Download How Increasing CO2 and Climate Change Affect Forests

How Increasing CO2 and Climate Change Affect Forests
Author(s): Robin Lambert Graham, Monica G. Turner, Virginia H. Dale
Source: BioScience, Vol. 40, No. 8 (Sep., 1990), pp. 575-587
Published by: University of California Press on behalf of the American Institute of Biological Sciences
Stable URL: http://www.jstor.org/stable/1311298 .
Accessed: 05/08/2011 23:03
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].
University of California Press and American Institute of Biological Sciences are collaborating with JSTOR to
digitize, preserve and extend access to BioScience.
http://www.jstor.org
How
Increasing
CO2
and
Climate Change Affect
Forests
At many spatial and temporalscales, therewill be forest
responsesthat will be affectedby humanactivities
Robin LambertGraham,Monica G. Turner,and VirginiaH. Dale
T
he strongrelationship
among
climate, atmosphere, soils,
biota, and human activities
providesa solid basis for anticipating
changes in terrestrialbiomes in response to changesin the global environment.A major impact of the projected doubling of atmosphericCO2
and increasesin trace gas concentrations is the potential for global climate change (e.g., Tangley 1988).
Currentpredictions are for an average rise in global temperatureof 1.5?
to 4.5? C, with greater warming in
winter than summer and increased
warmingwith increasedlatitude.But
the magnitude, rate, and spatiotemporal characteristicsof the climatic
responseremainuncertain.Increased
precipitationin high latitudesand decreasedsummerprecipitationand soil
moisture in middle latitudes of the
northern hemisphere are also predicted.
Global forests are likely to experience dramatic changes as a consequence of elevated CO2 and climate
changes. Predictingthese changes is
Forest responses will
determinethe fate of
many species
processes (Table 1). Of course, many
forest responses cross scales, and their
not only a social and economicneces- assignment may be somewhat arbisity, but it is also a scientificchallenge trary. For example, forest succession,
becauseof the many spatial and tem- which takes decades or centuries and
poral scales involved.Forestsdirectly
affect climate at the global scale by
alteringthe earth'salbedo, hydrological regimes, and atmosphericCO2.
Forests also affect climate at a local
scale by alteringtemperature,humidity, and solar radiation.
Similarly,forest responsesoccur at
many scales. For example, processes
that involveexchangesof water,heat,
and CO2 at the leaf surfacehave fast
n sponsetimes (secondsto days).Forest growth and communitycomposition have intermediateresponsetimes
(decades to centuries), whereas the
geographicdistributionof forestsis a
long-termresponse (centuriesto millennia; Figure 1). Furthermore,all
Robin LambertGraham,Monica G. these responsescan be alteredby huTurner,andVirginiaH. Daleareresearch man intervention.
scientistsin the Environmental
Sciences How can the fate of forests be
Division,Oak Ridge NationalLabora- predictedin the face of what appears
tory, Oak Ridge, TN 37831-6038. Gra- to be an inevitablerise in atmospheric
ham is a forest ecologist interested in CO2 and temperature?It requiresexlong-termforest productivityand the use plicit considerationof the many temof remotely sensed data and geographic
and spatial scales at which forinformation systems in assessing forest poral
resources.Turneris a terrestrialecologist est responsesoccur, incorporationof
interestedin the theoreticaland applied the links betweenthese scales in both
problemsof describinglandscapepattern models and experiments,and a clear
and processes. Dale is a mathematical understandingof human impacts on
ecologist interestedin modelingeffectsof forest responses.
disturbanceson forest succession.
Four levels of biotic organization
September1990
(the biosphere,the biome, the ecosystem, and the tree) provide a useful
framework for examining forest responses to CO2 and climate. Each
level has a characteristicspatial and
temporal scale and set of ecological
is generallystudied at the ecosystem
level of organization, is a function
of species-specificresponses to environmentalvariables(measuredat the
tree level); competitive interactions
amongindividualsfor light, nutrients,
and water (measuredat the ecosystem
level); and seed source availability,
which may depend on species migration rates (measured at the biome
level).
The objective of this article is to
examine potential forest responses to
elevated CO2 in conjunction with cli-
mate change. For each level of biotic
organization (Table 1), we discuss the
potential effects of changed climate
and elevated CO2 on key ecological
processes, how human intervention
can affect those processes, and the
role of modeling in elucidating and
predictingforest responses.We conclude with a discussion of future research needs.
Biosphere
Severalprocessesare of interestat the
biospherescale. They are the various
feedbacks among atmospheric CO2
concentrations,climate, and carbon
storage in global forests (Figure 2).
575
ORNLDWG90M11337
w
m
0-j
C/)
z
w
z
z
0
UJ
C,)
w
0
I-
(c)
port the hypothesis that the observed
increase in the annual variation of
atmospheric carbon, especially in the
northern latitudes, is due to an increased growth rate in temperate and
boreal forests (Barcastow et al. 1985,
D'Arrigo et al. 1987).
The regulation of CO2 is but one of
the ways global vegetation influences
global climate. Vegetation modifies
energy fluxes by altering the earth's
albedo; vegetation generally has a
lower albedo than soil and consequently absorbs more radiant energy,
thus warming the earth.
w
x
Hydrologic regime. Evapotranspiration from vegetation is a major component of the global hydrologic regime (Shuckla and Mintz 1982). As
vapor pressure deficit increases exponentially with temperature, a warmer
climate will increase evaporative deYEARS DECADES CENTURIES MILLENNIA mand if air moisture is not greatly
DAYS
increased. Although laboratory studTIME
ies have shown that increased CO2
generally reduces stomatal conductances and thus water losses from
Figure1. Spatialand temporalscales of biotic levels of organizationiin forests.
transpiration, it does not necessarily
follow that elevated atmospheric CO2
Carbon cycle. Carbon, a major con- ways whose photosynthiesis process is will reduce global evapotranspirastituent of plants, is taken up from the not saturated at current atmospheric tion.
The impact of stomatal conducatmosphere as CO2 through photo- CO2 concentrations (Figure 3). Such
on transpiration water losses
tance
necessarnot
an
increase
when
tisreleased
and
might (but
plant
synthesis
sue respires or dies and either decom- ily) result in more carbon storage in depends on the degree to which the
poses or is burned. Because forests forest biomass and thus could serve as foliage is coupled to the atmosphere
contain 90% of the carbon in terres- a drag on atmospheric CO2 rise. Lim- (Jarvis and McNaughton 1986). In
trial vegetation, the accumulation of ited tree ring evidence suggests that laboratory tests, there is good airflow
forest biomass is a major regulator of forest growth in some locations has over the foliage. Consequently, the
increased during the past century in- foliage is tightly coupled to the atmoatmospheric CO2.
Increased concentrations of atmo- dependent of climate (LaMarche et al. sphere, and transpiration is sensitive
spheric CO2 could increase photosyn- 1984), although this conclusion is to CO2-mediated changes in stomatal
thesis rates of existing forests, be- open to debate (Kienast and Lux- conductance. However, at the recause forests are composed largely of moore 1988). In addition, global 3-D gional scale, foliage is poorly coupled
plants with C3 photosynthetic path- atmospheric tracer model studies sup- to the atmosphere. As a result, regional evapotranspirational water
Table 1. Four biotic levels of organization that participate in forest response to climate and CO2
losses are mostly a function of the
change.
radiation balance. Thus it has been
Level of
Temporal
Spatial
that regional or larger
hypothesized
Human activities
scale
scale
Major processes
organization
scale water losses will probably be
Years to
Globe
Energy, carbon, water
Deforestation, fossil
Biosphere
insensitive to changes in atmospheric
fuel burning
fluxes
millennia
CO2 (Jarvis 1989).
Plant breeding, land
Biome
Subcontinental Centuries to Evolution/extinction,
In addition, the absolute amount of
millennia
management,
migration, disturbance
water returned to the atmosphere via
conservation
-I
Ecosystem
Tree
576
10 to 104 ha
Years to
centuries
10-2 to 103 m2 Minutes to
decades
Disturbance, nutrient
cycling, production,
water use, succession,
competition
Phenology, reproduction,
physiological processes,
death
Pollution, exotic
pests, fire
suppression, flood
control, forest
management, soil
management
Fertilizing, watering,
weeding, breeding
evapotranspiration
is a function
of
leaf area. If climate change or elevated CO2 alters leaf area, this change
would have a profound effect on hydrologic regimes and river flows
(Wigley and Jones 1985).
Human impacts. Recent human activ-
BioScienceVol. 40 No. 8
ORNL DWG 89M-1841
O I/
7'\
CO2 andthe forest.
betweenatmospheric
Figure2. Thecomplexrelationships
ities are affecting the feedbacks
among CO2 concentrations, global
climate, and global vegetation. The
largest global net fluxes of carbon
from 1800 to the present are from
fossil fuel burning(150 to 190 x 1015
grams; Rotty 1987) and land use
changes,in particularthe conversion
of forestland (highcarbonstorage)to
agriculture(low carbon storage; 135
to 228 x 1015 grams of carbon;
Houghton et al. 1983).
The use of fossil fuel for energy is
heaviestin the temperatezone of the
northern hemisphere, whereas current land-use changes are most dramatic near the equator, where large
areas of tropical forests are being
cleared or degraded. However, human activitieswithin any region that
cause increases in atmosphericCO2
concentrationscan affect global climate patterns.
Extensive planting of intensively
managed forests or tree plantations
has been discussed as a method for
recapturing this "excess" atmosphericCO2 and thus slowing down
the increasein CO2 and consequently
climate change (Marland 1988).
However, as Jarvis (1989) pointed
out, it would take a new plantation
with an area twice the size of Europe
to absorbthe currentglobal fossil fuel
emissions. Furthermore, after 80
September1990
years, the carbon uptake ability of
such a plantationwould begin to decrease due to stand maturity,a new
plantationwould have to be initiated,
and the wood from the previousone
would be stored in perpetuity.
Switchingfrom fossil fuels to biomass-based fuels is a much more
effective approach for using forests
to slow down the buildup of excess
atmosphericcarbon. The net release
of carbon from a biomass fuel is
close if not equal to zero. The US
Congress Office of Technology Assessment (OTA 1980) estimatedthat
biofuels could potentially generate
6-17 quads of energy annually in
the United States. In 1988, the most
recent data available, US fossil fuel
energy use was approximately 71
quads.
and the redistributionof forest types
in responseto global changeare complex issues (c.f. Brubaker1986, Davis
and Botkin 1985, Pennington1986).
Some organismstrack climatechange
closely, reacting to conditions each
year, whereas others respond so
slowly that only long-term climatic
trendshave any observableeffect(Davis 1984). Tree species are expected
to migrategraduallyto trackhospitable environmentsas climate changes.
However, although the average rate
of tree species migration in Europe
and North America during the Ice
Ages was approximately300 m/year
(Woodward 1987), future migration
rates are difficultto predict, because
the projectedrate of climatechangeis
an order of magnitude greater than
previousclimate changes.
Davis (in press) suggests that the
spatial displacementof habitats can
be visualizedif one assumesthe same
general configuration of climate as
today. With a latitudinallapse rate in
the Great Lakes region of approximately 100 km/?C,isothermswould
be displacednorthward300 km in the
next 100 years. This rate is ten times
greater than the documented range
extensions for trees in a similartime
interval.Furthermore,there are now
new barriersto migrations (e.g., cities, agriculture,and roads) and new
modes of migration(e.g., cars, trains,
transplantsfor horticulture,forestry,
and agriculture.)
Range extension in the futuremay
be less efficientthan in the past, because advancedisjunctcolonies have
been extirpated by human disturbances, and propagule sources have
ORNL-DWG 90M-1797
*I50-
Biome
I-
/
The majorecologicalprocessesof in25
terest at the biome level are species 0_
A
migration,evolution, and extinction.
Disturbances, such as introduced U
/.
blights, occur at both the biome and
IC
I
I I
the ecosystemlevel.To avoid duplica900 1100
700
300
500
100
tion, we will review these disturATMOSPHERICCO2 (ppm)
bances in the discussion of ecosystems.
response
Figure3. Typicalphotosynthesis
of a C3 plant(sugarbeet)to increasing
CO2.
Migration. The migrationof species atmospheric
577
tions (e.g., city parks) far south of
their naturalrange. Common garden
tests, widely used with both tropical
and temperatedomesticatedtree species, could provide more specific informationon geneticvariabilityin the
climate response (Kellisonand Weir
1987).
Tree specieswith broad ranges,includingmost commercialspecies, are
most likelyto surviveclimatechange.
The manytree speciesthat are rareor
restricted in occurrence, including
115 of the 679 native tree species in
the United States (Little 1979), will
likely be at greaterrisk of extinction.
The evolutionaryresponseof trees
to increasingatmosphericCO2 is unknown. The within-species genetic
variabilityof CO2 sensitivityin CO2responsivetraitssuch as photosynthesis and stomatal control is unknown
also. However,short-termdifferences
in physiological responsiveness of
seedlingsand leaves of differenttree
species have been documented (e.g.,
Tolley and Strain1984).
Tree speciesare likely to have high
variabilityin CO2sensitivity,because
naturaltree populationshave shown
high variabilityin other characteristics, genetic variabilityin CO2 sensitivity has been demonstratedin herbaceous agricultural species, and
earlyevolutionof hardwoodtree species took place in the Cretaceousera
when atmospheric CO2 concentrations were three to ten times higher
than today (Gammon et al. 1985).
Positive changes in individual tree
water-useefficiency(increasein biomass per unit of watertaken up), as a
resultof stomataland photosynthesis
responsesto elevatedCO2concentrations, could also extend some species'
Evolution.Evolutionor extinctionof ranges into warmer, drier climates
tree species may be affected by cli- than are acceptable under current
mate changeor increasingconcentra- CO2 conditions.
tions of atmospheric CO2. The genetic resilienceof most tree speciesto Modeling. Severaldifferentmodeling
climatechangemust be inferredfrom approaches have predicted biomethe currentclimaticconditionswithin level responsesto changes in climate
their natural ranges (Johnson and and to increased atmosphericCO2.
Sharpe 1982). However, the bound- Biome changes at the scale of 0.5aries of a species' range may reflect degree by 0.5-degree grid cells were
competitive ability rather than cli- simulatedby using the Holdridgelife
matic constraints (Woodward 1987, zone classification and the current
c.f. Michaels and Hayden 1987) or and predictedtemperatureconditions
climaticrestraintsthat areonly signif- undera doublingof CO2 (Emanuelet
icant during the reproductivephase. al. 1985a, 1985b). The simulations
Many northern tree species grow predicted biome type changes
quite well in noncompetitive situa- throughout34% of the world's land-
often been reduced (Davis 1989a).
The current spatial distributionand
abundanceof a speciesis expectedto
influence its ability to migrate successfullyto regionsof suitableclimate
and soils (Petersand Darling 1985).
The rate of habitat displacement
and the rate of species dispersaland
migrationwill influencethe composition and structureof forested landscapes, especially if time lags occur
(e.g., Davis 1984, Pennington1986,
Webb 1986). Time lags in species
responses to historical climatic
changeshave been well documented.
For example, beech (Fagusgrandifolia), which has animal-dispersedseeds
and tends to move as a front, showed
a time lag of 500 to 1000 years in
crossingfrom the easternto the western shore of Lake Michigan (Davis
1989a). In contrast, hemlock (Tsuga
canadensis), whose wind-dispersed
seeds can travel 100 km beyond the
main species front, showed no time
lags attributableto crossingthe Great
Lakes.
Davis (1989a) also proposes that
time lags in the adjustmentsof species
abundancesto climate will be small
(approximately 10-20 yrs) in the
heavily disturbed communities that
cover most of the landscape. The
most common species will be dispersed to new habitats by humans,
but time lags will be a problem for
unmanaged forests, natural areas,
and preserves. The climate change
during the next century may not be
enough to kill dominant species directly. Thus forestedlandscapesmay
appear superficiallyunchangedeven
though a differentcommunityis becoming establishedin the understory.
578
mass. Boreal forest decreased by
37%, tundradecreasedby 32%, subtropical forest decreased by 22%,
subtropical thorn woodland increased by 37%, and subtropical
desertsincreasedby 26%.
In addition to temperatureeffects,
CO2 and soil moisture effects were
includedin a model that predictedthe
rangesof eight commercialnorthwest
North Americantree species under a
doubling of CO2 (Leverenzand Lev
1987). The model combined a projectedpositive effect of elevatedCO2
on the water-use efficiencyand photosynthesis of individual trees, with
regional-scalecalculationsof site water-balance and temperatureconditions predicted under a doubling of
atmosphericCO2. The rangesof lower-elevationspecies were projectedto
expand to higher elevations at the
expense of the currenthigh-elevation
species.The requirementsfor chilling
in several of the species, ratherthan
their physiological tolerance to
drought and heat, determinedtheir
low-elevation and southern-range
limits. As in the Emanuelmodel, migration, succession, and competition
are not included.
The usefulnessof both these models lies in their ability to provide insight into the rangeof possibleconsequences of climate change. As the
authorsthemselveshave pointed out,
the actual predictive value of each
model is quite limited.
The effectsof temperatureand precipitation change were predictedfor
forests across easternNorth America
by using FORENA, an individualbased forest stand model (Solomon
1986). The ability of the model to
projectsuccessionalshifts in response
to climatehas been corroboratedwith
historicpollen records(Solomonand
Webb 1985). The effects of climate
change on forest growth of mixedaged, mature forests were simulated
for 21 sites across boreal, deciduous,
and tundra biomes. The FORENA
model predictedslower growth rates
of most deciduous tree species
throughout much of their range; a
universal but centuries-longdieback
of the original forest; an invasion of
the southernboreal forest by temperate deciduous trees; a shift in the
generalpatternof forest biomes similar to the patternobtainedby Emanuel et al. (1985a, b); and, most interBioScienceVol. 40 No. 8
esting, time lags in shifts of community species and biome boundariesof
up to 1000 years from the initial
change in climate. Species migration
is not considered in FORENA, and
none of the models included the potential effects of naturaldisturbances
or human activities.
Human impact. Human activitiesare
expectedto influencemany forest responses. Activities with biome-level
impact include breeding programs,
land management,and active conservation. Breedingprograms, a viable
option for many commercialspecies,
may permit certain species to maintain theircurrentranges(Kellisonand
Weir 1987).
Landuse and ownershipwill affect
not only the migrationrate of species
but also their ultimate range and
abundance. For example, the predicted range of loblolly pine (Pinus
taeda)undera warmerclimatewould
extend well into Indiana and Ohio
(Solomon1986). However,the actual
rangeis not likely to extend that far,
becausethe northernland would still
be used for agriculture(Miller et al.
1987). Furthermore,the spatial displacement of forest types may alter
their ownership and thus their commercial productivity (Wallace and
Newman 1986). Human use of forests for fuel wood could also change
with a warmerclimate,therebyaltering rates of forest degradation in
some areas.
The effectivenessof parks and wildernessareas in preservingrare species may change if the spatial distribution of a biome shifts. Humans
may need to assist the migration of
some species if those species are to
survive (Peters and Darling 1985).
Active conservationis most likely to
be necessaryif the range of a species
occurs at the edge of a biome, if the
species has limited dispersal ability,
or if the habitat of the species is
disjointly distributed within the biome.
Ecosystem
Severalecological phenomena are of
special interest within the ecosystem
level of organization. These include
disturbance,nutrientcycling, competition and succession, production,
and water use.
September1990
Disturbanceregimes.The interaction
between climatic change and disturbance regimes,which can rapidlyalter forest structure,potentiallycould
be as importantas the directeffectsof
global warming in controlling shifts
in species distributionsand local extinctions.Global warmingmay result
in an increasedfrequencyof dry years
and a consequentincreasein the risk
of large fires (Sandenburghet al.
1987), perhaps increasingthe likelihood of fires of the magnitude observedin the northernhemispherein
1987 (e.g., China and North America) and 1988 (e.g., YellowstoneNational Park and elsewhere in North
America; Turner and Romme in
press).Fireregimesmay be more sensitive to climate change than some
forest processes, because fire is responsive to fuel moisture, which depends in turn on precipitation and
relativehumidity.The directeffectsof
climate change on forest production
and decomposition might be overshadowed by more drastic effects of
climate on fire regimes (Clark in
press).
Past climaticchangesof small magnitude have caused significant
changes in fire regimes (e.g., Clark
1988, 1990). In northwesternMinnesota, for example, fire was most frequent (approximately8.6-yearreturn
interval) during the warm, dry fifteenth and sixteenth centuries and
less frequent (approximately 13.2year return interval) during the
cooler, moister periods from 12401440 A.D. and during the period
(1640-1840 A.D.) called the Little Ice
However,the projectionof changesin
fire frequency,intensity, and severity
with a changingclimateremainschallenging because of the complex relationship among weather, fuel availability, and ignition (Turner and
Romme in press).
Biotic disturbancessuch as forest
pests and pathogensmay also change.
Because of their short reproductive
cycles,such organismsmay be able to
adapt or evolve much more quickly
than forests. Furthermore,the ranges
of forest pests and pathogens are often limited by climatic factors, and
speciesdistributionsmay changewith
climate. For example, spruce budworm populationsmay increasewith
a warmer climate because they are
favored by warm, dry, early spring
conditions(Haynes1982). Insectherbivory may also increase because of
CO2-inducedchanges in plant tissue
quality.In controlledtests, leaf-eating
insects that fed on agriculturalplant
tissuegrown underelevatedCO2generally had higher consumptionrates
per individualbecause of the poorer
plant tissue quality (high C:N ratio)
(e.g., Butleret al. 1986). Trees experiencingstress due to climate change
may also have lower resistance to
pests such as bark beetles (e.g., reducedresinexudation).Possibleshifts
in the range and relative competitiveness of weedy plants may be important to managed forests and nature reserves (Peters and Darling
1985).
Because long-lived trees may survive short-termclimatic fluctuations,
forest speciesthat are best adaptedto
the currentclimate may only be able
to become establishedin open habitats after disturbance events. Thus
forest composition may respond to
climatic changes primarilyafter disturbance (Dunwiddie 1986). For example,althoughthe ultimatecauseof
postglacial vegetation change in the
Pacific Northwest was climate
change, the proximatecause of some
vegetationchangeswas an alteredfire
regime (Cwynar 1987). It is likely
that a small changeto a drierclimate
triggereda relativelylarge change in
the disturbance regime (Cwynar
1987). The implicationsfor the present are extremelyimportant,because
Age (Clark 1990). Almost all major
forest fires in Mt. Rainier National
Park, Washington, since 1300 A.D.
corresponded with major droughts
that were reconstructedfor regions
east of the Cascade Range (Hemstrom and Franklin1982).
Fire regimes are sensitive to the
averageclimate over decades to centuries,as well as the interannualvariabilityof water balance(Clark1990).
In northwestern Minnesota, fires
were clustered during times of extendedlow effectiveprecipitationand
soil moisturestoragein the 1880s and
1910s (Clark 1989). Similarly, the
durationof dry periods is one of the
best predictorsof the area burnedin an altereddisturbance
regimemaybe
contemporary Canadian forests one of the earliestsignalsof climate
(Flannigan and Harrington 1988). change.
579
Competition and succession. Experimental and paleoecological studies
that indicate that competitive abilities
of species will shift as a result of
respecies-specific physiological
sponses to climate change and elevated CO2. In a microcosm study of
bottomland and upland tree communities, seedlings of different tree species were grown together for 90 days
under elevated CO2 and two levels of
light. Elevated CO2 had no effect on
the overall growth of seedlings in
either simulated community, although there were differential species
responses to CO2 within the communities (Williams et al. 1986). In another growth chamber study with
seedlings, a fast-growing pioneer tree
species showed a 79% increase in
biomass under an enriched CO2 atmosphere, whereas a slower-growing
climax species showed only a 30%
increase (Oberbauer et al. 1985).
Such studies suggest that successional paths may change and new
community types may evolve under
conditions of elevated atmospheric
CO2, just as they have in areas subjected to chronic ozone air pollution
(Taylor 1984). However, the specific
results of such studies with seedlings
cannot be interpreted as reliable predictions of future stand responses,
because other factors, such as temperature and evaporative demand, that
will change concurrently with elevated CO2 were not considered in the
designs of the experiments, nor were
the studies designed to consider effects on seedling establishment or ecosystem feedbacks.
Paleoecological studies also strongly
support the hypothesis that individual
species response to climate change
and elevated CO2 and seed source
availability will work to create completely new community types (Davis
1989b). For example, such studies
have identified community types such
as a spruce-oak woodland in North
America that existed in the past but
cannot be found now. The rate at
which communities will change is difficult to predict, because it depends
on not only competitive differences
but also species longevity, changes in
disturbance patterns, and seed source
availability.
Productivity. Site-specific ecosystem
productivity will change as a conse-
580
ORNL-DWG90M-1796
50
(A
4
E 40
o
0
E
IUJ 30
z
0 20
-
0
0
o
0
900
300
600
CO2 (pbar)
INTERCELLULAR
1200
Figure 4. The relationshipbetween photosynthesis and intercellularCO2 partial
pressurein two C3 speciesgrown at atmosphericCO2 partial pressureof 300 pLbars
(filledsymbols)or 900-1000 ,xbars(open symbols) (300 txbarsis the normalpartial
pressureof CO2 in the atmospherein Reno, Nevada, where the experimentstook
place). The arrows indicate the intercellularpartial pressure of CO2 when the
experimentalatmosphericCO2 was set equal to the partialpressureof CO2 at which
the plants were grown. Note that the adaptionto elevatedCO2 resultedin a greatly
reducedrateof photosynthesisin the cabbageplant (Brassicaoleracea)but an increased
rateof photosynthesisin the potato plant (Solanummelongena).(AdaptedfromSageet
al. 1989 with permission.)
the boreal forest duringthe past century is evidence of this type of relationship (Josza and Powell 1987).
Long-termdecreases in productivity
will occur in ecosystems where the
current most limiting environmental
factor becomes even more limiting.
Forestecosystemsin the southeastern
United States appearto be quite vulnerableto a warming,becausethereis
crease. In the short term, ecosystem currentlya delicate balance between
productivity will be a function of the potential evapotranspirationand acresponses of the trees currently pre- tual precipitation (Neilson et al.
sent and of the abundance and 1989).
The degree to which elevated CO2
growth rates of invading individuals.
Thus the short-term response is not will directly alter forest responsesis
necessarily indicative of the long-term the wild card in all climate change
scenarios. Species adapt to elevated
condition.
Long-term positive increases in CO2 in differentways. Some species
productivity are likely where current display high assimilation rates at
quenceof changesin multiplefactors,
such as species composition, altered
climateconditionsfor growth, and/or
CO2 fertilization effects (Leverenz
and Lev 1987). Furthermore,the direction of the change may shift over
time. For example, if ecosystem
change is initiated by catastrophic
disturbance, productivity will undoubtedlyfirst decreaseand then in-
growth is strongly limited by length
of the growing season. The correlation between warmer climate and an
increase in stem wood productivity in
twice ambient CO2, regardless of the
CO2 concentrations under which the
individualshave beengrown (e.g., the
potato plant in Figure 4). In other
BioScience Vol. 40 No. 8
species, the positive assimilation response to elevated CO2 is lost if the
individuals have been grown under
elevatedCO2 (e.g., the cabbageplant
in Figure4). For an increasein productivityin responseto elevatedCO2,
not only must the species be responsive to elevatedCO2 even if adapted,
but there must also be a carbon sink
for the extra photosynthate,that is,
the ecosystem must be environmentally capableof increasingits biomass
(Drakeet al. in press).
The complexities of the CO2 response are apparentin the published
reports of two long-term ecosystemlevel studieson the impactof elevated
CO2.Arctictussocktundra,an exceptionally environmentallylimited ecosystem, showed no increase in productivityduringthree years of in situ
exposure to double ambient CO2
(Oechel and Reichers 1986, Tissue
and Oechel 1987). Furthermore,laboratory studies with tundra species
have shown no positive growth responses to elevated CO2, even with
the addition of nutrients (Billingset
al. 1984, Oberbauer et al. 1986).
Thus, in the arctic tundraecosystem,
both species characteristicsand environmentallimitationsappearto keep
ecosystemproductivityfromresponding positivelyto elevatedCO2.
In contrast,a salt marshdominated
by a C3 species showed increased
productivityduringeach of the three
contiguousyearsthat it was subjected
in situ to elevated CO2 (Curtiset al.
1989a,b, Drake et al. 1989, in press).
Both above-ground and belowground productivitywere higher underelevatedCO2. Quite interestingly,
the increase in productivitywas not
only a consequenceof increased assimilationrates but also of decreased
respiration rates (Drake et al. in
press).These two studiessuggestthat
potential forest ecosystem responses
to elevated CO2 are also likely to be
complex.
Even if all other things were constant (which, of course, will not
be the case under a changed climate), it is quite doubtful that forest
productivity responses to elevated
CO2 would be uniform. Young or
aggrading forests would be more
likely than mature forests to display
increased productivity in response
to elevated CO2, because younger
stands should provide more of
September1990
a carbon sink for extraphotosyn- mental evidence that tree litter
thate.
lignin:N ratios, which are also negatively correlatedwith decomposition
Wateruse. It is difficultto predicthow rates and generallya better indicator
water use will changein specificeco- of decay rates than C:N ratios, resystems.Althoughwater-useefficiency main the same or decreaseundereleof individual seedlings subjected to vated CO2 (Norby et al. 1986). The
elevated CO2 generallyincreasesdue interaction of all these positive and
to stomatalandphotosynthesiseffects, negativeeffects and the difficultiesin
total ecosystemwateruse maystaythe extrapolating results across ecosyssame or increase,becausewater use is tems limits currentability to predict
a functionof manyfactorsin addition the overall impact of climate change
and elevated CO2 on decomposition
to efficiency.
Leaf area is positively correlated processes.
The effectof elevatedCO2on many
with stand water use; thus a CO2induced increase in stand leaf area aspects of internalnutrientcycling is
would increase water use, all other largely unknown for forest ecosysfactorsbeingequal (Jarvis1989). Wa- tems. No studies have followed the
ter use is also strongly related to effect of elevated CO2 on woody
evaporative demand, which will in- plant nutrition for longer than one
crease under a warmer climate. In growingseason. Limiteddata suggest
addition, water use is a function of that the return of nitrogen through
precipitationand timing of precipita- litter fall may remainconstantor detion. If climate change alters either, creasedue to an observedincreasein
then forestwateruse will change,and nutrient-useefficiency (Norby et al.
this changecould have profoundcon- 1986). Species shifts associated with
sequenceson streamflow (Wigleyand climate change may also change ecoJones 1985). Finally,if the forest can- system nutrient cycling patterns and
opy is dense and deep, and thus rates at specificsites (Pastorand Post
poorly coupled with the atmosphere, 1988).
water use will be largelya functionof
net radiationabsorbedby the canopy Modeling. Experimental studies of
rather than of canopy conductance the impactof climatechangeand elevated CO2 on forest ecosystems are
(Jarvisand McNaughton 1986).
limited by inabilityto easily manipuNutrient cycles. Climate change and late climate and atmosphere at the
elevated CO2 will affect forest nutri- spatial scale of a forest ecosystem.
ent cycles by alteringlitter decompo- Thereare no experimentalstudiesexsition rates, plant nutrient uptake, posing intact forests ecosystems to
and/or internal cycling. Decomposi- elevatedCO2with or withoutwarmer
tion rates will change in responseto climate conditions. Thus predicting
alterations in the physical environ- forest ecosystemresponseto changes
ment, litter quantity and quality, or in CO2 concentrations and climate
abundanceand typesof decomposers. requires extrapolating results from
A warmerclimate would tend to in- finer to broader scales, probably
crease the rate of decompositionby through modeling (Johnson and
enhancing fungal and bacterial Sharpe1982, Solomon 1986).
The effects of climatic change on
growth (Meentemeyer1978), but a
drier climate would tend to retard processessuch as net primaryproducdecomposition.Forest arthropods,a tion and decompositionmay be elucikey componentof the decomposition dated by ecosystem simulationmodprocess, could have their ranges al- els. Individual-basedstandsimulation
tered and their activitylevels affected models, which incorporate climatic
by changingplanttissueor litterqual- factors as driving variables for tree
ity (Kimball1985).
growth, have been widely used to
If the higher C:N ratio found in predict the response of forest stands
plant tissues grown under elevated to climate change (e.g., Davis and
CO2 results in higher C:N ratios in Botkin 1985, Pastor and Post 1988).
litter, then decomposition may be Model results have supportedpaleoslowed because decomposition rates ecologicalevidencethat speciesabunare negatively correlated with litter dances are not always in equilibrium
C:N ratios. However, there is experi- with climate (Woodward1987), and
581
competitiveinteractionscan strongly
influence community response (Solomon 1986). Simulations also suggest that there may be a transition
period of up to 200 years during
which time stand biomass drops (often later to recover) as the forest
developedunder the original climate
is replacedby species bettersuited to
the modifiedclimate(Solomon1986).
Empiricalmodels of the relationship
between climate and ecosystem processes such as decomposition (e.g.,
Meentemeyer1978) may also be useful, particularly at broad spatial
scales.
The combined impact of climate
change and elevated CO2 on forest
stand structureand growth has been
simulated using individual-based
models by incorporatinghypothetical
effectsof elevated CO2 on individual
tree growth. For example, Solomon
and West (1987) modifiedFORENA
by increasingthe maximumpotential
growth by 20% for deciduousspecies
and 11% for coniferous species and
by decreasing deciduous-treewater
use by as much as 18%. These modifications were designed to account
for possible productivityand wateruse efficiencyeffectsassociatedwith a
doublingof CO2.
Three forest types were simulated:
boreal forest, coniferous-deciduous
transitionforest, and southerndeciduous forest. Temperature was assumed to increase by 6? to 9? C in the
winter and 4? to 5? C in the summer,
and precipitationwas assumedto decrease by 25% in all but the boreal
site. The incorporationof CO2effects
modified the projected forest responses to climate by shorteningthe
transitionperiod,decreasingthe associated decline in biomass, and increasing the projected equilibrium
stand biomass by 50% in the waterand heat-stressedsoutherndeciduous
forest and by 16% in the coniferousdeciduoustransitionforest.
Because understandingof ecosystem responsesto CO2 is so limited,
the simulation results must be regardedas highly speculativeand useful largely for their heuristic value.
Nonetheless, if reliable information
becomes available,these models will
help us examinethe potentialecosystem implications of climate change
underconditionsof elevatedCO2.
The impact of elevated CO2 on
582
short-termcarbon assimilationin Picea sitchensis plantations was modeled by using MAESTRO,a physiologically based canopy model (Jarvis
1989). Althoughthe stand structure,
soil respiration,and leaf area parameters of the model were not modified
for the elevated CO2 simulation,the
photosynthesisand stomatalparameters were alteredon the basis of physiologicaldatacollectedfromseedlings
adapted to elevated CO2 concentrations. Such seedlings display somewhat lower photosynthesisratesthan
unadaptedseedlingsat a givenlevel of
CO2.Nonetheless,the modelstill predictedan increasein carbonassimilation at the stand level underelevated
CO2. The modeling exercise demonstratedthe importanceof considering
CO2 adaption effects in modeling
even at the largerscale of the canopy.
Models at differentspatial or temporal scales can be linked in various
ways. Forexample,ecosystemmodels
can be joinedwith standmodels,with
the ecosystem model defining available resources and the stand model
allocatingresourcesand growing individualtrees.Pastorand Post (1988)
exploredthe impactof climatechange
on mature forests of eastern North
Americaon a varietyof soils with an
individual-based stand simulation
modelin which the nitrogencycle and
water balance were explicitly modeled. In the simulations,productivity
and biomass increasedon soils that
retained adequate water but decreasedon soils with inadequatewater. The model results illustrate the
importanceof spatialheterogeneityof
soils in predictingforest responsesto
climatic change at the ecosystem
scale.
Human impact. Humans influence
forest ecosystems primarilythrough
forest management activities, although local air and water pollution,
poor soil management (erosion or
compaction), introduction of new
pests, and human modifications of
natural disturbanceregimesthrough
activitiessuch as firesuppressionand
flood control can also be important.
Silviculturaltools such as planting
stock, density of planting, mulch,
fertilization, and thinning regimes
may be used in managed stands to
counteractnegativeeffectsof climate
changeor to augmentpositive effects
(Sandenburghet al. 1987).
Forest response to elevated CO2
and climate change may be easier to
manage in mixed-speciesstands under uneven-agedmanagementthan in
monospecificstands undereven-aged
management.Individualtreeswill respond to climatechangeon a gradual
basis and at uneven rates. Unevenaged managementallows harvesting
in accordance with the differential
timing of those responses. In stands
undereven-agedmanagement,species
composition and genotype can be
switched only at planting time. In
short-rotation,intensive-cultureplantations, where genotypes and/or cultural regimes can be easily changed
every 10 to 20 years, the risk may be
minimal.However, even-agedstands
having longer rotation times will require decisive management with
muchmore risk. Alreadythereis concern within the United States timber
industry about the fate of loblolly
pine stands planted to the south and
west of their natural range if the
climate in that region becomes even
drierand hotter (Grahamet al. 1986).
Regionalland-usechangeswill also
influence forest ecosystem responses
(e.g., Turner 1987). Hierarchy and
landscapeecology theorysuggestthat
an ecosystem's response to disturbance may be a function of the system's setting in the larger landscape
(e.g., Turneret al. 1989b).
Direction of changes. Although ecosystem phenomena are likely to
change in response to elevated CO2
and climate change, the directionof
the changes will depend on highly
specificcircumstances.The prediction
of the impact of these changes on
forest resourceswill depend on integrating experimentalresults from a
finer scale and linking them through
concepts of ecosystem function
(Woodward1987). The most significant long-termeffect of elevatedCO2
and climate change on forest ecosystems is likely to be changesin disturbanceregimes(Turnerand Rommein
press) and in successionalpatternsin
the unmanaged,mixed-speciesstands
that dominateglobal forest resources
(Solomon et al. 1984). The role of
elevated CO2 in these changes is
highly uncertain because of our inability to predict whole, maturetree
responses to elevated CO2 and the
BioScienceVol. 40 No. 8
many feedbacks between CO2 and
ecosystem function and structure.
Both currentmodelingresultsand paleoecological evidence suggest that
during times of climate change, the
responseof naturalforests in the absence of disturbanceis delayed because of the longevity of the individual trees. Thus we may not perceive
climateeffectson communitycomposition until long after the climatehas
changed. Managed mixed-species
stands and plantations will also be
sensitiveto elevatedCO2 and climate
change, but with foresight, human
interventioncan potentially mitigate
many of the economically and socially detrimentalresponses.
quency of such events increases.
responses to these factors (Kramer
Information on the effect of ele- and Sionit 1987). The lack of longvated CO2 on treesis limitedprimar- term studies precludesmuch data on
ily to leaf-levelstudies, a 3-week ex- the impact of elevated CO2 on tree
posure of the canopy of 12-year-old life-cycleevents or phenology.
trees (Wong and Dunin 1987), and
short-term (two growing seasons) Modeling. Even at the much finer
studieson seedlingsor young saplings spatialand temporalscale of the tree,
of less than 30 tree species (Eamus modeling is essentialfor understandandJarvis1989). The resultsfromthe ing and predictingresponse to CO2
few longer-term studies are con- and climate.Tree size, longevity,and
founded by experimental artifacts the large numberof species preclude
(Jarvis 1989). Under well-watered taking an exclusively experimental
and fertilizedconditions,nearlyall C3 approach to evaluating individual
species, includingtrees, show signifi- tree responses.Models of physiologicantly greater growth with elevated cal processes such as photosynthesis
CO2 (Kramerand Sionit 1987).
and respiration(e.g., Farquharet al.
The usefulnessof this information 1980) can be used to test hypotheses
is limited,however,becausetreesgen- about the interactionsof temperature,
erallygrow underless than ideal con- CO2, radiation, and moisture on a
Tree
ditionsand for manyyears.Thus data leaf. Thesemodelsmay help elucidate
Processesof interest when consider- on the long-termresponseof trees to the mechanisms of
whole-tree reing individual trees are phenology, CO2 under conditions of nutrient sponses to climate change and allow
life-cycle events (e.g., reproduction stress,water stress,and light depriva- prediction of tree responses under
and death), and physiological pro- tion are needed. These data will be conditions
beyondthose of the expercesses(e.g., photosynthesis,transpira- difficultto obtain and will likely be iments.
tion, respiration, carbon allocation, limited to a few species.
nutrientuptake, and nutrient allocaAlthoughinitiallysome of the eco- Humanimpact.Humanscan influence
tion). The key challengeis to identify logical literature invoked Liebeg's individual trees by modifying either
the whole-tree effects of climate fac- Law of the Minimumto suggestthat the immediateenvironmentof the tree
tors and/or atmospheric CO2 with CO2 will have no beneficialeffect if
(e.g.,watering,fertilization,weed conand without other stresses (e.g., otherenvironmentalfactorsare limit- trol,
or pollution abatement)or the
drought, flooding, cold, shade, poor ing, short-termseedling studies have geneticmakeupof the tree (e.g.,breedsoil, and pollution).
repeatedlyfound that elevated CO2 ing or geneticengineering).Of the four
enhances growth even under condi- levels consideredin this
the
Climate. Current understanding of tions of nutrient, soil moisture, or individualtree seems mostarticle,
amenable
the effect of climate variables on light deprivation (e.g., Tolley and to
experimental study. Currently,
trees is based largely on qualitative Strain1984). In many cases, the relalong-termdata and data on life-cycle
observations of mature trees and tive response of stressed plants to eventsand
phenologicalresponsesare
short-term(minutesto days) experi- elevated CO2 exceeds that of unneeded.
especially
ments on leaves or seedlings. Exper- stressed plants. Whether this enimental studies on whole, mature hancementcan be maintainedover a
trees have been limited because of longer period is, however, unknown. Implicationsfor
future research
tree size and longevity. Mature-tree
Severallong-termfield studies curresponses to climate are generally rently underway should help resolve
and detecting forest reinferredby relatingobservedgrowth this most important issue. Elevated Predicting
sponses to the projected changes in
responses to past weather patterns CO2can also inducein seedlingssub- CO2 and the global climate
present
(e.g., Holdaway 1988). Field data on tle morphologicalchangessuch as in- significantchallengesthat require
cregrowth responses to climate are of- creasedbranching,increasedratio of ative solutions. Major issues include
ten qualitative, and growth data are fine to coarseroots, and greaternodchanges in the areal distributionof
generally lacking from the tropics ule mass on nitrogen-fixingspecies forests and forest types,
changes in
(Solomon et al. 1984). In particular, (e.g., Norby 1987). Tissue quality productivityor other functions
of inthereis little informationon how the also changeswith elevatedCO2; C:N terest, and the influenceof
landscape
frequencyof extreme climatic events ratios increase,and lignin contentde- heterogeneityon these responses.The
affectstree-levelprocesses. This lack creases(e.g., Williamset al. 1986).
challenges arise primarily from the
is especiallyimportant,becauseafter
The combined effect of elevated broad-scalenature of the questions,
the predicted climate change trees CO2and otherenvironmentalfactors, the
projectedrate of change, and the
may be exposed more frequently to such as temperature,humidity,or air manyenvironmentalinteractions
that
extremeevents. Treesmay survivean pollution, have only been studied at must be considered. The
and
types
infrequent extreme event such as a the individualleaf level in trees.Agri- combinationsof experimentsthat can
severedroughtbut experiencedelete- culturalstudies suggest that elevated be conducted become
increasingly
rious effects-even death-if the fre- CO2 may partially mitigate negative limited as one
goes to successively
September1990
583
Table 2. Timingof forestresponsesto global climatechangeand elevatedCO2.
Feature
Likelyresponse
Earlychanges(yearsto decades)
Disturbanceregimes
Seedlingestablishment
Ecotonemovement
Hydrologicregimes
Intermediateto long responsetime
(decadesto centuries)
Productivityrates
Changesin pools of carbonand
otheressentialelements
Species composition in interior
biomes
they are not, why. More information
on below-ground responses to elevated CO2 is also needed, as is information on tropical forest responses.
Changesin the frequency,intensity,and extent of
It may be prudent to focus experidisturbancessuch as fire,pest or pathogen
mental work on tree species that are
outbreaks,and stormeventsmay lead to rapid
economically important or characterchangesin forestedlandscapes.
istic of particular forest ecosystem
Changesin the relativeabundanceof seedlingsof
differentspecies(i.e., failedor decreased
types. Research on water use and
recruitment)suggestthe initiationof a longCO2 response is particularly relevant
termchangein speciescomposition.
for species that reside in regions such
Biome-levelchangesin the locationof boundaries
as the southeastern United States that
betweenforestedand nonforestedsystemsmay
are
likely to be water limited under
be observedearly,becausespeciesat the edges
of theirrangesare likelyto be less tolerantof
global warming. Experimental work
climaticchanges.
should also focus on variables that
Changesin precipitation,temperature,and stand
provide linkages between spatial
wateruse will alterrunoffpatterns,although
scales and biotic levels. For example,
potentialregional-levelchangesin water use
et al. (1986) focused on paNorby
are
not
well
understood.
efficiency
rameters linking tree-level physiological phenomena with ecosystem-level
decomposition processes. Because of
Forestproductivityover largespatialareaswill
the increased likelihood of drought in
changeslowly, becausechangesin species
some of the major temperate forest
compositionmay reducethe rateof changein
areas and the societal importance of
productivity.Increasedwateruse efficiencyin
water availability, the interactions betreesmay help mitigatepotentialproductivity
losses due to drought.
tween CO2 and water use and forests
As productivity,decomposition,and nutrient
at larger spatial scales needs particucyclingchangewith climate,storageof
lar
exploration.
in
essentialelements forestswill also change.
Treemigrationwill be a long-termresponse
becauseof time lags in the responsesof mature
trees.Matureindividualsmay persistfor
decades,even thoughthe compositionof the
understorychanges.
larger spatial and temporal scales. Future research should include monitoring, experimentation, and modeling.
Detection. Attempts to detect changes
in forests should focus on sensitive
processes and areas, recognizing the
different time scales over which responses will be observed (Table 2).
One of the earliest indications of forest response to global change may be
alterations in the frequency and intensity of disturbances. Therefore, particular emphasis should be placed on
increasing ability to predict the occurrence and effects of disturbances and
on monitoring disturbance-prone forests. Seedling establishment may be
particularly sensitive to climate
change. Thus tracking tree recruitment in natural stands may provide
an earlier indication of forest response than monitoring changes in
the overstory composition, which
may exhibit substantial time lags.
Efforts to predict changes and monitor the distribution and boundaries
of major ecosystem types should con584
tinue. In the absenceof disturbance,
the firstperceptiblechangesin global
vegetation patterns may be at the
transitionsbetween major life zones
(e.g., Nielson et al. 1989). The boreal
forest transitionzone and its alpine
counterpartmay be particularlysensitive. It is also criticalthat the effects
of changeson the variancein temperatureandprecipitationbe considered.
Mean values may not be meaningful,
because extreme events may affect
forestedareas even though the mean
remainsunchanged.
Experimental research. Experimental
research needs include quantitative
data on the water use, productivity,
nutrient uptake, and phenological responses of mature trees to long-term
exposure to elevated CO2, the sensitivities of tropical species to climate,
and the interactions between exposure to elevated CO2 and other environmental stresses. We need to know
if the adaption and acclimation responses observed in seedlings are
characteristic of mature trees and, if
Scaling. Research is especially needed
on the scaling problems that permeate
the entire discussion of forest responses to climate change and elevated CO2. Methods to extrapolate
information from site-specific studies
to the regional scale must be developed if the broad-scale questions are
to be answered.
Changes in the importance of the
variables controlling ecological processes at different scales should be
used to our advantage in developing
hypotheses and designing experiments (Turner et al. 1989a). Ecological responses differ with the temporal
and spatial scale at which they are
studied, and the controlling variables
at each scale are likely to be different.
As spatial extent increases, for example, the level of discernible detail, the
number of important variables, and
the potential for experimental manipulation all decrease (Meentemeyer
and Box 1987).
Important variables at the broad
scale tend to be abiotic rather than
biotic, and the variance of parameters
in the landscape may increase with
scales (Meentemeyer and Box 1987).
For example, the rates of litter decomposition at a particular site can be
predicted knowing the chemical/
BioScienceVol. 40 No. 8
physical properties of litter species,
but at broad spatial scales, climatic
variablesare sufficientto predictrates
(Meentemeyer1978, 1984). Thus we
may be able to substantiallysimplify
experimentsif we can identifya priori
the key controllingvariablesat a particular scale. The development of
good correlative relationships between abiotic and biotic variables,for
example, could significantly reduce
the numberof parametersthat must
be measuredthrough time to predict
forest responses.
Large-scalestudies. The dynamicsof
heterogeneouslandscapemust be understood to predict regional forest
responses.Global change is likely to
have substantialeffects on the structure and function of forested landscapes,yet the implicationsof spatial
heterogeneity over large areas are
only beginning to be understood
(Turner 1989). The traditional process-orientedview of ecosystemsmust
be complemented with other approaches as we seek to understand
forest responses at larger and larger
spatial scales. For example, manipulation of the factors that constraina
species distributionor an ecological
process may be more insightfulthan
obtaininga more detailedview of all
factors that influence the phenomenon of interest.
Cross-site research is essential to
understanding forest responses to
global change. Observationsor manipulative experiments can be conducted in forests across a globally
significantgradient(e.g., temperature
or precipitation gradients in North
America)to test hypothesesabout the
mechanisms controlling ecological
processesat broad scales. For example, studies conducted along a climatic gradientmight use a varietyof
carefully chosen sites as surrogates
for climatic change. One might test
hypothesesabout the abilityof particulartree speciesto withstanddifferent
climates, the time for recovery from
disturbances,or changes in the variance of biological or physicalparameters.Identicalperturbationscould be
imposed in differentforests across a
major temperature or precipitation
gradientin the United States and responses compared. Experimentscan
also be conducted along altitudinal
gradients across which climate
September1990
changesrapidlyin responseto elevation. Comparativestudiesof the same
phenomenaat differentscales would
aid in the developmentof extrapolation rules.
alteredby the displacementof forests
at the biome scale, and local economies will certainlybe affected.Forest
responsesto climate change and elevated CO2 will determinethe fate of
many of the species dependent on
Modeling. Currentmodelingcapabil- them, includinghumans.
ities should continue to be used to
explore hypotheses about ecological Acknowledgments
responses, identify sensitive parameters that should be monitored, and We appreciatethe helpfuland insightprovide a means to extrapolatelocal ful comments of three anonymous
responses to regional scales. Model- reviewers, our colleagues at Oak
ing is particularlyimportantbecause Ridge National Laboratory,and the
of the time lags in forest responses. BioScience editors. Research was
Because of the potential economic sponsoredby the EcologicalResearch
consequencesof climate change, it is Division, Office of Health and Envialso importantto predictforesteffects ronmentalResearch,US Department
that are relevantto managementac- of Energy,under contractDE-AC05tions. From the perspective of the 840R21400 with Martin Marietta
forest industry,informationmust be Energy Systems, Inc. This article is
geographically and temporally spe- PublicationNo. 3476, Environmental
cific.
Sciences Division, Oak Ridge NaA comprehensiveresearchprogram tional Laboratory.
should avail itself of currentcapabilities in remote sensing. The extrapo- Referencescited
lation of empiricalresults and simulation studies to large spatial scales Barcastow,R. B., C. D. Keeling, and T. P.
Wharf.1985. Seasonalamplitudeincreasein
requires methods to measure and
atmosphericCO2 concentrationsat Mauna
track changes in broad patterns of
Loa, Hawaii, 1959-1982. J. Geophys.Res.
90: 10,529-10,540.
vegetation (Iversonet al. 1989a) and
processes (e.g., gas flux, evapotrans- Billings,W. D., K. M. Peterson,J. 0. Luken,
and D. A. Mortensen.1984. Interactionof
piration, and production;Luvall and
increasingatmosphericcarbon dioxide and
Holbo in press, Runninget al. 1989).
soil nitrogenon the carbonbalanceof tundra
Remote sensing holds great promise
microcosms.Oecologia (Berl.)B65: 26-29.
as a means of gatheringsynoptic in- Brubaker,L. B. 1986. Responsesof tree populations to climatic change. Vegetatio 67:
formation on forest composition,
structure,and processesat landscape, 119-130.
G. D. Jr., B. A. Kimball, and J. R.
biome, and global scales (Iversonet Butler,
Mauney. 1986. Populations of Bemisia
al. 1989b). Remote-sensingdata may
tabaci (Hymoptera:Aleyrodidae)on cotton
be invaluablein addressingquestions
grown in open-top field chambersenriched
with CO2. Environ.Entomol. 15: 61-63.
regardingthe effect of landscapehetClark,
J. S. 1988. Effectof climatechangeon fir
erogeneityon forest responses.A varegimesin northwesternMinnesota.Nature
of
riety indices may be used to quan334: 233-235.
tify landscape patterns (Turner
. 1990. Patterns,causes, and theoryof
1989). By repeating such measure- fire occurrenceduring the last 750 yr in
northwesternMinnesota.Ecol. Monogr.60:
ments throughtime with satelliteim135-169
agery, changes in the spatial pattern
. 1989. Ecologicaldisturbanceas a reof forests can be quantifiedat broad
newalprocess:theoryand applicationto fire
scales. Remotesensingshould also be
history.Oikos 56: 17-30.
combined more frequentlywith em. In press. Twentieth century climate
change,fire suppression,and forestproducpirical studies so that measuredvalues of processes can be correlated tion and decomposition in northwestern
Minnesota.Can.J. For. Res.
with the spectral data (Cook et al.
Cook, E. A., L. R. Iverson,and R. L. Graham.
1989).
1989. Estimatingforest productivitywith
Humansinfluenceforestsacrossall
ThematicMapperand biogeographicaldata.
RemoteSens. Environ.28: 131-141.
spatial and temporal scales. In turn,
forest resources are critical for hu- Curtis,P. S., B. G. Drake,P. W. Leadley,W. J.
and D. F. Whigham. 1989a. Growth
mans. Forestsinfluencethe availabil- Arp,
and senescence in plant communitiesexity of water, serve many recreational posed to elevatedCO2 concentrationson an
purposes, and provide fiber, timber,
estuarinemarsh.Oecologia78: 20-26.
and fuel. National economiesmay be Curtis,P. S., B. G. Drake,and D. F. Whigham.
585
1989b. Nitrogenand carbondynamicsin C3
and C4 estuarinemarshplantsgrown under
elevated CO2 in situ. Oecologia 78: 297301.
Cwynar,L. C. 1987. Fireand the foresthistory
of the North CascadeRange. Ecology 68:
791-802.
D'Arrigo,R., G. C. Jacoby, and I. Y. Fung.
1987. Boreal forests and atmospherebiosphereexchangeof carbondioxide. Nature 329: 321-323.
Davis, M. B. 1984. Climaticinstability,time
lags, and communitydisequilibrium.Pages
269-284 in J. Diamondand T. J. Case,eds.
CommunityEcology.HarperandRow, New
York.
. 1989a. Lagsin vegetationresponseto
greenhousewarming.Clim.Change15: 7582.
.1989b. Insightsfrompaleolecologyon
global change. Bulletin of the Ecological
Societyof America70: 222-228.
. In press. Climatic change and the
survivalof forest species. In G. M. Woodwell, ed. The Earth in Transition:Patterns
and Processes of Biotic Impoverishment.
CambridgeUniversityPress,New York.
Davis,M. B., and D. B. Botkin.1985. Sensitivity of cool-temperateforestsand their fossil
pollen recordto rapid temperaturechange.
Quat. Res. 23: 327-340.
Drake,B. G., P. W. Leadley,W. Arp,P. Curtis,
and D. Whigham. In press. The effect of
elevated atmosphericCO2 on C3 and C4
vegetationon ChesapeakeBay. Proceedings
of a Symposiumon PhysiologicalEcologyof
Aquatic Plants, August 30-September 2,
1988. Botanik Institute,Universityof Aarhus, Denmark.
Drake, P. G., P. W. Leadley,W. J. Arp, D.
Nassiry,and P. S. Curtis.1989. An open top
chamberfor field studiesof elevatedatmosphericCO2concentrationon saltmarshvegetation.Funct.Ecol. 3: 363-371.
Dunwiddie,P. W. 1986. A 6000-yearrecordof
forest history on Mount Rainier,Washington. Ecology 67: 58-68.
Eamus,D., and P. G. Jarvis.1989. The direct
effectsof increasein the global atmospheric
CO2concentrationon naturaland commercial temperatetrees and forest. Adv. Ecol.
Res. 19: 1-55.
Emanuel,W. R., H. H. Shugart,and M. P.
Stevenson.1985a. Climaticchangeand the
broad-scaledistributionof terrestrialecosystem complexes.Clim. Change7: 29-43.
. 1985b. Response to comment: climatic change and the broad-scaledistribution of terrestrial ecosystem complexes.
Clim.Change7: 457-460.
Farquhar,G. D., S. von Caemmerer,andJ. A.
Berry.1980. A biochemicalmodelof photosyntheticCO2 assimilationin leavesof a C3
species.Planta 149: 78-90.
Flannigan,M. D., andJ. B. Harrington.1988.
A study of the relation of meteorological
variablesto monthlyprovincialarea burned
by wildfirein Canada.Journal of Applied
Meteorology27: 441-452.
Gammon, R. H., E. T. Sundquist,and P. J.
Fraser.1985. History of carbondioxide in
the atmosphere.Pages 25-62 in J. R. Trabalka,ed. AtmosphericCarbonDioxide and
the Global Carbon Cycle. DOE/ER-0239.
National Technical Information Service,
586
Springfield,VA.
Graham,R. L., T. R. Fox, and P. M. Dougherty.1986. Multiplestressfactors:thepotential role of system models in assessingthe
impactof multiplestresseson forestproductivity.Pages151-170 in T. C. Hennessey,P.
M. Dougherty, S. V. Kossuth, and J. D.
Johnson,eds. Stress Physiologyand Forest
Productivity.MartinusNijhoff Publ., Dordrecht,The Netherlands.
Haynes,D. L. 1982. Effectsof climatechange
on agriculturalplant pests. Environmental
and Societal Consequencesof a Possible
C02-InducedClimateChange.vol. II, part
10. DOE/EV/10019-10.National Technical
InformationService,Springfield,VA.
Hemstrom,M. A., and J. F. Franklin.1982.
Fireand other disturbancesof the forestsin
Mount RainierNational Park. Quat. Res.
18: 32-51.
Holdaway,M. R. 1988. The relationshipbetween tree diametergrowth and climate in
the lakestates.Pages490-497 in R. A. Ek, S.
R. Shifley,andT. E. Birk,eds. Proceedingsof
the ForestGrowthModellingand Prediction
IUFRO Conference.vol. 1. Minneapolis,
MN, 23-28 August1987. USDAForestService General Technical Report NC-120,
North CentralExperimentStation,St. Paul,
MN.
Houghton,R. A., J. E. Hobbie,J. M. Mellillo,
B. Moore, B. J. Peterson,G. R. Shaver,and
G. M. Woodwell. 1983. Changes in the
carboncontentof terrestrialbiota and soils
between 1860 and 1980: a net release of
CO2 to the atmosphere.Ecol. Monogr.53:
235-262.
Iverson,L. R., E. A. Cook, and R. L. Graham.
1989a. A techniquefor extrapolatingand
validatingforest cover across large regions:
calibratingAVHRRdatawith TM data.Int.
J. RemoteSens. 10: 1805-1812.
Iverson,L. R., R. L. Graham,and E. A. Cook.
1989b. Applicationsof satelliteremotesensing to forestedecosystems.LandscapeEcol.
3: 131-143.
Jarvis,P. G. 1989. Atmosphericcarbondioxide
and forests. PhilosophicalTransactionsof
the RoyalSocietyof London332: 369-392.
Jarvis,P. G., and K. G. McNaughton.1986.
Stomatalcontrolof transpiration:scalingup
from leaf to region. Adv. Ecol. Res. 15:
1-49.
Johnson, W. C., and D. M. Sharpe. 1982.
Ecologicalconsequencesof CO2-inducedclimaticchangeon forestecosystems.Environmentaland SocietalConsequencesof a Possible C02-InducedClimateChange.vol. II,
part 13. DOE/EV/10019-12.NationalTechnical InformationService,Springfield,VA.
Josza, L. A., and J. M. Powell. 1987. Some
climaticaspectsof biomassproductivityof
white sprucestem wood. Can. J. For. Res.
17: 1075-1079.
Kellison,R. C., and R. J. Weir.1987. Breeding
strategiesin foresttreepopulationsto buffer
againstelevatedatmosphericcarbondioxide
levels.Pages285-294 in W. E. ShandsandJ.
S. Hoffman, eds. The GreenhouseEffect,
ClimateChange,and U.S.Forests.The ConservationFoundation,Washington,DC.
Kienast,F., and R. J. Luxmoore.1988. Treeringanalysisandconifergrowthresponsesto
increasedatmosphericCO2levels.Oecologia
76: 487-495.
Kimball,B. A. 1985. Adaptionof vegetation
and managementpracticesto a higher carbon dioxide world. Pages 185-204 in B. R.
Strainand J. D. Cure,eds. Direct Effectsof
IncreasingCarbon Dioxide on Vegetation.
DOE/ER-0238.NationalTechnicalInformation Service,Springfield,VA.
Kramer,P. J., and N. Sionit. 1987. Effectof
increasingcarbon dioxide concentrationon
the physiology and growth of forest trees.
Pages 219-246 in W. E. Shands and J. S.
Hoffman,eds. The GreenhouseEffect, Climate Change,and U.S.Forests.The Conservation Foundation,Washington,DC.
La Marche, V. C. Jr., D. A. Graybill,H. C.
Fritts, and M. R. Rose. 1984. Increasing
atmosphericcarbon dioxide: tree ring evidence for growth enhancementin natural
vegetation.Science225: 1019-1021.
Leverenz,J. W., and D. J. Lev. 1987. Effectsof
carbondioxide-inducedclimatechangeson
the naturalrangesof six majorcommercial
tree species in the western United States.
Pages 123-155 in W. E. Shandsand J. S.
Hoffman,eds. The GreenhouseEffect, Climate Change,and U.S. Forests.The Conservation Foundation,Washington,DC.
Little,E. L.Jr. 1979. Checklistof UnitedStates
Trees(Nativeand Naturalized).Agricultural
Handbook541. US Departmentof Agriculture,Washington,DC.
Luvall, J. C., and H. R. Holbo. In press.
Thermal remote sensing methods in landscape ecology. In M. G. Turnerand R. H.
Gardner, eds. Quantitative Methods in
New
LandscapeEcology. Springer-Verlag,
York.
Marland,G. 1988. The prospectof solvingthe
CO2 problemthroughglobal reforestation.
DOE/NBB-0082.National TechnicalInformationService,Springfield,VA.
Meentemeyer,V. 1978. Macroclimateand lignin control of litter decompositionrates.
Ecology59: 465-472.
. 1984. The geographyof organicdecompositionrates.Annalsof the Association
of AmericanGeographers74: 551-560.
Meentemeyer,V., and E. 0. Box. 1987. Scale
effectsin landscapestudies.Pages 15-34 in
M. G. Turner,ed. LandscapeHeterogeneity
and Disturbance. Springer-Verlag,New
York.
Michaels,P. J., and B. P. Hayden.1987. Modeling the climate dynamics of tree death.
BioScience37: 603-610.
Miller, W. F., P. M. Dougherty,and G. L.
Switzer.1987. Effectof risingcarbondioxide
andpotentialclimatechangeon loblollypine
distribution,growth, survival,and productivity.Pages157-187 in W. E. ShandsandJ.
S. Hoffman, eds. The GreenhouseEffect,
ClimateChange,and U.S.Forests.The ConservationFoundation,Washington,DC.
Neilson, R. P., G. A. King,R. L. DeVelice,J.
Lenihan,D. Marks,J. Dolph, B. Campbell,
and G. Glick. 1989. Sensitivityof ecological
landscape and regions to global climatic
change. US Environmental Protection
Agency,Corvallis,OR.
Norby,R.J. 1987. Nodulationandnitrogenase
activityin nitrogen-fixingwoodyplantsstimulated by CO2 enrichmentof the atmosphere.Physiol.Plant. 71: 77-82.
Norby,R.J., J. Pastor,andJ. M. Melillo.1986.
Carbon-nitrogen interactions in CO2-
BioScienceVol. 40 No. 8
enrichedwhite oak: physiologicaland longtermperspectives.TreePhysiol.2: 223-241.
Oberbauer,S. F., N. Sionit,S. J. Hastings,and
W. C. Oechel. 1986. Effectsof CO2 enrichmentand nutritionon growth,photosynthesis, and nutrient concentrationof Alaskan
tundraplant species.Can.J. Bot. 64: 29932998.
Oberbauer,S. F., B. R. Strain,and N. Fetcher.
1985. Effectof CO2-enrichment
on seedling
physiologyand growth of two tropicaltree
species.Physiol.Plant. 65: 352-356.
Oechel, W. C., and G. H. Reichers. 1986.
Impactsof increasingCO2 on naturalvegetation, particularlythe tundra.Pages36-42
in C. Rosenzweig and R. Dickinson, eds.
Climate-VegetationInteractions: Proceedings of a Workshop.27-29 January1986.
Report OIES-2. Office for Interdisciplinary
Earth Studies, University Corporation for
AtmosphericResearch,Boulder,CO.
Officeof TechnologyAssessment(OTA).1980.
Energyfrom BiologicalProcesses.US Congress,US GovernmentPrintingOffice,Washington,DC.
Pastor,J., and W. M. Post. 1988. Responseof
northern forest to C02-induced climate
change.Nature 344: 55-58.
Pennington,W. 1986. Lags in adjustmentof
vegetationto climatecaused by the pace of
soil development: evidence from Britain.
Vegetatio67: 105-118.
Peters,R. L., and J. D. S. Darling. 1985. The
greenhouseeffect and nature reserves.BioScience35: 707-717.
Rotty, R. M. 1987. Estimatesof seasonalvariation in fossil fuel CO2 emissions. Tellus
39B: 184-202.
Running, S. W., R. R. Ramakrishna,D. L.
Peterson,L. E. Band,D. F. Potts,L. L. Pierce,
and M. A. Spanner.1989. Mappingregional
forest evapotranspirationand photosynthesis by couplingsatellitedata with ecosystem
simulation.Ecology 70: 1090-1101.
Sage,R. F., T. D. Sharkey,and J. R. Seemann.
1989. Acclimationof photosynthesisto elevated CO2 in five C3 species. Plant Physiol.
89: 590-596.
R., C. Taylor,andJ. S. Hoffman.
Sandenburgh,
1987. Risingcarbondioxide,climatechange,
and forest management:an overview.Pages
113-121 in W. E. ShandsandJ. S. Hoffman,
eds. The Greenhouse Effect, Climate
Change,and U.S. Forests.The Conservation
Foundation,Washington,DC.
Shuckla,J., and Y. Mintz. 1982. Influenceof
land-surface evapotranspiration on the
earth'sclimate.Science215: 1498-1501.
Solomon,A. M. 1986. Transientresponseof
foreststo CO2-inducedclimatechange:simulation modeling experiments in eastern
North America.Oecologia (Berl.)68: 567579.
Solomon,A. M., M. L. Tharp,D. C. West, G.
E. Taylor, J. W. Webb, and J. L. Trimble.
1984. Response of unmanagedforests to
CO2-inducedclimate change: available information, initial tests and data requirements. DOE/NBB-053.National Technical
InformationService,Springfield,VA.
Solomon,A. M., and T. WebbIII.1985. Computer-aidedreconstructionof late Quaternary landscapedynamics.Ann. Rev. Ecol.
Syst. 16: 63-84.
Solomon,A. M., and D. C. West. 1987. Atmo-
September1990
sphericcarbondioxide change:agent of futureforestgrowthor decline?Pages23-38 in
J. G. Titus, ed. Effectsof Changesin StratosphericOzone and ClimateChange.vol. 3.
ClimateChange.Proceedingsof the International Conferenceon Health and Environmental Effects of Ozone Modificationand
ClimateChange.UNEPandUSEnvironmental ProtectionAgency, Paris and Washington, DC.
Tangley, L. 1988. Preparing for climate
change.BioScience38: 14-18.
Taylor,G. E. 1984. The significanceof elevated
levels of ozone on natural ecosystems of
North America.Pages152-175 in S. D. Lee,
ed. Evaluationsof the ScientificBasis for
Ozone/Oxidants Standards. Air Pollution
ControlAssociation,Houston,TX.
Tissue, D. T., and W. C. Oechel. 1987. Responseof Eriophorumvaginatumto elevated
CO2and temperaturein the Alaskantussock
tundra.Ecology68: 401-410.
Tolley,L. C., and B. R. Strain.1984. Effectsof
CO2 enrichmenton growthof Liquidamber
styracifluaand Pinus taeda seedlingsunder
differentirradiancelevels. Can. J. For. Res.
14: 343-350.
Turner,M. G. 1987. Landuse changesand net
primaryproductionin Georgia,USA, landscape: 1935-1982. Environ. Manage. 11:
237-247.
. 1989. Landscapeecology:the effectof
patternon process. Annu. Rev. Ecol. Syst.
20: 171-197.
Turner,M. G., V. H. Dale, and R. H. Gardner.
1989a. Predictingacrossscales:theorydevelopment and testing. Landscape Ecol. 3:
245-259.
Turner,M. G., R. H. Gardner,V. H. Dale, and
R. V. O'Neill. 1989b. Predictingthe spread
of disturbancein heterogeneouslandscapes.
Oikos 55: 121-129.
Turner,M. G., and W. H. Romme. In press.
Landscapedynamicsin crownfireecosystems.
In R. D. Lavenand P. M. Omi, eds. Pattern
and Processin CrownFireEcosystems.Princeton UniversityPress,Princeton,NJ.
Wallace, T. D., and D. H. Newman. 1986.
Measurementof ownershipeffectson forest
productivityin North Carolinafrom 1974 to
1984. Can.J. For. Res. 16: 733-738.
Webb,T., III.1986. Is vegetationin equilibrium
with climate?How to interpretlate-Quaternarypollendata. Vegetatio67: 75-91.
Wigley, T. M. L., and P. D. Jones. 1985.
Influenceof precipitationchangesand direct
CO2 effects on streamflow. Nature 314:
149-152.
Williams,W. E., K. Garbutt,F. A. Bazzaz,andP.
M. Vitousek.1986. The responseof plantsto
elevatedCO2. IV. Two deciduous-forest
tree
communities.Oecologia(Berl.)69: 454-459.
Wong,S. C., and F. X. Dunin. 1987. Photosynthesisandtranspirationof treesin a Eucalypt
forest stand: CO2, light and humidity responses. Aust. J. Plant. Physiol. 14: 619632.
Woodward, F. I. 1987. Climate and Plant
Distribution. CambridgeUniversity Press,
New York.
_f
?
Control the Environment,
Effectively and Efficiently
With PERCIVAL
Biological and
nsect rearing
incubators, plant
::::'"
?
, '
.
s
;
growth and dew
A
N-
..
...
chambers. Table
't
models or giant
walk-ins. Percival's
A
6
I
yr(i
extensive knowledge
and experience is
evidentthroughout
i
.;
I
.
-
i
our quality line. And
Percival flexibility
offers the options
and modifications for your specific needs.
Write today for our catalog, or call collect
515-432-6501 to discuss your requirements.
Since 1886
_
Percival Manufacturing Company
P. 0. Box 249 * 1805 East FourthStreet * Boone, Iowa 50036
587