This presentation was presented during the 3 Parallel session on Theme 1, Monitoring, mapping, measuring, reporting and verification (MRV) of SOC, of the Global Symposium on Soil Organic Carbon that took place in Rome 21-23 March 2017. The presentation was made by Ms. Skye Wills, from USDA - USA, in FAO Hq, Rome
Rapid Carbon Assessment of the U.S.: Project Overview and Summary
1. Rapid Carbon Assessment of US
Rapid Carbon Assessment
of the U.S. : Project
Overview and Summary
Skye Wills
Terry Loecke, Stephen Roecker,
Zamir Libohova, and Soil Survey Staff
March 21 – 23, 2017
Presentation for the Global Soil
Partnership, Global Symposium on
Soil Organic Carbon
2. Rapid Carbon Assessment of US
Objective: Base-line inventory of soil carbon
stocks for conterminous U.S.
Evaluate impact of ecosystems, land use and cover
3. Rapid Carbon Assessment of US
Multi-stage Stratified Random Sampling
► Not enough resources to sample all soil series
► Target sampling
► Stratified Random Sampling
– 17 regions
• (conterminous U.S.)
► Within region Stratification
► 8 – 20 groups of soil series*
► 6 LULC
► 5 types of Land use / cover (LULC): Cropland, Pastureland,
Rangeland, Forest land, Wetland
*Wills et al., 2013
4. Rapid Carbon Assessment of US
RaCA Data Collected
Information collected on
– 6,147 sites
– 32, 053 pedons
– 144,833 samples
5. Rapid Carbon Assessment of US
RaCA Data Collected
Site Information: landscape,
vegetation, management
Pedon Description and sample
collection
– 5 pedons described
– Samples collected by horizon to 1m
– Bulk density samples collected to
50cm
Lab Analysis
– All pedons
• Bulk Density*
• Visible – Near Infrared Spectra
– Central pedon (1 per site)
• Combustion – total carbon
• Calcimeter – inorganic carbon
*Sequiera et al., 2014
13. Rapid Carbon Assessment of US
Conclusions
• RaCA collected 6,000 pedons across the conterminous United
States
• SOC stocks are skewed with a few very high values
• Most SOC occurs within 5 – 30cm
• Wetlands have the highest typical SOC stocks and Rangelands
have the lowest
• A satisfactory map was made across the conterminous US
15. Rapid Carbon Assessment of US
References
– Wills, S., Seybold, C., Chiaretti, J., Sequeira, C., & West, L. (2013).
Quantifying tacit knowledge about soil organic carbon stocks using soil
taxa and official soil series descriptions. Soil Science Society of America
Journal, 77(5), 1711-1723.
– Sequeira, C. H., Wills, S. A., Seybold, C. A., & West, L. T. (2014).
Predicting soil bulk density for incomplete databases. Geoderma, 213,
64-73.
– Soil Survey Staff and T. Loecke. 2016. Rapid Carbon Assessment:
Methodology, Sampling, and Summary. S. Wills (ed.). U.S. Department
of Agriculture, Natural Resources Conservation Service.
– Kern, J. S. (1994). Spatial patterns of soil organic carbon in the
contiguous United States. Soil Science Society of America Journal,
58(2), 439-455.
– Guo, Y., Amundson, R., Gong, P., & Yu, Q. (2006). Quantity and spatial
variability of soil carbon in the conterminous United States. Soil Science
Society of America Journal, 70(2), 590-600.
16. Rapid Carbon Assessment of US
Data Availability
Search engine: RaCA soil
All methodology and data (including R scripts used to
make these figures)
https://go.usa.gov/xXCjD
– Full url:
https://www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/survey/?cid=nrcs142p2_054164
#methodology
R package: soilDB
– fetchraca()
https://go.usa.gov/xXCjD
Each region has its own groups
Soils that occur in multiple regions (most of them) were grouped separately for each region
The group numbers are meaningless outside of a region
The groups were chosen to get a range of soil carbon stock values
The data is not appropriate for producing fine-scale spatially explicit information about soil carbon stocks (no detailed, mu-level maps)
The data only make sense when aggregated
Sampling domain is all of US with SSURGO map in 2010
does not include water and developed (urban) lands
Uses National Resource Inventory as a framework
2 stage design ensures both spatial coverage and randomness
Covers all of US– no information for federal lands
Regional averages follow expected trends: based on known pedogenic relationships between climate and SOC
Using map of soil groups and land use/cover classes produces a relatively detailed map. Some local geographic features are visible such as: Low areas in the sand hills, and high values in norther Minnesota ‘boundary waters’
This plot contrasts the CONUS SOC pool estimates and ranges from Guo et al., 2006 (old) and RaCA (new).
The old/old line (old/old) is Guo median estimate (cross) and min/max range using STATSGO data and
the midpoint approach (old stat approach). The new/old line (orange) is RaCA data estimated using the
midpoint approach (old). The new/new line (blue) is RaCA data estimated using the RaCA sampling design.
The main point here is that RaCA advananced our understanding of the CONUS SOC pool by both providing
greater sampling density and a much better sampling design – which allowed for a hierarchical statistical
analysis instead of the midpoint approach.