WHAT DOES
NONFOREST LAND CONTRIBUTE TO THE GLOBAL C BALANCE?
Jennifer C. Jenkins[1]
and Rachel Riemann[2]
Citation: Proceedings, Third
Annual FIA Science Symposium, Traverse City, MI. Oct. 18-20, 2001. eds. Ron McRoberts and John Moser. St. Paul, MN: USDA Forest
Service General Technical Report NC-GTR-XX. In press.
ABSTRACT: Ð An inventory of land traditionally called ÒnonforestÓ and therefore not sampled by the FIA program was implemented by NE-FIA in 1999 for five counties in Maryland. Biomass and biomass increment were estimated from the nonforest inventory data using techniques developed for application to large-scale inventory data. Results were compared to estimates for forested land in Maryland. We conclude from this work that carbon (C) stocks and fluxes on nonforest land could add substantially to current estimates of local, regional, and national C balances, which are currently based on forest land only.
INTRODUCTION
Attempts
to quantify the global carbon (C) budget have focused heavily on the role of
forest growth and regrowth in C uptake (Caspersen
et al. 2000; Pacala et al. 2001; Wofsy 2001). The forest inventory approach, because it is typically based on
ground-measured data for a comprehensive, unbiased sample of forest land, has
widely been accepted as the most reliable approach for large-scale and
comprehensive estimation of forest C stocks and fluxes (Goodale
et al. 2001; Hicke et al. 2001; Pacala et al. 2001). The United States (US) forest C budget (Birdsey and Heath 1995, US
Government 2000), however, is based exclusively on land defined by the US
Department of Agriculture (USDA) Forest Service Forest Inventory and Analysis
(FIA) Program as Òforest.Ó
FIA
defines a stand of forestland as:
a) at least 1 acre in size; b) at least 120 feet wide; c) at least 10%
stocked; and d) not developed for another use (such as residential,
recreational, or agricultural) (Hansen
et al. 1992). Based on this definition, roughly two-thirds (67%) of the US land
base is considered ÒnonforestÓ (Smith
et al. 2001). This nonforest figure includes range and desert land in the arid
interior of the country; this arid land would not normally support forest
vegetation. Still, this definition of forest has critical gaps with respect to
large-scale C cycle estimation, especially for regions such as the northeastern
US, where trees and other vegetation are ubiquitous on land being used for all
types of purposes.
While
inventories of trees in urban areas do exist (Nowak
1994; Nowak and Crane 2001), these urban samples have been almost exclusively conducted within the
city limits. As a result, they do not include those areas missed from the FIA
ÒforestÓ sample in suburban, rural-residential, and rural-agricultural areas
outside the city limits. In this study, we examine the potential implications
of excluding nonforest land from the land base used to develop large-scale C
budgets.
METHODS
The Maryland
Nonforest Inventory
In
1999, a pilot study was undertaken to inventory the plots classified by FIA as
ÒnonforestÓ in 5 Maryland counties: Anne Arundel, Baltimore, Carroll, Harford,
and Howard (Figure 1). This 5-county area
covers 2237 mi2, and is home to 2,512,431 persons, according to the
2000 US Census. This 5-county region was selected to capture a gradient of
population density, urbanization, and landuse. In addition, the region is
identical to the 5-county area designated as the research site for the
Baltimore Ecosystem Study (BES), one of 24 Long-Term Ecological Research (LTER)
study sites across the US.
The
city of Baltimore is entirely urban, while large areas of suburban development
occur in 4 of the 5 counties. Population density ranges from 336 to 8059
persons per mi2 in rural-agricultural Carroll County and Baltimore
City, respectively. The pilot nonforest inventory was conducted concomitant
with the Maryland inventory in 1999: this timing increased the efficiency of
data collection, as the nonforest field crews collected the standard FIA plot
variables for nonforest plots as well as the additional variables required by
the nonforest inventory. By collecting data for the forest and nonforest
inventory simultaneously, we also ensured that the two inventory samples would
be comparable.
Details
of the nonforest inventory procedure are given in Riemann
(2001). Briefly, the nonforest inventory utilized the regular FIA plot ÔgridÕ
in Maryland. A 1/10th acre (37Õ radius circular) nonforest plot was
established around the center of subplot 1 if any nonforest condition occurred
on that subplot (Figure 2). The nonforest portion of that 1/10th
acre plot was then inventoried by the nonforest crew. A nonforest plot was not
established if the center subplot was entirely forested, even if nonforest
conditions did occur on any of the other subplots. The inventory methods and
protocols used by the FIA nonforest inventory crew were identical to the
standard FIA protocols wherever possible (USDA
Forest Service 2000).
In
1999, there were 243 forest and nonforest FIA plots in this 5-county
region. Of these, 146 were
classified as nonforest, 44 as forest, and 53 as mixed (i.e. containing both
forest and nonforest conditions). The mixed category contained 25 plots that
were entirely forested on subplot 1, and 28 plots that had some nonforest on
subplot 1. The nonforest crew inventoried 162 of these plots: 138 of the
nonforest plots and 24 of the mixed plots. Thus, eight of the nonforest plots
and four of the mixed plots were not sampled by the nonforest crew; these plots
are considered Òmissing,Ó and we assume that their exclusion does not bias this
analysis.
On
each plot, a subset of the standard FIA variables were collected, plus some
additional variables designed to better describe the tree health, biodiversity
and ground cover of trees in nonforest areas. Those regular FIA variables that
were considered to be less useful in nonforest areas, such as the
timber-related variables of cull and board feet, were excluded from the
nonforest sample. To better distinguish the types of areas in which the
nonforest plots and areas of high tree basal area were found, three additional
variables were added: detailed landuse class, detailed owner class, and reason
for nonforest status.
Biomass and NPP
from Nonforest Inventory Data
Net
primary production (NPP) is the rate at which C is accumulated by autotrophs
and is expressed as the difference between gross photosynthesis and autotrophic
respiration. Complete measurements of total NPP include annual above- and
below-ground production in both woody and non-woody biomass. In this study we
focus on annual biomass increment only, also known as wood NPP (WNPP), for two
reasons: a) we know of no data on litterfall and root production for nonforest
areas; and b) the wood component of NPP is the equivalent of annual C
sequestration and storage, since wood biomass turns over much more slowly than
the non-woody biomass compartments.
Total
tree biomass and wood net primary production (WNPP) were computed from plot-
and tree-level inventory data for the 162 nonforest plots in the 5-county area
in Maryland, using methods as described in Jenkins
et al. (2001a). WNPP was defined per tree as:
Wood
production per tree (kg yr-1) = [aboveground biomass (kg) (t1)
Ð aboveground biomass (kg) (t0)]/[t1 Ð t0
(yr)] (1)
where
t1 refers to the current year, and t0 refers to the year
at the beginning of the inventory period (in this analysis we assumed a
one-year sampling interval, and found dbht0 from dbht1 as
described below). Biomass estimates for current conditions (t1) were
found on a tree-by-tree basis from diameter at breast height (dbh) using
species-group regression equations as described by Jenkins
et al. (2001b). To find biomass and biomass increment on a per unit area basis from
tree-level measurements, the tree-level estimates were multiplied by the
expansion factor representing the number of trees per unit area represented by
that individual stem.
Because
these nonforest plots have only been censused once, to obtain estimates of
growth it was necessary to estimate dbh growth for all trees. This was accomplished using linear
algorithms developed for mid-Atlantic forests, which relate current diameter to
predicted diameter increment (Jenkins
et al. 2001a):
dbht0
(cm) = dbht1 (cm) Ð [average dbh growth rate (cm yr-1)] *
[remeasurement period (yr)]. (2)
Biomass
values were converted to C using 0.475 as the proportion C in biomass (Raich
et al. 1991).
For
comparison, biomass and WNPP were also computed for 316 remeasured forested
plots in Maryland from the 1985 inventory using the same techniques. These
values were further compared to biomass and WNPP data obtained using methods originally
described by (Birdsey 1992), applied to timber volume growth and mortality data for Maryland and
the entire northeastern region from the 1997 Resource Planning Act (RPA)
assessment (tables at http://www.fs.fed.us/fia/).
To
aggregate the nonforest inventory information from the 5 counties to the state
level, an average nonforest biomass and WNPP value per unit area was computed
for all 162 plots. No attempt was made to select plots with trees or on
particular land types; as a result, this sample is assumed to be representative
of the tree cover on an average piece of ÒnonforestÓ land in Maryland. This
average per-unit-area biomass and WNPP value was multiplied by the nonforest
land area in that state to approximate the aggregate state-level biomass and
WNPP totals on nonforest land. A parallel procedure was followed for forest
land in Maryland, except that the sample of forest plots was representative of
forest land over the entire state rather than the smaller 5-county region.
RESULTS
Maryland
Tree biomass stocks for Maryland forests computed using the Jenkins and Birdsey methods were comparable (Table 1). The larger biomass stocks computed from the RPA data most likely occurred because the RPA data apply exclusively to timberland, which is selected for its high productivity. Tree biomass stocks for nonforest land in Maryland were, per unit area, roughly 25% of the biomass computed for forested land as computed for all forests (Table 2). However, because there is a substantial amount of nonforest land in Maryland, the ratio of total biomass stocks on nonforest: forest land in Maryland was roughly 0.33 (Table 2).
Per-unit-area
wood production values for Maryland forests were also similar when computed
using the Jenkins and Birdsey methods (Table 1). The RPA data predicted
somewhat larger wood-biomass increments; this is probably (again) due to the
exclusion of nonproductive forests from the RPA sample. Per unit area, wood
production on nonforest land was approximately 22% of wood production on
forested land (Table 2). As with
forest C stocks, however, because of the large proportion of nonforest land in
Maryland, the ratio of total annual C storage on nonforest: forest land was
higher than this (Table 2).
Northeast
To
aggregate these values to the regional level for large-scale comparisons, we
assumed that the Maryland ratios of nonforest: forest C stocks and fluxes are
true for the region. Per-unit-area WNPP on all nonforest land in the region was
therefore assumed to be 22% of wood production on forested land, and
per-unit-area biomass on nonforest land was assumed to total 25% of biomass on
forest land. In the northeast
region (as defined by RPA), we calculate that nonforest land contributes about
280 million metric tons of C in tree biomass (one metric ton = 1 Mg = 106 g), or about 14
million metric tons every year (Table 1).
This adds to roughly 13% of the biomass and 24% of the WNPP on forested
land (Table 2).
DISCUSSION
There
is widespread consensus that a ÒmissingÓ carbon sink (i.e. the difference
between C emitted from anthropogenic and non-anthropogenic activities on the
surface of Earth, and the C sequestered in terrestrial and oceanic ecosystems
or stored in wood products) of up to 1-2 Pg C yr-1 (1 Pg = 1015
g) exists in terrestrial systems in the northern midlatitudes. Ongoing efforts
to find the missing C using different measurement methods have yielded
conflicting results (Birdsey and Heath 1995; Fan et al. 1998; Schimel et
al. 2000), though current estimates are converging toward a US sink between 0.35
and 0.90 Pg C yr-1 (Pacala
et al. 2001). Forest inventory measurements currently suggest that forest trees in
the United States remove between 0.11 and 0.15 Pg C yr-1 from the
atmosphere, but these estimates are currently based only on land classified by
FIA as Òforest.Ó If we assume that the ratios of forest: nonforest WNPP and
forest: nonforest land are similar for the rest of the United States, then
nonforest land could add an additional 24% to this value. In other words, based on the results of
this analysis it is possible that trees on nonforest land are storing an
additional 0.03 to 0.04 Pg C yr-1. This could amount to 10% of the existing ÒmissingÓ sink of
0.35 to 0.90 Pg C yr-1.
While
these results suggest that trees on nonforest land almost certainly contribute
to overall C sequestration, much more research is needed to understand the
dynamics of C stocks and fluxes in nonforest areas. For example, in this analysis we have excluded all
consideration of ornamental shrubs and grasses, which must sequester additional
C. Soils in gardens and other
cultivated non-agricultural areas are likely to harbor C as well, in
near-direct proportion to the types of management these lands experience.
The
diameter growth algorithms and the biomass regression equations used in this
analysis were developed for forest trees.
Research on urban trees suggests that open-grown urban trees have larger
crowns but lower biomass values than forest trees (Nowak
1996). Their diameter growth
rates are likely to be higher than those of forest trees, however. The chances are good that the WNPP
values presented here for nonforest land in Maryland are too low, but the
biomass values may be too high.
It
is difficult to extrapolate the results of this analysis to the entire United
States because the patterns of urbanization and land use change are likely to
differ from region to region. For
example, there may be very little nonforest land in rural states such as Maine.
In arid regions, there may be little difference between biomass and WNPP in
residential and non-residential areas. But these relationships may also be much
more complex: it is possible that irrigation in arid regions may increase
residential woody biomass and production. An analysis such as this one,
conducted for urban areas in different regions across the country, should help
to resolve the issue of nonforest, non-agricultural C sequestration.
ACKNOWLEDGMENTS
The
authors thank the entire NE-FIA unit, and especially the data acquisition
group, for their cooperation in implementing the 1999 inventory of nonforest
land in Maryland.
LITERATURE CITED
Birdsey,
R.A. 1992. Carbon storage and accumulation in United States forest ecosystems.
Washington, DC: USDA Forest Service. General Technical Report WO-59.
Birdsey, R.A.; Heath, L.S. Carbon changes in U.S.
forests. In: Joyce, L.A., editor; 1995; Fort Collins, CO. USDA Forest Service,
Rocky Mountain Forest and Range Experiment Station. General Technical Report RM-GTR-271. p 56-70.
Caspersen, J.; Pacala, S.; Jenkins, J.; Hurtt, G.;
Moorcroft, P.; Birdsey, R. 2000. Contributions of land-use history to carbon
accumulation in US forests. Science 290:1148-1151.
Fan, S.; Gloor, M.; Mahlman, J.; Pacala, S.;
Sarmiento, J.; Takahashi, T.; Tans, P. 1998. A large terrestrial carbon sink in
North America implied by atmospheric and oceanic carbon dioxide data and
models. Science 282:442-446.
Goodale, C.L.; Apps, M.J.; Birdsey, R.A.; Field, C.B.;
Heath, L.S.; Houghton, R.A.; Jenkins, J.C.; Kohlmaier, G.H.; Kurz, W.; Liu, S.
et al. . 2001. Forest carbon sinks in the northern hemisphere. Ecological
Applications, in press.
Hansen, M.H.; Frieswyk, T.; Glover, J.F.; Kelly, J.F.
1992. The Eastwide Forest Inventory Data Base: Users Manual. St. Paul, MN: United States Department of
Agriculture Forest Service North Central Experiment Station. General Technical
Report NC-151.
Hicke, J.; Asner, G.; Randerson, J.; Tucker, C.; Los,
S.; Birdsey, R.; Jenkins, J.; Field, C.; Holland, E. 2001. Satellite-derived
increases in net primary productivity across North America, 1982-1998.
Geophysical Research Letters, in press.
Jenkins, J.; Birdsey, R.; Pan, Y. 2001a. Biomass and
NPP estimation for the mid-Atlantic region (USA) using forest inventory data.
Ecological Applications 11:1174-1193.
Jenkins, J.; Chojnacky, D.; Heath, L.; Birdsey, R.
2001b. National-scale biomass estimators for United States tree species. Forest
Science, in press.
Nowak, D. 1994. Urban forest structure: The state of Chicago's urban forest.
In: McPherson, E.; Nowak, D.; Rowntree, R., editors. Chicago's urban forest
ecosystem: results of the Chicago
urban forest climate project. Radnor, PA: USDA Forest Service Northeastern
Forest Experiment Station. p 3-18.
Nowak, D. 1996. Estimating leaf area and leaf biomass
of open-grown deciduous urban trees. Forest Science 42(4):504-507.
Nowak, D.; Crane, D. 2001. Carbon storage and
sequestration by urban trees in the United States. Environmental Pollution, in
press.
Pacala, S.; Hurtt, G.; Houghton, R.; Birdsey, R.;
Heath, L.; Sundquist, E.; Stallard, R.; Baker, D.; Peylin, P.; Moorcroft, P. et
al. . 2001. Consistent land- and atmosphere-based US carbon sink estimates.
Science 292(5525):2316-2320.
Raich, J.W.; Rastetter, E.B.; Melillo, J.M.;
Kicklighter, D.W.; Steudler, P.A.; Peterson, B.J. 1991. Potential net primary
productivity in South America:
application of a global model. Ecological Applications 1:399-429.
Riemann, R. 2001. Pilot inventory of FIA plots
traditionally called 'nonforest'. USDA Forest Service, Northeastern Research
Station. Unpublished report.
Schimel, D.; Melillo, J.; Tian, H.; McGuire, A.;
Kicklighter, D.; Kittel, T.; Rosenbloom, N.; Running, S.; Thornton, P.; Ojima,
D. et al. 2000. Contribution of increasing CO2 and climate to carbon
storage by ecosystems in the United States. Science 287:2004-2006.
US Forest Service. 2000. Forest inventory and analysis
national core field guide, volume 1: field data collection procedures for phase
2 plots, version 1.4. Washington, DC: USDA Forest Service, Washington Office.
Internal report.
Smith, W.; Vissage, J.; Sheffield, R.; Darr, D. 2001.
Forest Resources of the United States, 1997. St. Paul, MN: USDA Forest Service
North Central Research Station. General Technical Report.
US Government (2000). United States Submission on Land Use, Land-Use Change, and Forestry. US Department of State, Washington, DC.
Wofsy,
S. 2001. Where has all the carbon gone? Science 292(5525):2261-2263.
Table 1.-- Biomass and
WNPP statistics for Maryland and the Northeast.
Maryland
(Jenkins) Northeast
(Birdsey) Maryland (Birdsey)
forest: (all forest) (RPA/
timberland) (RPA/
timberland)
land area
(thousand ac) 2701 78923 2423
average biomass
(Mg C/ha) 72.25 67.01 81.07
wood-biomass increment
(Mg C/ha/yr) 1.90 1.91 2.87
total C storage
(x 10^6 Mg C) 78.96 2141.31 79.50
annual C storage
(x 10^6 Mg C/yr) 2.08 61.06 2.82
nonforest: Maryland (Jenkins) Northeast
nonforest land area
(thousand ac) 3594 41330
average biomass
(Mg C/ha) 17.80 16.75
wood-biomass increment
(Mg C/ha/yr) 0.42 0.42
total C storage
(x 10^6 Mg C) 25.92 280.23
annual C storage
(x 10^6 Mg C/yr) 0.61 14.45
Table 2.-- Ratios of nonforest: forest statistics for
Maryland and the northeast.
Maryland Northeast
forest land (thousand ac) 2701 85484
nonforest land (thousand ac) 3594 41333
nonforest: forest land area 1.33 0.48
per-unit-area ratios
nonforest: forest biomass 0.25 assume
25%
nonforest: forest WNPP 0.22 assume
22%
aggregate state & region-level
nonforest: forest biomass 0.33 0.13
nonforest: forest WNPP 0.29 0.24
FIGURE
CAPTIONS
Figure 1. Ð The 5-county study area and all 243 FIA plots. Any FIA plot with nonforest occurring at the center subplot was visited by the nonforest crew. This included both the ÒnonforestÓ and Òmixed-nonforestÓ plots (i.e. some nonforest condition at the center subplot). From Riemann (2001).
Figure 2. Ð Nonforest inventory plot design compared to standard FIA plot design. From Riemann (2001).
Figure 1.
Figure 2.
