See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/238174789

Aggregate-Protected and Unprotected Organic Matter Pools in Conventional and

No-Tillage Soils

Article  in  Soil Science Society of America Journal · January 1994

DOI: 10.2136/sssaj1994.03615995005800030021x

CITATIONS READS
610 4,349

4 authors, including:

David C. Coleman
University of Georgia
377 PUBLICATIONS   31,559 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Impact of land-management of soil microbial communities View project

Physiological and biomolecular mechanisms involved in the effects of natural products and their active compounds. Plants of the genus Smallanthus (Smallanthus
sonchifolius -yacon-, Smallanthus macroscyphus) View project

All content following this page was uploaded by David C. Coleman on 31 October 2014.

The user has requested enhancement of the downloaded file.

Published May, 1994

Aggregate-Protected and Unprotected Organic Matter Pools

in Conventional- and No-Tillage Soils
M. H. Beare,* M. L. Cabrera, P. F. Hendrix, and D. C. Coleman
ABSTRACT adsorption to clay minerals (Oades, 1984; Ladd et al.,
No-tillage (NT) practices can result in greater soil aggregation and 1985) and the formation of microaggregates (Edwards
higher soil organic matter (SOM) levels than conventional-tillage (CT) and Bremner, 1967; Gregorich et al., 1989), by isolation
practices, but the mechanisms for these effects are poorly known. Our in micropores (Adu and Oades, 1978; Foster, 1981) and
objectives were to describe the size and quality of biologically active by physical protection within stable macroaggregates
pools of aggregate-associated SOM in long-term CT and NT soils of (Elliott, 1986; Gupta and Germida, 1988).
the southeastern USA. Samples were collected from replicated CT Several models of organic matter turnover in soil
and NT plots on a Hiwassee sandy clay loam (clayey, kaolinitic, thermic include pools of physically protected SOM (Jenkinson
Rhodic Kanhapludult) and separated into four aggregate size classes
(>2000, 250-2000, 106-250, 53-106 urn) by wet sieving. Potentially
and Rayner, 1977; van Veen et al., 1985; Parton et al.,
mineralizable C and N and N2O emissions were measured from 20-d 1987). Tisdall and Oades (1982) presented a conceptual
laboratory incubations of intact and crushed macroaggregates (>250 model of soil structure that describes the binding of
fun) and intact microaggregates (<250 um). Three primary pools of primary mineral particles into microaggregates (50-250-
aggregate-associated SOM were quantified: unprotected, protected, Hm diam.) and of microaggregates into macroaggregates
and resistant C and N. Aggregate-unprotected pools of SOM were 21 (>250-nm diam.). The more persistent agents (e.g.,
to 65% higher in surface soils of NT than of CT, with greater differences clay-polyvalent metal-humified organic matter com-
in the macroaggregate size classes. Disruption of macroaggregates plexes) that bind microaggregates are generally charac-
increased the mineralization of SOM in NT but had little effect in terized as older, more humified, or recalcitrant SOM.
CT. Rates of mineralization from protected and unprotected pools of The more temporary (e.g., roots and fungal hyphae)
C were higher in surface soils of CT than of NT. Macroaggregate-
protected SOM accounted for 18.8 and 19.1% of the total mineralizable
and transient (e.g., polysaccharides) agents that bind
C and N (0-15 cm), respectively, in NT but only 10.2 and 5.4% of microaggregates into macroaggregates are generally con-
the total mineralizable C and N in CT. Our results indicate that sidered to be relatively more labile or decomposable
macroaggregates in NT soils provide an important mechanism for the (van Veen and Paul, 1981; Elliott, 1986; Elliott and
protection of SOM that may otherwise be mineralized under CT Coleman, 1988).
practices. Elliott (1986) and Gupta and Germida (1988) attributed
much of the SOM lost during cultivation of grassland
soils to the mineralization of SOM binding microaggre-
T ONG-TERM CULTIVATION alters soil structure and in- gates into macroaggregates. Macroaggregates are much
J—/ creases the losses of SOM (Dalai and Mayer, 1986). less stable than microaggregates (Edwards and Bremner,
Apart from soil type and climatic variables, the magni- 1967; Oades, 1984; Beare et al., 1994), probably because
tude of these effects depends on the intensity of cultiva- of the nature of the binding agents involved. Macroaggre-
tion, in particular the type and frequency of tillage and gates are also more susceptible to the disruptive forces
the quantity and quality of fertilizers and organic residues of cultivation and to the dispersion that results from
returned to the soil (Jenkinson and Rayner, 1977; Ras- rapid wetting or raindrop impact (Tisdall and Oades,
mussen and Collins, 1991). Several studies have shown 1980, 1982).
that reduced-tillage practices can result in greater aggre- Although several studies have described the location
gation and higher standing stocks of SOM compared (Foster, 1981; Oades, 1984) and composition (Gupta
with CT practices (Doran, 1980; Lamb et al., 1985; and Germida, 1988; Dormaar, 1984; Beare et al., 1994)
Bruce et al., 1990; Havlin et al., 1990; Carter, 1992). of aggregate-associated SOM, few have provided experi-
Others report little or no effects of reduced-tillage prac- mental evidence for the physical protection of SOM in
tices on these properties (Hamblin, 1980; Carter and macroaggregates (Adu and Oades, 1978; Elliott, 1986;
Rennie, 1982), especially if the crop residues are burned Gupta and Germida, 1988). Christensen (1987) presented
(Carter and Mele, 1992). Where differences between results from fractionated and whole soils that showed
tillage practices are recognized, however, the mecha- no evidence of physically protected SOM. Gregorich et
nisms that regulate the accumulation or loss of SOM are al. (1989) used ultrasonication and particle-size fraction-
poorly known. ation to define labile and recalcitrant pools of microag-
Conventional tillage practices disrupt soil aggregates, gregate-protected C. There are, however, no studies that
exposing more organic matter to microbial attack. Or- define the size and quality of biologically active pools
ganic matter may be protected from microbial attack by of macroaggregate-associated SOM that, unlike microag-
gregate-associated SOM (Edwards and Bremner, 1967),
may be influenced by cultivation practices.
M.H. Beare, Inst. for Crop and Food Research (Lincoln), Private Bag The highly dispersible, kaolinitic-clay-based soils of
4704, Christchurch, New Zealand; and P.P. Hendrix and M.L. Cabrera,
Dep. of Agronomy and Inst. of Ecology, and D.C. Coleman, Inst. of
Ecology, Univ. of Georgia, Athens, GA 30602. Received 8 Mar. 1993. Abbreviations: NT, no-tillage; SOM, soil organic matter; CT, conven-
"Corresponding author. tional-tillage; HSB, Horseshoe Bend Experimental Area; WSA, water-
stable aggregates; WFP, water-filled porosity; ANOVA, analysis of vari-
Published in Soil Sci. Soc. Am. J. 58:787-795 (1994). ance.
787
788 SOIL SCI. SOC. AM. J., VOL. 58, MAY-JUNE 1994

the southeastern USA (Miller and Baharuddin, 1986) are Aggregate Mineralization Assays
particularly susceptible to the losses of SOM that result The potentially mineralizable C and N of aggregate fractions
from intensive cultivation (Giddens, 1957). Under re- was measured following methods adapted from Elliott (1986).
duced-tillage or NT management, soils of this region Six aggregate treatments were established for each replicate
can maintain greater water-stable aggregation and higher of the tillage and depth samples: (i) intact >2000-nm aggre-
standing stocks of SOM than do CT soils (Bruce et al., gates, (ii) crushed >2000-um aggregates (to pass a 250-um
1990, 1992; Beare et al., 1994). However, the importance screen), (iii) intact 250- to 2000-um aggregates, (iv) crushed
of aggregate formation and stabilization in regulating 250- to 2000-um aggregates (to pass a 250-um screen), (v)
the accumulation or loss of SOM in CT and NT soils intact 106- to 250-um aggregates, and (vi) intact 53- to 106-um
is not well known. Furthermore, understanding these aggregates. Aggregates <53 urn in diameter were not included
relationships may be particularly important for devel- in these assays because of their contamination with CaCk.
Subsamples of the aggregate treatments (5-20 g) were care-
oping environmentally sound and sustainable crop pro- fully mixed hi vials with equal quantities of pre-ashed sand
duction systems for the southeastern USA. (500 °C, 4 h) and tapped gently. Deionized water was added
Our objectives were to: (i) investigate the extent to to achieve 55% WFP. Vials were incubated (24 °C) in sealed
which the micro-macroaggregate model of SOM protec- jars containing alkali CO2 traps (0.05-0.1 M NaOH). The COa
tion in grassland soils can be extended to long-term CT traps were changed on days 3, 6, 10, 15, and 20, and the
and NT soils of the southeastern USA, (ii) quantify the respired C was measured by titration with standardized HC1.
size and quality of biologically active pools of aggregate- The jar lids were fitted with rubber septa for the collection
associated SOM, and (iii) describe the influence of CT of headspace ( = 400-mL volume) gas samples (5 mL, gas-tight
and NT management on these pools. syringe) prior to collecting the alkali traps. The efficacy of
the alkali traps was confirmed by CC<2 analysis of headspace
gases with a Hewlett-Packard GC equipped with a thermal
conductivity detector and a Porapak N column (Alltech Associ-
MATERIALS AND METHODS ates, Inc., Deerfield, IL), using He as the carrier gas. Nitrous
oxide concentrations were measured on a Tracor GC (Model
Site Description and Soils 550, Trametrics Analytical Division, Austin, TX) equipped
Soils were sampled from the long-term CT (moldboard with a 63Ni electron capture detector and two 1.8-m Porapak
plowed, disked, and rotary-tilled) and NT (direct-drilled) plots Q 80 to 100 mesh columns, using a mixture of Ar and CHt
at the HSB near Athens, GA. The soil at this site is a well- as the carrier. The NKtf (phenol method; Keeney and Nelson,
drained sandy'clay loam (66% sand, 21 % clay) in the Hiwassee 1982) and (NO2~ + NO3")-N (Gries-Ilosvay method; Keeney
series. Some general physical and chemical properties of these and Nelson, 1982) contents of the pre- and post incubation
soils are given in Beare et al. (1994). The tillage treatments (Day 20) aggregate treatments were measured from 1 M KC1
were established in 1978 on replicated (n = 4) plots (0.1 ha) extracts on an Alpkem RFA-300 Autoanalyzer (Alpkem Corp.,
assigned in a completely randomized design. For the 4 yr prior Clackamas, OR). Following extraction, all aggregate treat-
to sampling, all plots were double cropped to grain sorghum ments were dispersed in 0.05 M NaOH and washed over a
[Sorghum bicolor (L.) Moench] and winter rye (Secale cereale 53-um screen, and the dry (105 °C) mass of sand was recorded.
L.). Further details on the site and cultivation histories can
be found in Stinner et al. (1984) and Beare et al. (1992). Calculations and Statistical Analyses
Samples were taken from the replicate NT and CT plots on The results of the aggregate mineralization assays were used
2 May 1991, =2 wk prior to harvesting the winter cover to operationally define three primary forms of aggregate-
crop. No residues were removed. Six replicate soil cores were associated SOM: aggregate-unprotected, macroaggregate-
collected randomly from each plot with a thin-walled steel protected, and aggregate-resistant SOM. These forms of SOM
cylindrical corer fitted with a removable plastic insert (5.8-cm were calculated as follows:
i.d.). The intact cores were sectioned into 0- to 5- and 5- to
15-cm increments and composited within depths. Soils were unprotected C^f) = intact aggregate C^f),
stored field moist (3 °C) in crush-resistant, air-tight containers protected Cmin(0 = crushed aggregate C^a(f)
prior to analysis (<2 wk). The samples were prepared for wet
sieving as described previously (Beare et al., 1994). - intact aggregate Cmin(0
where C^t) is the cumulative C mineralized at time t (d)
Water-Stable Aggregates from the crushed and intact aggregate treatments. The unpro-
Intact WSA were collected following the methods of Beare tected and protected C^f) data were used to calculate the
et al. (1994). Briefly, field-moist soils were wetted under pools, Co (g C kg"1 sand-free) and corresponding mineralization
constant head (3.0 cm of HO) and separated into five aggregate rate constants, k (d~ l ) of protected and unprotected C using
size classes (>2000-, 250-2000-, 106-250-, 53-106-, and the following first-order equation: Cmin(0 = C0(l - e~ fa ). The
<53-u,m diam.) by wet sieving (using a 2-cm stroke for 300 s data were fitted to the model using the Marquardt method
at 0.52 cycles s"1). Aggregates were dried on the sieves in a (Bates and Watts, 1988) of nonlinear curve fitting (SAS Insti-
dehumidifying chamber (10 °C, 24 h), transferred to beakers, tute, 1992). Pools of aggregate-resistant C were calculated as
and dried at 35 °C (48 h) before the final dry masses were resistant Co = total C - unprotected C0 - protected Co
recorded. Subsamples of each fraction (0.25-2.0 g) were dried
at 105 °C to allow correction to a final dry weight. Additional based on the pools defined above.
subsamples (0.5-4.0 g) were ground in a mill (Spex Industries, The aggregate-unprotected, macroaggregate-protected, and
Edison, NJ) and total C and N content determined by Dumas aggregate-resistant pools of N were defined as the net N mineral-
combustion on a Carlo Erba C/N Analyzer (Carlo Erba Instru- ized (N™, N released as NHf + [NOf + NOf] + N2O) from
ments, Milano, Italy). intact aggregates (20 d), the difference in H™ from crushed
BEARE ET AL.: AGGREGATE-PROTECTED ORGANIC MATTER POOLS 789

and intact macroaggregates, and the difference between total however, there was little change in Cmin with depth except
aggregate N and the M,™ from crushed macroaggregates, re- for the 250- to 2000-n,m WSA, which mineralized 1.62
spectively. Standing stocks of aggregate-protected and unpro- times more C in 0- to 5-cm than in 5- to 15-cm samples.
tected C and N were calculated by multiplying the pools of In most cases, the Cmin from intact WSA by Day 20 was
protected and unprotected C and N (grams per kilogram of significantly (P < 0.05) higher for macroaggregates than
whole aggregate) by the quantity of whole soil in each aggregate for microaggregates regardless of tillage treatment or
size class, correcting for differences in bulk density between
depths and tillages (Beare et al., 1994). soil depth. In surface samples of NT, crushing macroag-
The data were analyzed statistically using ANOVA, with gregates (>250 um) resulted hi 27 to 30% increases (P <
nesting of plot within tillage. These analyses were accomplished 0.05) in Cmin compared with that of intact aggregates, but
using SuperANOVA and StateView SE 4-Graphics (Abacus there were no effects of crushing on Cmin from surface
Concepts, 1987, 1989). The C0 and k values calculated for macroaggregates of CT. For macroaggregates of the
the individual 'replicates of each aggregate-associated organic 5- to 15-cm samples of both tillages, however, crushing
matter pool were analyzed by ANOVA as described above. significantly increased (13-28%) Cmin compared with
intact aggregates, although the quantity of C^n was much
lower than that of surface samples.
RESULTS Aggregate-unprotected and macroaggregate-protected
The distribution of WSA from the NT and CT soils Cmin were calculated for each time point and fitted to a
at HSB and their total sand-free concentrations of C first-order model to estimate the size of the potentially
and N are reported in Beare et al. (1994). To describe mineralizable C pools (C0) and their rates of C minerali-
differences in the mineralization of aggregate-associated zation (k) (Table 1). In surface samples, aggregate-
C and N between aggregate sizes, all values were normal- unprotected C pools were significantly affected by tillage
ized to a sand-free basis. The quantity of C mineralized treatment (Fi,6 = 53.6; P = 0.0003) and aggregate size
during the 20-d incubation period was expressed as cumu- class (F3,i8 = 55.9; P = 0.0001). Unprotected pools of
lative respired CO2-C (Fig. 1). The resulting mineraliza- C from the 0- to 5-cm depth ranged from 1.20 (53-106
tion curves revealed significant effects (P < 0.05) of u,m) to 1.63 (>2000 um) times higher in NT than in
tillage, size classes, and soil depth on the quantity of C CT (P < 0.05). There were no differences between tillage
mineralized from intact WSA during the 20-d incubation. treatments in the 5- to 15-cm samples. In surface samples,
In general, the quantity of Cmin from 250- to 2000-p.m the largest pools of unprotected C were found in 250-
WSA was slightly higher than from >2000-nm WSA in to 2000-nm WSA and the smallest pools in 53- to 106-nm
both tillage treatments and soil depths. In NT, the Cmin aggregates of both tillage treatments. The decline in
from WSA was 1.7 to 2.9 times higher in surface samples unprotected C with depth was much more pronounced
(0-5 cm) than at depth (5-15 cm) (P < 0.0001). In CT, hi NT than in CT samples. In CT, unprotected C of the

>2000 um AGGREGATES 250-2000 AGGREGATES AGGREGATES

| £ 4.0n NT 106 |im
o> NT 53 \an
CT 106 |im
! 3.0- CT 53|im
o
m
IE
" £ 2.0
Q
ul LSD (0.051
I NT (20d) 1>0
1.0-
J CT (20d)
m o.O 0.0
0 2 4 6 8 10 12 14 16 18 20 0 2 4 8 10 12 14 16 18 20 4 6 8 10 12 14 16 18 20

LSD (0.051
I NT (20d)

I CT (20d)

0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20 0 2 4 6 8 10 12 14 16 18 20
INCUBATION TIME (d) INCUBATION TIME (d) INCUBATION TIME (d)

Fig. 1. Cumulative C mineralized (Crab,, g kg'1 sand-free aggregates) from intact and crushed macroaggregates (>2000 and 250-2000 um) and
intact microaggregates (106-250 and 53-106 um) of the 0- to 5- and 5- to 15-cm depths of no-tillage and conventional-tillage soils. Values are
means; n - 4.
790 SOIL SCI. SOC. AM. J., VOL. 58, MAY-JUNE 1994

Table 1. Mineralization potentials and rate constants for protected and unprotected pools of C in water-stable aggregates of conventional-
tillage (CT) and no-tillage (NT) soils at the Horseshoe Bend Experimental Area.
Carbon mineralization potential (C0)t Rate constants (fc)t
Aggregate-unprotected C Macroaggregate-protected C Aggregate-unprotected C Macroaggregate-protected C
Aggregate
size class NT CT P<0.05 NT CT P<0.05 NT CT P < 0.05 NT CT P < 0.05
———— g C k g - 4 —— d-'
0-5-cm depth
>2000 2.35 1.44 * 0.68 0.08 * 0.127 0.137 0.196 0.300 *
250-2000 3.17 2.26 * 0.92 0.28 * 0.124 0.146 * 0.183 0.203
106-250 2.85 1.62 * 0.137 0.129
53-106 1.28 1.07 0.217 0.236
LSD(0.05)§ 0.44 0.38 0.32 0.10 0.019 0.031 0.040 0.082
5-15-cm depth
>2000 0.82 1.08 0.16 0.15 0.157 0.153 0.201 0.303 *
250-2000 1.38 1.34 0.36 0.35 0.106 0.145 * 0.127 0.129
106-250 1.23 1.44 0.123 0.133
53-106 0.76 0.72 0.149 0.165
LSD(0.05)§ 0.31 0.37 0.10 0.09 0.031 0.017 0.058 0.080
* Significant differences (two-tailed (-test) at the P < 0.05 level between tillage treatments within aggregate size classes.
t Asymptotic parameter estimates based on nonlinear regression fitting of the first-order model: C^ = C0(l - e"*1)-
| Units are grams of CO2-C respired per kilogram of sand-free aggregates.
§ Fisher's least significant difference values for comparison of means of size classes within tillage and soil depth (ANOVA).

250- to 2000-nm WSA showed the greatest change (P < the unprotected C in CT. Rates of C mineralization from
0.05) with depth, being 1.62 times higher in the surface the macroaggregate-protected pools were significantly
samples than at depth. (P < 0.05) higher for >2000-um WSA of CT than of
The rate of C mineralization from the aggregate unpro- NT at both soil depths, but there were no differences
tected pools was significantly higher in the 250- to 2000- for 250- to 2000-um WSA. Except for 250- to 2000-nm
Hm WSA of CT than of NT at both depths (P < 0.05). WSA from 5- to 15-cm samples of CT, rates of C
No other tillage differences were observed. In surface mineralization were higher for macroaggregate-protected
samples, unprotected pools of C from 53- to 106-um than for unprotected pools of C from the same size class
WSA had significantly (P < 0.05) higher mineralization and sample depth.
rates than the other size classes in both NT (F3 6 = 51.8; As a percentage of total aggregate C, aggregate-
P = 0.0001) and CT (F3.i2 = 24.7; P = 0.0001). unprotected C was 1.2 to 1.8 times higher in WSA from
In surface samples, pools of macroaggregate-protected the surface samples of CT than of NT (Fi,6 = 35.8; P =
C were 3.3 (250-2000 urn) to 8.5 (>2000 um) tunes 0.001) (Table 2). There were also significant differences
higher in WSA of NT than of CT (F1>6 = 47.9; P = between size classes (F3,ig = 23.1; P = 0.0001) in the
0.0004). Protected pools of C were much lower in the surface samples of both NT and CT. Aggregate-
5- to 15- than the 0- to 5-cm samples of NT, but there unprotected C made up the highest percentage of total
were no significant differences with depth in CT. In aggregate C in 250- to 2000-nm WSA and the lowest
surface samples, macroaggregate-protected C was 29 to percentage in the 53- to 106-M.m WSA. In 5- to 15-cm
30% of the unprotected C in NT but only 5 to 12% of samples, there were similar differences between size

Table 2. Protected and unprotected pools of C and N as a percentage of total aggregate C and N, respectively, in water-stable aggregates
of conventional-tillage (CT) and no-tillage (NT) soils at the Horseshoe Bend Experimental Area.
Aggregate-unprotected C Aggregate-protected C Aggregate-unprotected N Aggregate-protected N
Aggregate
size class NT CT P<0.05 NT - CT P<0.05 NT CT P<0.05 NT CT P < 0.05
OJ
fim
0-5-cm depth
>2000 2.77 3.51 * 0.85 0.19 * 3.22 5.66 * 1.57 0.32 *
250-2000 3.11 4.52 * 0.90 0.51 * 4.10 5.15 * 1.44 0.40 *
106-250 2.62 3.16 3.72 4.24
53-106 1.52 2.77 * 4.52 6.34 *
LSD(0.05)t 0.44 0.81 0.32 0.11 0.66 1.25 0.57 0.23
S-15-cm depth
>2000 3.49 3.59 0.68 0.57 4.27 4.85 0.45 0.29
250-2000 3.89 4.05 1.05 1.00 4.02 5.03 0.49 0.36
106-250 2.98 3.12 3.50 3.24
53-106 1.92 2.14 5.56 5.84
LSD(0.05)t 0.88 1.04 0.36 0.27 1.15 2.09 0.30 0.20
* Significant differences (two-tailed f-test) at the P < 0.05 level between tillage treatments within aggregate size classes.
t Fisher's least significant difference values for comparison of size class means within tillage and soil depth (ANOVA).
BEARE ET AL.: AGGREGATE-PROTECTED ORGANIC MATTER POOLS 791

classes (F3,i8 = 29.9; P = 0.0001), but no significant There were also significant effects of tillage (F1;6 =
effects of tillage treatment (Fi,6 = 0.19; P = 0.68). 46.1; P = 0.0005) and aggregate treatment (F5,3o =
Macroaggregate-protected C made up a significantly 63.3; P = 0.0001) on the total N2O-N emissions from
(P < 0.05) higher percentage of total aggregate C in WSA of the surface samples (Fig. 2). In contrast to the
surface samples of NT than in surface samples of CT net N mineralized, intact >2000-nm WSA had much
(Fi,6 = 40.3; P = 0.0007). There were no significant higher total N2O-N emissions than did the other aggregate
differences between tillage treatments in the 5- to 15-cm treatments and were =5.5 times higher in NT than in
samples. Calculated by difference, aggregate-resistant C CT. Crushing significantly (P < 0.05) reduced emissions
pools composed 95 to 98% of total aggregate C and of NzO-N from the largest macroaggregates of surface
were slightly higher (»l%,P<0.05)in most aggregate soils but did not influence the emissions from the 250-
size classes of NT than of CT (data not shown). to 2000-|j,m WSA. Much lower NzO-N emissions were
The net N mineralized from WSA of the surface recorded from WSA of the 5- to 15-cm soil and there
samples was also significantly affected by tillage (Fi,6 = were no significant differences between tillage treatments
336; P - 0.001) and aggregate treatment (F3,i8 = 22.8; (Fi,6 = 0.15; P = 0.71). As in the surface samples,
P = 0.0001), but there were no significant differences macroaggregates of the deeper samples released more
at depth (Fig. 2). The N mineralized from intact WSA N2O-N than microaggregates and crushing the largest
of the soil surface in NT ranged from 1.4 to 1.7 times aggregates (>2000 urn) reduced NzO-N emissions to
greater than that of CT. Of these intact WSA, only the 9 and 36% of the intact aggregates in CT and NT,
>2000-u.m size class differed significantly (P < 0.05) respectively.
from the other size classes, mineralizing 62 to 71 and Consistent with the results for C, aggregate-unpro-
64 to 82% of the N released (on a sand-free basis) by tected N composed a significantly higher percentage of
the smaller size classes in NT and CT, respectively. total aggregate N in the surface samples of CT than hi
Crushing the macroaggregates of NT significantly (P < surface samples of NT (Fi,6 = 247; P = 0.0001; Table
0.05) increased (35-49%) the net N mineralized com- 2). Although unprotected N made up a higher percentage
pared with intact aggregates but had no significant affect of total aggregate N in all size classes from 5- to 15-cm
in CT. samples of CT than in samples of NT, the overall effects

NET NITROGEN MINERALIZED TOTAL N20-N EMISSIONS

7.0-
0-5 cm I NT LSD (0.05) 0-5 cm • No-Tillage
CT LSD (0.05) 6.0- D Conventional Tillage
0) 5.0- J NT LSD (0.05)
0)
(0 4.0- I CT LSD (0.05)
O>
£ 3.0-
D)
(0 2.0-
d>
0) 1.0-

T^ o.o-
TJ
C 500 (0 1.0
(0 5-15 cm B No-Tillage 5-15 cm NT LSD (0.05)
(0
D Conventional Tillage
400- 0.8-
CT LSD (0.05)
0) I NT LSD (0.05)
300- CT LSD (0.05) 0.6-

>2000 >2000 250- 250- 106- 53- >2000 >2000 250- 250- 106- 53-
2000
Intact Crushed Intact 2000 250 106 Intact Crushed 2000 2000 250 106
Crushed Intact Intact Intact Crushed Intact Intact
AGGREGATE TREATMENTS AGGREGATE TREATMENTS
Fig. 2. The net N mineralized (Nmi*, NHf + [NO2 + NOf ]-N) and total N2O-N emissions from intact and crushed macroaggregates (>2000
and 250-2000 fim) and intact microaggregates (106-250 and 53-106 jim) of the 0- to 5- and 5- to 15-cm depths of no-tillage and conventional-tillage
soils. Bars are means ± 1 standard error; « = 4.
792 SOIL SCI. SOC. AM. J., VOL. 58, MAY-JUNE 1994

Table 3. Standing stocks of aggregate-unprotected C and N and the C/N ratios in water-stable aggregates of conventional-tillage (CT) and
no-tillage (NT) soils at the Horseshoe Bend Experimental Area.
Aggregate-unprotected C Aggregate-unprotected N Aggregate-unprotected C/N
size class NT CT P<0.05 NT CT P<0.05 NT CT P < 0.05
———————— g m-2 —————————————————————
0-5-cm depth
>2000 24.9 13.2 * 2.50 1.45 * 10.0 9.1 «
250-2000 9.6 10.2 1.04 1.09 9.2 9.4
106-250 3.4 3.3 0.38 0.39 8.7 8.5
53-106 0.6 0.7 0.15 0.15 3.7 4.7
Total 38.4 27.4 * 4.07 3.08 * 9.4 8.8
S-lS-cm depth
>2000 22.5 33.3 * 2.77 4.18 * 8.1 8.0
250-2000 9.4 9.3 0.87 0.79 10.8 11.7
106-250 3.5 4.3 0.33 0.38 10.6 11.2
53-106 0.4 0.5 0.12 0.13 3.5 4.0
Total 35.8 47.3 « 4.09 5.48 * 8.8 8.6
0-15-cm depth
Total 74.2 74.7 8.16 8.56 9.1 8.7
* Significant differences (two-tailed t-test) at the P < 0.05 level between tillage treatments within aggregate size classes.

of tillage were not significant. Macroaggregate-protected contrast to unprotected SOM, standing stocks of pro-
N accounted for »1.4 to 1.6% of total aggregate N in tected C and N were consistently and significantly higher
surface samples of NT and was 3.6 to 4.9 times higher hi macroaggregates from surface samples of NT than of
than the protected N in CT samples. Macroaggregate- CT but generally did not differ in deeper samples (Table
protected N was much lower at depth and was unaffected 4). The C/N ratios of protected SOM were significantly
by tillage or size class. lower in 0- to 5- than 5- to 15-cm samples of both
Standing stocks (in grams per square meter) of unpro- tillage treatments. They were also significantly lower hi
tected C and N in >2000-M.m WSA accounted for >60% macroaggregates of NT than of CT at both sample depths
of the total (0-15 cm) unprotected C and N in both and throughout the total profile (0-15 cm).
NT and CT (Table 3). In NT, total standing stocks of
unprotected C were equally distributed between 0- to
5- (51.8%) and 5- to 15-cm (48.2%) depths but, in CT, DISCUSSION
more of the total unprotected SOM was found hi 5- to Results of Beare et al. (1994) showed differences hi
15- (63.3%) than hi 0- to 5-cm (36.7%) soils. Similar the distributions of WSA and aggregate-associated SOM
distributions were found for unprotected N. There were of CT and NT soils that contribute to understanding the
no differences hi the total standing stocks (0-15 cm) of factors that regulate SOM dynamics. Not only were the
unprotected C and N between NT and CT (P > 0.05). largest WSA (>2000 \nm) more abundant in surface
The C/N ratios of unprotected SOM were similar hi NT samples of NT, but they were also more stable and
and CT hi all aggregate size classes except for >2000-nm contained higher concentrations of C and N than did WSA
WSA of 0- to 5-cm samples, which were significantly of CT. Results of this study suggest that the formation and
higher hi NT than hi CT. The smallest microaggregates stabilization of macroaggregates hi NT soils represent
released much more unprotected N per unit of unpro- an important mechanism for the protection and mainte-
tected C than did the larger aggregate size classes. In nance of SOM that may be lost under CT practices. For

Table 4. Standing stocks of aggregate-protected C and N and the C/N ratios in water-stable aggregates of conventional-tillage (CT) and
no-tillage (NT) soils at the Horseshoe Bend Experimental Area.
Aggregate-protected C Aggregate-protected N Aggregate-protected C/N
size class NT CT P<0.05 NT CT P < 0.05 NT CT P < 0.05
__-l

0-5-cm depth
>2000 7.5 0.7 * 1.11 0.09 * 6.8 8.2 *
250-2000 2.8 1.2 * 0.38 0.08 * 7.5 13.6 *
Total 10.3 1.9 * 1.49 0.17 * 6.9 10.9 *
5- 15-cm depth
>2000 4.4 4.4 0.29 0.23 15.2 18.8
250-2000 2.5 2.3 0.15 0.09 16.2 25.5 *
Total 6.9 6.7 0.44 0.32 * 15.7 20.9 *
0-15-cm depth
Total 17.2 8.5 * 1.93 0.49 * 8.9 17.3 *

* Significant differences (two-tailed t-test) at the P < 0.05 level between tillage treatments within aggregate size classes.
BEARE ET AL.: AGGREGATE-PROTECTED ORGANIC MATTER POOLS 793

example, the potentially mineralizable C and N from of macroaggregate-protected N recovered during the 20-d
intact WSA ranged from 21 to 65% higher in surface incubation. In this case, protected N ranged from 6- to
samples of NT than in CT, with the greatest differences 10-fold higher in the surface samples of NT than of CT.
occurring in the macroaggregate size classes. Further- Although macroaggregate-protected C and N amounted
more, when normalized to a sand-free basis, intact mac- to only 0.9 to 1.5% of total aggregate C and N from
roaggregates (>250 u.m) had, on average, greater C and the surface samples of NT, these pools account for ~ 23
N mineralization potentials than did microaggregates and 29%, respectively, of the total mineralizable C and
(<250 urn). These findings are similar to those of Elliott N from macroaggregates. This compares with « 7.5 and
(1986) and Gupta and Germida (1988), who reported 6.5% of potentially mineralizable C and N, respectively,
higher C and N mineralization potentials in intact aggre- hi surface samples of CT.
gates from native sod compared with cultivated soils of Interestingly, our comparisons of mineralization rates
the Great Plains and greater mineralization potentials indicates that macroaggregate-protected pools of C are
from macro- than microaggregates. In contrast, other more labile than the unprotected pools. These differences
studies have reported greater N mineralization potentials probably result from the fact that unprotected pools of
in smaller aggregates (e.g., Craswell et al., 1970; Hat- SOM are more exposed and thus more highly processed
tori, 1973) but, in most cases, the results have not than the protected pools. The relatively large pools of
been normalized to a sand-free basis, making direct unprotected C in aggregates >106 \im in diameter also
comparisons between size classes difficult. had much slower rates of mineralization than did the
Crushing macroaggregates significantly increased the smaller microaggregates (53-106 urn). If larger microag-
mineralization of C and N from NT surface samples but gregates (106-250 nm) are actually formed of frag-
had little effect on the mineralization from CT macroag- mented paniculate organic matter encrusted with micro-
gregates. Elliott (1986) and Gupta and Germida (1988) bial mucilages and clays at the center of macroaggregates,
also found that physical disruption of macroaggregates as suggested by Oades (1984), then unprotected C in
markedly increased the mineralization of C, N, and this size class would be somewhat more occluded and
S from native sods but had much less effect on the thus have slower rates of mineralization, as observed hi
mineralization of SOM from long-term cultivated soils. this study. Furthermore, the smaller microaggregates
Disruption of macroaggregates also markedly reduced (53-106 nm), presumably those released during the de-
N2O emissions in our study. Intact aggregates of NT cay of larger microaggregates, had more labile pools of
surface samples released much more N2O than those of unprotected C, which are probably derived from pre-
CT, particularly the largest aggregates size classes where viously occluded, less humified organic matter (Beare
the concentrations of C (Beare et al., 1994) and NOa et al., 1994). In CT, both the macroaggregate-protected
(data not shown) were greater than hi CT. At least in and unprotected pools of C had generally higher rates
the case of the largest aggregates, disruption significantly of mineralization, indicating that these pools are more
reduced N2O emissions, presumably due to a reduction susceptible to microbial attack than those of NT. It is
in the availability of anaerobic microsites (Sexstone et these largest aggregates (>2000 p,m) that are the least
al., 1985). Although these effects can probably be attrib- stable and, as a result, the most heavily impacted by
uted to less than optimal conditions for denitrification, cultivation (Tisdall and Oades, 1982; Beare et al., 1994).
they may also be the result of a reduction in the N2O In addition to being more labile, unprotected pools
produced during nitrification. In contrast to the results of C and N represented a higher percentage of whole
of this study, Seech and Beauchamp (1988) reported aggregate SOM in surface samples of CT than for the
increasing N2O emissions with decreasing aggregate size same WSA of NT. This observation supports the conclu-
and greater emissions from crushed than intact aggre- sion that intact aggregates of CT maintain a somewhat
gates. Direct comparison with our results is difficult lower capacity to protect SOM than do aggregates of
because their assays were carried out under saturated NT. It is generally believed that microorganisms occupy
conditions and their results were not normalized to a sand- only about 0.1 % of the soil pore volume (Adu and Oades,
free basis. On the other hand, our measures of N2O 1978), limiting their proximity to SOM. If aggregates
emissions are probably underestimates due to the removal remain intact for relatively long periods of tune, the
of water-soluble C during wet sieving. Furthermore, we abundance of clay surfaces associated with small pores
made no attempt to account for other losses of gaseous N. (<l-um diam.) (Sills et al., 1974) may become increas-
When fitted to the first-order model, results of the C ingly occluded with extracellular byproducts of microor-
mineralization assays showed that aggregate-unprotected ganisms that further restrict the access of microorganisms
pools of C were 1.2 to 1.7 times higher in surface and extracellular enzymes to this physically isolated and
samples of NT than of CT, but there were no differences thus protected SOM. However, in CT soils, repeated
between tillage treatments in the unprotected C of deeper mechanical disruption of macroaggregates lowers their
samples. Macroaggregate-protected pools of C were also stability, leaving them more susceptible to the disruptive
much larger in the surface samples of NT than of CT; forces of drying and rewetting or raindrop impact (Adu
however, these pools were approximately one-half to and Oades, 1978; Tisdall and Oades, 1982). These dis-
one-third as large in the deeper samples and there were ruptive forces probably contribute to more frequent turn-
no differences between tillage treatments. Although this over of macroaggregates, exposing more of the physi-
study was not designed to estimate N mineralization cally protected, but relatively labile, SOM to microbial
rates, we did observe similar differences in the quantity attack and mineralization. Consistent with our find-
794 SOIL SCI. SOC. AM. J., VOL. 58, MAY-JUNE 1994

ings, these effects would be expected to reduce the quan- the C/N ratios of whole-soil SOM in fungal-dominated
tity of macroaggregate-protected SOM and increase the soils of NT. Differences in the compositions of soil biotic
quantity of unprotected SOM in proportion to total aggre- communities may be important in determining the size
gate SOM. and quality of aggregate-protected and unprotected pools
The CT soils retained much less macroaggregate- of SOM and may help to further explain the role of
protected SOM than did NT soils. Based on our stand- aggregation in the accumulation and loss of SOM from
ing stock estimates, macroaggregate-protected SOM CT and NT soils.
(>2000- plus 250-2000-u.m WSA) accounted for 18.8
and 19.1% of the total mineralizable C (91.4 g m~2) ACKNOWLEDGMENTS
and N (10.1 g m"2), respectively, in WSA of NT (0- We thank Karen Bushaw, Mary Weise, and Bert Lankester
15 cm) but only 10.2 and 5.4% of the total mineralizable for their field, laboratory, and technical assistance; R.R. Brace,
C (83.2 g m-2) and N (9.1 g nT2) in WSA of CT. D.A. Crossley, Jr., E.T. Elliott, and M. Wander provided
Compared with other size classes, >2000-nm macroag- valuable suggestions for and critiques of our research. Research
gregates accounted for much of the whole soil (0-15 was supported by National Science Foundation Grant BSR-
cm) protected SOM, ranging from 60.0 and 65.3% of 8818302 to die Univ. of Georgia Research Foundation.
the total protected C and N, respectively, in CT to 69.2
and 72.5% of the total protected C and N, respectively,
in NT. Furthermore, nearly all of the difference between
tillages in >2000-nm macroaggregate-protected C (6.8
g m~2) and N (1.02 g m~2) was found in the surface
soils. These differences are most likely attributable to a
somewhat lower abundance and higher turnover of these
largest aggregates (>2000 u,m) in the surface soils of
CT than of NT.
The C/N ratio of total mineralizable SOM (protected
plus unprotected, 0-15 cm) was slightly lower in NT
(9.0) than in CT (9.6); however, the C/N ratios of
whole-soil SOM (Beare et al., 1994) and of the resistant
pools of SOM were higher in NT («11.8) than in CT
(«10.6). Although the C/N ratios of total unprotected
SOM (0-15 cm) did not differ between tillages, the C/
N ratio of total protected SOM was much lower in NT
(8.9) than in CT (17.3). Furthermore, total standing
stocks of protected N reached =1.9 g m~ 2 in NT,
nearly fourfold higher than the protected N of CT. These
findings suggest that macroaggregates play a particularly
important role in the protection of N in NT soils.
The effects of CT and NT management on C and N
dynamics may be attributed to differences in the composi-
tion of the decomposer communities. Previous studies
from this site suggest that saprophytic fungi have a
greater influence on the decomposition of surface-applied
crop residues in NT while bacteria are more important
to the decomposition of incorporated residues in CT
(Hendrix et al., 1986; Beare et al., 1992). More recent
studies (Beare et al., 1993, unpublished data) have also
shown that fungal densities and fungal-mediated binding
of soil macroaggregates are considerably greater in NT
than in CT soils. Similarly, Gupta and Germida (1988)
reported greater populations of fungal-colony-forming
units and fungal hyphae in WSA of native sod than of
long-term (69-yr) cultivated soils, with higher population
densities in macro- than in microaggregate size classes.
It is likely that a fungal-dominated microflora produces
binding agents that differ chemically from those of other
microbial communities, and these differences probably
influence their biodegradability (Aspiras et al., 1971).
Furthermore, fungi have higher C assimilation efficien-
cies and higher C/N ratios than do bacteria (Holland
and Coleman, 1987), factors that would be expected to
lower the C/N ratios of mineralizable SOM and increase
NIELSEN & REVSBECH: COUPLED NITRIFICATION-DENITRIFICATION IN SOIL 795

View publication stats