1. GEO-XIII Plenary, 8/10/2016 St Petersburg, Russian Federation
Capacity Development for Stimulating Innovation in Global Agricultural Monitoring
GEO-XIII Plenary, 8/10/2016 St Petersburg, Russian Federation
Sven Gilliams, VITO-TAP
SIGMA Consortium

2. SIGMA - Facts http://www.geoglam-sigma.info/
Funded By The European Commission
Start 1 November 2013
Agriculture AND Environment
22 partners, 17 countries
VITO, CIRAD, JRC, IIASA, Alterra, RADI, NMSC, DEIMOS, GeoSAS, RCMRD, Aghrymet, RCMRD, Sarvision, Sarmap, INTA, Geoville , UCL, EFTAS, FAO, ITC, GISAT, IKI, SRI
Argentina, Ukraine, China, Russia, Burkina Faso, Ethiopia USA, Brazil, Vietnam, Belgium, …
11,2 M EUR
A Major European contribution to GEOGLAM-> Supporting JECAM
Coordinated by VITO

3. SIGMA - Goal
Improve Remote Sensing based methods and indicators to monitor and assess progress towards “sustainable agriculture”,
Inventory of Crop land distribution and its changes over time
Characterize changes in agricultural production levels
Assess environmental impact of agriculture over time

4. SIGMA Activities
Land cover & crop land assessment
Agricultural Productivity
Env. Impact Assessment of Land use change
Sites: IKI RAN, SRI, RADI, CIRAD, INTA, VITO, UCL, GEOSAS, AGHRYMET
Data Management
Capacity Building

5. SIGMA: Data Management
SIGMA distribution facility
SIGMA Analysis facility (VEGA)
SIGMA Validation facility(GeoWiki)
Agricultural database (STAC)
In Situ Data Storage

6. Cross Site Initiatives;
EO indicators in crop yield forecasting
How do (EO based) indicators from publically available, global data sets perform in crop yield forecasting across different agro-environmental landscapes (JECAM sites)?
Potential for agricultural intensification
Assess intensity of agricultural production via crop yield gaps and thus identifying areas that might have or will have environmental problems

7. HR CROPLAND mapping in South Africa
Upscaling of the method presented in Waldner et al, 2015 in ISPRS
Landsat-based statistical metrics
Local supervized classification with Support Vector Machines
Uncertainty map
Unsupervised local selection of reliable calibration pixels with self-organizing maps
mean 2013
0-10% ranked NDVI percentile
% ranked NDVI percentile
Cropland map
Existing land cover map
Validation data
National Land Cover 2000

8. Global Validation Effort
Core reference data set (~ 4000 samples)
Object-based validation samples of high quality
SIGMA project partners & invited experts
Geowiki crowd sourcing campaign
Large number of point validation samples of unknown quality (~ samples)
Expert validation tool
Crowd validation tool

9. Capacity Development
GOAL: to generate learning and resource materials and make
available as a global public good.
APPROACH:
Develop three curricula (modules) consist of a series of training-units.
training/e-learning materials must build on data, materials and experiences used/gained during WP 2-5;
On the job training -> internal SIGMA partners (cross site experiment)
The e-learning modules / ppt’s will be made available freely via the Internet
Additional tools, open source software for geo-spatial analysis, methodological guides will be made available

10. Curriculum Design
Based on the assessment and the results of the sigma WP (2-5), three training modules are proposed :
Methodological aspects of the use of geospatial technology for agriculture statistics;
Agricultural monitoring using hyper- temporal remote sensing;
Geospatial data for Monitoring Agricultural changes and environmental impacts
SIGMA
MODULE 1: Agricultural statistics and monitoring;
MODULE 2: Agricultural monitoring using hyper- temporal remote sensing;
MODULE 3: Geospatial data for Monitoring Agricultural changes and environmental impacts;

11. Module 1: Methodological aspects of the use of geospatial technology for agriculture statistics
The Lessons of the Module 1 are being developed by FAO;
Main topics:
Objectives, definitions and ground data collection
Basic types of agricultural surveys: main concepts
Agricultural probability sample surveys for exploiting geospatial information
Estimation methods using geospatial information
Data processing and analysis of survey results
Improving the precision of estimates with geospatial auxiliary variables
Resources required

12. Module II: Agricultural Monitoring using Multi temporal Remote Sensing Module:
Lesson 2.1: The Hyper-temporal time-space domain   
Lesson 2.2: Vegetation and Remote Sensing·       
Lesson 2.3: Acquisition and pre-processing of hyper-temporal vegetation (NDVI) data sets·       
Lesson 2.4:  Classification hyperspectral data·        
Lesson 2.5: Visualizing, interpreting and analyzing·      

13. Module 3: Geospatial data and tools for monitoring global trends, changes and environmental impacts of agriculture:
three e-learning lessons are in production:
Lesson 1: Remote Sensing for monitoring and change detection in agriculture
Lesson 2: Geospatial data and analysis tools for global environmental trends and changes related to agriculture
Lesson 3: Geospatial data and tools for analyzing large-scale Environmental Impacts of Agriculture

14. Dissemination of CD material
SIGMA training material ready by end 2016
Through SIGMA workshops:
Internal project workshops (training event connected to SIGMA meetings)
Regional Workshops
Through Web
E-learning and ppt’s will be made available through SIGMA Portal

15. Thank you!
VITO, CIRAD, JRC, IIASA, Alterra, RADI, NMSC, DEIMOS, GeoSAS, RCMRD, Aghrymet, RCMRD, Sarvision, Sarmap, INTA, Geoville , UCL, EFTAS, FAO, ITC, GISAT, IKI, SRI