110
rocal transplants can provide a valuable approach
for evaluating differences in substrate quality as an
explanation for decomposition rate differences
among divergent ecosystems.
Substituting a standard substrate for native litter
provides a similar and logistically simple means for
evaluating substrate quality versus environment
differences. In this technique (e.g., Binkley 1984;
Sinsabaugh et al. 1993; O'Lear et al. 1996) a single
organic substrate, usually one that is manufactured
and thus commercially available, is placed in litter
bags or buried across a range of two or more ecosystems. Studies have employed cellulose filter paper and hardwood dowels as substrate standards.
Both filter paper and dowels vary somewhat in their
substrate composition, however-filter paper can
contain small but varying amounts of nitrogen, and
dowels can contain different amounts of decayresistant heartwood-so it is necessary to select
materials carefully. Because of substrate differences, results may still be interpretable only within
the scope of a particular study. Harmon et al. (1999)
recommend dowels be made of basswood (Tilia
spp.), birch (Betula spp.), or ramin (Gonystylus
bancannus), 6 mm in diameter by 60 cm in length.
Placement, recovery, and analysis of standardized
substrates follow the same procedures as for fine
and coarse woody litter, above.
Soil Organic Matter Dynamics
Soil Organic Matter Stores
Soil organic matter is defined operationally as C
and associated nutrients entrained in soil mineral
horizons, in forms no longer easily identifiable as
plant residues. The 0 horizon or litter layer is by
convention excluded from SOM stores, because it
can be identified as specific plant residues that in
modeling are handled by separate equations. In
most soils, the mass of C in mineral horizons far
outweighs the mass of C in the litter horizon. Soil,
however, includes the entire solum, and especially
in ecosystem studies it is important to include C in
horizons both above and below the more commonly
analyzed A horizon when calculating total ecosystem SOM stores. This may mean sampling to a meter or more in many soils. Sites in which significant
decomposition occurs in the tree or grass canopy
G. Philip Robertson and Eldor A. Paul
present a special challenge, as this material also is
part of ecosystem decomposition.
Litter stores are easily evaluated by collecting
litter from known-area quadrants (e.g., 1 m 2 ) and
analyzing for C and nutrients as for leaf and woody
material, above. In most ecosystems, stores will
vary by season and collections should be staggered
accordingly. Stores of organic C in mineral soil are
usually evaluated by whole-soil combustion. Any
inorganic C that may be present is either removed
beforehand or analyzed separately and subtracted
from combustion results. Inorganic C can be a major fraction of total C in calcareous soils low in
SOM; this is often the case for soils in arid and
semiarid environments, for young soils developed
on carbonate parent materials, and for agricultural
soils limed for pH control.
Modem C and N (CN) analyzers oxidize small
samples at high temperatures (> 1000°C) in an O 2 -
enriched atmosphere, then measure the resulting
gases by gas chromatography (C02 and N gases)
or infrared gas analysis (C02 only) (Nelson and
Sommers 1996). Small sample size (usually <20
mg) means that great care must be taken to adequately subsample and powder the soil to be analyzed; inadequate grinding is the leading source of
analytical variability for most soils (Sollins et al.
1999). Conversion of gravimetric results (e.g.,
g C/g SOil-I) to an areal extent (e.g., g C m -2) is
crucial for including SOM C in total ecosystem carbon budgets and for making comparisons of SOM
across ecosystems with different soil types and
depths. It is thus critically important to have accurate values for both soil bulk density (g soil cm - 3),
stone content, and soil horizon depths. Bulk density
values are needed to convert from gravimetric values to areal values, stone volume must be excluded
from the conversion, and estimates must be
weighted by horizon depths, including the depth of
the 0 or litter layer horizon (Sollins et al. 1999).
Alternately, single soil cores taken to a depth of
minimal soil C content can be homogenized and
analyzed for total C content, and this total-core
value then divided by the sampled stone-free surface area to yield an areal estimate for SOM C.
Physical Fractionation of Soil
Organic Matter
The two most promising strategies for physically
separating SOM into differentially reactive fractions are fractionations based on density and size
rocal transplants can provide a valuable approach
for evaluating differences in substrate quality as an
explanation for decomposition rate differences
among divergent ecosystems.
Substituting a standard substrate for native litter
provides a similar and logistically simple means for
evaluating substrate quality versus environment
differences. In this technique (e.g., Binkley 1984;
Sinsabaugh et al. 1993; O'Lear et al. 1996) a single
organic substrate, usually one that is manufactured
and thus commercially available, is placed in litter
bags or buried across a range of two or more ecosystems. Studies have employed cellulose filter paper and hardwood dowels as substrate standards.
Both filter paper and dowels vary somewhat in their
substrate composition, however-filter paper can
contain small but varying amounts of nitrogen, and
dowels can contain different amounts of decayresistant heartwood-so it is necessary to select
materials carefully. Because of substrate differences, results may still be interpretable only within
the scope of a particular study. Harmon et al. (1999)
recommend dowels be made of basswood (Tilia
spp.), birch (Betula spp.), or ramin (Gonystylus
bancannus), 6 mm in diameter by 60 cm in length.
Placement, recovery, and analysis of standardized
substrates follow the same procedures as for fine
and coarse woody litter, above.
Soil Organic Matter Dynamics
Soil Organic Matter Stores
Soil organic matter is defined operationally as C
and associated nutrients entrained in soil mineral
horizons, in forms no longer easily identifiable as
plant residues. The 0 horizon or litter layer is by
convention excluded from SOM stores, because it
can be identified as specific plant residues that in
modeling are handled by separate equations. In
most soils, the mass of C in mineral horizons far
outweighs the mass of C in the litter horizon. Soil,
however, includes the entire solum, and especially
in ecosystem studies it is important to include C in
horizons both above and below the more commonly
analyzed A horizon when calculating total ecosystem SOM stores. This may mean sampling to a meter or more in many soils. Sites in which significant
decomposition occurs in the tree or grass canopy
G. Philip Robertson and Eldor A. Paul
present a special challenge, as this material also is
part of ecosystem decomposition.
Litter stores are easily evaluated by collecting
litter from known-area quadrants (e.g., 1 m 2 ) and
analyzing for C and nutrients as for leaf and woody
material, above. In most ecosystems, stores will
vary by season and collections should be staggered
accordingly. Stores of organic C in mineral soil are
usually evaluated by whole-soil combustion. Any
inorganic C that may be present is either removed
beforehand or analyzed separately and subtracted
from combustion results. Inorganic C can be a major fraction of total C in calcareous soils low in
SOM; this is often the case for soils in arid and
semiarid environments, for young soils developed
on carbonate parent materials, and for agricultural
soils limed for pH control.
Modem C and N (CN) analyzers oxidize small
samples at high temperatures (> 1000°C) in an O 2 -
enriched atmosphere, then measure the resulting
gases by gas chromatography (C02 and N gases)
or infrared gas analysis (C02 only) (Nelson and
Sommers 1996). Small sample size (usually <20
mg) means that great care must be taken to adequately subsample and powder the soil to be analyzed; inadequate grinding is the leading source of
analytical variability for most soils (Sollins et al.
1999). Conversion of gravimetric results (e.g.,
g C/g SOil-I) to an areal extent (e.g., g C m -2) is
crucial for including SOM C in total ecosystem carbon budgets and for making comparisons of SOM
across ecosystems with different soil types and
depths. It is thus critically important to have accurate values for both soil bulk density (g soil cm - 3),
stone content, and soil horizon depths. Bulk density
values are needed to convert from gravimetric values to areal values, stone volume must be excluded
from the conversion, and estimates must be
weighted by horizon depths, including the depth of
the 0 or litter layer horizon (Sollins et al. 1999).
Alternately, single soil cores taken to a depth of
minimal soil C content can be homogenized and
analyzed for total C content, and this total-core
value then divided by the sampled stone-free surface area to yield an areal estimate for SOM C.
Physical Fractionation of Soil
Organic Matter
The two most promising strategies for physically
separating SOM into differentially reactive fractions are fractionations based on density and size