Biogeographic variation in evergreen conifer needlelongevity and impacts on boreal forest carboncycle projectionsPeter B. Reicha,b,1, Roy L. Richa, Xingjie Luc, Ying-Ping Wangd, and Jacek Oleksyna,e

aDepartment of Forest Resources, University of Minnesota, St. Paul, MN 55108; bHawkesbury Institute for the Environment, University of Western Sydney,Penrith, NSW 2751, Australia; cCollege of Global Change and Earth System Science, Beijing Normal University, Beijing 100875, China; dCommonwealthScientific and Industrial Research Organisation Marine and Atmospheric Research, Aspendale, VIC 3195, Australia; and eInstitute of Dendrology, PolishAcademy of Sciences, PL-62-035, Kornik, Poland

Edited by Cyrille Violle, Centre National de la Recherche Scientifique, Montpellier, France, and accepted by the Editorial Board October 22, 2013 (received forreview January 30, 2013)

Leaf life span is an important plant trait associated with inter-specific variation in leaf, organismal, and ecosystem processes. Wehypothesized that intraspecific variation in gymnosperm needletraits with latitude reflects both selection and acclimation for traitsadaptive to the associated temperature and moisture gradient.This hypothesis was supported, because across 127 sites alonga 2,160-km gradient in North America individuals of Picea glauca,Picea mariana, Pinus banksiana, and Abies balsamea had longerneedle life span and lower tissue nitrogen concentration with de-creasing mean annual temperature. Similar patterns were notedfor Pinus sylvestris across a north–south gradient in Europe. Thesedifferences highlight needle longevity as an adaptive feature im-portant to ecological success of boreal conifers across broad climaticranges. Additionally, differences in leaf life span directly affect an-nual foliage turnover rate, which along with needle physiologypartially regulates carbon cycling through effects on gross primaryproduction and net canopy carbon export. However, most, if notall, global land surface models parameterize needle longevity ofboreal evergreen forests as if it were a constant. We incorporatedtemperature-dependent needle longevity and %nitrogen, and bio-mass allocation, into a land surfacemodel, Community AtmosphereBiosphere Land Exchange, to assess their impacts on carbon cyclingprocesses. Incorporating realistic parameterization of these varia-bles improved predictions of canopy leaf area index and gross pri-mary production compared with observations from flux sites.Finally, increasingly low foliage turnover and biomass fraction to-ward the cold far north indicate that a surprisingly small fraction ofnew biomass is allocated to foliage under such conditions.

biogeography | spruce | pine | ecotype

The boreal forest is an enormous terrestrial biome, charac-terized by long, cold winters, low rates of productivity and

nutrient cycling (1–4), and low richness of the dominant treespecies. All of these tree species have large geographic ranges(5), and ecological theory suggests that populations should varyboth genotypically and phenotypically in response to such vari-ation. In fact, it has been long recognized that terrestrial plantsin general, and boreal conifers specifically, exhibit ecotypicpopulation differentiation that reflects influence of site origin ona variety of chemical, morphological, physiological, and allo-metric traits (6–12). It is also well known that biogeographicvariation in the environment influences the phenotypic traitsplastically realized by species across their ranges. Both pheno-typic and genotypic processes therefore contribute to observedvariation in traits across environmental gradients. Hence, acrossthe boreal biome, conditions that lead to slow growth, such asshort growing season, cool summer temperatures, and low soilnutrient supply (1, 2, 13), should be reflected by shifts towardmore conservative traits (such as long leaf life span and lownutrient concentrations) considered to be advantageous under

such conditions (14, 15). However, although evidence that ecotypicvariation occurs is abundant (7, 8, 10), comprehensive character-ization of ecologically important intraspecific variation across bio-geographic gradients is lacking (16).Global land surface models that simulate carbon cycling of the

world’s terrestrial biomes face distinct challenges for both spe-cies-poor biomes, such as the boreal forest, and species-richbiomes, such as tropical rain forests. For the latter, the challengeis to sufficiently well parameterize leaf and canopy propertiesgiven enormous species diversity within and among sites (e.g.,>500 species for <1,000 individuals in a given several-hectarearea, with large species turnover at 100-km scale). In contrast,for boreal forests, the task is seemingly easy: Only a smallnumber of species dominate the entire boreal forest in eachcontinent. Moreover, the same genera dominate the circumborealregion. However, although leaf and canopy properties vary withclimate and geographic location, this variation is usually ignored inbiome-scale or global models, because of a lack of systematicunderstanding of that variation.Despite long recognition of ecotypic variation in boreal coni-

fers, certain broad-scale aspects of physiological macroecologyremain poorly quantified. One trait known to differ intraspecifi-cally in evergreen conifers with climate and latitude is the needlelife span (NL) (synonymous with needle longevity) (17, 18). Ingeneral, individuals growing in colder locations tend to havelonger NL (17–19), but important patterns and consequencesare unknown, such as the following. (i) What is the shape of the

Significance

How evergreen tree needle longevity varies from south tonorth in the boreal biome is poorly quantified and thereforeignored in vegetation and earth system models. This is prob-lematic, because needle longevity translates directly into nee-dle turnover rate and profoundly affects carbon cycling in bothnature and computer models. Herein we present data for fivewidespread boreal conifers, including pines and spruces, from>125 sites along a 2,000-km gradient. For each species, indi-viduals in colder, more northern environments had longerneedle life span, highlighting its importance to evergreenecological success. Incorporating biogeography of needle lon-gevity into a global model improved predictions of forestproductivity and carbon cycling and identified specific prob-lems for models that ignore such variability.

Author contributions: P.B.R. and R.L.R. designed research; R.L.R. performed research; P.B.R.,X.L., Y.-P.W., and J.O. analyzed data; and P.B.R. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission. C.V. is a guest editor invited by the EditorialBoard.1To whom correspondence should be addressed. Email: preich@umn.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1216054110/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1216054110 PNAS | September 23, 2014 | vol. 111 | no. 38 | 13703–13708

ECOLO

GY

SPEC

IALFEATU

RE

NL relation to geographic climate variation? (ii) Are proportionalshifts in NL for a given climate gradient similar among species?(iii) Do other leaf traits often associated interspecifically with leaflife span (e.g., refs. 15 and 20) covary with NL intraspecifically in asimilar fashion? (iv) What are the consequences of geographicpatterns of needle traits to ecosystem and regional carbon cycling?To improve our understanding of these issues, we measuredneedle traits of the four dominant evergreen conifers fromMinnesota to northern Canada, compiled similar data for thedominant evergreen species in Eurasia (Pinus sylvestris), character-ized the relations of needle traits with climate variables for all fivespecies, and incorporated such findings into a land surface model[Community Atmosphere Biosphere Land Exchange (CABLE)]that simulated biome-wide carbon and nitrogen cycling.We sampled foliage from naturally grown trees (range of

height 2.5–5 m) at 127 sites ranging from Minnesota to Ontario,Manitoba, Saskatchewan, Alberta, and the Northwest Territories,in central Canada. We sampled sunlit upper canopy branches oftrees ranging from 2.5 to 5 m in height, to standardize tree sizeand age across sites, as well as the light environment of the sampledbranches. The latter was of paramount concern, because needlelongevity in the crown of a conifer increases systematically withshading (by as much as 50–60% for spruce and fir). The fourspecies were present (and sampled) at 83 (Pinus banksiana),50 (Picea mariana), 45 (Picea glauca), and 21 (Abies balsamea)of the sites, respectively. We also use a literature compilationfor Pinus sylvestris across northern Fenno-Scandinavia to extendour analysis beyond North America. Because several climaticfactors covary with latitude and are hypothesized to be themechanism behind the latitudinal pattern, we focus our analysesand interpretation on climate, not latitude. Additionally, it isdifficult to ascertain which of the related climate metrics (includingannual and seasonal measures of temperature and precipitation)are responsible for the observed latitudinal patterns (Methods andSupporting Information). Because mean annual temperature(MAT) was identified in the model selection process as the cli-mate metric that explained the most variance in NL, we use it inthe results presented herein. However, we discuss the ways inwhich multiple aspects of climate, and associated soil resourceavailability, can combine to influence these results, below andin Supporting Information.

ResultsFor all five species, NL increases with increasing latitude (Fig. S1)and decreases with MAT (Fig. 1). The relationship is strongest forboth Picea species and P. sylvestris and weakest in P. banksiana.

The biogeographic range in NL is particularly pronounced in thePicea species—ranging from ∼6 y at their southern margins(∼46°N) to 15 y at the most northern locations sampled (∼62–63°N), which are not far from their northern limit in that partof central Canada. Abies has almost as strong a shift in NL perunit shift in MAT as the Picea species but has a narrower range,and thus NL varies from 5 y at its southern limit (∼46°N) to 9 y atits northern (∼55°N) limit. P. sylvestris shows a similar proportionalshift (more than doubling NL from 3.5 to 7.0 y from 58 to 69°N) asthe North America Picea species. In contrast, P. banksiana, whichhas as large a geographic range as the Picea species, has a muchweaker dependence of NL on MAT—it varies from 4.0 y inMinnesota to 5.5 in northern Canada. It is not clear whyP. banksiana has such a noisy and modest relation of NL toMAT, compared with the other NA species or its Europeancongener. It is possible that further to the south in its range it isrelegated by competition to particularly poor sites that supportslow growth and therefore longer NL than would otherwise oc-cur, but data to test this hypothesis are not available. Regardless,there is a strong signal that NL decreases with MAT for thesefive species that are among the most dominant species in muchof North American and Eurasian boreal forests.Needle %N increases significantly with MAT in P. banksiana,

P. glauca, and P. sylvestris, with nonsignificant relations in theother two species. For data pooled across all five species, thereare significant (P < 0.0001) relationships between %N and bothNL (Fig. S2) and MAT. This indicates that boreal evergreenforests dominated by individuals with long NL (owing to preva-lence of long NL species such as Picea, or to adaptive response toNL to cold temperatures, or both) also have low needle %N.In addition to illuminating intraspecific ecological strategies,

the biogeographic patterns documented here have importantimplications for ecosystem-scale function. We used the CABLEland surface model (21, 22) to explore the impact of bio-geographic variation in needle traits on evergreen boreal forest Ccycling (Methods and Table 1). This exploration is importantbecause most, perhaps all, widely recognized global land surfacemodels consider all boreal forest as a single plant functional type(PFT) with a single set of trait values, and land surface models ingeneral are inadequate with respect to biomass allocation (23, 24).To our knowledge, leading models such as Biome-BioGeochemicalCycles (Biome-BCG), Community Land Model version 4 (CLM-4),Lund–Potsdam–Jena Dynamic Global Model (LPJ) (and itsoffspring), and others do not include a temperate dependenceof needle turnover rate (or biomass allocation) within the borealevergreen PFT. Because CABLE does not recognize species

-4 -2 0 2 4 6

Nee

dle

life

span

(yea

rs)

-4 -2 0 2 4 6

Abies balsamea

black spruce

Mean annual temperature (oC)

Pinus banksiana

-4 -2 0 2 4 6

Pinus sylvestris

Mean annual temperature (oC)

Nee

dle

life

span

(yea

rs)

Picea glauca

Picea mariana

15

10

5

2.5

10

5

2.5

Fig. 1. Mean NL (years) for four North American and one European evergreen conifer species in relation to MAT along a large latitudinal gradient: A. balsamea(logNL = 0.860–0.041*MAT; P = 0.0083, R2 = 0.34, n = 19), P. glauca (logNL = 0.911–0.044*MAT; P < 0.0001, R2 = 0.65, n = 46), P. banksiana (logNL = 0.650−0.014*MAT;P = 0.0052, R2 = 0.10, n = 80), P. mariana (logNL = 0.932–0.056*MAT; P < 0.0001, R2 = 0.66, n = 51), and P. sylvestris (logNL = 0.792–0.040*MAT; P < 0.0001, R2 =0.65, n = 79).

13704 | www.pnas.org/cgi/doi/10.1073/pnas.1216054110 Reich et al.

differences, we used a generalized algorithm to modify NL withMAT in a fashion that averages responses across the speciesavailable (Methods).Structure, parameterizations, and functional routines of CABLE

are roughly similar to many other global land surface models,such as those mentioned above. Of relevance to this study, itsparameterizations of PFTs and their key traits are similar tomany other global land surface models. Thus, the examinationherein of how changes in parameters and functions alter modelperformance should be generally reflective and not exclusiveto CABLE.Temperature sensitivity of biomass allocation is likely an im-

portant but poorly understood regulator of C cycling in borealforests (25), and models at present do not predict allocation well,even ignoring biogeographic variation (23, 24). Standing biomassdistribution to leaves, stems, and roots is a joint function ofturnover rate and initial allocation of new biomass. Given this,and that our goal includes assessing ecosystem impacts of changesin foliage turnover rate, we jointly examine impacts of tempera-ture-sensitive biomass distribution (26) on CABLE predictions.Our modification of the biomass distribution in CABLE is basedon analyses that show that foliage and roots represent a greaterand smaller fraction, respectively, of total biomass in forests inincreasingly cold environments (26), which is consistent with in-dependent data showing proportionally greater carbon allocationbelow ground for Picea in colder, drier habitats (25). Incor-porating temperature-dependent biomass partitioning yields theresult that allocation of new biomass (NPP) to needles must bemarkedly low in the cold far north (Fig. 2), given that NLincreases in colder, higher latitudes (Fig. 1), but the fractionof standing biomass located in needles is low at low MAT (26)(Fig. S3).The CABLE model as previously published (21) (called de-

fault hereafter) assumed NL of 3.76 y, root life span of 18 y (thisincludes and averages across fine roots, coarse roots, and woodyroots), and wood life span of 70 y, needle %N of 1.23%, anda constant ratio of allocation of new biomass to needles, stems,and roots for evergreen needle leaf forest (21). To assess impactsof incorporating realistic biogeographic variation of biomass al-location (ALL), NL, and needle %N into the model, we comparethese default results to permutations (Table 1) that incorporatethese variables either as constants (cALLcNLc%N) or as func-tions of MAT (fALLfNLf%N), in various combinations. Theconstant values for ALL and %N are from the default CABLE,the constant value of NL is either the default value or the av-erage across all sites for our field data, and the functions thatvary with MAT are from our observations (Figs. 1 and 2 and Figs.S2 and S3), although for %N this is indirect, because we predict%N from temperature-dependent NL, to retain their covariance.We map these predictions spatially (Figs. 3 and 4) and also

compare predictions for four north–south climate zones to val-ues of canopy leaf area index (LAI) and gross primary pro-duction (GPP) from flux sites published by Luyssaert et al. (3)(Fig. 5). In the figures, moving from top to bottom (Figs. 3 and 4)or left to right (Fig. 5) generally moves from the default modelwith constant input parameter values, to new versions with oneor two modified parameters, to the model where values of allo-cation, NL, and needle %N are all biogeographically varied.The default version of CABLE successfully predicts the north–

south direction of LAI and GPP (Figs. 3–5) and slightly overpredictsGPP while underpredicting LAI, except in the coldest zone (Fig. 5).If a more realistic mean NL (6.02 y) as estimated from our datais used as a constant NL value everywhere (cALLcNLc%N), themodel still gets the direction of LAI and GPP correct, but over-predicts GPP and LAI. This directly results from model logic—witha constant allocation routine and a constant NL, greater NL resultsin greater LAI.Incorporating our general temperature-dependent relationship

for NL in the model instead of a constant value (cALLfNLc%N)also overpredicts GPP and LAI (Fig. 5) but additionally also failsto capture the directional signatures for GPP and LAI—withGPP stable and LAI decreasing with MAT. This is because witha constant allocation routine a realistically longer NL in the northunrealistically leads to higher LAI (and consequently higher GPPthan would otherwise occur), despite a shorter growing seasonand colder temperatures.Incorporating MAT-dependent biomass allocation (and with

NL constant, 6.02 y, fALLcNLc%N) also overpredicts GPP andLAI, but gets their directional variation correct (they are allhigher toward the warmer south). The permutation (fALLfNLc%N) that includes temperature-dependent NL as well as tem-perature-dependent biomass allocation makes predictions thatare similar. It is likely that the four new CABLE versions pre-sented above all overpredict GPP in part because the constant %N (1.23%) used in the model is higher than the empirically ob-served average (∼0.9%), and LAI and %N jointly influenceproductivity in the CABLE model, just as observed empiricallyfor both maximum instantaneous GPP and annual NPP (4). Thedefault CABLE predictions may approximate GPP (Fig. 5) ac-curately because unrealistically high %N offsets unrealisticallylow LAI.

Table 1. Versions used to assess CABLE model performance forboreal evergreen conifer forests

Version Allocation Needle life span %N

Default Constant Constant(3.76) Constant(1.23)cALLcNLc%N Constant Constant(6.02) Constant(1.23)cALLfNLc%N Constant NL(MAT) Constant(1.23)fALLcNLc%N ALL(MAT) Constant(6.02) Constant(1.23)fALLfNLc%N ALL(MAT) NL(MAT) Constant(1.23)cALLfNLf%N Constant NL(MAT) %N(MAT)fALLfNLf%N ALL(MAT) NL(MAT) %N(MAT)

The default version of CABLE represents the prior version (22, 33). Allo-cation (ALL) was either the default constant or a continuous function ofmean annual temperature (MAT) (25). Needle life span (NL) was either con-stant at the prior default value (3.76), constant at the mean value (6.02) forthe observed data, or modeled as a continuous function of MAT. Needle %Nwas either constant at the prior default value (1.23) or a continuous functionof NL, as adjusted using a canopy depth extinction coefficient.

0

0.2

0.4

0.6

0.8

1

-10 0 10

Allo

ca�o

n fr

ac�o

n to

folia

ge

MAT (0C)

50

100

200

300

Mg/ha

Fig. 2. Representation of the fraction of new total biomass allocated tofoliage in relation to MAT for evergreen conifer forests, for four above-ground biomass classes (50–300 Mg/ha). These curves result from the MAT-dependency of needle life span (Fig. 1) and fractional standing biomassdistribution to foliage (Fig. S1). The vertical dashed lines represent the rangeof data for which they were developed. The allocation fraction is mathe-matically derived as a function of life spans of leaf, wood, and root; biomassfraction of leaf, wood, and root; and their relations with MAT and totalbiomass. The idea of calculating allocation fraction in this fashion is froma semianalytic solution to accelerate biogeochemical model spin-up (39).

Reich et al. PNAS | September 23, 2014 | vol. 111 | no. 38 | 13705

ECOLO

GY

SPEC

IALFEATU

RE

The model permutation (cALLfNLf%N) in which both tem-perature-dependent NL and NL-dependent %N are incorporatedin the model (but with the default allocation scheme) does im-prove predictions on average compared with the model withtemperature-dependent NL but with the default constant %N,but the directionality of LAI and GPP are not captured. Themodel (fALLfNLf%N) that incorporates both of the tempera-ture-dependent functions (for NL and biomass partitioning) andthe NL-dependent %N seems to do well in matching not only thedirection of LAI and GPP but also the values observed empiri-cally, although least well in the coldest zone (Fig. 5).

DiscussionGiven the strong and consistent inverse relationship between leaflife span, nitrogen, and leaf carbon dioxide exchange ratesdocumented interspecifically, including for conifers (14, 15, 20),it is highly likely that evergreen conifers in cold, northern zonesof the boreal forest have low photosynthesis and respiration ratescompared with either conspecifics or unrelated taxa in warmer,southern habitats. A combination of direct and indirect effects oflow temperatures, including low nutrient availability and a shortgrowing season, likely contribute to selection for increasingly

conservative foliage traits at the highest latitudes (25, 27). Thus,patterns of NL and %N are consistent with the notion that undercold, nutrient-poor conditions and a short growing season, whichlimit realized carbon gain, a successful ecological strategy in-volves trading off potential for high carbon gain by having leaveswith low maximum photosynthetic capacity but associated lowrespiratory costs and having canopies that require only 6–12%replacement each year (14).Although we present results solely in relation to MAT, it is

unclear whether it is growing-season temperatures (or pre-cipitation), average annual temperatures (or precipitation), orrelative moisture and nutrient supply that are mechanisticallyresponsible for the low growth and associated long needle lifespan in high-latitude, cold locations (Supporting Information).Similarly, it is likely that low temperatures and dry conditionsboth contribute to high carbon allocation below ground in colderforests (25, 26). For the study sites on either continent, MAT wascorrelated strongly with mean annual precipitation (MAP), aswell as with seasonal climate metrics and with the ratio betweenMAP and potential evapotranspiration. MAT explained as muchor more of the variance in needle traits than any other climatemetric and was selected statistically in model selection, hence its

Fig. 3. Patterns of simulated GPP across the boreal forest zone in vegeta-tion under seven model versions, representing different model permutationswith panels corresponding to modifications described in Table 1.

Fig. 4. Patterns of simulated LAI across the boreal forest zone in vegetationunder seven model versions, representing different model permutationswith panels corresponding to modifications described in Table 1.

13706 | www.pnas.org/cgi/doi/10.1073/pnas.1216054110 Reich et al.

use herein. Nonetheless, it is likely that the documented patternsof needle trait variation with MAT are a response to seasonal aswell as annual aspects of the thermal and moisture regimes alongthe latitudinal gradients.Boreal forests play key roles in global carbon cycling, so de-

veloping accurate ecosystem understanding is critical for improvedcarbon cycle modeling. The region we studied here represents29% of global forests and was found to have the strongest positivecarbon feedback to future warming using earth system models(28). Empirical observations have found that needle life span andneedle %N (this paper) and the fraction of total stand biomassdistributed to foliage (26) all vary with MAT within this biome, butsuch patterns have not been accounted for by any of the earthsystem models that were used for estimating the sensitivity ofterrestrial carbon balance to warming.Using CABLE, we showed that implementing one or two of

these three model changes (Table 1) as shown in Figs. 3–5 willeither minimally alter or worsen the model predictions (comparedwith FLUXNET observations) compared with simulations withoutany of those changes as shown in the default output. Only with allthree changes implemented in the model (fALLfNLf%N, Table 1)does the performance of CABLE become quite similar to thedefault CABLE for GPP and improved for LAI (Fig. 5). There-fore, implementing ecological realism into CABLE, and per-haps other global land surface models, may not lead to betteragreements with observations or independent estimates of GPP.However, given that the default CABLE model uses unreal-istically low and unrealistically geographically stable values for

a key leaf parameter (turnover rate), it is possible, perhapslikely, that the default CABLE (and perhaps other models) mayagree with empirical observations because other aspects of themodel have compensating errors that were not previously identified.Such errors may be invisible in models gradually developed,parameterized, and compared with observations, because un-intentional tuning can occur that causes models to match eachother and observations under conditions for which they weredeveloped but diverge under novel conditions that additionallychallenge the models (29). We observed something similar hereusing a sensitivity analysis to 3 °C warming.The simulated sensitivity of annual GPP to a warming of 3 °C

is quite different for each of the seven versions (Table 1) ofCABLE (Fig. S4), even when their predictions for current con-ditions are similar. The sensitivity of the simulated annual GPPby the default CABLE to a 3 °C warming is quite small for eachof four zones. The sensitivity of the simulated annual GPP to a3 °C warming is negative if NL is constant but realistic (cALLcNLc%N) or is realistically varied with MAT (cALLfNLc%N). In con-trast, sensitivity to warming is positive if variable biomass allocationis used (Fig. S4). However, it has been observed that growth andproductivity of boreal trees and forests increase with warming, butmore further to the north (30–32), which is only mirrored in CA-BLE predictions if NL, needle %N, and biomass allocation are allbiogeographically varied (fALLfNLf%N).One take-home message of the CABLE modeling work (Figs.

3–5 and Fig. S4) is that incorporating more realistic parametersettings or functional equations into a model may initially makepredictions worse rather than better, but such exploration is stillcritically needed, because it has the potential to help identifyother ways in which a model could be, and should be, improved.In this particular case, modifying CABLE in three ways—withbiogeographically varying allocation, NL, and %N—improvedpredictions in some ways (compared with the original defaultversion of the model), such as LAI and sensitivity of GPP totemperature. This will not necessarily happen however, uponmaking models more realistic, because the complexity of manyland surface models could result in performance declines withimproved parameterization. We view such a process—of onestep backward and two steps forward—as positive rather thannegative, because it provides an opportunity to “get under thehood” of the model and find whatever else may need improving.In summary, our observations of intraspecific needle trait varia-

tion support the idea that some important vegetation attributeslikely vary in repeatable patterns along climate gradients, and thatidentifying such patterns can help to advance our general un-derstanding of biogeographic macroecology. That such patternscan assist the development of enhanced global models is furtherreason to encourage continued exploration in this domain.

MethodsNeedle traits for P. banksiana, P. mariana, P. glauca, and A. balsamea weresampled from individual 2.5- to 5-m-tall trees from 15- to 30-y-old forests,after full leaf expansion and before senescence in the summers of 2005 and2006. Sampling of canopy leaf traits was made at 127 sites encompassing anarea from 43.2° to 67.0°N, and 84.4° to 117.7°W. Dominant upland treespecies (including five deciduous angiosperms in addition to the four coni-fers) were sampled in zones of ∼0.5 ° latitude (∼111 km) delineated beforetravel. The first suitable site observed with at least three species was sam-pled and supplemental sites were added as feasible until as many of theavailable target species as possible were sampled from each zone. All targetspecies present were sampled at each site. At least two individuals weresampled from each site. Sampling occurred within forest stands and >25 mfrom disturbed areas such as roads. Only healthy/vigorous trees were sam-pled. Lateral branches with southern exposure from the lower portion of thetop 25% of the crown were cut using telescoping clippers. From thosebranches, prior year (1-y-old) needles were sampled and needle life span wasevaluated. Branches with male cones, cone scars (in P. bansksiana), disease,or insect damage were avoided. These species produce needles once per yearin spring. Thus, needle life span was calculated by counting all living annualcohorts (from terminal scar to terminal scar) up to and including the youn-gest cohort with >50% but <90% of needles remaining (18, 33). This was

Fig. 5. Comparison of GPP (A) and LAI (B) predictions from the CABLE landsurface to observed FLUXNET data from Luyssaert et al. (3). The comparisonsare shown for the predictions under seven versions, representing differentmodel permutations (Table 1).

Reich et al. PNAS | September 23, 2014 | vol. 111 | no. 38 | 13707

ECOLO

GY

SPEC

IALFEATU

RE

done on the canopy branch or highest accessible branch by hand if annualcohorts extended beyond what was present on an individual branch. Lifespan of the oldest cohort retaining >50% of the original live foliage wasestimated to the nearest 10th of a cohort (see refs. 18 and 33). Becausecohorts younger than this one tended to retain nearly all of their foliage,and cohorts older than this tended to retain few, this technique providesa reasonable estimate of both mean cohort and mean needle longevity.

Needles were scanned for leaf area and killed and dried in a 65 °C dryingoven within 12 h of their collection. Leaves or needles from individual treeswere ground into a composite sample and analyzed for analyzed for %N atthe University of California, Davis Stable Isotope Facility (Department ofPlant Sciences) using an Elementar Vario EL Cube or Micro Cube elementalanalyzer (Elementar Analysensysteme GmbH) interfaced to a PDZ Europa20-20 isotope ratio mass spectrometer (Sercon Ltd.).

CABLE is the Australian community land surface model (22, 34) and hasbeen implemented into the Australian community earth system model(ACCESS) (34). CABLE simulates the exchange of momentum, heat, water,and CO2 exchange between land surface and lower atmosphere at hourlytime steps, plant growth, litter fall, and biogeochemical cycling of carbon,nitrogen, and phosphorus at daily time steps. As in most global land surfacemodels, 13 plant functional types are used to approximate the variation ofvegetated surface on Earth and each plant functional type is defined by theunique set of parameters values that are estimated from calibration orpublished literature (see ref. 21). CABLE performs well compared with otherglobal land surface models at site level (see ref. 35) and can also reproducemajor global biogeochemical pools and fluxes quite well (see ref. 21). Themodel has recently been used to study the effects of correlations among keyleaf traits on the simulated gross primary production (36).

For this study, our goal was to explore how different assumptions aboutneedle life span, needle N concentration, and the fraction of NPP allocation tofoliage affect the simulated GPP, NPP, and canopy LAI of evergreen needleleaf forests in North America, Europe, and Asia. For some of the CABLEscenarios we generated temperature-dependent NL (logNL = 0.8246–

0.0320*MAT°C) based on regression of NL in relation to MAT that includedall available data for all five species (n = 845). In CABLE we also generatedfoliage N (mg/g) as a function of NL (logN = 0.1356–0.2252*logNL) based onregressing N vs. NL, including all available data for all five species (n = 391).To account for the decline in needle N with canopy depth within the canopy,as observed for evergreen conifer forests (e.g., refs. 37 and 38), we assumedthat needle N decreases exponentially with the cumulative canopy LAI (seeref. 39). This adjustment has a similar impact on whole-canopy N as adjustingN by weighting the canopy by proportion in different needle age classes andadjusting for differences in %N with needle age. We drove CABLE by reusingthe three-hourly 1° by 1° meteorological forcings for 1990 from the GlobalSoil Water Project II (40, 41) year after year until the simulated annual GPPand NPP reached their steady-state values. All of the simulated results asreported here are steady-state values. Additionally, to analyze the sensitiv-ities of simulated annual GPP, NPP, and canopy LAI to a 3 °C warming, werecalculated the parameter values (fraction of NPP allocated to needle,needle life span, and %N) that vary with MAT using MAT+3 if applicable forevery land point within the domain of this study for each of the sevenscenarios, and ran CABLE by adding 3 °C to the surface air temperatureacross every time step. CABLE was run by reusing the GSWPII meteorologicalforcing for 1990 until steady-state values of GPP, NPP, and canopy LAI wereobtained. The sensitivities of annual GPP, NPP, and canopy LAI to 3 °Cwarming were then calculated using the results from the simulations hereand previously without 3 °C warming.

ACKNOWLEDGMENTS. P.B.R. thanks the Wilderness Research Foundation;the Institute on the Environment (University of Minnesota), and the U.S.National Science Foundation (DEB-1234162); Y.-P.W. thanks the AustralianDepartment of Climate Change and Energy Efficiency for financial support.X.L. thanks the China Scholarship Council for financial support for his studyin Australia. J.O.’s work was supported by National Science Center (Poland)Grant 2011/02/A/NZ9/00108 and Ministry of Science and Higher Education(Poland) Grant N N304 375738.

1. McGuire AD, et al. (2002) Environmental variation, vegetation distribution, carbondynamics, and water/energy exchange in high latitudes. J Veg Sci 13:301–314.

2. Van Cleve K, Yarie J, Erickson R, Dyrness CT (1993) Nitrogen mineralization and ni-trification in successional ecosystems on the Tanana River floodplain, interior Alaska.Can J Res 23:970–978.

3. Luyssaert S, et al. (2007) Photosynthesis drives anomalies in net carbon-exchange ofpine forests at different latitudes. Glob Change Biol 13(10):2110–2127.

4. Reich PB (2012) Key canopy traits drive forest productivity. Proc Biol Sci 279(1736):2128–2134.

5. Burns RM, Honkala BH (1990) Silvics of North America: 1. Conifers; 2. Hardwoods.Agriculture Handbook 654 (US Dept Agric, Forest Service, Washington, DC), Vol 2.

6. Turesson G (1922) The genotypical response of the plant species to the habitat. He-reditas 3:211–350.

7. Clausen J, Keck DD, Hiesey WM (1948) Experimental studies on the nature of species.III. Environmental responses of climatic races of Achillea. Publication 581 (CarnegieInst Washington, Washington, DC).

8. Langlet O (1959) A cline or not a cline – a question of Scots pine. Silvae Genet 8:13–22.9. Wright JW, Bull WI (1963) Geographic variation in Scotch pine. Silvae Genet 12:1–25.10. Oleksyn J, Tjoelker MG, Reich PB (1992) Growth and biomass partitioning of pop-

ulations of European Pinus sylvestris L. under simulated 50° and 60° N daylengths:evidence for photoperiodic ecotypes. New Phytol 120:561–574.

11. Zhang JW, Marshall JD (1995) Variation in carbon isotope discrimination and pho-tosynthetic gas exchange among populations of Pseudotsuga menziesii and Pinusponderosa in different environments. Funct Ecol 9:402–412.

12. Jonas CS, Geber MA (1999) Variation among populations of Clarkia unguiculata(Onagraceae) along altitudinal and latitudinal gradients. Am J Bot 86(3):333–343.

13. Van Cleve K, Oliver L, Schlentner R, Viereck LA, Dyrness CT (1983) Productivity andnutrient cycling in taiga forest ecosystems. Can J Res 13:747–766.

14. Reich PB, Walters MB, Ellsworth DS (1992) Leaf lifespan in relation to leaf, plant andstand characteristics among diverse ecosystems. Ecol Monogr 62:365–392.

15. Wright IJ, et al. (2004) The worldwide leaf economics spectrum. Nature 428(6985):821–827.16. Albert CH, et al. (2010) Intraspecific functional variability: Extent, structure and

sources of variation. J Ecol 98:604–613.17. Pravdin LF (1969) Scots Pine Variation, Interspecific Taxonomy and Selection (Izd.

Nauka, Moscow, 1964) trans. Israel Program for Scientific Translations, Jerusalem.18. Reich PB, Oleksyn J, Modrzynski J, Tjoelker MG (1996) Evidence that longer needle

retention of spruce and pine populations at high elevations and high latitudes islargely a phenotypic response. Tree Physiol 16(7):643–647.

19. Schoettle AW (1990) The interaction between leaf longevity and shoot growth and foliarbiomass per shoot in Pinus contorta at two elevations. Tree Physiol 7(1–4):209–214.

20. Reich PB, Walters MB, Ellsworth DS (1997) From tropics to tundra: Global convergencein plant functioning. Proc Natl Acad Sci USA 94(25):13730–13734.

21. Wang YP, Law RM, Pak B (2010) A global model of carbon, nitrogen and phosphoruscycles for the terrestrial biosphere. Biogeosciences 7:2261–2282.

22. Wang YP, et al. (2011) Diagnosing errors in a land surface model (CABLE) in the timeand frequency domains. J Geophys Res 116:G01034.

23. Wolf A, Field CB, Berry JA (2011) Allometric growth and allocation in forests: Aperspective from FLUXNET. Ecol Appl 21(5):1546–1556.

24. Wolf A, et al. (2011) Forest biomass allometry in global land surface models. GlobalBiogeochem Cycles 25:GB3015.

25. Vogel JG, et al. (2008) Carbon allocation in boreal black spruce forests across regionsvarying in soil temperature and precipitation. Glob Change Biol 14:1503–1516.

26. Reich PB, et al. (2014) Temperature drives global patterns in forest biomass distri-bution in leaves, stems, and roots. Proc Natl Acad Sci USA 111:13721–13726.

27. Kikuzawa K, Onoda Y, Wright IJ, Reich PB (2013) Mechanisms underlying globaltemperature-related patterns in leaf longevity. Glob Ecol Biogeogr 22:982–993.

28. Yoshikawa C, Kawamiya M, Kato T, Yamanaka Y, Matsuno T (2008) Geographicaldistribution of the feedback between future climate change and the carbon cycle.J Geophys Res 113:G03002.

29. Friedlingstein P, et al. (2006) Climate-carbon cycle feedback analysis, results from theC4MIP model intercomparison. J Clim 19:3337–3353.

30. Nemani RR, et al. (2003) Climate-driven increases in global terrestrial net primaryproduction from 1982 to 1999. Science 300(5625):1560–1563.

31. Kimball JS, et al. (2007) Recent climate-driven increases in vegetation productivity forthe western arctic: Evidence of an acceleration of the northern terrestrial carboncycle. Earth Interact 11:1–30.

32. Reich PB, Oleksyn J (2008) Climate warming will reduce growth and survival of Scotspine except in the far north. Ecol Lett 11(6):588–597.

33. Perez Hardguindeguy N, et al. (2013) New handbook for standardised measurementof plant functional traits worldwide. Aust J Bot 61(3):167–234.

34. Kowalczyk EA, et al. (2013) The land surface model component of ACCESS: descriptionand impact on the simulated surface climatology. Aust Meteorol Oceanogr J 63:65–82.

35. Abramowitz G, Pitman AJ, Gupta H, Kowalczyk EA, Wang YP (2007) Systematic bias inland surface models. J Hydrometeorol 8:989–1001.

36. Wang YP, et al. (2012) Correlations among leaf traits provide a significant constrainton the estimate of global gross primary production. Geophys Res Lett 39:L19405.

37. Dang Q-L, et al. (1997) Profiles of PAR, nitrogen and photosynthetic capacity in theboreal forest: Implications for scaling from leaf to canopy. J Geophys Res BOREASSpecial Issue 102(D24):28845–28860.

38. Nippert JB, Marshall JD (2003) Sources of variation in ecophysiological parameters inDouglas-fir and grand fir canopies. Tree Physiol 23(9):591–601.

39. Lloyd J, et al. (2010) Optimisation of photosynthetic carbon gain and within-canopy gra-dients of associated foliar traits for Amazon forest trees. Biogeosciences 7:1833–1859.

40. Dirmeyer PA, et al. (2006) GSWP-2- multimodel analysis and implications for ourperception of the land surface. Bull Am Meteorol Soc 87:1381–1397.

41. Xia JY, Luo YQ, Wang YP, Weng ES, Hararuk O (2012) A semi-analytic solution toaccelerate spin-up of a coupled carbon and nitrogen land model to steady state.Geosci Model Dev 5:1259–1271.

13708 | www.pnas.org/cgi/doi/10.1073/pnas.1216054110 Reich et al.