Methods and Tools for Measurement, Monitoring and Verification for Soil Carbon Sequestration
Markets
Charles W. Rice
Department of Agronomy
1

•
Verifiable and transparent for reporting changes in soil carbon stocks
– (i.e., withstand reasonable scrutiny by an independent third party as to completeness, consistency, and correctness)
•
Cost efficient if soil C will be competitive with other C offsets
•
Based on best science possible
•
Meet requirements that are specified by international conventions
•
Designed to work with data currently available but compatible with different types of data or new methods of data collection
2

•
Provide accounts and associated uncertainties for soil C measurements
•
Flexible to accommodate new scientific developments (e.g., instrumentation, process or empirical models)
•
Reporting structures that are flexible to meet the needs of different users
3


Detecting soil C changes
•
Difficult on short time scales
•
Amount of change small compared to total C

Methods for detecting and projecting soil C changes
(Post et al. 2001)
•
Direct methods
– Field measurements
•
Indirect methods
– Accounting
- Stratified accounting
- Remote sensing
- Models
Post et al. (2001)
4

Level Resolution Cost
Practiced
Based
Producer
Acceptance
Individual
Fields
5

Estimating Changes in Soil Organic Carbon
Issues
 Choice of baseline
 Comparison to current practice
 Start and end time points
• Measure C sequestration or avoided C loss
 Uncertainty and cost in estimating soil C
• Measure and report mean and variation
 Seasonality
 Soil sampling
• Depths
• Roots
• Carbonates
• Rocks
6

40
35
30
25
20
15
0
Conventional
Management
Steady State
Improved Practice
Carbon Sequestering
Practice
Soil Measurement
D
30
O
Practice Change
60 90
Years of Cultivation 7
120
C
B
A
150
7

Sampling strategies: account for variable landscapes
8


Microsites reduces spatial variability

Simple and inexpensive

Used to improve models

Used to adopt new technology
7 m
Sampling location: initial subsequent electromagnetic marker

Soil C changes detected in 3 yr
•
0.71 Mg C ha -1 – semiarid
•
1.25 Mg C ha -1 – subhumid
4 m
Ellert et al. (2001)
9

Methods to Extrapolate Measurements and Model
Predictions from Sites to Regional Scales

Models
•
CENTURY
•
EPIC
•
RothC
•
Other models are also being developed

Spatial aggregation of soil carbon distribution
•
Remote sensing and climatic data
•
Indices:
– Productivity
– Practice monitoring
10

Resources Available for National-level Assessments
• NRCS/STATSGO soil data
• Daily Climate data from NOAA
• 1997 NRI area weights
• NRCS/ERS Cropping Practices Survey
• NRCS/National Soils Laboratory Pedon Database
11

Remote Sensing and Carbon Sequestration


Remote sensing useful for assessing
• Vegetation
– Type
– Cover
– Productivity
• Water, soil temperature
• Tillage intensity?
Remote sensing cannot be used to measure soil C directly unless soil is bare
Crop identification for spatial modeling. Courtesy: P Doraiswamy,
USDA-ARS, Beltsville, MD
12

Several satellite and airborne sensors can estimate LAI,
NPP, crop yields, and litter cover






Traditional sources of land cover data:
• AVHRR and Landsat
Increased resolution being obtained with MODIS
Good temporal resolution
• MODIS and AVHRR
Excellent spatial detail provided by
• Landsat and SPOT
IKONOS and Quickbird offer excellent spatial and temporal resolution
Two airborne sensors
• AVIRIS
• LIDAR
13



CSiTE and CASMGS terrestrial ecosystem models
Century
• Century
• DayCent
• C-STORE
EPIC
• EPIC
• APEX
Processes and drivers
Soil Properties, Management, Weather, CO
2
Soil Processes
Water movement
NH
3
, N
2
O, N
2
Plant Growth
Above Gr. Live
Above Gr. Dead
Below Gr. Live
Below Gr. Dead
Harvest
Erosion
Inorganic
Transformations
Nitrification
NH
3
Volatilization
Denitrification
P i
reactions
Temp & Moisture
Density Changes
Organic
Transformations
Pesticides
Surface residues
Subsoil residues
Humus
CO
2
Leaching
Carbon and nitrogen flows
Metabolic Litter Biomass C Passive C
Residue C
Structural Litter Slow C Leached C
14

In Situ Measurements of
Soil Carbon with Advanced Technologies
R.C. Izaurralde, M.H. Ebinger, J.B. Reeves,
C.W. Rice, L. Wielopolski,
B.A. Francis, R.D. Harris, S. Mitra, A.M.
Thomson, J.D. Etchevers,
K.D. Sayre, A. Rappaport, and B. Govaerts
15





Based on atomic emission spectroscopy
Portable
A laser pulse is focused on a soil sample, creating high temperatures and electric fields that break all chemical bonds and generating a white-hot plasma
The spectrum generated contains atomic emission peaks at wavelengths characteristic of the sample’s constituent elements
Cremers et al. (2001) J. Environ. Qual. 30:2202-2206
16


Mid Infrared / Near Infrared
Spectroscopy (MIR / NIR)
• Non-destructive method measurement of C in soils based on the reflectance spectra of illuminated soil
• Spectral regions
– NIR: 400–2500 nm
– MIR: 2500–25000 nm
• Excellent potential for assessment of spatial distribution of belowground C
MIR and NIR spectra of a calcareous soil before and after treatment with acid for removal of carbonates. Source: McCarty et al. (2002)
17




Conducted at CIMMYT on a 17-year old crop rotation, tillage, residue study
Treatments sampled:
• Maize (m) and wheat (w) grown in monoculture (M) or in rotation (R)
• Grown with conventional (CT) or no tillage (ZT), and with (+) or without (-
) removal of crop residues
• Each treatment is replicated twice
A composite soil sample made of 12 subsamples per soil depth (0-5, 5-10, and 10-20 cm) was taken from each of the 22 x 7.5 m plots.
General view of plots
No Till w/o residues
No Till with residues
18

-2
 Although LIBS and MIRS followed the C density trends detected by DC method
 Correlation between methods was low


LIBS vs. DC: R 2 = 0.174
MIRS vs. DC: R 2 = 0.329


DC LIBS MIRS
1.306
1.440
1.413
0.301
0.393
0.134
Max 2.315
2.300
1.791
Min 0.814
0.600
1.166
Range 1.500
1.700
0.625
n 112 112 112
19




Further calibration of CIMMYT data
Partial Least Squares method was used to improve calibration curves
A calibration curve was developed using 31 samples run 3 times each (1 missing value)
Re-estimation of data points improved significantly (see graph on the right)
2.5
2.0
1.5
1.0
0.5
0.5
1.0
1.5
LIBS y = 1.003x
R
2
= 0.919
2.0
2.5
20

Further calibration of MIRS of CIMMYT data




Original estimation of data using MIRS was developed with the calibration curve based on Maryland samples and 8 samples from Mexico
Eleven samples from the set of 112 were added to the calibration curve
Prediction of the remainder 101 points improved significantly with the revised calibration curve that used the Maryland data points plus the 19
Mexican data points
With the MIRS method, the greatest difficulty in predicting the correct values seems to be associated with high C samples
2.5
2.0
1.5
1.0
0.5
0.5
1.0
1.5
MIRS y = 0.7x + 0.4
R
2
= 0.8
2.0
21
2.5


•
Minimize spatial variability
•
Reduce number of samples
–Decrease costs
–Increase efficiency
•
Increase sensitivity for detecting change
•
Allow adoption of new technology

•
Modeling
•
Remote sensing
22

Chuck Rice
Phone: 785-532-7217
Cell: 785-587-7215 cwrice@ksu.edu

www.soilcarboncenter.k-state.edu/
23