An extensive suite of functional ... - hippnet.hawaii.edu · Resources and Environmental...

Functional Ecology 20181ndash23 wileyonlinelibrarycomjournalfec emsp|emsp1copy 2018 The Authors Functional Ecology copy 2018 British Ecological Society

Received13September2017emsp |emsp Accepted25September2018DOI1011111365-243513229

R E S E A R C H A R T I C L E

An extensive suite of functional traits distinguishes Hawaiian wet and dry forests and enables prediction of species vital rates

Camila D Medeiros1 emsp|emspChristine Scoffoni12emsp|emspGrace P John1 emsp|emsp Megan K Bartlett13 emsp|emspFaith Inman‐Narahari4emsp|emspRebecca Ostertag5emsp|emsp Susan Cordell6 emsp|emspChristian Giardina6 emsp|emspLawren Sack1

1DepartmentofEcologyandEvolutionaryBiologyUniversityofCaliforniaLosAngelesCalifornia2Department of Biological Sciences California State UniversityLosAngelesCalifornia3Department of Ecology and Evolutionary Biology Princeton University Princeton New Jersey 4Department of Natural Resources and Environmental Management University of Hawaii at Manoa Honolulu Hawaii 5Department of Biology University of Hawaii at Hilo Hilo Hawaii and 6InstituteofPacificIslandsForestryPacificSouthwestResearchStationUSDAForestServiceHiloHawaii

CorrespondenceLawren SackEmail lawrensackgmailcom

Funding informationNational Science Foundation Brazilian National Research Council (CNPq) Brazilian Science Without Borders Program GrantAwardNumber2028132014-2

Handling Editor Cyrille Violle

Abstract1 The application of functional traits to predict and explain plant speciesrsquo distribu-

tions and vital rates has been a major direction in functional ecology for decades yet numerous physiological traits have not yet been incorporated into the approach

2 Usingcommonlymeasuredtraitssuchasleafmassperarea(LMA)andwoodden-sity (WD) and additional traits related to water transport gas exchange and re-source economics including leaf vein stomatal and wilting traits we tested hypotheses for Hawaiian wet montane and lowland dry forests (MWF and LDF respectively) (1) Forests would differ in a wide range of traits as expected from contrasting adaptation (2) trait values would be more convergent among dry than wet forest species due to the stronger environmental filtering (3) traits would be intercorrelatedwithinldquomodulesrdquosupportinggivenfunctions (4)relativegrowthrate (RGR) and mortality rate (m) would correlate with a number of specific traits with(5)strongerrelationshipswhenstratifyingbytreesizeand(6)RGRandm can bestronglyexplainedfromtrait-basedmodels

3 The MWF speciesrsquo traits were associated with adaptation to high soil moisture and nutrient supply and greater shade tolerance whereas the LDF speciesrsquo traits were associated with drought tolerance Thus on average MWF species achieved higher maximum heights than LDF species and had leaves with larger epidermal cells higher maximum stomatal conductance and CO2 assimilation rate lower vein lengths per area higher saturated water content and greater shrinkage when dry lower dry matter content higher phosphorus concentration lower nitrogen to phosphorus ratio high chlorophyll to nitrogen ratio high carbon isotope discrimi-nation high stomatal conductance to nitrogen ratio less negative turgor loss point and lower WD Functional traits were more variable in the MWF than LDF were

2emsp |emsp emspenspFunctional Ecology MEDEIROS Et al

1emsp |emspINTRODUC TION

Functional traits influence plant growth reproduction and survival and thereby fitness (Lavorel amp Garnier 2002 Violle et al 2007) and thuscanbeusedtopredictvitalrates(Adleretal2014Poorteretal2008 Uriarte Lasky Boukili Chazdon amp Merow 2016) habitat pref-erences (Shipley et al 2017) and spatial distributions (Stahl Reu amp Wirth2014) Fordecadesmost studieshave focusedon relativelyfew commonly measured functional traits with some justification given that overall trait variation can be simplified statistically into a few fundamental dimensions (Diaz et al 2016 Messier et al 2017) However several have argued that more extensive suites of traits would enable strong predictive and explanatory power (Greenwood etal 2017Paine etal 2015Reich 2014YangCaoampSwenson2018a) and this argument has conceptual support because mechanis-tic models of growth and survival are sensitive to a broad set of traits as inputs (Marks amp Leichowicz 2006 Osborne amp Sack 2012 Sterck Markesteijn Schieving amp Poorter 2011) The traits measured in this studyincludewell-studiedfunctionaltraitswithintheleafandwoodldquoeconomics spectrardquo (LES and WES respectively) which describe trade-offs in plant carbon balancewith given traits contributing toeither fast growth and resource turnover or slow growth and longer tissuelifespansandstresstolerance(Chaveetal2009Wrightetal2004)Inadditionweincludedawidersetoftraitsrecognizedtohaveproximal physiological influence on water transport gas exchange and resource economics The aim of this study was to assess six key hy-potheses derived from first principles in trait physiology and ecology (Table1)utilizing45traitsexpectedtoshowcontrastingadaptationacross forests andor to influence relative growth rate (RGRdbh and RGRbiom) and mortality (m) (Table 2) We pursued this aim while rec-ognizing that many more traits than those we included play important roles and that species differ in the traits with most important influence on vital rates

First we tested the ability of an extensive suite of traits to resolve variation between Hawaiian wet and dry forest species given their contrasting adaptation We assessed traits which based on the previous literature would have specific mechanistic influences on resource acquisition growth and stress tolerance (Table 1 with detailed reasoning in Supporting Information Table S10) In particular we expected that relative to the dry forest the wet forest species would have shifted their traits values in

the direction beneficial to their adaptation to greater availabil-ity of water and soil nutrients Such trait shifts would include greater mean and maximum plant height (King Davies amp Noor 2006Koch Sillett JenningsampDavis 2004) lowerwoodden-sity(WDChaveetal2009Gleasonetal2016HackeSperryPockmanDavisampMcCulloh2001)andseedmass(Gross1984KhuranaampSingh2004)higheroverall ratesofphotosynthesisand rates of electron transport and carboxylation (all per unit leaf area andor dry mass) and higher values for the ratio of inter-nal to ambient CO2 (cica) related to higher values of carbon iso-tope discrimination (Δleaf Farquhar EhleringerampHubick1989Franks amp Beerling 2009 Donovan amp Ehleringer 1994 Wanget al 2017) larger and denser stomata and higher stomatal con-ductance (Beaulieu Leitch Patel Pendharkar amp Knight 2008 FranksampBeerling2009FranksampFarquhar2007HetheringtonampWoodward 2003 Sackamp Buckley 2016Wang etal 2015)higher densities of leaf major and minor veins and free ending veins (Brodribb Feild amp Jordan 2007 Iida et al 2016 Sack amp Frole 2006 Sack amp Scoffoni 2013 Scoffoni et al 2016) thinner and larger leaves of higher saturated water content and lower dry mass density lower water mass and dry mass per area and lower dry matter content with lesser shrinkage in area under dehydra-tion (Bartlett Scoffoni amp Sack 2012b Diaz et al 2016 Evans 1973Niinemets2001OgburnampEdwards2012SackampScoffoni2013ScoffoniVuongDiepCochardampSack2014Vendraminietal2002WestobyampWright2006Wrightetal2004)highfoliar concentrations of nitrogen (N) phosphorus (P) and chlo-rophyll (Chl) and lower concentration of carbon (Chatuverdi Raghubanshi amp Singh 2011 Lambers amp Poorter 2004Wrightetal2004) lowerNP(Elseretal2000)andgreaterstomatalopening relative to maximum aperture and relative to N (Franks ampBeerling 2009Wright ReichampWestoby 2001)Given thatspecies of the wet forest are adapted to lower understorey irra-diance also led to the expectation of lower rates of photosynthe-sis and greater Δleaf (DonovanampEhleringer 1994 Evans 2013Farquhar etal 1989 FranksampBeerling 2009) larger leaf area(Chatuverdi etal 2011 Niinemets 2001) lower LMA (SackGrubbampMarantildeoacuten2003bWaltersampReich1999)lowerNandP and higher C and Chlconcentrations(Givnish1988Niinemets2001 Lusk amp Warton 2007 Poorter Niinemets Poorter Wright amp Villar 2009 Chatuverdi etal 2011) higher Chl to N ratio

correlated within modules and predicted speciesrsquo RGR and m across forests with stronger relationships when stratifying by tree size Models based on multiple traits predicted vital rates across forests (R2 = 070ndash072 p lt 001)

4 Ourfindingsareconsistentwithapowerfulroleofbroadsuitesoffunctionaltraitsin contributing to forest speciesrsquo distributions integrated plant design and vital rates

K E Y W O R D S

drought tolerance endemic species forest tree demography growth analysis

emspensp emsp | emsp3Functional EcologyMEDEIROS Et al

TAB

LE 1

emspFrameworkofhypothesesderivedfromfirstprinciplesoftrait-basedphysiologyandecologytotesttheapplicationofanextensivesuiteoftraitstoresolvevariationamong

fore

sts

and

to e

nabl

e pr

edic

tion

of v

ital r

ates

acr

oss

spec

ies

Hyp

othe

sis

Expl

anat

ion

base

d on

firs

t pr

inci

ples

Refe

renc

esTe

stSu

ppor

t

1 W

et a

nd d

ry fo

rest

spe

cies

wou

ld d

iffer

in

num

erou

s tr

aits

as

expe

cted

from

con

tras

ting

adap

tatio

n

Adaptationtocontrastingclimate

and

soil

wou

ld le

ad to

var

iatio

n am

ong

spec

ies

in n

umer

ous

func

tiona

l tra

its im

port

ant i

n pl

ant p

erfo

rman

ce

Schimper(1903)MarksandLeichowicz(2006)

Lohbecketal(2015)andLevineBascompte

AdlerandAllesina(2017)

NestedANOVAsforindividual-leveltraits

and

ttestsforspecies-leveltraits

Yes

2 T

rait

valu

es w

ould

be

mor

e co

nver

gent

am

ong

dry

than

wet

fore

st s

peci

es d

ue to

the

selectivepressureimposedbylow-resource

avai

labi

lity

Envi

ronm

enta

l filt

erin

g is

ex

pect

ed to

redu

ce fu

nctio

nal

dive

rsity

by

cons

trai

ning

the

rang

e of

pos

sibl

e tr

ait s

tate

s ac

ross

hab

itats

Cor

nwel

l et a

l (2

006)

May

field

Bon

i an

d Ackerly(2009)Lebrija-Trejosetal(2010)

Kraftetal(2014)Nathanetal(2016)and

Asefaetal(2017)

t tes

t on

the

coef

ficie

nt o

f var

iatio

n in

trai

ts

from

MW

F an

d LD

F F

test

s on

the

varia

nce

of e

ach

trai

t bet

wee

n M

WF

and

LDF

Yes

3 T

raits

wou

ld b

e in

terc

orre

late

d w

ithin

fu

nctio

nal ldquo

mod

ules

rdquoSe

lect

ion

on m

ultip

le tr

aits

acr

oss

envi

ronm

ents

wou

ld le

ad to

tr

aitndash

trai

t cor

rela

tions

with

in

orga

ns a

nd fu

nctio

nal

mod

ules

du

e to

com

mon

dev

elop

men

tal

path

way

fun

ctio

n o

r ben

efit

in

give

n en

viro

nmen

ts

Sacketal(2003ab)Givnishetal(2005)

PoorterLambersandEvans(2014)andLietal

(2015b)

Pear

son

s co

rrel

atio

ns b

etw

een

trai

ts w

ithin

fu

nctio

nal ldquo

mod

ules

rdquoYes

4RGRand

m w

ould

cor

rela

te w

ith s

peci

fic

trai

tsTr

aits

con

trib

ute

mec

hani

stic

ally

di

rect

ly to

RG

Rs a

nd m

in g

iven

ha

bita

ts

Kitajima(1994)Grime(2001)Sacketal(2013)

Mar

ks a

nd L

eich

owic

z (2

006)

and

Wrig

ht e

t al

(201

0)

Pear

son

s co

rrel

atio

ns b

etw

een

spec

ific

trai

ts

and

vita

l rat

esYes

5RGRand

m w

ould

cor

rela

te w

ith tr

aits

mor

e fr

eque

ntly

whe

n st

ratif

ying

by

tree

siz

eOntogenetic-andsize-related

tren

ds in

trai

ts a

nd v

ital r

ates

m

ean

that

trai

tndashvi

tal r

ate

corr

elat

ions

wou

ld b

e re

duce

d gi

ven

com

paris

on o

f spe

cies

m

ean

valu

es w

hen

spec

ies

vary

in

siz

e di

strib

utio

ns s

trat

ifyin

g by

siz

e sh

ould

ther

efor

e st

reng

then

trai

tndashvi

tal r

ate

rela

tions

hips

Iidaetal(2014)Iidaetal(2016)andPrado-

Juni

or e

t al

(201

6)Ba

yesi

an m

odel

to e

stim

ate

vita

l rat

es a

t gi

ven

plan

t siz

es fo

llow

ed b

y Ke

ndal

ls

corr

elat

ions

bet

wee

n tr

aits

and

vita

l rat

es a

t ea

ch s

ize

Yes

6 R

GR

and

m c

an b

e pr

edic

ted

base

d on

trait-basedmodels

Giv

en re

latio

nshi

ps o

f vita

l rat

es

with

giv

en tr

aits

com

bina

tions

of

trai

ts s

houl

d be

str

ongl

y pr

edic

tive

Poor

ter e

t al

(200

8) U

riart

e et

al

(201

6) a

nd

Thom

as a

nd V

esk

(201

7)Li

near

regr

essi

onYes


TAB

LE 2

emspSt

udy

trai

ts re

latin

g to

sto

mat

al m

orph

olog

y le

af v

enat

ion

leaf

and

woo

d ec

onom

ics

and

stru

ctur

e le

af c

ompo

sitio

n a

nd e

stim

ated

pho

tosy

nthe

sis

and

plan

t siz

e a

nd th

e vi

tal

rate

s m

easu

red

for s

peci

es fr

om a

mon

tane

wet

fore

st (W

) and

a lo

wla

nd d

ry fo

rest

(D) i

n H

awai

i F

or th

e tr

aits

we

prov

ide

sym

bols

uni

ts h

ypot

hese

s fo

r giv

en tr

aits

for d

iffer

ence

s be

twee

n fo

rest

s an

d re

sults

from

sta

tistic

al te

sts

and

hyp

othe

ses

for c

orre

latio

ns w

ith v

ital r

ates

(rel

ativ

e gr

owth

rate

and

mor

talit

y) a

nd re

sults

from

Pea

rson

s co

rrel

atio

n te

sts

(whe

n on

e re

sult

is p

rese

nted

this

repr

esen

ts s

peci

es fr

om b

oth

fore

sts

toge

ther

and

whe

n tw

o re

sults

are

pre

sent

ed th

ese

repr

esen

t spe

cies

in th

e w

et a

nd d

ry fo

rest

s se

para

tely

) an

d re

fere

nces

su

ppor

ting

the

hypo

thes

es n

s in

dica

tes

no s

igni

fican

t diff

eren

ce a

t plt005ldquoW

rdquorepresentstheexpectationthatallelsebeingequalgiventhespecifichypothesisthewetforestwould

have

a h

ighe

r tra

it va

lue

than

the

dry

fore

st o

n av

erag

e ldquoD

rdquo tha

t the

dry

fore

st w

ould

hav

e th

e hi

gher

trai

t val

ue o

n av

erag

e a

nd ldquoe

ither

rdquo den

otes

the

exis

tenc

e of

mul

tiple

pub

lishe

d hy

poth

eses

whe

reby

eith

er M

WF

or L

DF

coul

d be

exp

ecte

d to

hav

e th

e hi

gher

trai

t val

ue (S

uppo

rtin

g In

form

atio

n Ta

ble

S10)

Pos

itive

sig

ns (+

) ind

icat

e th

e ex

pect

atio

n or

find

ing

of a

pos

itive

correlationwithrelativegrowthrateandmortalityratenegativesigns(minus)indicatetheoppositeFordetailedreasoningbehindeachhypothesisandreferencesseeSupportingInformation

Tabl

e S1

0 plt005p

lt 0

01

p lt

000

1

Trai

tvi

tal r

ate

Sym

bol

Uni

tH

ypot

hese

s W

or D

hi

gher

W

or D

hig

her

Hyp

othe

ses

trai

tndashvi

tal

rate

cor

rela

tion

Dire

ctio

n of

trai

tndashvi

tal

rate

cor

rela

tion

Refe

renc

e

Stom

atal

mor

phol

ogy

Stom

atal

den

sity

dst

omat

am

m2

eith

erns

++

1ndash5

Stom

atal

diff

eren

tiatio

n ra

te (o

r ind

ex)

i-

eith

erW

+

+

2ndash6

Stom

atal

are

as

μm2

WW

+

ns157

Gua

rd c

ell l

engt

hG

CL

μmW

W

157

Gua

rd c

ell w

idth

GC

Wμm

WW

157

Pore

leng

thSP

Lμm

WW

157

Epid

erm

al p

avem

ent c

ell a

rea

eμm

2W

W

8

Max

imum

sto

mat

al c

ondu

ctan

ceg m

axm

ol mminus2

sminus1

eith

erW

+

+2ndash59ndash10

Leaf

ven

atio

n

Maj

or v

ein

dens

ityVLA

maj

orm

m p

er m

m2

eith

erD

+ns

2ndash411ndash15

Min

or v

ein

dens

ityVLA

min

orm

m p

er m

m2

eith

erD

+ns

2ndash411ndash15

Tota

l vei

n de

nsity

VLA

tota

lm

m p

er m

m2

eith

erD

+ns

2ndash411ndash15

Free

end

ing

vein

den

sity

FEV

pe

r mm

2ei

ther

D

2ndash411ndash15

Leaf

and

woo

d ec

onom

ics

and

stru

ctur

e

Leaf

are

aLA

cm2

Wns

+WnsDminus

16ndash1

8

Leaf

mas

s pe

r are

aLMA

gm

2ei

ther

nsminus

W+Dminus

12 1

8ndash23

Leaf

thic

knes

sLT

mm

eith

erns

minusminus

12 1

8ndash23

Leaf

den

sity

LDg

cm3

eith

erns

minusW+Dminus

12 1

8ndash23

Leaf

dry

mat

ter c

onte

ntLD

MC

gg

DD

minus

ns1824

Satu

rate

d w

ater

con

tent

SWC

gg

eith

erW

25ndash27

Wat

er m

ass

per a

rea

WMA

gm

2D

ns25ndash27

Perc

enta

ge lo

ss a

rea

(dry

)PLA

dry

W

W

2829

Woo

d de

nsity

WD

gcm

3ei

ther

D

minus

ns530ndash32

(Con

tinue

s)


Trai

tvi

tal r

ate

Sym

bol

Uni

tH

ypot

hese

s W

or D

hi

gher

W

or D

hig

her

Hyp

othe

ses

trai

tndashvi

tal

rate

cor

rela

tion

Dire

ctio

n of

trai

tndashvi

tal

rate

cor

rela

tion

Refe

renc

e

Leaf

com

posi

tion

Nitr

ogen

con

cent

ratio

n pe

r lea

f are

aN

area

gm

2ei

ther

ns+

W+Dminus

2ndash520

Nitr

ogen

con

cent

ratio

n pe

r lea

f mas

sN

mas

sm

gg

eith

erns

++

2ndash520

Phos

phor

us c

once

ntra

tion

per l

eaf a

rea

P area

gm

2ei

ther

ns+

W+Dminus

2ndash520

Phos

phor

us c

once

ntra

tion

per l

eaf m

ass

P mas

sm

gg

eith

erW

+

+2ndash520

Chl

orop

hyll

conc

entr

atio

nCh

l area

SPAD

eith

erns

2ndash433

Chl

orop

hyll

per m

ass

Chl m

ass

SPADgminus1

m2

eith

erns

2ndash433

Car

bon

conc

entr

atio

n pe

r lea

f mas

sC

mas

sm

gg

Wns

34ndash35

Nitr

ogen

pho

spho

rus

ratio

NP

ndashei

ther

D

ndashns

35

Chl

orop

hyll

nitr

ogen

per

are

a ra

tioCh

l area

Nar

eaSPADgminus1

m2

Wns

5

Car

bon

isot

ope

disc

rimin

atio

nΔ

leaf

permilW

W

36ndash39

|Tur

gor l

oss

poin

t||π

tlp|

MPa

DD

28

Estim

ated

pho

tosy

nthe

sis

Elec

tron

tran

spor

t rat

e pe

r are

aJm

axar

eaμm

ol mminus2

sminus1

eith

erns

+W+Dminus

2ndash436ndash38

Elec

tron

tran

spor

t rat

e pe

r mas

sJm

axm

ass

nmol

gminus1

sminus1

eith

erns

++

2ndash436ndash38

Max

imum

rate

of c

arbo

xyla

tion

per a

rea

Vcm

axar

eaμm

ol mminus2

sminus1

eith

erns

+W+Dminus

2ndash436ndash38

Max

imum

rate

of c

arbo

xyla

tion

per m

ass

Vcm

axm

ass

nmol

gminus1

sminus1

eith

erns

++

2ndash436ndash38

Ratio

of i

nter

cellu

lar t

o am

bien

t CO

2 co

ncen

trat

ions

c ica

ndashei

ther

W

+

ns36

ndash38

Tim

e in

tegr

ated

leaf

CO

2 ass

imila

tion

rate

per

ar

ea Aarea

μmol

mminus2

sminus1

eith

erns

+W+Dminus

9

Tim

e in

tegr

ated

leaf

CO

2 ass

imila

tion

rate

per

m

ass

Amass

nmol

gminus1

sminus1

eith

erW

++

9

Tim

e in

tegr

ated

sto

mat

al c

ondu

ctan

cegcleaf

mm

ol mminus2

sminus1

eith

erW

+

W+Dminus

9

Tim

e in

tegr

ated

max

imum

sto

mat

al c

ondu

ct-

ance

ratio

gcleaf

g max

ndashei

ther

ns2ndash49

Max

imum

sto

mat

al c

ondu

ctan

cen

itrog

en p

er

area

ratio

g max

Nar

eam

ol gminus1

sminus1

WW

40

Plan

t siz

e

Mea

n he

ight

Hm

Wns

46ndash47

Max

imum

hei

ght

Hm

axm

WW

46ndash47

Seed

mas

sSM

mg

Dns

48ndash49

TAB

LE 2

emsp(C

ontin

ued)

(Con

tinue

s)


(Givnish1988) and lower stomatal andveindensities (Givnish1988SackampScoffoni2013Sacketal2012)Theliteraturealsosupports contrasting hypotheses in which dry forest species gain drought tolerance by achieving higher photosynthetic activity when water is available linked with smaller and more numerous stomata and epidermal pavement cells (Grubb 1998Maximov1931 Scoffoni RawlsMcKown Cochard amp Sack 2011Wanget al 2017) higher vein densities (Sack amp Scoffoni 2013) and high N and P per mass (Wright et al 2001) We also expected the dry forest species to have more negative turgor loss point (Bartlett et al 2012b) thick and small leaves (Sack et al 2012 Wright etal 2017) and highWD (Chave etal 2009 Gleasonet al 2016 Hacke et al 2001) and traits associated with high water use efficiency reflected in low cica and carbon isotope dis-crimination(DonovanampEhleringer1994Farquharetal1989)

Second we tested the hypothesis that on average species of the dry forest would have narrower ranges in trait values than the wetforest (NathanOsemShachakMeronampSalguero-Goacutemez2016) Two main processes of community assembly affect func-tional diversity at local scale environmental (or habitat) filter-ing and biotic interactions (Asefa etal 2017 Chesson 2000Cornwell Schwilk amp Ackerly 2006) In low-resource habitatsenvironmental filtering is expected to more strongly constrain trait diversity as would the reduction of biotic interactions which would promote greater niche overlap (Lebrija-Trejos MeavePoorter Peacuterez-Garciacutea amp Bongers 2010 Nathan etal 2016WeiherampKeddy1995)

Third we tested the hypothesis that traits would be intercorrelated in ldquomodulesrdquo due to their contributions to given functions (Li etal 2015b Sack Cowan Jaikumar amp Holbrook2003a) or ldquostrategiesrdquo (Westoby Falster Moles Vesk amp Wright 2002) Modules are defined as clusters of traits that show co-variation among themselves due to selection but are relatively independent of other clusters (Armbruster Pelabon Bolstad ampHansen2014WagnerampAltenberg1996)Suchco-selectionhasbeen a main explanation for why plant phenotypes are organized into dimensions (or axes) such as the leaf and wood economic spectra (Chave etal 2009 Wright etal 2004) Several of thenewly added traits are expected to be mechanistically related to traits from the LES and WES and are therefore grouped within the same trait modules (Table 2)

Fourth we hypothesized that across species RGR and m would bepositively correlateddue to life-history trade-offs andparallelassociationswithgiventraits(Kitajima1994Philipsonetal2014Russo et al 2010 Visser et al 2016 Wright et al 2010) Further we hypothesized that RGR and m would relate positively to pho-tosynthetic rate (Donovan amp Ehleringer 1994 Franks amp Beerling2009) leaf area (Iida etal 2016) N and P concentrations (Iidaet al 2016 Osone Ishida amp Tateno 2008) the sizes and numbers ofstomata (HetheringtonampWoodward2003Wangetal2015)maximum stomatal conductance and vein densities (Hetherington ampWoodward2003 Iidaetal2016)andnegatively toLMA (Iidaet al 2016 Osone et al 2008 Wright et al 2010) leaf thickness Tr

ait

vita

l rat

eSy

mbo

lU

nit

Hyp

othe

ses

W o

r D

high

er

W o

r D h

ighe

rH

ypot

hese

s tr

aitndash

vita

l ra

te c

orre

latio

nD

irect

ion

of tr

aitndash

vita

l ra

te c

orre

latio

nRe

fere

nce

Vita

l rat

es

Rela

tive

grow

th ra

te (d

iam

eter

incr

emen

t)RG

R dbh

cm c

mminus1

yea

rminus1

eith

erns

52041

Rela

tive

grow

th ra

te (b

iom

ass

incr

emen

t)RG

R biom

kg k

gminus1 y

earminus1

eith

erns

52041

Mor

talit

y ra

tem

p

er y

ear

eith

erD

42ndash45

References1HetheringtonandWoodward(2003)2Maximov(1931)3Grubb(1998)4Scoffonietal(2011)5Givnish(1988)6SackandBuckley(2016)7FranksandFarquhar(2007)8Beaulieu

etal(2008)9FranksandBeerling(2009)10Wangetal(2015)11SackandFrole(2006)12Brodribbetal(2007)13SackandScoffoni(2013)14Iidaetal(2016)15Scoffonietal(2016)16Sack

etal(2012)17Wrightetal(2017)18Niinemets(2001)19Evans(1973)20Wrightetal(2004)21WestobyandWright(2006)22LuskandWarton(2007)23Poorteretal(2009)24Diazetal

(2016)25Vendraminietal(2002)26SackTyreeandHolbrook(2005)27OgburnandEdwards(2012)28Bartlettetal(2012ab)29Scoffonietal(2014)30Hackeetal(2001)31Chaveetal

(2009)32Gleasonetal(2016)33Chatuverdietal(2011)34LambersandPoorter(2004)35Elseretal(2000)36Farquharetal(1989)37DonovanandEhleringer(1994)38Evans(2013)39

Wangetal(2017)40Wrightetal(2001)41Gibertetal(2016)42Wrightetal(2010)43McDowelletal(2008)44McDowelletal(2018)45KobeandCoates(1997)46Kochetal(2004)47

Kingetal(2006)48Gross(1984)49KhuranaandSingh(2004)

TAB

LE 2

emsp(C

ontin

ued)


density and dry matter content (Iida et al 2016 Niinemets 2001) NP(Elseretal2000)andWD(Philipsonetal2014Visseretal2016 Wright et al 2010) We also tested whether trait relationships withvitalratesdifferedbetweenforests(KobeampCoates1997LuskReichMontgomeryAckerlyampCavender-Bares2008)

Fifth we expected to uncover more relationships of traits with vital rates when accounting for tree size (Iida etal 2014 2016Prado-Junioretal2016)

Finally based on the expectations of strong traitndashvital rate asso-ciations we hypothesized that RGR and m can be predicted based on trait-basedmodels

Our study focused on Hawaiian forests with low species diver-sity located across highly contrasting environments (Table 3 Price amp Clague2002OstertagInman-NarahariCordellGiardinaampSack2014)Bytestingourframeworkofhypotheseswemoregenerallyaddressed the question of whether considering an extensive suite ofmechanistictraitshasvaluefortrait-basedecologicaltheoryandapplications

2emsp |emspMATERIAL S AND METHODS

For additional details for each methods section see correspondingly namedsectioninSupportingInformationMethodsAppendixS1

21emsp|emspStudy sites

The study was based in forest dynamics plots (FDPs) on Hawairsquoi Island within montane wet forest (MWF) and within lowland dry for-est (LDF) part of the Hawairsquoi Permanent Plot Network established in2008ndash09(HIPPNETFigure1Supporting InformationMethodsOstertagetal2014)TheMWFandLDFplotscontraststronglyinclimate and soil composition The substrate in the MWF is formed from weathered volcanic material and is old deep and moderately well drained while LDF has younger shallow and highly organic sub-strate (websoilsurveynrcsusdagov) The forests also have distinct species with only Metrosideros polymorpha common to both being thecanopyco-dominantintheMWFandlimitedtoafewindividualsin the LDF

Both FDPs were established using the standard methodology of the Center for Tropical Forest Science global FDP network (Condit 1998)From2008to2009alllivenativewoodyplantsge1cmdiam-eter at breast height (DBH at 130 cm) were tagged and mapped rel-ativeto5mtimes5mgridsinstalledthroughouttheplotsandmeasuredforDBH(Ostertagetal2014)

Some of our study questions were addressed by comparing these single forests that were selected to be highly represen-tative of their forest type an approach previously used in many ecophysiological comparisons of forests (eg Baltzer Davies Bunyavejchewin amp Noor 2008 Blackman Brodribb amp Jordan 2012Falcatildeoetal2015Markesteijn IraipiBongersampPoorter2010 Zhu Song Li amp Ye 2013)Notably statistical differencesbetween forests are not necessarily generalizable but enable

refined hypotheses for testing in future studies of replicate for-ests of each type However when predicting speciesrsquo vital rates from traits statistical significance is expected to reflect a higher generality as each species represents a replicate data point (Sokal amp Rohlf 2012)

22emsp|emspMeasurement of relative growth rate and mortality

Atotalof21805individualtreesof29speciesfrombothforestplotsweremeasuredforDBHinthefirstcensus2008andthe18745ofthose trees that were alive were remeasured in the second census in 2013 From individual plant DBH in both censuses we used the function ldquoAGBtreerdquo available in the ldquoCTFS R Packagerdquo (ctfssieduPublicCTFSRPackage)tocalculateabove-groundbiomassusingal-lometric equations specific for ldquowetrdquo and ldquodryrdquo forests that use DBH and wood density as species-specific inputs (Chave etal 2005)WethencalculatedrelativegrowthratesinDBHandabove-groundbiomass (RGRdbh and RGRbiom respectively) as ln (xt1)minusln (xt0)

Δt where x is

DBHorabove-groundbiomassand∆t is the time between measure-ments (in years) RGRdbh is the most commonly used in the literature but RGRbiom is arguably most relevant for relating mechanistically to traitsononehandandtoforestscaleprocessesontheother(Gil-PelegriacutenPeguero-PinaampSancho-Knapik2017)Annualmortalityrate (m)wascalculatedforeachofthesame29speciesusingsurvivaldata from both censuses as m= [1minus (N1∕N0)

(1∕ Δt)]times100 where N1 is the number of live individuals at census 2 N0 is the number of live individuals at census1 and∆t is the time between measure-ments(inyearsSheilBurslemampAlder1995)Duetothepotentialfor demographic stochasticity in small populations to affect vital rate estimatesspecieswithlt15individualswereexcludedfromanalysesof RGR and m (Fiske Bruna amp Bolker 2008) for RGRdbh the mean coefficient of variation was fivefold higher for species with nlt15than those with ngt15individuals(80and16respectively)

23emsp|emspSeed mass and maximum height

Speciesrsquo mean height (H) was calculated across all individuals in the plot estimated from allometries (Ostertag etal 2014) andmaxi-mum height (Hmax)wascalculatedas the95thpercentileheightofeach species Seed dry mass values were compiled from seed banks acrossHawairsquoi(LSackampAYoshinagaunpublisheddata)

24emsp|emspSampling for leaf and wood trait measurements

We sampled all native woody species from both FDPs that is 20 spe-ciesintheMWFand15speciesintheLDF(Table3Ostertagetal2014)Datawerecollectedforfiverandomlyselectedindividualsperspecies given availability in the plot but stomatal and venation traits were measured for only three randomly selected individuals for this study those three individuals per species were used for all trait anal-yses For each individual we used pole pruners to collect the most


TA B L E 3 emsp List of all species from the montane wet forest (MWF) and lowland dry forest (LDF) sites in Hawaii with family species code growth form leaf habit (evergreen E or deciduous D) and type (simple S compound C or phyllode P) and forest stratum Nomenclature followsWagnerHerbstandSommer(1999)withupdatesfromThePlantList(2013)andLuandMorden(2014)

Species Family Code Growth form Leaf habit and type Forest stratum

Montane Wet Forest (MWF)

Acacia koaAGray Fabaceae ACAKOA Tree E P Canopy

Broussaisia arguta Gaudich Hydrangeaceae BROARG Shrub E S Understorey

Cheirodendron trigynum(Gaudich)AHeller

Araliaceae CHETRI Tree E C Canopy

Cibotium chamissoi Kaulf Cibotiaceae CIBCHA Tree fern E C Understorey

Cibotium glaucum(Sm)HookampArn Cibotiaceae CIBGLA Tree fern E C Understorey

Cibotium menziesii Hook Cibotiaceae CIBMEN Tree fern E C Understorey

Clermontia parvifloraGaudichexAGray Campanulaceae CLEPAR Shrub E C Understorey

Coprosma rhynchocarpaAGray Rubiaceae COPRHY Tree E S Sub-canopy

Ilex anomalaHookampArn Aquifoliaceae ILEANO Tree E S Sub-canopy

Kadua axillaris (Wawra) WLWagner amp Lorence

Rubiaceae KADAXI ShrubSmall tree E S Understorey

Leptecophylla tameiameiae (Cham amp Schltdl) CM Weiller

Ericaceae LEPTAM Shrub E S Understorey

Melicope clusiifolia(AGray)TGHartleyamp BC Stone

Rutaceae MELCLU ShrubSmall tree E S Understorey

Metrosideros polymorpha Gaudich Myrtaceae METPOL_W ShrubTall tree E S Canopy

Myrsine lessertianaADC Primulaceae MYRLES Tree E S Sub-canopy

Myrsine sandwicensisADC Primulaceae MYRSAN ShrubSmall tree E S Understorey

Perrottetia sandwicensisAGray Dipentodontaceae PERSAN ShrubSmall tree E S Understorey

Pipturus albidus(HookampArn)AGray Urticaceae PIPALB Shrub E S Understorey

Psychotria hawaiiensis(AGray)Fosberg Rubiaceae PSYHAW Tree E S Sub-canopy

Trematolobelia grandifolia (Rock) O Deg Campanulaceae TREGRA Shrub E S Understorey

Vaccinium calycinum Sm Ericaceae VACCAL Shrub E S Understorey

Lowland dry forest (LDF)

Euphorbia multiformis Gaudich ex Hook ampArn

Euphorbiaceae EUPMUL Shrub D S Understorey

Chrysodracon hawaiiensis (O Degener amp IDegener)P-LLuampMorden

Asparagaceae CHRHAW Tree E S Sub-canopy

Diospyros sandwicensis(ADC)Fosberg Ebenaceae DIOSAN Tree E S Canopy

Dodonaea viscosa Jacq Sapindaceae DODVIS Shrub E S Understorey

Erythrina sandwicensis O Deg Fabaceae ERYSAN Tree D C Canopy

Metrosideros polymorpha Gaudich Myrtaceae METPOL_D ShrubTall tree E S Canopy

Myoporum sandwicenseAGray Scrophulariaceae MYOSAN ShrubSmall tree D S Understorey

Osteomeles anthyllidifolia (Sm) Lindl Rosaceae OSTANT Shrub E C Understorey

Pittosporum terminalioides Planch ex AGray

Pittosporaceae PITTER Tree E S Understorey

Psydrax odorata(GForst)ACSmampSP Darwin

Rubiaceae PSYODO ShrubSmall tree E S Understorey

Santalum paniculatumHookampArn Santalaceae SANPAN ShrubTree E S Canopy

Senna gaudichaudii(HookampArn)HSIrwin amp Barneby

Fabaceae SENGAU Shrub D C Understorey

Sophora chrysophylla (Salisb) Seem Fabaceae SOPCHR ShrubTree D C Canopy

Sida fallax Walp Malvaceae SIDFAL Shrub E S Understorey

Wikstroemia sandwicensis Meisn Thymelaeaceae WIKSAN ShrubTree E S Understorey


exposed mature branch grown in the current year with no signs of damage and herbivory Branches were carried to the laboratory in plastic with moist paper and rehydrated overnight under plastic be-fore harvesting stem sections and fully expanded leaves and stems forallsubsequentanalysesForcompound-leafedspecies(Table3)leaflets were used for Acacia koa phyllodes were used

25emsp|emspLeaf stomatal and venation traits

We measured stomatal and venation traits on one leaf from each of three individuals per species Stomatal measurements were ob-tained from microscopy images taken from nail varnish impressions of both leaf surfaces We measured stomatal density (d) and sto-matal index (ie differentiation rate the number of stomata per numbers of stomata plus epidermal pavement cells i) stomatal pore length (SPL) guard cell length and width (GCL GCW) stomatal area (s) and epidermal pavement cell area (e) (Sack Melcher Liu Middleton amp Pardee 2006) and calculated the maximum theoretical stomatal conductance (gmax Franks amp Farquhar 2007 Sack amp Buckley 2016)

For the venation traits fixed leaves were cleared stained and scanned formajorveindensity(VLAmajor) and the top middle and bottom of each leaf were imaged under light microscope for measurements of minor andfreeendingveindensities(VLAminor and FEV) (Scoffoni et al 2011) Euphorbia multiformis var microphylla (EUPMUL Table 3) the single C4 spe-ciesinthestudy(YangMordenSporck-KoehlerSackampBerry2018b)wasremovedfromanalysesofacross-speciescorrelationsofveintraitswith vital rates C4 species are known to differ from C3 species in the rela-tionship of photosynthetic rate to vein density and thus would be expected to differ in their relationships of vital rates to vein traits (Ogle 2003)

26emsp|emspLeaf and wood economics and structure and leaf composition

Leaf structure and composition traits were measured in three leaves per studied individual Leaf saturated mass was measured using an analytical balance (001mg XS205 Mettler-Toledo OH USA) and

leaf thickness (LT) using digital callipers (001 mm Fowler Chicago IL USA)The leafarea (LA)wasmeasuredusinga flatbedscannerandanalysedusing thesoftware ImageJ (httpimagejnihgovij)Afterscanningleaveswereoven-driedat70degfor72hrandtheirdrymassandareaweremeasuredagainLeafmassperarea(LMA)wascalcu-lated as lamina dry mass divided by saturated area leaf density (LD) as LMAdividedbyLTsaturatedwatercontent(SWC)as(saturatedmassminusdrymass)dividedbydrymasswatermassperarea(WMA)asthe (saturated mass minus dry mass) divided by saturated area leaf dry matter content (LDMC) as dry mass divided by saturated mass and percentagelossinareaafterdrying(PLAdry) as the per cent decline in areafromsaturatedtodry leaves(OgburnampEdwards2012Peacuterez-Harguindeguyetal2013WitkowskiampLamont1991)

Wemeasuredwooddensity (WD) fromone5-cm-branch seg-mentofeachofthestudiedindividualsafterbarkremovalbywater-displacement(Peacuterez-Harguindeguyetal2013)

The concentration of leaf nitrogen phosphorus carbon per mass (Nmass Pmass and Cmass) and carbon isotope ratio (δ13C) were deter-mined using oven-dried leaves of three individuals per species bytheUniversityofHawaiiatHiloAnalyticalLaboratory facility (Fryetal1996Peacuterez-Harguindeguyetal2013)Nmass and Pmass were converted into Narea and Parea bymultiplyingbyLMAThe carbonisotope discrimination (Δleaf in parts per thousand permil) was calcu-latedfollowing(FarquharampRichards1984)Thechlorophyllconcen-tration per area (Chl)wasmeasuredusingaSPADmeter (MonjeampBugbee1992SPAD-502KonicaMinolta Japan)and thechloro-phyllconcentrationpermasswasdeterminedbydividingbyLMA

Turgor loss point (πtlp) was measured in three leaves per studied in-dividualWeusedavapour-pressureosmometer(Vapro5520WescorUSA)toobtaintheosmoticconcentration(πo) of the leaves and used calibration equations to estimate πtlp (Bartlett et al 2012a)

27emsp|emspEstimating photosynthetic traits

We estimated maximum rate of carboxylation per mass (Vcmaxmass) and electron transport rate (Jmaxmass) from leaf N and P

F I G U R E 1 emsp Contour map of the Pālamanui(LDF)andLaupāhoehoe(MWF)4-haplotsonHawaiiIsland


concentrations per mass (Domingues et al 2010) The ratio of in-tercellular CO2 concentration (ci) to ambient CO2 concentration (ca) was estimated from Δleaf(FarquharOrsquoLearyampBerry1982Franksetal 2014) Estimatesof leaf lifetime integratedCO2 assimilation rate ( Amass) and stomatal conductance to CO2 (gcleaf ) were derived from Vcmaxmass Jmaxmass and isotope composition data using the Farquhar von Caemmerer and Berry model (Franks Drake amp Beerling2009)ToconvertVcmaxmass Jmaxmass and Amass to an area basiswemultipliedthetraitvaluesbyLMAWealsocalculatedtheratio between gcleaf and gmax an index of the degree that stomata are open on average relative to their anatomical maximum aperture (McElwainYiotisampLawson2016)andtheratiobetweengmax and Narea a lower value would indicate that at full stomatal opening the species has more conservative water use for a given investment in photosynthetic machinery (Wright et al 2001)

28emsp|emspStatistical analyses

Differences in traits between MWF and LDF species were deter-minedusingnestedANOVAswithspeciesnestedwithinforesttypefollowedbyaTukeytestat5probabilitywhendifferencesweredetected (Sokal amp Rohlf 2012) Differences between forests in traits established as species means (RGRs m H Hmax and SM) were tested using t tests Traits that did not fulfil the normality and homoscedas-ticityassumptionswere log-transformedprior toanalysesTo testwhether trait variation differed between forests we (a) performed F tests to compare the variances in each trait (Minitab Release 17 StateCollegePAUSA)and(b)calculatedthecoefficientofvariation(CV ) for each trait in each forest as CVforest=

120590forest

xforesttimes100 and ap-

plied a paired t test across all traitsFunctional traits were grouped into six ldquomodulesrdquo according to

their contributions to given functions or ldquostrategiesrdquo The ldquostomatal morphologyrdquo module included traits such as d and s the ldquoleaf ve-nationrdquomoduleincludedtraitssuchasVLAminor and FEVs the ldquoleaf and wood economics and structurerdquo module included traits such as LMAandWDtheldquoleafcompositionrdquomoduleincludedleafnutrientconcentrations and |πtlp| the ldquoestimated photosynthesisrdquo module in-cluded traits such as Amass and Vcmax and the ldquoplant sizerdquo module included traits such as Hmax and SM (Table 2)

To investigate traitndashtrait and traitndashvital rate relationships within and across modules we calculated Pearsonrsquos correlations for untransformed and log-transformed data to test for eitherapproximately linear or nonlinear (ie approximate power-law)relationships respectively and the higher correlation value is re-ported in the text These analyses were applied to all species from both forests (Supporting Information Table S4 described in themain text) and to species of each forest separately (Supporting InformationTablesS5andS6)

We focus on frequentist statistical approaches following the bulk of previous studies on traitndashvital rate relationships However inthecaseofanalysingsize-dependentchangesintherelationshipsbetween vital rates (RGRdbh and m) and functional traits we utilized a hierarchical Bayesian approach following (Iida etal 2014) the

most sophisticated previous approach for resolving such an influ-ence Detailed description of parameters priors and MCMC settings areprovidedintheSupportingInformationMethods(seeAppendixS1) and model code is available on GitHub (httpsgithubcomcamilamedeirosMedeiros_et_al_2018)

RGRdbh for each individual ith tree of species j (RGRdbhij) was modelled as a linear function of the natural logarithm of the initial diameter DBH1ij based on two parameters estimated for spe-cies j (αkj k = 1 2) and given the input of the initial stem diameter (DBH1i) the final stem diameter (DBH2i) and the census interval of the ith tree (Δt i)

To estimate m for each individual ith tree belonging to species j (mij) we first calculated the probability of survival of the ith indi-vidual tree (pi) from observations of whether the tree survived the census period (Si = 1) or not (Si = 0) We assumed that Si followed a Bernoulli distribution of the probability of survival (pi)

The pi of the ith tree was calculated from the per capita annual mortality rate mij adjusted to the census interval (Δti) which was a functionofthreespecies-specificparametersβkj (k = 1 2 3)

Posteriors were estimated via Markov chain Monte Carlo imple-mentedinJAGS(JustAnotherGibbsSamplerPlummer2003)fromR using the package ldquoR2Jagsrdquo These analyses were carried out in-cluding all species from both forests

To analyse traitndashdemographic rate relationships for given plant size classes we first calculated RGRdbh and m using Equations 1 and 5respectivelybyusingtheposteriordistributionofspecies-specificparameters α1 and α2 for RGRdbh (Supporting Information Table S7) and β1j β2j and β3j for m (Supporting Information Table S7) and substi-tutingtheDBH1termforareferencediameterat1-cmDBHclasses(Iidaetal2014)When theDBHofa sizeclassexceededagivenspeciesrsquoactualmaximumDBH(calculatedasthe95thpercentileofthe speciesrsquo individuals in the plots) that species was dropped from the analysis in larger size classes We then calculated the Kendall correlation coefficient (τ) between the RGRdbh and m (calculated for eachspeciesineach1-cmDBHclass)andspeciesrsquomeanvaluesforfunctional traits We decided to use Kendall correlation following (Iida etal 2014) becauseof the typical non-normality of the sizeclassstratifiedvitalrates(Prado-Junioretal2016)ThemaximumDBH class included in our analysis was 10 cm because analysis of

(1)RGRdbhij=1j+2jtimes ln (DBH1i)

(2)ln (DBH2i)= ln (DBH1i)+RGRdbhijtimesΔti

(3)SisimBernoulli(pi)

(4)pi=exp (minusmijtimesΔti)

(5)ln (mij)=1j+2jtimes ln (DBH1i)+3jtimesDBH1i


correlations lost power with lower species numbers available to test at larger plant sizes (nlt9)Toreducetherateoffalsepositivediscoveries the correlations were considered significant only when 99oftheprobabilitydistribution(usedascredibleinterval)ofτ did notincludezeroratherthan95asinpreviousstudies(Iidaetal2014)

Finally to test the ability of traits to predict plant RGRdbh RGRbiom and m we built multiple regression models that included as independent variables functional traits and a term for forest membership (site coded as 0 for MWF species and 1 for LDF spe-cies) We selected seven traits to include in the models based on consideration of the 26 traits hypothesized a priori to mechanisti-cally influence RGRdbh RGRbiom and m To avoid collinearity we did not choose traits that were partially redundant that is correlated and calculated in part from the same measurements and involved within similar physiological processes and within the same trait category (egwe consideredLMAandnot leaf thickness giventhat LMA equals leaf thicknesstimesdensity Table2)We selectedthe trait most strongly correlated with vital rates from each trait module in Table 2 except for the ldquoLeaf and wood economics and structurerdquo module from which we selected one leaf- and onewood-relatedtrait

To compare model performance we included only species that had complete observations for all traits (final sample size = 16 spe-cies Supporting Information Table S8) To select the trait-basedmodels that best predicted RGRdbh RGRbiom and m we used for-ward backward and bidirectional procedures of variable selec-tion and compared models using Akaike criterion (AIC) using theldquostepAICrdquo function in the ldquoMASSrdquo package (Supporting Information Table S9) and calculated theAIC corrected for small sample sizes(AICc)(HastieampPregibon1992HurvichampTsai1989VenablesampRipley 2002) To find the percentage contribution of each variable to the prediction of RGRdbh RGRbiom and m we performed a hierar-chical partitioning analysis using the ldquohierpartrdquo package (Chevan amp Sutherland1991)

All statistical analyses andplotswereperformedusingR soft-ware (RCoreTeam2016)andpackagesavailable fromtheCRANplatform

3emsp |emspRESULTS

31emsp|emspVariation in vital rates and functional traits between forests types

On average mwas39higherinspeciesfromLDFthaninspeciesfromMWFAlthoughseveralMWFspecieshadhighergrowthratesthan those of LDF species means for RGRdbh and RGRbiom were sta-tistically similar in the MWF and LDF (Figure 2)

Traits varied strongly between and within forests On average across the measured traits 16 of the total variation was accounted for by forest type 73 by species differences within forests and 11 by individuals within species (nested ANOVAs SupportingInformation Table S1) The MWF showed stronger trait variation

than the LDF the variance was higher in the MWF for 20 traits in the LDF for six traits and not different between forests on the remaining 19traits(F tests Supporting Information Table S3) and on average acrossalltraitsthecoefficientofvariation(CV)was135plusmn08intheMWFand101plusmn06intheLDF(pairedt test p lt 0001)

Species fromMWF and LDF differed in 24 of the 45 func-tional traits (53) used to test hypotheses (Table2 SupportingInformation Tables S1 and S2 Figure 2) MWF species had higher values on average for stomatal index (i) and area (s) dimensions of guard cells (GCL GCW and SPL) and epidermal pavement cells (e) and had on average a 70 higher gmax (Figure 2 Table 2 Supporting Information Table S1) Additionally SWC and PLAdry were47-49higherintheMWFthanintheLDFspeciesand Amass gcleaf and cica ratio were 28ndash33 higher for the MWF than the LDF species (Figure 2 Table 2 Supporting Information Table S1) Pmass gmaxNarea and Hmaxwere4917and82higher in theMWFspecies than in the LDF species respectively (Table 2 Supporting Information Tables S1 and S2 Figure 2)

ConverselyspeciesfromtheLDFhadvalues46ndash70higheron average than species from the MWF for VLAmajor VLAminor VLAtotal and FEVs and values 22-42 higher on average forLDMC WD and NP (Figure 2 Table 2 Supporting Information Table S1) The LDF species also had a πtlp more negative by 06 MPa on average and 25 lower Δleaf than MWF species (Figure 2 Table 2 Supporting Information Table S1)

32emsp|emspAssociations among vital rates

Acrossforests thetwomeasuresof relativegrowthrates (RGRdbh and RGRbiom) were strongly intercorrelated (r=097 p lt 0001) and both were correlated with m (r=055 and 057 respectivelyplt005Figure3ab)WithintheLDFbutnottheMWFm was posi-tively correlated with RGRdbh and RGRbiom (r=076and093respec-tively plt005SupportingInformationTablesS5andS6)

When using the Bayesian approach to account for plant sizes we found positive correlations across species between m and both RGRdbh and RGRbiom in all size classes (τ gt 0 Figure 3cd)

33emsp|emspTraitndashtrait coordination

Traits were highly intercorrelated within functional modules (ie stomatal morphology traits venation traits leaf and wood econom-ics and structure traits and compositional traits) when considering species from both forests together and in the MWF and LDF sepa-rately (Supplementary Results ldquoTrait‐trait coordinationrdquo Supporting InformationTablesS4ndashS6)

34emsp|emspTrait relationships with plant vital rates

Overall eight traits were correlated with RGRdbh andor RGRbiom and seven were correlated with m (Supporting InformationTableS4)Ofthe26traitshypothesizedtocorre-late with vital rates three traits were correlated with RGRdbh


seven with RGRbiom and two with m across all 35 species(Supporting InformationTableS4)ThusRGRdbh and RGRbiom were positively correlated with d i and gmax (r ranged from 057 to 064 plt005 Supporting Information Tables S4and S8 Figure4ab) RGRbiom was negatively correlated with LMAandVLAminor (r=minus05andminus056respectivelyplt005Supporting Information Tables S5 and S9 Figure4de) andpositively correlated with Pmass and Amass (r=048and051re-spectively plt005SupportingInformationTablesS4andS8Figure4cf) andm was positively correlated with both Nmass

and Amass (r=05and061respectivelyplt005SupportingInformationTableS8Figure4gh)

Given that speciesrsquo RGRs did not differ between forests traitndashRGR correlations within forests were tested but not ex-plored (Table2Supporting InformationTablesS2S5andS6)However the forests differed in m and in its trait correlations In the MWF mwas positively correlatedwith LMA LDNarea Parea and Pmass and with photosynthetic traits on both mass and area basis Jmaxarea Jmaxmass Vcmaxarea Vcmaxmass

Aarea Amass and gcleaf (r ranged from 072 and 089 plt005 Supporting

F I G U R E 2 emsp Radar graph illustrating per cent difference in trait means between MWF and LDF species The LDF species means were fixed arbitrarily as the 100 reference values (the dark red dashed line) and the black line indicates the per cent difference between MWF species and LDF species Traits are arranged according to putative traits modules previously defined (Table 1) Bold and indicate plt005

Stomatal morphology

Whole plant size and

growth and mortality

Leaf

ve

natio

n

Estim

ate

d p

ho

tosyn

the

sis

Leaf

and

woo

d ec

onom

ics

and

stru

ctur

e

Leaf composition

0

50

100

150

200

250

disGCL

GCW

SPL

e

gmax

VLAmajor

VLAminor

VLAtotal

FEV

LA

LMA

LT

LD

LDMC

SWC

WMA

PLAdry

WDNarea

Nmass Parea PmassChlareaChlmass

Cmass

N P

Chlarea Nmass

leaf

tlp

Jmaxarea

Jmaxmass

Vcmaxarea

Vcmaxmass

c i ca

Aarea

Amass

gcleaf

gcleaf gmax

gmax Narea

RGRdbh

RGRbiom

mH

HmaxSM


InformationTableS5) In theLDFm was negatively correlated with LA LMA LT Jmaxarea Vcmaxarea Aarea and gcleaf (r ranged from minus076 and minus091 plt005 Supporting Information TableS6) Notably the direction of the correlation across species be-tween mandLMAdifferedbetweenforestsresultinginpositiverelationships between m and area-based photosynthetic traitsintheMWFandnegativerelationshipsintheLDF(Figure5ab)Further m was positively correlated with Pmass in the MWF (r=089plt001Supporting InformationTableS5Figure5c)andnegativelycorrelatedwithLAintheLDF(r=minus076plt005SupportingInformationTableS6Figure5d)

35emsp|emspFunctional traits and size‐dependent plant relative growth and mortality rates

Many more trait correlations with relative growth rate were re-solved when accounting for tree size using the Bayesian approach

Whereas three traits were correlated with RGRdbh without account-ing for size class when using the Bayesian approach to account for plant sizes 18 traits were correlated with RGRdbh within at least one size class Within given size classes RGRdbh was positively correlated with d i gmax LDMC LD Cmass cica gcleaf gmaxNarea (Figure 6a) Hmean and Hmax and negatively correlated with eSWCWMALA PLAdry gcleafgmax and SM (τ gt 0)

When accounting for plant size we found correlations of m with 18 traits In all size classes m was positively correlated with Nmass Jmaxmass Vcmaxmass and Amass (τ gt 0) and negatively correlated with LT (τ lt 0) Within given size classes m was positively correlated with d(Figure6c)VLAmajor Narea Pmass (Figure 6d) NP and gcleaf and neg-atively correlated with s GCL GCWWMACmass ChlareaNarea and gmaxNarea (τ gt 0)

Notably the finding of a greater number of significant rela-tionships between traits and vital rates when stratifying by tree size was not based on the (appropriate) use of different correlation

F I G U R E 3 emsp Relationships between relative growth rate (RGR) and mortality rate (m) across species of Hawaiian wet and dry forest The top panels show the relationships across species between mean values for m and (a) relative growth rate in terms of diameter at breast height RGRdbhand(b)intermsofabove-groundbiomassRGRbiom The bottom panels show that the correlation of mortality with RGR is robust across size modules by plotting the Kendall correlation coefficient (τ) between m and (c) RGRdbh and (d) RGRbiom against plantsizeclasswiththegreylineshowingthenumberofspeciesineach1-cmdiameterclass(lowerinlargersizeclasses)plt005p lt 001 p lt 0001 Top row black symbols Montane Wet Forest (MWF) species grey symbols Lowland Dry Forest (LDF) RGRdbh = 001 + 0001 m RGRbiom = 003 m035 Bottom row Filled symbols represent significant correlations We use Pearson correlation coefficient in plots (a) and (b) because the species means for m RGRdbh and RGRbiom calculated across all individuals were normally distributedorbecamesoafterlog-transformationwhereasweusedKendallscorrelationcoefficientinplots(c)and(d)becauseafterstratifying by plant size mremainednon-normallydistributedevenaftertransformationNotablytheRGR‐m relationships can be discerned with either coefficient when calculating Kendalls coefficient for panels (a) and (b) Kendalls τ was 032 (p=007)and035(p=0048)respectivelyforpanels(c)and(d)correlationswereconsideredsignificantwhenthe99credibleintervalofτ did not include zero

0 5 10 15

m (year)

000

002

004

RG

Rdb

h( c

mcm

minus1ye

arminus1

)

(a)

r = 0550 5 10 15

m (year)

000

005

010

RG

Rbi

om(k

gkg

minus1ye

arminus1

)

(b)

r = 057

2 4 6 8 10Diameter (cm)

minus10

minus05

00

05

10

τ τ(mvs

RG

Rdb

h)

(c)


minus10

minus05

00

05

10

(mvs

R

GR

biom

)

1015

2025

Spe

cies

(d)


F I G U R E 4 emsp Traitndashvital rate relationships across Hawaiian wet and dry forest species including relationships between relative growth rate in terms of diameter at breast height (RGRdbh) and (a) stomatal density and (b) maximum stomatal conductance between relative growth rateintermsofabove-groundbiomass(RGRbiom) and (c) time integrated CO2 assimilation rate per mass (d) leaf mass per area (e) minor vein density and (f) phosphorus per mass and between mortality rate (m) and (g) time integrated CO2 assimilation rate per leaf dry mass (h) nitrogen per leaf dry mass Black symbols Montane Wet Forest (MWF) species grey symbols Lowland Dry Forest (LDF) species RGRdbh = 302eminus03+551eminus05 d RGRdbh=001+0004gmax RGRbiom = 10eminus03 A135

mass RGRbiom=253LMA

minus0858 RGRbiom = 017 VLAminus087

minor

RGRbiom=0041P097

mass m = 8eminus05 A240

mass m=004N167

mass plt005p lt 001

p lt 0001

100 300 500 700

d (stomatamm2)

000

002

004

RG

Rdb

h (c

m c

mminus1

year

minus1) (a)

r = 0640 2 4 6 8

gmax (mol mminus2sminus1)

(b)

r = 057

40 60 80 100 120

Amass (nmol gminus1sminus1)

000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (c) r = 051

50 100 150 200 250 300

LMA (gm2)

(d) r = minus 050

2 4 6 8 10 12

VLAminor (mm mmndash2)

000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (e) r = minus 056

05 10 15 20

Pmass (mgg)

(f) r = 048

40 60 80 100 120


05

1015

m (

yea

r)

(g) r = 061

5 10 15 20 25 30

Nmass (mg g)

(h) r = 050


methods selected according to the distribution of the data that is the Pearson r for the analyses of traitndashvital rate correlations when averaging across all individuals for each species and the Kendall tau when testing these correlations while stratifying by plant size (see Section 2) To test this we also determined the traitndashvital rate correlations using Kendall tau when averaging across all individu-als for each species and as for the Pearson test seven traits were correlated with RGRdbh andor RGRbiom and seven were correlated with m Thus the finding that more traitndashvital rate relationships are significant when stratifying by plant size is robust to the use of different correlation tests

36emsp|emspPredicting RGRdbh RGRbiom and m from functional traits

To predict RGRdbh RGRbiom and m we built multiple regression modelsthatincludedthesevennon-redundanttraitsmoststronglycorrelated with vital rates among the 26 hypothesized a priori to in-fluence vital rates (dVLAminorLMAWDNmass Pmass and Amass) and a term for forest membership (site coded as 0 for MWF species and 1 for LDF species) The variable selection procedures (Supporting Information Table S9) indicated that d VLAminor Pmass and Amass were the best predictors for RGRdbh (adjusted R2 = 072 p lt 0001

Table4Figure7a)dVLAminorLMAandPmass for RGRbiom (adjusted R2 = 070 plt001 Table3 Figure7b) and VLAminor LMA Pmass Amass and site for m (adjusted R2 = 071 plt0001Table4Figure7c)

4emsp |emspDISCUSSION

41emsp|emspTrait variation between Hawaiian wet and dry forests

We found strong novel trait variation between Hawaiian wet and dry forests demonstrating that these forests are highly distinct not only in climate and species composition but also in an extensive set of traits While previous studies have shown that wet and dry forestsdiffer in functional traits (Brenes-ArguedasRoddyKursaramp Tjoelker 2013 Lohbeck etal 2015 Markesteijn etal 2010SantiagoKitajimaWrightampMulkey2004Wrightetal2004)byincluding a far wider range of traits related to resource acquisitive-ness and stress tolerance our analyses highlight their power in mul-tiplecomparativeandpredictiveapplicationsoftrait-basedecology

The trait differences between forests aligned with their varia-tion in vital rates While the species of the two forests did not dif-fer on average in RGR the MWF species showed lower mortality

F I G U R E 5 emsp Contrasting relationships between mortality rate and functional traits across forests including (a) leaf mass per area (b) time integrated CO2 assimilation rate per leaf area (c) phosphorus concentration per leaf mass and (d) individual leaf area Black symbols and curve Montane Wet Forest (MWF) species grey symbols and curve Lowland Dry Forest (LDF) species In (e) the black and grey lines and r values represent the fit and Pearsons regression coefficients including only MWF species and LDF species respectively mMWF=5e

minus04LMA189 and mLDF=4e8

LMAminus374 mMWF=minus072+044Aarea and

mLDF=764699Aminus543area m = 173 P238

mass

m=7453LAminus146 plt005p lt 001 p lt 0001

50 100 200 300

LMA (gm2)

05

1015

m (

y

ear)

(a)

r = 085r = minus 091

0 5 10 15 20

Aarea (μmol mminus2 sminus1)

(b)

r = 085r = minus 085

05 10 15 20

Pmass (mgg)

05

1015

m (

y

ear)

(c)

r = 089r = ns

0 50 100 150

LA (cm2)

(d)

r = nsr = minus 076


rates than the LDF species consistent with previous work show-ing higher mortality in drier forests elsewhere (Gaviria Turner amp Engelbrecht 2017 Laura Suarez amp Kitzberger 2010) The lower mortality of the MWF species is consistent with the greater supply of water and soil nutrients related to greater accumulated weath-eringorganicmaterialformationN-fixationandnutrientretentioncapacity and its richer microbial community The positive relation-ship of RGRs and m across all species was consistent with that found across species in temperate (Iida etal 2014 Seiwa 2007) andtropicalforests(Kitajima1994Philipsonetal2014Wrightetal2010) Our finding of greater trait variation within the wet forest than the dry forest supports the expectations from first principles that the low-resource availability in the dry forestwould act as astrong environmental filter resulting in functional convergence andor promote greater niche overlap among species in the dry forest via fewer potential biotic interactions (Kraft Crutsinger Forrestel amp Emery2014Lebrija-Trejosetal2010Nathanetal2016WeiherampKeddy1995)

The greater soil resources in the MWF led to the expectation that species would possess traits associated with photosynthetic productivity and rapid growth Consistent with this expectation MWF species had higher values on average for i and s dimensions of guard cells (GCL GCW and SPL) and e gmax SWCPLAdry Pmass ChlareaNarea Amass gcleaf cica gmaxNarea Δleaf and Hmax and lower values for LDMC WD and NP By contrast the higher tempera-ture and lower rainfall of the LDF led to the expectation that spe-cies would possess drought tolerance traits Indeed LDF species had higher vein densities |πtlp| WD and LDMC and lower values for PLAdry stomatal dimensions SWC and cica ratio and Amass Finally

the greater understorey shade of the MWF led to expectations of shade adaptation confirmed for the lower values for vein densi-ties and LDMC (Baltzer etal 2008 Chave etal 2009 Farquharetal 1989 Li etal 2015aNiinemets 2001 StrattonGoldsteinampMeinzer 2000Wright etal 2004) Beyond these average dif-ferences among forests trait values were consistent with known life-historydifferencesamongspecieswithinandacrossforestsForexample Acacia koa the fastest growing species overall had notably high values for stomatal dimensions and index and estimated rates ofelectrontransportandgasexchangedrought-tolerantOsteomeles anthyllidifolia had high |πtlp| and WD and low cica ratio and Amass and shade-tolerantHedyotis hillebrandii had high values for stomatal di-mensionsandLAandlowveindensitiesandWD

42emsp|emspTrait correlations across species of wet and dry forests

Our work supported the hypothesis that traits would be intercor-related within modules corresponding to a given organ or function (Li etal 2015b Sack etal 2003a) These trait associations canindicate allometric relationships that arise developmentally such as those found among stomatal traits vein densities and leaf size (Brodribb Field amp Sack 2010 Sack et al 2012) Other traitndashtrait re-lationshipswithinmoduleswouldarisefromco-selectionforoptimalfunction for example traits potentially contributing to maximum gas exchange and RGR (Scoffoni et al 2016) such as high gmax and Pmass or to drought tolerance (Bartlett Klein Jansen Choat amp Sack 2016) such as high |πtlp| and Amass or to shade tolerance (Givnish etal2005)suchashighLAandlowWD

F I G U R E 6 emsp Estimating the influence of plant size on the correlation of relative growth rate and mortality with given functional traits Each panel shows theplotofthesize-dependentKendallcorrelation coefficient (τ) between (a) relative growth rate and the ratio of maximum stomatal conductance to leaf nitrogen per area gmaxNarea (b) relative growth rate and maximum height Hmax (c) mortality rate and stomatal density d and (d) mortality rate and phosphorus concentration Pmass Open symbols representnon-significantassociations(the99credibleintervalofτ included zero) and filled symbols significant correlations (the99credibleintervalofτ did not include zero) The grey line shows the numberofspeciesineach1-cmdiameterclass

10

00

05

10

(RG

Rdb

hvs

gm

axN

area) (a)

10

00

05

10

(RG

Rdb

hvs

Hm

ax)

1015

2025

Spe

cies

(b)


10

00

05

10

(mvs

d)

(c)


10

00

05

10

(mvs

Pm

ass)

1015

2025

Spe

cies

(d)


The numerous trait correlations across species result in a re-duced trait ldquodimensionalityrdquo by which most trait variation may be captured by few axes (Diaz et al 2016) However that finding does not in fact imply that traits are functionally redundant as correlated traitscancontributesemi-distinctlytofunctionandtheirconsider-ation as separate parameters improves predictive and mechanistic modelling (John et al 2017 Sterck et al 2011) For example while LMA is correlated with other traits that share structural or com-positional bases (Finegan etal 2015 John etal 2017) such asLDMCorWMAphotosyntheticratesandnutrientconcentrationsthesetraitscanplaynon-redundantrolesindeterminingfunctionssuch as shade and drought tolerance and in influencing RGR and m (SupportingInformationTablesS4ndashS6andS10)

43emsp|emspTrait associations with relative growth rates and mortality rates

Several novel trait correlations were found with mean RGRs and m across species that were expected from theory and that have poten-tial for generality including the relationships of RGRdbh RGRbiom andor m to Amass and d and several relationships were confirmed such as with Hmax LMAandWD thatwere reported in previous stud-ies of temperate (Iida et al 2016) andor tropical forests (Finegan

etal2015Heacuteraultetal2011Liuetal2016Wrightetal2010)The contrasting correlations of traits with m between the MWF andLDFsuchasLMAand Aarea (Figure5ab)and thecorrelationsof traits with m in one but not the other forest such as for Pmass and LA(Figure5cd)highlightthecontextdependenceoftraitndashvitalraterelationships IntheMWFahighLMAwasassociatedwithhighermasexpectedgivenistrepresentingthemoreshade-tolerantspe-cies in the understorey which tend to have higher mortality (Kobe ampCoates1997Lusketal2008)ConverselyintheLDFhighLMAwas related to lower m as expected given its potential contribution to greater drought tolerance via a lower surface area volume ratio andor a greater mechanical protection contributing to longer leaf lifespanandreducedrespirationcosts(Falcatildeoetal2015WrightWestobyampReich2002Wrightetal2004)

Hawaiian forests also showed contrasting relationships of certain traits to vital rates than previously reported For exam-ple vein density contributes mechanistically to greater hydraulic conductance photosynthetic productivity and RGR across diverse speciesallelsebeingequal(BrodribbampMcAdam2017Iidaetal2016Lietal2015aSackampFrole2006SackampScoffoni2013Sack et al 2013 Scoffoni et al 2016) However RGR was nega-tively related to vein density across the species of both forests This negative correlation may reflect the co-variation of vein

Model

Pearsons correlation coefficient

Multiple regression analyses coefficient estimate

Hierarchical partition analyses ()

(A)RGRdbh ~ d+VLAminor + Pmass + Amass

Intercept ndash 209eminus02 ndash

d 067043 776eminus05 532

VLAminor minus044minus049 minus332eminus03 276

Pmass 039025 minus243eminus02 109Amass 036 028 251eminus04 83

AdjustedmultipleR2 ndash 072 ndash

(B) RGRbiom ~ d+VLAminor+LMA+Pmass


d 065040 201eminus04 519


LMA minus035minus048 minus232eminus04 137

Pmass 039025 minus230eminus02 144


(C) m~VLAminor+LMA+Pmass + Amass + site

Intercept ndash minus259 ndashAmass 049073 018 391

Site ndash 665 312

LMA minus035minus038 minus003 159

Pmass 016042 minus339 77

VLAminor minus013minus017 minus060 61


plt005p lt 001 p lt 0001

TA B L E 4 emsp Models selected by maximum likelihood to estimate relative growth rate intermsofdiameteratbreastheight(ARGRdbh)orabove-groundbiomass(BRGRbiom) or mortality rate (C m) Independent variables included in the tested models were those of each module (Supporting Information Table S8) that were most correlated with each dependent variable We present the Pearsons coefficients for the relationships of each predicted variable vs each independent variable using untransformed andlog-transformeddatathemultipleregression coefficient estimates and per cent contribution of each trait to model fit Full models and detailed model selectionproceduresusingAICcsarepresented in Supporting Information Table S9


density with other traits negatively related to RGR including traits not considered such as root traits andor it may arise from the high values for LDF species which is consistent with their adapta-tion to higher RGR in the more limited periods when water is avail-able though this high RGR is not achieved integrated over time (Sack amp Scoffoni 2013)

Our study also confirmed the hypothesis that stratifying by plant size improved the frequency of correlations of vital rates with given traits Stratifying by size has previously been shown to improve reso-lution of correlations of RGR and m with traits such as vein densities LALMASWCLTNmass and Pmass WD and Hmax (Iidaetal20142016Prado-Junioretal2016)andourstudyexpandedthisfindingto a wider range of traits Stratifying by size reduces the confounding influenceofontogeneticshiftsinvitalratesoncross-speciescompar-isons (Heacuterault et al 2011) Notably when we examined trait correla-tions with RGR and m for plants of given sizes as in previous studies conducting this analysis our trait values were only for the sampled trees of typical mature size Future studies may further improve reso-lution of correlations by also considering ontogenetic variation in trait values

44emsp|emspTrait‐based predictions of vital rates

Our study showed the value of a broad suite of functional traits for predicting vital rates Models based on seven selected traits could explain more than 70 of the variation in RGRdbh RGRbiom and m (Table4 Supporting InformationTableS9Figure7)Themostpar-simonious models for all three vital rates retained minor vein density and P per mass and two of them included stomatal density time in-tegrated CO2assimilationrateandLMAThesefindingshighlightthepotential of an approach based on an extensive suite of functional traits and the continued need to refine our mechanistic understand-ing of how suites of traits drive processes at the scale of individuals and whole forests

45emsp|emspConclusions and limitations of the study

We conclude that the use of an extensive suite of functional traits contributes power to (a) discover and resolve variation across species expected from their contrasting adaptation (b) compare functional convergence across ecosystems (c) highlight novel traitndashtrait and (d) traitndashvital rate associations and (e) the mediating role of plant size in these associations and (f) to predict RGR and m across species Recent studies have applied trait data to mechanistic process mod-els to predict forest vital rates niche differentiation and productivity (Fyllasetal2014MarksampLeichowicz2006Stercketal2011)We propose that including an extensive suite of traits in such models will be a powerful avenue for future research on the functional ecol-ogy of contrasting communities including vital rates and ultimately their responses to climate change and shifts in speciesrsquo distributions Animportantavenueforfutureresearchistoconsidertheincorpo-ration of extended traits into estimating and testing speciesrsquo habitat preferences within and across forests extending from recent work showing substantial power even based on few traits such as leaf size wooddensityLMAandseedsize(Shipleyetal2017)

We note that some of our study questions were carried out by comparing single forests of each type and our findings suggest that the approach has value for further testing replicate forests of each typeAdditionallymodelsareneededofthespecificprocessesin-volved in vital rates in which traits can be included along with cli-mate to resolve how specific trait variation scales up to influencing RGR and m Our approach focused on the correlations of single traits and suites of traits with RGR and macentralapproachintrait-basedecology However given that upper level processes such as growth or species niche preferences depend on multiple traits given that correlations may not actually reflect causal mechanisms due to pat-ternsofco-variationwithothertraits(Johnetal2017Shipleyetal2017) Further while our models predicting vital rates included site as a factor that approach does not fully incorporate traitndashclimate

F I G U R E 7 emsp Relationship between observed growth rate in terms of diameter at breast height (RGRdbh)above-groundbiomass(RGRbiom) and mortality rate (m) and the values predicted from models using the plant traits most correlated with each dependent variable (a) RGRdbh=209e

minus02 + (776eminus05 d) ndash (332eminus03VLAminor)ndash(243eminus02 Pmass)+(251e

minus04 Amass) (b) RGRbiom=658eminus02 + (201eminus04 d) ndash

(478eminus03VLAminor) ndash (232eminus04LMA)ndash(23eminus02 Pmass) (c) m=minus259ndash(060VLAminor)ndash(003LMA)ndash(339Pmass) + (018 Amass) + (664site)Thedashedlinerepresentsthe11relationshipplt005p lt 001 p lt 0001

000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070


interactionssuggestingthevalueofmechanistictrait-basedmodelsthat include climatic factors

Including an extensive suite of functional traits can sharpen our char-acterization of species adaptation to their ecosystem and climatic pref-erences as well as predicting vital rates Including traits in mechanistic process models for growth and speciesrsquo distributions will increase pre-dictive power further Such prediction is increasingly critical for species conservation especially in ecosystems such as Hawaiian forests which are threatened in the face of development and ongoing climate change (Fortini et al 2013) Future work should also consider intraspecific varia-tion in the wider set of traits and its role in shaping species distributions within and between forests as well as trait determination of microsite differencesamongspecies(Inman-Naraharietal2014)Giventhepowerto predict vital rates this work can enable scaling up from the traits of component species to ecosystem and eventually global vegetation pro-cesses highlighting the enormous promise of increasing mechanistic informationmdashfrom measurements to analyses to modelsmdashfor clarifying and predicting processes in species and community ecology

ACKNOWLEDG EMENTS

Permits were obtained for work in the HETF through the Institute of Pacific Islands Forestry and the Hawairsquoi Division of Forestry and Wildlife Department of Land and Natural Resources This work was supported by the National Science Foundation the Brazilian National Research Council (CNPq) and Brazilian Science Without Borders Program (grant number2028132014-2)WethankAdamSibleyBrittneyChauNishaChoothakan Tiffany Dang Chirag Govardhan Jonnby Laguardia Jeffrey Lee Tram Nguyen Sara Rashidi Erin Solis and Dustin Wong for labora-tory and field assistance and Nathan Kraft and Marcel Vaz for discussion

AUTHORSrsquo CONTRIBUTIONS

CDM and LS conceived ideas and experimental design CDM CS FI GJ MB and LS collected trait data FI RO SC CG and LS organized and collected forest census data CDM and LS analysed the data and wrote the first draft of the manuscript and all authors contributed substantially to revisions

DATA ACCE SSIBILIT Y

All trait data collected for this paper are archived in theDryadDigitalRepositoryhttpsdoiorg105061dryadcq47n7s(Medeirosetal2018)

ORCID

Camila D Medeiros httpsorcidorg0000-0002-5822-5603

Grace P John httpsorcidorg0000-0002-8045-5982

Megan K Bartlett httpsorcidorg0000-0003-0975-8777

Susan Cordell httpsorcidorg0000-0003-4840-8921

Christian Giardina httpsorcidorg0000-0002-3431-5073

Lawren Sack httpsorcidorg0000-0002-7009-7202

R E FE R E N C E S

Adler P B Salguero-Goacutemez R Compagnonia A Hsud J S Ray-Mukherjeee J Mbeau-Ache C amp Franco M (2014) Functionaltraits explain variation in plant life history strategies Proceedings of the National Academy of Sciences of the United States of America 111 10019httpsdoiorg101073pnas1315179111

Armbruster W S Pelabon C Bolstad G H amp Hansen T F(2014) Integrated phenotypes Understanding trait covari-ation in plants and animals Philosophical Transactions of the Royal Society B Biological Sciences 369 20130245httpsdoiorg101098rstb20130245

AsefaMCaoMZhangGCiXLiJampYangJ(2017)Environmentalfiltering structures tree functional traits combination and lineages across space in tropical tree assemblages Scientific Reports 7 1ndash10 httpsdoiorg101038s41598-017-00166-z

Baltzer J L Davies S J Bunyavejchewin S amp Noor N S M (2008) The role of desiccation tolerance in determining tree species distri-butionsalongtheMalay-ThaiPeninsulaFunctional Ecology 22 221ndash231httpsdoiorg101111j1365-2435200701374x

Bartlett M K Klein T Jansen S Choat B amp Sack L (2016) The cor-relations and sequence of plant stomatal hydraulic and wilting re-sponses to drought Proceedings of the National Academy of Sciences of the United States of America 113 13098ndash13103 httpsdoiorg101073pnas1604088113

Bartlett M K Scoffoni C Ardy R Zhang Y Sun S Cao K ampSack L (2012a) Rapid determination of comparative drought tolerance traits Using an osmometer to predict turgor loss point Methods in Ecology and Evolution 3 880ndash888 httpsdoiorg101111j2041-210X201200230x

Bartlett M K Scoffoni C amp Sack L (2012b) The determinants of leaf turgor loss point and prediction of drought tolerance of species and biomesAglobalmeta-analysisEcology Letters 15393ndash405httpsdoiorg101111j1461-0248201201751x

Beaulieu J M Leitch I J Patel S Pendharkar A amp Knight C A(2008) Genome size is a strong predictor of cell size and stomatal density in angiosperms New Phytologist 179975ndash986httpsdoiorg101111j1469-8137200802528x

Blackman C J Brodribb T J amp Jordan G J (2012) Leaf hydraulic vul-nerability influences speciesrsquo bioclimatic limits in a diverse group of woody angiosperms Oecologia 168 1ndash10 httpsdoiorg101007s00442-011-2064-3

Brenes-ArguedasTRoddyABKursarTAampTjoelkerM (2013)Plant traits in relation to the performance and distribution of woody species in wet and dry tropical forest types in Panama Functional Ecology 27392ndash402httpsdoiorg1011111365-243512036

Brodribb T J Feild T S amp Jordan G J (2007) Leaf maximum photo-synthetic rate and venation are linked by hydraulics Plant Physiology 1441890ndash1898httpsdoiorg101104pp107101352

Brodribb T J Field T S amp Sack L (2010) Viewing leaf structure and evolution from a hydraulic perspective Functional Plant Biology 37 488ndash498httpsdoiorg101071FP10010

Brodribb T JampMcAdam SAM (2017) Evolution of the stomatalregulation of plant water content Plant Physiology 174 639ndash649httpsdoiorg101104pp1700078

ChatuverdiRKRaghubanshiASampSinghJS(2011)Leafattributesand tree growth in a tropical dry forest Journal of Vegetation Science 22917ndash931httpsdoiorg101111j1654-1103201101299x

ChaveJAndaloCBrownSCairnsMAChambersJQEamusDampYamakuraT(2005)Treeallometryandimprovedestimationofcarbon stocks and balance in tropical forests Oecologia 14587ndash99httpsdoiorg101007s00442-005-0100-x

ChaveJCoomesDJansenSLewisSLSwensonNGampZanneAE(2009)TowardsaworldwidewoodeconomicsspectrumEcology Letters 12351ndash366httpsdoiorg101111j1461-0248200901285x


Chesson P (2000) Mechanisms of maintenance of species diversity Annual Review in Ecology and Systematics 31343ndash366httpsdoiorg101146annurevecolsys311343

Chevan A amp Sutherland M (1991) Hierarchical partitioning The American Statistician 45 90ndash96 httpsdoiorg10108000031305199110475776

ConditR (1998)Tropical forest census plots (1st ed) Berlin Germany SpringerPublisherhttpsdoiorg101007978-3-662-03664-8

CornwellWKSchwilkDWampAckerlyDD(2006)Atrait-basedtestforhabitat filtering Convex hull volume Ecology 871465ndash1471httpsdoiorg1018900012-9658(2006)87[1465ATTFHF]20CO2

Diaz S Kattge J Cornelissen J H Wright I J Lavorel S Dray S amp Gorne L D (2016) The global spectrum of plant form and function Nature 529167ndash171httpsdoiorg101038nature16489

Domingues T F Meir P Feldpausch T R Saiz G Veenendaal E M Schrodt F amp Lloyd J O N (2010) Co-limitation of photo-synthetic capacity by nitrogen and phosphorus in West Africawoodlands Plant Cell amp Environment 33 959ndash980 httpsdoiorg101111j1365-3040201002119x

DonovanLAampEhleringerJR(1994)Carbonisotopediscriminationwater-useefficiencygrowthandmortalityinanaturalshrubpopula-tion Oecologia 100347ndash354httpsdoiorg101007BF00316964

ElserJJFaganWFDennoRFDobberfuhlDRFolarinAHubertyAamp Sterner R W (2000) Nutritional constraints in terrestrial and freshwa-ter foodwebs Nature 408578ndash580httpsdoiorg10103835046058

Evans G C (1973) The quantitative analysis of plant growth (1st ed) BerkeleyandLosAngelesCAUniversityofCaliforniaPress

Evans J R (2013) Improving photosynthesis Plant Physiology 162 1780ndash1793httpsdoiorg101104pp113219006

FalcatildeoHMMedeirosCDSilvaBLRSampaioEVSBAlmeida-CortezJSampSantosMG(2015)Phenotypicplasticityandecophys-iologicalstrategies inatropicaldryforestchronosequenceAstudycase with Poincianella pyramidalis Forest Ecology and Management 34062ndash69httpsdoiorg101016jforeco201412029

FarquharGDEhleringerJRampHubickKT(1989)Carbonisotopediscrimination and photosynthesis Annual Review in Plant Physiology and Plant Molecular Biology 40 503ndash537 httpsdoiorg101146annurevpp40060189002443

FarquharGDOLearyMHampBerryJA(1982)Ontherelationshipbetween carbon isotope discrimination and intercellular carbon di-oxide concentration in leaves Australian Journal of Plant Physiology 9 121ndash137httpsdoiorg101071PP9820121

FarquharGDampRichardsRA(1984)Isotopiccompositionofplantcar-boncorrelateswithwater-useefficiencyofwheatgenotypesFunctional Plant Biology 11539ndash552httpsdoiorg101071PP9840539

FineganBPentildea-ClarosMdeOliveiraAAscarrunzNBret-HarteMSCarrentildeo-RocabadoGampPoorterL (2015)Doesfunctionaltrait diversity predict above-ground biomass and productivity oftropical forests Testing three alternative hypotheses Journal of Ecology 103191ndash201httpsdoiorg1011111365-274512346

Fiske I J Bruna E M amp Bolker B M (2008) Effects of sample size on estimates of population growth rates calculated with matrix models PLoS One 3 e3080 httpsdoiorg101371journalpone0003080

FortiniLPriceJJacobiJVorsinoABurgettJBrinckKampPaxtonE (2013) A landscape‐based assessment of climate change vulnerability for all native Hawaiian plant Hilo HI University of Hawaii Publisher

FranksPJampBeerlingDJ(2009)MaximumleafconductancedrivenbyCO2 effects on stomatal size and density over geologic time Proceedings of the National Academy of Sciences of the United States of America 106 10343ndash10347httpsdoiorg101073pnas0904209106

Franks P J Drake P L amp Beerling D J (2009) Plasticity in maxi-mum stomatal conductance constrained by negative correlation between stomatal size and density An analysis using Eucalyptusglobulus Plant Cell amp Environment 32 1737ndash1748 httpsdoiorg101111j1365-30402009002031x

Franks P J amp Farquhar G D (2007) The mechanical diversity of sto-mataand its significance ingas-exchangecontrolPlant Physiology 14378ndash87httpsdoiorg101104pp106089367

Franks P J Royer D L Beerling D J Van de Water P K Cantrill D JBarbourMMampBerryJA (2014)Newconstraintsonatmo-spheric CO2 concentration for the Phanerozoic Geophysical Research Letters 414685ndash4694httpsdoiorg1010022014GL060457

Fry B Ganitt R Tholke K Neill C Michener R H Mersch F J amp Brand W (1996) Cryoflow Cryofocusing nano-mole amounts of CO2 N2 and SO2 from an elemental an-alyzer for stable isotopic analysis Rapid Communications in Mass Spectrometry 10 953ndash958 httpsdoiorg101002(SICI)1097-0231(19960610)108lt953AID-RCM534gt30CO2-0

FyllasNMGloorEMercadoLMSitchSQuesadaCADominguesTFampLloydJ(2014)AnalysingAmazonianforestproductivityusinganewindividualandtrait-basedmodel(TFSvol1)Geoscientific Model Development 71251ndash1269httpsdoiorg105194gmd-7-1251-2014

Gaviria J Turner B L amp Engelbrecht B M J (2017) Drivers of tree species distribution across a tropical rainfall gradient Ecosphere 8 e01712 httpsdoiorg101002ecs21712

Gibert A Gray E F Westoby M Wright I J Falster D S ampWilson S (2016) On the link between functional traits and growth rate meta-analysis shows effects change with plant sizeas predicted Journal of Ecology 104 1488ndash1503 httpsdoiorg1011111365-274512594

Gil-PelegriacutenEPeguero-PinaJJampSancho-KnapikD(2017)Oaks physio‐logical ecology Exploring the functional diversity of genus Quercus L (1st ed) NewYorkNYSpringerhttpsdoiorg101007978-3-319-69099-5

GivnishT J (1988)Adaptation to sunand shadeAwhole-plantper-spective Australian Journal of Plant Physiology 1563ndash92httpsdoiorg101071PP9880063

GivnishTJPiresJCGrahamSWMcPhersonMAPrinceLMPattersonTBampSytsmaKJ(2005)Repeatedevolutionofnetve-nation and fleshy fruits among monocots in shaded habitats confirms a priori predictions Evidence from an ndhF phylogeny Proceedings of the Royal Society B Biological Sciences 2721481ndash1490httpsdoiorg101098rspb20053067

Gleason S M Blackman C J Chang Y Cook A M Laws C A ampWestoby M (2016) Weak coordination among petiole leaf vein and gas-exchange traits acrossAustralian angiosperm species and itspossible implications Ecology and Evolution 6 267ndash278 httpsdoiorg101002ece31860

GreenwoodSRuiz-BenitoPMartiacutenez-VilaltaJLloretFKitzbergerTAllenCDampChave J (2017)Treemortalityacrossbiomes ispro-moted by drought intensity lower wood density and higher specific leaf area Ecology Letters 20539ndash553httpsdoiorg101111ele12748

Grime J P (2001) Plant strategies vegetation processes and ecosystem properties (2nd ed) West Sussex England John Wiley amp Sons Ltd

Gross K L (1984) Effects of seed size and growth form on seedlingestablishment of six monocarpic perennial plants Journal of Ecology 72369ndash387httpsdoiorg1023072260053

GrubbPJ(1998)Areassessmentofthestrategiesofplantswhichcopewith shortages of resources Perspectives in Plant Ecology Evolution and Systematics 13ndash31httpsdoiorg1010781433-8319-00049

Hacke U G Sperry J S Pockman W T Davis S D amp McCulloh K A (2001)Trends inwooddensityandstructureare linked topre-vention of xylem implosion by negative pressure Oecologia 126 457ndash461httpsdoiorg101007s004420100628

HastieT JampPregibonD (1992)Generalized linearmodels InJMChambers amp T J Hastie (Eds) Statistical models in S (pp195ndash246)PacificGroveCAWadsworthampBrooksCole

Heacuterault B Bachelot B Poorter L Rossi V Bongers F Chave J amp Baraloto C (2011) Functional traits shape ontogenetic growth trajectories of rain forest tree species Journal of Ecology 991431ndash1440httpsdoiorg101111j1365-2745201101883x


HetheringtonAMampWoodwardF I (2003)Theroleofstomata insensing and driving environmental change Nature 424 901ndash907httpsdoiorg101038nature01843

HurvichCMampTsaiC-L (1989)Regression and time seriesmodelselection in small samples Biometrika 76 297ndash307 httpsdoiorg101093biomet762297

Iida Y Poorter L Sterk F Kassim A R Potts M D Kubo T ampKohyamaTS(2014)Linkingsize-dependentgrowthandmortalitywith architectural traits across145co-occurring tropical tree spe-cies Ecology 95353ndash363httpsdoiorg10189011-21731

IidaYSun IFPriceCAChenC-TChenZ-SChiangJ-MampSwenson N G (2016) Linking leaf veins to growth and mortality ratesAnexample froma subtropical tree communityEcology and Evolution 66085ndash6096httpsdoiorg101002ece32311

Inman-NarahariFOstertagRAsnerGPCordellSHubbellSPampSackL(2014)Trade-offsinseedlinggrowthandsurvivalwithinand across tropical forest microhabitats Ecology and Evolution 4 3755ndash3767httpsdoiorg101002ece31196

John G P Scoffoni C Buckley T N Villar R Poorter H amp Sack L (2017) The anatomical and compositional basis of leaf mass per area Ecology Letters 20412ndash425httpsdoiorg101111ele12739

KhuranaEampSinghJS (2004)Germinationandseedlinggrowthoffive tree species from tropical dry forest in relation to water stress Impact of seed size Journal of Tropical Ecology 20385ndash396httpsdoiorg101017S026646740400135X

KingDADaviesSJampNoorNSM (2006)Growthandmortal-ity are related to adult tree size in a Malaysian mixed dipterocarp forest Forest Ecology and Management 223 152ndash158 httpsdoiorg101016jforeco200510066

KitajimaK(1994)Relativeimportanceofphotosynthetictraitsandallo-cation patterns as correlates of seedling shade tolerance of 13 tropical trees Oecologia 98419ndash428httpsdoiorg101007BF00324232

KobeRKampCoatesKD(1997)Modelsofsaplingmortalityasafunc-tion of growth to characterize interspecific variation in shade tolerance of eight tree species of northwestern British Columbia The Canadian Journal of Forest Research 27227ndash236httpsdoiorg101139x96-182

KochGW Sillett S C JenningsGM ampDavis SD (2004) Thelimits to tree height Nature 428851ndash854httpsdoiorg101038nature02417

KraftNJBCrutsingerGMForrestelEJampEmeryNC(2014)Functional trait differences and the outcome of community assem-blyAnexperimentaltestwithvernalpoolannualplantsOikos 123 1391ndash1399httpsdoiorg101111oik01311

LambersHampPoorterH(2004)Inherentvariationingrowthratebe-tweenhigherplantsAsearchforphysiologicalcausesandecologi-cal consequences Advances in Ecological Research Classic Papers 34 283ndash362httpsdoiorg101016S0065-2504(03)34004-8

Laura Suarez M amp Kitzberger T (2010) Differential effects of cli-mate variability on forest dynamics along a precipitation gradient in northern Patagonia Journal of Ecology 98 1023ndash1034httpsdoiorg101111j1365-2745201001698x

Lavorel S amp Garnier E (2002) Predicting changes in community com-position and ecosystem functioning from plant traits Revisiting the Holy Grail Functional Ecology 16 545ndash556 httpsdoiorg101046j1365-2435200200664x

Lebrija-Trejos E Meave J A Poorter L Peacuterez-Garciacutea E A ampBongers F (2010) Pathways mechanisms and predictability of veg-etation change during tropical dry forest succession Perspectives in Plant Ecology Evolution and Systematics 12267ndash275httpsdoiorg101016jppees201009002

Levine JMBascompte JAdlerPBampAllesinaS (2017)Beyondpairwise mechanisms of species coexistence in complex communi-ties Nature 54656ndash64httpsdoiorg101038nature22898

LiLMcCormackMLMaCKongDZhangQChenXampPentildeuelasJ(2015a)Leafeconomicsandhydraulictraitsaredecoupledinfive

species-rich tropical-subtropical forests Ecology Letters 18 899ndash906httpsdoiorg101111ele12466

LiRZhuSChenHYJohnRZhouGZhangDampYeQ(2015b)Are functional traitsagoodpredictorofglobalchange impactsontree species abundance dynamics in a subtropical forest Ecology Letters 181181ndash1189httpsdoiorg101111ele12497

LiuXSwensonNGLinDMiXUmanaMNSchmidBampMaK(2016)Linkingindividual-levelfunctionaltraitstotreegrowthinasubtropical forest Ecology 972396ndash2405httpsdoiorg101002ecy1445

LohbeckMLebrija-TrejosEMartiacutenez-RamosMMeaveJAPoorterLampBongersF(2015)Functionaltraitstrategiesoftreesindryandwettropical forests are similar but differ in their consequences for succession PLoS One 10e0123741httpsdoiorg101371journalpone0123741

Lu P -L amp Morden C L (2014) Phylogenetic relationships amongdracaenoid genera (Asparagaceae Nolinoideae) inferred fromchloroplast DNA loci Systematic Botany 39 90ndash104 httpsdoiorg101600036364414X678035

LuskCHReichPBMontgomeryRAAckerlyDDampCavender-Bares J (2008) Why are evergreen leaves so contrary about shade Trends in Ecology and Evolution 23 299ndash303 httpsdoiorg101016jtree200802006

LuskCHampWartonDI(2007)Globalmeta-analysisshowsthatrela-tionships of leaf mass per area with species shade tolerance depend on leaf habit and ontogeny New Phytologist 176764ndash774httpsdoiorg101111j1469-8137200702264x

Markesteijn L Iraipi J Bongers F amp Poorter L (2010) Seasonal vari-ation in soil and plant water potentials in a Bolivian tropical moist and dry forest Journal of Tropical Ecology 26497ndash508httpsdoiorg101017S0266467410000271

Marks CO amp LeichowiczM J (2006) Alternative designs and theevolution of functional diversity The American Naturalist 16755ndash66httpsdoiorg101086498276

MaximovNA (1931) The physiological significance of the xeromor-phic structure of plants Journal of Ecology 19279ndash282httpsdoiorg101016S0176-1617(86)80151-1

MayfieldMM BoniM F ampAckerlyDD (2009) Traits habitatsand clades Identifying traits of potential importance to environ-mental filtering American Naturalist 174 E1ndashE22 httpsdoiorg101086599293

McDowellNAllenCDAnderson-TeixeiraKBrandoPBrienenRChambersJXuX(2018)Driversandmechanismsoftreemor-tality in moist tropical forests New Phytologist 2191ndash19httpsdoiorg101111nph15027

McDowellNPockmanWTAllenCDBreshearsDDCobbNKolb T Yepez E A (2008)Mechanisms of plant survival andmortality during drought Why do some plants survive while others succumb to drought New Phytologist 178719ndash739httpsdoiorg101111j1469-8137200802436x

McElwainJCYiotisCampLawsonT(2016)Usingmodernplanttraitrela-tionships between observed and theoretical maximum stomatal conduc-tance and vein density to examine patterns of plant macroevolution New Phytologist 20994ndash103httpsdoiorg101111nph13579

MedeirosCDScoffoniCJohnGBartlettMInman-NarahariFOstertagRhellipSackL(2018)DatafromAnextensivesuiteoffunc-tional traits distinguishes wet and dry Hawaiian forests and enables prediction of species vital rates Dryad Digital Repository httpsdoiorg105061dryadcq47n7s

Messier J Lechowicz M J McGill B J Violle C Enquist B J amp Cornelissen H (2017) Interspecific integration of trait dimensions at local scales The plant phenotype as an integrated network Journal of Ecology 1051775ndash1790httpsdoiorg1011111365-274512755

MonjeOAampBugbeeB (1992) Inherent limitationsofnondestruc-tive chlorophyll meters A comparison of two types of metersHortScience 2769ndash71


NathanJOsemYShachakMMeronEampSalguero-GoacutemezR(2016)Linking functional diversity to resource availability and disturbance Amechanisticapproachforwater-limitedplantcommunitiesJournal of Ecology 104419ndash429httpsdoiorg1011111365-274512525

Niinemets Uuml (2001) Global-scale climatic controls of leaf dry massper area density and thickness in trees and shrubs Ecology 82 453ndash469 httpsdoiorg1018900012-9658(2001)082[0453GSCCOL]20CO2

Ogburn R M amp Edwards E J (2012) Quantifying succulence Arapid physiologically meaningful metric of plant water stor-age Plant Cell amp Environment 35 1533ndash1542 httpsdoiorg101111j1365-3040201202503x

Ogle K (2003) Implications of interveinal distance for quantum yield in C4grassesAmodelingandmeta-analysisOecologia 136532ndash542httpsdoiorg101007s00442-003-1308-2

Osborne C P amp Sack L (2012) Evolution of C4 plants A newhypothesis for an interaction of CO2 and water relations me-diated by plant hydraulics Philosophical Transactions of the Royal Society B Biological Sciences 367 583ndash600 httpsdoiorg101098rstb20110261

OsoneYIshidaAampTatenoM(2008)Correlationbetweenrelativegrowth rate and specific leaf area requires associations of specific leaf area with nitrogen absorption rate of roots New Phytologist 179 417ndash427httpsdoiorg101111j1469-8137200802476x

OstertagR Inman-Narahari FCordell SGiardinaCPampSack L(2014) Forest structure in low-diversity tropical forests A studyof Hawaiian wet and dry forests PLoS One 9 e103268 httpsdoiorg101371journalpone0103268

PaineCETAmissahLAugeHBaralotoCBaruffolMBourlandNampGibsonD(2015)Globallyfunctionaltraitsareweakpredic-tors of juvenile tree growth and we do not know why Journal of Ecology 103978ndash989httpsdoiorg1011111365-274512401

Peacuterez-Harguindeguy N Diacuteaz S Garnier E Lavorel S Poorter HJaureguiberry P amp Cornelissen J H C (2013) New handbook for standardised measurement of plant functional traits worldwide Australian Journal of Botany 61 167ndash234 httpsdoiorg101071BT12225_CO

Philipson C D Dent D H OBrien M J Chamagne J Dzulkifli DNilus RampHectorA (2014)A trait-based trade-off betweengrowthandmortalityEvidencefrom15tropicaltreespeciesusingsize-specific relative growth ratesEcology and Evolution 4 3675ndash3688 httpsdoiorg101002ece31186

PlummerM(2003)JAGSAprogramforanalysisofBayesiangraphicalmodels using Gibbs sampling In Proceedings of the 3rd International workshop on distributed statistical computingViennaAustria

PoorterHLambersHampEvansBJ(2014)TraitcorrelationnetworksAwhole-plantperspectiveontherecentlycriticized leafeconomicspectrum New Phytologist 201 378ndash382 httpsdoiorg101111nph12547

PoorterHNiinemetsUumlPoorter LWright I JampVillarR (2009)Causesandconsequencesofvariationinleafmassperarea(LMA)A meta-analysis New Phytologist 182 565ndash588 httpsdoiorg101111j1469-8137200902830x

Poorter L Wright S J Paz H Ackerly D D Condit R Ibarra-Manriacutequez G amp Wright I J (2008) Are functional traits goodpredictors of demographic rates Evidence from five neotropical forests Ecology 891908ndash1920httpsdoiorg10189007-02071

Prado-Junior JASchiavini IValeVSRaymundoDLopesSFampPoorterL(2016)Functionaltraitsshapesize-dependentgrowthand mortality rates of dry forest tree species Journal of Plant Ecology 10895ndash906httpsdoiorg101093jpertw103

Price J P amp Clague D A (2002) How old is the Hawaiian biotaGeology and phylogeny suggest recent divergence Proceedings of the Royal Society B Biological Sciences 2692429ndash2435httpsdoiorg101098rspb20022175

R Core Team (2016) A language and environment for statistical computing ViennaAustriaRFoundationforStatisticalComputing

Reich P B (2014) The world-wide lsquofast-slowrsquo plant economics spec-trumA traitsmanifesto Journal of Ecology 102275ndash301httpsdoiorg1011111365-274512211

Russo S E Jenkins K L Wiser S K Uriarte M Duncan R P amp Coomes DA(2010)Interspecificrelationshipsamonggrowthmortalityandxylem traits of woody species from New Zealand Functional Ecology 24253ndash262httpsdoiorg101111j1365-2435200901670x

Sack L amp Buckley T N (2016) The developmental basis of stoma-tal density and flux Plant Physiology 171 2358ndash2363 httpsdoiorg101104pp1600476

SackLCowanPDJaikumarNampHolbrookNM(2003a)Thelsquohy-drologyrsquoof leavesCo-ordinationofstructureandfunction intem-perate woody species Plant Cell amp Environment 26 1343ndash1356httpsdoiorg101046j0016-8025200301058x

Sack L amp Frole K (2006) Leaf structural diversity is related to hydraulic capacity in tropical rain forest trees Ecology 87483ndash491httpsdoiorg10189005-0710

Sack L Grubb P J amp Marantildeoacuten T (2003b) The functional morphology of juvenile plants tolerant of strong summer drought in shaded forest understories in southern Spain Plant Ecology 168139ndash163httpsdoiorg101023A1024423820136

Sack L Melcher P J Liu W H Middleton E amp Pardee T (2006) How strong is intracanopy leaf plasticity in temperate deciduous trees Journal of Botany 93829ndash839httpsdoiorg103732ajb936829

Sack L amp Scoffoni C (2013) Leaf venation Structure function develop-ment evolution ecology and applications in the past present and future New Phytologist 198983ndash1000httpsdoiorg101111nph12253

SackLScoffoniCJohnGPPoorterHMasonCMMendez-AlonzoRampDonovanLA(2013)Howdoleafveinsinfluencetheworldwideleaf economic spectrum Review and synthesis Journal of Experimental Botany 644053ndash4080httpsdoiorg101093jxbert316

SackLScoffoniCMcKownADFroleKRawlsMHavranJCampTran T (2012) Developmentally based scaling of leaf venation archi-tecture explains global ecological patterns Nature Communications 3837httpsdoiorg101038ncomms1835

Sack L Tyree M T amp Holbrook N M (2005) Leaf hydraulic ar-chitecture correlates with regeneration irradiance in tropi-cal rainforest trees New Phytologist 167 403ndash413 httpsdoiorg101111j1469-8137200501432x

SantiagoLSKitajimaKWrightSJampMulkeySS(2004)Coordinatedchanges in photosynthesis water relations and leaf nutritional traits of canopy trees along a precipitation gradient in lowland tropical forest Oecologia 139495ndash502httpsdoiorg101007s00442-004-1542-2

Schimper A F W (1903) Plant‐geography upon a physiological basis Oxford UK Clarendon Press

ScoffoniCChateletDSPasquet-Kok JRawlsMDonoghueMJ Edwards E J amp Sack L (2016) Hydraulic basis for the evolution of photosynthetic productivity Nature Plants 2 16072 httpsdoiorg101038nplants201672

Scoffoni C Rawls M McKown A Cochard H amp Sack L (2011)Decline of leaf hydraulic conductance with dehydration Relationship to leaf size and venation architecture Plant Physiology 156832ndash843httpsdoiorg101104pp111173856

Scoffoni C Vuong C Diep S Cochard H amp Sack L (2014) Leafshrinkage with dehydration Coordination with hydraulic vulnerabil-ity and drought tolerance Plant Physiology 164 1772ndash1788 httpsdoiorg101104pp113221424

SeiwaK(2007)Trade-offsbetweenseedlinggrowthandsurvivalinde-ciduous broadleaved trees in a temperate forest Annals of Botany 99 537ndash544httpsdoiorg101093aobmcl283

SheilDBurslemDFRPampAlderD(1995)Theinterpretationandmisinterpretation of mortality rate measures Journal of Ecology 83 331ndash333httpsdoiorg1023072261571


ShipleyBBelluauMKuumlhnISoudzilovskaiaNABahnMPenuelasJ amp Poschlod P (2017) Predicting habitat affinities of plant species using commonly measured functional traits Journal of Vegetation Science 281082ndash1095httpsdoiorg101111jvs12554

Sokal R R amp Rohlf F J (2012) Biometry The principles and practice of statis‐tics in biological research(4thed)NewYorkNYWHFreemanandCo

StahlUReuBampWirthC(2014)PredictingspeciesrsquorangelimitsfromfunctionaltraitsforthetreefloraofNorthAmericaProceedings of the National Academy of Sciences of the United States of America 111 13739ndash13744httpsdoiorg101073pnas1300673111

Sterck F Markesteijn L Schieving F amp Poorter L (2011) Functional traitsdetermine trade-offs andniches in a tropical forest commu-nity Proceedings of the National Academy of Sciences of the United States of America 108 20627ndash20632 httpsdoiorg101073pnas1106950108

Stratton L Goldstein G amp Meinzer F C (2000) Stem water storage capacity and efficiency of water transport Their functional signifi-cance in a Hawaiian dry forest Plant Cell and Environment 2399ndash106httpsdoiorg101046j1365-3040200000533x

The Plant List (2013) Version 11 Published on the Internet Retrieved from httpwwwtheplantlistorg (accessed 6th November)

ThomasFMampVeskPA (2017)Are trait-growthmodels transfer-ablePredictingmulti-speciesgrowthtrajectoriesbetweenecosys-tems using plant functional traits PLoS One 12e0176959httpsdoiorg101371journalpone0176959

Uriarte M Lasky J R Boukili V K Chazdon R L amp Merow C (2016) A trait-mediated neighbourhood approach to quantify climate im-pacts on successional dynamics of tropical rainforests Functional Ecology 30157ndash167httpsdoiorg101371journalpone0176959

Venables W N amp Ripley B D (2002) Modern applied statistics with S(4thed)NewYorkNYSpringerhttpsdoiorg101007978-0-387-21706-2

Vendramini FDiacuteaz SGurvichD EWilson P J ThompsonKampHodgsonJG(2002)Leaftraitsasindicatorsofresource-usestrat-egy in floras with succulent species New Phytologist 154147ndash157httpsdoiorg101046j1469-8137200200357x

ViolleCNavasM-LVileDKazakouEFortunelCHummelIampGarnier E (2007) Let the concept of trait be functional Oikos 116 882ndash892httpsdoiorg101111j20070030-129915559x

VisserMDBruijningMWrightSJMuller-LandauHCJongejansE Comita L S amp de Kroon H (2016) Functional traits as predictors of vital rates across the life cycle of tropical trees Functional Ecology 30168ndash180httpsdoiorg1011111365-243512621

WagnerGPampAltenbergL(1996)Complexadaptationsandtheevolutionof evolvability Evolution 50967ndash976httpsdoiorg1023072410639

WagnerW LHerbstDRampSommer SH (1999)A manual of the flowering plants of HawaiI (volumes I and II) Honolulu HI University of Hawaii Press

WaltersMBampReichPB(1999)Low-lightcarbonbalanceandshadetolerance in the seedlings of woody plants Do winter deciduous and broad-leaved evergreen species differNew Phytologist 143 143ndash154httpsdoiorg101046j1469-8137199900425x

Wang H Prentice I C Keenan T F Davis T W Wright I J Cornwell W K amp Peng C (2017) Towards a universal model for carbon di-oxide uptake by plants Nature Plants 3 734ndash741 httpsdoiorg101038s41477-017-0006-8

Wang R Yu G He NWangQ Zhao N Xu Z amp Ge J (2015)Latitudinal variation of leaf stomatal traits from species to commu-nity level in forests Linkage with ecosystem productivity Scientific Reports 51ndash11httpsdoiorg101038srep14454

WeiherEampKeddyPA(1995)Assemblyrulesnullmodelsandtraitdispersion New questions from old patterns Oikos 74 159ndash164httpsdoiorg1023073545686

WestobyMFalsterDSMolesATVeskPAampWrightIJ(2002)Plant ecological strategies Some leading dimensions of variation between species Annual Review of Ecology and Systematics 33125ndash159httpsdoiorg101146annurevecolsys33010802150452

WestobyMampWrightIJ (2006)Land-plantecologyonthebasisoffunctional traits Trends in Ecology and Evolution 21 261ndash268 httpsdoiorg101016jtree200602004

WitkowskiETFampLamontBB(1991)Leafspecificmassconfoundsleaf density and thickness Oecologia 88 486ndash493 httpsdoiorg101007BF00317710

Wright I J Dong N Maire V Prentice I C Westoby M Diaz S amp Wilf P (2017) Global climatic drivers of leaf size Science 357917ndash921httpsdoiorg101126scienceaal4760

Wright S J Kitajima K Kraft N J B Reich P B Wright I J Bunker DEampZanneAE(2010)Functionaltraitsandthegrowthndashmor-talitytrade-offintropicaltreesEcology 913664ndash3674httpsdoiorg10189009-23351

Wright I J Reich P B amp Westoby M (2001) Strategy shifts in leaf phys-iologystructureandnutrientcontentbetweenspeciesofhigh-andlow-rainfallandhigh-and low-nutrienthabitatsFunctional Ecology 15423ndash434httpsdoiorg101046j0269-8463200100542x

WrightIJReichPBWestobyMAckerlyDDBaruchZBongersF amp Villar R (2004) The worldwide leaf economics spectrumNature 428821ndash827httpsdoiorg101038nature02403

Wright I J Westoby M amp Reich P B (2002) Convergence towards higherleafmassperareaindryandnutrient-poorhabitatshasdiffer-ent consequences for leaf life span Journal of Ecology 90534ndash543httpsdoiorg101046j1365-2745200200689x

YangJCaoMampSwensonNG(2018a)Whyfunctionaltraitsdonotpredict tree demographic rates Trends in Ecology and Evolution 33 326ndash336 httpsdoiorg101016jtree201803003

YangYMordenCW Sporck-KoehlerM J Sack LampBerryPE(2018b) Repeated range expansion and niche shift in a volcanic hotspot archipelago Radiation of C4 Hawaiian Euphorbia (subgenus Chamaesyce Euphorbiaceae) Ecology and Evolution 8 8523ndash8536httpsdoiorg101002ece34354

ZhuS-DSongJ-JLiR-HampYeQ(2013)Planthydraulicsandpho-tosynthesisof34woodyspeciesfromdifferentsuccessionalstagesof subtropical forests Plant Cell amp Environment 36879ndash891httpsdoiorg101111pce12024

SUPPORTING INFORMATION

Additional supporting information may be found online in theSupporting Information section at the end of the article

How to cite this article Medeiros CD Scoffoni C John GP etalAnextensivesuiteoffunctionaltraitsdistinguishesHawaiian wet and dry forests and enables prediction of species vital rates Funct Ecol 2018001ndash23 httpsdoiorg1011111365-243513229


1emsp |emspINTRODUC TION

Functional traits influence plant growth reproduction and survival and thereby fitness (Lavorel amp Garnier 2002 Violle et al 2007) and thuscanbeusedtopredictvitalrates(Adleretal2014Poorteretal2008 Uriarte Lasky Boukili Chazdon amp Merow 2016) habitat pref-erences (Shipley et al 2017) and spatial distributions (Stahl Reu amp Wirth2014) Fordecadesmost studieshave focusedon relativelyfew commonly measured functional traits with some justification given that overall trait variation can be simplified statistically into a few fundamental dimensions (Diaz et al 2016 Messier et al 2017) However several have argued that more extensive suites of traits would enable strong predictive and explanatory power (Greenwood etal 2017Paine etal 2015Reich 2014YangCaoampSwenson2018a) and this argument has conceptual support because mechanis-tic models of growth and survival are sensitive to a broad set of traits as inputs (Marks amp Leichowicz 2006 Osborne amp Sack 2012 Sterck Markesteijn Schieving amp Poorter 2011) The traits measured in this studyincludewell-studiedfunctionaltraitswithintheleafandwoodldquoeconomics spectrardquo (LES and WES respectively) which describe trade-offs in plant carbon balancewith given traits contributing toeither fast growth and resource turnover or slow growth and longer tissuelifespansandstresstolerance(Chaveetal2009Wrightetal2004)Inadditionweincludedawidersetoftraitsrecognizedtohaveproximal physiological influence on water transport gas exchange and resource economics The aim of this study was to assess six key hy-potheses derived from first principles in trait physiology and ecology (Table1)utilizing45traitsexpectedtoshowcontrastingadaptationacross forests andor to influence relative growth rate (RGRdbh and RGRbiom) and mortality (m) (Table 2) We pursued this aim while rec-ognizing that many more traits than those we included play important roles and that species differ in the traits with most important influence on vital rates

First we tested the ability of an extensive suite of traits to resolve variation between Hawaiian wet and dry forest species given their contrasting adaptation We assessed traits which based on the previous literature would have specific mechanistic influences on resource acquisition growth and stress tolerance (Table 1 with detailed reasoning in Supporting Information Table S10) In particular we expected that relative to the dry forest the wet forest species would have shifted their traits values in

the direction beneficial to their adaptation to greater availabil-ity of water and soil nutrients Such trait shifts would include greater mean and maximum plant height (King Davies amp Noor 2006Koch Sillett JenningsampDavis 2004) lowerwoodden-sity(WDChaveetal2009Gleasonetal2016HackeSperryPockmanDavisampMcCulloh2001)andseedmass(Gross1984KhuranaampSingh2004)higheroverall ratesofphotosynthesisand rates of electron transport and carboxylation (all per unit leaf area andor dry mass) and higher values for the ratio of inter-nal to ambient CO2 (cica) related to higher values of carbon iso-tope discrimination (Δleaf Farquhar EhleringerampHubick1989Franks amp Beerling 2009 Donovan amp Ehleringer 1994 Wanget al 2017) larger and denser stomata and higher stomatal con-ductance (Beaulieu Leitch Patel Pendharkar amp Knight 2008 FranksampBeerling2009FranksampFarquhar2007HetheringtonampWoodward 2003 Sackamp Buckley 2016Wang etal 2015)higher densities of leaf major and minor veins and free ending veins (Brodribb Feild amp Jordan 2007 Iida et al 2016 Sack amp Frole 2006 Sack amp Scoffoni 2013 Scoffoni et al 2016) thinner and larger leaves of higher saturated water content and lower dry mass density lower water mass and dry mass per area and lower dry matter content with lesser shrinkage in area under dehydra-tion (Bartlett Scoffoni amp Sack 2012b Diaz et al 2016 Evans 1973Niinemets2001OgburnampEdwards2012SackampScoffoni2013ScoffoniVuongDiepCochardampSack2014Vendraminietal2002WestobyampWright2006Wrightetal2004)highfoliar concentrations of nitrogen (N) phosphorus (P) and chlo-rophyll (Chl) and lower concentration of carbon (Chatuverdi Raghubanshi amp Singh 2011 Lambers amp Poorter 2004Wrightetal2004) lowerNP(Elseretal2000)andgreaterstomatalopening relative to maximum aperture and relative to N (Franks ampBeerling 2009Wright ReichampWestoby 2001)Given thatspecies of the wet forest are adapted to lower understorey irra-diance also led to the expectation of lower rates of photosynthe-sis and greater Δleaf (DonovanampEhleringer 1994 Evans 2013Farquhar etal 1989 FranksampBeerling 2009) larger leaf area(Chatuverdi etal 2011 Niinemets 2001) lower LMA (SackGrubbampMarantildeoacuten2003bWaltersampReich1999)lowerNandP and higher C and Chlconcentrations(Givnish1988Niinemets2001 Lusk amp Warton 2007 Poorter Niinemets Poorter Wright amp Villar 2009 Chatuverdi etal 2011) higher Chl to N ratio

correlated within modules and predicted speciesrsquo RGR and m across forests with stronger relationships when stratifying by tree size Models based on multiple traits predicted vital rates across forests (R2 = 070ndash072 p lt 001)

4 Ourfindingsareconsistentwithapowerfulroleofbroadsuitesoffunctionaltraitsin contributing to forest speciesrsquo distributions integrated plant design and vital rates

K E Y W O R D S

drought tolerance endemic species forest tree demography growth analysis


TAB

LE 1


fore

sts

and

to e

nabl

e pr

edic

tion

of v

ital r

ates

acr

oss

spec

ies

Hyp

othe

sis

Expl

anat

ion

base

d on

firs

t pr

inci

ples

Refe

renc

esTe

stSu

ppor

t

1 W

et a

nd d

ry fo

rest

spe

cies

wou

ld d

iffer

in

num

erou

s tr

aits

as

expe

cted

from

con

tras

ting

adap

tatio

n


and

soil

wou

ld le

ad to

var

iatio

n am

ong

spec

ies

in n

umer

ous

func

tiona

l tra

its im

port

ant i

n pl

ant p

erfo

rman

ce





and


Yes

2 T

rait

valu

es w

ould

be

mor

e co

nver

gent

am

ong

dry

than

wet

fore

st s

peci

es d

ue to

the


avai

labi

lity

Envi

ronm

enta

l filt

erin

g is

ex

pect

ed to

redu

ce fu

nctio

nal

dive

rsity

by

cons

trai

ning

the

rang

e of

pos

sibl

e tr

ait s

tate

s ac

ross

hab

itats

Cor

nwel

l et a

l (2

006)

May

field

Bon

i an



Asefaetal(2017)

t tes

t on

the

coef

ficie

nt o

f var

iatio

n in

trai

ts

from

MW

F an

d LD

F F

test

s on

the

varia

nce

of e

ach

trai

t bet

wee

n M

WF

and

LDF

Yes

3 T

raits

wou

ld b

e in

terc

orre

late

d w

ithin

fu

nctio

nal ldquo

mod

ules

rdquoSe

lect

ion

on m

ultip

le tr

aits

acr

oss

envi

ronm

ents

wou

ld le

ad to

tr

aitndash

trai

t cor

rela

tions

with

in

orga

ns a

nd fu

nctio

nal

mod

ules

du

e to

com

mon

dev

elop

men

tal

path

way

fun

ctio

n o

r ben

efit

in

give

n en

viro

nmen

ts



(2015b)

Pear

son

s co

rrel

atio

ns b

etw

een

trai

ts w

ithin

fu

nctio

nal ldquo

mod

ules

rdquoYes

4RGRand

m w

ould

cor

rela

te w

ith s

peci

fic

trai

tsTr

aits

con

trib

ute

mec

hani

stic

ally

di

rect

ly to

RG

Rs a

nd m

in g

iven

ha

bita

ts


Mar

ks a

nd L

eich

owic

z (2

006)

and

Wrig

ht e

t al

(201

0)

Pear

son

s co

rrel

atio

ns b

etw

een

spec

ific

trai

ts

and

vita

l rat

esYes

5RGRand

m w

ould

cor

rela

te w

ith tr

aits

mor

e fr

eque

ntly

whe

n st

ratif

ying

by

tree

siz


tren

ds in

trai

ts a

nd v

ital r

ates

m

ean

that

trai

tndashvi

tal r

ate

corr

elat

ions

wou

ld b

e re

duce

d gi

ven

com

paris

on o

f spe

cies

m

ean

valu

es w

hen

spec

ies

vary

in

siz

e di

strib

utio

ns s

trat

ifyin

g by

siz

e sh

ould

ther

efor

e st

reng

then

trai

tndashvi

tal r

ate

rela

tions

hips


Juni

or e

t al

(201

6)Ba

yesi

an m

odel

to e

stim

ate

vita

l rat

es a

t gi

ven

plan

t siz

es fo

llow

ed b

y Ke

ndal

ls

corr

elat

ions

bet

wee

n tr

aits

and

vita

l rat

es a

t ea

ch s

ize

Yes

6 R

GR

and

m c

an b

e pr

edic

ted

base

d on

trait-basedmodels

Giv

en re

latio

nshi

ps o

f vita

l rat

es

with

giv

en tr

aits

com

bina

tions

of

trai

ts s

houl

d be

str

ongl

y pr

edic

tive

Poor

ter e

t al

(200

8) U

riart

e et

al

(201

6) a

nd

Thom

as a

nd V

esk

(201

7)Li

near

regr

essi

onYes


TAB

LE 2

emspSt

udy

trai

ts re

latin

g to

sto

mat

al m

orph

olog

y le

af v

enat

ion

leaf

and

woo

d ec

onom

ics

and

stru

ctur

e le

af c

ompo

sitio

n a

nd e

stim

ated

pho

tosy

nthe

sis

and

plan

t siz

e a

nd th

e vi

tal

rate

s m

easu

red

for s

peci

es fr

om a

mon

tane

wet

fore

st (W

) and

a lo

wla

nd d

ry fo

rest

(D) i

n H

awai

i F

or th

e tr

aits

we

prov

ide

sym

bols

uni

ts h

ypot

hese

s fo

r giv

en tr

aits

for d

iffer

ence

s be

twee

n fo

rest

s an

d re

sults

from

sta

tistic

al te

sts

and

hyp

othe

ses

for c

orre

latio

ns w

ith v

ital r

ates

(rel

ativ

e gr

owth

rate

and

mor

talit

y) a

nd re

sults

from

Pea

rson

s co

rrel

atio

n te

sts

(whe

n on

e re

sult

is p

rese

nted

this

repr

esen

ts s

peci

es fr

om b

oth

fore

sts

toge

ther

and

whe

n tw

o re

sults

are

pre

sent

ed th

ese

repr

esen

t spe

cies

in th

e w

et a

nd d

ry fo

rest

s se

para

tely

) an

d re

fere

nces

su

ppor

ting

the

hypo

thes

es n

s in

dica

tes

no s

igni

fican

t diff

eren

ce a

t plt005ldquoW


have

a h

ighe

r tra

it va

lue

than

the

dry

fore

st o

n av

erag

e ldquoD

rdquo tha

t the

dry

fore

st w

ould

hav

e th

e hi

gher

trai

t val

ue o

n av

erag

e a

nd ldquoe

ither

rdquo den

otes

the

exis

tenc

e of

mul

tiple

pub

lishe

d hy

poth

eses

whe

reby

eith

er M

WF

or L

DF

coul

d be

exp

ecte

d to

hav

e th

e hi

gher

trai

t val

ue (S

uppo

rtin

g In

form

atio

n Ta

ble

S10)

Pos

itive

sig

ns (+

) ind

icat

e th

e ex

pect

atio

n or

find

ing

of a

pos

itive


Tabl

e S1

0 plt005p

lt 0

01

p lt

000

1

Trai

tvi

tal r

ate

Sym

bol

Uni

tH

ypot

hese

s W

or D

hi

gher

W

or D

hig

her

Hyp

othe

ses

trai

tndashvi

tal

rate

cor

rela

tion

Dire

ctio

n of

trai

tndashvi

tal

rate

cor

rela

tion

Refe

renc

e

Stom

atal

mor

phol

ogy

Stom

atal

den

sity

dst

omat

am

m2

eith

erns

++

1ndash5

Stom

atal

diff

eren

tiatio

n ra

te (o

r ind

ex)

i-

eith

erW

+

+

2ndash6

Stom

atal

are

as

μm2

WW

+

ns157

Gua

rd c

ell l

engt

hG

CL

μmW

W

157

Gua

rd c

ell w

idth

GC

Wμm

WW

157

Pore

leng

thSP

Lμm

WW

157

Epid

erm

al p

avem

ent c

ell a

rea

eμm

2W

W

8

Max

imum

sto

mat

al c

ondu

ctan

ceg m

axm

ol mminus2

sminus1

eith

erW

+

+2ndash59ndash10

Leaf

ven

atio

n

Maj

or v

ein

dens

ityVLA

maj

orm

m p

er m

m2

eith

erD

+ns

2ndash411ndash15

Min

or v

ein

dens

ityVLA

min

orm

m p

er m

m2

eith

erD

+ns

2ndash411ndash15

Tota

l vei

n de

nsity

VLA

tota

lm

m p

er m

m2

eith

erD

+ns

2ndash411ndash15

Free

end

ing

vein

den

sity

FEV

pe

r mm

2ei

ther

D

2ndash411ndash15

Leaf

and

woo

d ec

onom

ics

and

stru

ctur

e

Leaf

are

aLA

cm2

Wns

+WnsDminus

16ndash1

8

Leaf

mas

s pe

r are

aLMA

gm

2ei

ther

nsminus

W+Dminus

12 1

8ndash23

Leaf

thic

knes

sLT

mm

eith

erns

minusminus

12 1

8ndash23

Leaf

den

sity

LDg

cm3

eith

erns

minusW+Dminus

12 1

8ndash23

Leaf

dry

mat

ter c

onte

ntLD

MC

gg

DD

minus

ns1824

Satu

rate

d w

ater

con

tent

SWC

gg

eith

erW

25ndash27

Wat

er m

ass

per a

rea

WMA

gm

2D

ns25ndash27

Perc

enta

ge lo

ss a

rea

(dry

)PLA

dry

W

W

2829

Woo

d de

nsity

WD

gcm

3ei

ther

D

minus

ns530ndash32

(Con

tinue

s)


Trai

tvi

tal r

ate

Sym

bol

Uni

tH

ypot

hese

s W

or D

hi

gher

W

or D

hig

her

Hyp

othe

ses

trai

tndashvi

tal

rate

cor

rela

tion

Dire

ctio

n of

trai

tndashvi

tal

rate

cor

rela

tion

Refe

renc

e

Leaf

com

posi

tion

Nitr

ogen

con

cent

ratio

n pe

r lea

f are

aN

area

gm

2ei

ther

ns+

W+Dminus

2ndash520

Nitr

ogen

con

cent

ratio

n pe

r lea

f mas

sN

mas

sm

gg

eith

erns

++

2ndash520

Phos

phor

us c

once

ntra

tion

per l

eaf a

rea

P area

gm

2ei

ther

ns+

W+Dminus

2ndash520

Phos

phor

us c

once

ntra

tion

per l

eaf m

ass

P mas

sm

gg

eith

erW

+

+2ndash520

Chl

orop

hyll

conc

entr

atio

nCh

l area

SPAD

eith

erns

2ndash433

Chl

orop

hyll

per m

ass

Chl m

ass

SPADgminus1

m2

eith

erns

2ndash433

Car

bon

conc

entr

atio

n pe

r lea

f mas

sC

mas

sm

gg

Wns

34ndash35

Nitr

ogen

pho

spho

rus

ratio

NP

ndashei

ther

D

ndashns

35

Chl

orop

hyll

nitr

ogen

per

are

a ra

tioCh

l area

Nar

eaSPADgminus1

m2

Wns

5

Car

bon

isot

ope

disc

rimin

atio

nΔ

leaf

permilW

W

36ndash39

|Tur

gor l

oss

poin

t||π

tlp|

MPa

DD

28

Estim

ated

pho

tosy

nthe

sis

Elec

tron

tran

spor

t rat

e pe

r are

aJm

axar

eaμm

ol mminus2

sminus1

eith

erns

+W+Dminus

2ndash436ndash38

Elec

tron

tran

spor

t rat

e pe

r mas

sJm

axm

ass

nmol

gminus1

sminus1

eith

erns

++

2ndash436ndash38

Max

imum

rate

of c

arbo

xyla

tion

per a

rea

Vcm

axar

eaμm

ol mminus2

sminus1

eith

erns

+W+Dminus

2ndash436ndash38

Max

imum

rate

of c

arbo

xyla

tion

per m

ass

Vcm

axm

ass

nmol

gminus1

sminus1

eith

erns

++

2ndash436ndash38

Ratio

of i

nter

cellu

lar t

o am

bien

t CO

2 co

ncen

trat

ions

c ica

ndashei

ther

W

+

ns36

ndash38

Tim

e in

tegr

ated

leaf

CO

2 ass

imila

tion

rate

per

ar

ea Aarea

μmol

mminus2

sminus1

eith

erns

+W+Dminus

9

Tim

e in

tegr

ated

leaf

CO

2 ass

imila

tion

rate

per

m

ass

Amass

nmol

gminus1

sminus1

eith

erW

++

9

Tim

e in

tegr

ated

sto

mat

al c

ondu

ctan

cegcleaf

mm

ol mminus2

sminus1

eith

erW

+

W+Dminus

9

Tim

e in

tegr

ated

max

imum

sto

mat

al c

ondu

ct-

ance

ratio

gcleaf

g max

ndashei

ther

ns2ndash49

Max

imum

sto

mat

al c

ondu

ctan

cen

itrog

en p

er

area

ratio

g max

Nar

eam

ol gminus1

sminus1

WW

40

Plan

t siz

e

Mea

n he

ight

Hm

Wns

46ndash47

Max

imum

hei

ght

Hm

axm

WW

46ndash47

Seed

mas

sSM

mg

Dns

48ndash49

TAB

LE 2

emsp(C

ontin

ued)

(Con

tinue

s)






ait

vita

l rat

eSy

mbo

lU

nit

Hyp

othe

ses

W o

r D

high

er

W o

r D h

ighe

rH

ypot

hese

s tr

aitndash

vita

l ra

te c

orre

latio

nD

irect

ion

of tr

aitndash

vita

l ra

te c

orre

latio

nRe

fere

nce

Vita

l rat

es

Rela

tive

grow

th ra

te (d

iam

eter

incr

emen

t)RG

R dbh

cm c

mminus1

yea

rminus1

eith

erns

52041

Rela

tive

grow

th ra

te (b

iom

ass

incr

emen

t)RG

R biom

kg k

gminus1 y

earminus1

eith

erns

52041

Mor

talit

y ra

tem

p

er y

ear

eith

erD

42ndash45








TAB

LE 2

emsp(C

ontin

ued)















Δt where x is










































































120590forest






















3emsp |emspRESULTS




















Stomatal morphology



Leaf

ve

natio

n

Estim

ate

d p

ho

tosyn

the

sis

Leaf

and

woo

d ec

onom

ics

and

stru

ctur

e

Leaf composition

0

50

100

150

200

250

disGCL

GCW

SPL

e

gmax

VLAmajor

VLAminor

VLAtotal

FEV

LA

LMA

LT

LD

LDMC

SWC

WMA

PLAdry

WDNarea


Cmass

N P

Chlarea Nmass

leaf

tlp

Jmaxarea

Jmaxmass

Vcmaxarea

Vcmaxmass

c i ca

Aarea

Amass

gcleaf

gcleaf gmax

gmax Narea

RGRdbh

RGRbiom

mH

HmaxSM









0 5 10 15

m (year)

000

002

004

RG

Rdb

h( c

mcm

minus1ye

arminus1

)

(a)

r = 0550 5 10 15

m (year)

000

005

010

RG

Rbi

om(k

gkg

minus1ye

arminus1

)

(b)

r = 057


minus10

minus05

00

05

10

τ τ(mvs

RG

Rdb

h)

(c)


minus10

minus05

00

05

10

(mvs

R

GR

biom

)

1015

2025

Spe

cies

(d)



mass RGRbiom=253LMA


minor

RGRbiom=0041P097


mass m=004N167

mass plt005p lt 001

p lt 0001

100 300 500 700

d (stomatamm2)

000

002

004

RG

Rdb

h (c

m c

mminus1

year

minus1) (a)

r = 0640 2 4 6 8


(b)

r = 057

40 60 80 100 120


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (c) r = 051

50 100 150 200 250 300

LMA (gm2)

(d) r = minus 050

2 4 6 8 10 12


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (e) r = minus 056

05 10 15 20

Pmass (mgg)

(f) r = 048

40 60 80 100 120


05

1015

m (

yea

r)

(g) r = 061

5 10 15 20 25 30

Nmass (mg g)

(h) r = 050














mass


50 100 200 300

LMA (gm2)

05

1015

m (

y

ear)

(a)

r = 085r = minus 091

0 5 10 15 20


(b)

r = 085r = minus 085

05 10 15 20

Pmass (mgg)

05

1015

m (

y

ear)

(c)

r = 089r = ns

0 50 100 150

LA (cm2)

(d)

r = nsr = minus 076








10

00

05

10

(RG

Rdb

hvs

gm

axN

area) (a)

10

00

05

10

(RG

Rdb

hvs

Hm

ax)

1015

2025

Spe

cies

(b)


10

00

05

10

(mvs

d)

(c)


10

00

05

10

(mvs

Pm

ass)

1015

2025

Spe

cies

(d)







Model






d 067043 776eminus05 532






d 065040 201eminus04 519







Site ndash 665 312





plt005p lt 001 p lt 0001














000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070




ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S
































































































































































TAB

LE 1


fore

sts

and

to e

nabl

e pr

edic

tion

of v

ital r

ates

acr

oss

spec

ies

Hyp

othe

sis

Expl

anat

ion

base

d on

firs

t pr

inci

ples

Refe

renc

esTe

stSu

ppor

t

1 W

et a

nd d

ry fo

rest

spe

cies

wou

ld d

iffer

in

num

erou

s tr

aits

as

expe

cted

from

con

tras

ting

adap

tatio

n


and

soil

wou

ld le

ad to

var

iatio

n am

ong

spec

ies

in n

umer

ous

func

tiona

l tra

its im

port

ant i

n pl

ant p

erfo

rman

ce





and


Yes

2 T

rait

valu

es w

ould

be

mor

e co

nver

gent

am

ong

dry

than

wet

fore

st s

peci

es d

ue to

the


avai

labi

lity

Envi

ronm

enta

l filt

erin

g is

ex

pect

ed to

redu

ce fu

nctio

nal

dive

rsity

by

cons

trai

ning

the

rang

e of

pos

sibl

e tr

ait s

tate

s ac

ross

hab

itats

Cor

nwel

l et a

l (2

006)

May

field

Bon

i an



Asefaetal(2017)

t tes

t on

the

coef

ficie

nt o

f var

iatio

n in

trai

ts

from

MW

F an

d LD

F F

test

s on

the

varia

nce

of e

ach

trai

t bet

wee

n M

WF

and

LDF

Yes

3 T

raits

wou

ld b

e in

terc

orre

late

d w

ithin

fu

nctio

nal ldquo

mod

ules

rdquoSe

lect

ion

on m

ultip

le tr

aits

acr

oss

envi

ronm

ents

wou

ld le

ad to

tr

aitndash

trai

t cor

rela

tions

with

in

orga

ns a

nd fu

nctio

nal

mod

ules

du

e to

com

mon

dev

elop

men

tal

path

way

fun

ctio

n o

r ben

efit

in

give

n en

viro

nmen

ts



(2015b)

Pear

son

s co

rrel

atio

ns b

etw

een

trai

ts w

ithin

fu

nctio

nal ldquo

mod

ules

rdquoYes

4RGRand

m w

ould

cor

rela

te w

ith s

peci

fic

trai

tsTr

aits

con

trib

ute

mec

hani

stic

ally

di

rect

ly to

RG

Rs a

nd m

in g

iven

ha

bita

ts


Mar

ks a

nd L

eich

owic

z (2

006)

and

Wrig

ht e

t al

(201

0)

Pear

son

s co

rrel

atio

ns b

etw

een

spec

ific

trai

ts

and

vita

l rat

esYes

5RGRand

m w

ould

cor

rela

te w

ith tr

aits

mor

e fr

eque

ntly

whe

n st

ratif

ying

by

tree

siz


tren

ds in

trai

ts a

nd v

ital r

ates

m

ean

that

trai

tndashvi

tal r

ate

corr

elat

ions

wou

ld b

e re

duce

d gi

ven

com

paris

on o

f spe

cies

m

ean

valu

es w

hen

spec

ies

vary

in

siz

e di

strib

utio

ns s

trat

ifyin

g by

siz

e sh

ould

ther

efor

e st

reng

then

trai

tndashvi

tal r

ate

rela

tions

hips


Juni

or e

t al

(201

6)Ba

yesi

an m

odel

to e

stim

ate

vita

l rat

es a

t gi

ven

plan

t siz

es fo

llow

ed b

y Ke

ndal

ls

corr

elat

ions

bet

wee

n tr

aits

and

vita

l rat

es a

t ea

ch s

ize

Yes

6 R

GR

and

m c

an b

e pr

edic

ted

base

d on

trait-basedmodels

Giv

en re

latio

nshi

ps o

f vita

l rat

es

with

giv

en tr

aits

com

bina

tions

of

trai

ts s

houl

d be

str

ongl

y pr

edic

tive

Poor

ter e

t al

(200

8) U

riart

e et

al

(201

6) a

nd

Thom

as a

nd V

esk

(201

7)Li

near

regr

essi

onYes


TAB

LE 2

emspSt

udy

trai

ts re

latin

g to

sto

mat

al m

orph

olog

y le

af v

enat

ion

leaf

and

woo

d ec

onom

ics

and

stru

ctur

e le

af c

ompo

sitio

n a

nd e

stim

ated

pho

tosy

nthe

sis

and

plan

t siz

e a

nd th

e vi

tal

rate

s m

easu

red

for s

peci

es fr

om a

mon

tane

wet

fore

st (W

) and

a lo

wla

nd d

ry fo

rest

(D) i

n H

awai

i F

or th

e tr

aits

we

prov

ide

sym

bols

uni

ts h

ypot

hese

s fo

r giv

en tr

aits

for d

iffer

ence

s be

twee

n fo

rest

s an

d re

sults

from

sta

tistic

al te

sts

and

hyp

othe

ses

for c

orre

latio

ns w

ith v

ital r

ates

(rel

ativ

e gr

owth

rate

and

mor

talit

y) a

nd re

sults

from

Pea

rson

s co

rrel

atio

n te

sts

(whe

n on

e re

sult

is p

rese

nted

this

repr

esen

ts s

peci

es fr

om b

oth

fore

sts

toge

ther

and

whe

n tw

o re

sults

are

pre

sent

ed th

ese

repr

esen

t spe

cies

in th

e w

et a

nd d

ry fo

rest

s se

para

tely

) an

d re

fere

nces

su

ppor

ting

the

hypo

thes

es n

s in

dica

tes

no s

igni

fican

t diff

eren

ce a

t plt005ldquoW


have

a h

ighe

r tra

it va

lue

than

the

dry

fore

st o

n av

erag

e ldquoD

rdquo tha

t the

dry

fore

st w

ould

hav

e th

e hi

gher

trai

t val

ue o

n av

erag

e a

nd ldquoe

ither

rdquo den

otes

the

exis

tenc

e of

mul

tiple

pub

lishe

d hy

poth

eses

whe

reby

eith

er M

WF

or L

DF

coul

d be

exp

ecte

d to

hav

e th

e hi

gher

trai

t val

ue (S

uppo

rtin

g In

form

atio

n Ta

ble

S10)

Pos

itive

sig

ns (+

) ind

icat

e th

e ex

pect

atio

n or

find

ing

of a

pos

itive


Tabl

e S1

0 plt005p

lt 0

01

p lt

000

1

Trai

tvi

tal r

ate

Sym

bol

Uni

tH

ypot

hese

s W

or D

hi

gher

W

or D

hig

her

Hyp

othe

ses

trai

tndashvi

tal

rate

cor

rela

tion

Dire

ctio

n of

trai

tndashvi

tal

rate

cor

rela

tion

Refe

renc

e

Stom

atal

mor

phol

ogy

Stom

atal

den

sity

dst

omat

am

m2

eith

erns

++

1ndash5

Stom

atal

diff

eren

tiatio

n ra

te (o

r ind

ex)

i-

eith

erW

+

+

2ndash6

Stom

atal

are

as

μm2

WW

+

ns157

Gua

rd c

ell l

engt

hG

CL

μmW

W

157

Gua

rd c

ell w

idth

GC

Wμm

WW

157

Pore

leng

thSP

Lμm

WW

157

Epid

erm

al p

avem

ent c

ell a

rea

eμm

2W

W

8

Max

imum

sto

mat

al c

ondu

ctan

ceg m

axm

ol mminus2

sminus1

eith

erW

+

+2ndash59ndash10

Leaf

ven

atio

n

Maj

or v

ein

dens

ityVLA

maj

orm

m p

er m

m2

eith

erD

+ns

2ndash411ndash15

Min

or v

ein

dens

ityVLA

min

orm

m p

er m

m2

eith

erD

+ns

2ndash411ndash15

Tota

l vei

n de

nsity

VLA

tota

lm

m p

er m

m2

eith

erD

+ns

2ndash411ndash15

Free

end

ing

vein

den

sity

FEV

pe

r mm

2ei

ther

D

2ndash411ndash15

Leaf

and

woo

d ec

onom

ics

and

stru

ctur

e

Leaf

are

aLA

cm2

Wns

+WnsDminus

16ndash1

8

Leaf

mas

s pe

r are

aLMA

gm

2ei

ther

nsminus

W+Dminus

12 1

8ndash23

Leaf

thic

knes

sLT

mm

eith

erns

minusminus

12 1

8ndash23

Leaf

den

sity

LDg

cm3

eith

erns

minusW+Dminus

12 1

8ndash23

Leaf

dry

mat

ter c

onte

ntLD

MC

gg

DD

minus

ns1824

Satu

rate

d w

ater

con

tent

SWC

gg

eith

erW

25ndash27

Wat

er m

ass

per a

rea

WMA

gm

2D

ns25ndash27

Perc

enta

ge lo

ss a

rea

(dry

)PLA

dry

W

W

2829

Woo

d de

nsity

WD

gcm

3ei

ther

D

minus

ns530ndash32

(Con

tinue

s)


Trai

tvi

tal r

ate

Sym

bol

Uni

tH

ypot

hese

s W

or D

hi

gher

W

or D

hig

her

Hyp

othe

ses

trai

tndashvi

tal

rate

cor

rela

tion

Dire

ctio

n of

trai

tndashvi

tal

rate

cor

rela

tion

Refe

renc

e

Leaf

com

posi

tion

Nitr

ogen

con

cent

ratio

n pe

r lea

f are

aN

area

gm

2ei

ther

ns+

W+Dminus

2ndash520

Nitr

ogen

con

cent

ratio

n pe

r lea

f mas

sN

mas

sm

gg

eith

erns

++

2ndash520

Phos

phor

us c

once

ntra

tion

per l

eaf a

rea

P area

gm

2ei

ther

ns+

W+Dminus

2ndash520

Phos

phor

us c

once

ntra

tion

per l

eaf m

ass

P mas

sm

gg

eith

erW

+

+2ndash520

Chl

orop

hyll

conc

entr

atio

nCh

l area

SPAD

eith

erns

2ndash433

Chl

orop

hyll

per m

ass

Chl m

ass

SPADgminus1

m2

eith

erns

2ndash433

Car

bon

conc

entr

atio

n pe

r lea

f mas

sC

mas

sm

gg

Wns

34ndash35

Nitr

ogen

pho

spho

rus

ratio

NP

ndashei

ther

D

ndashns

35

Chl

orop

hyll

nitr

ogen

per

are

a ra

tioCh

l area

Nar

eaSPADgminus1

m2

Wns

5

Car

bon

isot

ope

disc

rimin

atio

nΔ

leaf

permilW

W

36ndash39

|Tur

gor l

oss

poin

t||π

tlp|

MPa

DD

28

Estim

ated

pho

tosy

nthe

sis

Elec

tron

tran

spor

t rat

e pe

r are

aJm

axar

eaμm

ol mminus2

sminus1

eith

erns

+W+Dminus

2ndash436ndash38

Elec

tron

tran

spor

t rat

e pe

r mas

sJm

axm

ass

nmol

gminus1

sminus1

eith

erns

++

2ndash436ndash38

Max

imum

rate

of c

arbo

xyla

tion

per a

rea

Vcm

axar

eaμm

ol mminus2

sminus1

eith

erns

+W+Dminus

2ndash436ndash38

Max

imum

rate

of c

arbo

xyla

tion

per m

ass

Vcm

axm

ass

nmol

gminus1

sminus1

eith

erns

++

2ndash436ndash38

Ratio

of i

nter

cellu

lar t

o am

bien

t CO

2 co

ncen

trat

ions

c ica

ndashei

ther

W

+

ns36

ndash38

Tim

e in

tegr

ated

leaf

CO

2 ass

imila

tion

rate

per

ar

ea Aarea

μmol

mminus2

sminus1

eith

erns

+W+Dminus

9

Tim

e in

tegr

ated

leaf

CO

2 ass

imila

tion

rate

per

m

ass

Amass

nmol

gminus1

sminus1

eith

erW

++

9

Tim

e in

tegr

ated

sto

mat

al c

ondu

ctan

cegcleaf

mm

ol mminus2

sminus1

eith

erW

+

W+Dminus

9

Tim

e in

tegr

ated

max

imum

sto

mat

al c

ondu

ct-

ance

ratio

gcleaf

g max

ndashei

ther

ns2ndash49

Max

imum

sto

mat

al c

ondu

ctan

cen

itrog

en p

er

area

ratio

g max

Nar

eam

ol gminus1

sminus1

WW

40

Plan

t siz

e

Mea

n he

ight

Hm

Wns

46ndash47

Max

imum

hei

ght

Hm

axm

WW

46ndash47

Seed

mas

sSM

mg

Dns

48ndash49

TAB

LE 2

emsp(C

ontin

ued)

(Con

tinue

s)






ait

vita

l rat

eSy

mbo

lU

nit

Hyp

othe

ses

W o

r D

high

er

W o

r D h

ighe

rH

ypot

hese

s tr

aitndash

vita

l ra

te c

orre

latio

nD

irect

ion

of tr

aitndash

vita

l ra

te c

orre

latio

nRe

fere

nce

Vita

l rat

es

Rela

tive

grow

th ra

te (d

iam

eter

incr

emen

t)RG

R dbh

cm c

mminus1

yea

rminus1

eith

erns

52041

Rela

tive

grow

th ra

te (b

iom

ass

incr

emen

t)RG

R biom

kg k

gminus1 y

earminus1

eith

erns

52041

Mor

talit

y ra

tem

p

er y

ear

eith

erD

42ndash45








TAB

LE 2

emsp(C

ontin

ued)















Δt where x is










































































120590forest






















3emsp |emspRESULTS




















Stomatal morphology



Leaf

ve

natio

n

Estim

ate

d p

ho

tosyn

the

sis

Leaf

and

woo

d ec

onom

ics

and

stru

ctur

e

Leaf composition

0

50

100

150

200

250

disGCL

GCW

SPL

e

gmax

VLAmajor

VLAminor

VLAtotal

FEV

LA

LMA

LT

LD

LDMC

SWC

WMA

PLAdry

WDNarea


Cmass

N P

Chlarea Nmass

leaf

tlp

Jmaxarea

Jmaxmass

Vcmaxarea

Vcmaxmass

c i ca

Aarea

Amass

gcleaf

gcleaf gmax

gmax Narea

RGRdbh

RGRbiom

mH

HmaxSM









0 5 10 15

m (year)

000

002

004

RG

Rdb

h( c

mcm

minus1ye

arminus1

)

(a)

r = 0550 5 10 15

m (year)

000

005

010

RG

Rbi

om(k

gkg

minus1ye

arminus1

)

(b)

r = 057


minus10

minus05

00

05

10

τ τ(mvs

RG

Rdb

h)

(c)


minus10

minus05

00

05

10

(mvs

R

GR

biom

)

1015

2025

Spe

cies

(d)



mass RGRbiom=253LMA


minor

RGRbiom=0041P097


mass m=004N167

mass plt005p lt 001

p lt 0001

100 300 500 700

d (stomatamm2)

000

002

004

RG

Rdb

h (c

m c

mminus1

year

minus1) (a)

r = 0640 2 4 6 8


(b)

r = 057

40 60 80 100 120


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (c) r = 051

50 100 150 200 250 300

LMA (gm2)

(d) r = minus 050

2 4 6 8 10 12


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (e) r = minus 056

05 10 15 20

Pmass (mgg)

(f) r = 048

40 60 80 100 120


05

1015

m (

yea

r)

(g) r = 061

5 10 15 20 25 30

Nmass (mg g)

(h) r = 050














mass


50 100 200 300

LMA (gm2)

05

1015

m (

y

ear)

(a)

r = 085r = minus 091

0 5 10 15 20


(b)

r = 085r = minus 085

05 10 15 20

Pmass (mgg)

05

1015

m (

y

ear)

(c)

r = 089r = ns

0 50 100 150

LA (cm2)

(d)

r = nsr = minus 076








10

00

05

10

(RG

Rdb

hvs

gm

axN

area) (a)

10

00

05

10

(RG

Rdb

hvs

Hm

ax)

1015

2025

Spe

cies

(b)


10

00

05

10

(mvs

d)

(c)


10

00

05

10

(mvs

Pm

ass)

1015

2025

Spe

cies

(d)







Model






d 067043 776eminus05 532






d 065040 201eminus04 519







Site ndash 665 312





plt005p lt 001 p lt 0001














000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070




ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S
































































































































































TAB

LE 2

emspSt

udy

trai

ts re

latin

g to

sto

mat

al m

orph

olog

y le

af v

enat

ion

leaf

and

woo

d ec

onom

ics

and

stru

ctur

e le

af c

ompo

sitio

n a

nd e

stim

ated

pho

tosy

nthe

sis

and

plan

t siz

e a

nd th

e vi

tal

rate

s m

easu

red

for s

peci

es fr

om a

mon

tane

wet

fore

st (W

) and

a lo

wla

nd d

ry fo

rest

(D) i

n H

awai

i F

or th

e tr

aits

we

prov

ide

sym

bols

uni

ts h

ypot

hese

s fo

r giv

en tr

aits

for d

iffer

ence

s be

twee

n fo

rest

s an

d re

sults

from

sta

tistic

al te

sts

and

hyp

othe

ses

for c

orre

latio

ns w

ith v

ital r

ates

(rel

ativ

e gr

owth

rate

and

mor

talit

y) a

nd re

sults

from

Pea

rson

s co

rrel

atio

n te

sts

(whe

n on

e re

sult

is p

rese

nted

this

repr

esen

ts s

peci

es fr

om b

oth

fore

sts

toge

ther

and

whe

n tw

o re

sults

are

pre

sent

ed th

ese

repr

esen

t spe

cies

in th

e w

et a

nd d

ry fo

rest

s se

para

tely

) an

d re

fere

nces

su

ppor

ting

the

hypo

thes

es n

s in

dica

tes

no s

igni

fican

t diff

eren

ce a

t plt005ldquoW


have

a h

ighe

r tra

it va

lue

than

the

dry

fore

st o

n av

erag

e ldquoD

rdquo tha

t the

dry

fore

st w

ould

hav

e th

e hi

gher

trai

t val

ue o

n av

erag

e a

nd ldquoe

ither

rdquo den

otes

the

exis

tenc

e of

mul

tiple

pub

lishe

d hy

poth

eses

whe

reby

eith

er M

WF

or L

DF

coul

d be

exp

ecte

d to

hav

e th

e hi

gher

trai

t val

ue (S

uppo

rtin

g In

form

atio

n Ta

ble

S10)

Pos

itive

sig

ns (+

) ind

icat

e th

e ex

pect

atio

n or

find

ing

of a

pos

itive


Tabl

e S1

0 plt005p

lt 0

01

p lt

000

1

Trai

tvi

tal r

ate

Sym

bol

Uni

tH

ypot

hese

s W

or D

hi

gher

W

or D

hig

her

Hyp

othe

ses

trai

tndashvi

tal

rate

cor

rela

tion

Dire

ctio

n of

trai

tndashvi

tal

rate

cor

rela

tion

Refe

renc

e

Stom

atal

mor

phol

ogy

Stom

atal

den

sity

dst

omat

am

m2

eith

erns

++

1ndash5

Stom

atal

diff

eren

tiatio

n ra

te (o

r ind

ex)

i-

eith

erW

+

+

2ndash6

Stom

atal

are

as

μm2

WW

+

ns157

Gua

rd c

ell l

engt

hG

CL

μmW

W

157

Gua

rd c

ell w

idth

GC

Wμm

WW

157

Pore

leng

thSP

Lμm

WW

157

Epid

erm

al p

avem

ent c

ell a

rea

eμm

2W

W

8

Max

imum

sto

mat

al c

ondu

ctan

ceg m

axm

ol mminus2

sminus1

eith

erW

+

+2ndash59ndash10

Leaf

ven

atio

n

Maj

or v

ein

dens

ityVLA

maj

orm

m p

er m

m2

eith

erD

+ns

2ndash411ndash15

Min

or v

ein

dens

ityVLA

min

orm

m p

er m

m2

eith

erD

+ns

2ndash411ndash15

Tota

l vei

n de

nsity

VLA

tota

lm

m p

er m

m2

eith

erD

+ns

2ndash411ndash15

Free

end

ing

vein

den

sity

FEV

pe

r mm

2ei

ther

D

2ndash411ndash15

Leaf

and

woo

d ec

onom

ics

and

stru

ctur

e

Leaf

are

aLA

cm2

Wns

+WnsDminus

16ndash1

8

Leaf

mas

s pe

r are

aLMA

gm

2ei

ther

nsminus

W+Dminus

12 1

8ndash23

Leaf

thic

knes

sLT

mm

eith

erns

minusminus

12 1

8ndash23

Leaf

den

sity

LDg

cm3

eith

erns

minusW+Dminus

12 1

8ndash23

Leaf

dry

mat

ter c

onte

ntLD

MC

gg

DD

minus

ns1824

Satu

rate

d w

ater

con

tent

SWC

gg

eith

erW

25ndash27

Wat

er m

ass

per a

rea

WMA

gm

2D

ns25ndash27

Perc

enta

ge lo

ss a

rea

(dry

)PLA

dry

W

W

2829

Woo

d de

nsity

WD

gcm

3ei

ther

D

minus

ns530ndash32

(Con

tinue

s)


Trai

tvi

tal r

ate

Sym

bol

Uni

tH

ypot

hese

s W

or D

hi

gher

W

or D

hig

her

Hyp

othe

ses

trai

tndashvi

tal

rate

cor

rela

tion

Dire

ctio

n of

trai

tndashvi

tal

rate

cor

rela

tion

Refe

renc

e

Leaf

com

posi

tion

Nitr

ogen

con

cent

ratio

n pe

r lea

f are

aN

area

gm

2ei

ther

ns+

W+Dminus

2ndash520

Nitr

ogen

con

cent

ratio

n pe

r lea

f mas

sN

mas

sm

gg

eith

erns

++

2ndash520

Phos

phor

us c

once

ntra

tion

per l

eaf a

rea

P area

gm

2ei

ther

ns+

W+Dminus

2ndash520

Phos

phor

us c

once

ntra

tion

per l

eaf m

ass

P mas

sm

gg

eith

erW

+

+2ndash520

Chl

orop

hyll

conc

entr

atio

nCh

l area

SPAD

eith

erns

2ndash433

Chl

orop

hyll

per m

ass

Chl m

ass

SPADgminus1

m2

eith

erns

2ndash433

Car

bon

conc

entr

atio

n pe

r lea

f mas

sC

mas

sm

gg

Wns

34ndash35

Nitr

ogen

pho

spho

rus

ratio

NP

ndashei

ther

D

ndashns

35

Chl

orop

hyll

nitr

ogen

per

are

a ra

tioCh

l area

Nar

eaSPADgminus1

m2

Wns

5

Car

bon

isot

ope

disc

rimin

atio

nΔ

leaf

permilW

W

36ndash39

|Tur

gor l

oss

poin

t||π

tlp|

MPa

DD

28

Estim

ated

pho

tosy

nthe

sis

Elec

tron

tran

spor

t rat

e pe

r are

aJm

axar

eaμm

ol mminus2

sminus1

eith

erns

+W+Dminus

2ndash436ndash38

Elec

tron

tran

spor

t rat

e pe

r mas

sJm

axm

ass

nmol

gminus1

sminus1

eith

erns

++

2ndash436ndash38

Max

imum

rate

of c

arbo

xyla

tion

per a

rea

Vcm

axar

eaμm

ol mminus2

sminus1

eith

erns

+W+Dminus

2ndash436ndash38

Max

imum

rate

of c

arbo

xyla

tion

per m

ass

Vcm

axm

ass

nmol

gminus1

sminus1

eith

erns

++

2ndash436ndash38

Ratio

of i

nter

cellu

lar t

o am

bien

t CO

2 co

ncen

trat

ions

c ica

ndashei

ther

W

+

ns36

ndash38

Tim

e in

tegr

ated

leaf

CO

2 ass

imila

tion

rate

per

ar

ea Aarea

μmol

mminus2

sminus1

eith

erns

+W+Dminus

9

Tim

e in

tegr

ated

leaf

CO

2 ass

imila

tion

rate

per

m

ass

Amass

nmol

gminus1

sminus1

eith

erW

++

9

Tim

e in

tegr

ated

sto

mat

al c

ondu

ctan

cegcleaf

mm

ol mminus2

sminus1

eith

erW

+

W+Dminus

9

Tim

e in

tegr

ated

max

imum

sto

mat

al c

ondu

ct-

ance

ratio

gcleaf

g max

ndashei

ther

ns2ndash49

Max

imum

sto

mat

al c

ondu

ctan

cen

itrog

en p

er

area

ratio

g max

Nar

eam

ol gminus1

sminus1

WW

40

Plan

t siz

e

Mea

n he

ight

Hm

Wns

46ndash47

Max

imum

hei

ght

Hm

axm

WW

46ndash47

Seed

mas

sSM

mg

Dns

48ndash49

TAB

LE 2

emsp(C

ontin

ued)

(Con

tinue

s)






ait

vita

l rat

eSy

mbo

lU

nit

Hyp

othe

ses

W o

r D

high

er

W o

r D h

ighe

rH

ypot

hese

s tr

aitndash

vita

l ra

te c

orre

latio

nD

irect

ion

of tr

aitndash

vita

l ra

te c

orre

latio

nRe

fere

nce

Vita

l rat

es

Rela

tive

grow

th ra

te (d

iam

eter

incr

emen

t)RG

R dbh

cm c

mminus1

yea

rminus1

eith

erns

52041

Rela

tive

grow

th ra

te (b

iom

ass

incr

emen

t)RG

R biom

kg k

gminus1 y

earminus1

eith

erns

52041

Mor

talit

y ra

tem

p

er y

ear

eith

erD

42ndash45








TAB

LE 2

emsp(C

ontin

ued)















Δt where x is










































































120590forest






















3emsp |emspRESULTS




















Stomatal morphology



Leaf

ve

natio

n

Estim

ate

d p

ho

tosyn

the

sis

Leaf

and

woo

d ec

onom

ics

and

stru

ctur

e

Leaf composition

0

50

100

150

200

250

disGCL

GCW

SPL

e

gmax

VLAmajor

VLAminor

VLAtotal

FEV

LA

LMA

LT

LD

LDMC

SWC

WMA

PLAdry

WDNarea


Cmass

N P

Chlarea Nmass

leaf

tlp

Jmaxarea

Jmaxmass

Vcmaxarea

Vcmaxmass

c i ca

Aarea

Amass

gcleaf

gcleaf gmax

gmax Narea

RGRdbh

RGRbiom

mH

HmaxSM









0 5 10 15

m (year)

000

002

004

RG

Rdb

h( c

mcm

minus1ye

arminus1

)

(a)

r = 0550 5 10 15

m (year)

000

005

010

RG

Rbi

om(k

gkg

minus1ye

arminus1

)

(b)

r = 057


minus10

minus05

00

05

10

τ τ(mvs

RG

Rdb

h)

(c)


minus10

minus05

00

05

10

(mvs

R

GR

biom

)

1015

2025

Spe

cies

(d)



mass RGRbiom=253LMA


minor

RGRbiom=0041P097


mass m=004N167

mass plt005p lt 001

p lt 0001

100 300 500 700

d (stomatamm2)

000

002

004

RG

Rdb

h (c

m c

mminus1

year

minus1) (a)

r = 0640 2 4 6 8


(b)

r = 057

40 60 80 100 120


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (c) r = 051

50 100 150 200 250 300

LMA (gm2)

(d) r = minus 050

2 4 6 8 10 12


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (e) r = minus 056

05 10 15 20

Pmass (mgg)

(f) r = 048

40 60 80 100 120


05

1015

m (

yea

r)

(g) r = 061

5 10 15 20 25 30

Nmass (mg g)

(h) r = 050














mass


50 100 200 300

LMA (gm2)

05

1015

m (

y

ear)

(a)

r = 085r = minus 091

0 5 10 15 20


(b)

r = 085r = minus 085

05 10 15 20

Pmass (mgg)

05

1015

m (

y

ear)

(c)

r = 089r = ns

0 50 100 150

LA (cm2)

(d)

r = nsr = minus 076








10

00

05

10

(RG

Rdb

hvs

gm

axN

area) (a)

10

00

05

10

(RG

Rdb

hvs

Hm

ax)

1015

2025

Spe

cies

(b)


10

00

05

10

(mvs

d)

(c)


10

00

05

10

(mvs

Pm

ass)

1015

2025

Spe

cies

(d)







Model






d 067043 776eminus05 532






d 065040 201eminus04 519







Site ndash 665 312





plt005p lt 001 p lt 0001














000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070




ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S
































































































































































Trai

tvi

tal r

ate

Sym

bol

Uni

tH

ypot

hese

s W

or D

hi

gher

W

or D

hig

her

Hyp

othe

ses

trai

tndashvi

tal

rate

cor

rela

tion

Dire

ctio

n of

trai

tndashvi

tal

rate

cor

rela

tion

Refe

renc

e

Leaf

com

posi

tion

Nitr

ogen

con

cent

ratio

n pe

r lea

f are

aN

area

gm

2ei

ther

ns+

W+Dminus

2ndash520

Nitr

ogen

con

cent

ratio

n pe

r lea

f mas

sN

mas

sm

gg

eith

erns

++

2ndash520

Phos

phor

us c

once

ntra

tion

per l

eaf a

rea

P area

gm

2ei

ther

ns+

W+Dminus

2ndash520

Phos

phor

us c

once

ntra

tion

per l

eaf m

ass

P mas

sm

gg

eith

erW

+

+2ndash520

Chl

orop

hyll

conc

entr

atio

nCh

l area

SPAD

eith

erns

2ndash433

Chl

orop

hyll

per m

ass

Chl m

ass

SPADgminus1

m2

eith

erns

2ndash433

Car

bon

conc

entr

atio

n pe

r lea

f mas

sC

mas

sm

gg

Wns

34ndash35

Nitr

ogen

pho

spho

rus

ratio

NP

ndashei

ther

D

ndashns

35

Chl

orop

hyll

nitr

ogen

per

are

a ra

tioCh

l area

Nar

eaSPADgminus1

m2

Wns

5

Car

bon

isot

ope

disc

rimin

atio

nΔ

leaf

permilW

W

36ndash39

|Tur

gor l

oss

poin

t||π

tlp|

MPa

DD

28

Estim

ated

pho

tosy

nthe

sis

Elec

tron

tran

spor

t rat

e pe

r are

aJm

axar

eaμm

ol mminus2

sminus1

eith

erns

+W+Dminus

2ndash436ndash38

Elec

tron

tran

spor

t rat

e pe

r mas

sJm

axm

ass

nmol

gminus1

sminus1

eith

erns

++

2ndash436ndash38

Max

imum

rate

of c

arbo

xyla

tion

per a

rea

Vcm

axar

eaμm

ol mminus2

sminus1

eith

erns

+W+Dminus

2ndash436ndash38

Max

imum

rate

of c

arbo

xyla

tion

per m

ass

Vcm

axm

ass

nmol

gminus1

sminus1

eith

erns

++

2ndash436ndash38

Ratio

of i

nter

cellu

lar t

o am

bien

t CO

2 co

ncen

trat

ions

c ica

ndashei

ther

W

+

ns36

ndash38

Tim

e in

tegr

ated

leaf

CO

2 ass

imila

tion

rate

per

ar

ea Aarea

μmol

mminus2

sminus1

eith

erns

+W+Dminus

9

Tim

e in

tegr

ated

leaf

CO

2 ass

imila

tion

rate

per

m

ass

Amass

nmol

gminus1

sminus1

eith

erW

++

9

Tim

e in

tegr

ated

sto

mat

al c

ondu

ctan

cegcleaf

mm

ol mminus2

sminus1

eith

erW

+

W+Dminus

9

Tim

e in

tegr

ated

max

imum

sto

mat

al c

ondu

ct-

ance

ratio

gcleaf

g max

ndashei

ther

ns2ndash49

Max

imum

sto

mat

al c

ondu

ctan

cen

itrog

en p

er

area

ratio

g max

Nar

eam

ol gminus1

sminus1

WW

40

Plan

t siz

e

Mea

n he

ight

Hm

Wns

46ndash47

Max

imum

hei

ght

Hm

axm

WW

46ndash47

Seed

mas

sSM

mg

Dns

48ndash49

TAB

LE 2

emsp(C

ontin

ued)

(Con

tinue

s)






ait

vita

l rat

eSy

mbo

lU

nit

Hyp

othe

ses

W o

r D

high

er

W o

r D h

ighe

rH

ypot

hese

s tr

aitndash

vita

l ra

te c

orre

latio

nD

irect

ion

of tr

aitndash

vita

l ra

te c

orre

latio

nRe

fere

nce

Vita

l rat

es

Rela

tive

grow

th ra

te (d

iam

eter

incr

emen

t)RG

R dbh

cm c

mminus1

yea

rminus1

eith

erns

52041

Rela

tive

grow

th ra

te (b

iom

ass

incr

emen

t)RG

R biom

kg k

gminus1 y

earminus1

eith

erns

52041

Mor

talit

y ra

tem

p

er y

ear

eith

erD

42ndash45








TAB

LE 2

emsp(C

ontin

ued)















Δt where x is










































































120590forest






















3emsp |emspRESULTS




















Stomatal morphology



Leaf

ve

natio

n

Estim

ate

d p

ho

tosyn

the

sis

Leaf

and

woo

d ec

onom

ics

and

stru

ctur

e

Leaf composition

0

50

100

150

200

250

disGCL

GCW

SPL

e

gmax

VLAmajor

VLAminor

VLAtotal

FEV

LA

LMA

LT

LD

LDMC

SWC

WMA

PLAdry

WDNarea


Cmass

N P

Chlarea Nmass

leaf

tlp

Jmaxarea

Jmaxmass

Vcmaxarea

Vcmaxmass

c i ca

Aarea

Amass

gcleaf

gcleaf gmax

gmax Narea

RGRdbh

RGRbiom

mH

HmaxSM









0 5 10 15

m (year)

000

002

004

RG

Rdb

h( c

mcm

minus1ye

arminus1

)

(a)

r = 0550 5 10 15

m (year)

000

005

010

RG

Rbi

om(k

gkg

minus1ye

arminus1

)

(b)

r = 057


minus10

minus05

00

05

10

τ τ(mvs

RG

Rdb

h)

(c)


minus10

minus05

00

05

10

(mvs

R

GR

biom

)

1015

2025

Spe

cies

(d)



mass RGRbiom=253LMA


minor

RGRbiom=0041P097


mass m=004N167

mass plt005p lt 001

p lt 0001

100 300 500 700

d (stomatamm2)

000

002

004

RG

Rdb

h (c

m c

mminus1

year

minus1) (a)

r = 0640 2 4 6 8


(b)

r = 057

40 60 80 100 120


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (c) r = 051

50 100 150 200 250 300

LMA (gm2)

(d) r = minus 050

2 4 6 8 10 12


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (e) r = minus 056

05 10 15 20

Pmass (mgg)

(f) r = 048

40 60 80 100 120


05

1015

m (

yea

r)

(g) r = 061

5 10 15 20 25 30

Nmass (mg g)

(h) r = 050














mass


50 100 200 300

LMA (gm2)

05

1015

m (

y

ear)

(a)

r = 085r = minus 091

0 5 10 15 20


(b)

r = 085r = minus 085

05 10 15 20

Pmass (mgg)

05

1015

m (

y

ear)

(c)

r = 089r = ns

0 50 100 150

LA (cm2)

(d)

r = nsr = minus 076








10

00

05

10

(RG

Rdb

hvs

gm

axN

area) (a)

10

00

05

10

(RG

Rdb

hvs

Hm

ax)

1015

2025

Spe

cies

(b)


10

00

05

10

(mvs

d)

(c)


10

00

05

10

(mvs

Pm

ass)

1015

2025

Spe

cies

(d)







Model






d 067043 776eminus05 532






d 065040 201eminus04 519







Site ndash 665 312





plt005p lt 001 p lt 0001














000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070




ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S




































































































































































ait

vita

l rat

eSy

mbo

lU

nit

Hyp

othe

ses

W o

r D

high

er

W o

r D h

ighe

rH

ypot

hese

s tr

aitndash

vita

l ra

te c

orre

latio

nD

irect

ion

of tr

aitndash

vita

l ra

te c

orre

latio

nRe

fere

nce

Vita

l rat

es

Rela

tive

grow

th ra

te (d

iam

eter

incr

emen

t)RG

R dbh

cm c

mminus1

yea

rminus1

eith

erns

52041

Rela

tive

grow

th ra

te (b

iom

ass

incr

emen

t)RG

R biom

kg k

gminus1 y

earminus1

eith

erns

52041

Mor

talit

y ra

tem

p

er y

ear

eith

erD

42ndash45








TAB

LE 2

emsp(C

ontin

ued)















Δt where x is










































































120590forest






















3emsp |emspRESULTS




















Stomatal morphology



Leaf

ve

natio

n

Estim

ate

d p

ho

tosyn

the

sis

Leaf

and

woo

d ec

onom

ics

and

stru

ctur

e

Leaf composition

0

50

100

150

200

250

disGCL

GCW

SPL

e

gmax

VLAmajor

VLAminor

VLAtotal

FEV

LA

LMA

LT

LD

LDMC

SWC

WMA

PLAdry

WDNarea


Cmass

N P

Chlarea Nmass

leaf

tlp

Jmaxarea

Jmaxmass

Vcmaxarea

Vcmaxmass

c i ca

Aarea

Amass

gcleaf

gcleaf gmax

gmax Narea

RGRdbh

RGRbiom

mH

HmaxSM









0 5 10 15

m (year)

000

002

004

RG

Rdb

h( c

mcm

minus1ye

arminus1

)

(a)

r = 0550 5 10 15

m (year)

000

005

010

RG

Rbi

om(k

gkg

minus1ye

arminus1

)

(b)

r = 057


minus10

minus05

00

05

10

τ τ(mvs

RG

Rdb

h)

(c)


minus10

minus05

00

05

10

(mvs

R

GR

biom

)

1015

2025

Spe

cies

(d)



mass RGRbiom=253LMA


minor

RGRbiom=0041P097


mass m=004N167

mass plt005p lt 001

p lt 0001

100 300 500 700

d (stomatamm2)

000

002

004

RG

Rdb

h (c

m c

mminus1

year

minus1) (a)

r = 0640 2 4 6 8


(b)

r = 057

40 60 80 100 120


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (c) r = 051

50 100 150 200 250 300

LMA (gm2)

(d) r = minus 050

2 4 6 8 10 12


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (e) r = minus 056

05 10 15 20

Pmass (mgg)

(f) r = 048

40 60 80 100 120


05

1015

m (

yea

r)

(g) r = 061

5 10 15 20 25 30

Nmass (mg g)

(h) r = 050














mass


50 100 200 300

LMA (gm2)

05

1015

m (

y

ear)

(a)

r = 085r = minus 091

0 5 10 15 20


(b)

r = 085r = minus 085

05 10 15 20

Pmass (mgg)

05

1015

m (

y

ear)

(c)

r = 089r = ns

0 50 100 150

LA (cm2)

(d)

r = nsr = minus 076








10

00

05

10

(RG

Rdb

hvs

gm

axN

area) (a)

10

00

05

10

(RG

Rdb

hvs

Hm

ax)

1015

2025

Spe

cies

(b)


10

00

05

10

(mvs

d)

(c)


10

00

05

10

(mvs

Pm

ass)

1015

2025

Spe

cies

(d)







Model






d 067043 776eminus05 532






d 065040 201eminus04 519







Site ndash 665 312





plt005p lt 001 p lt 0001














000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070




ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S













































































































































































Δt where x is










































































120590forest






















3emsp |emspRESULTS




















Stomatal morphology



Leaf

ve

natio

n

Estim

ate

d p

ho

tosyn

the

sis

Leaf

and

woo

d ec

onom

ics

and

stru

ctur

e

Leaf composition

0

50

100

150

200

250

disGCL

GCW

SPL

e

gmax

VLAmajor

VLAminor

VLAtotal

FEV

LA

LMA

LT

LD

LDMC

SWC

WMA

PLAdry

WDNarea


Cmass

N P

Chlarea Nmass

leaf

tlp

Jmaxarea

Jmaxmass

Vcmaxarea

Vcmaxmass

c i ca

Aarea

Amass

gcleaf

gcleaf gmax

gmax Narea

RGRdbh

RGRbiom

mH

HmaxSM









0 5 10 15

m (year)

000

002

004

RG

Rdb

h( c

mcm

minus1ye

arminus1

)

(a)

r = 0550 5 10 15

m (year)

000

005

010

RG

Rbi

om(k

gkg

minus1ye

arminus1

)

(b)

r = 057


minus10

minus05

00

05

10

τ τ(mvs

RG

Rdb

h)

(c)


minus10

minus05

00

05

10

(mvs

R

GR

biom

)

1015

2025

Spe

cies

(d)



mass RGRbiom=253LMA


minor

RGRbiom=0041P097


mass m=004N167

mass plt005p lt 001

p lt 0001

100 300 500 700

d (stomatamm2)

000

002

004

RG

Rdb

h (c

m c

mminus1

year

minus1) (a)

r = 0640 2 4 6 8


(b)

r = 057

40 60 80 100 120


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (c) r = 051

50 100 150 200 250 300

LMA (gm2)

(d) r = minus 050

2 4 6 8 10 12


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (e) r = minus 056

05 10 15 20

Pmass (mgg)

(f) r = 048

40 60 80 100 120


05

1015

m (

yea

r)

(g) r = 061

5 10 15 20 25 30

Nmass (mg g)

(h) r = 050














mass


50 100 200 300

LMA (gm2)

05

1015

m (

y

ear)

(a)

r = 085r = minus 091

0 5 10 15 20


(b)

r = 085r = minus 085

05 10 15 20

Pmass (mgg)

05

1015

m (

y

ear)

(c)

r = 089r = ns

0 50 100 150

LA (cm2)

(d)

r = nsr = minus 076








10

00

05

10

(RG

Rdb

hvs

gm

axN

area) (a)

10

00

05

10

(RG

Rdb

hvs

Hm

ax)

1015

2025

Spe

cies

(b)


10

00

05

10

(mvs

d)

(c)


10

00

05

10

(mvs

Pm

ass)

1015

2025

Spe

cies

(d)







Model






d 067043 776eminus05 532






d 065040 201eminus04 519







Site ndash 665 312





plt005p lt 001 p lt 0001














000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070




ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S


































































































































































































































120590forest






















3emsp |emspRESULTS




















Stomatal morphology



Leaf

ve

natio

n

Estim

ate

d p

ho

tosyn

the

sis

Leaf

and

woo

d ec

onom

ics

and

stru

ctur

e

Leaf composition

0

50

100

150

200

250

disGCL

GCW

SPL

e

gmax

VLAmajor

VLAminor

VLAtotal

FEV

LA

LMA

LT

LD

LDMC

SWC

WMA

PLAdry

WDNarea


Cmass

N P

Chlarea Nmass

leaf

tlp

Jmaxarea

Jmaxmass

Vcmaxarea

Vcmaxmass

c i ca

Aarea

Amass

gcleaf

gcleaf gmax

gmax Narea

RGRdbh

RGRbiom

mH

HmaxSM









0 5 10 15

m (year)

000

002

004

RG

Rdb

h( c

mcm

minus1ye

arminus1

)

(a)

r = 0550 5 10 15

m (year)

000

005

010

RG

Rbi

om(k

gkg

minus1ye

arminus1

)

(b)

r = 057


minus10

minus05

00

05

10

τ τ(mvs

RG

Rdb

h)

(c)


minus10

minus05

00

05

10

(mvs

R

GR

biom

)

1015

2025

Spe

cies

(d)



mass RGRbiom=253LMA


minor

RGRbiom=0041P097


mass m=004N167

mass plt005p lt 001

p lt 0001

100 300 500 700

d (stomatamm2)

000

002

004

RG

Rdb

h (c

m c

mminus1

year

minus1) (a)

r = 0640 2 4 6 8


(b)

r = 057

40 60 80 100 120


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (c) r = 051

50 100 150 200 250 300

LMA (gm2)

(d) r = minus 050

2 4 6 8 10 12


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (e) r = minus 056

05 10 15 20

Pmass (mgg)

(f) r = 048

40 60 80 100 120


05

1015

m (

yea

r)

(g) r = 061

5 10 15 20 25 30

Nmass (mg g)

(h) r = 050














mass


50 100 200 300

LMA (gm2)

05

1015

m (

y

ear)

(a)

r = 085r = minus 091

0 5 10 15 20


(b)

r = 085r = minus 085

05 10 15 20

Pmass (mgg)

05

1015

m (

y

ear)

(c)

r = 089r = ns

0 50 100 150

LA (cm2)

(d)

r = nsr = minus 076








10

00

05

10

(RG

Rdb

hvs

gm

axN

area) (a)

10

00

05

10

(RG

Rdb

hvs

Hm

ax)

1015

2025

Spe

cies

(b)


10

00

05

10

(mvs

d)

(c)


10

00

05

10

(mvs

Pm

ass)

1015

2025

Spe

cies

(d)







Model






d 067043 776eminus05 532






d 065040 201eminus04 519







Site ndash 665 312





plt005p lt 001 p lt 0001














000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070




ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S




































































































































































3emsp |emspRESULTS




















Stomatal morphology



Leaf

ve

natio

n

Estim

ate

d p

ho

tosyn

the

sis

Leaf

and

woo

d ec

onom

ics

and

stru

ctur

e

Leaf composition

0

50

100

150

200

250

disGCL

GCW

SPL

e

gmax

VLAmajor

VLAminor

VLAtotal

FEV

LA

LMA

LT

LD

LDMC

SWC

WMA

PLAdry

WDNarea


Cmass

N P

Chlarea Nmass

leaf

tlp

Jmaxarea

Jmaxmass

Vcmaxarea

Vcmaxmass

c i ca

Aarea

Amass

gcleaf

gcleaf gmax

gmax Narea

RGRdbh

RGRbiom

mH

HmaxSM









0 5 10 15

m (year)

000

002

004

RG

Rdb

h( c

mcm

minus1ye

arminus1

)

(a)

r = 0550 5 10 15

m (year)

000

005

010

RG

Rbi

om(k

gkg

minus1ye

arminus1

)

(b)

r = 057


minus10

minus05

00

05

10

τ τ(mvs

RG

Rdb

h)

(c)


minus10

minus05

00

05

10

(mvs

R

GR

biom

)

1015

2025

Spe

cies

(d)



mass RGRbiom=253LMA


minor

RGRbiom=0041P097


mass m=004N167

mass plt005p lt 001

p lt 0001

100 300 500 700

d (stomatamm2)

000

002

004

RG

Rdb

h (c

m c

mminus1

year

minus1) (a)

r = 0640 2 4 6 8


(b)

r = 057

40 60 80 100 120


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (c) r = 051

50 100 150 200 250 300

LMA (gm2)

(d) r = minus 050

2 4 6 8 10 12


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (e) r = minus 056

05 10 15 20

Pmass (mgg)

(f) r = 048

40 60 80 100 120


05

1015

m (

yea

r)

(g) r = 061

5 10 15 20 25 30

Nmass (mg g)

(h) r = 050














mass


50 100 200 300

LMA (gm2)

05

1015

m (

y

ear)

(a)

r = 085r = minus 091

0 5 10 15 20


(b)

r = 085r = minus 085

05 10 15 20

Pmass (mgg)

05

1015

m (

y

ear)

(c)

r = 089r = ns

0 50 100 150

LA (cm2)

(d)

r = nsr = minus 076








10

00

05

10

(RG

Rdb

hvs

gm

axN

area) (a)

10

00

05

10

(RG

Rdb

hvs

Hm

ax)

1015

2025

Spe

cies

(b)


10

00

05

10

(mvs

d)

(c)


10

00

05

10

(mvs

Pm

ass)

1015

2025

Spe

cies

(d)







Model






d 067043 776eminus05 532






d 065040 201eminus04 519







Site ndash 665 312





plt005p lt 001 p lt 0001














000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070




ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S





































































































































































Stomatal morphology



Leaf

ve

natio

n

Estim

ate

d p

ho

tosyn

the

sis

Leaf

and

woo

d ec

onom

ics

and

stru

ctur

e

Leaf composition

0

50

100

150

200

250

disGCL

GCW

SPL

e

gmax

VLAmajor

VLAminor

VLAtotal

FEV

LA

LMA

LT

LD

LDMC

SWC

WMA

PLAdry

WDNarea


Cmass

N P

Chlarea Nmass

leaf

tlp

Jmaxarea

Jmaxmass

Vcmaxarea

Vcmaxmass

c i ca

Aarea

Amass

gcleaf

gcleaf gmax

gmax Narea

RGRdbh

RGRbiom

mH

HmaxSM









0 5 10 15

m (year)

000

002

004

RG

Rdb

h( c

mcm

minus1ye

arminus1

)

(a)

r = 0550 5 10 15

m (year)

000

005

010

RG

Rbi

om(k

gkg

minus1ye

arminus1

)

(b)

r = 057


minus10

minus05

00

05

10

τ τ(mvs

RG

Rdb

h)

(c)


minus10

minus05

00

05

10

(mvs

R

GR

biom

)

1015

2025

Spe

cies

(d)



mass RGRbiom=253LMA


minor

RGRbiom=0041P097


mass m=004N167

mass plt005p lt 001

p lt 0001

100 300 500 700

d (stomatamm2)

000

002

004

RG

Rdb

h (c

m c

mminus1

year

minus1) (a)

r = 0640 2 4 6 8


(b)

r = 057

40 60 80 100 120


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (c) r = 051

50 100 150 200 250 300

LMA (gm2)

(d) r = minus 050

2 4 6 8 10 12


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (e) r = minus 056

05 10 15 20

Pmass (mgg)

(f) r = 048

40 60 80 100 120


05

1015

m (

yea

r)

(g) r = 061

5 10 15 20 25 30

Nmass (mg g)

(h) r = 050














mass


50 100 200 300

LMA (gm2)

05

1015

m (

y

ear)

(a)

r = 085r = minus 091

0 5 10 15 20


(b)

r = 085r = minus 085

05 10 15 20

Pmass (mgg)

05

1015

m (

y

ear)

(c)

r = 089r = ns

0 50 100 150

LA (cm2)

(d)

r = nsr = minus 076








10

00

05

10

(RG

Rdb

hvs

gm

axN

area) (a)

10

00

05

10

(RG

Rdb

hvs

Hm

ax)

1015

2025

Spe

cies

(b)


10

00

05

10

(mvs

d)

(c)


10

00

05

10

(mvs

Pm

ass)

1015

2025

Spe

cies

(d)







Model






d 067043 776eminus05 532






d 065040 201eminus04 519







Site ndash 665 312





plt005p lt 001 p lt 0001














000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070




ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S







































































































































































0 5 10 15

m (year)

000

002

004

RG

Rdb

h( c

mcm

minus1ye

arminus1

)

(a)

r = 0550 5 10 15

m (year)

000

005

010

RG

Rbi

om(k

gkg

minus1ye

arminus1

)

(b)

r = 057


minus10

minus05

00

05

10

τ τ(mvs

RG

Rdb

h)

(c)


minus10

minus05

00

05

10

(mvs

R

GR

biom

)

1015

2025

Spe

cies

(d)



mass RGRbiom=253LMA


minor

RGRbiom=0041P097


mass m=004N167

mass plt005p lt 001

p lt 0001

100 300 500 700

d (stomatamm2)

000

002

004

RG

Rdb

h (c

m c

mminus1

year

minus1) (a)

r = 0640 2 4 6 8


(b)

r = 057

40 60 80 100 120


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (c) r = 051

50 100 150 200 250 300

LMA (gm2)

(d) r = minus 050

2 4 6 8 10 12


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (e) r = minus 056

05 10 15 20

Pmass (mgg)

(f) r = 048

40 60 80 100 120


05

1015

m (

yea

r)

(g) r = 061

5 10 15 20 25 30

Nmass (mg g)

(h) r = 050














mass


50 100 200 300

LMA (gm2)

05

1015

m (

y

ear)

(a)

r = 085r = minus 091

0 5 10 15 20


(b)

r = 085r = minus 085

05 10 15 20

Pmass (mgg)

05

1015

m (

y

ear)

(c)

r = 089r = ns

0 50 100 150

LA (cm2)

(d)

r = nsr = minus 076








10

00

05

10

(RG

Rdb

hvs

gm

axN

area) (a)

10

00

05

10

(RG

Rdb

hvs

Hm

ax)

1015

2025

Spe

cies

(b)


10

00

05

10

(mvs

d)

(c)


10

00

05

10

(mvs

Pm

ass)

1015

2025

Spe

cies

(d)







Model






d 067043 776eminus05 532






d 065040 201eminus04 519







Site ndash 665 312





plt005p lt 001 p lt 0001














000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070




ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S

































































































































































mass RGRbiom=253LMA


minor

RGRbiom=0041P097


mass m=004N167

mass plt005p lt 001

p lt 0001

100 300 500 700

d (stomatamm2)

000

002

004

RG

Rdb

h (c

m c

mminus1

year

minus1) (a)

r = 0640 2 4 6 8


(b)

r = 057

40 60 80 100 120


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (c) r = 051

50 100 150 200 250 300

LMA (gm2)

(d) r = minus 050

2 4 6 8 10 12


000

005

010

RG

Rbi

om (k

g kg

minus1ye

arminus1

) (e) r = minus 056

05 10 15 20

Pmass (mgg)

(f) r = 048

40 60 80 100 120


05

1015

m (

yea

r)

(g) r = 061

5 10 15 20 25 30

Nmass (mg g)

(h) r = 050














mass


50 100 200 300

LMA (gm2)

05

1015

m (

y

ear)

(a)

r = 085r = minus 091

0 5 10 15 20


(b)

r = 085r = minus 085

05 10 15 20

Pmass (mgg)

05

1015

m (

y

ear)

(c)

r = 089r = ns

0 50 100 150

LA (cm2)

(d)

r = nsr = minus 076








10

00

05

10

(RG

Rdb

hvs

gm

axN

area) (a)

10

00

05

10

(RG

Rdb

hvs

Hm

ax)

1015

2025

Spe

cies

(b)


10

00

05

10

(mvs

d)

(c)


10

00

05

10

(mvs

Pm

ass)

1015

2025

Spe

cies

(d)







Model






d 067043 776eminus05 532






d 065040 201eminus04 519







Site ndash 665 312





plt005p lt 001 p lt 0001














000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070




ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S












































































































































































mass


50 100 200 300

LMA (gm2)

05

1015

m (

y

ear)

(a)

r = 085r = minus 091

0 5 10 15 20


(b)

r = 085r = minus 085

05 10 15 20

Pmass (mgg)

05

1015

m (

y

ear)

(c)

r = 089r = ns

0 50 100 150

LA (cm2)

(d)

r = nsr = minus 076








10

00

05

10

(RG

Rdb

hvs

gm

axN

area) (a)

10

00

05

10

(RG

Rdb

hvs

Hm

ax)

1015

2025

Spe

cies

(b)


10

00

05

10

(mvs

d)

(c)


10

00

05

10

(mvs

Pm

ass)

1015

2025

Spe

cies

(d)







Model






d 067043 776eminus05 532






d 065040 201eminus04 519







Site ndash 665 312





plt005p lt 001 p lt 0001














000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070




ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S






































































































































































10

00

05

10

(RG

Rdb

hvs

gm

axN

area) (a)

10

00

05

10

(RG

Rdb

hvs

Hm

ax)

1015

2025

Spe

cies

(b)


10

00

05

10

(mvs

d)

(c)


10

00

05

10

(mvs

Pm

ass)

1015

2025

Spe

cies

(d)







Model






d 067043 776eminus05 532






d 065040 201eminus04 519







Site ndash 665 312





plt005p lt 001 p lt 0001














000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070




ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S





































































































































































Model






d 067043 776eminus05 532






d 065040 201eminus04 519







Site ndash 665 312





plt005p lt 001 p lt 0001














000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070




ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S











































































































































































000 002 004

RGRdbh observed

000

002

004

RG

Rdb

hpr

edic

ted

(a)

R2 = 072

000 005 010

RGRbiom observed

000

005

010

RG

Rbi

ompr

edic

ted

(b)

R2 = 070

0 5 10 15

mobserved

05

1015

mpr

edic

ted

(c)

R2 = 070




ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S


































































































































































ACKNOWLEDG EMENTS






ORCID







R E FE R E N C E S































































































































































An extensive suite of functional ... - hippnet.hawaii.edu · Resources and Environmental...

Documents

Transcript of An extensive suite of functional ... - hippnet.hawaii.edu · Resources and Environmental...