3 DETECTING AND PREDICTING SOIL ORGANIC CARBON CHANGES: SAMPLING, ANALYSIS, MODELS, AND REMOTE SENSING

Measuring and Monitoring Soil Carbon Sequestration



475



either in one or two dimensions. In principle, systematic sampling would yield more precise results than the first two

sampling schemes (Petersen and Calvin, 1996).

The number of samples (n) to be taken from a field

depends on the variability in SOC content as well as on

minimum difference that needs to be detected. For example,

Izaurralde et al. (1998) used a one-tailed t test to calculate

the number of soil samples needed to detect, with a 90%

confidence, a 0.1% increase in SOC with a known variance of

3.3 (g C kg−1)2. They calculated that for each representative

parcel of land the baseline sampling would require 54 samples, while the final sampling would require another 54 samples, a large number indeed. Similar calculations were carried

out by Garten and Wullschleger (1999), who evaluated the

statistical power to detect significant SOC differences under

switchgrass (Panicum virgatum L.). They calculated the

smallest difference in SOC that could be detected between

two means for a given variance, significance level, statistical

power, and numbers of samples. They concluded that while

differences of about 5 Mg SOC ha−1 were detectable with

reasonable numbers of samples (n = 16) and good statistical

power (1 − β = 0.90), the smallest difference in SOC inventories (1 Mg SOC ha−1) would be detectable only with large

numbers of samples (n > 100).

In order to reduce the number of samples required and

to minimize soil variability, Ellert et al. (2001) proposed a

high-resolution method to detect temporal changes in SOC

storage by comparing the quantities from a sampling microsite (4 × 7 m) at two sampling times separated by periods of

4 to 8 years. In this method, one to six microsites are selected

in such as way so as to represent the dominant soils found in

fields ranging from 30 to 65 ha. Guidance for the location of

the microsites is obtained from experienced pedologists. The

location of each microsite is recorded by survey methods,

including geographic positioning systems. The authors also

made useful recommendations regarding core size and number, time of sampling, depth of sampling, and ancillary measurements. This methodology was successfully applied to the

PSCBP in 1997 to 2000, and allowed for the statistically



© 2005 by Taylor & Francis Group, LLC



476



Izaurralde



significant detection of SOC storage gains as small as 1.2 Mg

ha−1, only 3 years after the implementation of a SCS practice.

Upscaling point measurements of SOC storage to the

field level requires confidence in the assumption that the

properties of the point measurements, including their measurement errors, will hold across the area of prediction. This

confidence has been growing by an increased understanding

of the relationships of soils in the landscape. The spatial

dependence of soil attributes, including SOC content, has been

studied with a variety of techniques or tools, including soil

and topographical surveys, geostatistical techniques,

remotely sensed data interpretation, as well as ground and

monitoring devices. Like other disciplines, soil science has

greatly benefited from advances in computation and information technology (e.g., McBratney et al., 2003). A few examples

of these approaches follow.

Pennock et al. (1987) proposed a segmentation procedure

to describe landscapes into functional units (i.e., landform segments such as shoulder, backslope, footslope, and depression).

Pennock and Corré (2001) used it to study the comparative

effects of cultivation on soil distribution and SOC storage, and

to understand the main landscape features controlling soil

emissions of N2O. This approach was used in the PSCBP to

help delineate the sampling areas for monitoring SOC changes.

MacMillan et al. (2000) expanded on Pennock’s approach and

developed a model, which based on digital elevation models

(DEMs) and fuzzy rules, could identify up to 15 morphologically

defined landform facets. A consolidation in the number of landforms can be obtained to provide units at a farm field scale

that are relevant for benchmark soil testing, application of

simulation models, and precision farming.

Geostatistical methods are being increasingly used to

predict soil attributes. Odeh et al. (1994) compared various

interpolation methods (e.g., multilinear regression, kriging,

co-kriging, and regression kriging) in their ability to predict

soil properties from landform attributes derived from a DEM.

The two regression-kriging procedures tested performed best,

and thus showed promise for predicting sparsely located soil

properties from dense observations of landform attributes



© 2005 by Taylor & Francis Group, LLC



Measuring and Monitoring Soil Carbon Sequestration



477



derived from DEM data. Triantafilis et al. (2001) had success

in using regression kriging to predict soil salinity in cotton

(Gossipium hirsutum L.) fields with electromagnetic induction

data. They attributed the success of the method to the incorporation of regression residuals within the kriging system.

Hengl et al. (2004) tested a framework based on regressionkriging to predict SOM, soil pH, and topsoil depth from 135

soil profile observations from the Croatian national survey.

These research results are promising, as they anticipate the

possibility of implementing these algorithms in a GIS, thus

enabling the interpolation of soil profile data from existing

data sets (Hengl et al., 2004). The challenge remains, however,

of developing rapid methods to accurately estimate SOC

stocks in space and time (including uncertainties) at a relatively low cost.

19.3.2



Bulk Density



Soil bulk density (ρb, Mg m–3) is the ratio of the mass of dry

solids to a bulk volume of soil (Blake and Hartge, 1986). Its

determination is essential to calculate the mass of soil organic

carbon (SOCm, Mg C m–3) from SOC concentration (SOCc, Mg

C Mg–1):

SOCm = SOCc × ρb



(19.1)



Although ρb is a relatively straightforward measurement,

its evaluation can be subject to errors. Blake and Hartge

(1986) and Culley (1993) offer excellent descriptions of the

various methods that can be used to determine ρb. In the

extractive methods, a soil sample of known (core method) or

unknown volume (clod and excavation methods) is extracted,

dried, and weighed (Blake and Hartge, 1986). Bulk density

can also be determined in situ with the use of gamma radiation methods (Blake and Hartge, 1986). Instrument cost and

radiation hazard may limit the utilization of gamma radiation

methods in carbon sequestration projects.

For determination of ρb at various depths, which will be

the case for carbon sequestration projects, Blake and Hartge

(1986) recommend the use of hydraulically driven probes



© 2005 by Taylor & Francis Group, LLC



478



Izaurralde



mounted on pickup trucks, tractors, or other vehicles, but

certainly, hand-driven samplers are appropriate as well. The

obvious goal with any sampling method for determining ρb is

to avoid compressing the soil in the confined space of the

sampler. Challenges are encountered when trying to determine ρb in soils containing coarse fragments, soils with large

swell-shrink capacity, or high organic matter content (Lal and

Kimble, 2001). Each of these challenges must be answered

with specific solutions. Lal and Kimble (2001) briefly review

these and other cases and recommend solutions. For example,

the excavation method might work best for determining ρb in

soils containing significant amounts of coarse fragments or

soils with high organic matter content. The clod method might

be the best method for soils that develop large cracks upon

drying.

Can ρb be estimated by in situ measurements other than

the gamma radiation probe? Time domain reflectometry

(TDR), a technique originally designed for detecting failures

in coaxial transmission lines, was first applied in soil science

to measure soil water content (Topp et al., 1980). Theory and

applications for TDR technology have expanded quickly since

then as a way to measure mass and energy in soil (Topp and

Reynolds, 1998). Ren et al. (2003) used a thermo-TDR probe

to make simultaneous field determinations of soil water content, temperature, electrical conductivity, thermal conductivity, thermal diffusivity, and volumetric heat capacity.

Knowledge of volumetric heat capacity (ρc) and soil water

content (η) further allowed them to calculate other soil physical parameters such as ρb, air-filled porosity, and degree of

saturation. They calculated ρb, as in Ochsner et al. (2001):

ρc – ρ w c w θ

ρ b = ---------------------------cs



(19.2)



where cs is the specific heat capacity of soil solids (kJ kg–1

K–1), ρw is the density of water (kg m–3), and cw is the specific

heat of water (kJ kg–1 K–1). They tested their procedure in the

laboratory with six column-packed soils ranging in texture

from sand to silty clay loam with ρb ranging from 0.85 to 1.52



© 2005 by Taylor & Francis Group, LLC



Measuring and Monitoring Soil Carbon Sequestration



479



Mg m–3. The ρb predicted with the thermo-TDR was able to

explain slightly more than half of the variation in measured

ρb, which suggests a method that, when improved, could

deliver rapid measurements of ρb in the field.

Soil bulk density is a dynamic property; its value changes

in response to applied pressure, soil water content, and SOM

content. Up to a 20% change in ρb can occur with changes in

soil water potential from 0.03 to 1.5 MPa (Lal and Kimble,

2001). Reporting ρb at standardized soil water content of 0.03

MPa is recommended. SOM content has a strong effect on ρb.

Adams (1973) developed an equation to estimate ρb:

100

ρ b = -------------------------------------------------%OM 100 – %OM

------------- + ----------------------------0.244

ρm



(19.3)



where %OM is percent SOM, ρm is mineral bulk density (Mg

m−3), and the value 0.244 is the bulk density of organic matter

(Mg m−3). The bulk density of organic matter is fairly constant.

However, the formula is difficult to apply because ρm is not

usually known. Mann (1986) rearranged Adams’s equation to

calculate ρm from 121 pairs of soil samples with known values

of SOM and ρb.

The Adams equation is difficult to solve directly because

it has two unknowns (ρb, ρm). In principle, ρb could be estimated from knowledge of soil texture, soil particle density

(ρs), and the packing arrangement of mineral particles. Here,

a simple method is proposed for estimation of ρb based on soil

texture, ρs, and packing arrangement information. Soil particle density is usually assumed to be 2.65 Mg m–3, but there

are slight variations depending on the textural composition.

While the sand fraction has a ρs of 2.65 Mg m–3, the clay and

silt fractions have ρs of about 2.78 Mg m–3.

If the fractional values of sand (sandf), silt (siltf), and

clay (clayf) are known, then ρs can be calculated as:

ρs = ρsa × sandf + ρsc × (siltf + clayf)



(19.4)



where ρsa is the soil particle density of sand, while ρsc is the soil

particle density of silt and clay. The next problem is to estimate

a possible arrangement of these particles in the soil matrix.



© 2005 by Taylor & Francis Group, LLC



480



Izaurralde



Assuming a spherical shape for soil particles, there are various

geometrical arrangements in which these particles can accommodate when packed. Sphere packing can be done in two and

three dimensions, but only three-dimensional packing applies

to soils. The densest packing is provided by the cubic close and

the hexagonal close geometries (http://mathworld.wolfram.com/SpherePacking.html). These and other types of packing are defined by the packing density (η), which is the fraction

of a volume filled by a given collection of solids. The packing

density can be solved analytically for some types of arrangements; for others it cannot. For example, η for a cubic lattice

arrangement is 0.524; it is 0.64 for a random arrangement, and

0.74 for a hexagonal close packing arrangement. (See

http://mathworld.wolfram.com/SpherePacking.html for additional information on this topic.)

After selecting a value for η, mineral bulk density can

be estimated as:

ρm = ρs × η



(19.5)



A modified Equation 19.1 is then used to calculate ρb at a

given SOC concentration:

100

ρ b = ------------------------------------------------------------------------------------- ;

SOC × 1.724 100 – SOC × 1.724

------------------------------- + ---------------------------------------------0.244

ρm



(19.6)



0 ≤ SOC ≤ 58 (g C kg−1 × 10−1)

where 1.724 is the conversion factor generally used to convert

SOC into SOM. A theoretical example is shown in Figure 19.1

for three types of sphere packing (cubic lattice, random, and

hexagonal). A test of the model is shown in Figure 19.2

against soil taxonomy data (U.S. Department of Agriculture,

1999). Gupta and Larson (1979) described a model that uses

the same principles of sphere packing described here. This

model is theoretically very good because it accounts for various particle sizes, but it requires complete information on soil

fractions (very coarse sand, coarse sand, medium sand, fine

sand, very fine sand, coarse silt, fine silt, and clay).



© 2005 by Taylor & Francis Group, LLC



Measuring and Monitoring Soil Carbon Sequestration



481



2.0



Soil bulk density (Mg m-3)



1.8

1.6

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0.0

0.0



Cubic

Hexagonal

Random

1.0



2.0



3.0

SOC (g



4.0

kg-1



5.0

x



6.0



7.0



8.0



10-1)



Figure 19.1 Soil bulk density estimated with three packing density models at different soil organic carbon concentrations and constant texture (0.33 clay, 0.33 silt, and 0.34 sand).



19.3.3



Analysis of Soil Organic Carbon



Soils contain carbon in two forms: organic and inorganic.

Organic C is the main constituent of SOM. Inorganic C

appears largely in carbonate minerals. Soil organic C is very

dynamic, intensively reflects management influences, and

exhibits turnover times that range from tens to hundreds of

years (Six and Jastrow, 2002). Soil inorganic C, instead, is

less responsive to management, and has greater turnover

times than SOC. Thus, the emphasis in this section is to

summarize methodologies to determine SOC concentration

and to provide an update on emerging methodologies for quick

and in situ determinations of soil C.

Detailed methodologies of organic C, inorganic C, and

total C are provided by Nelson and Sommers (1996) and

Tiessen and Moir (1993). Basically, SOC concentration can be

determined by either wet or dry combustion. In the wet



© 2005 by Taylor & Francis Group, LLC



482



Izaurralde

2.0



Soil bulk density (Mg m-3)



1.8

1.6

1.4

1.2

1.0

0.8

0.6

0.4



Predicted

Measured



0.2

0.0

0.0



2.0



4.0



6.0



8.0



10.0



12.0



14.0



SOC (g kg-1 x 10-1)



Figure 19.2 Predicted vs. measured soil bulk density using soil

taxonomy data (Typic Haplustert, Typic Haplustalf, Typic Kandiudult, Pachic Argiustoll, Aeric Haplaquox, Typic Dystrudept, Typic

Molliorthel, Eutric Fulvudand).



combustion procedure, a soil sample is treated with acid

dichromate solution in a heated vessel, and then the CO 2

generated due to the oxidation of organic matter is evaluated

either by titrimetric (indirect) or gravimetric (direct) methods.

In titrimetric methods, the amount of organic C present in a

soil sample is obtained by back titration of the unused dichromate with ferrous ammonium sulfate solution. This method

is relatively easy to implement and has been used worldwide

for many years. However, in this method, the digestion of

organic matter is usually incomplete due to insufficient heating. Correction factors have been reported to correct for this

incomplete oxidation, but these factors are soil dependent.

Nelson and Sommers (1996) reported correction factors for 15

studies that averaged 1.24 ± 0.11, or a mean recovery of

81%. Improvements in recoveries are obtained with the wet



© 2005 by Taylor & Francis Group, LLC



Measuring and Monitoring Soil Carbon Sequestration



483



oxidation method, with determination of CO 2 due to the

higher digestion temperatures achieved.

In dry combustion, all forms of C (organic and inorganic,

if present) are converted to CO2 at high temperatures achieved

in resistance (~1000°C) or induction furnaces (>1500°C) (Nelson and Sommers, 1996). Once generated, the CO2 can be then

assessed with a variety of spectrophotometric, volumetric, titrimetric, gravimetric, or conductimetric techniques. Dry combustion methods, instrumented in automated systems capable

or performing multiple elemental analysis (C, H, N, or S), have

become the standard in many laboratories worldwide. They

are very accurate and exhibit minimal variability and low

operational errors. Dry combustion instruments have a detection limit of about 10 mg C kg−1, and their relative deviation

(accuracy) decreases as soil carbon concentration increases.

Figure 19.3 presents a comparison of total soil C (%), as measured by two dry combustion instruments. Because dry combustion determines total carbon, extra steps are required for

reporting organic C concentration when carbonates are

present in the soil sample. If this is the case, then the fraction

of total C that is inorganic can be estimated with either an

independent measurement of carbonate C or the total C analysis conducted on a carbonate-free soil sample previously

treated with an acid solution. Dry combustion should be the

preferred methodology for measuring SOC concentration in

SCS projects. However, due to its relatively high initial cost

(>$20,000), the dry combustion methodology may be difficult

to implement in developing countries participating in SCS

projects and programs.

Assessment of SCS due to the implementation of alternative practices worldwide will require a technical effort to

ensure that the results obtained are accurate and comparable,

and include an estimation of the uncertainty associated with

the measurements. Assessment of these changes will occur

under many environmental conditions, and will have to be

provided at a relatively low cost and may have to include

numerous measurements within a field in order to detect more

continuously the response of SOC to changes in management.

With this in mind, various research groups in the United



© 2005 by Taylor & Francis Group, LLC



484



Izaurralde

3.0



Instrument 2 (g C kg-1 x 10-1)



2.5



2.0



1.5

y = 1.0216x + 0.0342

R2 = 0.9453**, n = 171

1.0



0.5



0.0

0.0



0.5



1.0



1.5



2.0



2.5



3.0



Instrument 1 (g C kg-1 x 10-1)



Figure 19.3 Total soil C as measured by two dry combustion

instruments. (Data from Izaurralde et al. 2001b. Soil Sci. Soc. Am. J.

65:431–441.)



States have been advancing and developing instrumentation

for fast, in situ measurements of soil C. Three methodologies

have been advanced (adapted) so far to measure soil C: (1)

laser-induced breakdown spectroscopy (LIBS) (Cremers et al.,

2001); (2) mid-infrared (MIRS) and near-infrared (NIRS) spectroscopy (McCarty et al., 2002); and (3) inelastic neutron

scattering (INS) (Wielopolski, 2002).

The LIBS method is based on atomic emission spectroscopy (Cremers et al., 2001). In this method, a laser is applied

to a (soil) sample, converting it into plasma that emits light

whose colors are spectrally resolved. Cremers et al. (2001)

calibrated a LIBS instrument that measured total C in soils

from east-central Colorado against measurements with a dry

combustion apparatus, and used the calibration curve



© 2005 by Taylor & Francis Group, LLC



Measuring and Monitoring Soil Carbon Sequestration



485



obtained to predict the total C of additional soil samples. Their

results indicated that LIBS has a detection limit of 300 mg

C kg−1 with a precision of 4% to 5%, and an accuracy ranging

from 3% to 14%. The laboratory version of LIBS tested was

capable of analyzing samples in less than a minute, with a

daily throughput of more than 200 samples. The authors also

reported the development of a field version of the LIBS instrument capable of analyzing soil C over large areas and also in

depth. Martin et al. (2003) also used LIBS to measure total

C and N in samples of soils that had or had not received acid

washing to destroy carbonates. Like Cremers et al. (2001),

Martin et al. (2003) obtained high correlations between total

C measured by LIBS and dry combustion (r2 = 0.962). The

latter team, however, reported increased variability in C

determinations in soils low in organic matter content due to

spectral interference with iron whose peak (248.4 nm) appears

very close to that of carbon (247.9 nm).

NIRS is a widely used technique used to characterize

organic and inorganic compounds in the chemical, pharmaceutical, agricultural, semiconductor, and other industries.

Dalal and Henry (1986) pioneered the use of near-infrared

reflectance spectroscopy to determine water content, organic

C, and total N in soils. Ben-Dor and Banin (1994) used NIRS

to characterize the spectral reflectance of 91 Israeli soils for

several soil properties, including carbonate and organic matter content. Because the NIRS approach is empirical, it

requires the availability of calibration sets to match the spectral characteristics of the sample. They used 39 soils to calibrate the method, and 52 to validate it. Although predicted

and measured SOM values were significantly correlated (r 2 =

0.51) within a range of 0% to 12%, NIRS underpredicted SOM

concentration at the high end. More recently, McCarty and

Reeves (2001) used NIRS and pyrolysis analysis to quantify

SOC content from soils under conventional and no-tillage

management in central Maryland. One objective of the study

was to test whether these methods could be used to understand the spatial structure of SOC distribution in agricultural

fields by sacrificing some accuracy in the point measurements.

Their findings confirmed that NIRS offers a simple and rapid



© 2005 by Taylor & Francis Group, LLC