Presented by Pierre C.S. Traore (ICRISAT) at the CGIAR-CSI Annual Meeting 2009: Mapping Our Future. March 31 - April 4, 2009, ILRI Campus, Nairobi, Kenya
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
[Day3] Agcommons Quickwin: Seeing is Believing
1. yenko ye foko bo
be be wule bibile wob e
seeing is believing:
unlocking precision agriculture in West
African smallholder communities with very
high resolution imagery
nif yaab la sida
gani ya kori ji
AMEDD, Fuma Gaskiya, ICRISAT, IER, INERA, INRAN,
KMG, SARI, UACT, UPN
SIBWA – CSI – Nairobi, 02APR09
2. The idea
• Precision ag. irrelevant to smallholders? FALSE
• West African farmers = PA pioneers
• Why would they want VHRI then?
– We’re not sure, but they want it for sure
– Field acreages, reveal less visible patterns of change
– Map hotspots, bright spots, other spots
– Field-level metrics for rainwater management
– The “conscious” (and ambitious) side: decision
support for productivity enhancement technology
(field level)
SIBWA – CSI – Nairobi, 02APR09
3. More ideas
• The “unconscious” (or safe) side: discussion support for
agricultural landscape design (community level)
• Land tenure, community arbitration, decentralization
• The urban analogy – why should VHRI concentrate on
urban areas? At such a low cost, shouldn’t rural
communities equally benefit from it? YES, THEY
SHOULD
• VHR mapping = precursor of land security & SLM (when
you realize that you should own and invest)
• VHR mapping = precursor of intensification (when space
is limited and resources need better organization)
• VHR mapping = precursor of demystification (when
climate gets back into Pandora’s box)
SIBWA – CSI – Nairobi, 02APR09
4. Objective(s)
• Demonstrate the value of VHRI to help scale up a few
quick-win productivity enhancement technologies in 6
smallholder communities across Burkina Faso, Ghana,
Mali and Niger
– Focus on the last 8 km
– Candidate technologies: spatially optimized soil and
water management practices
– Show a variety of value-added products
– Demonstrate real-world deployability and potential
impact
– Upload GIS datasets to shared online AgCommons
repository
– Publish metadata on GeoNetwork
SIBWA – CSI – Nairobi, 02APR09
5. 4 phases, 7 tasks
• 1: Organize resources: human, methods, tools
• VHRIbox
• VHRIex2
• 2: Share information at sites
• ROLLout (x 6)
• 3: CRT support functions
• CRTtopo (x 5)
• CRTverif (x 2)
• 4: FEED and CAP
• FEEDback (x 6)
• FASTfwd (x 6)
SIBWA – CSI – Nairobi, 02APR09
6. Task 1: Build site-specific VHR information
containers and proto-maps (VHRIbox)
• 1 information container and 1 set of proto-maps per site
• Containers: geodatabase shells with initial matching
ingredients. Will later host project generated information
• Laminated printouts will crystallize initial VHRIbox content
into proto-maps to engage VHR information exchange
with farmers
• Proto-maps: field boundaries overlaid on i/ VHR color
composites, ii/ VHR NDVI, iii/ toposequence, slope
class, iv/ hotspots (field-level NDVI anomaly), v/ field-
level CRT potential
• Deliverables: 6 containers, 6x5 proto-maps
SIBWA – CSI – Nairobi, 02APR09
7. Task 2: Build a human interface for VHR
information extraction and exchange
(VHRIex2)
• Assemble, train team of VHRI-conversant staff including
1 gender-aware quintet per site: junior local extension
specialist literate in local languages, junior field GIS
technician or student , farmer representative, senior local
NGO or extension personnel, a scientist/backstopper
• Equip team with standard interfacing tools and
procedures to interact with different stakeholders
• Deliverables: 6 quintets, 1-week crash retreat held
for training on VHRIex2, 6 experimental protocols
with toolkits, procedures, etc. uploaded
SIBWA – CSI – Nairobi, 02APR09
8. Task 3: Roll out VHRI to farmer fields
(ROLLout)
• Sequential site exposure from wet to dry
• proto-maps presented to FOs through focus groups
following a stratified sampling protocol (tbd): i/ farmers
exposed to proto-maps and productivity enhancement
technologies and ii/ farmers exposed to productivity
enhancement technologies only (control group)
• Movies and on-site demos on productivity enhancement
technologies
• IER and AMEDD to lead
• Deliverables: 6 sites covered from 01MAY-11JUN (1
week/site)
SIBWA – CSI – Nairobi, 02APR09
9. Task 4: Derive field-level topography metrics
to assess potential (CRTtopo)
• develop semi-automated VHRI analysis method to
extract dominant furrow azimuth in cattle plowed fields
• Drape results on DEM to estimate average departure
from the dominant field slope
• Interpreted in terms of local priority for infiltration or
drainage (function of field position on the toposequence)
• only applies to sites with significant cattle plowing: all but
Serkin Hawsa
• Deliverables: 1 report on VHRI processing methods
to assess field-level CRT potential, 5 field-level
suitability maps (1 per site)
SIBWA – CSI – Nairobi, 02APR09
10. Task 5: Test VHRI as an objective
verification tool (CRTverif)
• test the potential of VHRI for CRT impact assessment
using historical measures of biomass productivity (NDVI-
based) as an alternative to detailed household and field
surveys
• Compare 30 CRT-equipped farmer fields in Fansirakoro
and Sukumba to control fields that lie in the same
toposequence class
• Deliverables: 1 report on VHRI processing methods
to assess CRT impact in 2 sites
SIBWA – CSI – Nairobi, 02APR09
11. Task 6: Collect user feedback on site
(FEEDback)
• deploy team in 6 sites to update, for all collaborating
farmers: field boundaries, field ownership, cropping
histories (field level), information resource allocation and
management of abiotic stresses (household level).
• For best tradeoff between crop differentiation and farmer
time constraints, will take place towards peak biomass
(September)
• will involve collection of farmer feedback on productivity
enhancement technologies they may have tested (or
not), and on VHRI derived maps they may have used for
purposes of technology targeting or for other purposes
(or not).
• largest concurrent deployment of human resources.
• Deliverables: 6 sites covered concurrently (on-site
presence: 1 month/site)
SIBWA – CSI – Nairobi, 02APR09
12. Task 7: Forward maps of farmer screened
technology to sites (FASTfwd)
• Collate user feedback and encode information into site
geodatabases, update proto-maps and ship back to FOs,
LGAs, community leaders
• Laminate individual A4 farm maps for collaborating
farmers, synthesizing key learnings, recommendations
for technology deployment
• Elements of fine tuned information that will be forwarded:
field acreage, intra-field hotspots of abiotic stress for
targeting of (organic) fertilizer inputs, field fitness for
CRT implementation (or other promising technology
indentified during the course of the project).
• Deliverables: 6x5 laminated maps finalized and
forwarded to local communities. 6xn individual
farmer maps finalized and forwarded to local
recipients, All geospatial material generated during
project lifetime uploaded for online serving.
SIBWA – CSI – Nairobi, 02APR09
13. 2 productivity enhancement technologies in
mind… or more?
• Water management: CRT (+20%!)
• ISFM: fertilizer micro-doses? Composting/manure?
(+20%!)
• Flexibility built-in for game time decisions depending on
local experts, farmers
SIBWA – CSI – Nairobi, 02APR09
23. • Top: 2003 and cover
(crop type) on pan
imagery
• Bottom: 2003 crop-
specific estimates of
mean biomass production,
in g.m-2 on ASTER digital
elevation model
SIBWA – CSI – Nairobi, 02APR09
24. Multispectral false color composite (NIR reflectance
appears in red) displaying cotton stand establishment in
CRT (left) and non-CRT fields (right) on Diakaria Konaté's
farm, 40 days after sowing
SIBWA – CSI – Nairobi, 02APR09
25. Same, with regressed biomass estimates overlayed.
Yellowish to greenish colors display field areas with at least
10 g.m-2 of crop dry matter equivalent
SIBWA – CSI – Nairobi, 02APR09
26. Who and how many will be affected
• ~3,000 farmers & dependents (est. 50 farms or
households per site, 10 people per farm/household)
• ~ 6 to 12 farmer organizations (1/site, plus women’s
groups)
• ~3 to 6 local NGOs
• ~30 research personnel (scientists and extension staff
included)
• ~ 3 to 6 institutions from donor community and policy
making arena
SIBWA – CSI – Nairobi, 02APR09
27. Ideas for phase 2 (if 1 works)
• Scale up: do the same in more communities
– All 703 communes of Mali? ~2M USD
– Larger selection in more countries? ~x M USD
• Test new applications:
– Test SOWO carbon accounting protocol for carbon trade projects
– Scale down aflatoxin risk early warning products in Ghana, Mali
(in partnership with CCLF-aflatoxin project)
• Data-wise:
– Get that GeoEye
– Get that SRTM-30m
SIBWA – CSI – Nairobi, 02APR09