Validates the method for improving measurements of agricultural productivity by combining household level and soil fertility data. This is achived by:
developing soil-plant spectral analytical methods and diagnostic tools for rapid, low cost and reliable assessment of soil samples using light (Technology); and
demonstrating applicability of integrating scientific method of soil quality assessment along with the socio-economic panel data
Collecting the Dirt on Soils: Advancements in Plot-Level Soil Testing and Implications for Agricultural Statistics
1. Collecting the Dirt on Soils:
Advancements in Plot-Level Soil Testing
and Implications for Agricultural Statistics
http://www.worldagroforestry.org/research/land-health
Ermias Betemariam (e.betemariam@cgiar.org)
Sydney Gourlay
Keith Shepherd
2. Context
1Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015|
Innovations in Agricultural Statistics
• Administrative data is often seen as methodologically flawed and
lacks credibility
• Lack of consistent, good quality data on soil health
• Data collection methods
• The role of technology
• What works in developing countries?
3. Context (2)
2Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015|
Methodological improvements
in smallholder agricultural
statistics for decision making
http://www.csa.gov.et/index.php/component/co
ntent/article/14-survey-reports/123-latest-
national-statistics-abstract
4. Objectives
3
Improving measurements of agricultural productivity by
combining household level and soil fertility data
• Develop soil-plant spectral analytical methods and diagnostic tools for
rapid, low cost and reliable assessment of soil samples using light
(Technology)
• Demonstrate applicability of integrating scientific method of soil quality
assessment along with the socio-economic panel data
Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015|
5. Field data
4Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015|
3661 soil samples collected from three
agro-ecologica zones
• Household socio-economic data
• Subjective assessment of soil quality
• Soil sampling
• Objective assessment of soil quality
6. Lab analyses
Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015| 5
Laser Diffraction Particle
Size Analysis(LDPSA)
Total X-ray Fluorescence spectroscopy
(TXRF) for elemental analysis
FT Diffuse reflectance MIR Alpha MIR spectrometer
7. Instrumentation
Dispersive VNIRFT-NIR FT-MIR
Handheld
NIR/MIR
•Portable
•Repeatability?
•External service
•No validation
•Benchtop
•Repeatability
•Self serviceable
•Validation in-built
•ISO compliant
•Industry proven
•Multipurpose
•Benchtop
•Repeatability
•No gas purging
•Some servicing
•Robotic
•Validation in-built
•ISO compliant
•Outperforms NIR
•Handheld
•Sample
homogeneity?
•Variable
moisture?
•Repeatability?
•Still expensive
•Rapidly
developing
•Need to prepare
by developing
soil reference
libraries
Working on Rural labs
Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015| 6
Lab analyses (2)
8. Why infrared spectroscopy?
Rapid
Low cost
Reproducible
Predicts many soil functional properties
Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015| 7
9. 8
Land Health applications
Africa Soil Spectroscopy Lab Network
EthioSIS
97 Sentinel sites
Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015|
12. 1
1
Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015|
Objective vs subjective soil quality measurements
Farmers use soil texture and color to describe soil qualities
13. 1
2
Soil carbon and pH values
Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015|
Objective vs subjective soil quality measurements
SOC (95% CI): 3.18 - 3.30%
pH(95% CI): 6.28 - 6.34
14. 1
3
Objective vs subjective soil quality measurements
Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015|
Soil quality indicators
USDA:
http://www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/health/assessment/?cid=stelprdb1237387
15. 1
4
Subjective soil quality assessment poorly correlated to the scientific output
Science: about 17% of the soils
have SOC < 2% but only 5%=
poor soils
Local respondents identified
only 5% of the as poor soils
Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015|
Objective vs subjective soil quality measurements
16. 1
5
Age Sex Literacy
Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015|
Objective vs subjective soil quality measurements
Age class, sex, and literacy did not significantly influence the subjective
Grading of soil quality
Farmers better identify poor and not-poor soils but not “Fair soils” that future
Questions should consider
17. 0
2000
4000
6000
10 50 100 150 200 250
Carbonmeasurment
cost(USD)
Number of samples
NIR spectroscopy
Thermal oxidation
Sample preparation
0
3
6
9
12
15
Carbonmeasurement
costpersample(USD)
NIR spectroscopy
Thermal oxidation
Sample preparation
Soil sampling
Cost –error analysis
0
2000
4000
6000
8000
0 500 1000 1500
Carbonmeasurement
cost(USD)
Number of samples
Thermal oxidation
Comparisons of costs of measuring SOC using a commercial lab and NIR
Cost
IR is cheaper (<~ 56%) than combustion
method for large number of samples
Throughput
Combustion ~ 30-60 samples/day
NIR ~ 350 samples/day
MIR ~ 1000/day
Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015| 1
6
18. Surveillance science
Land health metrics
Consistent field
protocol
Soil spectroscopy
Coupling with
remote sensing
Sentinel sites
Randomized sampling schemes
Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015| 1
7
19. AfSIS: Soil functional properties
From polygon-based to probabilistic mapping
+
Probability of observing
cultivation
Current lime requirement ? ~ min
[prob(pH < 5.5), prob(cult)]
Probability topsoil pH < 5.5
... very acid soils
Grid-based probabilistic maps increases the reliability of the map and its
power to be combined with other data sources (remote sensing & terrain data)
(Walsh, 2013)
=
1
8
Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015|
21. • More research on cost-effective measurement tools- the future is bright
• Subjective and objective measurement results do not much well
• It was possible to integrate soil health measurement into the panel
household survey
• Enhancing our understanding of household attributes, land management
practices and soil health management
• Enable decision makers have clear understanding of soil status and trends
Finally…
2
0
Ermias Betemariam| ISI2005 2015, | Rio de Janeiro| July 2015|
Smart data - Smart decisions
Lack of consistent, good quality data on soil health
We won a grant to incorporate a field-scale soil monitoring component in the World Bank’s Living Standards Measurement Study, which has been helping a number of Governments establish household panel surveys and agricultural monitoring over several decades. We will be piloting a soil fertility monitoring component in two African countries.
Our Soil-Plant Spectral Diagnostics lab continues to attract much interest – we received over 500 visitors in 2011, over half of which received some training.
Requests for capacity building are growing exponentially and we need to increase our staff to deal with it.
Global libraries are needed to be able to quickly bring new instruments into calibration
Black red light/white/yellow
Fine, medium, coarse
A quick reminder of our conceptual framework and tools, we work by a set of surveillance science principles, which are similar to those used in public health surveillance – which emphasize quantifying health problems and associated risk factors in populations.
We implement those science principles through a set of tools, which encompass use of randomized, landscape level sampling schemes. The use of consistent field sampling protocols so we collect data on land health indicators in the same way everywhere. The use of soil spectroscopy methods to provide high throughput low cost analysis of key soil health metrics, centred on soil functional properties. Coupling of the field and lab observations with remote sensing data, to provide consistent data on the population distributions and prevalence of land health problems, associated risk factors and digital mapping of indicators.