JOURNAL OF GEOPHYSICAL
RESEARCH, VOL. 102, NO. D24, PAGES 29,029-29,041, DECEMBER
26, 1997
Carbon distribution and aboveground net primary production
in aspen, jack pine, and black spruce stands in Saskatchewan
and Manitoba, Canada
S. T. Gower,• J. G. Vogel, 1 j. M. Norman, 2 C. J. Kucharik, 2 S. J. Steele,•'3
and T. K. Stow •
Abstract. The objectivesof this studyare to (1) characterizethe carbon(C) content,leaf
area index,and abovegroundnet primary production(ANPP) for mature aspen,black
spruce,and young and mature jack pine standsat the southernand northern Boreal
Ecosystem-Atmosphere
Study(BOREAS) areasand (2) comparenet primary production
and carbon allocationcoefficientsfor the major boreal forest typesof the world. Direct
estimatesof leaf area index, defined as one half of the total leaf surfacearea, range from
a minimum of 1.8 for jack pine foreststo a maximum of 5.6 for black spruceforests;stems
comprise5 to 15% of the total overstoryplant area. In the BOREAS study,total
ecosystem(vegetationplus detritusplus soil) carbon content is greatestin the black spruce
forests(445,760-479,380
kg C ha-l), with87 to 88%of the C in the soil,andis lowestin
thejackpinestands
(68,370-68,980
kg C ha-1) witha similardistribution
of carbonin the
vegetationand soil. Forestfloor carboncontentand mean residencetime (MRT) alsovary
more amongforest typesin a studyarea than between studyareas for a forest type; forest
floor MRT range from 16 to 19 years for aspen standsto 28 to 39 years for jack pine
stands.ANPP differs significantlyamong the mature forestsat each of the BOREAS study
areas,ranging
froma maximum
of 3490to 3520kg C ha-• yr-1 for aspenstands
to 1170
to 1220kg C ha-l yr-• forjackpinestands.
Bothnetprimaryproduction
(NPP)and
carbon allocation differ between boreal evergreen and deciduousforests in the world,
suggestingglobal primary productionmodels should distinguishbetween these two forest
types. On average,56% of NPP for boreal forestsoccursas detritus and illustratesthe
need to better understandfactorscontrolling abovegroundand below-grounddetritus
productionin boreal forests.
Polglase[1995] coupleda soil and vegetationcarbonmodel and
examined the effects of climate on carbon budgetsof tundra,
Boreal forestsare of great interestto Earth systemscientists boreal, and tropical forests;their analysessuggestthat boreal
becausethey are important in the global carbon cycle.Boreal forestsoscillate from being a carbon sink to a carbon source.
forestscontainthe secondlargestpercentageof the total world
A major focus of BOREAS is to examine the exchangeof
soil carbon of any biome and contain 31% of all forest soil
CO2 betweenterrestrialecosystems
and the atmosphere[Sellcarbon (calculatedfrom Schlesinger,
1991). North temperate
ers et al., 1995]. A better understandingof the influence of
and boreal forests are believed to be a net C sink (i.e., net
environmentaland ecologicalfactorson the major components
storageof carbon in terrestrial ecosystems),thus helping to
of NEE (net primary production (NPP) and heterotrophic
offset risingatmosphericCO2 concentration[Tanset al., 1990;
respiration) is required to determine the causesfor temporal
Ciaset al., 1995;Keelinget al., 1996]. Other scientists,however,
and spatialvariation in NEE. Despite the importanceof boreal
have questioned these conclusionsand suggestthat boreal
forests
in the globalcarbonbudgetthere are few completenet
forests may not be a large carbon sink [Quay et al., 1992;
primary productionbudgetsfor boreal forestsand no synthesis
Siegenthaler
and Sarmiento,1993].
The role of boreal forestsin the global carbon cycle is de- of the data is currentlyavailable.This paper focusesprimarily
on NPP.
termined by the net exchangeof CO2 between the terrestrial
Species/soilinterrelationships,climate, and stand age influecosystemand the atmosphere,commonlyreferred to as net
ecosystemexchange(NEE). Net ecosystemexchangeis the ence NPP of forests.Speciescompositionof the boreal forest
balance between net primary production,both above ground is simplerelative to other forest biomes[Landsbergand Gower,
and below ground, and heterotrophicrespiration. Wang and 1997] and is stronglyinfluencedby soil type. Aspen occurson
fine textureduplandsoils,while jack pine occurson excessively
•Departmentof ForestEcologyand Management,Universityof drained, sandyupland soils.Black sprucecommonlyoccurson
Wisconsin, Madison.
poorly drained, organiclowland soils,but it can also occur on
2Department
of SoilScience,
Universityof Wisconsin,
Madison.
upland mineral soils. Water and nutrient availability differ
SNowat KempNaturalResourceStation,Woodruff,Wisconsin.
among these contrastingsoils and influences both NPP and
Copyright 1997 by the American GeophysicalUnion.
carbon allocation [Gower et al., 1995]. Large differences in
carbonallocationmay affect the soil carboncyclebecausethe
Paper number 97JD02317.
0148-0227/97/97J D-02317509.00
different plant tissuesdecomposeat very different rates. CarIntroduction
29,029
29,030
Table
GOWER
1.
Select Climate
Characteristics
ET AL.: BOREAL
FOREST CARBON
for the Southern
studyarea(SSA) andthe northernstudyarea(NSA). The SSA
is near the southern boundary of the boreal forest in
Saskatchewan,
Canada.Permafrostdoesnot occurin the top
2 m of soil in any of the stands.The NSA is located --•40-60
km west of Thompson,Manitoba, Canada. The soilsare derivedfrom parentmaterialdepositedby GlacialLake Agassiz.
Permafrostoccursin poorly drained forestsand fens in the
StudyArea (SSA) and Northern StudyArea (NSA)
Climate Characteristic
Mean Januaryair temperature,øC
Mean July air temperature,øC
Mean annual precipitation,mm
SSAa
NSA b
-19.8
+17.6
405
-25.0
+15.7
536
BUDGETS
aData basedon a 48 year averagefrom PrinceAlbert, Saskatchewan. NSA. Winters are less harsh and summers are hotter and drier
bDatabasedon a 23yearaverage
fromThompson,
Manitoba.
bon allocationpatterns change as a stand ages, and these
changesmay directlyor indirectlycauseforestproductivityto
declinein older stands[Goweret al., 1996;Ryanet al., 1996].
Climate influencescanopyarchitectureand leaf physiology
which determine radiation interceptionand utilization efficiency,importantdeterminantsof canopyphotosynthesis
[McMurtrie et al., 1994;Landsberget al., 1997].
The primary objectiveof this study is to characterizethe
distributionof carbonand aboveground
net primaryproduction (ANPP) for three dominantforest typesat the southern
and northern Boreal Ecosystem-Atmosphere Study
(BOREAS) areas.A secondobjectiveis to compareNPP data
for BOREAS and other borealforeststo determineif general
NPP or carbon allocation "principles"exist for major functional typesof boreal forests.
Methods
Study Areas
The studywas conductedat two major studyareas:Prince
Albert National Park (PANP), Nipawin ProvincialForest region in centralSaskatchewan,
and Thompson,Manitoba, Canada. For brevitywe refer to the two studyareasasthe southern
Table
2.
Select Stand Characteristics
in the SSA than NSA (Table 1).
At each studyarea, we establishedfour replicateplots in
mature aspen and black spruceand young and mature jack
pine stands.One objectiveof the BOREAS studyis to estimate
NEE for the different terrestrial ecosystemsand understand
the processescontrollingthe exchangeof CO2 between the
BOREAS tower sites and the atmosphere.To achievethis
objectivewe locatedfour replicateplots immediatelyoutside
the footprintof the eddyfluxtowers,in forestrepresentative
of
that insidethe footprint. The size of the plots in each stand
vary dependingon the tree density.The numberof treesper
plot rangefrom 60 to 140 (Table 2). The largenumberof trees
and the randomlocationof the four replicateplotsare sufficientto adequatelycharacterizethe forest.Species,crownclass
[Danielet al., 1979],and diameterat breastheight(dbh) (1.37
m) were tallied for all trees in eachplot. Table 2 summarizes
selectstandcharacteristics
for the eight stands.
The young(25-27 years) and old (60-65 years)jack pine
(PinusbanksianaLamb.) standsare even-agedand originated
from wildfires. Bearberry (Arctostaphylosuva-ursi [L.]
Sprenge.)and reindeerlichen(Cladinamitis [Sandst.]Hale &
Culb) are the dominantgroundcoverin the youngstandsand
reindeerlichen forms nearly a completegroundcoverin the
mature standsat both the SSA and the NSA. Isolatedclumps
of greenalder (Alnuscrispa[Ait.] Pursh)with a feathermoss
for Forest Tower Flux Sites at NSA and SSA
Mean
ForestType
Tree
Basal
PlotSize,
Dominant
Ageb
Height,
Trees,
Area,
m
Vegetation
a
years
mc
ha-1
m2 ha-l
13.8
1,960
DBH, cm
Mean
Range
NSA
Aspen
25 x 25
TAd
15 x 15
W
JP, GA ½,PB½
BS
TA, Wc
JP
(NA)
Old blackspruce
(NBS)
Youngjack pine
7.5 x 7.5
53
155
26.7(96)f
12.5
280
10
5,450
20
15,160
O.7(2)
0.4 (2)
35.6(98)
0.8(2)
7.2 (100)
5.2
23.5
8.5
20.2
2.1
(4.5-23.2)
(3.0-13.7)
(--)
(2.5-18.6)
(15.4-25.0)
(0.5-6.1)
10.3
1,280
30
13.3(97)
0.4(3)
11.1
12.3
(3.0-18.7)
(7.1-18.0)
980
33.5(100)
20.5
(9.9-31.5)
27.2(87)
2.7 (8)
1.6(5)
11.9(100)
7.1
18.3
11.3
3.2
(2.1-21.1)
(14.1-22.7)
(5.1-16.1)
(0.8-10.9)
16.9(100)
12.9
(3.0-23.2)
9.1
2.9
(NYJP)
Old jackpine
(NOJP)
25 x 25
JP
PB
63
Old aspen
(SA)
Black spruce
(SBS)
30 x 30
TAd
67
20.1
15 x 15
115
7.2
Young jack pine
(SYJP)
Old jack pine
(SOJP)
7.5 x 7.5
BS,
JP
ET, Wc
JP, GA•
25
3.7
5,900
100
150
10,670
25 x 25
JP
65
12.7
1,190
SSA
•TA, tremblingaspen;W, willow;JP,jack pine;GA, greenalder;PB, paperbirch;BS,blackspruce;ET, easterntamarack.
bDeterminedfrom stem diskstaken at the soil surfacefor three dominant and codominanttrees in each stand.
CEstimated
from dbh/heightrelationshipappliedto standinventorydata.
dAspen
andbalsam
poplarnotdifferentiated
in standinventory.
½Less than 1% of total basal area.
fPercent of total stand basal area.
GOWER ET AL.: BOREAL FOREST CARBON BUDGETS
29,031
(Pleuroziumspp.) ground cover occur in the mature standsat
The stemsof youngjack pine treeswere cut into 1 m sections
both studyareas [Vogel,1997]. Paper birch (Betulapapyrifera and weighedon an electronicbalance,while the stemsof large
Marsh.) occursin green alder clumpsin the NSA. The soil in trees(all other stands)were cut into 2 m sectionsand weighed.
both jack pine standsis an excessivelydrained coarse sand. A 2 cm thick disk was cut from the base of each stem section
Coarse fragments occupy 20-30% of the soil volume in the and weighedon an electronicbalance.Stem disksalsowere cut
NSA but are absent in the SSA. The soil in the NSA is a
from the base of the tree, breastheight (1.37 m), and baseof
Dystrocreptand the soil in the SSA is a Eutrocrept.
live crownto determinesapwoodcross-sectional
area. The dry
The aspenstandsare even-agedand consistlargelyof trem- biomassof eachstemsectionwascalculatedby multiplyingthe
bling aspen(Populustrernuloides
Michx.), althougha few bal- wet mass of each stem section by the dry-wet ratio of each
sampoplar (PopulusbalsarniferaL.) are present.Hazel (Cory- stemdisk.The total dry biomassof eachstemwascalculatedby
lus cornutaMarsh) forms a continuousunderstory,reachinga summingthe dry biomassof all stem sections.Stem area was
height of 1.5-2.0 m, at both studyareas.Willow (Salix spp.) calculatedfor eachtree by assumingthat the stemfrom dbh to
also occursin the understoryat the NSA. The soilsat both the base of live crown was a paraboloid and from dbh to the soil
SSA and the NSA are moderately drained loam to clay loam surfacewas a cylinder. Equations for each stem shape were
and are classified as Boralfs.
obtainedfrom Husch e! al. [1982].
The black sprucestand at the SSA is dominated by black
All fresh sampleswere storedat 4ø-6øCin walk-in coolersat
spruce(Picearnariana[Mill.] BSP), but tamarack(Larix lari- the end of eachday. Sampleswere transportedfrom Canada to
cina [Du Roi] K. Koch) and jack pine occur in the better Madison, Wisconsin, in an insulated, ice-cooled truck. In Maddrained areas.Black spruceis the only overstoryspeciesin the ison, all sampleswere immediatelyplacedin a walk-in, forced
NSA. Several different black spruce forest communities are air dryingoven and dried at 70øCto a constantmass.Sapwood
present at both the SSA and the NSA, but we restricted our area diskswere stored at 2øC until processed.The sapwoodanalysesto the black spruce-feathermoss
community.Domi- heartwood boundary was determined by placing a Q-beam
nant shrub speciesinclude Labrador tea (Ledurn groenlandi- 2001 (Quantum Devices Inc, Barneveld, Wisconsin) light
curn Oeder) and wild rose (Rosa spp.). Feathermoss(Pleu- sourcebehind each sapwooddisk and marking the boundary
roziurnspp.) is the dominantgroundcover at both studyareas. with a permanentpen; this techniqueprovideda more reliable
The soil at the SSA is a 20-30 cm deep peat layer over a and rapid estimateof sapwoodarea than stains.The sapwood
coarse-textured mineral soil, and the soil at the NSA black image was photocopiedon clear acetate sheetsand measured
sprucestandis a 30-50 cm peat layer over a clay mineral soil. by using a Li-Cor 3100 area meter (Li-Cor Inc., Lincoln, Nebraska). Sapwoodvolume was calculatedfollowingprocedures
Allometric Equations
describedby Gower et al. [1993]. A disk of known area was
Site specificallometric equationswere developed for the copied periodically with the sapwooddisks and measured to
dominant overstoryspeciesin each stand.In August 1994, two ensurethat copyingdid not introduce a bias in sapwoodarea
dominant, three codominant,three intermediate, and two suppressedtrees were destructivelysampledimmediatelyoutside
the replicateplots.Trees with severedefectswere avoided.All
treeswere harvestedfrom late July to early August.Trees were
cut at the soil surface,and total tree height and length of the
estimates.
Aboveground Vegetation Biomass and Net Primary
Production
Abovegroundbiomassof tree components(stem, new and
old branch, and new and old foliage), sapwoodvolume, and
thirds (top, middle, and bottom), and all live and dead stem and leaf area were determinedfrom allometricequations
branchesfrom each positionwere cut and weighedseparately. that were developed for the major overstory speciesin each
One live branch from each canopypositionwas randomly se- stand. Standing dead biomass(stem plus branch) was estilected for detailed analysisin the field immediately after the mated from the stem and branch allometric equations.The
tree was felled.
allometric equationsfor aspenwere used for balsam poplar,
Aspen branch sampleswere divided into foliated twig and willow, and paper birch, while the allometric equationsfor
nonfoliage-bearing
branches.Jackpine brancheswere divided tamarackin northernWisconsin[Kloeppel,1997] were usedto
into current,1, 2, 3, and _>4year old shoots(needlesplustwig) estimate biomass and leaf area for tamarack in the SSA. The
and nonfoliagebearing branches.Black sprucebrancheswere use of nonsitespecificallometric equationsintroducedonly a
divided into current, 1-2, 3-4, and >5 year old shoots and small error becausethe minor speciestogether comprisedless
nonfoliage-bearingbranches.Approximately30 to 50 foliage- than 3% of the total standbasal area in each stand, exceptfor
bearing shootswere selectedfrom each shoot age classand the SSA black spruce stand where tamarack and jack pine
canopypositionfor detailed measurements.The fresh massof comprised13% of the total standbasalarea. Vegetation wood
each componentwas determined by using an electronic bal- and foliage biomasswere multiplied by 0.50 and 0.45, respecanceand the samplewasplacedin a labeledbag and storedin tively, to estimate C content [Atjay et al., 1977].
Aboveground NPP was calculated as the sum of annual
a cold room (4øC) until transportedto Wisconsin.Five to ten
shootsfrom each subsamplewere used for specificleaf area biomassincrement(AB) and detritusproduction(D). Overmeasurement
and moisture content determination
of the leaf
storybiomassincrementwasmeasuredin variableradiusplots
and wood tissuefor each subsample,and the remainingshoots usinga 10 basal area factor prism. One to two variable radius
were dried at 70øC to a constantmass.Specificleaf area was plots, depending upon tree density,were establishedinside
measuredusing a volume displacementmethod [Chen et al., eachof the four fixedarea replicateplots.Variable radiusplots
thisissue].The needlesurfacearea (one-halfthe total leaf area were used to estimate NPP becausethey provide an unbiased
or hemisurfacearea) of each individualtree was calculatedas subsampleof trees. We did not core all the trees in each plot
the product of foliage mass and specificleaf area for each becausewe wantedto maintain a portion of the fixed area plot
needle age classand canopyposition[Fassnachtet al., 1994]. free from tree coringfor long-term measurements.Two increlive crown
were
measured.
The
live crown
was marked
into
29,032
ment
GOWER
cores
were
removed
ET AL.: BOREAL
at 1.37 m from
each
tree
FOREST CARBON
in the
variable radiusplots in October 1994.Annual stemwoodradial
increment for each year between 1986 and 1994 was measured
using a dual axis optical micrometer (M-2001-D series)coupled to a SpaldingB5 digital position displaysystem(7109C225, Gaertner Scientific Corp., Chicago, Illinois).
Abovegroundwood (stem plus branch) and foliage biomass
were calculatedfor each year using allometric equationsand
the recreated tree dbh for the past 10 years. Aboveground
detrituswas measuredusingten 40 x 60 cm litter screensthat
were placed in random locationsin each of the four replicate
plots. Litter screenswere deployedin August 1993 and litter
was collectedin May, July,August,and October 1994 and May,
July, August 1995.
Understory Aboveground Biomass and NPP
BUDGETS
tissuewas ground to passthrough a 40 mm wire mesh and a
10 g subsamplewasdry ashedat 450øCfor 24 hoursto estimate
percent mineral soil content in each forest floor sample;this
value was used to correct
each forest floor biomass value for
mineral soil contamination. Forest floor organic matter was
assumedto be 50% carbon [Atjayet al., 1977].
Mineral
soil carbon
data were obtained
from
BOREAS
in-
vestigators.Soil carbon content was calculated for each soil
horizon from horizon depth, bulk density,and percent organic
carbon data and summed
for all horizons
to estimate
total soil
carbon content. If more than one soil pit was establishedin a
stand,we computedan averagefor the pits. Soil carbon content was standardized
to 70 cm because this was the minimum
soil depth usedby all BOREAS scientists.For mostsitesthe 70
cm depth includedthe A and B horizonsand occasionallythe
C horizon.
We did not differentiate
between
forest floor and
Understory abovegroundbiomassand net primary produc- mineral soil carbon in the black sprucestandsbecauseof the
tivity were measuredin small plots that were randomlylocated difficultiesin developinga satisfactoryprotocolfor distinguishinsideor immediatelyoutsideeachof the four replicateplotsin ing surfacedetritus from dead feathermoss.
each stand. The area of the understory plots ranged in size
from 2 x 2 m for black spruceand jack pine standsto 1 x 1 m Statistical Analysis
The equation, lOglo Y = a + b(log•o X), was used to
for aspenstands.All abovegroundunderstoryvegetationtissue
was removed from each small plot and stored in a cold room correlatethe dependentvariable(Y, componentbiomass,sap(2ø-4øC) until the tissuewas processed.Vegetationwas sepa- wood volume,or leaf and stemarea) to stemdiameter (X). A
transformationwas used to correct for nonhomograted into annual herbs, new twig, and new foliage and old 1Oglo-1og10
foliage and old twig. Sampleswere dried at 70øCand weighed enousvariance of the independentvariable. Equations of the
to the nearest0.1 g. Total biomasswascalculatedasthe sumof form Y - a + bX and Y = a + b(loglo X) were also
the dry biomassof all componentsand ANPP wascalculatedas examined, but a plot of the residuals suggestedthat these
the sum of the dry biomassof herbs,new twig, and new foliage models provided an inadequate fit and/or did not correct for
components.Except in the aspen stands,stem growth of un- nonhomogeneousvariance.
Levene's test was used to test for equal variance between
derstoryvegetationwas ignored becauseit was so small that it
was difficult to obtain reliable estimates of stemwood radial
specieswithin a study area or within a speciesbetween study
increment.
areas.We did not detect any evidenceof unequalvariance.An
Ten hazel shrubs,encompassing
a wide range in stem diam- analysisof covariance[Zar, 1983]was usedto determine if the
eters, were destructivelysampledin each aspenstand. A disk allometric coefficientsdiffered within a speciesbetween sites.
was removed
from the base of each stem and its diameter
All stand values reported in the paper are the averagesfor
measured.A regressionequation for stem masswas developed the four replicate plots. We used an analysisof variance to
(stemmass(g) = 1.73+ 2.58lOglo(basaldiameter,cm),r 2 = determinethe effect of forest type, studyarea, and their interaction on ecosystemcharacteristics.If an ecosystemcharacter0.90, p < 0.001, mean square error = 0.008, n = 20).
Equations for alder (Alnus rubra) were developedfrom an istic differed among the forest types, a Fisher's least square
allometry harvestin August 1995 [Vogel,1997].The difference difference(LSD) test (c• = 0.05) was usedto separateforest
in biomassbetween years was calculated from the increment types.
data and the regressionequation. Foliage litterfall was measured in the litterbaskets.
Moss net primary production was measured in the black
spruce stands using screen ingrowth plots. In late May, five
30 x 30 cm wire mesh screens(0.5 x 0.5 cm mesh) were
carefully placed on the moss surface and bent to conform to
the
surface
of the
moss.
The
screens
were
secured
to the
ground with 25 cm long aluminum wire pegsplaced in each
corner of the screen.At the end of the growingseason(midOctober) all the mossfronds stickingthrough the screenwere
clipped, collectedin paper bags,dried at 70øCto a constant
mass and weighed. AbovegroundNPP was calculatedas dry
matter increment times 0.5 to convert dry biomassto carbon
content.
Soil Carbon
Stem, branch, and foliage biomass,stem and leaf area, and
sapwoodvolume are all highly correlated to dbh for each
speciesat each site (Table 3). Allometric equationsfor stem
and branch
Distribution
and the forest floor tissue was dried at 70øC to
a constant mass and weighed to the nearest 0.1 g. The dried
biomass
do not differ
between
SSA and NSA
for
each species,but foliage biomass and leaf area allometric
equationsfor black spruceand jack pine differ (p < 0.05)
between study areas. All allometric equations relating tissue
massand area to stem dbh differ (p < 0.01) betweenyoung
and old jack pine trees at both the SSA and NSA. Because
allometric
Three forest floor samples(25.2 cm diameter) were collected from random locationsin each replicate plot in the
aspen and jack pine stands.Live vegetation and coarseroots
were removed
Results
Allometry
coefficients
differ between
the SSA and the NSA we
usedsite specificallometric equationsto estimateaboveground
biomass
Stand
and NPP.
Structure
The LAI (one-half total leaf surfacearea) of the BOREAS
forestsrange from 1.8 for youngjack pine at the NSA to 5.6 for
black spruceat the SSA (Table 4). At both the SSA and the
GOWER ET AL.: BOREAL FOREST CARBON BUDGETS
29,033
Table3. Regression
Coefficients,
SampleSize(n), Adjusted
r2 (Coefficient
of Determination),
MeanSquareError (MSE),
and CorrectionFactor (CF) for Boreal Tree Species
RegionSite
Species
DBHa Range
Tissue
n
a
b
r2
MSE
CF
NSA
NA
aspen
5.8-23.7
(T2Q6A) '•
NBS
black spruce
2.8-17.0
(T3R8T)
NYJP
jack pine
0.9-6.7
(T8S9T)
NOJP
jack pine
4.5-16.8
(T7Q8T)
Sb
12
-2.312
4.184
0.993
0.002
1.005
SAP c
TF
TBR
TLA
10
8
8
8
- 5.211
-4.742
- 3.002
-3.349
4.345
4.018
3.265
3.798
0.996
0.761
0.964
0.752
0.001
0.204
0.017
0.191
1.004
"'
1.046
'-'
S
10
- 1.294
2.584
0.996
0.001
1.003
SAP
NF
OF
TF
NT
BR
TBR
NLA
OLA
TLA
10
9
9
9
9
9
9
9
9
9
-4.309
-3.051
- 2.478
- 2.400
-4.232
- 2.384
- 2.379
- 1.861
- 1.676
- 1.542
2.514
2.219
2.706
2.668
2.742
2.642
2.646
1.875
2.631
2.549
0.982
0.599
0.990
0.988
0.543
0.843
0.842
0.545
0.974
0.974
0.006
0.057
0.001
0.002
0.108
0.024
0.024
0.050
0.003
0.003
1.016
...
1.003
1.005
.. ß
1.066
1.066
1.142
1.008
1.008
S
10
-0.956
1.956
0.989
0.002
1.005
SAP
NF
10
10
-3.940
-2.185
2.205
1.843
0.991
0.931
0.002
0.014
1.005
1.038
1.675
0.903
0.017
1.705
0.921
0.014
1.597
2.028
0.856
0.948
0.024
0.013
1.046
1.038
1.066
1.035
OF
10
TF
10
NT
BR
10
10
- 1.448
- 1.374
-2.782
- 1.566
TBR
10
- 1.541
2.012
0.948
0.013
1.035
NLA
OLA
TLA
10
10
10
- 1.271
-0.629
-0.538
1.768
1.610
1.642
0.908
0.916
0.932
0.018
0.014
0.011
1.049
1.037
1.030
S
10
- 1.148
2.440
0.991
0.002
1.005
SAP
NF
OF
TF
NT
BR
TBR
NLA
OLA
TLA
10
10
10
10
10
10
10
10
10
10
-4.208
-3.654
-2.332
- 2.327
-4.134
-2.440
- 2.430
- 2.742
- 1.535
- 1.523
2.626
3.021
2.560
2.613
2.816
2.799
2.798
2.890
2.519
2.568
0.968
0.966
0.970
0.974
0.944
0.951
0.952
0.956
0.972
0.977
0.007
0.010
0.006
0.005
0.014
0.012
0.011
0.011
0.005
0.005
1.019
1.027
1.016
1.013
1.038
1.032
1.030
1.030
1.013
1.013
S4
10
- 2.505
4.212
0.988
0.002
1.004
SAP
TF
TBR
10
10
10
-4.127
-3.621
-3.668
2.739
2.971
3.679
0.874
0.933
0.890
0.021
0.012
0.032
1.057
1.032
1.089
TLA
S
SAP
NF
OF
TF
NT
BR
TBR
10
10
9
9
9
9
9
9
9
-2.011
- 1.177
-4.081
-2.566
- 1.364
- 1.336
-3.506
- 1.583
- 1.575
2.611
2.418
2.226
1.956
1.800
1.811
2.264
2.068
2.070
0.905
0.991
0.979
0.755
0.866
0.859
0.671
0.819
0.818
0.014
0.005
0.011
0.121
0.050
0.053
0.239
0.093
0.094
1.038
1.013
1.030
'"
1.142
"'
'"
'"
'"
NLA
OLA
TLA
S
SAP
NF
OF
TF
NT
BR
TBR
NLA
OLA
TLA
9
9
9
10
10
10
10
10
10
10
10
10
10
10
- 1.843
-0.675
-0.645
-0.921
-3.695
- 1.879
- 1.354
- 1.239
- 2.395
- 1.743
- 1.662
- 1.041
-0.597
-0.462
1.937
1.796
1.806
1.888
1.879
1.621
1.592
1.606
1.495
2.277
2.204
1.545
1.575
1.575
0.772
0.867
0.861
0.994
0.965
0.917
0.941
0.957
0.884
0.959
0.961
0.901
0.944
0.959
0.108
0.049
0.052
0.002
0.011
0.021
0.014
0.010
0.026
0.020
0.018
0.023
0.013
0.010
"'
1.139
'"
1.005
1.030
1.057
1.038
1.027
1.071
1.054
1.049
1.063
1.035
1.027
SSA
SA
aspen
11.3-29.8
(C3B7T)
SBS
black spruce
1.9-17.6
(G814T)
SYJP
(F8L6T)
jack pine
0.8-6.4
29,034
GOWER ET AL.: BOREAL FOREST CARBON BUDGETS
Table 3. (continued)
RegionSite
Species
DBHa Range
Tissue
n
a
b
10
10
9
- 1.292
- 4.648
- 3.200
2.570
2.904
2.404
OF
TF
9
9
-2.799
- 2.661
NT
9
BR
TBR
NLA
OLA
TLA
9
9
9
9
9
r2
MSE
CF
SSA
SOJP
(G2L3T)
jack pine
6.6-18.8
S
SAP
NF
0.002
0.005
0.013
1.005
1.013
1.035
2.621
2.578
0.991
0.979
0.925
0.963
0.966
0.007
0.006
1.019
1.016
-4.308
2.671
0.752
0.062
-'.
-3.460
- 3.443
- 2.444
- 2.054
- 1.911
3.697
3.688
2.412
2.609
2.568
0.951
0.951
0.924
0.961
0.964
0.019
0.019
0.013
0.007
0.007
1.052
1.052
1.035
1.019
1.019
All equations
(except
wherenoted)followtheformlog10
Y = a + b(log•o)XwhereX isstemdiameter
I (cm)at breastheight(1.37m) and
Y ismass(kg),stemsapwood
volume(m3) or leafarea(m2).Tissuesymbols
areasfollows:
S,stem(wood+ bark);SAP,stemsapwood
volume;
NF, newfoliage;OF, old foliage;TF, total foliage;NT, newtwig;BR, branch;TBR, branch+ newtwig;NLA, newleaf area (currentyear);OLA,
old leaf area (>1 year); TLA, total leaf area.
aBOREAS
site code.
bRegression
formis log•oY = a + b(log•0X) + cX wherethe C coefficient
is -0.052.
CRegression
form is log•oY = a + b(log•0 X) + cX where the C coefficientis -0.051.
dRegression
form is log•0Y = a + b(log•oX) + cX wherethe C coefficient
is -0.039.
NSA, LAI is significantlygreater (p < 0.01) in the black
spruce than aspen or jack pine stands.LAI did not differ
consistentlybetweenyoungand old jack pine standsat the two
study areas. Stem surface area (one-half total stem surface
area)rangefrom 0.1 m2 m-2 for the-young
jack pineat the
NSA to 0.6 for black spruceat the NSA and comprise5-15%
of one-half total tree area (excludingbranches)for the eight
stands.Stem sapwood
volumesrangefrom 13.3m3 for the
youngjackpineat theNSA to 327m3for theaspenstandat the
SSA (Table 4). Stem sapwoodvolumeis 2-4 timesgreaterfor
the old than for the youngjack pine stands.At both the SSA
and the NSA the leaf area:sapwoodvolume(LA:SA•,) ratio is
significantlygreater(p < 0.01) for youngversusold jack pine
stands. The
lower
ratio
for the mature
stand in the NSA
is
becauseof an increasein sapwoodvolume in the mature stand,
are 1900,600,and450kg C m-2 yr-l, respectively.
Woody
biomassincrementduring the 1993-1994 BOREAS field campaignsis below the 10 year averagefor aspen and jack pine
standsin the SSA and is above the 10 year averagefor black
sprucein the SSA for all the forestsin the NSA.
ANPP and carbonallocationdiffer more amongforest types
in a similar climate than for a similar forest type in the two
contrastingclimates.For example,at both studyareas,woody
biomassincrementis 2 to 3 times greater for aspenthan black
spruceor jack pine stands.Leaf and total abovegrounddetritus
production(i.e., litterfall) are significantly
greater(p < 0.01)
for aspenthanjack pine or blacksprucestandsat both the SSA
and the NSA (Tables 6 and 7). On a relative basis,however,
detritus production comprisesa similar or greater fraction of
ANPP in the conifer than deciduousaspen forests.Averaged
but in the $SA, the lower ratio for the mature stand is because
of an increasein sapwoodvolume and a decreasein LAI.
Carbon contentand its relative distributiondiffer amongthe Table 4. SapwoodVolume, Leaf and Stem Area Index,
eight stands(Table 5). For the SSA and NSA, total ecosystem Stem:Total Tree Area Index Ratio for Aspen, Black Spruce,
carboncontent (vegetationplus detritusplus soil) is 2.8 to 2.9 and Jack Pine Stands at SSA and NSA
times greater for black sprucethan for aspenstands,and both
Leaf
Stem
are greater than the mature jack pine ecosystem.Soil contains
Sapwood
Area
Area
Stem:Total
Volume
Index
Index
Tree Surface
the greatestpercentageof the total ecosystemcarbon content
Study Area
Forest Type
(m3 ha-l) (ha ha-1) (ha ha-1) Area Ratio
in black spruceforests(87-88%) but vegetationand soil contain a similar percentageof total ecosystemcarboncontent in SSA
the mature jack pine and aspen stands.Forest floor carbon
aspen
327.2
3.3
0.4
11
content estimates for some stands vary greatly among
(72.5)
(0.7)
(<0.1)
black spruce
60.8
5.6
0.4
7
BOREAS scientists(Table 5); potential explanationsare examined
in the discussion.
jack pine (young)
Aboveground Net Primary Production (ANPP)
ANPP variesby lessthan 20% between 1993 and 1994 for all
the mature stands, except for the NSA aspen stand which
variesby 29% (Table 6). The lower ANPP in 1993 is, in part,
becauseunderstoryNPP was not measured in 1993. Interannual variation (annual maximum - annual minimum/10year
average from 1985 to 1994) in woody biomass increment
rangesfrom 21% for aspenat the SSA to 83% for mature jack
pine at the NSA (Figure 1). Average(1985-1994woodybiomassincrementfor aspen,black spruce,and maturejack pine
at the SSAare 1860,720,600kg C m-2 yr-• andat the NSA
jack pine (old)
(9.8)
26.0
(8.4)
56.9
(16.0)
(1.7)
2.8
(0.8)
2.4
(0.3)
(0.1)
0.2
(<0.1)
0.2
(<0.1)
196.0
(13.3)
82.5
(9.3)
13.3
(5.7)
54.7
(5.6)
2.2
(0.2)
4.2
(0.3)
1.8
(0.6)
2.2
(0.3)
0.4
(<0.1)
0.6
(<0.1)
0.1
(<0.1)
0.2
(<0.1)
5
13
NSA
aspen
black spruce
jack pine (young)
jack pine (old)
15
12
5
8
Valuesin parentheses
are 1 standarddeviationof the mean(n = 4).
GOWER ET AL.: BOREAL FOREST CARBON BUDGETS
29,035
Table5. Aboveground
CarbonDistribution
(kgC ha-1) byEcosystem
Component
for Aspen,BlackSpruce,
andYoung
and Old Jack Pine Stands at SSA and NSA
Aspen
Ecosystem
Component
C
BlackSpruce
%
C
%
JackPine (Young)
C
%
JackPine (Old)
C
SouthernStudyArea (SSA)
Vegetation
stem
branch
newfoliage
old foliage
understory
bryophyte
total living
82,710
8,970
940
0
720
0
93,340
(52)
(6)
(<1)
(0)
(< 1)
(0)
(59)
36,330
6,750
450
4,600
510
600
49,240
(8)
(2)
(<1)
(1)
(< 1)
(<1)
(11)
7,650
2,430
530
1,660
na
na
12,260
(15)
(5)
(1)
(3)
('- ')
(..-)
(24)
25,760
4,130
190
840
160
3,470
34,550
(37)
(6)
(<•)
(•)
(<•)
standingdead
forestfloor
9,680
19,430
(6)
(12)
6,160
--.
(1)
0
18,100
(0)
(36)
5,670
14,560
(s)
(2•)
mineralsoil
35,990
(23)
390,360
(88)
20,160
(40)
Detritus
90,2902
Total ecosystem
16,780a
158,440
445,760
50,530
14,200
(20)
68,980
Northern StudyArea (NSA)
Vegetation
stem
branch
newfoliage
old foliage
understory
bryophyte/lichen
total living
49,780
5,720
720
40
690
0
56,950
(28)
(3)
(< 1)
(0)
(< 1)
(0)
(32)
47,370
4,910
350
3,610
0
970
57,210
(10)
(!)
(< 1)
(1)
(0)
( < 1)
(12)
4,820
1,370
230
1,070
310
na
7,800
(6)
(2)
(< 1)
(2)
(< 1)
...
(10)
19,060
2,520
240
1,490
580
5,100
28,990
standingdead
forestfloor
6,390
15,880
70,510
•'
(4)
(9)
3,810
......
(1)
0
40,260c
128,010
b
(0)
(53)
2,120
11,480
223,400
b
mineralsoil
97,170
(55)
418,360
(87)
28,430
(37)
25,780
(28)
(4)
(<•)
(2)
(1)
(7)
(42)
Detritus
(3)
ooo
9,420 c
Total ecosystem
176,390
479,380
76,490
(38)
68,370
Valuesin parentheses
are the contributionof eachcomponentto the total ecosystem
carboncontent.
"T. Norbusand D. Anderson(unpublished
data, Universityof Saskatchewan,
1996).
bH.Veldius(unpublished
data,University
of Manitoba,
1996).
•Data from Trumborectal. [this issue].
for the twoyears,fine woodydetrituscomprises23-37% of the
total abovegrounddetritus. In both 1993 and 1994, leaf and
total detritusproductionare significantlygreater (p < 0.01)
in the SSA than in the NSA for all three forest types.
or near-maximumLAI of control, irrigated and fertilized Scots
pine (Pinus•ylvestris)were 1.0, 1.3, and 3.6, respectively[Waring and Schlesinger,1985]. Low nutrient availability of most
boreal forestsis causedby cold soils and poor litter quality
which adverselyaffect litter decompositionand nutrient minDiscussion
eralization[Flanaganand Van Cleve,1983]. Severewinter environmental
conditionsalsorestrictLAI [Gholz, 1982].
Canopy Architecture
A secondimportantcanopyarchitecturecharacteristicof the
Canopyarchitectureinfluenceslight interceptionand there- borealforestsis foliageclumping--thenonrandomdistribution
fore effectsCO2 exchangebetweenterrestrial ecosystems
and
of foliagein the canopy.The distributionof foliage elementsin
the atmosphere.The canopyarchitectureof boreal forestsdifthe canopyinfluencesthe fractionof the canopythat is sunlit
fers from mosttemperateforestsin two importantways,both
versusshadedand total radiation interception,especiallyin
of which effect radiation absorptionand canopyphotosyntheforeststhat have low LAIs. Lessradiation is absorbedif foliage
sis. On average,LA! is lower in boreal than in temperate
is clumped versusrandomly distributed.Direct estimatesof
forests [Landsbergand Gower, 1997]. The LAI for the dominant foresttypesin the two BOREAS studyareasrangesfrom LAI (see Table 4) agreewith optical estimatesonly when the
1.8 to 5.6 and averages3.1 (Table 4). In contrast,the LAI of indirectestimatesof LAI are correctedfor foliageclumpingat
the six major forest types forming the transition from cold the shoot, branch, and tree level [Chen et al., this issue;Kutemperateto boreal forestsin northern Wisconsinrangesfrom chariket al., this issue].Ecosystemprocessmodelsthat ignore
1.9 to 8.4 and averages4.9 [Fassnacht
and Gower, 1997]. Com- the largefoliageclumpingin forestswith low LAIs (e.g.,boreal
[Kuchariket
parative and experimental studieshave shown that nutrient forests)will overestimatecanopyphotosynthesis
al.,
this
issue].
Future
global
terrestrial
carbon
models
should
availabilitylimits the LAI of boreal forests.For example,the
LAI of jack pine forestscontaininggreen alder, a nitrogen- accountfor the nonrandom distributionof foliage in boreal
fixing shrub,is 29 to 34% greater than for adjacentjack pine forests to accuratelysimulate CO2 exchangebetween terresstandswhere greenalder is absent[Vogel,1997].The maximum trial ecosystemsand the atmosphere.
29,036
GOWER ET AL.: BOREAL FOREST CARBON BUDGETS
Table 6. Aboveground
Net PrimaryProduction
(ANPP) (kg C ha-• yr-•) for 1993and
1994by Componentfor Aspen (A), Black Spruce(BS), and Old (OJP) and Young (YJP)
Jack Pine Stands at NSA
and SSA
A
VegetationComponent
BS
1993
1994
1993
OJP
1994
1993
1994
YJP, 1994
380a
SouthernStudyArea
Biomass
increment
wood
1,640
780
800
570
590
(190)
0
(250)
0
(170)
70
(20)
(130)
80
(20)
(70)
20
(<•0)
(60)
20
(<•0)
understory
240
530
nab
130
na
50
na
bryophytes
(40)
0
(50)
0
na
(60)
120
(60)
na
(<10)
< 10
(<10)
na
520
510
(50)
(50)
foliage
Detritus
1,760
1240
1230
(150)
ANPP
(60)
590
530
(100)
(100)
3120
3520
1440
1660
(210)
(280)
(190)
(160)
1110
1170
(20)
10
(<•0)
530 c
(20)
920
(80)
(80)
(80)
170'•
Northern StudyArea
Biomass
increment
wood
1,750
foliage
(230)
0
understory
na
bryophytes
2,140
(320)
0
330
(60)
0
0
Detritus
740
730
690
650
(10)
50
(<10)
na
(20)
50
(<10)
< 10
(<10)
120
(60)
(80)
50
(<10)
na
(70)
50
(<10)
160
(40)
< 10
(<10)
400
460
240
360
(20)
(40)
(30)
(20)
1020
(60)
ANPP
660
(60)
2490
3490
(280)
(380)
na
1110
(40)
1360
(60)
na
980
(100)
(10)
40
(<10)
70
(10)
na
230 c
(20)
1220
510
(100)
(S0)
Values in parenthesesare _+1 standarderror.
aCalculatedas standingbiomass/standage.
bna,not available.
•Estimated from new foliage production.
(thisstudy)to 223,400kg C ha • (H. Veldhuis,unpublished
Forest Floor C Dynamics
data, 1996). The large differencesin forest floor carboncondiffer greatly amongBOREAS investigators.Forest floor car- tent are disturbingbecausesoil carboncyclingprocessmodels
boncontentin aspenstands
at theSSA(19,430kgC ha-l) and that simulateheterotrophicrespiration,a major componentof
NSA (15,880kg C ha-•) are roughlyone-fifththe valuesof NEE, require forestfloor carboncontentas an input variable.
90,290kgC ha-1 reportedfor theSSAaspenstand(D. Ander- We cannot explain the large discrepancies,but note that our
sonand T. Nerbas,personalcommunication,1996) and 70,510 estimatesare comparableto valuesreported for similar forest
kg C ha-• reportedfor the NSA aspenstand(H. Veldhuis, types in Canada and Alaska. For example, our estimatesfor
personal communication,1996). Forest floor carbon content forest floor carbon content for the aspenstandsare similar to
estimatesfor the NSA old jack pine stand differ even more 23,750kg C ha • (assuming
organicmatteris 50% C) for
amongBOREAS scientists,
rangingfrom 11,480kg C ha-• tremblingaspenstandsin Alaska [Van Cleveet al., 1983], and
our estimatesof forest floor carbon content for mature jack
Estimates
of forest
floor
carbon
content
for
some
stands
pinestands
aresimilarto 10,150kg C ha-• (assuming
organic
2500
..........
ß SA
0 SBS
ß
SJP
V
NA
[] NBS
[] NJP
Year
Figure 1. Interannualvariation (annual maximum - annual
minimum/10year average)in wood biomassincrementfor the
aspen,black spruce,and jack pine standsat the southernstudy
area (SSA) and northernstudyarea (NSA).
matter is 50% C) for jack pine forestsin northern Ontario
[Fosterand Morrison, 1976].
Forest floor carbon content is the long-term integrated difference between detritus inputs (e.g., litterfall and fine root
turnover) and litter decomposition.Mean residencetimes
(MRT) for carbonin the forestfloor, calculatedasthe ratio of
forest floor carboncontent and abovegroundlitterfall carbon,
providesa simpleindex of the balanceof thesetwo processes.
The calculation
assumes that forest floor and litterfall
carbon
are in steady state; this assumptionappears to be valid for
these standsbecauseleaf litterfall does not greatly differ betweenyears(Table 7) and is similar to new foliageproduction
(Tables6 and 7). Forestfloor carbonMRT variesmore among
GOWER ET AL.' BOREAL FOREST CARBON BUDGETS
29,037
Table 7. Aboveground
DetritusProduction
(kg C ha-• yr-•) byTissueComponent
for
Aspen, Jack Pine and Black SpruceStandsat SSA and NSA
SouthernStudy Area
Litter Component
Aspen
JackPine
Northern StudyArea
BlackSpruce
Aspen
JackPine
BlackSpruce
1993
Foliage
dominant overstory
660
380
260
other overstory
(60)
0
(40)
10
(10)
(20)
90a
(40)
understory
260
(60)
total
Woody
Total
450
160
240
(10)
10b
(<10)
(20)
10c
(<10)
(20)
0
10
< 10
130
(<10)
(<10)
(10)
10
< 10
(<10)
(<10)
180
240
920
400
350
590
320
120
240
150
60
(100)
(20)
(100)
(60)
(20)
(10)
1240
520
740
240
400
(lS0)
(S0)
(100)
(60)
(30)
(20)
300
590
160
1994
Foliage
dominant overstory
other overstory
understory
total
Woody
Total
730
380
330
620
260
(80)
(30)
(60)
(10)
(10)
10
20c
<10
(<10)
(10)
(<10)
<10
<10
90
(<10)
(<10)
(40)
260
<10
0
130
(0)
(20)
(30)
(•0)
Total
0
0
280
300
990
390
420
760
240
130
110
260
80
(20)
(20)
(10)
(70)
(10)
(30)
510
530
360
460
(60)
(20)
(40)
675
205
230
70
270
160
880
300
430
1230
(60)
Foliage
Woody
(10)
955
280
1235
(50)
395
125
(100)
Average
385
175
515
560
1020
160
Valuesin parentheses
are +1 standard
error.All valuesareroundedto thenearest10kg C ha-• yr-•
•Tamarack.
bjackpine.
CPaperbirch.
dWillowspp.
forest typesin a similar climate (i.e., studyarea) than for a
similar speciesin the two contrastingclimates.The forest floor
carbonMRT for the jack pine stands(28-39 years)are greater
than the calculatedvalues of 5 years reported for jack pine
forests in Minnesota [Perala and Alban, 1982] and Ontario
[Fosterand Morrison,1976] and 7 yearsfor jack pine standsin
northern Wisconsin[Fassnachtand Gower, 1997]. The forest
Cleveet al., 1983]. ANPP doesnot differ betweenthe mature
jack pine standsin the SSA and NSA and the averagein 1994
is smallerthan the averageANPP of 2050 kg C ha-• yr-•
(assumingorganicmatter is 50% C) for eightjack pine stands
in northern Wisconsin[Fassnachtand Gower, 1997].
ANPP is the mostcommonlymeasuredcomponentof forest
productionbudgets,but belowgroundnet primary production
floor carbonMRT for the aspenstands(16-19 years) in this (BNPP) can equal or exceedANPP in forests[Landsberg
and
studyare lessthan the calculatedvaluesof 30 yearsfor aspen Gower, 1997]. Moreover, fine root turnover is a large C input
standsin Alaska [Van Cleveet al., 1983]but are greaterthan 8 to the soil. Ruesset al. [1996] reported that fine root producyearsfor aspenstandsin northern Wisconsin[Ruarkand Bock- tion, measuredusing soil cores, accountedfor 49 and 32% of
heim, 1988] and 6 yearsfor aspenstandsin Minnesota [Perala total NPP in coniferous and deciduous boreal forests in
and Alban, 1982].
Alaska, respectively.Using the below-ground NPP data for
these stands[Steeleet al., 1997], we estimatethat BNPP comNet Primary Production
prises41-46% of total NPP in the conifer standsand 10-19%
ANPP for the mature foreststandsrangesfrom 1170to 3520 in the aspenstands(Figure 2). The data from Ruesset al. and
kg C ha-• yr-• in 1994.The averageANPP for the black this study suggestthat a greater percentageof total NPP is
sprucestandsis abouttwofoldgreaterthanthe 565kg C ha- •
yr-• reportedfor blackspruceforestsin Alaska[VanCleveet
allocated
to roots in boreal
coniferous
than deciduous
forests.
Clearly, fine root production must be included to construct
al., 1983]. The averageANPP for the aspen standsis lower completecarbonbudgetsfor boreal forests.
One consistentpattern at both study areas in ANPP is
thanthe5548kgC ha-• yr-• (assuming
organic
matteris50%
C) for an upland birch-aspenforest in Alaska [Ruesset al., greater for the deciduousthan coniferousforests. Other sci1996]butis similarto 2825kg C ha-• yr- • (assuming
organic entistshave also reported that ANPP is greater for deciduous
matter is 50% C) for a tremblingaspenstandin Alaska [Van than evergreenboreal forests[Van Cleveet al., 1983;Ruesset
29,038
GOWER
ET AL.: BOREAL
FOREST CARBON
4000
3000
2000
1000
SA
SBS SJP NA
NBS NJP
Figure 2. Net primary production budgets for the mature
aspen(A), black spruce(BS), andjack pine (JP) standsin the
SSA (S) and NSA (N). The youngjack pine standsare not
included
because
fine root
NPP
was not measured
in these
BUDGETS
deciduousand evergreenconiferousforestsallocate different
amountsof carbonto root production.Including BNPP in the
productionefficiencycalculationsdoesnot diminishthe difference in production efficiency between deciduousand evergreenforests(Figure 4). The fractionof total NPP allocatedto
root production,however,is smallerfor deciduous(0.14) than
for coniferousstands(0.44) in this study(Figure 2 and Table
8). We speculatethat productionefficiencyand NPP differ
between deciduousand evergreen forests becausedeciduous
forestshave a greater light use efficiency[Hunt and Running,
1993] and experienceless environmental constraintson photosynthesis(see below).
The slope of the regressionline describingthe relationship
stands.Root NPP data are from Steeleet al. [1997] and ANPP
between
data are for 1994.
studyand conifer forests(largelyjack pine) in northern Wisconsin.However, the Y intercept is greater for temperate than
for boreal coniferousforests, suggestingthe production efficiency is greater for temperate than for boreal forests.The
most likely explanationfor the lower productionefficiencyof
boreal than temperateforestsis that environmentalconstraints
on canopyphotosynthesisare greater in boreal environments.
McMurtie et al. [1994] used a forest ecosystemprocessmodel
BIOMASS to elucidate the relative importanceof variousenvironmentalcontrolson pine forestproductivityand found that
incompletelight interception and short growing seasonwere
the major constraintson leaf photosynthesis
for boreal Scots
pine forestsin Sweden.Low LAI and large foliage clumping
al., 1996]. Possiblecausesfor the greater ANPP of deciduous
than evergreenconifer boreal forestsinclude greater capacity
to absorblight (i.e., LAI), greater intrinsicphysiologicalcapacity of the canopyto convert solar radiation to dry matter,
fewer environmental constraintson leaf photosynthesisand
less allocation of C to roots and mycorrhizae.Aboveground
NPP is positivelycorrelatedto LAI in many temperateforests
[Gholz,1982;Goweret al., 1992;Runyonet al., 1994;Fassnacht
and Gower,1997].LAI differsby more than twofold amongthe
BOREAS forests(Table 4), but ANPP is not significantlycorrelated to LAI exceptwhen only conifer standsare considered
ANPP
are common
and LAI
is similar
characteristics
for the conifer
stands in this
of boreal forests and contribute
to
(1'2 = 0.70, n = 6) (Figure3). Production
efficiency
(ANPP/ incompletelight interceptionin theseforests.Long, harshwinLAI) is significantlygreater (p < 0.01) for aspenthan for ters restrictthe growingseasonin the SSA and NSA to 150 and
conifer forests in both the SSA and the NSA (Figure 4). 120 days,respectively[Steeleet al., 1997]. In addition, frozen
Fassnachtand Gower [1997] observedgreater productioneffi- soil is an important constrainton the carbonbalanceof everciencyfor deciduousthan evergreenconifer temperateforests green conifers,especiallyin black spruceforests,in the spring
in northern Wisconsin.
becausesolar radiation and air temperatures are often suffiIt is unclear why productionefficiencyis greater for decid- cient for photosynthesisby evergreen conifers at this time
uous than for evergreen conifer forests in boreal and cold [Frolkinget al., 1996].Canopyphotosynthesis
for all three tree
temperate regions. One possibleexplanationfor the "appar- speciesin the NSA is adverselyaffectedby atmosphericvapor
ent" higher production efficiencyof deciduousforests is that pressuredeficit during the growing season[Dang et al., this
issue].
6000
i
t
Global
I
I
I
I
I
I
-'-'5000
4000
I V
BD I
[V
BCI
2000
z
and Carbon
Allocation
Patterns
for Boreal
Global terrestrial primary production models are useful
tools to examine the role terrestrial ecosystemsplay in the
global C budget; however, few NPP data are available for
comparison.Information on carbon allocation is even more
scarce,and as a result, many modelsuse a constantallocation
coefficient(i.e., abovegroundand below-groundNPP) for all
3000
<
NPP
Forests
1000
1500
0
1
4
5
6
7
8
9
1000
LAI
500
Figure 3. Relationshipbetween abovegroundnet primary
production(ANPP) andleaf areaindex(LAI) for aspen,black
0
spruce,and jack pine standsat the SSA and NSA. Solid and
SA SBS SJP NA NBS NJP
opentrianglesdepictborealdeciduous(BD) and borealevergreen conifer (BC), respectively.For comparison,data for Figure 4. Comparisonof productionefficiencybasedon net
temperatedeciduous(TD) and temperateevergreenconifer primaryproduction(NPP/LAI) and abovegroundnet primary
(TC) from northernWisconsin[Fassnacht
and Gower,1997] production
(ANPP/LAI).NPP is expressed
in kg C ha-J yr-•
are included.
and LAI is m 2 m -2
GOWER ET AL.: BOREAL FOREST CARBON BUDGETS
29,039
Table 8. Net Primary Production Rates for Boreal Forests of the World
Tree
ANPP
UnderstoryANPP
Latitude
Species
Location
BNPP,
D,
BNPP
NPP
NPP
AB
---
1550
2850
0.54
1.71
2600
......
1100
3700
0.29
1.00
850
3100
......
500
700
4300
0.16
0.69
2,3
350
300
650
......
550
1050
2250
0.47
3.86
2,3,4
800
530
1330
0
130
130
1200
2660
0.45
2.32
5,6
730
460
1190
0
100h
100
900
2190
0.41
2.00
5,6
900
550
1450
200
650
85O
1.09
7
200
150
350
450
800
1250
1.46
7
400
1050
1450
150
50
200
2.00
8, 9
200
250
450
650
1750
2400
......
960
1080
2990
(120)
(260) (0.04)
Longitude
AB
D
total
62 51 N
1050
250
1300
1850
750
2250
AB
D
total
Reference
Evergreen
Picea abies
Ilomantsi,
Finland
Piceaglauca,FP a
30 53 E
Alaska,
64 45 N
United
Piceaglauca,UPb
States
148 15 W
States
148 15 W
64 45 N
Alaska,
64 45 N
United
Picea mariana
148 15 W
Alaska,
United
Picea manana
States
c
Saskatchewan,
Canada
Picea manana
c
Manitoba,
55 93 N
Canada
Piceamanana,Bd
98 62 W
Manitoba,
49 53 N
Canada
Picea manana, M c
95 54 W
Manitoba,
49 53 N
Canada
Piceamanana, pt
Piceamanaria (R g)
53 92 N
104 69 W
95 54 W
Minnesota,
49 53 N
United
95 54 W
States
Minnesota,
49 53 N
United
95 54 W
States
1390
Picea (average)
(280)
Pinus banksianac
Saskatchewan,
Canada
Pinus banksiana•
Pinussylvestris
Pinussylvestris
0.39
8,9
1.93
(0.46)
650
510
1160
0
50
5O
1050
2260
0.46
2.48
5,6
700
360
1060
0
150
150
1000
2210
0.45
2.16
5, 6
850
600
1450
"-
1150
2600
0.44
2.06
1
700
500
1200
--'
550
1750
0.31
1.50
1
1550
750
2300
-..
1000
3300
0.30
1.13
10
0.39
104 69 W
Manitoba,
55 93 N
Canada
Pinussylvestris
53 92 N
(230)
0.27
98 62 W
Ilomantsi,
62 51 N
Finland
30 53 E
Ilomantsi,
62 51 N
Finland
30 53 E
Jadraas,
60 49 N
Sweden
16 30 E
Pinus(average)
1430
100
950
2420
(220)
(50)
(80)
(260) (0.04)
1.87
(0.24)
Deciduous
Alnus-Populus
Alaska,
United
Betulapapyrifera
Lark gmelini
148 15 W
States
148 15 W
Alaska,
United
Betulapubescens
64 45 N
States
64 45 N
Llomantsi,
62 51 N
Finland
30 53 E
Inner,
62 51 N
Mongolia
Populusbalsamifera
Alaska,
United
Populustremuloides
• Saskatchewan,
Canada
Populustremuloides
c Manitoba,
Canada
Deciduous(average)
1550
6350
......
ß'-
1600
7950
0.20
0.66
2
4050
1350
5400
......
<50
1250
6650
0.19
0.64
2,3
750
350
1100
......
ß--
400
1500
0.27
1.00
1
2050
650
2700
..-
800
............
3250
900
4150
......
1760
970
2730
530
260
790
2140
890
3030
330
130
1
ß'-
2000
6150
0.32
0.89
2
400
3920
0.10
0.71
5,6
4600
650
3490
0.19
0.68
5,6
525
1050
4740
0.21
(270)
(840) (0.03)
1020
2730
148 15 W
53 92 N
104 69 W
55 93 N
98 62 W
3640
(560)
Evergreen(average)
800
30 53 E
64 45 N
States
4800
1400
(200)
(180)
620
(230)
(70)
0.39
(230) (0.03)
0.76
(0.09)
1.90
(0.26)
All valueshavebeenconverted
to kg C m-2 yr-1 assuming
woodis 50% C andfoliageandfinerootsare45% C. All valuesareroundedto
the nearest 10 kg C.
1,Finer[1989,1991];2, Ruessetal. [1996];3, Oecheland l/an Cleve[1986];4, l/an Cleveetal. [1983];5, thisstudy;6, Steeleet al. [1997];7, Reader
and Stewart[1972];8, Grigalet al. [1985];9, Grigal [1985];10, Linder andAxelsson[1982].
aFP, flood plain.
bUp,upland.
CData from
1994.
dB,bog.
øM, muskeg.
tp, perched.
gR, raised.
hMossonly.
29,040
GOWER ET AL.: BOREAL FOREST CARBON BUDGETS
forestbiomes.A secondobjectiveof this paper is to relate the
results of this study to other boreal forests to determine if
general NPP or carbon allocation paradigmsexist. Table 8
summarizesNPP by major componentfor borealforestsof the
world.AverageNPP is 3510kg C ha-• yr-• (n = 17) andis
verysimilarto 3600kgC ha- • yr- • calculated
byWhittaker
and
Likens [1973], cited by Schlesinger
[1991]. Few completeprimary productionbudgetsexistfor boreal forests,but the available data suggestseveralinterestingpatternsthat warrant further investigation. ANPP is greater for deciduous than
evergreen
borealforests(3640versus1400kg C m-2 yr-•),
corroborating the results from this study and Ruess et al.
[1996]. Carbon allocation patterns also differ between boreal
deciduous and evergreen forests. Evergreen conifer forests
allocate a greater fraction of carbon to root production
(mean -- 0.39, standarderror = 0.03,n = 15) than deciduous
forests(mean = 0.21, standarderror -- 0.03,n = 6). Detritus
production(i.e., litterfall and fine root turnover) is a major
componentof NPP for boreal forests,and the fraction appears
to differ among major forest types. The ratio of detritus
production:biomassincrementis greater for evergreenconifer
(1.90) than for deciduousforests(0.76), especiallyfor Picea
(spruce)forests.Bryophytescan comprisea large percentage
of the total detritusproductionin somesprucestands[Oechel
and l/an Cleve,1986].
In summary,LAI, carboncontentand distribution,and NPP
and carbon allocation vary more among forest types in each
study area than for a similar speciesin the two study areas.
Many of the differencesin NPP and carbonallocationbetween
evergreenand deciduousforestsin the BOREAS studyappear
to be general patterns for boreal forests.Becausemany of the
componentsof carbon budgetsdiffer between boreal evergreen and deciduousforests,we suggestthat global terrestrial
primary productionmodelscould be improvedby distinguishing betweendeciduousand evergreenboreal forestsand incorporating carbon allocation coefficientsthat reflect apparent
differencesbetween these two functional groups. Finally, a
better understandingof the factorsthat control leaf litter production
and fine
root
turnover
in boreal
forests
is needed
becausedetritusproductionconstitutesa largefractionof NPP
for boreal forests and is the major pathway for C input in
boreal
forest soil.
Chen, J., P. W. Rich, S. T. Gower, J. M. Norman, and S. Plummer,
Leaf area indexof boreal forests:Theory, techniques,and measurements,J. Geophys.Res., this issue.
Cias, P., P. P. Tans, M. Troller, J. W. C. White, and R. J. Francey,A
large northern hemisphere terrestrial CO2 sink indicated by the
13C/•2C
ratioof atmospheric
CO2,Science,
269, 1098-1102,1995.
Dang, Q.-L., H. A. Margolis, M. Sy, M. R. Coyea, G. J. Collatz, and
C. L. Walthall, Profilesof photosynthetically
activeradiation,nitrogen, and photosyntheticcapacityin the boreal forest:Implications
for scalingfrom leaf to canopy,J. Geophys.Res.,this issue.
Daniel, T. W., J. A. Helms, and F. S. Baker, Principlesof Silviculture,
2nd ed., McGraw-Hill, New York, 1979.
Fassnacht,K. S., and S. T. Gower, Interrelationshipsamong the
edaphicand standcharacteristics,
leaf area index,and aboveground
net primary productionof uplandforest ecosystems
in north central
Wisconsin, Can. J. For. Res., 27, 1058-1067, 1997.
Fassnacht,K. S., S. T. Gower, J. M. Norman, and R. E. McMurtrie, A
comparisonof optical and direct methods for estimatingfoliage
surfacearea indexin forests,Agric.For. Meteorol.,71,183-207, 1994.
Finer, L., Biomassand nutrient cyclein fertilized and unfertilizedpine,
mixedbirch and pine and sprucestandson a drained mire,Acta For.
Fenn., 208, 6-54, 1989.
Finer, L., Effect of fertilization on dry massaccumulationand nutrient
cyclingin Scotspine on an ombrotrophicbog,Acta For. Fenn., 223,
1-42, 1991.
Flanagan, P. W., and K. Van Cleve, Nutrient cyclingin relation to
decompositionand organicmatter quality in taiga ecosystems,
Can.
J. For. Res., 13, 795-817, 1983.
Foster,N. W., and I. Morrison, Distribution and cyclingof nutrientsin
a natural Pinus banksianaecosystem,Ecology,57, 11-120, 1976.
Frolking, S., et al., Modelling temporal variability in the carbon balanceof a spruce/moss
borealforest,GlobalChangeBiol.,2, 343-366,
1996.
Gholz, H. L., Environmentallimits on abovegroundnet primary production, leaf area, and biomassin vegetationzones of the Pacific
Northwest,Ecology,63, 469-481, 1982.
Gower, S. T., K. A. Vogt, and C. C. Grier, Carbon dynamicsof Rocky
Mountain Douglas-fir:Influence of water and nutrient availability,
Ecol. Monogr.,62, 43-65, 1992.
Gower, S. T., B. E. Haynes, K. S. Fassnacht,S. W. Running, and E. R.
Hunt Jr., Influence of fertilization on the allometric relations for two
pinesin contrastingenvironments,Can. J. For. Res.,23, 1704-1711,
1993.
Gower, S. T., J. G. lsebrands, and D. W. Sheriff, Carbon allocation and
accumulationin conifers, in ResourcePhysiologyof Conifers,edited
by W. Smith and T. M. Hinckley, Academic, San Diego, Calif.,
217-254, 1995.
Gower, S. T., R. E. McMurtrie, and D. Murty, Aboveground net
primaryproductiondeclinewith standage:Potentialcauses,Trends
Ecol. Evol., 11,378-382, 1996.
Grigal, D. F., Sphagnumproduction in forested bogs of northern
Minnesota, Can. J. Bot., 63, 1204-1207, 1985.
Acknowledgments. The study was jointly funded by NSF (DEB
9221668)and NASA (NAG5-2601) grantsto Gowerand Norman.The
authorsexpresstheir deepestappreciationto all the NASA staff scientistsat Goddard SpaceFlight Center who made BOREAS a huge
successand added humor to the long daysin the field. David Archer
assistedwith numerousfield measurements.ForestryCanada assisted
with the destructivesamplingof the trees that were usedto develop
allometric equationsto calculateabovegroundC distributionin the
vegetationand abovegroundnet primary production.NumerousUniversityof Wisconsin-Madisonundergraduateshelped establishthe researchsites,conductfield measurements,
and processsamplesin the
laboratory;two studentsthat deservemention by name, Jon Martin
and JoeHouse,assistedwith all facetsof the research.Finally,it is only
right that we acknowledgeour spousesand childrenwho made sacrificesand enduredhardshipswhilewe were in Canadaduringmuchof
the summer and fall of 1994.
References
Atjay, G. L., P. Ketner, and P. Duvigneaud,Terrestrial primary productionand phytomass,in The Global CarbonCycle,editedby B. E.
Bolin, T. Degens, S. Kempe, and P. Ketner, scopevol. 13, pp.
129-181, John Wiley, New York, 1977.
Grigal, D. F., C. G. Buttleman, and L. K. Kernik, Biomassand productivityof the woodystrataof forestedbogsin northernMinnesota,
Can. J. Bot., 63, 2416-2424, 1985.
Hunt, E. R., Jr., and S. W. Running, Simulated dry matter yieldsfor
aspenand sprucestandsin the North American boreal forests,Can.
J. Remote Sens., 18, 126-133, 1992.
Husch, B., C. I. Miller, and T. W. Beers, ForestMensuration, 3rd ed.,
402 pp., John Wiley, New York, 1982.
Keeling, C. D., J. F. S. Chin, and T. P. Whorl, Increasedactivity of
northernvegetationinferred from atmosphericCO2 measurements,
Nature, 382, 146-149, 1996.
Kloeppel, B. D., Abovegroundnet primary productionand resource
use of larch and sympatricevergreenconifers in contrastingclimates, Ph.D. dissertation, Univ. of Wisconsin, Madison, 1997.
Kucharik, C. J., J. M. Norman, L. M. Murdock, and S. T. Gower,
Characterizingcanopynonrandomness
with a multibandvegetation
imager (MVI), J. Geophys.Res.,this issue.
Landsberg,J. J., and S. T. Gower,Applicationsof Physiological
Ecology
to ForestManagement,Academic,San Diego, Calif., 1997.
Landsberg, J. J., S. D. Prince, P. G. Jarvis, R. E. McMurtrie,
R. Luxmoore, and B. E. Medlyn, Energy conversionand use in
forests:The analysisof forest productionin terms of radiation utilization efficiency(e), in The Useof RemoteSensingin the Modeling
of ForestProductivity
at ScalesFrom Standsto Globe,editedby H. L.
GOWER ET AL.: BOREAL FOREST CARBON BUDGETS
Gholz, K. Nakane, and H. Shirada, Kluwer Acad., Norwell, Mass.,
1997.
McMurtrie, R. E., H. L. Gholz, S. Linder, and S. T. Gower, Climatic
factors controlling the productivityof pine stands:A model-based
analysis,Ecol. Bull., 43, 173-188, 1994.
Oechel, W. C., and K. Van Cleve, The role of bryophytesin nutrient
cyclingin the taiga,in ForestEcosystems
in theAlaskan Taiga,edited
by K. Van Cleve, F. S. Chapin III, P. W. Flanagan,L. A. Viereck,
and C. T. Dyrness,Springer-Verlag,pp. 121-137, New York, 1986.
Perala, D. A., and D. H. Alban, Biomass, nutrient distribution, and
litterfall in Populus,Pinus and Picea standson two different soilsin
Minnesota, Plant Soil, 64, 177-192, 1982.
29,041
Steele, S. J., S. T. Gower, J. G. Vogel, and J. M. Norman, Root
biomass,net primary productionand turnoverof aspen,jack pine
and black spruce standsin Saskatchewanand Manitoba, Canada,
Tree Physiol.,17, 577-587, 1997.
Tans, P. P., I. Y. Fung, and T. Takahashi,Observationalconstraintson
the global atmosphericCO2 budget, Science,247, 1431-1438, 1990.
Van Cleve, K., L. K. Oliver, P. Schlentner, L. A. Viereck, and C. T.
Dyrness,Productivityand nutrient cyclingin taigaforestecosystems,
Can. J. For. Res., 13, 747-766, 1983.
Vogel, J. G., Carbon and nitrogendynamicsof borealjack pine stands
with different understoryvegetation,M.S. thesis,Univ. of Wisconsin, Madison, 1997.
Quay, P. D., B. Tilbrook, and C. S. Wong, Oceanicuptake of fossilfuel Wang, Y. P., and P. J. Polglase,Carbon balance in the tundra, boreal
CO2:Carbon-13 evidence, Science,256, 74-79, 1992.
forest and humid tropical forest during climate change:Scalingup
Reader, R. J., and J. M. Stewart,The relationshipbetweennet primary
from leaf physiologyand soil carbon dynamics,Plant Cell Environ.,
18, 1226-1244, 1995.
productionand accumulationfor a peatlandin southeasternManitoba, Ecology,53, 1024-1037, 1972.
Waring, R. H., and W. H. Schlesinger,ForestEcosystems:
Conceptsand
Ruark, G. A., and J. G. Bockheim,Biomass,net primary production,
Management,Academic, San Diego, Calif., 1985.
and nutrient distributionfor an age sequenceof Populustremuloides Whittaker, R. H., and G. E. Likens, Carbon in the biota, in Carbon and
ecosystems,Can. J. For. Res., 18, 435-443, 1988.
the Biosphere,edited by G. M. Woodwell and E. V. Pecan, Conf.
Ruess, R. W., K. Van Cleve, J. Yarie, and L. A. Viereck, Contributions
720510,pp. 281-302, Natl. Tech. Inf. Serv., Springfield,Va., 1973.
of fine root production and turnover to the carbon and nitrogen Zar, J. H., BiostatisticalAnalysis, Prentice Hall, Englewood, Cliffs,
N.J., 1983.
cyclingin taiga forestsof the Alaskan interior, Can. J. For. Res.,26,
1326-1336, 1996.
Runyon, J., R. H. Waring, S. N. Goward, and J. M. Welles, Environmental limit on net primary production and light use efficiency
acrossthe Oregon transect,Ecol. Applica., 4, 226-237, 1994.
Ryan, M. G., D. Binkley, and J. H. Fownes, Age-related decline in
forest productivity:Pattern and process,Adv. Ecol. Res., 27, 213262, 1996.
Schlesinger,W. H., Biogeochemistry:An Analysis of Global Change,
S. T. Gower (correspondingauthor), T. K. Stow, and J. G. Vogel,
Department of Forest Ecologyand Management,1630 Linden Drive,
University of Wisconsin, Madison, WI 53706. (e-mail:
[email protected].
edu)
C. J. Kucharik and J. M. Norman, Department of Soil Science,
University of Wisconsin,Madison, WI 53706.
S. J. Steele, Kemp Natural ResourceStation, Woodruff, WI 54568.
Academic Press, 1991.
Sellers, P., et al., The Boreal Ecosystem-Atmosphere study
(BOREAS): An overviewand early resultsfrom the 1994 field year,
Bull. Am. Meteorol. Soc., 76, 1549-1577, 1995.
Siegenthaler,U., and J. L. Sarmiento,Atmosphericcarbondioxideand
the ocean, Nature, 365, 119-125, 1993.
(ReceivedJune 10, 1996;revisedJuly 29, 1997;
acceptedAugust 14, 1997.)