- Email: [email protected]
Aspect differences in above- and belowground carbon allocation: a Montana case-study
Environmental Pollution 116 (2002) S149–S155 www.elsevier.com/locate/envpol
Aspect differences in above- and belowground carbon allocation: a Montana case-study D. Coblea,*, J. Marshallb a
Arthur Temple College of Forestry, Stephen F. Austin State University, Box 6109, SFA Station Nacogdoches, TX, 75962, USA b Forest Resources Department, University of Idaho, Moscow, ID 83844-1133, USA
‘‘Capsule’’: Aboveground net primary production and belowground gross primary production were found to be greater on the north aspect for all leaf area indices in young forest stands in western Montana. Abstract Aboveground net primary production (ANPP) and belowground gross primary production (BGPP) of all vegetation were measured in eight young, paired plots on a north and south aspect in western Montana. Stands of high and low overstory tree leaf area index (LAI) were compared. BGPP increased with ANPP, though they were not directly proportional. ANPP ranged from 1550 to 4400 kg C ha1 year1 and BGPP ranged from 1360 to 3500 kg C ha1 year1. ANPP and BGPP were both significantly related to LAI and aspect, where both were greater on the north aspect at any given LAI. Litterfall represented the largest share of ANPP; increases in overstory biomass represented the next largest share. Soil CO2 flux was higher on the north aspect. We conclude that growth differences were not simply a matter of re-allocating carbon between root production and ANPP. Rather, both production and allocation were different among the sites. # 2001 Elsevier Science Ltd. All rights reserved. Keywords: Carbon allocation; Net primary production; Aspect; Soil respiration; Site productivity
1. Introduction Carbon storage or carbon sequestration has emerged as a possible land use activity to offset greenhouse gas emissions of CO2 in the atmosphere (Intergovernmental Panel on Climate Change, 1991). Forests can contribute to the reduction of greenhouse gases through carbon sequestration, but estimates of the size of the contribution can only be as reliable as the measurements upon which they are based (Birdsey et al., 1993). The amount of carbon fixed into biomass at a site is typically measured as the sum of above and belowground production of all the vegetation. Ecologists refer to these fixation processes as net primary production (NPP, kg C ha1 year1; Waring and Schlesinger, 1985. However, measurement of belowground production is difficult and uncertain, in part because the methodologies used to measure it are indirect and prone to errors (Singh et al., 1984; Nadelhoffer and Raich, 1992). Yet, accurate assessment of belowground production is important because this production can represent up to 76% of the annual total net primary production (Grier * Corresponding author. Tel.: +1-936-468-3301; fax: +1-936-4682489. E-mail address: [email protected] (D. Coble).
et al., 1981; Vogt, 1991; Nadelhoffer and Raich, 1992; Gower et al., 1995, 1996). The proportion of the production allocated belowground is often increased as belowground resources more strongly limit growth (Grier et al., 1981; Keyes and Grier, 1981; Running, 1983; Axelsson and Axelsson, 1986; Santantonio, 1989). Little research has addressed the question of aspect differences in allocation patterns. However, one might expect that the higher light intensity and lower soil moisture availability on south aspects might enhance belowground allocation. Whether the total production would be similar on north and south aspects is difficult to predict, but the longer growing season and higher light intensity on the south aspect lead some models to predict higher productivity (Wykoff et al., 1982; Milner et al., 1996). We assessed differences in allocation and total production at a study site in western Montana. The study sampled natural vegetation, both woody and nonwoody, on north and south aspects. We purposely sought out stands of high and low leaf area index (LAI) on both aspects. Although this is a case study, the results highlight aspect differences in production and allocation. The results of this case study lead to general hypotheses regarding the topographic distribution of carbon sequestration in high-latitude, mountainous terrain.
0269-7491/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved. PII: S0269-7491(01)00268-8
S150
D. Coble, J. Marshall / Environmental Pollution 116 (2002) S149–S155
2. Materials and methods 2.1. Objectives Using data from a case study at a site with varying aspect and varying leaf area index in western Montana, we examined the relationship between above and belowground carbon production. We tested the hypothesis that losses in aboveground productivity would be directly offset by increased carbon allocated belowground.
overstory tree density [low (L) and high (H) trees per hectare] and replicated twice for a total of eight plots (Table 1). Since aspect is not replicated at other locations, this project was considered a case study. Plots were square and 0.04 ha (0.1 acre) in size. On each plot, total annual aboveground biomass accumulation (AGBA) was measured by a combination of destructive sampling and dimensional analysis (Whittaker and Woodwell, 1968; Grier et al., 1989). Measurement procedures varied by type of vegetation. Measurements of the physical site included: slope ( 1%), elevation ( 1 m), and aspect ( 1 ).
2.2. Sampling design and site description 2.3. Vegetation measurements The study site comprised eight plots installed on opposing north and south aspects in a young (< 30 year old) forest located in west-central Montana at 1394 m (Coble et al., 2001). The location was chosen primarily because it had contrasting aspects, which created a contrasting moisture regime over a short distance (Coble, 1997). Plots were installed in close proximity on each aspect to minimize plot-to-plot variability (e.g. differences in soils, slope, etc.). The area was clear-cut harvested in 1976, planted in 1977, and precommercially thinned in 1980. However, the regeneration was predominantly natural because most of the planting stock died. The vegetative community on the south aspect consisted of patchy ponderosa pine (Pinus ponderosa Dougl. ex Laws.) and Douglas-fir [Pseudotsuga menziesii (Mirb.) Franco] in the overstory, and grasses (Agropyron spicatum, Calamagrostis rubescens), sedges (Carex spp.), snowberry (Symphoricarpos albus), and knapweed (Centaurea maculosa) in the understory. Vegetation on the north aspect consisted of mostly homogeneous western larch (Larix occidentalis Nutt.) and lodgepole pine (Pinus contorta Dougl. ex Loud.) with some Douglas-fir and ponderosa pine in the overstory, and grasses, sedges, snowberry, huckleberry (Vaccinium spp.), spirea (Spiraea betulifolia var. lucida), serviceberry (Amelanchier alnifolia), and various forbs in the understory. Four plots were located on each of the two aspects [north (N) and south (S)] in two different levels of
AGBA of the understory vegetation was estimated via clipping. Deciduous shrubs, forbs, grasses, and trees less than or equal to 3 years old were sub-sampled on four (for the north aspect) and eight (for the south aspect) subplots. The subplots were square, 4 m2 in area, and were randomly located on the plot. On each sub-plot, current year’s foliage and lateral stem growth (if applicable) of all non-tree individuals were clipped, combined by species, placed in bags, labelled, and transported to the laboratory for drying. Radial increment of shrubs was ignored (Alaback, 1986). The trees were clipped, grouped by 1-year age classes (years 1, 2, 3), placed in bags, labelled, and transported to the laboratory for drying. The foliage, branches, and stems were not separated because these trees were so small and rare (total dry weight less than 2 g). Foliage and stem samples from the sub-plot vegetation were oven-dried for 24 h at 70 C and weighed to the nearest 0.01 g (Alaback, 1986). Herbivory was not explicitly measured in this study because cattle were excluded from the study area and deer browsing was minimal, and we saw no evidence of significant insect herbivory. All trees greater than 3 years of age were tallied on each plot. Diameter ( 0.25 cm), height ( 0.1 m), and crown ratio ( 5% of live crown on the bole) were measured and recorded for all standing trees. Allometric equations were applied to this tally to estimate
Table 1 Site and density characteristics of four plots on each of two aspects Aspects
Plot
Aspect (degrees)
Slope (%)
Trees per hectare
Overstory tree LAIa (m2 m2)
Overstory density
North North North North South South South South
1 2 3 4 1 2 3 4
340 356 354 314 220 218 214 224
18 11 12 12 25 24 27 24
667.2 444.8 963.7 963.7 321.2 222.4 1532.1 1705.0
0.9 1.3 3.3 3.4 0.4 0.2 3.4 4.6
Low Low High High Low Low High High
a
Note: all-sided leaf area index (LAI) values were based on specific leaf area values not measured on the study site and should therefore be interpreted as rough estimates.
D. Coble, J. Marshall / Environmental Pollution 116 (2002) S149–S155
AGBA for trees greater than 3 years of age (procedure described in next section). 2.4. Aboveground data summarization The tree tally data were used with local allometric equations (Coble, 1997) to find the AGBA in leaf, branch, stem, and bark components for standing trees greater than 3 years old on each plot. The AGBA values for trees destructively sampled to build the allometric equations were added to the AGBA values of the tally trees. The AGBA values for all trees greater than 3 years old were then expanded to a per hectare basis for each plot. Dry-weight AGBA of the understory plants on each sub-plot was summed on a total, life form, and specieslevel basis to determine the change in understory biomass for the sub-plot. For trees less than or equal to 3 years old on the sub-plots, the total dry weight was divided by the age of the tree to determine its average AGBA, then added to the sub-plot AGBA totals. The AGBA values for each sub-plot were then averaged by plot to find the average plot-level understory AGBA. Mean AGBA values for each plot were then expanded to a per hectare basis. The dry weight AGBA values for tree and understory vegetation were multiplied by 0.5 to convert grams of dry weight to grams of carbon (Vogt, 1991). All-sided LAI (m2 m2) for overstory trees on each plot was calculated as the product of leaf carbon and specific leaf area (SLA, m2 kg C1). LAI was calculated for each tree and summed to obtain the plot-level allsided overstory tree LAI. Since SLA was not measured in this study, the following SLAs were used: Douglas-fir and western larch SLA=25.0 m2 kg C1 (Pierce et al., 1994), lodgepole pine and ponderosa pine=15.0 m2 kg C1 (Pierce et al., 1994). Plot-level LAI values were only used as covariates for overstory tree density in statistical analyses (see ANCOVA description later); they were not used to estimate or measure the amount of carbon. 2.5. Soil CO2 flux measurements Belowground CO2 efflux (soil respiration) was measured with a LiCor 6200 gas exchange analyzer with the soil chamber attachment (LiCor Inc., Lincoln, Nebraska, USA). The LiCor 6200 is a closed-system infrared gas analyzer (IRGA) that measures the amount of CO2 flux from the soil. Ten sampling points were randomly located in the buffer of each study plot for a total of 80 samples. The points were located in the plot buffers to avoid sampling in areas where earlier destructive sampling was conducted. Each sampling point was prepped by removing all vegetation inside of a 0.00856 m2 sampling collar imbedded 2 cm in the ground. Measurements were taken with the LiCor 6200
S151
twice a day (morning and evening) at seven different times throughout the sampling period. This sampling period approximately encompassed snowmelt/ground thaw in the spring until the first heavy snowfall in the fall (206 days for the south aspect and 188 days for the north aspect). Soil temperature ( C) at a 10 cm depth was also measured with each LiCor 6200 CO2 flux measurement. Soil CO2 fluxes (mmol m2 s1) at each sampling point were determined by taking four readings at 6–10 ppm differentials, then averaging the two flux values immediately above and below the average ambient CO2 value for the plot. The daily average soil CO2 flux for each sampling point was calculated as the average of the morning and evening flux measurements. 2.6. Litterfall measurement Litterfall was also measured adjacent to each sampling point. Litterfall (including branches and twigs) was collected at the end of the growing season in 0.00856 m2 traps (the same size as the soil collars), which were located in a random direction within 10 cm of the soil collar. Litterfall samples were bagged, transported to the lab, dried for 48 h at 70 C, then removed and weighed to the nearest 0.1 g. The dry mass was multiplied by 0.5 to convert grams of dry weight to grams of carbon (Vogt, 1991). 2.7. ANPP calculation Total aboveground net primary productivity (ANPP) for all vegetation on each plot was calculated as the sum of the component biomass weights (kg C) for all the life forms plus litterfall: ANPP ¼ Tðl; b; sÞ þ tðl; b; sÞ þ Sðl; sÞ þ GðlÞ þ FðlÞ þ D where ANPP=aboveground net primary production (kg C ha1 year1); T (l, b, s)=leaf, branch, and stem production of trees greater than 3 years old; t(l, b, s)=leaf, branch, and stem production of trees less than or equal to 3 years old; S(l, s)=leaf and stem production of shrubs; G(l)=leaf production of grasses; F(l)=leaf production of forbs; and D=detritus production or litterfall. 2.8. BGPP calculation A carbon budget method was used to estimate the upper limit for total annual carbon allocation to roots (Raich and Nadelhoffer, 1989). This technique is based on the conservation of mass; i.e. all carbon entering the soil must either exit the soil or enter the soil carbon pool. The method assumes that the soil carbon pool is in steady state on an annual basis; i.e. the annual amount of carbon entering the soil equals the annual amount
S152
D. Coble, J. Marshall / Environmental Pollution 116 (2002) S149–S155
of carbon leaving the soil. Under this assumption, heterotrophic respiration (Rh) is assumed to equal the annual amount of carbon allocated into the soil from above (Pa=litterfall) and belowground (Pb): Rh ffi Pa þ Pb
ð1Þ
Rh is often impossible to measure directly because it is mixed with respiration from the roots (Rr). Therefore, the method estimates annual carbon allocation to roots by subtracting litterfall from total soil respiration to estimate the sum of root respiration and root production: Rs Pa ffi Pb þ Rr
ð2Þ
We refer to this sum as belowground gross primary production (BGPP). The Raich and Nadelhoffer (1989) technique was used in this study to estimate the total carbon allocation to roots (BGPP) at each sampling point for each measurement period. Though the forest sampled in this study was young, it had reached canopy closure on the plots with high tree densities, which makes us believe that the site had reached, or was approaching, steady state. From the measured BGPP values, weighted averages (the weights were the number of days in each measurement period) were calculated across the seven measurement periods to find the seasonal average BGPP for each sampling point. These weighted averages were then averaged to find the seasonal average BGPP for each study plot. 2.9. Comparison of ANPP and BGPP Simple linear regression (SLR; Zar, 1999) was used to test two alternative hypotheses about the relationship between BGPP and ANPP. The two were related according to the following linear model: BGPP ¼ b0 þ b1 ANPP
hypotheses: Ho1: (0,1)=(0,2), Ha1: not Ho1, and Ho2: (0,1)=(0,2), Ha2: not Ho2. An analysis of covariance (ANCOVA) was also used to test if mean BGPP and ANPP were equal for the two levels of overstory density (the covariate=overstory tree LAI) and the two aspects (the class variable), and that the density–aspect interaction does not significantly affect production (Zar, 1999). The Fmax test was used to test for homogeneity among variances in production values between aspects prior to the ANCOVA (Sokal and Rolf, 1995). ANCOVA was employed to test for differences in carbon production rather than ANOVA or a Kruskal– Wallis test because overstory tree density was not an experimentally controlled variable in this study. Therefore, it could not be assumed that measured carbon production was affected equally for the two levels of overstory density on the north versus south aspect installations. Though overstory tree LAI was highly correlated with overstory tree density1 (R2=0.861), overstory tree LAI was used as the covariate rather than tree density (TPH; trees per hectare) in the ANCOVA. Tree LAI was chosen because it was considered to be a better measure of the competitive effect of tree density on site productivity than numbers of trees, simply because it represents photosynthetic capacity of the trees on a site (Waring and Schlesinger, 1985).
3. Results Belowground carbon allocation to roots (BGPP, kg C ha1 period1; 206 days on south aspect, 188 days on north aspect) increased with total ANPP (kg C ha1 year1; Fig. 1). However, allocation was not constant across plots because the actual slope value of the SLR model was less than 1 (P < 0.0001) and the intercept was significantly different from zero (P < 0.0001). Furthermore, the slope was also quite different from 2 (P < 0.0001), which leads us to conclude that root
ð3Þ
If allocation were constant across the plots, then the intercept (b0) would not be different from zero and the slope (b1) would be approximately two. (The value of two was chosen to compare BGPP and ANPP; since ANPP does not include respiration and respiration represents roughly half the amount of gross primary production, we needed to use the value of two in our hypothesis tests). If, however, BGPP were traded off directly against ANPP, and BGPP were constant, then the value of b0 should be equal to BGPP and the slope (b1) should equal approximately 2. If both BGPP and allocation were changing, then the result should be intermediate between these two extremes. Simultaneous F-tests (Neter et al., 1985) were used to evaluate the two
Fig. 1. Comparison of above (ANPP) and belowground (BGPP) primary production.
D. Coble, J. Marshall / Environmental Pollution 116 (2002) S149–S155
production was not directly traded off against aboveground production. Instead, both BGPP and allocation were apparently different among the plots. Given the positive slope of the relationship, it appears that variation in BGPP was quantitatively more important than reallocation. This should perhaps not be surprising considering that the plots varied in LAI from 1.7 to 5.4 (Fig. 2). ANPP was significantly (P=0.0031) greater on the north versus south aspects (Fig. 3). ANPP also significantly (P=0.0336) increased with tree LAI (Fig. 3). Also, there was no aspect/overstory tree LAI interaction nor any heterogeneity of variances (P < 0.05) for either ANPP or BGPP, so the assumptions of ANCOVA were not violated. Thus, ANPP varied among sites rather than simply re-routed between overstory trees and understory vegetation. Based on this result, we conclude that the reduction in ANPP on the south aspect was not a result of shifts in allocation to BGPP. We also conclude that production was not re-routed belowground on sites with low overstory tree density. Soil CO2 flux and litterfall, both of which were used to calculate BGPP, were greater for north versus south aspects (Figs. 4 and 5). This was an expected result
Fig. 2. Leaf area index (LAI) for tree and non-tree vegetation by aspect (N=North, S=South) and density (H=High, L=Low). Error bars represent one standard deviation. Note: all LAI values calculated from specific leaf areas not measured on the site.
Fig. 3. Above (ANPP) and belowground (BGPP) primary production by aspect (N=North, S=South) and density (H=High, L=Low). Error bars represent one standard deviation.
S153
because the north aspect plots contained more plant biomass, which in turn likely contributed to greater amounts of root respiration and litterfall than on the south aspect. However, the high variability in the flux measurements on both aspects (see error bars in Fig. 4) was not expected because 10 sampling points per plot was believed to be an adequate sample of soil CO2 flux. These results confirm that soil CO2 flux is a heterogeneous process that requires intensive sampling to accurately measure.
4. Discussion As mentioned in Section 1, methodologies used to measure belowground carbon production can be prone to errors. The Raich–Nadelhoffer model used to estimate BGPP in this study assumed that carbon inputs to the soil equaled carbon outputs from the soil. As stated earlier, we believe this is a reasonable assumption for this forest. We do not believe this would be the case had we sampled immediately after planting, but we do believe that 30 years is an adequate time for this site to reach a steady-state. The steady-state assumption is also questionable in the western United States where detritus (i.e. litter) inputs usually exceed outputs (Olson, 1963; Harvey et al., 1978), unless they are removed by fire (Wright and Heinselman, 1973). Since this study site was clearcut
Fig. 4. Soil CO2 flux by aspect (N=North, S=South) and density (H=High, L=Low). Error bars represent one standard deviation.
Fig. 5. Litterfall by aspect (N=North, S=South) and density (H=High, L=Low). Error bars represent one standard deviation.
S154
D. Coble, J. Marshall / Environmental Pollution 116 (2002) S149–S155
harvested in 1976, numerous large stumps and large downed logs remain from previous logging activity, though much of the smaller material had decayed. Coarse root increment was not measured for this study, so it was not added to the BGPP estimates (Ryan et al., 1996). Other studies found that coarse root production as a percent of total belowground production equaled 12–40% for Pinus radiata (Ryan et al., 1996), 6–15% for Abies amabilis (Grier et al., 1981, their Table 3), and 14–40% for Douglas-fir (Keyes and Grier, 1981). Therefore, the BGPP measurements based on the RN model alone were conservative, possibly upwards of 40 or 50%. BGPP is likely even more conservative because winter soil CO2 flux was not measured in this study. Considerable amounts of CO2 can be produced from the soil during the winter months under snow packs (Harvey et al., 1978; McDowell et al., 2000). McDowell et al. (2000) found that 17% of the total annual soil CO2 flux was through a snow pack, and the snow pack lasted 36% of the year. Coxson and Parkinson (1987) reported that soil CO2 lost during the winter was nearly 60% of the annual biomass input for an aspen forest. Dormaar et al. (1981) found a 50% loss during winter for a grassland. Thus, annual BGPP estimates at this study site could actually be twice the reported values; i.e. +50% for winter respiration and +50% for coarse roots and woody debris.
5. Conclusions We used conventional measurement techniques to estimate ANPP and BGPP at a site in western Montana. We then used these production estimates to test the hypothesis that BGPP and ANPP are related on a 2/ 1 basis. Since we found that ANPP and BGPP values were higher on the north versus south aspects and high versus low overstory tree density levels, we concluded that these correlated increases argue against differential allocation among life forms or between roots and shoots for this case study site. Acknowledgements This paper was presented at the USDA Forest Service Southern Global Change Program sponsored Advances in Terrestrial Ecosystem: Carbon Inventory, Measurements, and Monitoring Conference held 3–5 October 2000 in Raleigh, North Carolina.
References Alaback, P.B., 1986. Biomass regression equations for understory plants in coastal Alaska: effects of species and sampling design on estimates. NW Science 60 (2), 90–103.
Axelsson, E., Axelsson, B., 1986. Changes in carbon allocation in spruce and pine trees following irrigation and fertilization. Tree Physiology 2, 189–204. Birdsey, R.A., Plantinga, A.J., Heath, L.S., 1993. Past and prospective carbon storage in the United States. Forest Ecology and Management 58, 33–40. Coble, D.W., 1997. Trends in above and below ground production of trees and non-tree vegetation on contrasting aspects in western Montana. Unpublished PhD dissertation, School of Forestry, University of Montana, Missoula, Montana, USA. Coble, D.W., Milner, K.S., Marshall, J.D., 2001. Above and below ground production of trees and other vegetation on contrasting aspects in western Montana: a case study. Forest Ecology and Management 142, 231–241. Coxson, D.S., Parkinson, D., 1987. Winter respiratory activity in aspen woodland forest floor litter and soils. Soil Biology and Biochemistry 19 (1), 49–59. Dormaar, J.F., Smoliak, S., Johnston, A., 1981. Seasonal fluctuations of blue grama roots and chemical characteristics. Journal of Range Management 34, 62–64. Gower, S.T., Isebrands, J.G., Sheriff, D.W., 1995. Carbon allocation and accumulation in conifers. In: Smith, W.K., Hinckley, T.M. (Eds.), Resource Physiology of Conifers. Academic Press, San Diego, California, USA, pp. 217–254. Gower, S.T., Pongracic, S., Landsberg, J.J., 1996. A global trend in belowground carbon allocation: can we use the relationship at smaller scales? Ecology 77 (6), 1750–1755. Grier, C.C., Vogt, K.A., Keyes, M.B., Edmonds, R.L., 1981. Biomass distribution and above- and below-ground production in young and mature Abies amabilis zone ecosystems of the Washington Cascades. Canadian Journal of Forest Research 11, 155–167. Grier, C.C., Lee, K.M., Nadkarni, N.M., Klock, G.O., Edgerton, P.J., 1989. Productivity of forests of the United States and its relation to soil and site factors and management practices: a review (General Technical Report PNW-222). USDA Forest Service, Pacific Northwest Forest and Range Experiment Station, Portland, OR. Harvey, A.E., Jurgensen, M.F., Larsen, M.J., 1978. Seasonal distribution of ectomycorrhizae in a mature Douglas-fir/larch forest soil in western Montana. Forest Science 24, 203–208. Intergovernmental Panel on Climate Change, 1991. Climate Change— the IPCC Response Strategies. Island Press, Washington, DC. Keyes, M.B., Grier, C.C., 1981. Above and belowground net production in 40-year-old Douglas-fir stands on low and high productivity sites. Canadian Journal of Forest Research 11, 599–605. McDowell, N.G., Marshall, J.D., Hooker, T.D., Musselman, R., 2000. Estimating CO2 flux from snowpacks at three sites in the Rocky Mountains. Tree Physiology 20, 745–753. Milner, K.S., Running, S.W., Coble, D.W., 1996. A biophysical soil-site model for estimating potential productivity of forested landscapes. Canadian Journal of Forest Research 26, 1174– 1186. Nadelhoffer, K.J., Raich, J.W., 1992. Fine root production estimates and belowground carbon allocation in forest ecosystems. Ecology 73 (4), 1139–1147. Neter, J., Wasserman, W., Kutner, M.H., 1985. Applied Linear Statistical Models, 2nd Edition. R.D. Irwin, Inc, Homewood, IL. Olson, J.S., 1963. Energy storage and the balance of producers and decomposers in ecological systems. Ecology 44, 321–322. Pierce, L.L., Running, S.W., Walker, J., 1994. Regional-scale relationships of leaf area index to specific leaf area and leaf nitrogen content. Ecological Applications 4 (2), 313–321. Raich, J.W., Nadelhoffer, K.J., 1989. Belowground carbon allocation in forest ecosystems: global trends. Ecology 70 (5), 1346–1354. Running, S.W., 1983. A biophysical assessment of the regeneration and growth potential of forested land in Montana. In: O’Loughlin, J., Pfister, R.D. (Eds.), Management of Second-growth Forests: The State of Knowledge and Research Needs, 14 March 1983, Missoula,
D. Coble, J. Marshall / Environmental Pollution 116 (2002) S149–S155 Montana, USA. Montana Forest and Conservation Experiment Station, School of Forestry, University of Montana, Missoula, Montana, USA, pp. 58–75. Ryan, M.G., Hubbard, R.M., Pongracic, S., Raison, R.J., McMurtrie, R.E., 1996. Foliage, fine-root, woody-tissue and stand respiration in Pinus radiata in relation to nitrogen status. Tree Physiology 16, 333–343. Santantonio, D., 1989. Dry-matter partitioning and fine-root production in forests—new approaches to a difficult problem. In: Periera, J.S., Landsberg, J.J. (Eds.), Biomass Production by Fast-Growing Trees. Kluwer Academic Publishers, Dordrecht, Netherlands, pp. 57–72. Singh, J.S., Lauenroth, W.K., Hunt, H.W., Swift, D.M., 1984. Bias and random errors in estimators of net root production: a simulation approach. Ecology 65, 1760–1764. Sokal, R.R., Rolf, F.J., 1995. Biometry, 3rd Edition. W.H. Freeman and Company, New York.
S155
Vogt, K.A., 1991. Carbon budgets of temperate forest ecosystems. Tree Physiology 9, 69–86. Waring, R.H., Schlesinger, W.H., 1985. Forest Ecosystems: Concepts and Management. Academic Press, Orlando, FL. Whittaker, R.H., Woodwell, G.M., 1968. Dimension and production relations of trees and shrubs in the Brookhaven Forest, New York. Journal of Ecology 56, 1–25. Wright, H.E., Heinselman, M.L., 1973. The ecological role of fire in natural conifer forests of western and northern North America: introduction. Quaternary Research 3, 319–328. Wykoff, W.R., Crookston, N.L., Stage, A.R., 1982. User’s Guide to the Stand Prognosis Model. General Technical Report INT-133. USDA Forest Service, Intermountain Forest and Range Experiment Station, Ogden, UT. Zar, J.H., 1999. Biostatistical Analysis, 4th Edition. Prentice-Hall, NJ.
Aspect differences in above- and belowground carbon allocation: a Montana case-study D. Coblea,*, J. Marshallb a
Arthur Temple College of Forestry, Stephen F. Austin State University, Box 6109, SFA Station Nacogdoches, TX, 75962, USA b Forest Resources Department, University of Idaho, Moscow, ID 83844-1133, USA
‘‘Capsule’’: Aboveground net primary production and belowground gross primary production were found to be greater on the north aspect for all leaf area indices in young forest stands in western Montana. Abstract Aboveground net primary production (ANPP) and belowground gross primary production (BGPP) of all vegetation were measured in eight young, paired plots on a north and south aspect in western Montana. Stands of high and low overstory tree leaf area index (LAI) were compared. BGPP increased with ANPP, though they were not directly proportional. ANPP ranged from 1550 to 4400 kg C ha1 year1 and BGPP ranged from 1360 to 3500 kg C ha1 year1. ANPP and BGPP were both significantly related to LAI and aspect, where both were greater on the north aspect at any given LAI. Litterfall represented the largest share of ANPP; increases in overstory biomass represented the next largest share. Soil CO2 flux was higher on the north aspect. We conclude that growth differences were not simply a matter of re-allocating carbon between root production and ANPP. Rather, both production and allocation were different among the sites. # 2001 Elsevier Science Ltd. All rights reserved. Keywords: Carbon allocation; Net primary production; Aspect; Soil respiration; Site productivity
1. Introduction Carbon storage or carbon sequestration has emerged as a possible land use activity to offset greenhouse gas emissions of CO2 in the atmosphere (Intergovernmental Panel on Climate Change, 1991). Forests can contribute to the reduction of greenhouse gases through carbon sequestration, but estimates of the size of the contribution can only be as reliable as the measurements upon which they are based (Birdsey et al., 1993). The amount of carbon fixed into biomass at a site is typically measured as the sum of above and belowground production of all the vegetation. Ecologists refer to these fixation processes as net primary production (NPP, kg C ha1 year1; Waring and Schlesinger, 1985. However, measurement of belowground production is difficult and uncertain, in part because the methodologies used to measure it are indirect and prone to errors (Singh et al., 1984; Nadelhoffer and Raich, 1992). Yet, accurate assessment of belowground production is important because this production can represent up to 76% of the annual total net primary production (Grier * Corresponding author. Tel.: +1-936-468-3301; fax: +1-936-4682489. E-mail address: [email protected] (D. Coble).
et al., 1981; Vogt, 1991; Nadelhoffer and Raich, 1992; Gower et al., 1995, 1996). The proportion of the production allocated belowground is often increased as belowground resources more strongly limit growth (Grier et al., 1981; Keyes and Grier, 1981; Running, 1983; Axelsson and Axelsson, 1986; Santantonio, 1989). Little research has addressed the question of aspect differences in allocation patterns. However, one might expect that the higher light intensity and lower soil moisture availability on south aspects might enhance belowground allocation. Whether the total production would be similar on north and south aspects is difficult to predict, but the longer growing season and higher light intensity on the south aspect lead some models to predict higher productivity (Wykoff et al., 1982; Milner et al., 1996). We assessed differences in allocation and total production at a study site in western Montana. The study sampled natural vegetation, both woody and nonwoody, on north and south aspects. We purposely sought out stands of high and low leaf area index (LAI) on both aspects. Although this is a case study, the results highlight aspect differences in production and allocation. The results of this case study lead to general hypotheses regarding the topographic distribution of carbon sequestration in high-latitude, mountainous terrain.
0269-7491/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved. PII: S0269-7491(01)00268-8
S150
D. Coble, J. Marshall / Environmental Pollution 116 (2002) S149–S155
2. Materials and methods 2.1. Objectives Using data from a case study at a site with varying aspect and varying leaf area index in western Montana, we examined the relationship between above and belowground carbon production. We tested the hypothesis that losses in aboveground productivity would be directly offset by increased carbon allocated belowground.
overstory tree density [low (L) and high (H) trees per hectare] and replicated twice for a total of eight plots (Table 1). Since aspect is not replicated at other locations, this project was considered a case study. Plots were square and 0.04 ha (0.1 acre) in size. On each plot, total annual aboveground biomass accumulation (AGBA) was measured by a combination of destructive sampling and dimensional analysis (Whittaker and Woodwell, 1968; Grier et al., 1989). Measurement procedures varied by type of vegetation. Measurements of the physical site included: slope ( 1%), elevation ( 1 m), and aspect ( 1 ).
2.2. Sampling design and site description 2.3. Vegetation measurements The study site comprised eight plots installed on opposing north and south aspects in a young (< 30 year old) forest located in west-central Montana at 1394 m (Coble et al., 2001). The location was chosen primarily because it had contrasting aspects, which created a contrasting moisture regime over a short distance (Coble, 1997). Plots were installed in close proximity on each aspect to minimize plot-to-plot variability (e.g. differences in soils, slope, etc.). The area was clear-cut harvested in 1976, planted in 1977, and precommercially thinned in 1980. However, the regeneration was predominantly natural because most of the planting stock died. The vegetative community on the south aspect consisted of patchy ponderosa pine (Pinus ponderosa Dougl. ex Laws.) and Douglas-fir [Pseudotsuga menziesii (Mirb.) Franco] in the overstory, and grasses (Agropyron spicatum, Calamagrostis rubescens), sedges (Carex spp.), snowberry (Symphoricarpos albus), and knapweed (Centaurea maculosa) in the understory. Vegetation on the north aspect consisted of mostly homogeneous western larch (Larix occidentalis Nutt.) and lodgepole pine (Pinus contorta Dougl. ex Loud.) with some Douglas-fir and ponderosa pine in the overstory, and grasses, sedges, snowberry, huckleberry (Vaccinium spp.), spirea (Spiraea betulifolia var. lucida), serviceberry (Amelanchier alnifolia), and various forbs in the understory. Four plots were located on each of the two aspects [north (N) and south (S)] in two different levels of
AGBA of the understory vegetation was estimated via clipping. Deciduous shrubs, forbs, grasses, and trees less than or equal to 3 years old were sub-sampled on four (for the north aspect) and eight (for the south aspect) subplots. The subplots were square, 4 m2 in area, and were randomly located on the plot. On each sub-plot, current year’s foliage and lateral stem growth (if applicable) of all non-tree individuals were clipped, combined by species, placed in bags, labelled, and transported to the laboratory for drying. Radial increment of shrubs was ignored (Alaback, 1986). The trees were clipped, grouped by 1-year age classes (years 1, 2, 3), placed in bags, labelled, and transported to the laboratory for drying. The foliage, branches, and stems were not separated because these trees were so small and rare (total dry weight less than 2 g). Foliage and stem samples from the sub-plot vegetation were oven-dried for 24 h at 70 C and weighed to the nearest 0.01 g (Alaback, 1986). Herbivory was not explicitly measured in this study because cattle were excluded from the study area and deer browsing was minimal, and we saw no evidence of significant insect herbivory. All trees greater than 3 years of age were tallied on each plot. Diameter ( 0.25 cm), height ( 0.1 m), and crown ratio ( 5% of live crown on the bole) were measured and recorded for all standing trees. Allometric equations were applied to this tally to estimate
Table 1 Site and density characteristics of four plots on each of two aspects Aspects
Plot
Aspect (degrees)
Slope (%)
Trees per hectare
Overstory tree LAIa (m2 m2)
Overstory density
North North North North South South South South
1 2 3 4 1 2 3 4
340 356 354 314 220 218 214 224
18 11 12 12 25 24 27 24
667.2 444.8 963.7 963.7 321.2 222.4 1532.1 1705.0
0.9 1.3 3.3 3.4 0.4 0.2 3.4 4.6
Low Low High High Low Low High High
a
Note: all-sided leaf area index (LAI) values were based on specific leaf area values not measured on the study site and should therefore be interpreted as rough estimates.
D. Coble, J. Marshall / Environmental Pollution 116 (2002) S149–S155
AGBA for trees greater than 3 years of age (procedure described in next section). 2.4. Aboveground data summarization The tree tally data were used with local allometric equations (Coble, 1997) to find the AGBA in leaf, branch, stem, and bark components for standing trees greater than 3 years old on each plot. The AGBA values for trees destructively sampled to build the allometric equations were added to the AGBA values of the tally trees. The AGBA values for all trees greater than 3 years old were then expanded to a per hectare basis for each plot. Dry-weight AGBA of the understory plants on each sub-plot was summed on a total, life form, and specieslevel basis to determine the change in understory biomass for the sub-plot. For trees less than or equal to 3 years old on the sub-plots, the total dry weight was divided by the age of the tree to determine its average AGBA, then added to the sub-plot AGBA totals. The AGBA values for each sub-plot were then averaged by plot to find the average plot-level understory AGBA. Mean AGBA values for each plot were then expanded to a per hectare basis. The dry weight AGBA values for tree and understory vegetation were multiplied by 0.5 to convert grams of dry weight to grams of carbon (Vogt, 1991). All-sided LAI (m2 m2) for overstory trees on each plot was calculated as the product of leaf carbon and specific leaf area (SLA, m2 kg C1). LAI was calculated for each tree and summed to obtain the plot-level allsided overstory tree LAI. Since SLA was not measured in this study, the following SLAs were used: Douglas-fir and western larch SLA=25.0 m2 kg C1 (Pierce et al., 1994), lodgepole pine and ponderosa pine=15.0 m2 kg C1 (Pierce et al., 1994). Plot-level LAI values were only used as covariates for overstory tree density in statistical analyses (see ANCOVA description later); they were not used to estimate or measure the amount of carbon. 2.5. Soil CO2 flux measurements Belowground CO2 efflux (soil respiration) was measured with a LiCor 6200 gas exchange analyzer with the soil chamber attachment (LiCor Inc., Lincoln, Nebraska, USA). The LiCor 6200 is a closed-system infrared gas analyzer (IRGA) that measures the amount of CO2 flux from the soil. Ten sampling points were randomly located in the buffer of each study plot for a total of 80 samples. The points were located in the plot buffers to avoid sampling in areas where earlier destructive sampling was conducted. Each sampling point was prepped by removing all vegetation inside of a 0.00856 m2 sampling collar imbedded 2 cm in the ground. Measurements were taken with the LiCor 6200
S151
twice a day (morning and evening) at seven different times throughout the sampling period. This sampling period approximately encompassed snowmelt/ground thaw in the spring until the first heavy snowfall in the fall (206 days for the south aspect and 188 days for the north aspect). Soil temperature ( C) at a 10 cm depth was also measured with each LiCor 6200 CO2 flux measurement. Soil CO2 fluxes (mmol m2 s1) at each sampling point were determined by taking four readings at 6–10 ppm differentials, then averaging the two flux values immediately above and below the average ambient CO2 value for the plot. The daily average soil CO2 flux for each sampling point was calculated as the average of the morning and evening flux measurements. 2.6. Litterfall measurement Litterfall was also measured adjacent to each sampling point. Litterfall (including branches and twigs) was collected at the end of the growing season in 0.00856 m2 traps (the same size as the soil collars), which were located in a random direction within 10 cm of the soil collar. Litterfall samples were bagged, transported to the lab, dried for 48 h at 70 C, then removed and weighed to the nearest 0.1 g. The dry mass was multiplied by 0.5 to convert grams of dry weight to grams of carbon (Vogt, 1991). 2.7. ANPP calculation Total aboveground net primary productivity (ANPP) for all vegetation on each plot was calculated as the sum of the component biomass weights (kg C) for all the life forms plus litterfall: ANPP ¼ Tðl; b; sÞ þ tðl; b; sÞ þ Sðl; sÞ þ GðlÞ þ FðlÞ þ D where ANPP=aboveground net primary production (kg C ha1 year1); T (l, b, s)=leaf, branch, and stem production of trees greater than 3 years old; t(l, b, s)=leaf, branch, and stem production of trees less than or equal to 3 years old; S(l, s)=leaf and stem production of shrubs; G(l)=leaf production of grasses; F(l)=leaf production of forbs; and D=detritus production or litterfall. 2.8. BGPP calculation A carbon budget method was used to estimate the upper limit for total annual carbon allocation to roots (Raich and Nadelhoffer, 1989). This technique is based on the conservation of mass; i.e. all carbon entering the soil must either exit the soil or enter the soil carbon pool. The method assumes that the soil carbon pool is in steady state on an annual basis; i.e. the annual amount of carbon entering the soil equals the annual amount
S152
D. Coble, J. Marshall / Environmental Pollution 116 (2002) S149–S155
of carbon leaving the soil. Under this assumption, heterotrophic respiration (Rh) is assumed to equal the annual amount of carbon allocated into the soil from above (Pa=litterfall) and belowground (Pb): Rh ffi Pa þ Pb
ð1Þ
Rh is often impossible to measure directly because it is mixed with respiration from the roots (Rr). Therefore, the method estimates annual carbon allocation to roots by subtracting litterfall from total soil respiration to estimate the sum of root respiration and root production: Rs Pa ffi Pb þ Rr
ð2Þ
We refer to this sum as belowground gross primary production (BGPP). The Raich and Nadelhoffer (1989) technique was used in this study to estimate the total carbon allocation to roots (BGPP) at each sampling point for each measurement period. Though the forest sampled in this study was young, it had reached canopy closure on the plots with high tree densities, which makes us believe that the site had reached, or was approaching, steady state. From the measured BGPP values, weighted averages (the weights were the number of days in each measurement period) were calculated across the seven measurement periods to find the seasonal average BGPP for each sampling point. These weighted averages were then averaged to find the seasonal average BGPP for each study plot. 2.9. Comparison of ANPP and BGPP Simple linear regression (SLR; Zar, 1999) was used to test two alternative hypotheses about the relationship between BGPP and ANPP. The two were related according to the following linear model: BGPP ¼ b0 þ b1 ANPP
hypotheses: Ho1: (0,1)=(0,2), Ha1: not Ho1, and Ho2: (0,1)=(0,2), Ha2: not Ho2. An analysis of covariance (ANCOVA) was also used to test if mean BGPP and ANPP were equal for the two levels of overstory density (the covariate=overstory tree LAI) and the two aspects (the class variable), and that the density–aspect interaction does not significantly affect production (Zar, 1999). The Fmax test was used to test for homogeneity among variances in production values between aspects prior to the ANCOVA (Sokal and Rolf, 1995). ANCOVA was employed to test for differences in carbon production rather than ANOVA or a Kruskal– Wallis test because overstory tree density was not an experimentally controlled variable in this study. Therefore, it could not be assumed that measured carbon production was affected equally for the two levels of overstory density on the north versus south aspect installations. Though overstory tree LAI was highly correlated with overstory tree density1 (R2=0.861), overstory tree LAI was used as the covariate rather than tree density (TPH; trees per hectare) in the ANCOVA. Tree LAI was chosen because it was considered to be a better measure of the competitive effect of tree density on site productivity than numbers of trees, simply because it represents photosynthetic capacity of the trees on a site (Waring and Schlesinger, 1985).
3. Results Belowground carbon allocation to roots (BGPP, kg C ha1 period1; 206 days on south aspect, 188 days on north aspect) increased with total ANPP (kg C ha1 year1; Fig. 1). However, allocation was not constant across plots because the actual slope value of the SLR model was less than 1 (P < 0.0001) and the intercept was significantly different from zero (P < 0.0001). Furthermore, the slope was also quite different from 2 (P < 0.0001), which leads us to conclude that root
ð3Þ
If allocation were constant across the plots, then the intercept (b0) would not be different from zero and the slope (b1) would be approximately two. (The value of two was chosen to compare BGPP and ANPP; since ANPP does not include respiration and respiration represents roughly half the amount of gross primary production, we needed to use the value of two in our hypothesis tests). If, however, BGPP were traded off directly against ANPP, and BGPP were constant, then the value of b0 should be equal to BGPP and the slope (b1) should equal approximately 2. If both BGPP and allocation were changing, then the result should be intermediate between these two extremes. Simultaneous F-tests (Neter et al., 1985) were used to evaluate the two
Fig. 1. Comparison of above (ANPP) and belowground (BGPP) primary production.
D. Coble, J. Marshall / Environmental Pollution 116 (2002) S149–S155
production was not directly traded off against aboveground production. Instead, both BGPP and allocation were apparently different among the plots. Given the positive slope of the relationship, it appears that variation in BGPP was quantitatively more important than reallocation. This should perhaps not be surprising considering that the plots varied in LAI from 1.7 to 5.4 (Fig. 2). ANPP was significantly (P=0.0031) greater on the north versus south aspects (Fig. 3). ANPP also significantly (P=0.0336) increased with tree LAI (Fig. 3). Also, there was no aspect/overstory tree LAI interaction nor any heterogeneity of variances (P < 0.05) for either ANPP or BGPP, so the assumptions of ANCOVA were not violated. Thus, ANPP varied among sites rather than simply re-routed between overstory trees and understory vegetation. Based on this result, we conclude that the reduction in ANPP on the south aspect was not a result of shifts in allocation to BGPP. We also conclude that production was not re-routed belowground on sites with low overstory tree density. Soil CO2 flux and litterfall, both of which were used to calculate BGPP, were greater for north versus south aspects (Figs. 4 and 5). This was an expected result
Fig. 2. Leaf area index (LAI) for tree and non-tree vegetation by aspect (N=North, S=South) and density (H=High, L=Low). Error bars represent one standard deviation. Note: all LAI values calculated from specific leaf areas not measured on the site.
Fig. 3. Above (ANPP) and belowground (BGPP) primary production by aspect (N=North, S=South) and density (H=High, L=Low). Error bars represent one standard deviation.
S153
because the north aspect plots contained more plant biomass, which in turn likely contributed to greater amounts of root respiration and litterfall than on the south aspect. However, the high variability in the flux measurements on both aspects (see error bars in Fig. 4) was not expected because 10 sampling points per plot was believed to be an adequate sample of soil CO2 flux. These results confirm that soil CO2 flux is a heterogeneous process that requires intensive sampling to accurately measure.
4. Discussion As mentioned in Section 1, methodologies used to measure belowground carbon production can be prone to errors. The Raich–Nadelhoffer model used to estimate BGPP in this study assumed that carbon inputs to the soil equaled carbon outputs from the soil. As stated earlier, we believe this is a reasonable assumption for this forest. We do not believe this would be the case had we sampled immediately after planting, but we do believe that 30 years is an adequate time for this site to reach a steady-state. The steady-state assumption is also questionable in the western United States where detritus (i.e. litter) inputs usually exceed outputs (Olson, 1963; Harvey et al., 1978), unless they are removed by fire (Wright and Heinselman, 1973). Since this study site was clearcut
Fig. 4. Soil CO2 flux by aspect (N=North, S=South) and density (H=High, L=Low). Error bars represent one standard deviation.
Fig. 5. Litterfall by aspect (N=North, S=South) and density (H=High, L=Low). Error bars represent one standard deviation.
S154
D. Coble, J. Marshall / Environmental Pollution 116 (2002) S149–S155
harvested in 1976, numerous large stumps and large downed logs remain from previous logging activity, though much of the smaller material had decayed. Coarse root increment was not measured for this study, so it was not added to the BGPP estimates (Ryan et al., 1996). Other studies found that coarse root production as a percent of total belowground production equaled 12–40% for Pinus radiata (Ryan et al., 1996), 6–15% for Abies amabilis (Grier et al., 1981, their Table 3), and 14–40% for Douglas-fir (Keyes and Grier, 1981). Therefore, the BGPP measurements based on the RN model alone were conservative, possibly upwards of 40 or 50%. BGPP is likely even more conservative because winter soil CO2 flux was not measured in this study. Considerable amounts of CO2 can be produced from the soil during the winter months under snow packs (Harvey et al., 1978; McDowell et al., 2000). McDowell et al. (2000) found that 17% of the total annual soil CO2 flux was through a snow pack, and the snow pack lasted 36% of the year. Coxson and Parkinson (1987) reported that soil CO2 lost during the winter was nearly 60% of the annual biomass input for an aspen forest. Dormaar et al. (1981) found a 50% loss during winter for a grassland. Thus, annual BGPP estimates at this study site could actually be twice the reported values; i.e. +50% for winter respiration and +50% for coarse roots and woody debris.
5. Conclusions We used conventional measurement techniques to estimate ANPP and BGPP at a site in western Montana. We then used these production estimates to test the hypothesis that BGPP and ANPP are related on a 2/ 1 basis. Since we found that ANPP and BGPP values were higher on the north versus south aspects and high versus low overstory tree density levels, we concluded that these correlated increases argue against differential allocation among life forms or between roots and shoots for this case study site. Acknowledgements This paper was presented at the USDA Forest Service Southern Global Change Program sponsored Advances in Terrestrial Ecosystem: Carbon Inventory, Measurements, and Monitoring Conference held 3–5 October 2000 in Raleigh, North Carolina.
References Alaback, P.B., 1986. Biomass regression equations for understory plants in coastal Alaska: effects of species and sampling design on estimates. NW Science 60 (2), 90–103.
Axelsson, E., Axelsson, B., 1986. Changes in carbon allocation in spruce and pine trees following irrigation and fertilization. Tree Physiology 2, 189–204. Birdsey, R.A., Plantinga, A.J., Heath, L.S., 1993. Past and prospective carbon storage in the United States. Forest Ecology and Management 58, 33–40. Coble, D.W., 1997. Trends in above and below ground production of trees and non-tree vegetation on contrasting aspects in western Montana. Unpublished PhD dissertation, School of Forestry, University of Montana, Missoula, Montana, USA. Coble, D.W., Milner, K.S., Marshall, J.D., 2001. Above and below ground production of trees and other vegetation on contrasting aspects in western Montana: a case study. Forest Ecology and Management 142, 231–241. Coxson, D.S., Parkinson, D., 1987. Winter respiratory activity in aspen woodland forest floor litter and soils. Soil Biology and Biochemistry 19 (1), 49–59. Dormaar, J.F., Smoliak, S., Johnston, A., 1981. Seasonal fluctuations of blue grama roots and chemical characteristics. Journal of Range Management 34, 62–64. Gower, S.T., Isebrands, J.G., Sheriff, D.W., 1995. Carbon allocation and accumulation in conifers. In: Smith, W.K., Hinckley, T.M. (Eds.), Resource Physiology of Conifers. Academic Press, San Diego, California, USA, pp. 217–254. Gower, S.T., Pongracic, S., Landsberg, J.J., 1996. A global trend in belowground carbon allocation: can we use the relationship at smaller scales? Ecology 77 (6), 1750–1755. Grier, C.C., Vogt, K.A., Keyes, M.B., Edmonds, R.L., 1981. Biomass distribution and above- and below-ground production in young and mature Abies amabilis zone ecosystems of the Washington Cascades. Canadian Journal of Forest Research 11, 155–167. Grier, C.C., Lee, K.M., Nadkarni, N.M., Klock, G.O., Edgerton, P.J., 1989. Productivity of forests of the United States and its relation to soil and site factors and management practices: a review (General Technical Report PNW-222). USDA Forest Service, Pacific Northwest Forest and Range Experiment Station, Portland, OR. Harvey, A.E., Jurgensen, M.F., Larsen, M.J., 1978. Seasonal distribution of ectomycorrhizae in a mature Douglas-fir/larch forest soil in western Montana. Forest Science 24, 203–208. Intergovernmental Panel on Climate Change, 1991. Climate Change— the IPCC Response Strategies. Island Press, Washington, DC. Keyes, M.B., Grier, C.C., 1981. Above and belowground net production in 40-year-old Douglas-fir stands on low and high productivity sites. Canadian Journal of Forest Research 11, 599–605. McDowell, N.G., Marshall, J.D., Hooker, T.D., Musselman, R., 2000. Estimating CO2 flux from snowpacks at three sites in the Rocky Mountains. Tree Physiology 20, 745–753. Milner, K.S., Running, S.W., Coble, D.W., 1996. A biophysical soil-site model for estimating potential productivity of forested landscapes. Canadian Journal of Forest Research 26, 1174– 1186. Nadelhoffer, K.J., Raich, J.W., 1992. Fine root production estimates and belowground carbon allocation in forest ecosystems. Ecology 73 (4), 1139–1147. Neter, J., Wasserman, W., Kutner, M.H., 1985. Applied Linear Statistical Models, 2nd Edition. R.D. Irwin, Inc, Homewood, IL. Olson, J.S., 1963. Energy storage and the balance of producers and decomposers in ecological systems. Ecology 44, 321–322. Pierce, L.L., Running, S.W., Walker, J., 1994. Regional-scale relationships of leaf area index to specific leaf area and leaf nitrogen content. Ecological Applications 4 (2), 313–321. Raich, J.W., Nadelhoffer, K.J., 1989. Belowground carbon allocation in forest ecosystems: global trends. Ecology 70 (5), 1346–1354. Running, S.W., 1983. A biophysical assessment of the regeneration and growth potential of forested land in Montana. In: O’Loughlin, J., Pfister, R.D. (Eds.), Management of Second-growth Forests: The State of Knowledge and Research Needs, 14 March 1983, Missoula,
D. Coble, J. Marshall / Environmental Pollution 116 (2002) S149–S155 Montana, USA. Montana Forest and Conservation Experiment Station, School of Forestry, University of Montana, Missoula, Montana, USA, pp. 58–75. Ryan, M.G., Hubbard, R.M., Pongracic, S., Raison, R.J., McMurtrie, R.E., 1996. Foliage, fine-root, woody-tissue and stand respiration in Pinus radiata in relation to nitrogen status. Tree Physiology 16, 333–343. Santantonio, D., 1989. Dry-matter partitioning and fine-root production in forests—new approaches to a difficult problem. In: Periera, J.S., Landsberg, J.J. (Eds.), Biomass Production by Fast-Growing Trees. Kluwer Academic Publishers, Dordrecht, Netherlands, pp. 57–72. Singh, J.S., Lauenroth, W.K., Hunt, H.W., Swift, D.M., 1984. Bias and random errors in estimators of net root production: a simulation approach. Ecology 65, 1760–1764. Sokal, R.R., Rolf, F.J., 1995. Biometry, 3rd Edition. W.H. Freeman and Company, New York.
S155
Vogt, K.A., 1991. Carbon budgets of temperate forest ecosystems. Tree Physiology 9, 69–86. Waring, R.H., Schlesinger, W.H., 1985. Forest Ecosystems: Concepts and Management. Academic Press, Orlando, FL. Whittaker, R.H., Woodwell, G.M., 1968. Dimension and production relations of trees and shrubs in the Brookhaven Forest, New York. Journal of Ecology 56, 1–25. Wright, H.E., Heinselman, M.L., 1973. The ecological role of fire in natural conifer forests of western and northern North America: introduction. Quaternary Research 3, 319–328. Wykoff, W.R., Crookston, N.L., Stage, A.R., 1982. User’s Guide to the Stand Prognosis Model. General Technical Report INT-133. USDA Forest Service, Intermountain Forest and Range Experiment Station, Ogden, UT. Zar, J.H., 1999. Biostatistical Analysis, 4th Edition. Prentice-Hall, NJ.