Transcript Decision support tools for GEC/Food Systems research

Developing Iterative Decision Support
Tools for Exploring GEC and Food System
Scenarios for the Indo-Gangetic Plain
Greg Kiker
Agricultural & Biological Engineering Dept
University of Florida
Phone: (352) 392-1864 ext 291
Email: [email protected]
Overview
•
•
•
•
Decision Support Tools and Analysis
QnD Model and its Use
QnD:Jamaica Version 0.0
Discussion
Traditional Decision Support Systems
• Many definitions, many visions, many versions…
• DSS Integrate Tools and Methodologies
–
–
–
–
–
Method Descriptions and Tools
Information Storage and Databases
Models
Visualization (examples: maps, graphics, charts…)
Decision analysis (examples: Adaptive management,
Multi-Criteria Decision Analysis, Trade-off analysis,
Linear Programming)
Challenges in Building DSS
• DSS almost always focus on one sector’s (or
discipline’s) tools and concepts
• Model integration is difficult and technically
challenging
• DSS are not always helpful in unearthing assumptions
• Transparency versus Ease of Use
• Integration with scenario planning and communication
- “More gaps than links” (Mike Brklacich, 2005)
Scenarios, predictions & projections?
The probability that …
Facts
Causality
known
What if?
Prediction
Decision Support Scenarios
Projection
unknown
low
Uncertainty
high
(data availability)
Source: modified after R. Leemans 2003
Useful Questions
• Who will use these IGP Decision Support Tools?
• What decisions need to be made?
– Over time?
– Over space?
– Adaptively?
• How do you or your constituents currently make decisions?
• What new techniques are you willing to explore?
• What are your information, support and communication
needs?
– Training/Workshops? Reports? Web Tools?
• What does transparency in decisions mean to you?
– What does it look like?
– What level of transparency is appropriate?
And remember…you can always change your mind…
QnD Model
• QnD™ – “Questions and Decisions™” or
“Quick n Dirty”
• A fully integrated Graphic User Interface
(GUI) with a flexible model engine
• One model - Many ecosystems
– Java code / XML inputs
– Dynamically instantiated objects
• Java-based deployment in web browsers
• “Fast Deployment” (weeks/months)
• Spatial simulation with GIS linkage
• Multiple time steps
• Multiple maps/graphs/files for output
variables
“Simulation Engine”
“Game View”
 Developer’s point of contact
 User/Player’s point of contact
 Integrates information
 Communicates information
 Objects: Components, Processes and Data
“Widgets”: Maps, Charts, Warning Lights,
Text, Sliders, Icons, Buttons
 Calculation of state variables
 User choices – management settings,
simulate fast or slow time step, reset
subComponent
subProcess
QnD:Game View
User-Defined
“Slow and Fast”
Time Steps
Management Options
Reset Button
“Warning Lights”
Tabbed panes with several
time series charts
Mouse-driven
Chart and Text
Display
Spatial Maps of Ecosystem Indicators
“Game View”
Actors
Players: Interact mostly with the game view.
• Explore management responses, adaptive opportunities,
trade-offs for different scenarios.
• Provide reality checks
• Have some interest in the engine structure in their area of
interest
“Simulation Engine”
• Provide ideas and directions for further iterations
subComponent
Developers: Interact mostly with the engine.
• Design and implement engine/game view through XML
files.
• Provide formal calibration/validations
subProcess
• Implement ideas and directions of Players
• Have some interest in the model code
QnD Java Source Code
GameDriver.java
ModelCreator.java

Coders: Interact mostly with the QnD source code.
• Develop java code to expand engine and game view utility
PrimaryGameFrame.java
• Create new programming code for ideas from Players and
Developers…
QnD: Development Methodology
Genesis Session
Prototype QnD Game View and Simulation Engine
• Talk about the system, goals, desires
• Explore current management options
• Gather initial maps/data
• Brainstorm about desired management options, relevant
information and socio-economic realities
• Rough estimate of components, processes and data
• Simple information
• Deployed in limited circulation for calibration/reality checks
Iterative Sessions 1…n
Deployed QnD Model
• Refine goals, objectives
• Explore current and possible management options
• Calibrate/Validate engine performance
• Revise Game View for relevant management information
• Make changes concerning management options, relevant
information and socio-economic factors
• Player/Developer reviewed components, processes and data
• More relevant information
• Brainstorm about desired management options, relevant
information and socio-economic realities
QnD Web Site/Deployment
Launch QnD JNLP deployment
Launch QnD Applet Version
View QnD XML Files Style Sheets
QnD:Jamaica – Overview
Moving Money, Food and People
Jamaica
Climate
Rural People
Domestic Food
Production
International
Economic Activity
Foreign
Remunerations
Middle Folk
Food Imports
Key
Urban People
External Driver
Influences
Food Transfer
QnD Objects
Money Transfer
Population Transfer
Parish Spatial Unit
Habitats
• Urban Habitat
• Rural Habitat
Marine Spatial Unit
Habitats
• Default Habitat
(100 %)
J$
Exchange
Rate
US$
Key
Class A Road
Class B Road
Import Market
• Food
• Other Goods
Export Market
• Cash Crop
• Other Goods
Rural Area
Unsuitable Land
Convert
Land
Plant
Available Land
Livestock
Farmers
Cash Crops
Harvest
$
Food Crops
Sell Crop
$
Cash Crop
Sell Fish
$
Fishers
Catch
Buy Food
Export Market
Buy Food
$
Tourist Workers
Buy Food
Other Worker
$
Unemployed
Buy Import
Imported Food
Middle Folk
Marine Spatial Unit
$
Sell Fish
Fish
$
Catch
Fisher
Export Market
Buy Food
Key
$
Influences
$
Food Transfer
Tourist Worker
Urban Area
Buy Food
Other Worker
Unemployed
Money Transfer
Rural Area
Farmers
Harvest
$
Sell Crop
$
Payments
from
Overseas
Sell Fish
$
Fishers
Buy Food
Buy Food
$
Tourist Workers
Buy Food
Other Worker
$
Unemployed
Buy Import
Imported Food
Middle Folk
$
Sell Fish
$
Catch
Fisher
Key
Buy Food
Population Transfer
$
Tourist Worker
Food Transfer
Money Transfer
Buy Food
Urban Area
Other Worker
Unemployed
Send Money
$
Change Location
Socio-Economic Drivers
Millennium Scenarios
• Global Orchestration
• Order From Strength
• Techno Garden
• Adapting Mosaic
Tourist visits, foreign investment,
external prices, etc.
QnD:Jamaica Version 0.0
• Stochastic relationships
• Time series values
• Incremental change
Climatic Drivers
Climate Scenarios
• Precipitation
• Hurricane Intensity/ Frequency
• Sea Level Rise
QnD Scenarios
• Techno Garden with mild Climate Change
• Techno Garden with intense Climate Change
• Etc…
IGP Decision Support Tools
Ludihana, Central Punjab, India:
wheat and rice predominate, slow to
stagnant productivity growth,
groundwater dependent, lots of
investment, high income levels,
functional policy support.
Gujarat, Punjab, Pakistan:
wheat dominates, food selfsufficient, mixed irrigation,
high level of infrastructure,
moderate income levels,
policies function.
Vaisahali District, Bihar, India:
rice preferred, low infrastructure
investment, flooding, low income
levels, out migration, little
government policy support.
Ruhani Basin, Terai of Nepal:
rice preferred, transition zone,
seasonal flooding, out-migration,
sharecropping dominates,
urbanization increasing.
Greater Faridpur, Bangladesh:
rice dominates, flooding and concern
over salt water intrusion, low income
levels, government institutions fail.
Source: www.gecafs.org
IGP Decision Support Tools
• Spatial Areas of Interest ?
– 5 zones? Districts?
– Rural vs Urban ?
• Scenarios?
– Land Use, Socio-Economic, Climate
Change Scenarios?
– Drivers (Floods, Rainfall, Climate Change,
Trade Policies?
• Decision Metrics / Well Being Metrics
?
– “Bright Lines”?
– Local metrics vs regional metrics
• Time Steps?
– Growers/Citizens (Monthly?)
– Decision/Policy Folk (1-5 years?)
– GEC (10 – 50 years?)
• People?
– Landless , Landholders (.1-.2.5ha),
Landholders (2.5-5ha), Landholders (>5 ha),
Wage Earners, Unemployed
• Food?
– Wheat? Rice? Imports?
• Resources?
– “Bright Lines”?
– Local metrics vs regional metrics
• Policies?
– Food Aid?
– Crop Insurance?
– Rural Enterprise development?
QnD Model: Further Information
• General QnD information (available as pdf files)
– Kiker, G.A., Rivers-Moore, N.A., Kiker, M.K. and Linkov, I. (2006). QnD: A modeling game
system for integrating environmental processes and practical management decisions. (Chapter in
Morel, B. Linkov, I., (Eds) “Environmental Security and Environmental Management: The Role of
Risk Assessment.” Springer, Netherlands. Pp:151-185.
– Kiker, G.A. and Linkov, I. 2006. The QnD Model/Game System: Integrating Questions and
Decisions for Multiple Stressors. (Chapter in Arapis, G., Goncharova, N. and Baveye, P. (Eds)
“Ecotoxicology, Ecological Risk Assessment and Multiple Stressors ” Springer, Netherlands.
Pp:203-225.
• Other QnD Papers
– Kiker, G.A., Kiker, M.K., Mosier, A. and Jones, J. (2006). Design and Development of a ScenarioBased, Decision Support System for Integrating Food, Market and Ecological Systems in the
Caribbean Region. GECAFS White Paper. (In final preparation, Agricultural Systems).
– Kiker, G.A., Justice, L. (2006). QnD:Jamaica v0.0 – Design and Implementation of a decision
support model for food, finance and population dynamics in Jamaica. (In preparation, Agricultural
Systems).
– Kiker, G.A., Kim, J.B., Justice, L. and Linkov, I. (2006). Integrating Simulation Models and
Multicriteria Decision Tools for Estimating the Socio-Economic Impacts of Natural Hazards:
Hurricanes and Sea Level Rise in Jamaica. (In preparation, Risk Analysis).
Discussion
• Who will use these IGP Decision Support Tools?
• What decisions need to be made?
– Over time?
– Over space?
– Adaptively?
• How do you or your constituents currently make decisions?
• What new techniques are you willing to explore?
• What are your information, support and communication
needs?
– Training/Workshops? Reports? Web Tools?
• What does transparency in decisions mean to you?
– What does it look like?
– What level of transparency is appropriate?
And remember…you can always change your mind…
Questions?