Australian soil carbon stocks: a summary of the Australian Soil Carbon Research Program
the Australian Soil Carbon Research Program
LAND AND WATER/SUSTAINABLE AGRICULTURE FLAGSHIP
Soil Carbon Research Program (SCaRP)
Objectives
 Apply consistent methodology to quantify soil carbon stocks
 Assess MIR as a rapid and cost‐effective means for quantifying soil carbon stocks and composition
k
d
 Test automated devises for measuring soil bulk density
 Quantify soil carbon stocks under different land management strategies at regional levels
 Provide temporal soil carbon stock data for FullCAM/NGGI development
Q
y
p
p
p
y
 Quantify the inputs of carbon to soils under perennial pasture systems
Issues beyond the scope of SCaRP
• The sampling methodology used in SCaRP was not appropriate to: The sampling methodology used in SCaRP was not appropriate to:
• quantify soil carbon stocks in a paddock
• quantify rates or amounts of carbon sequestration in Australian soils
Background ‐ Composition of soil carbon and definitions
Soil carbon ≤2 mm OC = organic carbon
IC = inorganic carbon
TC = Total carbon = (OC + IC)
Carbon accounting is focused on OC, may Australian soils contain IC
Soil organic carbon ≤2 mm – complex mixture of many materials
POC = particulate organic carbon (2000–50 µm excluding charcoal)
HOC = humus organic carbon (≤50 µm excluding charcoal)
ROC = resistant organic carbon (≤2000 µm charcoal)
Why are we interested in fractions of soil carbon?
• Quantifying fractions is not a requirement for soil carbon accounting
• Provides data for initialising and calibrating FullCAM ‐ used in the NGGI
• Provides an indication of the vulnerability of soil carbon to subsequent change POC
HOC
ROC
Vulnerability (V)
V
POC
HOC  ROC
V
POC
HOC
Composition of POC and HOC
Index of extent of decomposition
(Alkyl C/O‐Alkyl C)
POC
HOC
Sampling locations and soil samples collected
Samples collected and analysed by SCaRP
‐ 17,721 samples
17 721 samples
‐ 3,836 sites
Additional samples
Additi
l
l
‐ 2774 samples
‐ 690 sites
Totals
20 495 samples
‐ 20,495 samples
‐ 4,526 sites
>92% from farmer paddocks
Soil sampling methodology
Some variance in the process used to select sampling locations
Some variance in the process used to select sampling locations
Once the locations of sampling sites were defined – everything was consistent from that point
p
Sampling site
Processing and Analyses
Air dry
Crush
Sieve ≤2 mm
Split ≤2 mm subsamples using riffle boxes
iffl b
Fine grind
– TC, OC, IC, TN and MIR
– Fractions
– POC (2000‐50µm)
– HOC (<50µm)
– ROC (<2000µm)
(
µ )
Variations in 0‐30 cm soil carbon stocks 0.20
All sites
0 16
0.16
0.14
0 8
0.18
0.08
0.16
0.08
0 06
0.06
0.04
0.02
Soil OC stocks (Mg C/ha)
Soil OC stocks (Mg C/ha)
100
95
90
85
80
75
70
65
60
55
50
45
40
30
25
0 00
0.00
20
0.00
0.10
Soil OC stocks (Mg C/ha)
15
0.02
Central Tablelands
NSW
0 12
0.12
10
0.04
0.14
5
0.06
0
0
0.00
0.10
40
0
0
60
0
5
80
0
10
Relat
tive freque
ncy
100
0
15
20
120
0
25
140
0
30
160
0
35
180
0
40
200
0
45
220
0
50
240
0
55
60
260
0
65
280
0
70
300
0
75
80
85
90
95
100
0.05
NSW sites
0.12
20
0
0.10
35
0.15
Relaative frequeency
Relative frequ
uency
0.25
Variations in 0‐30 cm soil carbon stocks
0.30
Relative frequencyy
0.25
Central Tablelands, NSW
0.20
0.15
Central Slopes, NSW
Central Plains, NSW
l l i
0.10
0.05
0.00
10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100
Soil OC stocks (Mg C/ha)
Creation of cumulative probability distributions
Cumulative prob
bability (%)
100
Central Tablelands
NSW
80
60
40% of the time <38 Mg C/ha
60% of the time >38 Mg C/ha
40
20
0
0
20
40
60
80
Soil OC stocks (Mg C/ha)
100
120
Assessing the ability of MIR/PLSR to predict soil C and its allocation to different forms
MIR as rapid method of analysis for soil organic carbon
MIR as rapid method of analysis for TC, OC, IC and TN
TC
IC
OC
TN
MIR as rapid method of analysis for soil OC fractions
OC
HOC
POC
ROC
Rapid assessment of bulk density – gamma radiometrics
WA Project – soil surface
•Impact of NDM vs cores on soil C stocks small
•Bulk density accounted for <4% of variation in C stock
CSIRO – core scanner
•Required correction for water content
•NIR predicted water contents sufficed
•Potential to take this technology to the field
Soil carbon under perennial pasture : C3/C4 transition
-1
SOC (M
Mg C ha )
50
SA - FP + KI
WA - ND
WA - SD
40
30
Regional SOC sequestration rates:
Regional SOC sequestration rates:
• 0.9 ± 0.3 Mg C/ha/y for kikuyu in WA
20
p
/
• No trend for panic/Rhodes in WA
10
• 0.3 ± 0.1 Mg C/ha/y for kikuyu in SA
0
-10
-20
20
-1
C4
4-C (Mg C ha
a )
25
•
Kikuyu more responsive than pastures with a mix of panic/Rhodes grasses •
Approximately 80% of change due to Approximately
80% of change due to
C4‐SOC
20
15
10
5
0
0
10
20
30
40
Perennial pasture age (yr)
50
Input of carbon to soil under kikuyu perennial pasture: fate of 14C CO2‐C
Soil
Depth
Time
since
labelling 0
14
% of C CO2 fed to pasture
5
10
15
10 d
days
20
25
2 1
25.1
62
6.2
17.9
6 wks
30
6.4
35
11
1.1
1.2
0-10cm
1 year
7.2
7.5
1.2
>2mm Roots
POM fraction
10 days
y
10-20cm
<50µm fraction (humus)
6 wks
1 year
10 days
20-30cm
6 wks
Annual addition: 1 year
30-50cm
POC + HOC = 1052 kg/ha
10 days
>2mm Roots
6 wks
ks
POM fraction
1 year
10 days
50-70cm
6 wks
1 year
HOC = 160 kg/ha
<50µm fraction (humus)
Murray CMA – soils of the slopes and plains
• Sl
Slopes: introduced and voluntary i t d d d l t
pastures contained more carbon than cropping
• Si
Significant variations were confined to 0‐
ifi t i ti
fi d t 0
10cm
• Increasing trend with rainfall, decreasing t d ith A O t t
trend with Apr‐Oct temperature
t
NSW – Northwest slopes and plains
• SOC stocks higher in pastures than cropping
• Tillage differences were not significant
• Pasture type – may be an impact of land use history
• Natural gradients are important
g
p
• Correlations were found with soil type, clay content, rainfall, temperature and elevation
Queensland cropping lands
Hermitage long‐term trial (1981‐2008)
• No evidence that no‐till or stubble retention can increase soil carbon stocks
• NT and stubble retention decreased extent of loss for 0‐
10 cm layer but not 0‐30 cm layer
Queensland rangeland systems
• Non linear trends with increasing shoot dry matter production, NDVI, and annual rain for 0‐10 cm SOC stocks
• Soil type was also important
• Potential for remote sensing to provide key information
Potential for remote sensing to provide key information
60
50
40
30
20
0
10
0
Mid‐North
Eyre Peninsula
cropped crop‐pasture
pasture
70
0‐30
0 cm, tonnes/hectaree
Soil orgganic carbon sttock
70
(0‐30
0 cm, tonnes/hectaree)
Soil orgganic carbon sttock South Australia – dry land cropping soils
60
50
40
30
20
10
0
350
400
450
500
Rainfall zone, < mm
600
Victoria – soil carbon in cropping and pasture systems
Location
Management practices
Duration
(y)
SOC difference
((Mg OC/ha)
g / )
Hamilton
P rate and grazing intensity
27
nsd
Ararat
Gra ing management
Grazing management
5
nsd
Horsham, LR1
Cropping systems
94
≤5
Horsham, SCRIME
Cropping systems
13
≤3
Walpeup
Cropping systems
Cultivation
27
27
≤2
nsd
nsd= no statistical difference
Western Australia
Soil Orrganic Carbon (t C/h
ha; 0‐0.3 m)
250
Albany Sand Plain
200
150
Modelled attainable carbon
• Some pasture systems are at or above modelled attainable soil carbon stocks
carbon stocks
100
50
0
350 400 450 500 550 600 650 700 750 800
Annual Rainfall (mm)
Continuous cropping Mixed crop/livestock
Annual pasture
Perennial pasture
• Cropping and mixed crop/livestock were lower than pasture systems and generally p
y
g
y
50% of modelled attainable values
Summary and future directions
• Significant progress in measurement technologies
•
•
•
•
•
Ability to generate a suite of soil carbon data with one analysis (assess vulnerability of soil carbon change, provide input to models)
How to extend the capability developed?
Need to develop a coordinated nationally consistent approach – central MIR spectral library accessible to regional labs
We need to catch and analyse samples that do not fit current calibrations –
build the prediction system through time
How do we optimise predictions and enhance certainty?
• Sampling was not comprehensive
•
•
Some important regions, managements, soil types were not included
Some
important regions managements soil types were not included
New management strategies have not been assessed
Summary and future directions
• Temporal measurements are required to quantify sequestration
Temporal measurements are required to quantify sequestration
•
•
Establishment of monitoring system
Verification requirements – expensive ongoing requirement (are there alternatives) alternatives)
• More to be extracted from the data collected and collation with other datasets
•
•
Creation of a national repository for data and new data
Further examination of covariates and ability to predict spatial distributions of soil organic carbon stocks and composition.
of soil organic carbon stocks and composition.
• Many regions did not show a statistically significant effect of defined management practices on soil carbon stocks
•
•
At least partially due to variability that existed within management classes Are we doing the right thing by binning up on broad management classes –
should we be considering alternative metrics?
Thank you
h k
Land and Water/ Sustainable Agriculture Flagship
Jeff Baldock
Research Scientist
t +61 8 8303 8537
e [email protected]
w www.csiro.au/lorem
LAND AND WATER/SUSTAINABLE AGRICULTURE FLAGSHIP