VIRTUAL WATERSHED: A SPATIAL DECISION SUPPORT SYSTEM FOR AN

AGRICULTURAL WATERSHED

Seth Soman, Girmay Misgna, andSteven Kraft

ACES 2018Washington D.C. , December 4- 6

Based upon work supported by the National Science Foundation under Grant No. 0410187

Introduction

Watershed management takes place on a landscape controlled by private landowners. Their decisions will, in large part, reflect economic criteria like profit maximization.

To maintain or enhance ecological integrity, as well as avoid conflict with the land users, watershed management plans should reflect the economic uses to which the privately held land can be put.

Virtual Watershed

• Virtual Watershed is a prototype web-based agricultural watershed planning tool based on the Big Creek watershed in Southern Illinois

• Aimed at helping to explore and gain insight in to tradeoffs among – agricultural and environmental policies, – landowner decision-making processes, – and environmental and economic outcomes.

Big Creek watershed issues• Identified by ISWS as primary source of sediment

in the Lower Cache River (Demissie et al., 1992). • More than 70% of sediment inflows in to the

Lower cache based on 1985-1988 data (Demissieet al., 2001)

• Significant amount of nutrient pollution(NPS)

Integrated System

• The tool combines several important systems-related models – multi-objective optimization model

(evolutionary algorithms)– agent-based model, and – environmental/hydrologic simulation model

Conceptual Framework

Prices/Policy instruments/ Available technology

Land use, input , and management decisions at field/farm level1 to N

Land use pattern and distribution at Watershed level

Ecosystem service output

Societal Goal?(Tradeoff )

Economic output

Decision Environment

Scientists,Policy makers,Interest groups,

Public

Decision maker/Stake holder

Modify

Evaluate

Land managers/Agents

Hydrologic and biophysical characteristics

Translation from Concept to ModelFour essential modeling requirements:1. Representing the socio-economic driving forces or the decision

environment? Using scenario analysis and formulating scenarios expressed by relevant parameters

2. Representing farmers/farm operators response to specific decision environment? Using Agent based model

3. Simulating the economic and environmental outcomes? Using Environmental/Hydrologic simulation model.

4. Evaluating performance of each outcome? Using a tradeoff curve or Production Possibility Frontier (PPF)

Farmers as Agents

SWAT NSGA-II

Policy scenario formulation

Land use

+

PPF generator (Optimization) model

Biophysicalparameters

Agent based model

Hydrologic & Environmental model

PPF

Scenario parameters

Farm input and

Biophysicalparameters

Biophysicalparameters

Env/Ecooutcome SWAT

Integrated modeling Framework

Eval

uati

on

NSGA-II : Non-dominated Sorting Genetic Algorithm II (Deb, 2002)

Production Possibility Frontier (PPF)• The Production Possibility Frontier is a graph

that shows all the combinations of goods (or services) that can be produced at maximum efficiency given a set of inputs (resources, labor, etc.)

• PPF for Virtual Watershed constructed based on:– Two competing alternatives

• Production of Agricultural commodities (indicated by Crop production index) and

• Production of Ecosystem services (indicated by Hydrologic water quality index)

HydrologicSimulation Model

(SWAT)

MultiobjectiveOptimization Model(Genetic Algorithm)

•Crop yields•Nitrogen•Phosphorus•Sediment yield•Peak flow

•Crop type•Tillage practice• 3 year rotation PPF for Bigcreek watershed

0

200

400

600

800

0.000 1.000 2.000Cro

p pr

oduc

tion

Inde

x

Water quality index

PPF Generator Model for optimal land uses

200.000

300.000

400.000

500.000

600.000

700.000

800.000

0.000 0.200 0.400 0.600 0.800 1.000 1.200Cro

p pr

oduc

tion

Inde

x

Water quality index

A

C

Two dimensional PPF where each point represents a discrete land use pattern with considerably different levels of economic and ecological performance.

Legend

CornForestHay/PastureSoybeanUrban

Chart5

0.1573010946

0.1660285767

0.1359842875

0.1806213125

0.196610553

0.2778157813

0.2951489489

0.3210667813

0.3601440828

0.376745177

0.4595645451

0.4192886218

0.5727458248

0.5974770329

0.943961522

0.9651213797

1

Crop Index

Water quality index

Crop production Index

687.896

687.4392

683.5699

684.2851

683.401

654.1852

676.2807

683.7414

663.6782

650.848

686.8313

683.3808

538.0247

537.809

264.9954

254.12

236.26

Sheet1

New HWICrop Index

0.157687.896

0.166687.439

0.136683.570

0.181684.285

0.197683.401

0.278654.185

0.295676.281

0.321683.741

0.360663.678

0.377650.848

0.460686.831

0.419683.381

0.573538.025

0.597537.809

0.944264.995

0.965254.120

1.000236.260

Sheet1

Crop Index

Sheet2

Crop Index

Ecosystem Services Index

Crop Index

Sheet3

Agent Based Model

Endogenous•Labor & machinery

Exogenous variables(Scenario parameters)•Crop prices•Policies•Biophysical parameters

EnvironmentalModel

•Crop management•Field operations•Biophysical parameters

•Crop yields•Nitrogen•Phosphorus•Sediment yield•Peak flow•Production Index•Water quality index

Land useOutput

Land use Input

0

100

200

300

400

500

600

700

800

0.000 0.200 0.400 0.600 0.800 1.000 1.200

Cro

p Pr

oduc

tion

Inde

x

Water quality index

Scenarios1996-2002 farm bill

CRP rental increase

High commodity

•Crop yields•Nitrogen•Phosphorus•Sediment yield•Peak flow

•Crop Production Index•Water quality index

The management problem involves user determination of how policy (e.g., public subsidization and regulation) and price structures can be altered to provide incentives so that to move the landscape closer to the PPF through the improvement space.

Virtual Watershed Web Application Demo

• Virtual Watershed can be accessed at http://vws.erp.siu.edu:90/vws/

• Users define scenarios and submit through the scenario entry form

• Policy scenarios are represented by parameters like crop prices, CRP rental rates and level of soil loss

• Simulation results are then displayed in various views as maps, graphs, and tables .

Output

Land useLand useLand use

Use

r Int

erfa

ce

Fron

t-en

dBa

ck-e

nd 1

Web server(Apache)

•Raster files•Vector files

RDBMSPostgres/PostGIS

Config. Files

Web Map Service(Mapserver/Geoserver)Ba

ck-e

nd 2

Agent Based Model

SWAT Model

____ Openlayers/Extjs/JQuery/Openfl

Pyth

on c

ontr

olle

r

VWS Implementation Architecture: Front-end to Back-end interactionBack-end inter application communication

AJAX

Google map service Yahoo map service

AJAX

Python ontrollerinput

input

1

2

3. Hover

4. Click

Farm 19

Advantages of Web Applications No special configuration or hardware requirements for the

user. Lower costs. Centralized data is secure and easy to backup. Updates can be made quickly and easily. Information is accessible to a wide audience anywhere in the

world. Everybody has a browser. Familiar interface encourages use. Web-applications make collaboration easy, as basically

everyone is using one “instance” of an application. Because all activity takes place on your servers you can see

how people are using your application.

Source: http://www.pssuk.com/AdvantagesWebApplications.htm

Slide Number 1IntroductionVirtual WatershedBig Creek watershed issuesIntegrated SystemConceptual FrameworkTranslation from Concept to ModelIntegrated modeling FrameworkProduction Possibility Frontier (PPF)Slide Number 10Slide Number 11Slide Number 12Slide Number 13Virtual Watershed Web Application DemoSlide Number 15Slide Number 16Slide Number 17Slide Number 18Slide Number 19Slide Number 20Slide Number 21Slide Number 22Slide Number 23Slide Number 24Slide Number 25Slide Number 26Advantages of Web Applications