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


     

Published online 8 June 2007
Published in Soil Sci Soc Am J 71:1215-1224 (2007)
DOI: 10.2136/sssaj2006.0015
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
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 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 Google Scholar
Google Scholar
Right arrow Articles by Zinn, Y. L.
Right arrow Articles by Resck, D. V. S.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Zinn, Y. L.
Right arrow Articles by Resck, D. V. S.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Zinn, Y. L.
Right arrow Articles by Resck, D. V. S.
Related Collections
Right arrow Soil Organic Matter
Right arrow Carbon Sequestration
Right arrow Tropical Soil Management

FOREST, RANGE & WILDLAND SOILS

Edaphic Controls on Soil Organic Carbon Retention in the Brazilian Cerrado: Soil Structure

Yuri L. Zinna,*, Rattan Lala, Jerry M. Bighama and Dimas V. S. Resckb

a School of Natural Resources, Ohio State Univ., 2021 Coffey Rd., Columbus, OH 43210-1085
b Embrapa Cerrados Agric. Research Center, P.O. Box 08223, 73310-970, Planaltina-DF, Brazil

* Corresponding author (zinn.14{at}osu.edu).


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Soil structure can be an important factor affecting soil organic carbon (SOC), but it is also a dynamic property affected by texture, mineralogy, land use, soil fauna, and also SOC. Assuming that structure affects SOC mostly by occlusion of particulate organic matter (POM) within aggregates, it was hypothesized that structure exerts a major control on SOC retention in soils of the Brazilian Cerrado region. Water-stable aggregates (WSA) were obtained from the 1-m depth of three different-textured, uncultivated soils. The mean weight diameter (MWD) of WSA was used as a structural indicator, and SOC concentrations were determined in intact WSA and their respective sand fractions (estimating occluded POM). Clay + silt content in bulk soils was correlated with MWD in all depths but more strongly in the top 10 cm. Although equally correlated with clay + silt contents, SOC concentrations were well correlated with MWD only in the 0- to 5-cm layer. Sand-free SOC concentrations in WSA fractions were proportional to sand content, indicating that the SOC dilution effect reported in particle size fractionations occurs naturally in the soil fabric. Occlusion of POM within aggregates was proportional to clay + silt contents, but this did not result in larger total POM pools, and the weak correlations obtained did not warrant predictive models. Aggregates produced by macrofauna comprised a minor but significant part of the soil and were mostly SOC enriched. We concluded that the structural control on SOC retention is less significant than the textural and mineralogical controls, since aggregation depends on those properties and is not as strongly correlated with SOC concentration and POM occlusion.

Abbreviations: MWD, mean weight diameter • POM, particulate organic matter • SOC, soil organic carbon • WSA, water-stable aggregates


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The intricate relationship between soil structure and SOC (e.g., Jenny, 1941; Carter, 1996) is a recurrent topic in soil science. Soil structure can be considered an edaphic control on SOC retention, such as soil texture and mineralogy (Zinn et al., 2007). The literature on this structural control is scanty for subsoil layers, often contradictory (Feller et al., 1996), and mostly based on nonstandardized research methods (Ashman et al., 2003; Loveland and Webb, 2003). The study of the structural control is conceptually complex because of the presumed feedback relation between aggregation and SOC, which does not occur for the textural and mineralogical controls. Additionally, soil structure (like SOC) results from the interaction of numerous factors including SOC (Bronick and Lal, 2005), and also texture and mineralogy. Baver (1934) explained how aggregation is affected by clay and silt contents, and Jenny (1941) stated that the amount and type of colloidal clay is intrinsically related to soil structure. Specifically, texture controls structure because aggregate size and stability depend on the balance between plasma (mostly clay and silt) and skeletal (mostly sand and gravel) soil constituents (Buol et al., 1997). Consequently, stabilization of SOC through aggregation is unlikely in sandy soils (Eusterhues et al., 2003), since low clay contents limit aggregation and thence POM occlusion. Thus, SOC retention can be affected by soil texture and mineralogy directly, through sorption by clays (Zinn et al., 2007), and indirectly through soil structure.

The structural control on SOC retention has been commonly approached as the relative protection of coarse SOC components (i.e., POM) by occlusion in aggregates (e.g., Beare et al., 1994; Golchin et al., 1994; Besnard et al., 1996). In accord with this conceptual background, the structural control does not refer to adsorbed, colloidal SOC, although it is known that SOC-depleted clays are more dispersible than those rich in SOC (e.g., Nelson et al., 1999). Obviously, all of these are working simplifications: Tisdall (1996) proposed that SOC occlusion may also occur by diffusion and deposition of organic compounds within micropores small enough to prevent microbial attack. Those SOC forms are probably equivalent to the clay-bound SOC obtained in particle-size fractionation experiments. Guggenberger and Kaiser (2003) also agreed that stabilization of altered organic matter by clays occurs by sorption and aggregation, with no clear distinction between them. Considerable visual evidence of POM occlusion exists in the literature, especially for disturbed soil samples. Mineral particles totally or partially occluding decayed organic particles can form microaggregates (Oades and Waters, 1991; Angers and Chenu, 1998) and macroaggregates (Oades and Waters, 1991; Golchin et al., 1994, 1998; Angers and Chenu, 1998). Fitzpatrick (1993) reported several examples of POM in soil thin sections, but the location of POM inside or outside aggregates is usually uncertain in undisturbed samples.

Although POM occlusion within aggregates is well documented (e.g., Besnard et al., 1996), its mechanisms and importance are neither homogeneous nor well established. The interactive effect of soil texture, mineralogy, management, and SOC on soil structure results necessarily in high variability of patterns and intensity of the structural control, as reflected in some reviews (Loveland and Webb, 2003; Blanco-Canqui and Lal, 2004). Some researchers (e.g., Elliott, 1986; Cambardella and Elliott, 1993; Puget et al., 1995) have reported that macroaggregates (>250-µm diameter) are SOC rich compared with microaggregates (<250 µm), while others have reported the opposite (e.g., Bossuyt et al., 2002). These apparent contradictions are due not only to the interactions described above, but also to widely different experimental procedures. For example, SOC concentrations can vary with aggregate size in tropical and temperate soils (Feller et al., 1996; Jastrow, 1996), but not after the correction for sand particles suggested by Elliott et al. (1991). Thus, differences in SOC concentration of whole fractions are mainly due to sand grains and POM occlusion, and SOC sorption is similar for all aggregate size classes. Also, separation of POM not occluded in aggregates (free POM) by flotation or other techniques has a critical effect. In summary, comparative interpretation of the literature must consider the effect of different climate, texture, and mineralogy, but also: (i) dry or wet sieving of aggregates; (ii) degree of slaking during wet sieving; (iii) post-sieving separation of free POM; and (iv) correction of SOC concentrations for sand particles.

In soils of the humid tropics, the structural control on SOC retention probably differs considerably from that for temperate climates. High contents of Fe and Al oxides in these soils greatly increase aggregate stability (Oades and Waters, 1991; Feller et al., 1996; Neufeldt et al., 1999) and may also affect the type of structure. For B horizons of clayey Oxisols in Brazil, gibbsitic clay results in small granular peds, whereas kaolinitic soils have small subangular blocky peds and much higher bulk density (Ferreira et al., 1999; Schaefer et al., 2004). Also, the concept of aggregate hierarchy states that highly stable microaggregates bind to form less stable macroaggregates (Edwards and Bremner, 1967), which has been validated for Alfisols and Mollisols (Oades and Waters, 1991; Tisdall, 1996) but not for Oxisols (Oades and Waters, 1991). The latter conclusion is partially undermined by the comparison between very clayey Oxisols and much coarser soils of other orders, which only adds to the uncertainty about the relationship between SOC and aggregation in the humid tropics. A final consideration is that soil structure and SOC are also affected by the activity of soil fauna, to an extent that is rarely assessed quantitatively (e.g., Humphreys, 1994; Lavelle et al., 2001).

The objective of this study was to assess the structural control on SOC retention in three soils of the Brazilian Cerrado, with due consideration to the interactive effects of soil texture, mineralogy, SOC, and faunal activity. The hypothesis tested was that soil structure predictably controls SOC retention, or at least POM forms, to a depth of 1 m in different-textured soils.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Composite, disturbed soil samples were obtained in triplicate from 0- to 5-, 5- to 10-, 10- to 20-, 20- to 30-, 30- to 40-, 50- to 60-, and 90- to 100-cm depths of three different-textured soils (clayey and loamy Haplustox, and sandy soils) under native Cerrado vegetation similar in composition and aboveground biomass. Climate (megathermic, ustic) and topography (top of interfluves) on all sites are the same or very similar. All soils are highly weathered, dominated by kaolinitic and oxidic clays, and with the exception of the clayey Haplustox, acidic with low base saturation. A detailed description of the sampling sites, particle size distribution, and soil properties is available in Zinn et al. (2007). Size distribution of WSA was measured by wet sieving (Yoder, 1936) using previously dry sieved (<8-mm) soil samples and a five-sieve set (2, 1, 0.5, 0.25, and 0.106 mm). After capillary wetting, the set was oscillated for 30 min (16 cycles min–1), and the WSA separates were oven dried at 40°C and weighed to calculate MWD (Youker and McGuiness, 1957).

In the present work, direct quantification of free POM from WSA separates would not be representative because part of free POM in the <8-mm soil sample was floatable and thus removed from the top sieve during the wet sieving. To quantify occluded and free POM in three selected depths (0–5, 30–40, and 90–100 cm), SOC concentrations were determined in bulk soil and WSA size separates, and their respective sand (>20-µm) fractions (adapted from Beare et al., 1994). Free POM was estimated by the difference between total sand-sized SOC determined in soil <2 mm (total POM, Zinn et al., 2007) and the weighed sum of the sand-sized SOC within WSA size fractions (occluded POM). This estimate is only as efficient as the separation of all free POM from the WSA, which in turn must cause minimal aggregate disruption. For WSA >1 mm, free POM was separated manually on a petri dish. For WSA <1 mm, free POM was separated by flotation in water as follows: a 3- to 4-g sample was placed in a 250-mL beaker, slowly wetted with a water spray, then immersed in approximately 100 mL of deionized water. To allow free POM to disentangle from overlying aggregates and float, the beakers were gently shaken, and WSA were gently stirred with water using a 20-mL polyethylene transfer pipette. The floating and slowly settling free POM was removed by decantation and siphoning with the transfer pipette, and the WSA samples were then oven dried at 40°C. The procedure was repeated until a free POM removal efficacy of about 99% was achieved, verified by observation of dry samples under microscope (20x magnification). Flotation in dense solutions was avoided for several reasons: (i) there is an unsolved conceptual problem of very high densities commonly ascribed to free POM (Christensen, 1992); (ii) no agreement exists about a standard density for POM; (iii) the chemical used to prepare the dense solution needs to be fully washed from the WSA fractions before the automated C and N analyses, which would cause aggregate disruption; and (iv) there is a need to identify a cost-effective procedure for use in simple laboratories in developing countries. From the WSA fractions thus obtained, a 1-g subsample was separated for C and N analyses, and the remainder dispersed in 0.1 M NaOH and wet sieved (20 µm) to collect sand fractions. The convenience in using NaOH rather than sonication for sample dispersion is explained in detail by Zinn et al. (2007); for this study, it must be added that no artifacts due to SOC dissolution are likely, since by convention only POM forms are retained in the sand fraction. The >20-µm fraction (mineral sand and occluded POM) was oven dried at 105°C, and homogenized with a bar mill. Total C and N concentrations in 1-g samples of intact WSA and sand fractions of WSA were determined by the dry combustion method in a Variomax CNHOS analyzer (Hanau, Germany). Free POM and occluded POM were estimated as follows:

Formula 1[1]

Formula 2[2]
where wsasand w and wsasand-SOC are the total weight and SOC content of particles >20 µm in the ith WSA size class (i varying from 1 to 5, or >2 mm to 0.25–0.106 mm). Statistical analyses were performed as mathematical regressions using a standard electronic spreadsheet and the software JMP IN 5.1 (SAS Institute, Cary, NC), whereas variability in quantitative data were presented as standard errors of the mean (refer to Zinn et al., 2007, for more detailed information on sampling design and analysis).

Micromorphology of Water-Stable Aggregate Larger than Two Millimeters
In general, WSA >2 mm of all soils had a predominantly subangular blocky structure, but a considerable part of the 0- to 5-cm-depth samples consisted of faunal peds with distinct morphological features. Thus, WSA >2 mm representing the common blocky aggregates and the three most common faunal peds were selected for micromorphological description and also for SOC and total N analyses. Thin sections were prepared using epoxy resin blocks mounted on glass slides.


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Soil Structure and Soil Organic Carbon Retention
The mean bulk SOC concentration and MWD of WSA along the profiles of the three soils under native Cerrado are shown in Fig. 1a and 1b, respectively. It is notable that SOC concentrations decline gradually with soil depth, whereas MWDs reach maximal values in the top 10 cm and decline sharply below that depth, with little variation. These discrepant profiles represent initial evidence that SOC and MWD may not be strongly correlated in the studied soils, at least in subsoil layers.


Figure 1
View larger version (25K):
[in this window]
[in a new window]

  		Fig. 1. Organic C and aggregation in soil profiles: (a) mean bulk soil organic C concentration; and (b) mean weight diameter of water-stable aggregates. Bars represent standard error (n = 3).

 
In addition, Fig. 1 shows that increasingly finer textures are associated with higher SOC concentrations throughout the 1-m profiles, and with higher MWDs for the top 20-cm depth only. The effect of texture on structure was also reflected in aggregate slaking during wet sieving, negligible in the clayey Haplustox, moderate in the loamy Haplustox, and extremely high in the sandy soils, even during capillary wetting. The texture–structure relation is mathematically depicted in Fig. 2a, which shows that MWD is directly and strongly correlated with clay + silt contents in all depths, but with higher intercept and slopes in the upper 10 cm, where most aggregation factors (e.g., SOC, biota, wetting–drying cycles, etc.) are more active. On the other hand, Fig. 2b shows that SOC concentration, shown earlier to be highly correlated with clay + silt contents in all depths (Zinn et al., 2005), is visibly correlated with MWD only in the 0- to 10-cm depth. The parameters and fit of the correlations in Fig. 2 are shown in Table 1. While the determination coefficients between MWD and clay + silt are generally high and decrease below 30-cm depth, those between SOC and MWD are high only in the 0- to 5-cm layer. The interpretation of Fig. 1 and 2 and Table 1 suggest that: (i) aggregation is strongly controlled by texture, as SOC concentration (Zinn et al., 2005); and (ii) the structural control on SOC retention, but also any potential aggregation effects of SOC, may be expressed only in the surface layer of those soils. Detailed assessment of aggregate size classes is necessary, however, for proper appreciation of the structural control.


Figure 2
View larger version (30K):
[in this window]
[in a new window]

  		Fig. 2. Linear relations between mean weight diameter and (a) clay + silt contents in bulk soil and (b) soil organic C concentration for 10 depths.

 

View this table:
[in this window]
[in a new window]

  		Table 1. Parameters for the linear relations for mean weight diameter (MWD) and soil organic C (SOC) shown in Fig. 2 (n = 9).

 
When individual WSA fractions are considered, the amount of WSA >2 mm increases as clay + silt increases for all depths, but declines markedly in subsoil (not shown, see Zinn, 2005). For sandy soils, WSA >2 mm decreased from >10% of the total soil sample from the 0- to 10-cm depth to approximately 1% at 1-m depth; for the loamy Haplustox this decline was from >20% to about 3%. For the clayey Haplustox, WSA >2 mm comprised about half of the total samples from 0- to 10-cm, declining to <10% below that depth. The dominant WSA size classes were: (i) for the sandy soils, 0.25 to 0.5 mm in all depths; (ii) for the loamy Haplustox, >2 mm in the top 10 cm and <0.25 mm below that depth; and (iii) for the clayey Haplustox, >2 mm in the top 10 cm, and <1 mm below that depth, with a nearly homogeneous distribution. Put more simply, the MWD reflects in great part the percentage of WSA >2 mm, which predominates in the top 10 cm and greatly increases in finer textured soils.

In all depths, large macroaggregates (WSA >1 mm) in the sandy and loamy soils always had higher total SOC concentrations than smaller size classes (not shown, see Zinn, 2005), partly because of relatively lower contents of mineral sand. Similar trends were reported by Lima and Anderson (1997) for two clayey Oxisols in the Cerrado region. After correction for the sand fraction, however, SOC concentrations showed little variation among WSA size classes for the same soil and depth (Fig. 3), as reported earlier for other soils of tropical and temperate climates (Feller et al., 1996; Jastrow, 1996). This similarity suggests that SOC partitioning throughout the finer soil components (silt and clay) is homogeneous for all aggregate sizes in the soil matrix. Thus, the five WSA size classes were averaged in Fig. 4, which shows that sand-free SOC concentrations in WSA were inversely proportional to clay + silt contents in each depth. From the mathematical regressions used to explain these relations, natural log functions consistently produced the best determination coefficients. These important trends indicate that the SOC dilution effect demonstrated by Christensen (1992) and other researchers occurs in natural soil environments and is not merely an artifact of soil dispersion. As reported separately for the clay, silt, and sand fractions (Zinn et al., 2007), this SOC dilution effect is stronger in the 0- to 5-cm depth.


Figure 3
View larger version (31K):
[in this window]
[in a new window]

  		Fig. 3. Soil organic C (SOC) concentration in individual water-stable aggregate size classes corrected for mineral sand and particulate organic matter in three selected soil depths. Bars represent standard error (n = 3).

 

Figure 4
View larger version (20K):
[in this window]
[in a new window]

  		Fig. 4. Graphic depiction of the soil organic C (SOC) dilution effect in sand-free water-stable aggregates (WSA, mean of five size classes) as a function of clay + silt contents in bulk soil (n = 9 for each depth).

 
The mechanism for in situ SOC dilution effect can be explained with the help of micromorphology. Figure 5 shows how the s-matrix and coarse to fine (c/f) distribution (Eswaran and Baños, 1976) of pedogenic, nonfaunal WSA >2 mm from the 0- to 5-cm layer varies with soil texture. The c/f distribution is close-spaced dermatic (plasma skins around skeletal grains) to intertextic (plasma bridges among grains) in the Quartzipsamment (Replicate 1 of sandy soils, Fig. 5a), evolving to a pronounced intertextic type in the sandy Haplustox (Replicates 2–3 of sandy soils, Fig. 5b). In the loamy and clayey Haplustoxes (Fig. 5c and 5d, respectively), the c/f distribution is open-spaced porphyric (grains within a plasmatic groundmass), common in Oxisols (Stoops and Buol, 1985). The clay and silt particles in the sandy soils, occurring in the form of thin skins and bridges, are relatively much more exposed to diffusion of colloidal and soluble SOC compounds, as denoted by their dark color. The clay and silt that form most of the groundmass in the loamy and clayey Haplustoxes are, conversely, relatively much less subject to this SOC diffusion, resulting in their overall lower SOC concentration.


Figure 5
View larger version (112K):
[in this window]
[in a new window]

  		Fig. 5. S-matrix and coarse to fine (c/f) distribution of pedogenic water-stable aggregates >2 mm, 0- to 5-cm depth: (a) Quartzipsamment—note dark, isotropic plasma and large packing voids; b) sandy Haplustox—note dark, isotropic plasma; (c) loamy Haplustox—note hollow root remains and surrounding void; and (d) clayey Haplustox—note opaque nodules. All images between crossed polars, except for (d) (plane polarized light).

 
The overall partitioning of POM outside and inside the combined aggregate size classes is shown in relative (as a percentage of bulk SOC) and absolute units in Fig. 6a and 6b, respectively. In the 0- to 5-cm depth, most POM occurs in the free form, i.e., outside aggregates, indicating that any protective effects of aggregation in surface soils are restricted to a minor part of total POM. For clayey Cerrado Oxisols, Roscoe et al. (2001) reported even lower occluded-POM/free-POM ratios than those in this study, because of the use of soil <2 mm instead of WSA size classes, which probably released POM occluded in large macroaggregates. Nevertheless, when all depths are considered, the relative POM protection within aggregates (Fig. 6a) is proportional to the clay content: occluded POM comprises approximately 40 to 50% of total POM in the clayey Haplustox, but only about 33 and 20 to 25% in the loamy Haplustox and sandy soils, respectively. This trend persists also in absolute units (Fig. 6b), supporting the idea that POM protection within aggregates is enhanced as the clay content increases (Balesdent et al., 2000).


Figure 6
View larger version (26K):
[in this window]
[in a new window]

  		Fig. 6. Total free and occluded particulate organic matter (POM) C of selected soil depths (a) as a percentage of total soil organic C (SOC), and (b) in absolute units. Bars represent standard error (n = 3).

 
Considering that occluded POM can be an important C pool partly protected from decomposition, and that this protection depends on soil texture, it could be interesting to develop a predictive model based on soil texture and depth. A procedure similar to those in previous models (Zinn et al., 2005, 2007) was used: different correlations were computed using data of free- and occluded-POM pools, and particle size distribution. The best results obtained were linear functions between free POM (as a percentage of total SOC) and clay contents (not shown, R2 = 0.72–0.90). Since the intercept and slopes of those functions vary randomly with depth (not shown), a profile pedotransfer function was not feasible. A significant relationship for the combined 0- to 5-, 30- to 40-, and 90- to 100-cm depths was obtained, however:

Formula 3[3]
where clay content is in grams per kilogram soil; however, the wide 95% confidence interval for Eq. [3] (Fig. 7) shows that the predictive value of this equation is very low.


Figure 7
View larger version (22K):
[in this window]
[in a new window]

  		Fig. 7. Linear relation between free POM C percentage and clay content for the combined 0- to 5-, 30- to 40-, and 90- to 100-cm depths (R2 = 0.74, n = 27). Dotted line is 95% confidence interval.

 
Although POM occlusion is often considered a major mechanism for SOC retention, data from soils under temperate (e.g., Beare et al., 1994; Besnard et al., 1996) and tropical (Roscoe et al., 2001) climates support this study and the suggestion that most POM is not protected within aggregates, or at least within WSA. Consequently, the structural control on SOC retention is probably restricted to only part of the total POM pool, and depends strongly on soil texture. On the other hand, it is important to note that, for a specific depth, total POM in the same depth did not vary with soil texture (Fig. 6b; Zinn et al., 2007). Assuming that POM occlusion results in protection from decomposition and is proportional to clay content, then total POM in clayey soils must be more than that in sandy soils. That not being the case suggests that POM occlusion within aggregates of the soils studied does not effectively protect these forms from decomposition, despite the data in Fig. 6. This has been recognized earlier in the literature: Baldock and Skjemstad (2000) stated that POM protection inside aggregates does not stop decomposition, but rather reduces its rate. Micromorphology can also give some insight into the effectiveness of POM protection by occlusion. Figure 8 provides a detail of Fig. 5c, showing a root remnant occluded in a WSA from the loamy Haplustox (0–5-cm depth). The inner core underwent active decomposition, whereas the epidermal tissues remained much longer for two probable reasons: (i) its cells are naturally hardened in comparison to the cortex; and (ii) clay particles closely adhere to external epidermal cells, as seen by the red hues, which may effectively delay decomposition. The latter conclusion is reinforced by the high porosity within and around the root, allowing unimpeded access to microbial decomposers. Further research must also include an additional approach, i.e., the protection of colloidal and soluble SOC within aggregates, developing a methodological procedure capable of differentiating this effect from the simple sorption by clay domains.


Figure 8
View larger version (94K):
[in this window]
[in a new window]

  		Fig. 8. Loamy Haplustox, 0- to 5-cm depth (detail of Fig. 5c): occluded particulate organic matter as a partially decomposed root. Note adhesion of soil particles to root surface (between crossed polars).

 
The reverse side of the structure–SOC relation, that is, the role of SOC as an aggregation factor, must also be treated with caution in the studied soils. Aggregation and SOC concentration are the highest in the top 10 cm, but so are other aggregation factors such as biotic activity and swell–shrink cycles (Bronick and Lal, 2005), so it is difficult to isolate the effect of SOC on aggregation. Additionally, Tisdall and Oades (1982) concluded that SOC is not the major binding agent in aggregates, not all SOC is involved in stabilization, and there is a threshold SOC concentration beyond which there is no more effect on aggregation. This does not mean, however, that SOC and POM are not important to aggregation in soils of the Cerrado region. For the 0- to 40-cm depth, WSA >1 mm in sandy soils contained more occluded POM than in loamy and clayey Oxisols (Zinn, 2005), suggesting that POM may promote some aggregate stabilization where low clay contents are limiting, as also reported for soils of the Niger floodplain (Igwe and Stahr, 2004). Finally, the data suggests that bulk SOC and also POM do not effectively promote aggregation in subsoil layers, where mineral plasma must be the main aggregation factor.

Carbon/Nitrogen Ratios of Water-Stable Aggregate Size Classes
The C/N ratios of intact, undispersed WSA vary little among size classes (between 10 and 15) and soil type, and only in the clayey Haplustox do C/N ratios decline in the subsoil (data not shown; see Zinn, 2005). Carbon/N ratios of sand fractions extracted from WSA (i.e., occluded POM), however, are much higher and follow a very different pattern (Fig. 9). Only in the 0- to 5-cm depth are C/N ratios comparable to those in intact WSA, and for the 30- to 40- and 90- to 100-cm depths, the C/N ratios of occluded POM in all soils are extremely high and variable, especially in the sandy and loamy soils. In comparison, bulk C/N ratios of all soils decline in subsoil (Zinn et al., 2007), but even in the 0- to 5-cm depth are lower than those of their respective intact WSA. The sand fractions extracted from bulk soils had C/N ratios >20 for the 0- to 5-cm layer, which increased in the subsoil (except for the clayey Haplustox). This trend is also observed for sand extracted from WSA, but the C/N ratios reach almost 200. Since C/N ratios of silt and clay fractions from bulk soils are <15 and vary little with depth, POM is responsible for higher bulk C/N ratios at or near the soil surface. For the lower depths of the sandy and loamy soils, the extreme C/N ratios of occluded POM may reflect root tissues with initially low C/N ratios that in time became devoid of N, or charcoal particles (Teixeira et al., 2002). In any case, the C/N ratios are evidence that POM occluded within aggregates is chemically different from free POM, probably more recalcitrant.


Figure 9
View larger version (38K):
[in this window]
[in a new window]

  		Fig. 9. Mean C/N ratio of occluded particulate organic matter (POM) of sand fractions from dispersed water-stable aggregate size classes of selected soil depths. Bars represent standard error (n = 3).

 
Faunal Peds and Soil Organic Carbon Retention
The occurrence of faunal or zoogenic micropeds as irregular patches in Oxisols was briefly described by Stoops and Buol (1985), and the role of fauna in the structure has been noted by many researchers in surface (Lavelle et al., 2001) or subsoil horizons (Schaefer, 2001) of those soils. Faunal peds were not originally included as part of the structural control of SOC retention, and no specific samplings for fauna or faunal structures were conducted in this study. Their unexpected and common occurrence as large macroaggregates, however, suggested not only a micromorphological study, but also that they could differ from bulk soil and other WSA in SOC retention. Figure 10 shows that the main types of faunal peds occurring in the WSA >2-mm fractions can be classified as: (i) aggrotubules; (ii) fecal pellets; and (iii) cocoons.


Figure 10
View larger version (102K):
[in this window]
[in a new window]

  		Fig. 10. Faunal peds in water-stable aggregates >2 mm, 0- to 5-cm depth: (a) loamy Haplustox, aggrotubule; (b) clayey Haplustox, fecal pellet; (c) Quartzipsamment, cocoon (a–c, in oblique incident light); and (d) clayey Haplustox, detail of aggrotubule with unknown euhedral crystal (between crossed polars). Scale in millimeters.

 
Aggrotubules, defined by Brewer (1964) as pedotubules in which the infilling of tunnels and chambers has been performed by organisms, are the most striking and common faunal peds: they comprise 12% of the weight of WSA >2 mm in the Quartzipsamment, and 6, 11, and 26% in the sandy, loamy, and clayey Haplustoxes, respectively. They are basically clusters (~1-cm diameter) of ellipsoidal crumbs of 0.5 to 1 mm in diameter (Fig. 10a), cemented with moderate to large packing porosity, sometimes including coarse quartz grains (Fig. 10d) and POM (Zinn, 2005). Their porphyric s-matrix is very different from that of pedogenic, subangular blocks, with much smaller and fewer skeletal grains (as often reported for faunal peds) and no internal voids (Fig. 10d). This concentration of plasma is partially responsible for the higher C and N concentrations compared with bulk soils and pedogenic WSA >2 mm, except for the clayey Haplustox (Table 2). They fit the Variety 3 aggrotubule of Brewer (1964), composed of intensely reworked soil material, and the rounded, smooth outline of individual grains indicates they were egested in a semiliquid state. In other studies, aggrotubules are often named differently (e.g., as "granular infilling" in Stoops et al., 1994), but their faunal origin is easily recognizable. In Cerrado soils, their occurrence was reported by Balbino et al. (2002) and Gomes et al. (2004).


View this table:
[in this window]
[in a new window]

  		Table 2. Soil organic C (SOC) concentration and C/N ratios of faunal peds, blocky water-stable aggregates >2 mm, and bulk soil under Cerrado.

 
The occurrence of aggrotubules in all soils indicates that their builders are widely distributed and a major factor in the structure of neotropical savanna soils—most likely social arthropods such as mound termites or leaf-cutting ants, both common in the sampled areas. The fact that aggrotubules are more common in finer textured soils agrees with the higher frequency of termite mounds in clayey soils (Sys, 1955, as cited by Lal, 1987), but can also result from stronger ped stability. It is generally recognized that faunal structures are stable and persistent in soils (Lavelle et al., 2001), but the fate of the aggrotubules apparently differs among soil types. In the loamy and clayey Haplustoxes, some aggrotubules were degraded and compacted until reaching a faintly granular surface almost indistinct from pedogenic aggregates. Indeed, some thin sections of WSA from deep layers, initially considered pedogenic, showed aggrotubule features internally (not shown, Zinn, 2005), which generally does not occur in sandy soils. Aggrotubules may also break into smaller clusters or individual crumbs, however, which can be seen in micrographs of B horizons of sandy and clayey Haplustoxes (Balbino et al., 2002). Schaefer (2001) observed many aggrotubule remains in deep horizons of Brazilian clayey Oxisols, and proposed a theory about their relation with the granular structure commonly observed in these soils.

The fecal pellets (Fig. 10b) comprise <0.1% of the mass of WSA >2 mm of all soils, are mostly composed of mineral particles, and are very hard when dry. These also have an ellipsoidal, smooth shape, but are much larger than individual crumbs in aggrotubules. The porphyric s-matrix contains less and smaller skeletal grains than pedogenic peds. Their SOC concentration is always higher than in bulk soils, varying little with soil type (Table 2). Their producers are also unknown, but one possibility is Coleopterae larvae that feed on decaying termite mounds in the Cerrados (Matthews, 1977), and whose pellets are also characterized by high mineral contents (Fitzpatrick, 1993). Another alternative are cicada larvae, common in the Cerrados (especially Quesada spp., Majeorona spp., or Fidicina spp.), which in Australia produce pellets of similar size to those described here (Humphreys, 1994).

Cocoons (Fig. 10c) were also observed in all soil types, but are much more common and larger in the sandy soils, where they comprise about 0.5% of the mass of WSA 2 to 8 mm. They consist of soil tubes, hollow or backfilled, 2 to 3 mm in diameter and up to 1 cm long, and strongly concentrated in plasma compared with the surrounding soil. In thin sections, cocoons are almost opaque, probably because of high humus concentration; however, their SOC concentration is lower than that of aggrotubules and fecal pellets (Table 2), because of a significant content of coarse quartz. Their origin is also unknown.


   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The structural control on SOC retention, defined here as the occlusion of POM within aggregates stable in water, can be briefly described as: (i) mainly restricted to a minor part of the total POM pool; (ii) dependent on soil texture, since POM occlusion is favored in finer textured soils; and (iii) partly effective in protecting occluded POM, as indicated by C/N ratios considerably higher than those of total POM and bulk soil. On the other hand, evidence against a strong structural control include the following: (i) correlations between SOC concentration and MWD were strong only in the 0- to 5-cm depth, although aggregation was significant below that depth; (ii) higher POM occlusion in finer textured soils did not result in larger total POM pools; and (iii) correlations among POM occlusion and textural and depth data were weak and no predictive models could be developed. Therefore, one can conclude that structural controls on SOC retention in Cerrado soils are not deterministic or fully effective in protecting POM from decay, and thus not as strong as the textural and mineralogical controls. Additionally, preliminary evidence suggests that faunal activity generates peds that may be enriched in SOC, but whose stability and abundance also depend on soil texture.


   ACKNOWLEDGMENTS
 
This work is part of the doctoral dissertation of Y.L. Zinn, sponsored by the Capes Foundation (Ministry of Education, Brazil) and by a Presidential Fellowship from the Graduate School of the Ohio State University (OSU). We thank Embrapa Cerrados (in special Mr. Jesuíno S. Caldas and Mr. Wantuir C. Vieira) and V&M Florestal Co. (in special Dr. Hélder B. Andrade) for their support with the sampling operations. We also express our thanks to Mr. Franklin S. Jones and Yogi Raut (Soil Science, OSU) for their important help in the laboratory procedures.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Current address: The Capes Foundation, Ministry of Education, Cx. Postal 365 Brasília DF 70375, Brazil

All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.

Received for publication January 11, 2006.


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





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 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 Google Scholar
Google Scholar
Right arrow Articles by Zinn, Y. L.
Right arrow Articles by Resck, D. V. S.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Zinn, Y. L.
Right arrow Articles by Resck, D. V. S.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Zinn, Y. L.
Right arrow Articles by Resck, D. V. S.
Related Collections
Right arrow Soil Organic Matter
Right arrow Carbon Sequestration
Right arrow Tropical Soil Management

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