temperatures in boreal systems than what we found here could be because of the use of weather station air temperatures as opposed to the canopy air temperatures that were used in this study. It is possible that day-time weather station temperatures in the boreal systems are cooler during the day than surrounding forests due to both the lower albedo and lower latent heat exchange in the forests (Baldocchi et al. 2000, Bonan 2008) as compared to cleared areas where standard weather stations are typically located. Such a scenario would certainly have biased the comparisons of H&R. Based on the current analysis, we conclude that canopy temperatures in the boreal forest are as warm as those in temperate systems because day-time-growing-season air temperatures are similarly warm, yet this conclusion does not preclude the possibility of limited homeothermy.

The relationship between leaf and air temperature has been the subject of research for more than 50 years, and while air temperature is clearly the reference point to which leaf temperature tends (Jones 2014), there has long been a question of a systematic deviation from air temperature due to changes in leaf/branch boundary layers or latent heat flux (Gates 1980). Recent work has suggested that plants exist on a continuum of limited homeothermy (Michaletz et al. 2015, 2016; Dong et al. 2017), where leaves can, through adjustment of absorbed radiation, transpiration, convective heat loss and the thermal time constant, maintain an offset above or below air temperature. Our results show that across different sites and species, canopy temperatures do not deviate much from air temperatures when both are photosynthesis-weighted. However, when considering the errors associated with the Tcanδ approach—and because we homogenized tree rings across multiple years—our results are too coarse and may be masking small offsets between air and canopy temperatures. To fully examine the existence of limited homeothermy, and perhaps more importantly whether limited homeothermy is adaptive and/or relevant in terms of productivity, it is probably better to examine photosynthesis-weighted canopy temperatures at faster response times using the combination of thermal imagery and eddy-covariance flux measurements.

Photosynthesis and the energy balance of leaves are indelibly linked, yet it has been difficult to obtain measurements in one without precluding the measurement of the other, but it is surely the combination of canopy temperatures and photosynthesis that matters over the long term. We have confirmed here that the stable isotope approach to resolve photosynthesis-weighted canopy temperatures successfully integrates these processes. At the current time, our lack of understanding of small, temporal scale variability in a variety of isotopic fractionation factors and/or within-plant carbon cycling may limit the application of Tcanδ to at least annual resolution, and possibly even multi-year integrations as were done here. This still leaves a broad range of open applications, however, such as the examination of plants growing along altitudinal transects or ecotones where climatic changes are most acute, and co-occurring plants with differing life-history strategies.

Notes

Acknowledgements

BRH would like to acknowledge the intellectually stimulating and unfailingly supportive environment provided by Jim Ehleringer during those young and formative years, despite the ongoing and fruitless quest to locate Baby Jesus. BRH and XS were supported by the National Science Foundation under award number IOS-0950998. The National Center for Atmospheric Research (NCAR) is sponsored by NSF.

Author contribution statement

BRH and XS conceived and designed the data collection and modeling approach. BRH, XS, MLG, KC, PB, JWM, JC, AN, DH, SW, TM, DB, EE, AD and SPB collected and analyzed data. BRH and XS wrote the manuscript; other authors provided editorial advice.

Supplementary material

442_2018_4198_MOESM1_ESM.xlsx (45 kb)
Supplementary material 1 (XLSX 44 kb)

References

  1. Anderson WT, Bernasconi SM, Mckenzie JA, Saurer M (1998) Oxygen and carbon isotopic record of climatic variability in tree ring cellulose (Picea abies): an example from central Switzerland (1913–1995). J Geophys Res 103:31625–31636CrossRefGoogle Scholar
  2. Aubrecht DM, Helliker BR, Goulden ML, Roberts DA, Still CJ, Richardson AD (2016) Continuous, long-term, high-frequency thermal imaging of vegetation: uncertainties and recommended best practices. Agric For Meteorol 228–229:315–326CrossRefGoogle Scholar
  3. Baldocchi D (1997) Measuring and modelling carbon dioxide and water vapour exchange over a temperate broad-leaved forest during the 1995 summer drought. Plant Cell Environ 20:1108–1122CrossRefGoogle Scholar
  4. Baldocchi D, Sturtevant C (2015) Does day and night sampling reduce spurious correlation between canopy photosynthesis and ecosystem respiration? Agric For Meteorol 207:117–126CrossRefGoogle Scholar
  5. Baldocchi D, Kelliher FM, Black TA, Jarvis P (2000) Climate and vegetation controls on boreal zone energy exchange. Glob Change Biol 6:69–83CrossRefGoogle Scholar
  6. Barbour MM, Farquhar GD (2000) Relative humidity- and ABA-induced variation in carbon and oxygen isotope ratios of cotton leaves. Plant Cell Environ 23:473–485CrossRefGoogle Scholar
  7. Barbour MM, Fischer RA, Sayre KD, Farquhar GD (2000) Oxygen isotope ratio of leaf and grain material correlates with stomatal conductance and grain yield in irrigated wheat. Aust J Plant Physiol 27:625–637Google Scholar
  8. Barton CVM, Ellsworth DS, Medlyn BE et al (2010) Whole-tree chambers for elevated atmospheric CO2 experimentation and tree scale flux measurements in south-eastern Australia: the Hawkesbury forest experiment. Agric For Meteorol 150:941–951CrossRefGoogle Scholar
  9. Berry J, Bjorkman O (1980) Photosynthetic response and adaptation to temperature in higher plants. Annu Rev Plant Physiol 31:491–543CrossRefGoogle Scholar
  10. Bonan GB (2008) Forests and climate change: forcings, feedbacks, and the climate benefits of forests. Science 320:1444–1449CrossRefPubMedGoogle Scholar
  11. Brendel O, Iannetta PPM, Stewart D (2000) A rapid and simple method to isolate pure alpha-cellulose. Phytochem Anal 11:7–10CrossRefGoogle Scholar
  12. Buck AL (1981) New equations for computing vapor pressure and enhancement factor. J Appl Meteorol 20:1527–1532CrossRefGoogle Scholar
  13. Cernusak LA, Farquhar GD, Pate JS (2005) Environmental and physiological controls over oxygen and carbon isotope composition of Tasmanian blue gum, Eucalyptus globulus. Tree Physiol 25:129–146CrossRefPubMedGoogle Scholar
  14. Cheesman AW, Cernusak LA (2017) Infidelity in the outback: climate signal recorded in Δ18O of leaf but not branch cellulose of eucalypts across an Australian aridity gradient. Tree Physiol 37:554–564PubMedGoogle Scholar
  15. Craig H, Gordon LI (1965) Deuterium and oxygen 18 variations in the ocean and marine atmosphere. In: Tongiorigi E (ed) Stable isotopes in oceanographic studies and paleotemperatures. Consiglio Nazionale Delle Ricerche Laboratorio di Geologia Nucleare, Pisa, pp 9–130Google Scholar
  16. Desai AR, Richardson AD, Moffat AM et al (2008) Cross-site evaluation of eddy covariance GPP and RE decomposition techniques. Agric For Meteorol 148:821–838CrossRefGoogle Scholar
  17. Dong N, Prentice IC, Harrison SP, Song QH, Zhang YP (2017) Biophysical homoeostasis of leaf temperature: a neglected process for vegetation and land-surface modelling. Glob Ecol Biogeogr 26:998–1007CrossRefGoogle Scholar
  18. Ehleringer JR (1989) Temperature and energy budgets. In: Pearcy RW, Ehleringer JR, Mooney HA, Rundel PW (eds) Plant physiological ecology field methods and instrumentation. Chapman and Hall, London, pp 117–135CrossRefGoogle Scholar
  19. Ehleringer JR, Bjorkman O, Mooney HA (1976) Leaf pubescence: effects on absorptance and photosynthesis in a desert shrub. Science 192:376–377CrossRefPubMedGoogle Scholar
  20. Epstein S, Thompson P, Yapp CJ (1977) Oxygen and hydrogen isotopic ratios in plant cellulose. Science 198:1209–1215CrossRefPubMedGoogle Scholar
  21. Falge E, Baldocchi D, Olson R et al (2001) Gap filling strategies for defensible annual sums of net ecosystem exchange. Agric For Meteorol 107:43–69CrossRefGoogle Scholar
  22. Farquhar GD, Lloyd J (1993) Carbon and oxygen isotope effects in the exchange of carbon dioxide between terrestrial plants and the atmosphere. In: Ehleringer JR, Hall AE, Farquhar GD (eds) Stable isotopes and plant carbon/water relations. Academic Press, San Diego, pp 47–70CrossRefGoogle Scholar
  23. Farquhar GD, Ehleringer JR, Hubick KT (1989) Carbon isotope discrimination and photosynthesis. Annu Rev Plant Physiol Mol Biol 40:503–537CrossRefGoogle Scholar
  24. Flanagan L, Farquhar G (2014) Variation in the carbon and oxygen isotope composition of plant biomass and its relationship to water-use efficiency at the leaf- and ecosystem-scales in a northern Great Plains grassland. Plant Cell Environ 37:425–438CrossRefPubMedGoogle Scholar
  25. Flanagan L, Comstock J, Ehleringer J (1991) Comparison of modeled and observed environmental influences on the stable oxygen and hydrogen isotope composition of leaf water in Phaseolus vulgaris L. Plant Physiol 96:588–596CrossRefPubMedPubMedCentralGoogle Scholar
  26. Gates DM (1962) Energy exchange in the biosphere. Harper and Row, New YorkGoogle Scholar
  27. Gates DM (1965) Energy, plants, and ecology. Ecology 46:1–13CrossRefGoogle Scholar
  28. Gates DM (1980) Biophysical Ecology. Springer, New York, p 611CrossRefGoogle Scholar
  29. Gaudinski JB, Dawson TE, Quideau S et al (2005) Comparative analysis of cellulose preparation techniques for use with 13 C, 14 C, and 18 O isotopic measurements. Anal Chem 77:7212–7224CrossRefPubMedGoogle Scholar
  30. Gessler A, Peuke AD, Keitel C, Farquhar GD (2007) Oxygen isotope enrichment of organic matter in Ricinus communis during the diel course and as affected by assimilate transport. New Phytol 174:600–613CrossRefPubMedGoogle Scholar
  31. Gessler A, Brandes E, Buchmann N, Helle G, Rennenberg H, Barnard R (2009) Tracing carbon and oxygen isotope signals from newly assimilated sugars in the leaves to the tree-ring archive. Plant Cell Environ 32:780–795CrossRefPubMedGoogle Scholar
  32. Gessler A, Ferrio JP, Hommel R, Treydte K, Werner RA, Monson RK (2014) Stable isotopes in tree rings: towards a mechanistic understanding of isotope fractionation and mixing processes from the leaves to the wood. Tree Physiol 34:796–818CrossRefPubMedGoogle Scholar
  33. Gray J, Thompson P (1976) Climatic information from O18/O16 ratios of cellulose in tree rings. Nature 262:481–482CrossRefGoogle Scholar
  34. Helliker BR (2014) Reconstructing the δ18O of atmospheric water vapour via the CAM epiphyte Tillandsia usneoides: seasonal controls on δ18O in the field and large-scale reconstruction of δ18Oa. Plant Cell Environ 37:541–556CrossRefPubMedGoogle Scholar
  35. Helliker BR, Griffiths H (2007) Toward a plant-based proxy for the isotope ratio of atmospheric water vapor. Glob Change Biol 13:723–733CrossRefGoogle Scholar
  36. Helliker BR, Richter SL (2008) Subtropical to boreal convergence of tree-leaf temperatures. Nature 454:511–514CrossRefPubMedGoogle Scholar
  37. Jarvis PG, Mcnaughton KG (1986) Stomatal control of transpiration: scaling up from leaf to region. Adv Ecol Res 15:1–49CrossRefGoogle Scholar
  38. Jones HG (2014) Plants and Microclimate. Cambridge University Press, CambridgeGoogle Scholar
  39. Körner C (2008) Winter crop growth at low temperature may hold the answer for alpine treeline formation. Plant Ecol Divers 1:3–11CrossRefGoogle Scholar
  40. Larcher W (1995) Physiological plant ecology. Springer, New YorkCrossRefGoogle Scholar
  41. Leuzinger S, Körner C (2007) Tree species diversity affects canopy leaf temperatures in a mature temperate forest. Agric For Meteorol 146:29–37CrossRefGoogle Scholar
  42. Long SP, Woodward FI (1988) Plants and temperature. The Society for Experimental Biology, CambridgeGoogle Scholar
  43. Michaletz ST, Weiser MD, Zhou J, Kaspari M, Helliker BR, Enquist BJ (2015) Plant thermoregulation: energetics, trait–environment interactions, and carbon economics. Trends Ecol Evol 30:714–724CrossRefPubMedGoogle Scholar
  44. Michaletz ST, Weiser MD, McDowell NG, Zhou J, Kaspari M, Helliker BR, Enquist BJ (2016) The energetic and carbon economic origins of leaf thermoregulation. Nat Plants 2:16129CrossRefPubMedGoogle Scholar
  45. Miller PC (1971) Sampling to estimate mean leaf temperatures and transpiration rates in vegetation canopies. Ecology 52:885–889CrossRefGoogle Scholar
  46. Moffat AM, Papale D, Reichstein M et al (2007) Comprehensive comparison of gap-filling techniques for eddy covariance net carbon fluxes. Agric For Meteorol 147:209–232CrossRefGoogle Scholar
  47. Moser L, Fonti P, Buntgen U, Esper J, Luterbacher J, Franzen J, Frank D (2010) Timing and duration of European larch growing season along altitudinal gradients in the Swiss Alps. Tree Physiol 30:225–233CrossRefPubMedGoogle Scholar
  48. Porter TJ, Pisaric MFJ, Kokelj SV, Edwards TWD (2009) Climatic signals in δ 13 C and δ 18 O of tree-rings from white spruce in the Mackenzie Delta Region, Northern Canada. Arct Antarct Alp Res 41:497–505CrossRefGoogle Scholar
  49. Ramesh R, Bhattacharya SK, Gopalan K (1985) Dendroclimatological implications of isotope coherence in trees from Kashmir Valley, India. Nature 317:802–804CrossRefGoogle Scholar
  50. Raschke K (1960) Heat transfer between the plant and the environment. Annu Rev Plant Physiol 11:111–126CrossRefGoogle Scholar
  51. Reynolds-Henne CE, Siegwolf RTW, Treydte KS, Esper J, Henne S, Saurer M (2007) Temporal stability of climate–isotope relationships in tree rings of oak and pine (Ticino, Switzerland). Glob Biogeochem Cycles 21:GB4009Google Scholar
  52. Richter SL, Johnson AH, Dranoff MM, LePage BA, Williams CJ (2008) Oxygen isotope ratios in fossil wood cellulose: isotopic composition of Eocene- to Holocene-aged cellulose. Geochim Cosmochim Acta 72:2744–2753CrossRefGoogle Scholar
  53. Robertson I, Waterhouse JS, Barker AC, Carter AHC, Switsur VR (2001) Oxygen isotope ratios of oak in east England: implications for reconstructing the isotopic composition of precipitation. Earth Planet Sci Lett 191:21–31CrossRefGoogle Scholar
  54. Roden JS, Ehleringer JR (1999) Hydrogen and oxygen isotope ratios of tree-ring cellulose for riparian trees grown long-term under hydroponically controlled environments. Oecologia 121:467–477CrossRefPubMedGoogle Scholar
  55. Roden JS, Ehleringer JR (2000) There is no temperature dependence of net biochemical fractionation of hydrogen and oxygen isotopes in tree-ring cellulose. Isot Environ Health Stud 36:303–317CrossRefGoogle Scholar
  56. Roden JS, Farquhar GD (2012) A controlled test of the dual-isotope approach for the interpretation of stable carbon and oxygen isotope ratio variation in tree rings. Tree Physiol 32:490–503CrossRefPubMedGoogle Scholar
  57. Roden JS, Lin G, Ehleringer JR (2000) A mechanistic model for interpretation of hydrogen and oxygen isotope ratios in tree-ring cellulose. Geochim Cosmochim Acta 64:21–35CrossRefGoogle Scholar
  58. Saurer M, Aellen K, Siegwolf R (1997) Correlating d13C and d18O in cellulose of trees. Plant Cell Environ 20:1543–1550CrossRefGoogle Scholar
  59. Saurer M, Cherubini P, Siegwolf R (2000) Oxygen isotopes in tree rings of Abies alba: the climatic significance of interdecadal variations. J Geophys Res 105:12461–12470CrossRefGoogle Scholar
  60. Schimper AFW (1903) Plant-geography upon a physiological basis. Clarendon Press, Oxford, p 839CrossRefGoogle Scholar
  61. Smith WK (1978) Temperatures of desert plants: another perspective on the adaptability of leaf size. Science 201:614CrossRefPubMedGoogle Scholar
  62. Smith WK, Carter GA (1988) Shoot structural effects on needle temperatures and photosynthesis in conifers. Am J Bot 75:496–500CrossRefGoogle Scholar
  63. Song X, Farquhar GD, Gessler A, Barbour MM (2014a) Turnover time of the non-structural carbohydrate pool influences δ 18 O of leaf cellulose. Plant Cell Environ 37:2500–2507CrossRefPubMedGoogle Scholar
  64. Song X, Clark KS, Helliker BR (2014b) Interpreting species-specific variation in tree-ring oxygen isotope ratios among three temperate forest trees. Plant Cell Environ 37:2169–2182CrossRefPubMedGoogle Scholar
  65. Song X, Barbour MM, Saurer M, Helliker BR (2011) Examining the large-scale convergence of photosynthesis-weighted tree leaf temperatures through stable oxygen isotope analysis of multiple data sets. New Phyto 192:912–924CrossRefGoogle Scholar
  66. Sternberg LSL (1989) Oxygen and hydrogen isotope ratios in plant cellulose: mechanisms and applications. In: Rundel PW, Ehleringer JR, Nagy KA (eds) Stable isotopes in ecological research. Springer, New York, pp 124–141CrossRefGoogle Scholar
  67. Sternberg L, Ellsworth PFV (2011) Divergent biochemical fractionation, not convergent temperature, explains cellulose oxygen isotope enrichment across latitudes. PLoS One 6:e28040CrossRefPubMedPubMedCentralGoogle Scholar
  68. Szczepanek M, Pazdur A, Pawełczyk S et al (2006) Hydrogen, carbon and oxygen isotopes in pine and oak tree rings from southern poland as climatic indicators in years 1900–2003. Geochronometria 25:67–76Google Scholar
  69. Walter H, Harnickell E, Mueller-Dombois D (1975) Climate-diagram maps of the individual continents and the ecological climatic regions of the Earth. Springer, BerlinGoogle Scholar
  70. West AG, Patrickson SJ, Ehleringer JR (2006) Water extraction times for plant and soil materials used in stable isotope analysis. Rapid Commun Mass Spectrom 20:1317–1321CrossRefPubMedGoogle Scholar
  71. Wright WE, Leavitt SW (2006) Boundary layer humidity reconstruction for a semiarid location from tree ring cellulose δ18O. J Geophys Res Atmos 111:D18105CrossRefGoogle Scholar
  72. Yakir D (1992) Variations in the natural abundance of oxygen-18 and deuterium in plant carbohydrates. Plant Cell Environ 15:1005–1020CrossRefGoogle Scholar
  73. Yakir D, DeNiro MJ (1990) Oxygen and hydrogen isotope fractionation during cellulose metabolism in Lemna gibba L. Plant Physiol 93:325–332CrossRefPubMedPubMedCentralGoogle Scholar
  74. Zech M, Mayr C, Tuthorn M, Leiber-Sauheitl K, Glaser B (2014) Reply to the comment of Sternberg on “Zech et al. (2014) Oxygen isotope ratios (18O/16O) of hemicellulose-derived sugar biomarkers in plants, soils and sediments as paleoclimate proxy I: Insight from a climate chamber experiment. GCA 126, 614–623”. Geochimica et Cosmochimica Acta 141:680–682CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2018

Authors and Affiliations

  • Brent R. Helliker
    • 1
  • Xin Song
    • 1
    • 13
  • Michael L. Goulden
    • 2
  • Kenneth Clark
    • 3
  • Paul Bolstad
    • 4
  • J. William Munger
    • 5
  • Jiquan Chen
    • 6
  • Asko Noormets
    • 7
  • David Hollinger
    • 8
  • Steve Wofsy
    • 5
  • Timothy Martin
    • 9
  • Dennis Baldocchi
    • 10
  • Eugenie Euskirchenn
    • 11
  • Ankur Desai
    • 12
  • Sean P. Burns
    • 14
    • 15
  1. 1.Department of BiologyUniversity of PennsylvaniaPhiladelphiaUSA
  2. 2.Department of Earth System ScienceUniversity of CaliforniaIrvineUSA
  3. 3.USDA Forest ServiceNorthern Research StationNew LisbonUSA
  4. 4.Department of Forest ResourcesUniversity of MinnesotaSaint PaulUSA
  5. 5.Department of Earth and Planetary Sciences, School of Engineering and Applied SciencesHarvard UniversityCambridgeUSA
  6. 6.Department of Geography, Center for Global Change and Earth Observations (CGCEO)Michigan State UniversityEast LansingUSA
  7. 7.Department of Ecosystem Science and ManagementTexas A & M UniversityCollege StationUSA
  8. 8.USDA Forest ServiceNorthern Research StationDurhamUSA
  9. 9.School of Forest Resources and ConservationUniversity of FloridaGainesvilleUSA
  10. 10.ESPM, University of California, BerkeleyBerkeleyUSA
  11. 11.Institute of Arctic BiologyUniversity of Alaska-FairbanksFairbanksUSA
  12. 12.Department of Atmospheric and Oceanic SciencesUniversity of Wisconsin-MadisonMadisonUSA
  13. 13.College of Life Sciences and OceanographyShenzhen UniversityShenzhenChina
  14. 14.Department of GeographyUniversity of ColoradoBoulderUSA
  15. 15.National Center for Atmospheric ResearchBoulderUSA

Personalised recommendations