B21F-01 INVITED
Soil Carbon Vulnerability and Black Carbon Stability
The vulnerability of soil organic carbon (SOC) to destabilization is an increasingly important topic in the context of global warming, soil C sequestration, and biofuel development. Many conceptual models of soil C cycling treat chemically recalcitrant SOC as fairly invulnerable to changes in climate and land use. Black carbon (BC), for example, is widely considered to be highly recalcitrant in soils and a possible avenue for long term CO2 sequestration in soil. We estimated in situ turnover time of BC, using (1) BC and radiocarbon content of soil samples collected 100 years apart from the same Russian Steppe soil and (2) the rate of decline in BC stock after cessation of biomass burning and associated BC inputs. BC stocks declined 25 percent between 1900 and 1997, from 2.5 kg m-2 in to 1.9 kg m-2. The implied BC turnover time is about 300 years, much faster than so-called inert or passive C in ecosystem models. Moreover, this was faster turnover than other SOC components had in the same soil. The BC was not homogeneous, however. While the amount of less-condensed BC structures decreased, highly condensed structures persisted. Although BC was not recalcitrant in this soil, it appears to be persistent elsewhere. We conclude that there may be no universally recalcitrant or stable SOC, but rather the dominant mechanisms of SOC stabilization and destabilization vary according to environment and disturbance, and soil carbon cycling is best understood viewing soils as distinct ecosystems.
B21F-02
Using Cs-137, C-14 and biomarker compounds to identify reasons for C and N losses in resampled profiles
A New Zealand data set of archived and resampled pasture soil profiles has identified a systematic pattern
large soil C and N losses and gains that appear to be related to land-use intensity. We use isotope and
organic geochemistry techniques in selected archived and resampled soil horizons to identify reasons for the
observed large soil C and N losses and gains in intensive flat non-allophanic pasture and hill country soil
profiles, respectively. These techniques allow us to examine 3 of the ~10 hypotheses proposed to explain the
large losses initially observed in intensive pasture soils. These three hypotheses are: (1) soil C and N
changes may be due to erosion and deposition; (2) pre-European forest-derived organic matter is being lost;
and (3) changes in litter quality are reducing the amount of plant C and N stabilized in soil.
To test hypothesis (1), we use 137Cs, accumulated in the soil clay fraction from nuclear fallout between
1945 and 1965. Measurements comparing archived (post-1965) and resampled horizons show losses or
gains of 137Cs, which we interpret as erosion and deposition, respectively. Apparent wind erosion of up
to ~6 cm of surface soil explains large surface soil C losses in 2 flat profiles, while apparent deposition
explains soil C gains in two hill country profiles.
Measurements of 14C assist in the evaluation of hypothesis (2) by suggesting that, after accounting for
137Cs-estimated erosion or deposition, surface soils are mainly losing C fixed since bomb 14C was
injected into the atmosphere (post-1950). In contrast, soil C losses below 40 cm depth are dominated by C
derived from pre-European forests.
Biomarker compounds, particularly lignin-derivatives, allow us to evaluate hypotheses (2) and (3). Results to
date suggest that failure to stabilize grass-derived C is more important than losses of forest-derived C in
explaining soil C losses in the upper 30 cm. More broadly, biomarker and 137Cs measurements suggest
that steady-state assumptions must be carefully applied in models of C and N in New Zealand pasture soils,
and will be inappropriate in some circumstances. Based on these studies of 3–5 selected profiles, we
examine approaches to broaden the use of these techniques to identify reasons for C and N losses and
gains as the resampling of archived New Zealand soil profiles continues.
http://www.gns.cri.nz/who/staff/2234.html
B21F-03
Soil Organic Matter Mean Residence Time Measurement: Characterization of Analyitical and Numerical Approaches for the Bombspike Model Resolution
Anthropogenic actions, fossil fuel use and land use change, are drastically altering the global carbon cycle.
CO2 in the atmosphere, the most sensitive compartment of the entire cycle, has never been exceeded
in the last 650 ka. Terrestrial ecosystems preserve in the soil organic matter(SOM) 1600 PgC, an amount of
C twice with respect to the atmosphere. The amounts of carbon in SOM, its exchange fluxes with the
atmosphere, the observed sequestering times(ranging from years to centuries)and the possibility of
regulation of C in future land management, has led the scientific community to look at a more precise soil C
cycle. Recent studies pointed out how, because of the multi compartment nature of soils, SOM cannot be
considered as a homogeneous reservoir. The coexistence of several phases in the soil compartment is
evidenced by the presence of different SOM carbon pools (fractions) each one characterized by a
homogeneous C mean residence time. Considering SOM as a composite compartment drastically increases
difficulties in the methodological approach to the study of these pools resulting in complex feedback
responses to changing climate conditions. Atmospheric tests of nuclear weapons (banned in 1963 after the
test ban treaty) enriched atmospheric 14CO2 doubling its background levels. The signal
(bombspike) decreases over time, with an exponential trend (annual rate of 0.4%), because of the net
uptake of CO2 from oceans and vegetation; and the dilution caused by fossil CO2 releases.
Observed bombspike decrease rate allows dating of atmospheric CO2 with a precision of ± 1 year.
Once globally distributed in the atmosphere the bombcarbon signal becomes a SOM marker allowing a
sensitive mean residence time measurement of C by means of dynamic models simulating its fate in the soil
(bombspike models). Bombspike models rely on the SOM box model, in which first derivative of soil C content
over time is expressed by the net balance between net inputs of C and outputs mainly due to decomposition.
Decomposition flux is related to the carbon content by means of the decomposition constant (k) implying a
first degree kinetic relationship among outputs and C content. SOM C over time can be analytically
expressed only under some particular edge conditions (i.e. constant input scenario). Combining the bomb
spike properties (i.e. each annual 14C input) and the box model it is possible to model SOM 14C
content as the sum of a contribution of each annual input and hence express SOM 14C/12C ratio.
Annual radiocarbon input over time are not constant (because of the bomb spike) hence the constant input
scenario is not verifiable for 14C and model predicted SOM 12C are not be free of bias.The main
goal of this contribution is to quantify the biases introduced during the analytical resolution of the bombspike
model by comparing it with the alternative numerical approach. The same results will be shown for the
resolution of more complex model simulating the relationships occurring among fractions as described by
different soil fractionation protocols (i.e. cascade three boxes model).
http://www.agu.org/marzaioli
B21F-04
Characterising soil surface condition and carbon vulnerability using spatial statistics and directional reflectance
Soils can experience rapid structural degradation in response to land cover changes, resulting in reduced soil productivity, increased erodibility and a loss of Soil Organic Matter (SOM). The breakdown of soil aggregates through slaking and raindrop impact is linked to soil organic matter turnover, with subsequently eroded material often displaying proportionally more SOM. A reduction in aggregate stability is reflected in a decline in soil surface roughness, indicating that a physical soil structural change can be used to highlight soil vulnerability to SOM loss through mineralisation or erosion. Remotely sensed data can provide a cost- effective means of monitoring changes in soil surface condition over broad spatial extents. Growing recognition of the importance of the directional reflectance domain has highlighted their potential application for monitoring changes in soil surface roughness, associated with the breakdown of macro-aggregates and therefore SOM release. This is particularly relevant for soil condition monitoring because during soil structural degradation, changes in the self-shadowing effects of soil aggregates has a measurable effect on directional reflectance factors measured by proximal remote sensing devices. Field and laboratory data are therefore required for an empirical understanding of soil directional reflectance, underpinning subsequent model development. This experiment details the use of hyperspectral multiple view angle, proximal reflectance data (400-2500 nm) for describing changes in soil surface structure. Five different soil crusting states were produced, simulating a progressive decline in soil surface structure using artificial rainfall. Each stage was characterised using a close-range laser scanning device with a 2 mm spatial sampling methodology. Data were analysed within a geostatistical framework, where variogram analysis quantitatively confirmed the change in soil surface structure during crusting (sill variance = 0.350 (control); 0.274 (crusted soil)). Each was measured using an ASD FieldSpec Pro, fitted with an 8° foreoptic and attached to an A frame device which permitted measurement of directional reflectance factors at 5° sampling angles in the solar principal plane. The measurement angles (θr) used were in the range -60° to +60°, with illumination angles (θi) in the range 28° to 74°. Reflectance measurements were compared to geostatistically-derived indicators of surface roughness, derived from laser profile data. The results showed a strong relationship between directional measurements and surface roughness (R2 = 0.94; θr = -60°, θi = 67°-74°). The results provide an empirical and theoretical basis for the future retrieval of coarser spatial scale, distributed assessments of soil surface structure, and therefore, important information on carbon turnover and vulnerability in a landscape context.
B21F-05 INVITED
A Resource Assessment Approach to Carbon Sequestration: Implications for Soil Carbon Science and Assessment
Carbon sequestration has emerged as an important concern in decisions affecting land, water, and ecosystem resources. There is a need for comprehensive carbon sequestration resource assessments that address (1) the full range of sequestration options (biological, geological, and oceanic sequestration; and actions that affect the vulnerability of existing carbon storage); (2) both potential rates and potential capacities of carbon storage; and (3) the broad information needs of managers and decision makers who are responsible for the array of both carbon and non-carbon resources that are involved in carbon sequestration decisions. The need for this comprehensive approach is well illustrated by the fact that biological carbon sequestration is most often evaluated in terms of potential rates, whereas geological sequestration is most often evaluated in terms of potential capacities. Non-specialists have no basis for comparing these two commonly used sequestration metrics. A further difficulty is that carbon sequestration options are often described in a context of advocacy for particular choices. Thus, non-specialists may be reluctant to make decisions based simply on reported information without further evaluation. Carbon sequestration resource assessment inherently requires evaluating the potential future availability and vulnerability of carbon storage. Soil carbon sequestration assessment is particularly challenging. The natural heterogeneity of soil properties makes soil carbon monitoring and accounting extremely difficult, especially given the need for resource assessments that are scalable from local projects to regional, national, and global estimates. Basic research is needed to understand the dynamic interactive processes that affect potential rates and capacities of soil carbon storage, including its permanence in both managed and unmanaged settings. Economic analysis is necessary to anticipate relevant costs, benefits, risks, and tradeoffs that are involved in decisions that affect soil carbon storage. Uncertainties of potential effects due to climate and land-use change are large and difficult to quantify, requiring innovative scenario-based evaluation. Comprehensive carbon sequestration resource assessments are needed to evaluate soil carbon sequestration in the context of other sequestration options and other resource priorities.
B21F-06 INVITED
Monitoring Soil C Stocks and Turnover in Agricultural Lands
Soils are a large reservoir of carbon in the biosphere, and have the potential to act as a sink or source of carbon to the atmosphere, depending on a variety of driving variables such as land use, management, and climate change. Given the potential for carbon change in the future, modeling and measurement approaches are needed to monitor C stocks and turnover in soils, with a reasonable level of accuracy and precision for informing decision makers. We have developed an approach combining simulation modeling with input driving data to monitor soil organic C stocks in agricultural lands. The approach is applied using the USDA National Resource Inventory providing land use and management data for the past few decades at about 400,000 points locations in the conterminous US. Uncertainties are addressed using a combination of a Monte Carlo simulation approach and an empirically-based method comparing model results to measurements. In the past decade, soil organic C stock change has ranged from 15 to 18 Tg C per year for the conterminous US. Uncertainties range from several 100 percent at local NRI sites to 20 percent at the national scale. Reducing uncertainties will depend on model improvements, but also depend on an expanded network of measurements to evaluate uncertainties in the model results. Moreover, analyses suggest that the network should include at least 3000 locations to minimize uncertainties in the soil organic C change estimates. Ultimately, modeling along with a measurement network to assess uncertainties can provide the framework and confidence in results to support policy and management decisions.
B21F-07 INVITED
Creating an Effective Network: The GRACEnet Example
Networking activities require time, work, and nurturing. The objective of this presentation is to share the experience gained from The Greenhouse gas Reduction through Agricultural Carbon Enhancement network (GRACEnet). GRACEnet, formally established in 2005 by the ARS/USDA, resulted from workshops, teleconferences, and other activities beginning in at least 2002. Critical factors for its formation were to develop and formalize a common vision, goals, and objectives, which was accomplished in a 2005 workshop. The 4-person steering committee (now 5) was charged with coordinating the part-time (0.05- to 0.5 SY/location) efforts across 30 ARS locations to develop four products; (1) a national database, (2) regional/national guidelines of management practices, (3) computer models, and (4) "state-of-knowledge" summary publications. All locations are asked to contribute to the database from their field studies. Communication with everyone and periodic meeting are extremely important. Required to populate the database has to be a common vision of sharing, format, and trust. Based upon the e-mail list, GRACEnet has expanded from about 30 to now nearly 70 participants. Annual reports and a new website help facilitate this activity.
B21F-08
Development and Application of the GRACEnet Database
More carbon is stored in soil than in the atmosphere. This reservoir is vulnerable to change, and can be a source or sink of atmospheric CO2. It is difficult to precisely quantify the impacts of land management on soil C levels because weather, biological community and other factors act as controls on soil C. Consequently, simulation models that represent the control mechanisms on soil C (plant growth, organic matter decomposition, etc.) have been developed. Field measurements of soil C have been limited because sampling methods are not uniform and model development is constrained by the availability of model input and testing data. A primary goal of the GRACEnet project is to provide a database for soil C data collected from various agricultural systems across the US using standard measurement protocols. The database also includes measurements of nitrous oxide, methane, crop yields, and input data needed to develop and test ecosystem models. Researchers contributing to the database are required to adhere to prescribed data reporting formats. Database users can access complete data sets required to drive and test models and perform queries to extract data subsets. The database has been used to validate and improve ecosystem models. These improved models are currently being used to calculate CO2 and N2O fluxes from agricultural soils for the US National Greenhouse Gas Inventory and to quantify the greenhouse gas mitigation potential of different land management strategies.