+ All Categories
Home > Documents > Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10%...

Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10%...

Date post: 31-Dec-2015
Category:
Upload: flora-walton
View: 213 times
Download: 0 times
Share this document with a friend
Popular Tags:
26
Areas of Operational R&D for GLAM Enhancements
Transcript
Page 1: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

Areas of Operational R&D for GLAM Enhancements

Page 2: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

Development of EO-based Yield Forecasting

RMSE= 10%R= 0.94Y=0.9934X

Kansas Estimates within 8%, 6 weeks prior to harvest

Ukraine Estimates within 10%, 6 weeks prior to harvest

Becker-Reshef I, Vermote E, Lindeman M, Justice C. 2010. In Remote Sensing of Environment, 114, 1312–1323.

Page 3: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

MODIS data for corn area indicator mapping – per state model

NASS AWiFS CDL 2008 corn MODIS 2008 corn

Annual Within Season Cropland Area Indicators

0 20 40 60 80 100

100

80

60

40

20

0

CDL %

MO

DIS

%

Single State Corn Single Year– 2008 – 5km

R-Squared 0.9238

Hansen et al.

Page 4: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

Top-of-atmosphere TOA Surface Reflectance

Next Steps for GLAM:

Transferring Atmospheric correction algorithm from MODIS to Landsat domain

• Integration of Landsat and preparation for LDCM• System preparation for VIIRS continuity• Continued R&D on yield forecasting, area indicators, crop

mapping

Page 5: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

• Currently multiple operational agricultural monitoring systems – operate independently in a poorly coordinated

way and without standardized methods

• Increasing synergies, and improving satellite and in-situ observations would enhance our ability to effectively monitor agriculture worldwide

Need for International Collaboration & Coordination: The Programmatic Context GEO

Page 6: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

Group on Earth ObservationsGlobal Agricultural Monitoring Task: AG 07-03

Goals: • Bring the international agricultural monitoring community together to build a

system of systems (GEOSS) for effective global agricultural monitoring• Building on existing assets and the systems that are in place• Promoting data sharing• Promoting methods/modeling sharing and inter-comparisons leading to best

practices guidelines• Assessing current state of the art and articulating gaps, requirements and needs

GEO Plenary, Beijing. November 2010

Page 7: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

The GEO Global Agricultural Monitoring (Task: AG-07-03)

Task Co-Leads: Chris Justice, University of Maryland, USAWu Bingfang, Institute of Remote Sensing

Applications, CAS, Beijing, ChinaOlivier Leo, Joint Research Centre, European

Commission, Ispra, ItalyDerrick Williams, USDA FAS, USA

Task Executive Director: Jai Singh Parihar, Space Applications Centre (ISRO), India

JECAM Sub-task Lead: Ian Jarvis, Agriculture and Agri-Food Canada PAY Sub-task: Lead Inbal Becker-Reshef UMD, Meng Jihua CASGEO Secretariat PoC: Joao Soares, GEO Secretariat, Geneva

CEOS GEO Agriculture POC: Prasad Thenkabail, USGS , USA

Page 8: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

Agricultural Monitoring Systems Contributing to the GEO Community of Practice

• Most countries have a national agricultural monitoring system • Similar data needs, need for coordination and cooperation in sharing of data and

methods inter-comparison

Page 9: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

GIEWS- Global Information and Early Warning System (UN-FAO)

Provides global information on food supply and demand provides early warnings of impending food crises in individual countries

Page 10: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

MARS-FOOD Crop Monitoring SystemEuropean Commission Joint Research Center (JRC)

Crop Assessment Process

Data collection & retrieval

Earth Observation Data

Meteorological Data

Agronomic Database

WEB InformationEuropean Media Monitor

Processing &Analysis

#Y

Bay

Mudug

Gedo

Hiran

Lower Juba

Galgadud

Bakool

LowerShabelle

MiddleJuba

MiddleShabelle

Mogadishu

Ethiopia

Somalia

#Y

Tog-Dheer

Djibouti

AwdalGalbeed

Bakool class "isolated fields/ rainfed"Agro-pastoral sorghum

0.0

0.1

0.2

0.3

0.4

0.5

0.6

M A M J J A S O N D J F

Time

0

10

20

30

40

50

60

70

80

mm

rain 2003-04Average 98-022002-032003-04

Cumulated PrecipitationIran (south)

0

100

200

300

400

500

600

Sep1 Oct1 Nov1 Dec1 Jan1 Feb1 Mar1 Apr1 May1 Jun1 Jul1 Aug1

dekads

mm

2003/2004 2002/2003

Kenya

Percentage of maize area in each

class

0%

20%

40%

60%

80%

100%

2001 2002

C-NDVI

Crop GMS

WS Index

Dissemination

- EU Delegations- National EW Agencies- Int. Institutions (FAO, …)

Bulletin Dissemination

Data Dissemination

Reporting

Page 11: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

GEO Agricultural Monitoring Task Near Term Initiatives

• Initiative 1 : Joint Experiments on Crop Assessment and Monitoring (JECAM)

• Initiative 2 : A Multi-source Production, Acreage and Yield (PAY) database

• Initiative 3 : Coordinated Data Initiatives for Global Agricultural Monitoring (CDIGAM)

• Initiative 4 : GLAMSS Thematic Workshop Series (GTWS).

• Initiative 5 : Agricultural Land Use and Climate Change.

Page 12: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

JECAM: Joint Experiments on Crop Assessment and Monitoring

• GOAL: facilitate the inter-comparison of monitoring and modeling methods, product accuracy assessments, data fusion and product integration, for agricultural monitoring

• setting up a network of regional experiments in cropland pilot sites around the world• Ongoing discussion for a site in Ukraine to be led by NASU (National Space

Agency of Ukraine)

• Time series datasets from a variety of earth observing satellites and in-situ data sources will be acquired for each of the sites

• synthesis of the results from JECAM will enable:– development of international standards for monitoring and reporting protocols– a convergence of the approaches to define best practices for different

agricultural systems– identify requirements for future EO systems for agricultural monitoring.

Page 13: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

Initial Countries participating in the JECAM InitiativeSubtask led by Agriculture and Agri-Food Canada

JECAM Website http://www.umanitoba.ca/outreach/aesb-jecam/index.html

USA

Paraguay

JECAM activities are being undertaken at a series of study sites which represent the world’s main cropping systems and agricultural practices.

12 sites currently exist. Additional sites will be added to meet science objectives and ensure all major crop systems are addressed.

At a joint CEOS-JECAM meeting in Ottawa in Sept 2011 the space agencies and commercial providers pledged support for JECAM, JECAM was afforded a high priority by all data providers

Page 14: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

The PAY- a Production, Acreage, Yield multi-source online database initiative

• Goal: provide a platform for comparisons between crop statistics generated by different agencies, through a common centralized online database of Production, Area, and Yield (PAY) – enable identification of agreements and disagreements in

national level crop statistics to guide methods development and best practices guidelines

• Potential interface with G20 AMIS Initiative

Page 15: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

Example query for results comparing yields from the different agencies:

Lines in blue indicate reported statistics, white indicate estimates

Page 16: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

Graphing functionality: Inter-comparison of Crop Statistics

Squares indicate official statisticsCircles indicate in-season estimates

Page 17: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

GLAMSS Thematic Workshop Series (GTWS)• April 2011, ISRSE, Sydney: Workshop on Rangelands and Pasture Monitoring• May 2011, Curtiba Brazil (SBSR): JECAM South America Workshop• June 2011, Vienna Austria: Agricultural Land Cover Mapping Workshop• September 2011, Nairobi Kenya: JRC CRAM workshop• October 2012, China: Workshop on Agricultural Water Availability

Brussels 2010 – AGRISAT WorkshopBeijing 2009 – System of Systems Components

Kananaskis 2009 - SAR to support Agriculture

Page 18: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

Coordinated Data Initiatives for Global Agricultural Monitoring (CDIGAM)

– Ensure the on-going, frequent and timely acquisition, accessibility of satellite data during crop growing season and the continuity of those observations necessary for agricultural monitoring

– Compile the best available information on agricultural areas, crop calendars and cropping systems – define a global acquisition strategy

– To fill the gaps in the current in-situ observations

– Near Term CoP Contributions:- Dynamic Global Croplands Likelihood Map (250m) - Near Real Time data from MODIS (NASA LANCE System)- Compilation of Enhanced Global Crop Calendars (ISRO)- NASA VIIRS data- SR landsat data

Page 19: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

Defourny 2010Black - Global Operational Grey – Regionally ImplementedWhite – Research / Local Domain

Crop Monitoring and Famine Early Warning EO System Schematic

Page 20: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

GEO Agriculture Monitoring Community of Practice Website:http://www.earthobservations.org/cop_ag_gams.shtml

Page 21: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

GEO Ag 0703 CoP Brochure

Page 22: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

G20 GEO-GLAM Initiative• G20 Ministers adopted the GEO GLAM proposal in June

2011 Four components are envisioned for GEO-GLAM: 1. Improving Global Agricultural Monitoring Systems with a focus on :

a) Large Producer/Exporter Countriesand

b) Countries at Risk

2 Enhancing National and Regional Capacity for Agricultural Monitoring and the timely dissemination of monitoring results

3 Improving availability, access to, timeliness and use of EO data for agricultural monitoring (Satellite, In-situ and EO parameterized Models)

4 Undertaking innovative Research and Development in support of Operational Monitoring Systems

Page 23: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.
Page 24: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

GEO-GLAM Components

Coordinated Satellite and In-

Situ Earth Observations

Strengthening National

Capacity for Agricultural Monitoring

Earth Observations Satellite / Ground Data / Models

Operational Research and Development Techniques/Methods/Best Practices

Improved Reporting and Information and Timely Dissemination SystemsCondition/Area/ Yield / Statistics

FAO STAT AMIS Public

MONITORING SYSTEM OF SYSTEMS

Meteorological Expertise and

Info

Agricultural Expertise

(GEO CoP+)

Enhancing Global Agricultural

Monitoring Systems

1

Monitoring Countries and Regions at Risk

(EWS)

2 3

Govts

Page 25: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

Phased Implementation

• Initial Planning Phase (June 2011-2012)• Phase 1 (2012-2015) Focus on:

i) Coordination of Earth Observationsii) Cereal Crops for Major Producer/Exporter Countriesiii) Agricultural food supply for Countries at Risk

(Karamoja project)

iv) National Capacity Building for monitoring primary national crops

• Phase 2 (2015-2020) Expanding program focus e.g. to include Rangeland Productivity Monitoring, Climate Change Adaptation

Page 26: Areas of Operational R&D for GLAM Enhancements. Development of EO-based Yield Forecasting RMSE= 10% R= 0.94 Y=0.9934X Kansas Estimates within 8%, 6 weeks.

Summary • Changing climate, competing demands for agricultural land, changing diets

and changing energy and food prices will mean more volatility in food supply and demand

• Timely agricultural monitoring is becoming increasingly important and ensuring continuity and coordination of Earth observations is fundamental

• International community will continue to partner– For ensuring enhancements and continuity of effective global monitoring and

information dissemination

• The G20 GEO-GLAM initiative WILL ADDRESS A CRITICAL NEED FOR IMPROVED INFORMATION FOR GLOBAL FOOD SECURITY and market stability

• THE TASK OF GLOBAL AGRICULTURAL MONITORING IS TECHNICALLY FEASIBLE AND ASSUMING THAT WE HAVE THE POLITICAL WILL FROM THE G20 COUNTRIES AND INTERNATIONALFUNDING COMMITMENT IT CAN BE IMPLEMENTED.

• THE EFFORT IS TOO LARGE AND IMPORTANT FOR ANY ONE COUNTRY TO IMPLEMENT AND THE INTERNATIONAL SATELLITE ASSETS WILL BE NEEDED TO PROVIDE THE NECESSARY OBSERVATION FREQUENCY OF COVERAGE AND A COMMON POLICY OF FREE AND OPEN DATA (AS ADOPTED BY THE US) WILL BE NEEDED INTERNATIONALLY.


Recommended