SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Homann, P. S.
Right arrow Articles by Bormann, B. T.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Homann, P. S.
Right arrow Articles by Bormann, B. T.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Homann, P. S.
Right arrow Articles by Bormann, B. T.
Related Collections
Right arrow Soil Organic Matter
Right arrow Carbon Sequestration
Right arrow Frozen Soils
Published in Soil Sci. Soc. Am. J. 68:2023-2030 (2004).
© 2004 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA

Division S-7—Forest & Range Soils

Carbon Storage in Coarse and Fine Fractions of Pacific Northwest Old-Growth Forest Soils

P. S. Homanna,*, S. M. Remillardb, M. E. Harmonc and B. T. Bormannb

a Dep. of Environmental Sciences, Huxley College of the Environment, Western Washington University, Bellingham WA, 98225-9181
b USDA Forest Service Pacific Northwest Research Station, Corvallis OR, 97331
c Department of Forest Science, Oregon State University, Corvallis OR, 97331

* Corresponding author (homann{at}cc.wwu.edu)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Many assessments of soil C have been restricted to the <2-mm fraction, but C has recently been identified in >2-mm fractions of forest mineral soils. Our objective was to determine the importance of the >2-mm fraction to whole-soil C pools in Pacific Northwest old-growth coniferous forests. Seventy-nine pedons in 18 western Washington and Oregon forests were sampled to a depth of 100 cm. The <2-mm fraction was separated from the >2-mm fraction by air-drying, physically crushing soil, and sieving; C was determined by Leco combustion. The >2-mm fraction contained up to 46% of the whole-soil C and averaged 23% for the seven forests that had C in that fraction. Following treatment with sodium hexametaphosphate to disaggregate soil material, up to 20% of whole-soil C remained in the >2-mm fraction. Thus, the >2-mm fraction C appears to be in stable and unstable aggregates, as well as concretions. The whole-soil C in the surface 100 cm of mineral soil ranged from 30 to 400 Mg C ha–1. Multiple regression analysis indicated this C pool was positively related to available water capacity, annual precipitation, and coarse woody debris (r2 = 0.63 to 0.66, n = 18 forests). Similar results were obtained with only the <2-mm soil C, which is the basis of previous regional evaluations. This suggests consideration of the >2-mm fraction does not alter our understanding of the importance of climate and soil texture as controls of soil C pools, but it does affect the quantification of soil C pools in many old-growth forests in the Pacific Northwest.

Abbreviations: LOI, loss on ignition


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
SOIL IS A MAJOR POOL in regional budgets and the global cycles of C (Jobbágy and Jackson, 2000; Zinke and Stangenberger, 2000; Bernoux et al., 2002; Goodale et al., 2002). Soil C content varies considerably within landscapes and between regions. It is influenced by climate, soil texture, topography, vegetation, and land management (Homann et al., 1995; Jobbágy and Jackson, 2000; Parker et al., 2001). There is considerable uncertainty associated with estimates of soil C pools and their association with controlling variables, because of sampling and analytical difficulties, soil spatial variability, and methods of extrapolation to regions (Homann et al., 1995, 1998, 2001; Brejda et al., 2001).

Most studies of soil C pools have measured only fine fractions of the soil, such as <2, <3, or <4 mm (Grigal and Ohmann, 1992; Canary et al., 2000; Parker et al., 2001; Brejda et al., 2001; Homann et al., 2001). Recent investigations of forests indicate the coarse fractions can also contain substantial C, and consideration of only the fine fraction would substantially underestimate the whole-soil C pool. For example, the <2-mm fraction contains only 63% of whole-soil C in a coastal Oregon forest (Cromack et al., 1999), 53 to 80% in three Australian forest sites (Bauhus et al., 2002), and 37% in one western Washington forest but 97% in another (Harrison et al., 2003).

Some >2-mm C may be associated with charcoal or rocks (Bauhus et al., 2002; Harrison et al., 2003). Alternatively, it may be contained in aggregates that are not broken down during routine sample processing. It is not clear to what extent the coarse fractions are physically stable, or if they readily reduce to <2-mm material under standard soil preparation procedures (USDA-NRCS, 1996).

Quantitative empirical relations between soil C and environmental controls are critical in parameterizing and testing dynamic models, and in estimating deeper soil C pools in areas where only the surface soil C has been measured. In forests of the Pacific Northwest, USA, Homann et al. (1995) found soil C increased with annual temperature, annual precipitation, actual evapotranspiration, clay, and available water-holding capacity, and decreased with slope. In a global assessment, Jobbágy and Jackson (2000) found the vertical distribution of soil organic C to be related to climate and vegetation type, with forests having a higher proportion of soil C concentrated nearer the surface. In assessments like these, data are compiled from various sources (Homann et al., 1995; Jobbágy and Jackson, 2000), which may have used a variety soil preparation and analytical methods (Homann et al., 1995). It is not clear if the results from these assessments are representative of whole-soil C, or only some fraction of it.

Our objectives were (i) to determine the extent to which the >2-mm fraction contributes to whole mineral-soil C pools in old-growth forests of the Pacific Northwest, (ii) to determine the stability of the >2-mm fraction, and (iii) to determine how the >2-mm C affects our interpretation of the relation of soil C pools to environmental controls. We also provide forest floor C pools to complement the mineral soil information.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Site Description
Soil samples were collected from one to eight pedons in each of 18 forests distributed over five physiographic provinces in western Washington and Oregon (Table 1). The pedons were associated with previously established plots, which differ in number among the forests. The specific stands sampled within the forests are associated with the Andrews Long Term Ecological Research (LTER) program and its network of permanent plots designed to observe and monitor changes in composition, structure, and functions of forest ecosystems over long time periods. A description of the history and characteristics of the network is provided in Acker et al. (1998). Mature to old-growth coniferous trees dominate each forest (Smithwick et al., 2002). Mean annual temperature ranges from 5.6 to 10.4°C, and mean annual precipitation from 40 to 370 cm (Table 1).


View this table:
[in this window]
[in a new window]
  		Table 1. Characteristics of Pacific Northwest old-growth forests.

 
Forest Floor Pools
Forest floor samples were taken at five points above the sampling face of a pedon with a vertically oriented 5-cm diameter corer. The samples were composited for each pedon. The samples included woody debris <1 cm in diameter on the surface and all material within the forest floor that was not visible from the surface, including all woody debris irrespective of size.

Forest floor samples were oven dried (70°C), weighed, coarsely ground in a blender (Braun AG, Frankfurt, West Germany), and finely ground (<850 µm, <20 mesh) in a mill (IKA-A 10, Staufen, Germany). Samples were randomized before analysis of total C and N with a LECO CNS 2000 analyzer (Leco Corp., St. Joseph, MI).

The forest floor C pool (kg C m–2 dry mass) was determined as

[1]
where C is the C concentration (g C kg–1 oven-dried mass), M is oven-dried sample mass (g), A is total cross-sectional area of the five subsamples (cm2), and 10 is a units-conversion factor.

Mineral Soil Pools
Three mineral soil layers were sampled (0–20, 20–50, and 50–100 cm). Depth strata, as opposed to horizon, sampling was chosen because it is more repeatable and comparable with other studies that have assessed C pools. Mineral soil samples were obtained by collecting material in three swaths, 5- to 10-mm deep, across the pedon face within each layer. For the few pedon faces that were too rocky to make swaths, samples were collected as part of bulk density sampling.

The samples were air-dried and then processed with 2- and 4-mm sieves. Sieves were repeatedly shaken by hand and material was broken up with a rubber stopper repeatedly run over the surfaces of the sieves. Based on observed morphology and color, the resulting 2- to 4- and >4-mm fractions were sorted by hand into 2- to 4-mm C-bearing soil fraction, >4-mm C-bearing soil fraction, >2-mm rock (assumed non-C bearing), and >2-mm buried wood, roots, charcoal. The 2- to 4-mm C-bearing and >4-mm C-bearing soil fractions appeared to be spherical concretions or aggregates of the <2-mm material. Each component was weighed. The >2-mm buried wood, roots, and charcoal accounted for <3% of the sample masses, and were not considered as part of the soil C pools. Carbon bearing fractions >2 mm were analyzed for C only if they were >10%, by weight, of the whole sample. Otherwise, the weight of the >2-mm C-bearing fraction was incorporated into the rock mass used to estimate soil volume (see below).

Subsamples (50–100 g) of <2-, 2- to 4-, and >4-mm C-bearing fractions were obtained with a sample splitter (CL-280 series, SoilTest, Loveland, CO). These subsamples were ground to 850 µm (<20 mesh) using a 20-cm disc pulverizer (BICO Inc., Burbank, CA), dried at 60°C, and analyzed for total C and N concentration using a LECO CNS 2000 analyzer. Concentrations were converted to a 105°C basis based on moisture determination of subsamples.

Bulk density methods varied with soil properties. For non-rocky soils, a 5-cm diam. by 5-cm deep soil-core bulk-density sampler with sampling ring inserts (AMS Inc., American Falls, ID) or a 5-cm diam. tube was used. Three cores were taken at various depths within each layer and composited by layer, resulting in one bulk-density sample for each of the three layers per pedon. For rocky pit faces, a cube was cut (approximately 20 by 10 cm by layer depth), the material excavated, and dimensions of the hole measured to determine excavated volume. The excavated material was processed in the field with a 20-mm sieve; the <20- and >20-mm fractions were weighed and subsampled for moisture content and further sorting and analysis. The estimated excavated volume may deviate from actual excavated volume by as much as 15%, based on an independent evaluation of 71 small excavated pits (P.S. Homann and B.T. Bormann, 2004, unpublished data).

Each air-dried bulk density sample was sieved and hand-sorted into the same components as the samples for C and N analysis, dried to 105°C, and weighed. Bulk density (g cm–3) of C-bearing material, Dc, was calculated as

[2]
where Mc (g) is the oven-dried mass of all C-bearing material; V (cm3) is the sample volume; M2–75 (g) is the mass of 2- to 75-mm rocks in the sample; and D2–75 (g cm–3) is the particle density of 2- to 75-mm rocks, which was assumed to be 2.65 g cm–3 for all pedons except the pumice at Pringle Falls which was assumed to be 2.1 g cm–3 (Flint and Childs, 1984). Because of their small contribution, buried wood, roots, and charcoal were omitted from the calculation.

The fractional volume of all C-bearing material, Vc, was calculated by rearranging the volume-estimate equation of USDA-NRCS (1996) to yield

[3]
where V>75 is the fractional rock volume of >75-mm rocks. The V>75 was estimated in the field by examination of the pit face and making a visual estimate; or in cases where large rocks were abundant, the material extracted from the pit was separated into <75- and >75-mm fractions, and the relative amount of each fraction was estimated visually for each layer. Other variables are as defined in Eq. [2].

The mineral soil C pool (kg C m–2) for each of three soil layers (0–20, 20–50, 50–100 cm) and each of three size classes (<2, 2–4, >4mm) was determined as

[4]
where Ci is the total C concentration (g C kg–1 dry mass) of size class i; i refers to the <2-, 2- to 4-, and >4-mm C-bearing size classes; Mi is the oven-dried mass (g sample–1) of material in size class i; T is layer thickness (cm), and 10 is a units conversion factor. Other variables are as defined in Eq. [2] and [3]. Whole mineral-soil C pools were obtained by summing the C pools of the three size classes.

Stability of 2- to 4-mm Fraction
The air-dried C-bearing 2- to 4-mm fraction from five samples from each province was treated with sodium hexametaphosphate solution [35.7 g (NaPO3)6 and 7.94 g Na2CO3 L–1; USDA-NRCS, 1996] by two different methods. In the first method, a subsample was added to sodium hexametaphosphate solution, swirled for 1 min every 15 min for 1 h, allowed to set for 10 h, and swirled for 1 min every 15 min for an additional 1 h. In the second method, a subsample was added to sodium hexametaphosphate solution and shaken constantly for 12 h on an orbital shaker. After either treatment, the material was placed on a 2-mm sieve, rinsed with running water for 15 s, and air-dried. Before and after treatment, the 2- to 4-mm material was weighed, and analyzed for moisture content (105°C) and for loss-on-ignition (LOI) by heating at 450°C for 12 h. The LOI values and pretreatment LOI/C ratios were used to convert all values to C masses.

Statistical Methods
Comparisons of soil C pools among forests were made with a single-factor ANOVA, in which the forest was the factor and the pedon was the observational unit. Before analysis, soil C values were log-transformed to reduce heterogeneity of variance among forests. The influence of treatment on the stability of the 2- to 4-mm fraction was determined with a randomized block ANOVA, in which the soil sample was the block and the treatment was the factor. The least significant difference (LSD) test (P < 0.05) was used to separate means.

The relation of mineral soil C pools to site factors was determined with step-wise regression analysis (Homann et al., 1995). To avoid pseudoreplication, the average values for each of the 18 forests were used. A separate analysis was performed for each of four dependent variables: <2-mm C pool in 0- to 20-cm mineral soil, whole-soil C pool in 0- to 20-cm mineral soil, <2-mm C pool in 0- to 100-cm mineral soil, and whole-soil C pool in 0- to 100-cm mineral soil. These dependent variables were log-transformed to reduce heterogeneity of variance. The potential independent variables were potential evapotranspiration; actual evapotranspiration; mean annual precipitation; mean annual temperature; masses of sand, silt, and clay to 20- and 100-cm depths; available water holding capacities to 20- and 100-cm depths; and slope (Homann et al., 1995); as well as stand age, and masses of fine and coarse woody debris (Smithwick et al., 2002). To avoid multicolinearity, highly correlated (r > 0.7) independent variables were not included in the same regression model. Variables in final models were accepted at P < 0.05.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Whole-Soil C
Forests with relatively small forest floor C pools (approximately 10 Mg C ha–1) occurred in all five provinces, while forests with high amounts were found only in Oregon Coast, Washington Cascades, and Oregon Cascades (Fig. 1) . The magnitude and variability of forest floor C pools is consistent with the assessment of western Oregon forests, including second-growth stands, in which forest floor C pools ranged from 1 to 110 Mg C ha–1 and averaged 20 Mg C ha–1 (Homann et al., 1995).



View larger version (36K):
[in this window]
[in a new window]
  		Fig. 1. Forest floor C pools of old-growth forests of the Pacific Northwest. Each bar represents a forest and is the average of 1 to 8 pedons (see Table 2). Pooled within-forest coefficient of variation (CV) is 77%. With n = 4 pedons per forest, a ratio of a higher mean to a lower mean >2.8 indicates two forests are significantly different, based on ANOVA of log-transformed values followed by the LSD test ({alpha} = 0.05); exact cutoff value of ratios differs among pairs of forests because of different numbers of pedons at different forests.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Average whole mineral-soil characteristics of Pacific Northwest old-growth forests.

 
The C pool sizes of mineral-soil (Fig. 2) were more consistent within provinces than those of the forest floor (Fig. 1). Whole mineral-soil C pools in the upper 100 cm of soil ranged from 30 to 400 Mg C ha–1 (Fig. 2). The values were within the range of 23 to 880 Mg C ha–1 for western Oregon pedons (Homann et al., 1995), even though many of those assessments were for C in <2-mm fraction rather than both <2- and >2-mm fractions, as measured in this study.



View larger version (37K):
[in this window]
[in a new window]
  		Fig. 2. Whole mineral-soil C pools of old-growth forests of the Pacific Northwest. Each bar represents a forest and is the average of 1 to 8 pedons (see Table 2). Pooled within-forest CV is 34%. With n = 4 pedons per forest, a ratio of a higher mean to a lower mean >1.5 indicates two forests are significantly different, based on ANOVA of log-transformed values followed by the LSD test ({alpha} = 0.05); exact cut-off values of ratios differ among pairs of forests because of different numbers of pedons at different forests.

 
The Oregon coast had the largest mineral-soil C pools and Oregon eastside had the lowest amounts (Fig. 2). This pattern is consistent with the spatial distribution of soil C based on the state soil geographic database (Homann et al., 1998). Variation in C concentration, bulk density, and rock volume all contributed to the differences in C pools among forests (Table 2). For example, Washington coast forest A had higher C concentrations that more than offset low bulk density and high rock volume, to yield a higher mineral-soil C pool than other Washington coast forests. In contrast, low bulk density and high rock volume of Oregon Cascades forest L resulted in a low mineral-soil C pool in spite of C concentrations comparable to other Oregon Cascade forests (Table 2).

The nature of the C in soil samples likely depends on both natural processes and analytical procedures. We assume that most of the C in our study is organic. The high precipitation in most forests (Table 1) and the volcanic and glacial parent material (Heilman et al., 1981) make it unlikely the C is from carbonate, although we did not specifically test for this. Furthermore, small pieces (<2 mm) of charcoal may have contributed C in our samples, although large identifiable pieces of charcoal had been removed. Finally, we did not make C measurements of the material we classified as >2-mm rocks, although organic C may be present in sedimentary rocks (Des Marais et al., 1992), and rock has been mentioned as a possible pool of organic C in forest soils (Bauhus et al., 2002; Harrison et al., 2003).

Carbon in >2-mm Fractions
Seven of the 18 forests had C in the >2-mm mineral-soil fractions (Fig. 3) . In the 0- to 20-cm layer, the >2-mm fractions contained as much as 46% of the whole-soil C, as observed in Oregon Cascade forest P. In general, the >2-mm fractions tended to contain lower proportions of whole-soil C at greater depths (Fig. 3). The occurrence and magnitude of >2-mm C-containing fractions in more than a third of the forests evaluated in this study supports the conclusion of Cromack et al. (1999) that these fractions can contribute substantially to soil C pools. They found 37% of soil C to be in >2-mm fractions. Harrison et al. (2003) found 63% of whole-soil C in >2-mm fractions at one western Washington site, but only 3% at another site. Bauhus et al. (2002) determined 20 to 47% of whole-soil C was in >2-mm fractions in three Australian forests.



View larger version (64K):
[in this window]
[in a new window]
  		Fig. 3. Proportion of whole mineral-soil C in size fractions of Pacific Northwest old-growth forests. Only forest provinces with C in >2-mm fractions are shown. Each bar represents 1 to 8 pedons within a forest (see Table 2). Pooled within-cell (forest–depth combination) CV of <2-mm proportion is 15%.

 
In our study, much of the C in the >2-mm fractions appears to have been in aggregates or concretions that had composition similar to <2-mm material. This conclusion is supported by the sorting procedures and chemical characteristics of the fractions. The concentrations of C in the 2- to 4- and >4-mm C-bearing material were highly positively correlated with those in the <2-mm fractions (r > 0.95); similarly the concentrations of N in the 2- to 4- and >4-mm C-bearing material were highly positively correlated with those in the <2-mm fractions (r > 0.95). The C and N concentrations in the 2- to 4-mm C-bearing material averaged 15 to 16% less than in the <2-mm fraction (P < 0.0001, paired t tests), while there were no differences in concentrations between the <2 and >4-mm C-bearing material (P > 0.4, paired t tests).

The USDA-NRCS (1996) has developed detailed standard procedures to separate coarse fragment (>2-mm) fractions from the <2-mm fraction. The procedures differ among soil types. Soils without coarse fragments (>2 mm) are crushed in a laboratory jaw crusher. Clayey soils with many coarse fragments are sieved (2 mm) in a mechanical shaker. Soils with easily crushable coarse fragments are crushed with a rubber roller, and other soils are crushed with a wooden rolling pin, to pass a 2-mm sieve; the soils are rolled and sieved until only the coarse fragments that do not slake in a sodium metaphosphate solution remain on the sieve. The <2-mm fraction is saved for chemical analyses, while the >2-mm fractions are discarded following measurement of their masses. Hence, the implicit premise is that the >2-mm fractions are not chemically important for purposes undertaken by USDA-NRCS, such as soil classification and fertility assessments.

Our assessment indicated that the >2-mm fraction could contain a substantial amount of C, even following exposure to sodium hexametaphosphate to disaggregate soil materials (Fig. 4) . Hence, the applicability of procedures designed for one purpose, for example, standard procedures to separate fine and coarse fractions, may bias results of other uses, for example, the estimation of total soil C. The importance of C in the >2-mm fraction varied among provinces. For Washington coast samples, the contribution of the 2- to 4-mm fraction to soil C became progressively lower with harsher treatment, while there was no effect of this treatment on the Washington Cascade samples (Fig. 4). Approximately 5 to 20% of the <4-mm soil C remained in the 2- to 4-mm fraction even with the severe treatment of shaking in sodium hexametaphosphate. These results indicate all size fractions need to be assessed for C (Harrison et al., 2003), irrespective of whether or not disaggregation procedures are used to separate the size fractions. Alternatively, the whole soil can be measured without fractionation or sieving (Binkley et al., 1992; Gerzabek et al., 2001).



View larger version (55K):
[in this window]
[in a new window]
  		Fig. 4. Proportion of <4-mm C in size fractions of Pacific Northwest old-growth forest soils, as affected two different treatments: periodic swirling with sodium hexametaphosphate for 12 h, or constant shaking with sodium hexametaphosphate for 12 h. Each value is based on five soil samples that have C in both the <2 and 2- to 4-mm fractions. The LSD bars are based on randomized block ANOVA (soil is block) for each province, followed by LSD test ({alpha} = 0.05).

 
Other studies that have identified C in >2-mm fractions have used a variety of soil preparation methods. Cromack et al. (1999) separated <2-mm from >2-mm fractions by air-drying (40°C) and sieving for 2 min on a motor-driven shaker sieve. Bauhus et al. (2002) obtained >2-mm fractions of air-dried soil that "did not disintegrate easily from the hand-applied pressure during sieving." Harrison et al. (2003) air-dried samples to constant weight and then screened to 2 mm. Many studies of forest soil C, however, do not detail soil preparation and sieving procedures, other than to indicate analyses were made on <2-, <3-, or <4-mm fractions (Grigal and Ohmann, 1992; Canary et al., 2000; Parker et al., 2001; Homann et al., 2001). Hence, it is not clear how much effort was made to physically breakdown aggregates into smaller units that would pass through sieves and, hence, be counted in estimates of soil C. Further, it is not clear to how much soil C was missed by not measuring the coarser fractions.

The elucidation of the contribution of various soil physical fractions to C and nutrient dynamics is an ongoing challenge (Christensen, 1992). Advances in understanding might occur with long-term laboratory incubations of separated and recombined fractions, such as those performed with light and heavy density fractions of Pacific Northwest second-growth forest soils (Swanston et al., 2002, 2004).

Regional Controls on Soil Carbon
Regional evaluations can provide estimates of horizontal and vertical spatial patterns of soil C (Homann et al., 1998; Jobbágy and Jackson, 2000). The spatial patterns provide a basis to estimate extant soil C pools over large areas, which in turn, serve as a foundation for potential increases or decreases in response to pedogenesis, land use, and climate change (Sollins et al., 1983; Burke et al., 1989; Jenkinson et al., 1991; Bouwman and Leemans, 1995; Liski and Westman, 1997a). Regional evaluations also establish relations between soil C and possible controlling factors, such as climate and soil texture (Homann et al., 1995), which can be used in formulating and testing simulation models (Schimel et al., 1996; Thompson et al., 1996; Den Elzen et al., 1997; Post et al., 1997).

Consideration of soil C in >2-mm fractions did not influence our qualitative understanding of regional soil C distribution or its controls, but quantitative relations were influenced slightly. The vertical distribution of soil C was similar for the <2-mm fraction and the whole soil (Table 3). For the <2-mm fraction, the percentage of the 0- to 100-cm C in the 0- to 20-cm layer averaged 41%. For the whole soil, which included the >2-mm fractions, this percentage was only slightly higher at 42%. These values are in agreement with an average of 42% for western Oregon forests (Homann et al., 1995). However, the percentages are substantially lower than the average of 50% for forests around the globe (Jobbágy and Jackson, 2000). Relations of vertical soil C distribution to climate exacerbate the discrepancy. In an analysis of global forest soils (Jobbágy and Jackson, 2000), the average percentage in the 0- to 20-cm layer increased from 45 to 55% as average annual temperature decreased from 25 to 5°C. Furthermore, in low temperature (0–10°C) forests, the average percentage contributed by the 0- to 20-cm layer increased from 52 to 58% as annual precipitation increased to >1000 mm. Based on these trends, and the low annual temperatures and high precipitation of Pacific Northwest forests (Table 1), an average value >50% would be expected, in contrast to our observed average of 42%. The causes of this discrepancy between Pacific Northwest forests and relationships developed from global forests are unclear, but they could include seasonality of climate patterns, topographic influences, and soil textural effects.


View this table:
[in this window]
[in a new window]
  		Table 3. Proportion of the 0- to 100-cm mineral-soil C [mean (s.d.)] in soil layers of Pacific Northwest USA old-growth forests compared with global forests.

 
The inclusion of >2-mm mineral-soil C did not qualitatively influence the relation of soil C to climate and site variables. The best stepwise regression equations for the whole-soil C and the <2-mm C were quite similar (Table 4), although the regression coefficients differed slightly. Available water capacity was a significant independent variable in accounting for the variation in soil C in both the 0- to 20- and 0- to 100-cm layers (Table 4). In combination with mean annual temperature, coarse woody debris mass, or precipitation, up to 84% of the soil C variation was explained.


View this table:
[in this window]
[in a new window]
  		Table 4. Stepwise-regression relations of mineral soil C pools to site variables in Pacific Northwest old-growth forests (n = 18 forests).{dagger}

 
These results are consistent with processes of soil organic matter production, stabilization, and decomposition (Sollins et al., 1996). The positive relation of soil C to available water capacity has also been observed in Lake States forests (Grigal and Ohmann, 1992), while the positive relation to precipitation has been found in a variety of forests and grasslands (Nichols, 1984; Sims and Neilsen, 1986; Burke et al., 1989; Grigal and Ohmann, 1992; Conant et al., 1998). Both precipitation and available water capacity are important for supplying water for primary production; in Pacific Northwest forests water storage is important because of potential water limitation during the dry summers (Waring and Franklin, 1979). Some of this production eventually becomes above- and belowground detritus, which contributes to soil C following decomposition. Available water capacity also reflects soil texture, which is important because of clay and silt stabilization of soil C (Sollins et al., 1996).

The negative relation of soil C to temperature is opposite that found in western Oregon forests (Homann et al., 1995). One possible explanation is that the relation found in this study is a statistical error, given the high P values of 0.04 for <2-mm C and 0.06 for whole-soil C. Alternatively, the interaction of temperature with other independent variables may cause different results; available water capacity was coupled with temperature in this study (Table 4), while precipitation and clay were the other independent variables coupled with temperature for western Oregon (Homann et al., 1995).

The positive relation of soil C to coarse woody debris may occur from direct causation or through indirect processes. The role of coarse woody debris in providing organic matter to the mineral soil is poorly understood. Coarse woody debris might provide direct inputs of organic matter to the mineral soil via leaching or embedding of wood into soil at the time of tree or snag fall. Alternatively, mechanisms of organic matter production and removal, such as by fire, might influence coarse woody debris and soil organic matter simultaneously.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Substantial soil C was present in the >2-mm fractions of one third of the Pacific Northwest old-growth forests we examined. The >2-mm fractions contained up to 46% of whole-soil C. The separation of C between <2- and >2-mm fractions is likely influenced by soil-preparation procedures; however, up to 20% of whole-soil C remained in the >2-mm fraction even after treatment with sodium hexametaphosphate, a standard NRCS procedure to disaggregate soil materials. Therefore, determination of soil C pools should include the >2-mm fraction, at least in initial assessments to determine the magnitude of that fraction at individual sites. In spite of the importance of C in the >2-mm fraction in some soils, consideration of this C does not substantially affect our understanding of the vertical distribution or the controls on soil C on a regional basis. Climate and soil texture variables continue to be important in explaining the spatial variation in soil C. Additionally, coarse woody debris may influence soil C in these old-growth forests.


   ACKNOWLEDGMENTS
 
This research was made possible by funding from the following sources: H.J. Andrews Long Term Ecological Research (LTER) program, under NSF grant number DEB-9632921; NASA grant NAG5-6242; Pacific Northwest Research Station Long-Term Ecosystem Productivity Program, Corvallis, OR; Interagency Agreement DW 12936179 between USEPA and the Pacific Northwest Research Station; Bureau for Faculty Research, Western Washington University. We thank Jerry Franklin and Steve Acker for old-growth plot establishment and oversight.

Received for publication January 5, 2004.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Soil Sci.Home page
H. S. Jones, L. G. Garrett, P. N. Beets, M. O. Kimberley, and G. R. Oliver
Impacts of Harvest Residue Management on Soil Carbon Stocks in a Plantation Forest
Soil Sci. Soc. Am. J., September 30, 2008; 72(6): 1621 - 1627.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Homann, P. S.
Right arrow Articles by Bormann, B. T.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Homann, P. S.
Right arrow Articles by Bormann, B. T.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Homann, P. S.
Right arrow Articles by Bormann, B. T.
Related Collections
Right arrow Soil Organic Matter
Right arrow Carbon Sequestration
Right arrow Frozen Soils

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome