forest ecology pervasive shifts in forest dynamics in a ......review summary forest ecology...

REVIEW SUMMARY

FOREST ECOLOGY

Pervasive shifts in forest dynamicsin a changing worldNate G McDowell Craig D Allen Kristina Anderson-Teixeira Brian H Aukema Ben Bond-LambertyLouise Chini James S Clark Michael Dietze Charlotte Grossiord Adam Hanbury-BrownGeorge C Hurtt Robert B Jackson Daniel J Johnson Lara Kueppers Jeremy W LichsteinKiona Ogle Benjamin Poulter Thomas A M Pugh Rupert Seidl Monica G Turner Maria UriarteAnthony P Walker Chonggang Xu

BACKGROUND Forest dynamics arise from theinterplay of chronic drivers and transient dis-turbances with the demographic processes ofrecruitment growth andmortality The result-ing trajectories of vegetation development drivethe biomass and species composition of terres-trial ecosystems Forest dynamics are changingbecause of anthropogenic-driven exacerbationof chronic drivers such as rising temperatureandCO2 and increasing transient disturbancesincluding wildfire drought windthrow bioticattack and land-use change There are wide-spread observations of increasing tree mor-tality due to changing climate and land useas well as observations of growth stimulationof younger forests due to CO2 fertilizationThese antagonistic processes are co-occurringglobally leaving the fate of future forests un-certain We examine the implications of chang-ing forest demography and its drivers for bothfuture forest management and forecasting im-pacts of global climate forcing

ADVANCES We reviewed the literature of for-est demographic responses to chronic driversand transient disturbances to generate hy-

potheses on future trajectories of these factorsand their subsequent impacts on vegetationdynamics with a focus on forested ecosystemsWe complemented this reviewwith analyses ofglobal land-use change and disturbance data-sets to independently evaluate the implicationsof changing drivers and disturbances on global-scale tree demographics Ongoing changes inenvironmental drivers anddisturbance regimesare consistently increasing mortality and forc-ing forests toward shorter-statured and youn-ger stands reducing potential carbon storageAcclimation adaptation and migration maypartiallymitigate these effects These increasedforest impacts are due to natural disturbances(eg wildfire drought windthrow insect orpathogenoutbreaks) and land-use change bothof which are predicted to increase in magni-tude in the future Atmospherically derivedestimates of the terrestrial carbon sink andremote sensing data indicate that tree growthand potentially recruitment may have in-creased globally in the 20th century but thegrowth of this carbon sink has slowed Vari-ability in growth stimulation due to CO2 fer-tilization is evident globally with observations

and experiments suggesting that forests ben-efit from CO2 primarily in early stages of sec-ondary succession Furthermore increased treegrowth typically requires sufficient water andnutrients to take advantage of rising CO2Collectively the evidence reveals that it ishighly likely that tree mortality rates willcontinue to increase whereas recruitment andgrowth will respond to changing drivers in aspatially and temporally variable manner Thenet impact will be a reduction in forest canopycover and biomass

OUTLOOK Pervasive shifts in forest vegetationdynamics are already occurring and are likelyto accelerate under future global changeswith consequences for biodiversity and cli-mate forcing This conclusion is robust withrespect to the abundant literature evidence andour global assessment of historical demographicchanges but it also forms the basis for hypothe-

ses regarding the patternsand processes underlyingthe shifts in forest dynam-ics These hypotheses willbe directly testable usingemerging terrestrial andsatellite-based observation

networks The existing evidence and newlymade observations provide a critical test ofEarth systemmodels that continue to improvein their ability to simulate forest dynamics andresulting climate forcing Ultimately forestmanagers and natural resource policies mustconfront the consequences of changing climateand disturbance regimes to ensure sustainableforests and accrue their associated benefits

RESEARCH

McDowell et al Science 368 964 (2020) 29 May 2020 1 of 1

The list of author affiliations is available in the full article onlineCorresponding author Email natemcdowellpnnlgovCite this article as N G McDowell et al Science 368eaaz9463 (2020) DOI 101126scienceaaz9463

Old Disturbed Recovering Old

Novel shorter ecosystemRecruitment and growth dominateDemographic drivers Drought LUC wildfire wind insect outbreaks

Driv

ers

CO

2 te

mpe

ratu

re V

PD

A conceptual diagram of the components of forest dynamics and thedisturbances that drive them In the far-left panel a mature ecosystem isresponsive primarily to localized mortality and the primary drivers of demographyare chronically changing variables such as CO2 temperature and vapor pressuredeficit (VPD) In the next panel the system is disturbed by fire insect outbreakor another large-scale perturbation that removes most of the overstory trees

and species adapted to rapid postdisturbance recruitment become establishedIn the third panel recruitment and growth dominate demographic processeswith mortality increasing over time as competition leads to self-thinningIn the last panel a mature ecosystem is dominated by species that have replacedthe original community in response to chronic environmental changes leading toa novel ecosystem

ON OUR WEBSITE

Read the full articleat httpsdxdoiorg101126scienceaaz9463

on May 28 2020

httpsciencesciencem

agorgD

ownloaded from

REVIEW

FOREST ECOLOGY

Pervasive shifts in forest dynamicsin a changing worldNate G McDowell1 Craig D Allen2 Kristina Anderson-Teixeira34 Brian H Aukema5Ben Bond-Lamberty6 Louise Chini7 James S Clark8 Michael Dietze9 Charlotte Grossiord10Adam Hanbury-Brown11 George C Hurtt7 Robert B Jackson12 Daniel J Johnson13Lara Kueppers1114 Jeremy W Lichstein15 Kiona Ogle16 Benjamin Poulter17 Thomas A M Pugh1819Rupert Seidl2021 Monica G Turner22 Maria Uriarte23 Anthony P Walker24 Chonggang Xu25

Forest dynamics arise from the interplay of environmental drivers and disturbances with the demographicprocesses of recruitment growth and mortality subsequently driving biomass and species compositionHowever forest disturbances and subsequent recovery are shifting with global changes in climate and landuse altering these dynamics Changes in environmental drivers land use and disturbance regimes areforcing forests toward younger shorter stands Rising carbon dioxide acclimation adaptation andmigrationcan influence these impacts Recent developments in Earth system models support increasingly realisticsimulations of vegetation dynamics In parallel emerging remote sensing datasets promise qualitatively newand more abundant data on the underlying processes and consequences for vegetation structure Whencombined these advances hold promise for improving the scientific understanding of changes in vegetationdemographics and disturbances

The interplay of vegetation demographymdashrecruitment growth andmortalitymdashwithenvironmental conditions and distur-bances drives forest dynamics of bio-mass function and species composition

(see Box 1 for definitions) In old-growth for-ests that approximate steady-state demograph-ics the recruitment growth and mortality oftrees are approximately balanced in contrastrapid recruitment often follows widespreaddisturbance-inducedmortality (1) Vegetationdynamics may now be changing because theenvironmental context in which plant demog-raphy and disturbances interact is shiftingwith anthropogenic change The interactionbetween episodic forest disturbances such aswindthrow or wildfire and chronically chang-ing drivers such as rising temperature vaporpressure deficit (VPD) and CO2 together withland-use change (LUC) (2) leads to both com-pounding and antagonistic impacts that alterdemographic rates (3) with consequences forterrestrial biogeochemical cycles and climate(4 5) Understanding the drivers of vegetation

dynamics is thus critical for accurate predic-tion of global terrestrial biogeochemistry underfuture conditions (6)The impacts of global change on forest

demographic rates may already be materializ-ing Inmature ecosystems tree mortality rateshave doubled throughoutmuch of theAmericasand in Europe over the past four decades (7ndash9)Simultaneously global carbon budgets indicateeither a growing or constant terrestrial carbonsink (10ndash12) which implies increased or con-stant vegetation production rates (13 14) How-ever satellite evidence suggests that forestsmight be switching from a CO2 fertilizationndashdominated period to a VPDndashdominated period(15) Terrestrial greening indices indicate a shiftfrom a CO2-driven increase in greenness inthe late 20th century to aVPD-drivendecrease inthe past decade (16) Thus increasing mortalitydue to anthropogenic changes and potentiallyincreasing or stable growth and recruitmentdue to CO2 fertilization (5) represent opposingprocesses that are co-occurring globally leav-ing the fate of future forests uncertain

In addition to changing vegetation dynam-ics in intact or relatively undisturbed forestsepisodic disturbances are tending to be largermore severe and in some regions more fre-quent under global climate change (17ndash20)Similarly the rates and types of LUC varywidely (21) but have on average increasedglobally in the past few centuries (2 22 23)Thus at the global scale disturbances and LUChave likely amplified tree mortality beyondwhat is suggested by the doubling of back-ground mortality rates in undisturbed forests(7ndash9) Current understanding of the net balanceof tree losses (mortality) and gains (recruitmentand growth) under a changing environmentcharacterized by more-extreme drivers anddisturbances is limited preventing predictionof whether recruitment and growth can bal-ance increased mortality rates in the futureTo evaluatewhether environmental changes

and increasing disturbances are causing glob-ally widespread shifts in vegetation demog-raphy we reviewed global observations ofrecruitment growth andmortality of forestsand woodlands Our expert-derived com-pilation of the state-of-the-art knowledge onvegetation dynamics their drivers and distur-bances allowed us to address four questions(i) Is there evidence for shifts in demographyover recent decades (ii) What physiologicaland disturbance-mediated processes underliethese demographic shifts (iii) What are the po-tential consequences of disturbance-mediatedchanges in demography for climate forcing(iv) How can global predictions of future vege-tation dynamics best be improved

Evidence for changing drivers and disturbancesand their impact on demography

Determining the impacts of changing driverson demography is difficult given the lack ofglobal observation platforms However evi-dence abounds from individual publishedstudies on the drivers and their impacts onplant communities and new modeling andobservational efforts now enable a more com-plete picture of disturbances and forest de-mography (24ndash26) In this section we firstexamine whether there are global trends instand ages and test the sensitivity of the stand-age distribution to changes in disturbance rate

RESEARCH

McDowell et al Science 368 eaaz9463 (2020) 29 May 2020 1 of 10

1Pacific Northwest National Laboratory Richland WA 99354 USA 2US Geological Survey Fort Collins Science Center New Mexico Landscapes Field Station Los Alamos NM 87544 USA3Conservation Ecology Center Smithsonian Conservation Biology Institute Front Royal VA 22630 USA 4Forest Global Earth Observatory Smithsonian Tropical Research Institute Republic ofPanama 5Department of Entomology University of Minnesota St Paul MN 55108 USA 6Joint Global Change Research Institute Pacific Northwest National Laboratory College Park MD 20740USA 7Department of Geographical Sciences University of Maryland College Park MD 20742 USA 8Nicholas School of the Environment Duke University Durham NC 27708 USA 9Departmentof Earth and Environment Boston University Boston MA 02215 USA 10Swiss Federal Institute for Forest Snow and Landscape Research WSL 8903 Birmensdorf Switzerland 11Energyand Resources Group University of California Berkeley Berkeley CA 94720 USA 12Department of Earth System Science Woods Institute for the Environment and Precourt Institute for EnergyStanford University Stanford CA 94305 USA 13School of Forest Resources and Conservation University of Florida Gainesville FL 32611 USA 14Division of Climate and Ecosystem SciencesLawrence Berkeley National Laboratory Berkeley CA 94720 USA 15Department of Biology University of Florida Gainesville FL 32611 USA 16School of Informatics Computing and CyberSystems Northern Arizona University Flagstaff AZ 86001 USA 17Biospheric Sciences Laboratory NASA Goddard Space Flight Center Greenbelt MD 20771 USA 18School of GeographyEarth and Environmental Sciences University of Birmingham B15 2TT Birmingham UK 19Birmingham Institute of Forest Research University of Birmingham B15 2TT Birmingham UK20Department of Forest and Soil Sciences University of Natural Resources and Life Sciences Vienna 1180 Vienna Austria 21School of Life Sciences Technical University of Munich 85354Freising Germany 22Department of Integrative Biology University of WisconsinndashMadison Madison WI 53706 USA 23Department of Ecology Evolution and Environmental Biology ColumbiaUniversity New York NY 10027 USA 24Environmental Sciences Division and Climate Change Science Institute Oak Ridge National Laboratory Oak Ridge TN 37831 USA 25Earth andEnvironmental Sciences Division Los Alamos National Laboratory Los Alamos NM 87545 USACorresponding author Email natemcdowellpnnlgov

on May 28 2020

httpsciencesciencem

agorgD

ownloaded from

using global datasets on LUC (27) and non-LUC (25 28) disturbances We subsequentlydraw upon the wealth of published studieson changes in forest demographics and theirdrivers to investigate the potential changesleading to global stand-age trends Ultimatelythe combination of our global estimates andthe large literature base allows us to generatetestable hypotheses regarding trends and im-pacts of the drivers of forest demographics

Is disturbance changing forest demography atthe global scale

We reanalyzed the Land-use Harmonization(LUHv2) dataset (27) with respect to forestage revealing that the area of young foreststands (here defined as stands younger than140 years old) resulting directly from LUC(conversion of forest to nonforest) or woodharvest (forest retentionwith reduction of bio-mass and age) has increased from 48 millionkm2 in 1900 to 125 million km2 in 2015 (orfrom 113 to 336 of forest area) (Fig 1A)The results were insensitive to assumptionsregarding the link of disturbance likelihoodto stand age (Fig 1A) These forest stand-agedistributions exhibit different trajectories indifferent regions Tropical forests have pro-

gressively lost old-growth area owing to LUCover the course of the 20th century (Fig 2Ablack dashed line) Wood harvest has increasedfrom a minor driver of tropical forest age dis-tribution in 1900 to a major one in 2015 (dif-ference between solid and dashed lines) Thesplit between deforestation and shifting cul-tivation drivers is broadly consistent with asatellite-based analysis for the period 2001ndash2015(29) Temperate and Mediterranean forestages are strongly influenced by wood harvestwhich has made old-growth forests increas-ingly scarce in these regions LUC has hadminimal influence on stand age in borealforests but wood harvest has substantiallyshifted boreal forest age distribution towardyoung growthIn reality old-growth forests aremade scarcer

bymore than just LUC and wood harvest (Figs1A and 2) they are threatened by other dis-turbances that have shifted landscapes fromold to young growthndashdominated stands (14)such as wildfire (29) windthrow (30) andbiotic outbreaks (31) To address these ad-ditional disturbances we integrated recentobservation-based estimates of non-LUC dis-turbance for closed-canopy forests (25 28) withLUC from LUHv2 to obtain a first-principles

estimate of the combined effect of human andnatural disturbances on forest age structure(Fig 1B) A twofold increase in non-LUC distur-bance rates over the period 2015ndash2050 wouldresult in a substantial increase in the fractionof young forests (Fig 1 B andC) Thus realisticshifts in disturbance rates can substantiallyaffect the age structure of forests in the futureAs discussed below such an increase in dis-turbance rate is consistentwith themagnitudeof changes observed or predicted in individualecosystemsNotably calculations based on the Global

Forest AgeDataset (GFAD) v11 (14 32) yielded165 million km2 of old-growth forest and263 million km2 of young forest (32) whichdiffers from what is shown in Fig 1 B and CThis disparity is likely attributable to consid-eration of different forest types (closed-canopyforests versus all forests) and to differences indefinitions of stand size and age used in in-ventories versus those used in satellite-basedestimates

Chronically changing driversAtmospheric CO2

Atmospheric CO2 has risenmore than 125 partsper million (ppm) since the industrial revolu-tion (11) and is projected to rise an additional50 to 200 ppm by 2100 Higher CO2 increasesleaf-level water-use efficiency and rising CO2

has positive but uncertain feedbacks on plantdemographic rates (Fig 3 A and B) Matura-tion and seed production can be acceleratedunder elevated CO2 (33) however seedlinggrowth is not always stimulated by CO2 (34)Recruitment response to rising CO2 is variable(35 36) Forest inventory and tree-ring studiesshow limited evidence for CO2 fertilization ofgrowth (37ndash43) potentially because of the over-whelming influence of increasing drought andnutrient limitations (44) Ecosystem-scale CO2

enrichment experiments in young forests sug-gest a 30 gain in decadal biomass increment(45) but experiments in mature forests havefound minimal growth stimulation (46 47)This is consistent with evidence for an initiallystrongCO2-related growth stimulation in youngforests that decreaseswith tree age and size (39)perhaps due to nutrient (7 48) and hydraulicpath-length limitations (49)A limited number of studies suggest that

elevated CO2 causes increased mortality or nochange inmortalityMortality rates of saplingsduring experimental drought were not miti-gated by elevatedCO2 (50 51) while acceleratedself-thinning due to CO2 fertilizationndashinducedstand density increasesmay lead to highermor-tality (6 52 53) (Fig 3B) The latter processwould be consistent with increases in recruit-ment at large scales Because tree mortality isdominated by large size classes [ie (54)] (fordetails see section on size-related mortalitybelow) faster growth via CO2 fertilizationmay


Box 1 Vegetation dynamics definitions

We focus on three main plant demographic processes recruitment growth and mortality Recruitment(including reproduction) determines the seedling and sapling composition of a plant community afterdisturbance (152) Growth from sapling to mature plant results in development of mature forests andincludes competitive processes Mortality is a key rate controlling carbon storage and species compositionin a plant community and is a dominant demographic rate during a pulse disturbance (153 154)Abiotic drivers Physical factors that cause changes in demography and that respond to global change orto disturbances such as light CO2 soil moisture humidity temperature etcBiotic drivers Biological factors that may drive changes in demography such as pathogens insectsherbivores or competition with other individualsChronic environmental change Persistently changing drivers of demographic rates These drivers have anonstable and directional trajectory such as rising CO2 temperature and VPDDemographic rate Any individual- population- or community-level parameter that affects the age andorsize structure of a population or community including rates of recruitment growth and deathDemographic driver An abiotic or biotic factor that when undergoing a change itself also leads to change(s)in one or more demographic ratesDisturbance The destruction of live plant biomass in a discrete event (155 156)Disturbance regimes Spatial and temporal characteristics of disturbances in a landscape over a long timeperiod including frequency return interval duration intensity severity and sizeGrowth The rate of biomass production over time at the individual or ecosystem scale (ie net primaryproduction in grams of carbon per square meter per year)Land-use and land-cover change Anthropogenic shifts in forms of cultivation or in vegetation cover dueto for example forestry or conversion of woodlands to crop ecosystemsMortality Defined herein as the complete loss of a plantrsquos ability to reproduce and ultimately the loss ofcellular metabolismRecruitment The rates of transition of plants from one size class to another (typically in units of individualsper square meter per year) Recruitment results from the birth and growth of individuals Herein weconsider recruitment from the stage of seed dispersal through seedling growth into the sapling stageSelf-thinning Reduction in the number of live plants within a stand occurring via competition forresourcesVegetation dynamics The net outcome of the interplay between disturbances and vegetationdemographic rates

RESEARCH | REVIEWon M

ay 28 2020

httpsciencesciencemagorg

Dow

nloaded from

expose trees to size-related mortality risksearlier (7) Such CO2-induced increases inmortality may be global (55) Furthermorefaster growth is often associated with lowerwood density (56) rendering fast growingtrees more susceptible to high winds Thusfuture CO2 fertilization could increase re-cruitment growth and mortality (Fig 4B)although there is considerable uncertaintyabout these effects

Temperature and vapor pressure deficit

Temperature and VPD are rising globally andwill continue to rise into the future (57) Bothtemperature and VPD can have impacts ondemographic rates Rising temperature forcesan exponential rise in VPD which promptsstomatal closure and limits photosynthesisleading to lower growth highermortality (58)and reduced regeneration (59) and ultimatelydriving community shifts (60 61) These ob-

servations are consistent with hydraulic theorywhich suggests that as VPD rises potentialmaximum tree height declines (62) (Fig 4)This results from the dependency of watertransport limitations on tree size (49) that areexacerbated by elevated VPD (Fig 4) makingshort stature advantageous with rising VPDBecause most plants cannot reduce their size(beyond limited reductions in leaf area orcrown dieback) forests respond through in-creased mortality of large plants which arereplaced by smaller ones (62) as has beenobserved inmany studies (26 54)While risingair temperature may also increase respira-tory carbon loss leaving less carbon for growth(63) warming inwetter and cooler regionsmayactually stimulate reproductive output re-cruitment and growth (3 64 65) Changesin temperature and VPD also can produceasynchrony in floral and pollinator phenol-ogy (66) and can reduce cold stratification (67)

both of which reduce seed abundance (68) andnegatively affect recruitment (69 70) Saplingmortality is accelerated by elevated temper-ature (70 71) but recruitment has increasedin moist areas (72) Thus rising temperatureand VPDmay be beneficial in cooler or wetterareas but most evidence suggests negativeimpacts on plant demographic rates (Figs 3C and D and 4)

Changing disturbance regimesDroughts

Droughts are anticipated to increase in fre-quency duration and severity globally (Fig 3E and F) and are more stressful to plantsowing to increases in temperature VPD andassociatedwater loss (57) Drought can directlycause tree death or indirectly lead to mortalitythrough associated increases in insect or path-ogen attack (51) Hydraulic failure and car-bon starvation remain themost likelymutually


1900 1920 1940 1960 1980 2000Year

0

10

20

30

40

50F

ores

t are

a (m

illio

n km

2 )A

lt140 y LUC+WH140 y LUC+WH

lt140 y LUC140 y LUC

1900 1950 2000 2050 2100Year

5

10

15

20

CC

fore

st a

rea

(mill

ion

km2 )

B

lt140 y Baseline140 y Baseline

lt140 y Inc dist140 y Inc dist

1-1011-2021-3031-4041-5051-6061-7071-8081-9091-100101-110111-120121-130131-140O

G

Age class (years)

0

05

10

15

20

25

CC

you

ng fo

rest

are

a (m

illio

n km

2 )C

0

5

10

15

CC

OG

forest area (million km

2)

Baseline 2000Baseline 2050Baseline 2100Inc dist 2050Inc dist 2100

Fig 1 Both anthropogenic and wild disturbances have reducedforest ages over the past century (A) Human activities have increasedthe amount of young forest area (stands lt 140 years old) over the 20thcentury as a result of both land-use change (LUC) and wood harvest (WH)Forest stand-age distribution was reconstructed using forest covertransitions from LUHv2 initialized using forest cover fractions in1750 and incrementing forest cover each year tracking the forest age up to140 years Solid lines show the combined effect of LUC and WH anddashed lines show the effect of LUC alone Total forest area is based onLUHv2 The nominal minimum size of a stand is assumed to be ca 01 ha(B) Sensitivity of age distribution in closed-canopy (CC) forests to plausiblechanges in disturbance rate Forest stand-age distribution was recon-structed using forest cover transitions due to LUC from LUHv2 alongsidenon-LUC observation-based disturbance rates (25) In the baseline scenario(solid lines) non-LUC disturbance is assumed to be constant at observed2001ndash2014 values throughout In the incremented distribution scenario(ldquoInc distrdquo dashed lines) disturbance rates are incremented linearly to 200of the 2001ndash2014 values over the period 2015ndash2050 and held constant atthat level thereafter The underlying LUC scenario is Global Change AssessmentModel Representative Concentration Pathway 34 (GCAM RCP 34) whichincludes land-based mitigation for CO2 emissions Results are presented for CCforests only (25) which is why total forest area is lower in (B) than in (A) asnon-LUC disturbance rate information is not currently available for open-canopyforests The shaded areas in (A) and (B) indicate the effect of assuming thatdisturbances are five times more likely to affect young forests than old-growth

forests or vice versa as opposed to an even probability across ages(solid lines) The apparent large dampening of this assumption in(A) versus (B) is primarily due to the different y-axis scales (C) Changes inthe disturbance regime propagate through forest age structure at decadaltime scales CC young (lt140 years old) forest area is shown on the left-handy axis Old-growth (OG gt140 years old) forest area is shown on the right-hand y axis (same units) and refers to the data points in the upper right-handcorner of the panel Scripts used and additional methods can be accessedat httpsgithubcompughtamAgeClassReconst_relgit


ay 28 2020


Dow

nloaded from

inclusive underlying physiological mecha-nisms for drought-induced mortality (73)and both processes are likely to increase treesusceptibility to biotic agents (74) Evidencesuggests that drought-inducedmortality occursmore rapidly under warmer conditions (51 71)Consistent with these empirical results mod-els suggest far greater mortality of temperateconifer trees in the future (75) Reproductiveoutput is often reduced by drought [but see(64)] which combined with drought impactson seedling survival leads to reduced recruit-ment (76) However growth was relatively sta-ble across a drought in Amazonia (77) whilemortality increased Thus like rising temper-ature and VPD it appears that drought mayincreasemortality regardless of location whilehaving variable impacts on recruitment andgrowth (Fig 3F)

Land-use change

LUC and forest management have reducedvegetation stature and biomass and havealtered species composition with profoundconsequences for forest dynamics (Figs 1A

and 3 G and H) Todayrsquos global vegetationbiomass stocks may amount to only ~50 oftheir potential because of LUC (78) Woodharvest and shifting cultivation are the land-use activities primarily responsible for theconversion from primary to secondary vegeta-tion cover and associated demographic shifts(2) In systems that return to wild vegetationor to managed forest after human clearingdemographic rates are typically acceleratedThe increased resource availability after for-est removal facilitates establishment of earlysuccessional species reduces species diversity(79 80) and triggers a transition to youngersmaller plants (81) Post-deforestation recruit-ment is often prolific even in the absence ofmanagement (82) Globally the recovery ofharvested forests and abandoned agriculturalland along with establishment of new planta-

tions has resulted in younger forests (Fig 1A)with associated reductions in tree size and bio-mass (83) Such post-deforestation recruitmentmay be limited by elevated VPD or drought asis the case with recruitment after all-naturaldisturbances Overall the net effect of histor-ical LUC and wood harvest has resulted in asubstantial loss of forest area along with al-tered demographic rates leading to youngershorter less diverse ecosystems (Fig 3H)

Wildfire

Wildfire is increasing in many forests world-wide (84) (Fig 3I) although human man-agement of landscapes has led to wildfiresuppression in some biomes (85) Given suffi-cient fuel burned area increases exponentiallywith aridity (86) and future fire frequenciesmay exceed those documented over the past


0

2

4

6

8

10

12

14

For

est a

rea

(mill

ion

km2)

A Tropical



0

1

2

3

4

5

6

7

For

est a

rea

(mill

ion

km2)

B Temperate and Mediterranean

1900 1920 1940 1960 1980 2000Year

0

2

4

6

8

10

12

For

est a

rea

(mill

ion

km2)

C Boreal

Fig 2 Human activities have increased the amountof young forest area irrespective of biome As inFig 1A but broken down by biome (157) (A) tropical(B) temperate and Mediterranean and (C) boreal

PresentPast Future PresentPast FutureTime

DRIVERS AND DISTURBANCES DEMOGRAPHIC RATES

A

C

E

G

I

K

M N

L

J

H

F

D

B

VPD

Temp

[CO2]

VPD Temperature

Drought

Deforestation

Wildfire

Wind

Insect outbreaks

Growth

Growth

Growth

Growth

Growth

Growth

Growth

Mortality

Mortality

Mortality

Mortality

Mortality

Mortality

Mortality

Recruitment

Recruitment

Recruitment

Recruitment

Recruitment

Recruitment

Recruitment

Fig 3 Drivers disturbances and demographics are changing both historically and into the futureA graphical summary of the literature evidence of changing drivers and disturbances (left-hand column)and subsequent demographic rates (right-hand column) Shown are the chronically changing drivers(A and B) CO2 and (C and D) VPD and temperature as well as the more transient disturbances of(E and F) drought (low precipitation) (G and H) deforestation (I and J) wildfire (K and L) wind and(M and N) insect outbreaks Each driver or disturbancersquos corresponding demographic responses(shown as carbon fluxes per unit area over time) are shown in the corresponding right-hand panels


ay 28 2020


Dow

nloaded from

10000 years (87) Increased fire activity causesincreased mortality and potentially higher re-cruitment and growth of either preexisting ornewly introduced species but rates of re-cruitment and growth may be slowed underclimate warming Forests characterized bystand-replacing fire regimes are dominated byobligate seeders and typically have effectiveseedling recruitment (88) However high-severity and high-frequency fires can reducerecruitment by reducing seed supply throughthe repeated and severe loss of reproductivelymature vegetation (89) and high-frequencyfires can cause recruitment losses via directmortality of the seedbank seedlings andsaplings (90) which is worsened by elevatedVPD (59) Woody species that can resprout af-ter fire including shrubs that suppress treeregeneration (59)may be favored by increasedfire frequency and severity Increased fire se-verity results in high tree mortality in forestshistorically adapted to low-severity fires andsubsequent recruitment and growth may beslow or absent resulting in conversion of for-ests to low-biomass ecosystems (91) Thuswildfire can result in higher demographicrates although rising temperature and VPDcan negatively affect recruitment and growth(Fig 3J)

Windthrow

Windthrow from cyclonic storms representsthe dominant natural disturbance in coastalforests across the globe (92) Cyclonic stormsare expected to increase in frequency windvelocities and precipitation intensity (93) (Fig3K) resulting in more extreme flooding thatpromotes tree instability Windthrow alsoresults from convective thunderstorms andtopographically mediated winds and warm-ing is expected to increase the frequency ofatmospheric conditions conducive to severethunderstorms (94) Canopy damage and

whole-tree mortality are the most immediateimpacts of windthrow (95) (Fig 3L) Storm-induced mortality is greatest for larger trees(96) and the loss of large canopy trees duringwind disturbance favors growth of survivingtrees (96 97) and advances regeneration re-cruitment of early successional species (98)or resprouting of trees broken by wind (99)Depending on the resprouting or seedingcapacity of surviving species wind damagemay either slow or accelerate succession (100)We note that storms may also be associatedwith lightning which may be a prominentcause of large-tree mortality (101) Thus wind-storms should result in changes in all threedemographic rates although with large uncer-tainty at the global scale (Fig 3L)

Biotic agents

Bioticdisturbances frominsects insectndashpathogencomplexes and other biotic agents have beenincreasing in frequency severity and extent inrecent decades (17 31 102) (Fig 3M) Suchtrends reflect a changing climate (103) alteredland use (104) and introductions of nonindig-enous insects and pathogens (105) Climatechange is expected to further amplify bioticdisturbances (106) in part through enhancedhost vulnerability (Fig 3M) However shiftsin frequency or dampening of disturbanceregimes could also emerge (107) leading tosome uncertainty in outbreak dynamics underfuture conditions (Fig 3M) Whereas insectsand associated pathogens are globally wide-spread lianas or vines that use other plants ashost structures are increasing in abundanceand are thought to be causing increasing mor-tality in the tropics (7 108)Response of insects and pathogens to cli-

mate change is likely to increase plant mor-tality (4) with variable impacts on growth andrecruitment (Fig 3N) Tree mortality can re-sult from girdling of the phloem and xylem by

bark beetles (74) and from repeated defolia-tion events that exhaust the capacity of treesto recover (109) Tree mortality during out-breaks is usually partial at the stand level be-cause many biotic agents preferentially attacktrees of specific size or health classes or arehost-specific (16) Suppressed smaller trees andnonhost tree species may survive and growrapidly when released from competition forresources (110 111) Thus similar tomany otherdisturbances mortality increases while recruit-ment and growth show variable responses tobiotic disturbances including a dependencyon post-disturbance temperature VPD anddrought

On size and age demographics

The combination of LUC disturbances andchronic drivers is likely to have already shiftedforests to younger and shorter stands withthese impacts increasing under expected fu-ture changes in drivers and disturbances (Fig 1A to C) These results are consistent with ourreview of the literature (Fig 3) Large trees arethe most susceptible to die from LUC-inducedforest fragmentation (112 113) drought (26)rising temperature or VPD (54 62) (Fig 4)windthrow (114 115) biotic attacks (116) andlightning (101) with variable size impacts offire (117) The abundance of size-dependentmortality drivers and disturbances shouldlogically push stands toward younger andsmaller distributions of trees and shorter-statured species assemblages (118)There are exceptions to the pattern of cli-

mate drivers and disturbances reducing treeheight and stand age Non-stand-replacing firesthat kill smaller trees but spare the larger oldertrees will shift forests toward larger size distri-butions Similarly on occasions when droughtspreferentially kill younger but fast-growingtrees subsequent size distribution and rateof carbon accumulation would be affectedRising CO2 and increased precipitation in someareas also counter the general decrease in sizebecause they may lead to faster growth andhence taller trees (119) Thus the antagonisticdrivers promoting larger trees (eg rising CO2)and smaller trees (eg rising VPD increasingdisturbances) co-occur but the general patternof decreasing size and younger ages reveals thatprocesses driving down size and age (Figs 1 to4) are dominant globally

Mitigation of demographic-disturbance impacts

The literature patterns suggest that most driv-ers and disturbances will increase tree mortal-ity now and in the future with variable effectson recruitment and growth (Fig 3) This sup-position becomes uncertain however whenwe consider multiple feedbacks that canmitigate the changes in forest demographyinduced by chronically changing drivers anddisturbance regimes These processes include


Fig 4 Rising VPD forces declinesin potential plant staturePredictions of plant height inresponse to rising VPD from thehydraulic corollary to Darcyrsquos lawThe equation is h = (As times ks timesDY)(G times Al times VPD) where his height As is sapwood area ks isspecific conductivity DY is theleaf-to-soil water potential gradientG is stomatal conductance and Al isleaf area (53) The different linesrepresent different levels of accli-mation of As ks DY G and Al allallowed to adjust simultaneouslyfrom 0 to 60 of their initial valuesIn the case of G it is assumed to decrease because of rising atmospheric CO2 Acclimation can help but notcompletely mitigate the impact of rising VPD on plant stature


ay 28 2020


Dow

nloaded from

acclimation adaptation migration and com-pensatory mechanisms of resource use Withglobal change forests will be influenced by acombination of phenotypic plasticity [ie ac-climation (120)] adaptation to novel biotic andabiotic stresses (121) and the ability tomigrateas conditions change (122) Failure to accli-mate adapt or migratemdashincluding failure dueto human infrastructure (123)mdashcould lead torecruitment and growth reductions and localextinctions Plants have demonstrated accli-mation of phenology seed longevity andmeta-bolic processes to single andor multiplestressors (124ndash127) Acclimation and adapta-tion will likely depend on an array of factorsincluding genetic variation fecundity disper-sal population size and environmental varia-bility (120) Many tree species have migratedin response to past climatic cycles but at ratesslower than the current pace of climate change(128) Regarding resource use reductions instand density as a result of increasedmortalityor reduced recruitment should allow greaterresource availability to surviving individualsand therefore subsequently higher growth andsurvival rates (129) Such stand resourcemech-anisms can manifest at the landscape scaleas most disturbances are patchy (130) andthe size shape and arrangement of survivingforest patches can play a key role in recovery ofthe disturbed landscape (20) Taken togetherthe mitigating factors can play a substantialrole in buffering the impacts of changing driv-ers on plant survival but it remains unclearwhether these factors will enable trees to keeppace with ongoing climate change (50 120)Ultimately the uncertainty surrounding futuredemographic rates shown in Fig 3 is partiallydue to the influence of these mitigating factors

Consequences for community assembly andfor climate forcing

The widespread shift in vegetation dynamicsbegets questions regarding consequences forcommunity assembly and climate forcing Hy-draulic theory suggests that under rising VPDfunctional traits of high conductance lowstature and low leaf area should best enablesurvival all of which are characteristics of pio-neer shrub and weed species (62) Consistentwith this theory diversity (eg species rich-ness) temporarily increases post-disturbancefor many systems as short-statured oppor-tunistic species invade (131) If forest com-munities shift toward trait assemblages bettersuited to the new disturbance regime suchshifts may confer some resistance to futuredisturbances (131 132) Alternatively if distur-bance regimes shift faster than recruitmentgrowth and subsequent community assemblycan respond resistance to future disturbanceswill likely declineClimate forcing responds to changing vege-

tation dynamics in complex ways Changes

in forest disturbance regimes and vegetationdynamics can affect climate via biogeochem-ical hydrological and land-surface energybudgets (133) Reductions in biomass resultin a loss of carbon to the atmosphere despiteyounger shorter stands often having highergross photosynthesis this is due to the lossof carbon through decomposition of necro-mass which is a particularly large flux frommortality of older larger trees such as those inold-growth forests (134) and reduced landscape-mean carbon storage under an intensified dis-turbance regime (135) The time required foran ecosystem to reachieve the same live car-bon storage after disturbance can be decadesto centuries particularly if the disturbancecycle is increased thus the net effect of thebiomass loss is increased CO2 to the atmo-sphere and hence greater climate forcingThis impact may be mitigated by increasedcarbon uptake due to CO2 fertilization (119)or enhanced recruitment Calculations ofthe terrestrial carbon sink from atmosphericinversions indicate that the sink grew overrecent decades (12) in part because of in-creased leaf area (13) which is consistent withincreased recruitment and growth Howeverevidence suggests that forests are switchingfrom a CO2 fertilizationndashdominated periodto a VPD-dominated period (15 16) despitesustained high gross photosynthesis at theglobal scale (136) The increased mortalitythroughout much of the terrestrial biosphere(7ndash9) further minimizes potential carbonstorage through enhanced biomass loss Ul-timately the terrestrial contribution to cli-mate forcing through carbon uptake andrelease results from the antagonistic processof rising CO2 and forest recovery from LUCwhich enhance the carbon sink and risingVPD and disturbances that reduce the car-bon sinkChanging vegetation dynamics also influ-

ence regional and global surface energy bud-gets and hydrological cycles Disturbancesfrequently shift albedo of ecosystems fromdarker to lighter resulting in a decline in ra-diative forcing through less light absorption(137) The rate of recruitment after distur-bance influences the temporal period of thisnegative feedback (138) The impact of chang-ing vegetation dynamics on the water cycleis particularly complex Evaporation fromcanopies shifts as stands become taller be-cause taller trees transpire less (per unit leafarea) than smaller trees (49) but larger treesoften have better rooting access to watersources and have greater total leaf area Thenet effect of disturbance is a transient de-crease in evaporative loading to the atmo-sphere along with albedo shifts causing afeedback of decreasing precipitation down-wind (139 140) Ultimately carbon storage isat least transiently reduced by disturbances

with mixed impacts on the water and energybudgets

The path to improved prediction

Changes in global drivers (temperature CO2and VPD) and disturbances (including LUCdrought wildfire windstorms and insect out-breaks) should all force forests toward shorteryounger lower-biomass ecosystems This trendis supported by hydraulic theory (62) (Fig 4)and by abundant empirical evidence dem-onstrating a consistent increase in mortalityacross the global spectrum of drivers anddisturbances and variable but often declin-ing recruitment and growth (Fig 3)While thebulk of the evidence points to reduced plantstature owing to changing drivers large un-certainty remains in the magnitude and slopeof demographic trajectories in the future (Fig3) Given these trajectories and the large un-certainties around them what are the criticalnext steps to allow improved global predic-tion Continued long-term observations (bothon the ground and remotely sensed) are essen-tial to reveal the patterns of demographic re-sponses to drivers and disturbances Likewisemanipulative experiments are needed that alterconditions such as CO2 or drought to providecause-and-effect understanding of the inter-actions among mechanisms of demographicresponses However for global-scale predic-tion of responses and climate consequenceswe need to mainstream insights from obser-vations and experiments into Earth systemmodels (ESMs)ESMs simulate the exchange of fluxes be-

tween the atmosphere land and ocean andstores of carbon water and energy the land-surface modules of ESMs simulate vegetationESMs have made great progress in simulat-ing land use disturbances and demographyincluding representation of wildfire (141)drought-induced mortality (142) and cohort-age structured models that enable represen-tation of succession and associated shifts inphysiological traits (6) The global CoupledModel Intercomparison Project CMIP6 nowincludes a dedicated model intercomparisonactivity focused on the effects of changes ofland use on carbon and climate (143) Ad-vances in remote sensing and forest inventoryintegration are enhancing global datasets offorest structure (144) and age (32) that can beused in model initialization data assimilationbenchmarking and sensitivity analyses (Fig 1A to C) These advancements set the stage fordevelopments in ESMs such as the predic-tion of disturbances and demographic rateresponses under climate and LUC scenariosThe newest generation of ESMs uses size or

age-structured approaches to explicitly mod-el demography in the Earth system (6) whichshould ultimately enable model-based repre-sentation of observed shifts in age structure



ay 28 2020


Dow

nloaded from

(eg Fig 1) However representation of vege-tation demographic rates remains relativelysimplistic Simulation of growth responses toglobal change requires model refinement inlight capture belowground water and nutri-ent acquisition and responses of respirationto temperature (6) Recruitment includingreproduction and dispersal is the most un-developed demographic process in ESM simu-lations Reproductive allocation is invariantwith plant functional type (PFT) and seedis assumed to mix evenly throughout a gridcell [but see (145)] Environmental constraintsto PFT establishment are derived from priordistributions of major taxa and while recruit-ment rates can be influenced by light or spaceavailability they are not responsive to tem-perature CO2 or soil moisture (146 147)Simplistic dispersal assumptions are typicallyeither overly permissive or overly restrictiveImprovements in representing recruitmentunder global change are critical for improv-ing predictions of vegetation dynamics Theseadvancements will require data synthesis andadditional data collection to support PFT-specific environmentally sensitive param-eterizations of regeneration processes suchas reproductive allocation effective dispersalseedling establishment survival and growthand post-disturbance recovery strategies (egserotiny and resprouting)Disturbance-inducedmortality is better de-

veloped for landscape-scale models than forESMs ESMmodeling of disturbance-inducedmortality exists for wildfire and drought(141 142) although considerable challengesremain to reliably represent both disturbancesglobally while ESMs are underdeveloped forwind and insect mortality To our knowledgeonly one ESM currently represents canopydamage (148) this causes ESMs to potentiallyunderestimate the impacts of drought andwind as both disturbances cause lagged treemortality associated with canopy loss yearsafter the inciting event (149 150) As for in-sects there have been prescriptive studiesexamining the impact of insect outbreaks onland processes within ESMs but no ESM hasyet explicitly considered the interaction be-tween plant defense and insect populationdynamics for prediction of large-scale insect-induced treemortality For predictingwildfiremodels should be sensitive to both fuels andclimate interactions and represent spatialpatterns of burn severity because the burnmosaic strongly influences postfire vegetationdynamics (141) Next-generation demographicmodels are evolving to include explicit mech-anistic representations of drought-associatedmortality including carbon starvation andhydraulic failure (151) The evaluation of newhydraulics models (151) for prediction of mor-tality is an essential next stepUltimatelymodelformulations that include environmentally sen-

sitive PFT-specific processes compatible withthe cohort-based approach are likely to pro-vide the best compromise between processdetail and parsimony and are therefore mostlikely to capture changes in large-scale forestdynamics under future conditions

Outlook

Forest vegetation dynamics are already stronglyinfluenced by global change (Fig 1) and willcontinue to be affected in the future (Figs 1 to4) by changes in land use chronic drivers suchas CO2 and VPD and increasing frequencyand severity of transient disturbances suchas windthrow wildfire and insect outbreaksEffects on forests are driven largely by con-sistent increases in tree mortality from thesedrivers and variable responses of recruitmentand growth depending on stand age distur-bance type and geographic location (Fig 3)The consequences of changing demographicssuggest an increasing constraint in terrestrialcarbon storage due at least to the consistentincrease in mortality Any declines in recruit-ment or growth especially when disturbancendashrecovery cycles are disrupted will exacerbatethis carbon-cycle constraint Shifts in otherterrestrial radiative forcing terms such as en-ergy andwater budgets are also likely Althoughwell supported by the literature data andsensitivity analysis (Fig 1) the trends in Fig 3represent hypotheses to be tested by the nextgeneration of observational platforms bothterrestrial and spaceborne Forestmanagementmust ultimately confront the elevated mortal-ity and uncertainty in recruitment and growthwhen considering options for sustaining thesocietal benefits of forests into the future

REFERENCES AND NOTES

1 M G Turner T G Whitby D B Tinker W H RommeTwenty-four years after theYellowstone Fires Are postfirelodgepole pine stands converging in structure and functionEcology 97 1260ndash1273 (2016) doi 10189015-15851pmid 27349102

2 G C Hurtt et al Harmonization of land-use scenarios for theperiod 1500ndash2100 600 years of global gridded annualland-use transitions wood harvest and resulting secondarylands Clim Change 109 117ndash161 (2011) doi 101007s10584-011-0153-2

3 K J Anderson-Teixeira et al Altered dynamics of forestrecovery under a changing climate Glob Change Biol 192001ndash2021 (2013) doi 101111gcb12194 pmid 23529980

4 M Reichstein et al Climate extremes and the carbon cycleNature 500 287ndash295 (2013) doi 101038nature12350pmid 23955228

5 R Seidl et al Forest disturbances under climate changeNat Clim Change 7 395ndash402 (2017) doi 101038nclimate3303 pmid 28861124

6 R A Fisher et al Vegetation demographics in Earth SystemModels A review of progress and priorities Glob Change Biol24 35ndash54 (2018) doi 101111gcb13910 pmid 28921829

7 N McDowell et al Drivers and mechanisms of tree mortalityin moist tropical forests New Phytol 219 851ndash869 (2018)doi 101111nph15027 pmid 29451313

8 J Carnicer et al Widespread crown condition decline foodweb disruption and amplified tree mortality with increasedclimate change-type drought Proc Natl Acad Sci USA108 1474ndash1478 (2011) doi 101073pnas1010070108pmid 21220333

9 C Senf et al Canopy mortality has doubled in Europersquostemperate forests over the last three decades Nat Commun

9 4978 (2018) doi 101038s41467-018-07539-6pmid 30478255

10 M Fernaacutendez-Martiacutenez et al Global trends in carbon sinksand their relationships with CO2 and temperature Nat ClimChange 9 73ndash79 (2019) doi 101038s41558-018-0367-7

11 P Friedlingstein et al Global carbon budget 2019 Earth SystSci Data 11 1783ndash1838 (2019) doi 105194essd-11-1783-2019

12 P Ciais et al Five decades of northern land carbon uptakerevealed by the interhemispheric CO2 gradient Nature 568221ndash225 (2019) doi 101038s41586-019-1078-6pmid 30944480

13 J M Chen et al Vegetation structural change since 1981significantly enhanced the terrestrial carbon sink NatCommun 10 4259 (2019) doi 101038s41467-019-12257-8pmid 31534135

14 T A M Pugh et al Role of forest regrowth in global carbonsink dynamics Proc Natl Acad Sci USA 116 4382ndash4387(2019) doi 101073pnas1810512116 pmid 30782807

15 J Pentildeuelas et al Shifting from a fertilization-dominated to awarming-dominated period Nat Ecol Evol 1 1438ndash1445(2017) doi 101038s41559-017-0274-8 pmid 29185529

16 W Yuan et al Increased atmospheric vapor pressure deficitreduces global vegetation growth Sci Adv 5 eaax1396(2019) doi 101126sciadvaax1396 pmid 31453338

17 K F Raffa et al Cross-scale drivers of natural disturbancesprone to anthropogenic amplification The dynamics of barkbeetle eruptions Bioscience 58 501ndash517 (2008)doi 101641B580607

18 M K Tippett C Lepore J E Cohen More tornadoes in themost extreme US tornado outbreaks Science 3541419ndash1423 (2016) doi 101126scienceaah7393pmid 27934705

19 G R van der Werf et al Global fire emissions estimatesduring 1997ndash2016 Earth Syst Sci Data 9 697ndash720 (2017)doi 105194essd-9-697-2017

20 A Sommerfeld et al Patterns and drivers of recentdisturbances across the temperate forest biome NatCommun 9 4355 (2018) doi 101038s41467-018-06788-9pmid 30341309

21 R A Houghton A A Nassikas Global and regional fluxes ofcarbon from land use and land cover change 1850ndash2015Global Biogeochem Cycles 31 456ndash472 (2017) doi 1010022016GB005546

22 J A Foley et al Global consequences of land use Science309 570ndash574 (2005) doi 101126science1111772pmid 16040698

23 M P Perring et al Global environmental change effects onecosystems The importance of land-use legacies GlobChange Biol 22 1361ndash1371 (2016) doi 101111gcb13146pmid 26546049

24 H Hartmann et al Monitoring global tree mortality patternsand trends Report from the VW symposium lsquoCrossingscales and disciplines to identify global trends of treemortality as indicators of forest healthrsquo New Phytol 217984ndash987 (2018) doi 101111nph14988 pmid 29334597

25 T A M Pugh A Arneth M Kautz B Poulter B SmithImportant role of forest disturbances in the global biomassturnover and carbon sinks Nat Geosci 12 730ndash735 (2019)doi 101038s41561-019-0427-2 pmid 31478009

26 A E L Stovall H Shugart X Yang Tree height explainsmortality risk during an intense drought Nat Commun 104385 (2019) doi 101038s41467-019-12380-6pmid 31558795

27 G C Hurtt et al Harmonization of Global Land-Use Changeand Management for the Period 850ndash2100 (LUH2) forCMIP6 Geosci Model Dev Discuss 105194gmd-2019-360(2020)

28 M C Hansen et al High-resolution global maps of 21st-century forest cover change Science 342 850ndash853 (2013)doi 101126science1244693 pmid 24233722

29 P G Curtis C M Slay N L Harris A TyukavinaM C Hansen Classifying drivers of global forest lossScience 361 1108ndash1111 (2018) doi 101126scienceaau3445 pmid 30213911

30 S Frolking et al Forest disturbance and recovery A generalreview in the context of spaceborne remote sensing ofimpacts on aboveground biomass and canopy structureJ Geophys Res 114 G00E02 (2009) doi 1010292008JG000911

31 M Kautz A J H Meddens R J Hall A Arneth Bioticdisturbances in Northern Hemisphere forestsmdashA synthesis ofrecent data uncertainties and implications for forest



ay 28 2020


Dow

nloaded from

monitoring and modelling Glob Ecol Biogeogr 26 533ndash552(2017) doi 101111geb12558

32 B Poulter et al The Global Forest Age Dataset and itsUncertainties (GFADv11) NASA National Aeronautics andSpace Administration PANGAEA (2019) doi 101594PANGAEA897392

33 S L LaDeau J S Clark Rising CO2 levels and the fecundityof forest trees Science 292 95ndash98 (2001) doi 101126science1057547 pmid 11292871

34 J E Mohan J S Clark W H Schlesinger Long-term CO2

enrichment of a forest ecosystem Implications for forestregeneration and succession Ecol Appl 17 1198ndash1212(2007) doi 10189005-1690 pmid 17555228

35 L G Perry P B Shafroth D M Blumenthal J A MorganD R LeCain Elevated CO2 does not offset greater waterstress predicted under climate change for native and exoticriparian plants New Phytol 197 532ndash543 (2013)doi 101111nph12030 pmid 23171384

36 N Saintilan K Rogers Woody plant encroachment ofgrasslands A comparison of terrestrial and wetland settingsNew Phytol 205 1062ndash1070 (2015) doi 101111nph13147pmid 25729806

37 S M McMahon G G Parker D R Miller Evidence for arecent increase in forest growth Proc Natl Acad Sci USA107 3611ndash3615 (2010) doi 101073pnas0912376107pmid 20133710

38 J J Camarero A Gazol J D Galvaacuten G Sanguumlesa-BarredaE Gutieacuterrez Disparate effects of global-change drivers onmountain conifer forests Warming-induced growthenhancement in young trees vs CO2 fertilization in old treesfrom wet sites Glob Change Biol 21 738ndash749 (2015)doi 101111gcb12787 pmid 25362899

39 S L Voelker R M Muzika R P Guyette M C StambaughHistorical CO2 growth enhancement declines with age inQuercus and Pinus Ecol Monogr 76 549ndash564 (2006)doi 1018900012-9615(2006)076[0549HCGEDW]20CO2

40 J Pentildeuelas J G Canadell R Ogaya Increased water-useefficiency during the 20th century did not translate intoenhanced tree growth Glob Ecol Biogeogr 20 597ndash608(2011) doi 101111j1466-8238201000608x

41 P van der Sleen et al No growth stimulation of tropical treesby 150 years of CO2 fertilization but water-use efficiencyincreased Nat Geosci 8 24ndash28 (2015) doi 101038ngeo2313

42 M P Girardin et al No growth stimulation of Canadarsquosboreal forest under half-century of combined warming andCO2 fertilization Proc Natl Acad Sci USA 113 E8406ndashE8414(2016) doi 101073pnas1610156113 pmid 27956624

43 R J W Brienen et al Long-term decline of the Amazoncarbon sink Nature 519 344ndash348 (2015) doi 101038nature14283 pmid 25788097

44 Z Gedalof A Berg Tree ring evidence for limited direct CO2

fertilization of forests over the 20th century GlobalBiogeochem Cycles 24 GB3027 (2010) doi 1010292009GB003699

45 A P Walker et al Decadal biomass increment in earlysecondary succession woody ecosystems is increased byCO2 enrichment Nat Commun 10 454 (2019) doi 101038s41467-019-08348-1 pmid 30765702

46 M K F Bader et al Central European hardwood trees in ahigh-CO2 future Synthesis of an 8-year forest canopy CO2

enrichment project J Ecol 101 1509ndash1519 (2013)doi 1011111365-274512149

47 D S Ellsworth et al Elevated CO2 does not increase eucalyptforest productivity on a low-phosphorus soil Nat ClimChange 7 279ndash282 (2017) doi 101038nclimate3235

48 R J Norby et al Net primary productivity of a CO2‐enricheddeciduous forest and the implications for carbon storageEcol Appl 12 1261ndash1266 (2002) doi 1018901051-0761(2002)012[1261NPPOAC]20CO2

49 N G McDowell B J Bond L T Dickman M G RyanD Whitehead ldquoRelationships between tree height andcarbon isotope discriminationrdquo in Size- and Age-RelatedChanges in Tree Structure and Function F C MeinzerB Lachenbruch T E Dawson Eds (Springer 2011)pp 255ndash286

50 H Duan et al Elevated [CO2] does not ameliorate thenegative effects of elevated temperature on drought-inducedmortality in Eucalyptus radiata seedlings Plant Cell Environ37 1598ndash1613 (2014) doi 101111pce12260pmid 24372529

51 C D Allen D D Breshears N G McDowell Onunderestimation of global vulnerability to tree mortality and

forest die‐off from hotter drought in the AnthropoceneEcosphere 6 129 (2015) doi 101890ES15-002031

52 C Koumlrner A matter of tree longevity Science 355 130ndash131(2017) doi 101126scienceaal2449 pmid 28082545

53 U Buumlntgen et al Limited capacity of tree growth to mitigatethe global greenhouse effect under predicted warming NatCommun 10 2171 (2019) doi 101038s41467-019-10174-4pmid 31092831

54 A C Bennett N G McDowell C D AllenK J Anderson-Teixeira Larger trees suffer most duringdrought in forests worldwide Nat Plants 1 15139 (2015)doi 101038nplants2015139 pmid 27251391

55 K Yu et al Pervasive decreases in living vegetation carbonturnover time across forest climate zones Proc Natl AcadSci USA 116 24662ndash24667 (2019) doi 101073pnas1821387116 pmid 31740604

56 H Pretzsch P Biber G Schuumltze J Kemmerer E Uhl Wooddensity reduced while wood volume growth accelerated inCentral European forests since 1870 For Ecol Manage 429589ndash616 (2018) doi 101016jforeco201807045

57 K E Trenberth et al Global warming and changes indrought Nat Clim Change 4 17ndash22 (2014) doi 101038nclimate2067

58 A Park Williams et al Temperature as a potent driver ofregional forest drought stress and tree mortality Nat ClimChange 3 292ndash297 (2013) doi 101038nclimate1693

59 A J Tepley J R Thompson H E EpsteinK J Anderson-Teixeira Vulnerability to forest loss throughaltered postfire recovery dynamics in a warming climatein the Klamath Mountains Glob Change Biol 23 4117ndash4132(2017) doi 101111gcb13704 pmid 28447370

60 J M Serra-Diaz et al Disequilibrium of fire-prone forestssets the stage for a rapid decline in conifer dominance duringthe 21st century Sci Rep 8 6749 (2018) doi 101038s41598-018-24642-2 pmid 29712940

61 M Uriarte J R Lasky V K Boukili R L Chazdon A trait-mediated neighbourhood approach to quantify climateimpacts on successional dynamics of tropical rainforestsFunct Ecol 30 157ndash167 (2016) doi 1011111365-243512576

62 N G McDowell C D Allen Darcyrsquos law predicts widespreadforest mortality under climate warming Nat Clim Change 5669ndash672 (2015) doi 101038nclimate2641

63 J Liu et al Contrasting carbon cycle responses of thetropical continents to the 2015ndash2016 El Nintildeo Science 358eaam5690 (2017) doi 101126scienceaam5690pmid 29026011

64 S J Wright O Calderoacuten Seasonal El Nintildeo and longer termchanges in flower and seed production in a moist tropicalforest Ecol Lett 9 35ndash44 (2006) pmid 16958866

65 T F Keenan W J Riley Greening of the land surface in theworldrsquos cold regions consistent with recent warming NatClim Change 8 825ndash828 (2018) doi 101038s41558-018-0258-y pmid 30319714

66 J R Forrest Plantndashpollinator interactions and phenologicalchange What can we learn about climate impacts fromexperiments and observations Oikos 124 4ndash13 (2015)doi 101111oik01386

67 J L Walck S N Hidayati K W Dixon K ThompsonP Poschlod Climate change and plant regeneration fromseed Glob Change Biol 17 2145ndash2161 (2011) doi 101111j1365-2486201002368x

68 J Memmott P G Craze N M Waser M V Price Globalwarming and the disruption of plant-pollinator interactionsEcol Lett 10 710ndash717 (2007) doi 101111j1461-0248200701061x pmid 17594426

69 L M Kueppers et al Warming and provenance limit treerecruitment across and beyond the elevation range ofsubalpine forest Glob Change Biol 23 2383ndash2395 (2017)doi 101111gcb13561 pmid 27976819

70 W D Hansen M G Turner Origins of abrupt changePostfire subalpine conifer regeneration declines nonlinearlywith warming and drying Ecol Monogr 89 e01340 (2019)doi 101002ecm1340

71 H D Adams et al Temperature sensitivity of drought-induced tree mortality portends increased regional die-offunder global-change-type drought Proc Natl Acad Sci USA106 7063ndash7066 (2009) doi 101073pnas0901438106pmid 19365070

72 R A Hember et al Accelerating regrowth of temperate-maritime forests due to environmental change Glob ClimChange 18 2026ndash2040 (2012) doi 101111j1365-2486201202669x

73 H D Adams et al A multi-species synthesis of physiologicalmechanisms in drought-induced tree mortality Nat EcolEvol 1 1285ndash1291 (2017) doi 101038s41559-017-0248-xpmid 29046541

74 M L Gaylord et al Drought predisposes pintildeon-juniperwoodlands to insect attacks and mortality New Phytol 198567ndash578 (2013) doi 101111nph12174 pmid 23421561

75 N G McDowell et al Multi-scale predictions of massiveconifer mortality due to chronic temperature rise Nat ClimChange 6 295ndash300 (2016) doi 101038nclimate2873

76 B M J Engelbrecht T A Kursar Comparative drought-resistance of seedlings of 28 species of co-occurring tropicalwoody plants Oecologia 136 383ndash393 (2003) doi 101007s00442-003-1290-8 pmid 12811534

77 C E Doughty et al Drought impact on forest carbondynamics and fluxes in Amazonia Nature 519 78ndash82 (2015)doi 101038nature14213 pmid 25739631

78 K H Erb et al Land management Data availability andprocess understanding for global change studies GlobChange Biol 23 512ndash533 (2017) doi 101111gcb13443pmid 27447350

79 P A Martin A C Newton M Pfeifer M Khoo J M BullockImpacts of tropical selective logging on carbon storage andtree species richness A meta-analysis For Ecol Manage356 224ndash233 (2015) doi 101016jforeco201507010

80 A Chaudhary Z Burivalova L P Koh S Hellweg Impact offorest management on species richness Global meta-analysis and economic trade-offs Sci Rep 6 23954 (2016)doi 101038srep23954 pmid 27040604

81 M J Duveneck J R Thompson E J Gustafson Y LiangA M G de Bruijn Recovery dynamics and climate changeeffects to future New England forests Landsc Ecol 321385ndash1397 (2017) doi 101007s10980-016-0415-5

82 D Thom W Rammer R Garstenauer R Seidl Legacies ofpast land use have a stronger effect on forest carbonexchange than future climate change in a temperate forestlandscape Biogeosciences 15 5699ndash5713 (2018)doi 105194bg-15-5699-2018

83 T Vileacuten et al Reconstructed forest age structure in Europe1950ndash2010 For Ecol Manage 286 203ndash218 (2012)doi 101016jforeco201208048

84 W M Jolly et al Climate-induced variations in global wildfiredanger from 1979 to 2013 Nat Commun 6 7537 (2015)doi 101038ncomms8537 pmid 26172867

85 N Andela et al A human-driven decline in global burnedarea Science 356 1356ndash1362 (2017) doi 101126scienceaal4108 pmid 28663495

86 J T Abatzoglou A P Williams Impact of anthropogenicclimate change on wildfire across western US forests ProcNatl Acad Sci USA 113 11770ndash11775 (2016) doi 101073pnas1607171113 pmid 27791053

87 A L Westerling M G Turner E A H Smithwick W H RommeM G Ryan Continued warming could transform GreaterYellowstone fire regimes by mid-21st century Proc Natl AcadSci USA 108 13165ndash13170 (2011) doi 101073pnas1110199108 pmid 21788495

88 D Bowman G J Williamson L D Prior B P MurphyThe relative importance of intrinsic and extrinsic factors inthe decline of obligate seeder forests Glob Ecol Biogeogr25 1166ndash1172 (2016) doi 101111geb12484

89 J F Johnstone et al Changing disturbance regimesecological memory and forest resilience Front Ecol Environ14 369ndash378 (2016) doi 101002fee1311

90 M G Turner K H Braziunas W D Hansen B J HarveyShort-interval severe fire erodes the resilience of subalpinelodgepole pine forests Proc Natl Acad Sci USA 11611319ndash11328 (2019) doi 101073pnas1902841116pmid 31110003

91 T Kitzberger et al Fire-vegetation feedbacks and alternativestates Common mechanisms of temperate forestvulnerability to fire in southern South America andNew Zealand N Z J Bot 54 247ndash272 (2016) doi 1010800028825X20161151903

92 A E Lugo Visible and invisible effects of hurricanes on forestecosystems An international review Austral Ecol 33368ndash398 (2008) doi 101111j1442-9993200801894x

93 K Balaguru G R Foltz L R Leung Increasing magnitude ofhurricane rapid intensification in the central and easterntropical Atlantic Geophys Res Lett 45 4238ndash4247 (2018)doi 1010292018GL077597

94 N S Diffenbaugh M Scherer R J Trapp Robust increasesin severe thunderstorm environments in response togreenhouse forcing Proc Natl Acad Sci USA 110



ay 28 2020


Dow

nloaded from

16361ndash16366 (2013) doi 101073pnas1307758110pmid 24062439

95 M Uriarte C D Canham J Thompson J K ZimmermanA neighborhood analysis of tree growth and survival in ahurricane-driven topical forest Ecol Monogr 74 591ndash614(2004) doi 10189003-4031

96 B Gardiner P Berry B Moulia Review Wind impacts onplant growth mechanics and damage Plant Sci 24594ndash118 (2016) doi 101016jplantsci201601006pmid 26940495

97 M Uriarte et al Natural disturbances and human land use asdeterminants of tropical forest dynamics Results from aforest simulator Ecol Monogr 79 423ndash443 (2009)doi 10189008-07071

98 L S Comita et al Abiotic and biotic drivers of seedlingsurvival in a hurricane-impacted tropical forest J Ecol 971346ndash1359 (2009) doi 101111j1365-2745200901551x

99 M Uriarte et al Multidimensional trade-offs in speciesresponses to disturbance Implications for diversity in asubtropical forest Ecology 93 191ndash205 (2012) doi 10189010-24221 pmid 22486099

100 D F B Flynn et al Hurricane disturbance alters secondaryforest recovery in Puerto Rico Biotropica 42 149ndash157(2010) doi 101111j1744-7429200900581x

101 S P Yanoviak et al Lightning is a major cause oflarge tree mortality in a lowland neotropical forestNew Phytol 225 1936ndash1944 (2020) doi 101111nph16260 pmid 31610011

102 R Seidl M J Schelhaas W Rammer P J VerkerkIncreasing forest disturbances in Europe and their impact oncarbon storage Nat Clim Change 4 806ndash810 (2014)doi 101038nclimate2318 pmid 25737744

103 K F Raffa et al ldquoResponses of tree-killing bark beetles to achanging climaterdquo in Climate Change and Insect PestsC Bjoumlrkman P Niemelauml Eds vol 7 of CABI Climate ChangeSeries (CABI International 2015) pp 173ndash201

104 J S Schurman et al Large-scale disturbance legacies andthe climate sensitivity of primary Picea abies forests GlobChange Biol 24 2169ndash2181 (2018) doi 101111gcb14041pmid 29322582

105 D W Rosenberger R C Venette B H AukemaDevelopment of an aggressive bark beetle on novel hostsImplications for outbreaks in an invaded range J Appl Ecol55 1526ndash1537 (2018) doi 1011111365-266413064

106 R Seidl et al Invasive alien pests threaten the carbon storedin Europersquos forests Nat Commun 9 1626 (2018)doi 101038s41467-018-04096-w pmid 29691396

107 D M Johnson et al Climatic warming disrupts recurrentAlpine insect outbreaks Proc Natl Acad Sci USA 10720576ndash20581 (2010) doi 101073pnas1010270107pmid 21059922

108 M di Porcia e Brugnera et al Modeling the impact of lianainfestation on the demography and carbon cycle of tropicalforests Glob Change Biol 25 3767ndash3780 (2019)doi 101111gcb14769 pmid 31310429

109 D S Pureswaran R Johns S B Heard D QuiringParadigms in eastern spruce budworm (LepidopteraTortricidae) population ecology a century of debate EnvironEntomol 45 1333ndash1342 (2016) doi 101093eenvw103pmid 28028079

110 G V J Virgin D A MacLean Five decades of balsam firstand development after spruce budworm-related mortalityFor Ecol Manage 400 129ndash138 (2017) doi 101016jforeco201705057

111 M Macek et al Life and death of Picea abies after bark-beetle outbreak Ecological processes driving seedlingrecruitment Ecol Appl 27 156ndash167 (2017) doi 101002eap1429 pmid 28052495

112 W F Laurance P Delamocircnica S G LauranceH L Vasconcelos T E Lovejoy Rainforest fragmentationkills big trees Nature 404 836 (2000) doi 10103835009032 pmid 10786782

113 D B Lindenmayer W F Laurance J F Franklin Globaldecline in large old trees Science 338 1305ndash1306 (2012)doi 101126science1231070 pmid 23224548

114 C J Peterson Catastrophic wind damage to North Americanforests and the potential impact of climate change SciTotal Environ 262 287ndash311 (2000) doi 101016S0048-9697(00)00529-5 pmid 11087033

115 C D Canham M J Papaik E F Latty Interspecific variationin susceptibility to windthrow as a function of tree sizeand storm severity for northern temperate tree speciesCan J For Res 31 1ndash10 (2001) doi 101139x00-124

116 J Boone B H Aukema J Bohlmann A L CarrollK F Raffa Efficacy of tree defense physiology varies withbark beetle population density A basis for positive feedbackin eruptive species Can J For Res 41 1174ndash1188 (2011)doi 101139x11-041

117 N G McDowell et al Predicting chronic climate-drivendisturbances and their mitigation Trends Ecol Evol 3315ndash27 (2018) doi 101016jtree201710002pmid 29146414

118 D J Johnson et al Climate sensitive size-dependent survivalin tropical trees Nat Ecol Evol 2 1436ndash1442 (2018)doi 101038s41559-018-0626-z pmid 30104751

119 H Pretzsch P Biber G Schuumltze E Uhl T Roumltzer Foreststand growth dynamics in Central Europe have acceleratedsince 1870 Nat Commun 5 4967 (2014) doi 101038ncomms5967 pmid 25216297

120 A B Nicotra et al Plant phenotypic plasticity in a changingclimate Trends Plant Sci 15 684ndash692 (2010) doi 101016jtplants201009008 pmid 20970368

121 S N Aitken S Yeaman J A Holliday T WangS Curtis-McLane Adaptation migration or extirpationClimate change outcomes for tree populations Evol Appl 195ndash111 (2008) doi 101111j1752-4571200700013xpmid 25567494

122 A L Angert et al Do speciesrsquo traits predict recent shifts atexpanding range edges Ecol Lett 14 677ndash689 (2011)doi 101111j1461-0248201101620x pmid 21535340

123 K M Miller B J McGill Land use and life history limitmigration capacity of eastern tree species Glob EcolBiogeogr 27 57ndash67 (2018) doi 101111geb12671

124 M Slot K Kitajima General patterns of acclimation of leafrespiration to elevated temperatures across biomes andplant types Oecologia 177 885ndash900 (2015) doi 101007s00442-014-3159-4 pmid 25481817

125 I Rieu D Twell N Firon Pollen development at hightemperature From acclimation to collapse Plant Physiol173 1967ndash1976 (2017) doi 101104pp1601644pmid 28246296

126 H D Adams et al Experimental drought and heat can delayphenological development and reduce foliar and shootgrowth in semiarid trees Glob Change Biol 21 4210ndash4220(2015) doi 101111gcb13030 pmid 26149972

127 C Grossiord et al Warming combined with more extremeprecipitation regimes modifies the water sources used bytrees New Phytol 213 584ndash596 (2017) doi 101111nph14192 pmid 27612306

128 J S McLachlan J J Hellmann M W Schwartz A frameworkfor debate of assisted migration in an era of climate changeConserv Biol 21 297ndash302 (2007) doi 101111j1523-1739200700676x pmid 17391179

129 F Lloret A Escudero J M Iriondo J Martiacutenez‐VilaltaF Valladares Extreme climatic events and vegetation Therole of stabilizing processes Glob Change Biol 18 797ndash805(2012) doi 101111j1365-2486201102624x

130 K B Kemp P E Higuera P Morgan Fire legacies impactconifer regeneration across environmental gradients inthe US northern Rockies Landsc Ecol 31 619ndash636 (2016)doi 101038nature15374 pmid 26466564

131 D Thom W Rammer R Seidl The impact of future forestdynamics on climate Interactive effects of changingvegetation and disturbance regimes Ecol Monogr 87665ndash684 (2017) doi 101002ecm1272 pmid 29628526

132 T L Powell et al Variation in hydroclimate sustains tropicalforest biomass and promotes functional diversity NewPhytol 219 932ndash946 (2018) doi 101111nph15271pmid 29923303

133 G B Bonan Forests and climate change Forcings feedbacksand the climate benefits of forests Science 320 1444ndash1449(2008) doi 101126science1155121 pmid 18556546

134 S Luyssaert et al Old-growth forests as global carbon sinksNature 455 213ndash215 (2008) doi 101038nature07276pmid 18784722

135 M E Harmon Carbon sequestration in forests Addressingthe scale question J For 99 24ndash29 (2001)

136 J E Campbell et al Large historical growth in globalterrestrial gross primary production Nature 544 84ndash87(2017) doi 101038nature22030 pmid 28382993

137 J T Randerson et al The impact of boreal forest fire onclimate warming Science 314 1130ndash1132 (2006)doi 101126science1132075 pmid 17110574

138 X Lee et al Observed increase in local cooling effect ofdeforestation at higher latitudes Nature 479 384ndash387(2011) doi 101038nature10588 pmid 22094699

139 N Devaraju G Bala A Modak Effects of large-scaledeforestation on precipitation in the monsoon regions Remoteversus local effects Proc Natl Acad Sci USA 112 3257ndash3262(2015) doi 101073pnas1423439112 pmid 25733889

140 Q Lejeune E L Davin B P Guillod S I SeneviratneInfluence of Amazonian deforestation on the future evolutionof regional surface fluxes circulation surface temperatureand precipitation Clim Dyn 44 2769ndash2786 (2015)doi 101007s00382-014-2203-8

141 Y Le Page D Morton B Bond-Lamberty J M C PereiraG Hurtt HESFIRE A global fire model to explore the role ofanthropogenic and weather drivers Biogeosciences 12887ndash903 (2015) doi 105194bg-12-887-2015

142 N G McDowell et al Evaluating theories of drought-inducedvegetation mortality using a multimodel-experimentframework New Phytol 200 304ndash321 (2013) doi 101111nph12465 pmid 24004027

143 D Lawrence et al The Land Use Model IntercomparisonProject (LUMIP) contribution to CMIP6 Rationale andexperimental design Geosci Model Dev 9 2973ndash2998(2016) doi 105194gmd-9-2973-2016

144 M Simard N Pinto J B Fisher A Baccini Mapping forestcanopy height globally with spaceborne lidar J Geophys ResBiogeosci 116 G04021 (2011)

145 V Lehsten M Mischurow E Lindstroumlm D LehstenH Lischke LPJ-GM 10 Simulating migration efficiently in adynamic vegetation model Geosci Model Dev 12 893ndash908(2019) doi 105194gmd-12-893-2019

146 B Smith I C Prentice M T Sykes Representation ofvegetation dynamics in the modelling of terrestrialecosystems Comparing two contrasting approaches withinEuropean climate space Glob Ecol Biogeogr 10 621ndash637(2001) doi 101046j1466-822X200100256x

147 R A Fisher et al Taking off the training wheels Theproperties of a dynamic vegetation model without climateenvelopes CLM45(ED) Geosci Model Dev 8 3593ndash3619(2015) doi 105194gmd-8-3593-2015

148 Y Chen et al Simulating damage for wind storms in the landsurface model ORCHIDEE-CAN (revision 4262) Geosci ModelDev 11 771ndash791 (2018) doi 105194gmd-11-771-2018

149 T K Henkel J Q Chambers D A Baker Delayed treemortality and Chinese tallow (Triadica sebifera) populationexplosion in a Louisiana bottomland hardwood forestfollowing Hurricane Katrina For Ecol Manage 378 222ndash232(2016) doi 101016jforeco201607036

150 J P Roccaforte et al Delayed tree mortality bark beetleactivity and regeneration dynamics five years following theWallow Fire Arizona USA Assessing trajectories towardsresiliency For Ecol Manage 428 20ndash26 (2018)doi 101016jforeco201806012

151 D Kennedy et al Implementing plant hydraulics in theCommunity Land Model version 5 J Adv Model Earth Syst11 485ndash513 (2019) doi 1010292018MS001500

152 P J Grubb The maintenance of species-richness in plantcommunities The importance of the regeneration niche BiolRev Camb Philos Soc 52 107ndash145 (1977) doi 101111j1469-185X1977tb01347x

153 B D Amiro et al Ecosystem carbon dioxide fluxes afterdisturbance in forests of North America J Geophys ResBiogeosci 115 G00K02 (2010)

154 J A Hicke et al Effects of biotic disturbances on forestcarbon cycling in the United States and Canada GlobChange Biol 18 7ndash34 (2012) doi 101111j1365-2486201102543x

155 J P Grime Plant Strategies and Vegetation Processes(Wiley 1979)

156 A Jentsch P White A theory of pulse dynamics anddisturbance in ecology Ecology 100 e02734 (2019)doi 101002ecy2734 pmid 31018013

157 D M Olson et al Terrestrial ecoregions of the world A newmap of life on Earth Bioscience 51 933ndash938 (2001)doi 1016410006-3568(2001)051[0933TEOTWA]20CO2

ACKNOWLEDGMENTS

Funding This manuscript was derived from the Department ofEnergy (DOE) Workshop ldquoDisturbance and vegetation dynamics inEarth System Modelsrdquo held March 2018 in Washington DCFunding was provided by the DOErsquos Next Generation EcosystemExperiment-Tropics (NGEE-Tropics) and Pacific NorthwestNational Labrsquos (PNNLrsquos) LDRD program (NGM) McIntire-StennisMIN-17-095 (BHA) the US Geological Surveyrsquos Ecosystemsand Land Resources mission areas (CDA) the Joint Fire ScienceProgram (16-3-01-4) and the University of WisconsinndashMadison



ay 28 2020


Dow

nloaded from

Vilas Trust (MGT) sabbatical fellowship support from sDiv(JWL) the Synthesis Centre of iDiv (DFG FZT 118) DOErsquosNGEE-Tropics and the University of Californiarsquos Laboratory FeesResearch Program (grant LFR-18-542511 to CX) the EuropeanResearch Council (ERC) under the European Unionrsquos Horizon 2020research and innovation program (grant 758873 TreeMort to

TAMP) paper number 43 of the Birmingham Institute of ForestResearch PNNLrsquos LDRD program (BB-L) sabbatical supportfrom Stanfords Center for Advanced Study in the BehavioralSciences and financial support from the Gordon and Betty MooreFoundation (RBJ) the NASA Carbon Monitoring System andNASA Interdisciplinary Science Programs (GCH and BP) Oak

Ridge National Laboratory is operated by UT-Battelle LLC undercontract DE-AC05-00OR22725 to the United States Department ofEnergy Competing interests The authors have no competinginterests to declare

101126scienceaaz9463



ay 28 2020


Dow

nloaded from

Pervasive shifts in forest dynamics in a changing world

Maria Uriarte Anthony P Walker and Chonggang XuLara Kueppers Jeremy W Lichstein Kiona Ogle Benjamin Poulter Thomas A M Pugh Rupert Seidl Monica G TurnerClark Michael Dietze Charlotte Grossiord Adam Hanbury-Brown George C Hurtt Robert B Jackson Daniel J Johnson Nate G McDowell Craig D Allen Kristina Anderson-Teixeira Brian H Aukema Ben Bond-Lamberty Louise Chini James S

DOI 101126scienceaaz9463 (6494) eaaz9463368Science

this issue p eaaz9463Sciencestable dynamics are dwindlingemerging pattern is that global forests are tending toward younger stands with faster turnover as old-growth forest withcontext of global climate change The authors show that shifts in forest dynamics are already occurring and the

thereview recent progress in understanding the drivers of forest dynamics and how these are interacting and changing in et althe forest community These processes are driven by disturbances both natural and anthropogenic McDowell

Forest dynamics are the processes of recruitment growth death and turnover of the constituent tree species ofShifting forest dynamics

ARTICLE TOOLS httpsciencesciencemagorgcontent3686494eaaz9463

REFERENCES

httpsciencesciencemagorgcontent3686494eaaz9463BIBLThis article cites 152 articles 25 of which you can access for free

PERMISSIONS httpwwwsciencemagorghelpreprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAASScienceScience 1200 New York Avenue NW Washington DC 20005 The title (print ISSN 0036-8075 online ISSN 1095-9203) is published by the American Association for the Advancement ofScience

Science No claim to original US Government WorksCopyright copy 2020 The Authors some rights reserved exclusive licensee American Association for the Advancement of

on May 28 2020

httpsciencesciencem

agorgD

ownloaded from

REVIEW

FOREST ECOLOGY

Pervasive shifts in forest dynamicsin a changing worldNate G McDowell1 Craig D Allen2 Kristina Anderson-Teixeira34 Brian H Aukema5Ben Bond-Lamberty6 Louise Chini7 James S Clark8 Michael Dietze9 Charlotte Grossiord10Adam Hanbury-Brown11 George C Hurtt7 Robert B Jackson12 Daniel J Johnson13Lara Kueppers1114 Jeremy W Lichstein15 Kiona Ogle16 Benjamin Poulter17 Thomas A M Pugh1819Rupert Seidl2021 Monica G Turner22 Maria Uriarte23 Anthony P Walker24 Chonggang Xu25

Forest dynamics arise from the interplay of environmental drivers and disturbances with the demographicprocesses of recruitment growth and mortality subsequently driving biomass and species compositionHowever forest disturbances and subsequent recovery are shifting with global changes in climate and landuse altering these dynamics Changes in environmental drivers land use and disturbance regimes areforcing forests toward younger shorter stands Rising carbon dioxide acclimation adaptation andmigrationcan influence these impacts Recent developments in Earth system models support increasingly realisticsimulations of vegetation dynamics In parallel emerging remote sensing datasets promise qualitatively newand more abundant data on the underlying processes and consequences for vegetation structure Whencombined these advances hold promise for improving the scientific understanding of changes in vegetationdemographics and disturbances

The interplay of vegetation demographymdashrecruitment growth andmortalitymdashwithenvironmental conditions and distur-bances drives forest dynamics of bio-mass function and species composition

(see Box 1 for definitions) In old-growth for-ests that approximate steady-state demograph-ics the recruitment growth and mortality oftrees are approximately balanced in contrastrapid recruitment often follows widespreaddisturbance-inducedmortality (1) Vegetationdynamics may now be changing because theenvironmental context in which plant demog-raphy and disturbances interact is shiftingwith anthropogenic change The interactionbetween episodic forest disturbances such aswindthrow or wildfire and chronically chang-ing drivers such as rising temperature vaporpressure deficit (VPD) and CO2 together withland-use change (LUC) (2) leads to both com-pounding and antagonistic impacts that alterdemographic rates (3) with consequences forterrestrial biogeochemical cycles and climate(4 5) Understanding the drivers of vegetation

dynamics is thus critical for accurate predic-tion of global terrestrial biogeochemistry underfuture conditions (6)The impacts of global change on forest

demographic rates may already be materializ-ing Inmature ecosystems tree mortality rateshave doubled throughoutmuch of theAmericasand in Europe over the past four decades (7ndash9)Simultaneously global carbon budgets indicateeither a growing or constant terrestrial carbonsink (10ndash12) which implies increased or con-stant vegetation production rates (13 14) How-ever satellite evidence suggests that forestsmight be switching from a CO2 fertilizationndashdominated period to a VPDndashdominated period(15) Terrestrial greening indices indicate a shiftfrom a CO2-driven increase in greenness inthe late 20th century to aVPD-drivendecrease inthe past decade (16) Thus increasing mortalitydue to anthropogenic changes and potentiallyincreasing or stable growth and recruitmentdue to CO2 fertilization (5) represent opposingprocesses that are co-occurring globally leav-ing the fate of future forests uncertain

In addition to changing vegetation dynam-ics in intact or relatively undisturbed forestsepisodic disturbances are tending to be largermore severe and in some regions more fre-quent under global climate change (17ndash20)Similarly the rates and types of LUC varywidely (21) but have on average increasedglobally in the past few centuries (2 22 23)Thus at the global scale disturbances and LUChave likely amplified tree mortality beyondwhat is suggested by the doubling of back-ground mortality rates in undisturbed forests(7ndash9) Current understanding of the net balanceof tree losses (mortality) and gains (recruitmentand growth) under a changing environmentcharacterized by more-extreme drivers anddisturbances is limited preventing predictionof whether recruitment and growth can bal-ance increased mortality rates in the futureTo evaluatewhether environmental changes

and increasing disturbances are causing glob-ally widespread shifts in vegetation demog-raphy we reviewed global observations ofrecruitment growth andmortality of forestsand woodlands Our expert-derived com-pilation of the state-of-the-art knowledge onvegetation dynamics their drivers and distur-bances allowed us to address four questions(i) Is there evidence for shifts in demographyover recent decades (ii) What physiologicaland disturbance-mediated processes underliethese demographic shifts (iii) What are the po-tential consequences of disturbance-mediatedchanges in demography for climate forcing(iv) How can global predictions of future vege-tation dynamics best be improved

Evidence for changing drivers and disturbancesand their impact on demography

Determining the impacts of changing driverson demography is difficult given the lack ofglobal observation platforms However evi-dence abounds from individual publishedstudies on the drivers and their impacts onplant communities and new modeling andobservational efforts now enable a more com-plete picture of disturbances and forest de-mography (24ndash26) In this section we firstexamine whether there are global trends instand ages and test the sensitivity of the stand-age distribution to changes in disturbance rate

RESEARCH


1Pacific Northwest National Laboratory Richland WA 99354 USA 2US Geological Survey Fort Collins Science Center New Mexico Landscapes Field Station Los Alamos NM 87544 USA3Conservation Ecology Center Smithsonian Conservation Biology Institute Front Royal VA 22630 USA 4Forest Global Earth Observatory Smithsonian Tropical Research Institute Republic ofPanama 5Department of Entomology University of Minnesota St Paul MN 55108 USA 6Joint Global Change Research Institute Pacific Northwest National Laboratory College Park MD 20740USA 7Department of Geographical Sciences University of Maryland College Park MD 20742 USA 8Nicholas School of the Environment Duke University Durham NC 27708 USA 9Departmentof Earth and Environment Boston University Boston MA 02215 USA 10Swiss Federal Institute for Forest Snow and Landscape Research WSL 8903 Birmensdorf Switzerland 11Energyand Resources Group University of California Berkeley Berkeley CA 94720 USA 12Department of Earth System Science Woods Institute for the Environment and Precourt Institute for EnergyStanford University Stanford CA 94305 USA 13School of Forest Resources and Conservation University of Florida Gainesville FL 32611 USA 14Division of Climate and Ecosystem SciencesLawrence Berkeley National Laboratory Berkeley CA 94720 USA 15Department of Biology University of Florida Gainesville FL 32611 USA 16School of Informatics Computing and CyberSystems Northern Arizona University Flagstaff AZ 86001 USA 17Biospheric Sciences Laboratory NASA Goddard Space Flight Center Greenbelt MD 20771 USA 18School of GeographyEarth and Environmental Sciences University of Birmingham B15 2TT Birmingham UK 19Birmingham Institute of Forest Research University of Birmingham B15 2TT Birmingham UK20Department of Forest and Soil Sciences University of Natural Resources and Life Sciences Vienna 1180 Vienna Austria 21School of Life Sciences Technical University of Munich 85354Freising Germany 22Department of Integrative Biology University of WisconsinndashMadison Madison WI 53706 USA 23Department of Ecology Evolution and Environmental Biology ColumbiaUniversity New York NY 10027 USA 24Environmental Sciences Division and Climate Change Science Institute Oak Ridge National Laboratory Oak Ridge TN 37831 USA 25Earth andEnvironmental Sciences Division Los Alamos National Laboratory Los Alamos NM 87545 USACorresponding author Email natemcdowellpnnlgov

on May 28 2020

httpsciencesciencem

agorgD

ownloaded from

















ay 28 2020


Dow

nloaded from









1900 1920 1940 1960 1980 2000Year

0

10

20

30

40

50F

ores

t are

a (m

illio

n km

2 )A



1900 1950 2000 2050 2100Year

5

10

15

20

CC

fore

st a

rea

(mill

ion

km2 )

B



1-1011-2021-3031-4041-5051-6061-7071-8081-9091-100101-110111-120121-130131-140O

G

Age class (years)

0

05

10

15

20

25

CC

you

ng fo

rest

are

a (m

illio

n km

2 )C

0

5

10

15

CC

OG


2)





ay 28 2020


Dow

nloaded from


Land-use change




Wildfire



0

2

4

6

8

10

12

14

For

est a

rea

(mill

ion

km2)

A Tropical



0

1

2

3

4

5

6

7

For

est a

rea

(mill

ion

km2)


1900 1920 1940 1960 1980 2000Year

0

2

4

6

8

10

12

For

est a

rea

(mill

ion

km2)

C Boreal




A

C

E

G

I

K

M N

L

J

H

F

D

B

VPD

Temp

[CO2]

VPD Temperature

Drought

Deforestation

Wildfire

Wind

Insect outbreaks

Growth

Growth

Growth

Growth

Growth

Growth

Growth

Mortality

Mortality

Mortality

Mortality

Mortality

Mortality

Mortality

Recruitment

Recruitment

Recruitment

Recruitment

Recruitment

Recruitment

Recruitment



ay 28 2020


Dow

nloaded from


Windthrow



Biotic agents












ay 28 2020


Dow

nloaded from














ay 28 2020


Dow

nloaded from




Outlook












9 4978 (2018) doi 101038s41467-018-07539-6pmid 30478255

























ay 28 2020


Dow

nloaded from







































































ay 28 2020


Dow

nloaded from

































































ACKNOWLEDGMENTS




ay 28 2020


Dow

nloaded from







ay 28 2020


Dow

nloaded from








REFERENCES






on May 28 2020

httpsciencesciencem

agorgD

ownloaded from

















ay 28 2020


Dow

nloaded from









1900 1920 1940 1960 1980 2000Year

0

10

20

30

40

50F

ores

t are

a (m

illio

n km

2 )A



1900 1950 2000 2050 2100Year

5

10

15

20

CC

fore

st a

rea

(mill

ion

km2 )

B



1-1011-2021-3031-4041-5051-6061-7071-8081-9091-100101-110111-120121-130131-140O

G

Age class (years)

0

05

10

15

20

25

CC

you

ng fo

rest

are

a (m

illio

n km

2 )C

0

5

10

15

CC

OG


2)





ay 28 2020


Dow

nloaded from


Land-use change




Wildfire



0

2

4

6

8

10

12

14

For

est a

rea

(mill

ion

km2)

A Tropical



0

1

2

3

4

5

6

7

For

est a

rea

(mill

ion

km2)


1900 1920 1940 1960 1980 2000Year

0

2

4

6

8

10

12

For

est a

rea

(mill

ion

km2)

C Boreal




A

C

E

G

I

K

M N

L

J

H

F

D

B

VPD

Temp

[CO2]

VPD Temperature

Drought

Deforestation

Wildfire

Wind

Insect outbreaks

Growth

Growth

Growth

Growth

Growth

Growth

Growth

Mortality

Mortality

Mortality

Mortality

Mortality

Mortality

Mortality

Recruitment

Recruitment

Recruitment

Recruitment

Recruitment

Recruitment

Recruitment



ay 28 2020


Dow

nloaded from


Windthrow



Biotic agents












ay 28 2020


Dow

nloaded from














ay 28 2020


Dow

nloaded from




Outlook












9 4978 (2018) doi 101038s41467-018-07539-6pmid 30478255

























ay 28 2020


Dow

nloaded from







































































ay 28 2020


Dow

nloaded from

































































ACKNOWLEDGMENTS




ay 28 2020


Dow

nloaded from







ay 28 2020


Dow

nloaded from








REFERENCES






on May 28 2020

httpsciencesciencem

agorgD

ownloaded from









1900 1920 1940 1960 1980 2000Year

0

10

20

30

40

50F

ores

t are

a (m

illio

n km

2 )A



1900 1950 2000 2050 2100Year

5

10

15

20

CC

fore

st a

rea

(mill

ion

km2 )

B



1-1011-2021-3031-4041-5051-6061-7071-8081-9091-100101-110111-120121-130131-140O

G

Age class (years)

0

05

10

15

20

25

CC

you

ng fo

rest

are

a (m

illio

n km

2 )C

0

5

10

15

CC

OG


2)





ay 28 2020


Dow

nloaded from


Land-use change




Wildfire



0

2

4

6

8

10

12

14

For

est a

rea

(mill

ion

km2)

A Tropical



0

1

2

3

4

5

6

7

For

est a

rea

(mill

ion

km2)


1900 1920 1940 1960 1980 2000Year

0

2

4

6

8

10

12

For

est a

rea

(mill

ion

km2)

C Boreal




A

C

E

G

I

K

M N

L

J

H

F

D

B

VPD

Temp

[CO2]

VPD Temperature

Drought

Deforestation

Wildfire

Wind

Insect outbreaks

Growth

Growth

Growth

Growth

Growth

Growth

Growth

Mortality

Mortality

Mortality

Mortality

Mortality

Mortality

Mortality

Recruitment

Recruitment

Recruitment

Recruitment

Recruitment

Recruitment

Recruitment



ay 28 2020


Dow

nloaded from


Windthrow



Biotic agents












ay 28 2020


Dow

nloaded from














ay 28 2020


Dow

nloaded from




Outlook












9 4978 (2018) doi 101038s41467-018-07539-6pmid 30478255

























ay 28 2020


Dow

nloaded from







































































ay 28 2020


Dow

nloaded from

































































ACKNOWLEDGMENTS




ay 28 2020


Dow

nloaded from







ay 28 2020


Dow

nloaded from








REFERENCES






on May 28 2020

httpsciencesciencem

agorgD

ownloaded from


Land-use change




Wildfire



0

2

4

6

8

10

12

14

For

est a

rea

(mill

ion

km2)

A Tropical



0

1

2

3

4

5

6

7

For

est a

rea

(mill

ion

km2)


1900 1920 1940 1960 1980 2000Year

0

2

4

6

8

10

12

For

est a

rea

(mill

ion

km2)

C Boreal




A

C

E

G

I

K

M N

L

J

H

F

D

B

VPD

Temp

[CO2]

VPD Temperature

Drought

Deforestation

Wildfire

Wind

Insect outbreaks

Growth

Growth

Growth

Growth

Growth

Growth

Growth

Mortality

Mortality

Mortality

Mortality

Mortality

Mortality

Mortality

Recruitment

Recruitment

Recruitment

Recruitment

Recruitment

Recruitment

Recruitment



ay 28 2020


Dow

nloaded from


Windthrow



Biotic agents












ay 28 2020


Dow

nloaded from














ay 28 2020


Dow

nloaded from




Outlook












9 4978 (2018) doi 101038s41467-018-07539-6pmid 30478255

























ay 28 2020


Dow

nloaded from







































































ay 28 2020


Dow

nloaded from

































































ACKNOWLEDGMENTS




ay 28 2020


Dow

nloaded from







ay 28 2020


Dow

nloaded from








REFERENCES






on May 28 2020

httpsciencesciencem

agorgD

ownloaded from


Windthrow



Biotic agents












ay 28 2020


Dow

nloaded from














ay 28 2020


Dow

nloaded from




Outlook












9 4978 (2018) doi 101038s41467-018-07539-6pmid 30478255

























ay 28 2020


Dow

nloaded from







































































ay 28 2020


Dow

nloaded from

































































ACKNOWLEDGMENTS




ay 28 2020


Dow

nloaded from







ay 28 2020


Dow

nloaded from








REFERENCES






on May 28 2020

httpsciencesciencem

agorgD

ownloaded from














ay 28 2020


Dow

nloaded from




Outlook












9 4978 (2018) doi 101038s41467-018-07539-6pmid 30478255

























ay 28 2020


Dow

nloaded from







































































ay 28 2020


Dow

nloaded from

































































ACKNOWLEDGMENTS




ay 28 2020


Dow

nloaded from







ay 28 2020


Dow

nloaded from








REFERENCES






on May 28 2020

httpsciencesciencem

agorgD

ownloaded from




Outlook












9 4978 (2018) doi 101038s41467-018-07539-6pmid 30478255

























ay 28 2020


Dow

nloaded from







































































ay 28 2020


Dow

nloaded from

































































ACKNOWLEDGMENTS




ay 28 2020


Dow

nloaded from







ay 28 2020


Dow

nloaded from








REFERENCES






on May 28 2020

httpsciencesciencem

agorgD

ownloaded from







































































ay 28 2020


Dow

nloaded from

































































ACKNOWLEDGMENTS




ay 28 2020


Dow

nloaded from







ay 28 2020


Dow

nloaded from








REFERENCES






on May 28 2020

httpsciencesciencem

agorgD

ownloaded from

































































ACKNOWLEDGMENTS




ay 28 2020


Dow

nloaded from







ay 28 2020


Dow

nloaded from








REFERENCES






on May 28 2020

httpsciencesciencem

agorgD

ownloaded from







ay 28 2020


Dow

nloaded from








REFERENCES






on May 28 2020

httpsciencesciencem

agorgD

ownloaded from








REFERENCES






on May 28 2020

httpsciencesciencem

agorgD

ownloaded from

forest ecology pervasive shifts in forest dynamics in a ......review summary forest ecology...

Documents