Spatial Variability of Soil Organic Carbon: Implications for Detecting Change at Different Scales

1
Spatial Variability of Soil Organic
CarbonImplications for Detecting Change at
Different Scales
Richard T. Conant Keith Paustian Natural
Resource Ecology Laboratory Colorado State
University
2
Needs and Uncertainty

• National vs Project requirements differ
• Uncertainty
• influences international acceptance
• influences market-value of carbon
• affected by spatial and temporal scale
• cost versus uncertainty tradeoffs

3
What are the major challenges to quantifying soil
C sequestration?

• Many controlling factors (e.g. climate,
  vegetation, topography, soil properties,
  drainage, management and landuse history).
• Can exhibit large spatial variability.
• Background amounts are large relative to rate of
  change.

4
Century soil organic matter model
Harvest C removal CGRAIN
Harvest
Standing dead C STDEDC
NPP
Death
Death
Tillage
Fall rate
Tillage
LigninN
Tillage
LigninN
Surface structural C STRUCCC(1)
Surface metabolic C METABC(1)
Belowground Structural C STRUCCC(2)
Belowground Metabolic C METABC(2)
M
M
Lignin M
Lignin M
Surface microbe C SOM1C(1)
Active organic C SOM1C(2)
M
Slow organic C SOM1C(3)
M
Sand M
M
Sand Leaching
M clay
Leached C STREAM(5)
Passive organic C SOM1C(2)
M
5
Controls on soil C

• Management
• Tillage
• Crop rotation
• Fertilization
• Residue management
• Organic amendments

• Abiotic
• Temperature
• Moisture
• Aeration

• Soil properties
• Soil texture
• Clay mineralogy
• Soil depth
• pH

6

• Distribution of soil organic matter in a
  cultivated field
• Spatially variable in visually uniform field
• Large soil C range (0.85-1.93)
• Significant spatial structure
• Substantial fine-scale variability (2-5m)

Data from Robertson et al. (1997)
7
Average Cultivated soil C 35 Mg C ha-1
Accumulation rate 0.5 Mg C ha-1 yr-1
Soil C pool
2 years
25 years
8
US uncultivated grassland and grassland-derived
soils
9
NE Uncultivated grassland-derived soils
(mollisols)
10
Dundy County uncultivated grasslands soils
11
Mean
Coefficient of variation
Standard error
samples
1.78
0.63
1.13
2717
1.39
0.54
0.75
394
0.71
0.39
0.28
15
12
(No Transcript)
13
Measuring and modeling approach
Local, Regional National Estimates
Soil C model
Spatial data
Survey data

• Climate (PRISM)
• Soils (STATSGO)
• Landcover/landuse

• CSRA
• CTIC
• NRI

Field validation
Monitoring

• Long-term field experiments
• USDA pedon database

• On-farm field plots

14
Sampling for projects or monitoring networks

• Sample to maximize sensitivity to changes over
  time
• Must incorporate plot- or field-scale variability
• Sample locations should be georeferenced to
  enable plot relocation
• - Limits incorporation of additional spatial
  variation
• Must measure soil C per unit area
• Can measure indicators (e.g. POM C)
• Field sites should be stratified
• Land use, management, soils, climate, etc.
• Cost vs. accuracy tradeoff

15
Pasture management soil C sampling sites

• Designed for
• Comparative sampling
• Future resampling

• Grassland management
• Grazing
• Fertilization
• Legumes/grass species
• Land use change

• Attributes measured
• Soil C (in/organic)
• Bulk density
• Texture
• POM C

16
Design of microsites

Sample location
5m
Magnetic marker
2m
17
6 sites x 3 cores
18
(No Transcript)
19
Conclusions

• Quantifying soil C presents unique
• challenges, but
• we have an extensive knowledge base on soil
  carbon dynamics and good resource databases
• efficient sampling designs can be developed to
  verify soil C changes at various scales
• integrated modeling and measurement approaches
  offer the best alternative for accurate,
  cost-effective quantification as well as decision
  support capabilities.