The goal for pro-poor mitigation activity, is to develop a low-cost protocol to quantify greenhouse gas emissions and to identify mitigation options for smallholders at whole-farm and landscape levels. Learn more: www.ccafs.cgiar.org

1. Identifying pro-poor mitigation options for
smallholder agriculture in the developing world:
a multi-criteria and across-scales assessment
Mariana Rufino, Todd Rosenstock, Lini Wollenberg, Klaus Butterbach-Bahl

2. • Why multiple scales?
• Why multi-criteria?

3. The concerns
• Mitigation not linked to livelihoods
• Fragmented and diverse landscapes
• No data on mitigation
• Multi-criteria approaches missing

4. The goal
Develop a low-cost protocol to quantify greenhouse gas
emissions and to identify mitigation options for
smallholders at whole-farm and landscape levels

5. How to identify mitigation options at farm and
landscape level?

6. Phase I: Targeting, priority setting and infrastructure
Landscape analysis Set-up of state-of-the-art
and targeting laboratory facilities
Landscape
implementation Training of laboratory
and field staff
Capacity building
Phase II: Data acquisition
Productivity GHG Profitability Social acceptability
assessment measurements evaluation assessment
Joint
scientific &
stakeholder
Phase III: Multi-dimensional evaluation of
evaluation
Development of systems-level mitigation options
mitigation options
System-level estimation of
mitigation potential
Phase IV:
Implementation with Scalable and social acceptable
(UPCOMING)
development partners
mitigation options

7. GIS analysis,
Complex landscape: f (m, n, o, p, q)
remote
sensing,
landuse
trends Physical
environment
m Landscape units
Land
Food
Livestock
security,
poverty
Other assets n Farm types o Common lands
Sources of
levels incomes
Productivity,
GHG Characterise
emissions, fertility x p Field types q Land types
crop management
preferences

8. Landscape
patterns
Nyando,
western
Kenya

9. Cultivated (subsistence)
Cultivated (cash crops) id 15
id 11
Mixed
id 2
Landscape patterns
simplified

10. Landscape
structure
Nyando,
western
Kenya

11. 2001 2002 2003 2004
Mean NDVI
2005 2006 2007 2008
2009 2010 2011 2012
MODIS time series – Nyando, western Kenya

12. NDVI across landscape units (Timesat)
Nyando,
western
Kenya
Baldi et al. 2013

13. Landscape
units +
household
survey
Nyando,
western
Kenya

14. Landscape units and
land users
Sampling intensity
(sites: area)
In terms of a 250 m
square grid
class sites area (km2) sites:area
cultivated (cash and subsistence) 28 2.74 10.23
cultivated (cash) 47 5.94 7.91
cultivated (grasslands and pastures) 47 12.69 3.70
cultivated (subsistence) 141 41.54 3.39
mixed 93 34.69 2.68
uncultivated vegetation 4 2.39 1.67

15. Landscape units and landusers
Nyando, Kenya

16. GIS analysis,
Complex landscape: f (m, n, o, p, q)
remote
sensing,
landuse
trends Physical
environment
m Landscape units
Land
Food
Livestock
security,
poverty
Other assets n Farm types o Common lands
Sources of
levels incomes
Productivity,
GHG Characterise
emissions, fertility x p Field types q Land types
crop management
preferences

17. Targeting and upscaling: from
landscape to fields and back…

18. Step 1. Landscape analysis
Targeting:
- Landscape units, farm types,
field types, soils
- Site selection
Step 2. Installing measurement
stations
Site characterization:
- Soils, crops, biomass
Installation of Informing and
chamber frames interviewing farmers

19. Step 3. Measurements applying
gas pooling
Field work:
- Overcoming spatial variability
by gas pooling method
Gas sampling(closed Storage of gas
Arias-Navarro et al., Soil Biol. Biochem. submitted chamber method) samples in vials
Step 4. Lab analysis and flux
calculations
Determination of trace
Lab work: gas concentrations via
- Analyzing gas samples gas chromatography
- Calculating concentrations and
fluxes
Flux
b * Mw * VCh * 60 * 106
calculation F
formula ACh * Vm * 109

20. Step 5. Intepretation and 500
250
Forest individual chambers
gas pooling
upscaling 100
75
50
N2O flux [µg N m h ]
-1
25
0
-2
500 Grassland
250
100
75
Temporal variability of N2O 50
25
fluxes at three sites 0
500
differing in land use at 250
Cropland
Maseno, Kenya. 100
75
50
25
0
30 Oct 4 Nov 9 Nov 14 Nov 19 Nov 24 Nov 29 Nov
2012
Arias-Navarro et al., Soil Biol. Biochem. submitted
Synthesis of GHG measurements: information useful to derive emission factors,
empirical models, calibrating and validating of detailed models
Upscaling: using the targeting approach (assigning emissions to landscape elements)
and/or of GIS coupled biogeochemical models

21. GIS analysis,
Complex landscape: f (m, n, o, p, q)
remote
sensing,
landuse
trends Physical
environment
m Landscape units
Land
Food
Livestock
security,
poverty
Other assets n Farm types o Common lands
Sources of
levels incomes
Productivity,
GHG Characterise
emissions, fertility x p Field types q Land types
crop management
preferences

22. Multi-dimensional assessment of mitigation options
Farm Field Profit Production Emissions Emissions Social acceptability
type type ($/ha) (kg/ha) (t CO2eq (kg CO2 per (ranking)
per ha) kg product)
1 1 50 500 0.6 1.2 1
1 2 140 5000 3 0.6 2
1 3 120 2000 2 1.0 2
1 4 40 4500 3 0.7 1
2 1 30 800 0.7 0.9 3
2 3 180 8000 3 0.4 2
2 4 250 300 0.5 1.7 1
n m Vn,m Wn,m Xn,m Yn,m Zn,m
Trade-off analysis on multiple dimensions

23. Phase I: Targeting, priority setting and infrastructure
Landscape analysis Set-up of state-of-the-art
and targeting laboratory facilities
Landscape
implementation Training of laboratory
and field staff
Capacity building
Phase II: Data acquisition
Productivity GHG Profitability Social acceptability
assessment measurements evaluation assessment
Joint
scientific &
stakeholder
Phase III: Multi-dimensional evaluation of
evaluation
Development of systems-level mitigation options
mitigation options
System-level estimation of
mitigation potential
Phase IV:
Implementation with Scalable and social acceptable
(UPCOMING)
development partners
mitigation options

24. • Why multiple scales? -> landscape redesign
• Why multi-criteria? -> landusers are (often) poor

25. Thanks for your attention
m.rufino@cgiar.org