112
~
....
I
~ 14
"'0
• 12
....
I
·0 10
C/)
Cl
• 8
()
Cl
6
2:
c
0
4
~
. !::! 2
~
CD
C
0
~
()
0
•
•
• •
50 100 150 200 250 300 350
Days
FIGURE 7.3. Long-term carbon mineralization in a cultivated soil from an agricultural site in southwest Michigan. The solid line is fit with the 2-pool nonlinear model
noted in text. The model provides estimates of the fast
and slow turnover pools; the passive or resistant pool can
be estimated by acid hydrolysis. (From Paul et al.
[1999].)
els (see Fig. 7.1)(Paul et al. 1999, in press). Models
make it possible to correct for the effects of microbial growth and the knowledge that materials in soil
move from one pool to the other. Analytically determined SOM decomposition kinetics greatly assist model improvement and in tum benefits our
empirical understanding and application of SOM
knowledge (Paustian et al. 1992; Paul et al. 1997)
A variety of short-term incubation techniques
have been developed that involve measuring CO2
release from incubated soil over 1- to 24-hour incubation periods. This release provides a quick
snapshot of the quantity of readily mineralizable C
in one soil versus another. Methods range from simple incubations of intact soil cores brought to some
optimal proportion of water-holding capacity in the
laboratory (e.g., Robertson et al. 1988; Collins et
al. 1999) to more elaborate shaken-slurry incubations that-when accompanied by a glucoseamended companion slurry-can provide a relative
measure of microbial biomass (e.g., Anderson and
Domsch 1989; Beare et al. 1990).
In many soils, an initial CO2 flush associated
with soil sampling, mixing, and sieving can distort
estimates of the size of the fast turnover pool, so
G. Philip Robertson and Eldor A. Paul
some methods-especially long-term incubationsemploy a preincubation interval of several days.
Preincubation is not important for within-site comparisons of immediately available SOM C or for
intact-core samplings, but for cross-study comparisons in which samples may be dried, differentially
mixed, sieved, or brought to a new moisture content, preincubation can be important to normalize
for differential sampling effects and for differences
in rainfall and temperature patterns in the days immediately prior to sampling .
Robertson et al. (1999a) recommend a technique
in which unamended soil is brought to the laboratory and, after sieving, is incubated at optimal water
content (approximately 60% water-holding capacity) in an aerobic atmosphere for 7 days. At the end
of 7 days, the soil's respiration potential is assessed
by measuring the rate of CO 2 release over a 2 to 3hour period during which soil head-space is sealed
from the atmosphere. Carbon dioxide is readily
measured by introducing I-ml samples into a
stream of N2 carrier gas connected to an infrared
gas analyzer. This technique can be amended to follow long-term CO 2 release by repeating CO 2 analyses at geometrically increasing intervals (e.g., 14,
28,42,63, 84, 105, 140, 196, and 252 days). Longterm N release can be followed concomitantly by
sacrificing subsamples at each interval for inorganic
N analysis.
Long-term incubations can be used to estimate
multiple SOM pools by fitting the following model
to a graph of CO2 production (e.g., CO 2 -C
g - j day - j or COrC cm - 2 day - j) versus incubation time:
C Mineralizatiol\ = kj(Cje-klt)
+ k2(C2e-k2t) (7.2)
where Cj is carbon content of the active or fast
turnover pool, kj is the rate constant for the C 1 pool,
C2 is carbon content of the intermediate or slow
turnover pool, k2 is the rate constant for the intermediate pool, and t is incubation time in days.
The third, resistant pool can be very roughly estimated as total C (SOM stores, above) less C 1 and
C2. A more accurate estimate of the resistant pool
can be made by using acid hydrolysis to define its
size and by using the 14C age of the hydrolysis residue to estimate the rate constant for this pool (Paul
et al. 1998). Because of the great age of this pool,
an average age of 500 y can be assumed or one can
~
....
I
~ 14
"'0
• 12
....
I
·0 10
C/)
Cl
• 8
()
Cl
6
2:
c
0
4
~
. !::! 2
~
CD
C
0
~
()
0
•
•
• •
50 100 150 200 250 300 350
Days
FIGURE 7.3. Long-term carbon mineralization in a cultivated soil from an agricultural site in southwest Michigan. The solid line is fit with the 2-pool nonlinear model
noted in text. The model provides estimates of the fast
and slow turnover pools; the passive or resistant pool can
be estimated by acid hydrolysis. (From Paul et al.
[1999].)
els (see Fig. 7.1)(Paul et al. 1999, in press). Models
make it possible to correct for the effects of microbial growth and the knowledge that materials in soil
move from one pool to the other. Analytically determined SOM decomposition kinetics greatly assist model improvement and in tum benefits our
empirical understanding and application of SOM
knowledge (Paustian et al. 1992; Paul et al. 1997)
A variety of short-term incubation techniques
have been developed that involve measuring CO2
release from incubated soil over 1- to 24-hour incubation periods. This release provides a quick
snapshot of the quantity of readily mineralizable C
in one soil versus another. Methods range from simple incubations of intact soil cores brought to some
optimal proportion of water-holding capacity in the
laboratory (e.g., Robertson et al. 1988; Collins et
al. 1999) to more elaborate shaken-slurry incubations that-when accompanied by a glucoseamended companion slurry-can provide a relative
measure of microbial biomass (e.g., Anderson and
Domsch 1989; Beare et al. 1990).
In many soils, an initial CO2 flush associated
with soil sampling, mixing, and sieving can distort
estimates of the size of the fast turnover pool, so
G. Philip Robertson and Eldor A. Paul
some methods-especially long-term incubationsemploy a preincubation interval of several days.
Preincubation is not important for within-site comparisons of immediately available SOM C or for
intact-core samplings, but for cross-study comparisons in which samples may be dried, differentially
mixed, sieved, or brought to a new moisture content, preincubation can be important to normalize
for differential sampling effects and for differences
in rainfall and temperature patterns in the days immediately prior to sampling .
Robertson et al. (1999a) recommend a technique
in which unamended soil is brought to the laboratory and, after sieving, is incubated at optimal water
content (approximately 60% water-holding capacity) in an aerobic atmosphere for 7 days. At the end
of 7 days, the soil's respiration potential is assessed
by measuring the rate of CO 2 release over a 2 to 3hour period during which soil head-space is sealed
from the atmosphere. Carbon dioxide is readily
measured by introducing I-ml samples into a
stream of N2 carrier gas connected to an infrared
gas analyzer. This technique can be amended to follow long-term CO 2 release by repeating CO 2 analyses at geometrically increasing intervals (e.g., 14,
28,42,63, 84, 105, 140, 196, and 252 days). Longterm N release can be followed concomitantly by
sacrificing subsamples at each interval for inorganic
N analysis.
Long-term incubations can be used to estimate
multiple SOM pools by fitting the following model
to a graph of CO2 production (e.g., CO 2 -C
g - j day - j or COrC cm - 2 day - j) versus incubation time:
C Mineralizatiol\ = kj(Cje-klt)
+ k2(C2e-k2t) (7.2)
where Cj is carbon content of the active or fast
turnover pool, kj is the rate constant for the C 1 pool,
C2 is carbon content of the intermediate or slow
turnover pool, k2 is the rate constant for the intermediate pool, and t is incubation time in days.
The third, resistant pool can be very roughly estimated as total C (SOM stores, above) less C 1 and
C2. A more accurate estimate of the resistant pool
can be made by using acid hydrolysis to define its
size and by using the 14C age of the hydrolysis residue to estimate the rate constant for this pool (Paul
et al. 1998). Because of the great age of this pool,
an average age of 500 y can be assumed or one can