EPA /600/R-12/008
January 2012
Decision Support Framework Implementation of the Web-Based
Environmental Decision Analysis Application DASEES: Decision Analysis for
a Sustainable Environment, Economy, and Society
U.S. EPA Office of Research and Development Ecosystem Services
Research Program (ESRP)
-------
Notice
This report is based in part on work prepared by Neptune and Company, Inc. (contract EP-C-08-
007/ task order 45). The information presented and the views expressed herein are strictly the
opinions of the authors and in no manner represent or reflect current or planned policy by the
USEPA. Mention of trade names or commercial products does not constitute endorsement or
recommendation of use.
The correct citation for the document is:
Stockton, T.1, B. Dyson2, W. Houghteling1, K. Black3, M. Buchholtz ten Brink4, T. Canfield5, A.
9 f\ f\
Vega 'M.Small , A. Rehr 2011. Decision Support Framework Implementation of DASEES:
Decision Analysis for a Sustainable Environment, Economy, and Society. U.S. Environmental
Protection Agency, Cincinnati, OH, EPA /600/R-12/008
1 Neptune and Company, Inc. 1505 Suite B, Los Alamos, NM 87544
2 U. S. Environmental Protection Agency, National Risk Management Research Laboratory,
26 W. Martin Luther King Drive, Cincinnati, OH 45268
3 Neptune and Company, Inc., 8550 W. 14th Ave. Suite 100, Lakewood, CO 80215
4 U. S. Environmental Protection Agency, National Health and Environmental Effects Research
Laboratory, 27 Tarzwell Drive, Narragansett, RI 02882
5 U.S. Environmental Protection Agency, National Risk Management Research Laboratory,
919 Kerr Research Drive, P.O. Box 1198, Ada, OK 74820
6 Department of Engineering and Public Policy, Carnegie-Mellon University, Baker Hall 129
5000 Forbes Ave., Pittsburgh, PA 12513
II
-------
Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the
Nation's land, air, and water resources. Under a mandate of national environmental laws, the
Agency strives to formulate and implement actions leading to a compatible balance between
human activities and the ability of natural systems to support and nurture life. To meet this
mandate, EPA's research program is providing data and technical support for solving
environmental problems today and building a science knowledge base necessary to manage our
ecological resources wisely, understand how pollutants affect our health, and prevent or reduce
environmental risks in the future.
The National Risk Management Research Laboratory (NRMRL) is the Agency's center for
investigation of technological and management approaches for preventing and reducing risks
from pollution that threaten human health and the environment. The focus of the Laboratory's
research program is on methods and their cost-effectiveness for prevention and control of
pollution to air, land, water, and subsurface resources; protection of water quality in public water
systems; remediation of contaminated sites, sediments, and ground water; prevention and control
of indoor air pollution; and restoration of ecosystems. NRMRL collaborates with both public
and private sector partners to foster technologies that reduce the cost of compliance and to
anticipate emerging problems. NRMRL's research provides solutions to environmental
problems by: developing and promoting technologies that protect and improve the environment;
advancing scientific and engineering information to support regulatory and policy decisions; and
providing the technical support and information transfer to ensure implementation of
environmental regulations and strategies at the national, state, and community levels.
This publication has been produced as part of the Laboratory's strategic long-term research plan.
It is published and made available by EPA's Office of Research and Development to assist the
user community and to link researchers with their clients.
Cynthia Sonich-Mullin, Director
National Risk Management Research Laboratory
III
-------
Table of Contents
NOTICE II
FOREWORD Ill
LIST OF FIGURES VI
ACKNOWLEDGEMENTS VII
NOTE TO READERS VII
1 INTRODUCTION 1
1.1 Environmental Applications Research 1
1.2 Decision Support Framework (DSF) 1
1.2.1 Conceptual Model 1
1.3 Decision Support Systems (DSS) 3
1.3.1 Web-based Decision Support 3
1.3.2 Decision Analysis for a Sustainable Environment, Economy, and Society (DASEES) 4
2 DASEES WEB APPLICATION FRAMEWORK 4
2.1 Functional Design 4
2.2 Requirements 5
2.2.1 Site Navigation 5
2.2.2 Content Presentation 5
2.2.3 Graphical Tool User Interface (Ul) 5
2.2.4 Data Management 5
2.2.5 Data Analysis 5
2.3 Hardware and Operating System Requirements 7
2.4 Secondary Data Requirements 7
3 SYSTEM DESIGN 7
3.1 Overview 7
3.2 User Interface 8
3.3 Server Scripting 10
3.4 Relational Data Managment System (ROMS) 10
3.5 GIS Layers Management and Web Serving 10
3.6 Probabilistic Modeling and Statistical Analysis 11
3.7 Advantages of Design Using Open Standards 11
4 SOFTWARE DEVELOPMENT AND IMPLEMENTATION 12
4.1 Development Process 12
4.1.1 Module Skeleton 12
4.1.2 Module Interfaces 12
4.1.3 Tool Development 12
IV
-------
4.2 Deployment 13
4.3 Change Control and Archiving 13
5 QUALITY ASSURANCE (QA) AND REVIEW ACTIVITIES FOR
VALIDATION, VERIFICATION, AND TESTING 13
5.1 Review Cycles 13
5.2 Testing DASEES Components 13
5.2.1 Technical Guidance 14
5.2.2 Databases 14
5.2.3 Technical Analysis 14
5.2.4 User Interface 15
5.3 DASEES Peer Review 15
6 DOCUMENTATION, MAINTENANCE, AND USER SUPPORT 15
6.1 Documentation 15
6.2 Maintenance and User Support 16
6.3 Security 16
7 REPORTING 16
8 REFERENCES 18
9 GLOSSARY OF WEB APPLICATION DEVELOPMENT TERMINOLOGY 20
V
-------
List of Figures
Figure 1. DSF Conceptual Model (USEPA, 2009) 2
Figure 2. Generic Architecture for a Decision Support System (Adapted from Black and Stockton, 2009). 3
Figures. DASEES Webpage Design 6
Figure 4. DASEES Web Application Framework 9
Figure 5. Maintenance and User Support Workflow 17
VI
-------
Acknowledgements
We thank the contributors to the decision support framework (DSF) who provided input to the
conceptual model driving the development of the environmental web application Decision
Analysis for a Sustainable Environment, Economy, and Society (DASEES). Members of the
framework team include: Ann Vega, Pat Bradley, Dave Burden, Tim Canfield, and Verle
Hansen, U.S. Environmental Protection Agency; Ken Reckhow, Duke University; Mitch Small
and Amanda Rehr, Carnegie-Mellon University; Tom Stockton and Kelly Black, Neptune and
Company, Inc.
The project is currently funded by EPA's Ecosystem Services Research Program (ESRP). The
DASEES development Team includes Pat Bradley, John Carriger, Brian Dyson, Marilyn
Buchholtz ten Brink, Tim Canfield, of the U.S. Environmental Protection Agency. Neptune and
Company, Inc. is responsible for web-development under the direction of the U.S.
Environmental Protection Agency as part of contract EP-C-08-007. We thank Barbara Butler and
Douglas Grosse for their comments and suggestions for improving an earlier draft of the
document. Brian Dyson, USEPA NRMRL is a DASEES Project Co-Lead and is the
corresponding author for this report.
Note to Readers
The information in this report is intended for web-based application developers and those
interested in open source web-design. As such, the greater part of this report is of a technical
nature. For those finding the terminology unfamiliar, a glossary is included at the end of this
report.
VII
-------
1 Introduction
1.1 Environmental Applications Research
Pervasive environmental challenges such as climate change, coastal eutrophication, habitat and
species' loss are rarely the result of natural processes solely, but linked to human activities and
their resource management behaviors (Shultz, 2011). The implication of this is that solutions will
not always have a straightforward, "tame" application of science-based actions, but are often
"wicked" problems. "Wicked" problems (Balint, et al., 2011) are large-scale environmental
policy questions where environmental concerns, economic constraints, and societal values
conflict causing seemingly intractable political situations.
Over the last few decades, advances in information technology have enabled the environmental
sciences to evolve from relatively distinct scientific disciplines, e.g., geology, meteorology,
hydrology, etc., to a more inter-related systems-based science. When human behavior is factored,
the analysis of complex phenomena, such as climate change becomes more feasible (Dozier and
Gail, 2009). Dozier and Gail (2009) argue that these analyses drive the need for research into the
development of environmental applications that use systems approaches and decision science to
enable society to make decisions that address "wicked' problems.
1.2 Decision Support Framework (DSF)
The EPA Science Advisory Board (SAB) recommended that the EPA (U.S. Environmental
Protection Agency, 2000) should develop a decision-making framework for addressing such
complex multi-faceted problems. They specified that this framework should focus on the
interaction between analytic and deliberative processes in decision-making and be capable of
assessing cumulative risk, evaluating competing management options, including societal values,
and clarifying potential tradeoffs (United States, 1997, National Research Council, 1996). In
response, the Decision Support Framework (DSF) team developed a research implementation
plan (U. S. Environmental Protection Agency, 2009), with the goal of creating a flexible decision
framework for structuring complex problems that helps decision-makers and stakeholders
transparently evaluate scientific and technical analyses within an economic and societal values
context.
7.2.7 Conceptual Model
The DSF conceptual model envisions a five-step decision process of defining context,
identifying objectives, formulating alternatives, evaluating and deciding among alternatives, and
implementation (Fig. 1). Solicited advice from decision experts and review of literature
-------
identified four key overarching ideas that were considered important in developing a successful
decision-making model.
1. The framework should support a process that follows the concepts of behavioral decision
theory and decision analysis (French et al., 2009; Gregory and Keeney, 2002) to ensure all the
necessary elements are considered for making an informed decision.
2. The framework should use analytic-deliberation (National Research Council, 1996), which
combines rigorous, scientific approaches with an emphasis on communication to facilitate
mutual understanding among stakeholders.
3. The framework should encourage adaptive management as a strategy for managing complex
environmental systems having uncertainty (Williams, 2001).
4. The framework should use an accepted conceptual approach for establishing decision context
and problem framing. The Drivers-Pressures-States-Impacts-Response (DPSIR) conceptual
model (UNEP, 2007) is a widely recognized approach and was adopted for the DSF.
Goals &
Objectives
Values &
Preferences
Decision Scientific
Makers Input
\ \
Stake-
holders
Problem
Formulation
Alternative
Management
or Policy
Options
Adaptive
Management
Objectives
Options can be
used to manage
Pressures & States
Drivers are human needs
that generate Pressures
on the Ecological and
Environmental State
Options generate
implementation
costs that have
Impacts
Selected
Management
or Policy Action
Response
•f Societal Sustainability
^ Economic Sustainability
•s Environmental Sustainability
•/ Meet environmental regulations
•/ Acceptable level of uncertainty
Environmental, Economic,
Societal & Health Impacts
are evaluated against the
Objectives to generate a
Response
The change in the
Environmental State
generates Impacts
Pressures generate
stresses that change the
Environmental State
Figure 1. DSF Conceptual Model (USEPA, 2009)
-------
1.3 Decision Support Systems (DSS)
Implementing the DSF entails development of a computerized decision support system (DSS). A
generic DSS architecture (Fig. 2) must include a knowledge base, analysis tools, and inference
engine (Black and Stockton, 2009). Other components such as user interface and stakeholder
participation tools greatly improve the DSS function but are not considered fundamental.
Specific structures for a DSS are contingent on its intended purpose. An information-based DSS
(Black and Stockton, 2009) provides structures that support the intended functions of the DSF
conceptual model. Information-based DSSs maintain database architectures such as relational
and datatypes e.g., geographical information systems (GIS), of varying complexity, to provide
information appropriate to the DSS needs. They can incorporate a range of analytical tools, e.g.,
multi-criteria decision analysis (MCDA) to analyze database information.
Construction
Literature
Data Collection
Expert Judgment
Storage
Flat Files
ROMS
— ->
— < —
Retrieval
Search Engine
Data Mining
—
Knowledge
Base
Model
Building
Expert
System
Probability
Uncertainty Ana lysis
Multi-criteria
Decision Analysis
J
Inference
Engine
User Interface
f "^
Decision Support
^ J
Figure 2. Generic Architecture for a Decision Support System (Adapted from Black and
Stockton, 2009)
1.3.1 Web-based Decision Support
From its inception, the implementation of the DSF was envisioned to be a web-based application
of the principles and ideas elucidated by the framework conceptual model. The reasons for this
are technical and functional. Technically, there is a large and expanding resource pool of
-------
software and hardware options for making customizable DSSs and associated tools capable of
accessing information on the web (Black and Stockton, 2009). Of particular note is the rise of
open source developer options, such as the statistical programming language R (www.r-
project.org), that allow an online community of developers to contribute to the continual
development and maintenance of web-based decision support systems. Functionally, a web-
based DSS supports a key aim of the DSF of promoting stakeholder involvement and
participatory decision-making by using community-based deliberative processes to better resolve
"wicked" problems (Bayley and French, 2007; Insua et al., 2007; Panagiotopoulos et al., 2010).
1.3.2 Decision Analysis for a Sustainable Environment, Economy, and Society (DASEES)
Web-based decision and statistical analysis tool developers at Neptune and Company, Inc., have
facilitated the implementation of the Decision Support Framework through the web-based
application DASEES (Decision Analysis for a Sustainable Environment, Economy, and
Society). The following discussion is specific to the platform architecture and software design
approaches being taken by the developers.
2 DASEES Web Application Framework
2.1 Functional Design
DASEES is organized in five steps according to the DSF conceptual model:
1. Understand context
2. Define objectives
3. Develop options
4. Evaluate options
5. Take action
At the top level, DASEES will consist of a set of guidance and software tools designed both to
educate decision-makers in using this conceptual model and to allow them to create their own
decision-specific model using interactive tools that allow them to input data and generate graphs,
charts, and statistical analyses. By using these tools, different decision options can be quantified
and evaluated in the larger context of the conceptual model. In addition, DASEES will house
case studies that demonstrate how the tools and guidance can be applied to specific real-world
decisions. The case studies will be used as the building blocks for the upper levels of DASEES.
In terms of site navigation, the structure described above will be implemented as a series of tabs
(Fig 3). Each of these top-level tabs will contain sub-tabs housing tools useful in the decision
process. The "DASEES steps" will contain an overview tab, which will provide an introduction
-------
to the individual steps. Each of these sub-tabs may contain its own sub-tabs, housing guidance or
tools.
Requirements for each tab and sub-tab will depend on whether they contain guidance or house a
software tool, or both.
2.2 Requirements
2.2.1 Site Navigation
The web site should be organized such that different areas of the site are easy to find and access
(with minimal clicks). Different site sections must display promptly in response to mouse clicks.
It is important that the organization of the site must reflect the five steps of decision-making (see
Section 1.2.1) that form the basis of the DSF conceptual model.
2.2.2 Content Presentation
Site content should be clearly written and factually accurate. Guidance must be appropriate to the
overall conceptual model. Guidance on using the embedded software tools should address all the
functionality of the tool in question in language that is understandable by the user without special
scientific or technical knowledge.
2.2.3 Graphical Tool User Interface (UI)
Software tools should have interfaces that are dynamic, responsive, and intuitive. Interacting
with these tools should not lead to page reloads, require the user to go to a different page on the
site or open a new browser window.
2.2.4 Data Management
The user should be able to input data relevant to their decision process and have those data stored
in a persistent database. They should be able to modify or delete data after they have been
entered. Data input interfaces should be responsive and intuitive.
2.2.5 Data Analysis
DASEES should be able to perform technical analysis of user-entered data. This will include
performing statistical analysis of the data using Bayesian methodology and generating plots,
charts, and graphs of the data. Images generated by analysis tools must be displayed promptly
and without requiring the user to go to a different page on the site or to open a new browser
window.
-------
' Decisions and Ecosystems - Windows Internet Explorer provided by EPA
CJ^ -, T £3* http://zeus,nept:uneinc
arg/dasees/content/gijanica. html
File Edit View Favorites Tools Help
Links £j Customize Links
"Si & Si Decisions and Ecosystems
*
Decision Analysis for a
Sustainable
E ^Environment* Economy & Society
Guanica Bay Watershed
Getting Started , Step 1: Understand Context Step 2: Define Objectives Step 3: Develop Options Step 4: Evaluate Options ! Step 5: Take Action
Overview Decsion Landscape Interactive Map DPSIR Social Network
Understand Context
The first step in DASEES is to develop an understanding of the scope of the scientific and decision setting of the management problem.
Political, Regulatory, Social, and Institutional Setting
The decision context will be characterized through development of a decision diagram, showing the political, regulatory, social, and institutional setting of the
environmental management problem. This will provide participants with the critical context for their studies, including (for example): Are decision metrics specified by
law or prior agreement? Are management options limited to a set of predefined alternatives, or is there flexibility to propose new approaches? Do the various
stakeholders trust and utilize common sources for data and scientific assessment, or are there competing studies? Are mechanisms in place to include ecosystem
services and externality costs in economic accounts for project evaluation?
Social Network Analysis is a tool that can help to developing an understanding of how and with whom stakeholders, decision-makers, and the scientific community
interact.
Scientific Setting
The scientific setting will be addressed through the development of an information framework called DPSIR that shows the relationship between Driving forces (e.g.
need for food, demand for biofuels), Pressures (e.g., fishing, pollutant discharges). States (the response of environmental compartments and ecological variables).
Impacts (changes in ecosystem services, economic systems, social systems), and Responses (actions taken). The DPSIR process facilitates defining a broad set of
Responses from which specific potential management options can developed.
Decision Landscape
Taken together the Political, Regulatory, Social, Institutional and Scientific context provides the Decision Landscape of the decision. DASEES includes several tools
to aid in the building of the Decision Landscape
Local intranet
•\ 100%
Figure 3. DASEES Webpage Design
-------
2.3 Hardware and Operating System Requirements
Since DASEES is a web-based application, user hardware and operating system requirements are
minimal. DASEES will support any Web 2.0 compliant web browser that also supports Adobe
FLASH technology. However, due to resource constraints, DASEES only will be tested
comprehensively with Internet Explorer versions 7 and higher running on Windows and Firefox
version 3.5 and eventually on Windows and Mac OS 10.6.
DASEES will be served from computers running the Linux operating system, with a minimum of
2 Gigabytes of RAM and 300 Gigabytes available hard disk space.
2.4 Secondary Data Requirements
A key component of the DASEES approach is the propagation of uncertainty at each step of the
decision process. Data from a wide variety of different sources, as well as stakeholder values and
preferences, will be combined as part of the decision analysis process. Under the Bayesian
decision analysis paradigm (French et al., 2009) the quality of the information and data is
represented by its uncertainty. Highly uncertain information and data are essentially weighted
inversely to their uncertainty. Thus secondary data of low quality may be included with a high
degree of uncertainty but would have little weight in the analysis. The amount of uncertainty in a
particular area will help guide stakeholders and decision-makers in their decisions regarding how
better to utilize scarce resources when collecting additional data.
3 System Design
3.1 Overview
Note: The remainder of this report is of a technical nature specific to web design and open
source web development. For those not familiar with the terminology a glossary is included at
the end of this report.
DASEES is a web application framework containing a collection of linked decision analysis
tools to implement a comprehensive and coherent decision analysis framework for
environmental decision making. The DASEES tools are designed to be used independently or
within the DASEES 5-step decision analysis framework. For example, the Land User Scenario
tool allows a user to add or edit land use polygons and the associated attributes, to create a new
land use scenario. This tool can be used as a stand-alone GIS visualization tool or in conjunction
with the DASEES framework to inform watershed modeling land use input requirements.
This system uses open source technology to mirror the capabilities of a system constructed with
proprietary technologies such as S-PLUS, .Net, ArcGIS Server, and Flex. The resulting system
is open source, can be installed easily and run on any desktop computer (such that internet access
is not required to run the system) or server, is cross-browser, cross-platform compatible, and
-------
requires no browser plug-ins. This platform can incorporate a variety of GIS data types,
including ESRI shape files, Google Earth Keyhole Markup Language (KML), GeoJSON, and
flat files with geographical coordinates.
The web framework for DASEES tools leverages the Ext JS (JavaScript) Cross-Browser Rich
Internet Application (RIA) Framework to provide a desktop-like look, feel, and responsiveness
via its built-in support for Asynchronous JavaScript And XML (AJAX). Ext JS also has native
support for representational state transfer (RESTful) data storage for managing create, read,
update, delete (CRUD) interactions through PHP with the back-end database. The GeoExt
project provides an extension of the Ext JS RIA that brings together OpenLayers and Ext JS to
build desktop-style interactive mapping. Combining GeoExt with R, Geoserver, and
PostgreSQL/PostGIS allows for a full-featured, web-based, open source GIS and spatial
modeling system.
Due to recent advances, web-based visualization of information via vector-based/scriptable
graphs, charts, and map and data interaction can be dynamically built and presented on a web
page without requiring a plug-in. The HTML5 Canvas specification provides cross-browser
support for JavaScript-based graphics. Canvas is supported by the Internet Explorer 9, Firefox,
Chrome, and Safari web browsers. Backward capability for earlier versions of Internet Explorer
is provided through the ExplorerCanvas JavaScript library. Although not a requirement,
DASEES applications have been developed around the Google Earth plug-in.
Figure 4 depicts the overall client-server web application design for DASEES that will be
discussed in detail in the following sections.
3.2 User Interface
The DASEES graphical user interface (UI) will be programmed in JavaScript (officially known
as ECMAScript) using the ExtJS open source cross-browser JavaScript library. The ability of
individual UI components to interact asynchronously with the server allows page content to be
dynamically updated without reloading the page, providing a level of responsiveness more
commonly associated with desktop applications running on the user's computer than with
traditional web applications. Site navigation is implemented via a tab-based layout, which
allows users to drill down to more specific layers of content while still being able to visualize the
overall structure of the site.
The content portion of the web interface typically is created in the open source word processing
program Open Office and exported to HTML, while the JavaScript code will reside in separate
source files with a js file extension. Presentation and styling is controlled using Cascading Style
Sheets (CSS). CSS source code will also be stored in separate .ess files. All three kinds of files
will be served by the Apache web server in response to requests from the user's web browser; the
browser then executes JavaScript. The server can display dynamically generated images of plots
-------
and graphs created by the R module or interactively display data and information in one of the
many JavaScript based charting packages (e.g., Google Charts).
Client
Web Browser
User Interface
AJAX
Maps
Geoserver
CIS data layers
R
I
SQL
A.
Maps,
Data,
Graphics
PostgreSQL Server
PostGIS Module
PL/R Module
Project &
Stakeholder
Schemas
Figure 4. DASEES Web Application Framework
The interactive mapping interface will be developed in OpenLayers (www.openlayers.org), and
GeoExt (www.geoext.org). Openlayers provides an API for incorporating and interacting with
GIS data layers from several data sources and formats including Google Map base layers, Web
Map Service (WMS) requests from a map server, ArcEVIS requests, Google Earth Keyhole
Markup Language (KML), GeoJSON, and PostGIS SQL requests. This creates a flexible
environment for managing GIS data layers, by taking advantage of data products from external
sources (e.g., Google, EPA), and allowing for data layers to be dynamically generated, analyzed,
and updated. DASEES currently is using Geoserver, an open source map server application
-------
(www.geoserver.org), as its map server. However, map server applications that support standard
specifications generally are interchangeable.
3.3 Server Scripting
PHP: Hypertext Preprocessor (PHP) is an open source server-side scripting language, which will
run as a plug-in module to the open source Apache web server. The PHP module will implement
the model and controller components (the view is implemented in the user interface) of a
model/view/controller (MVC) architecture, which will allow users to interact directly with tables
in the PostgreSQL database. Supported operations will include displaying table data and
creating, deleting, and editing rows in the table. The PHP module also will allow the web
interface to directly call stored procedures in the database. These stored procedures will be
written in PL/R. PL/R is a PostgreSQL Procedural Language for R, a statistical computing,
programming, and graphics language environment (www.r-project.org) that allows R functions
to be stored and run in the PostgreSQL database using SQL.
3.4 Relational Data Managment System (RDMS)
Databases for the DASEES will be managed in PostgreSQL, a highly scalable, Structured Query
Language (SQL) compliant open source object-relational database management system. Data
will be organized using schemas. Schemas provide namespacing, which allows multiple tables
with the same name to exist as long as they are in separate schemas. This provides great
flexibility in database design. For example, multiple users each can have a set of tables that are
identical in structure but contain user-specific data. Schemas also can be used on a per-project
(rather than a per-user) basis. GIS and statistical technical analysis of the data will be
implemented in stored procedures written in PL/R, a procedural language for PostgreSQL that
provides an interface to the R statistical programming language. All geospatial data incorporated
into DASEES will be Federal Geographic Data Committee (FGDC) compliant.
3.5 GIS Layers Management and Web Serving
GIS data layers will be developed and stored natively as ESRI shapefiles. This will allow static
GIS data to be developed under a Windows operating system. ESRI shapefiles will be imported
into Geoserver or PostGIS. Geoserver is used to serve maps that are static, while PostGIS is
used to store spatial information that is being dynamically generated or edited by the user. The
advantage of PostGIS is that open source spatial analysis tools available in R Packages can be
called directly, interact with spatial data, and return results to the database through SQL requests
from the AJAX/PHP/Apache engine. This allows for a responsive interactive web-based spatial
analysis environment that can take advantage of the wide array of open source analysis tools
available in R.
10
-------
GIS and statistical analysis, computations, and algorithms will be accessed or programmed in R.
R code will be accessed from PL/R functions written as stored procedures in the PostgreSQL
database. The key workhorse packages for GIS analysis in R will be sp and rgdal. These
packages provide the tools needed to provide full GIS functionality in a scriptable, open source,
web interactive framework. Another potentially a key package for the GIS analysis in DASEES
will be RSAGA. RSAGA provides an interface from R to the System for Automated
Geoscientific Analyses, SAGA (www.saga-gis.org). SAGA is an open source spatial analysis
tool designed to provide 1) an effective implementation of spatial algorithms, 2) an approachable
user interface with many visualization options, and 3) run under Windows and Linux operating
systems.
3.6 Probabilistic Modeling and Statistical Analysis
Bayesian probabilistic modeling and statistical analysis of data will be implemented in R using
existing graphical modeling R packages such as gBase and gRain, Monte Carlo integration R
packages such as MCMCpack, visualization R packages such as googleVis, and, when needed,
custom DASEES R code. An influence diagram interface will provide the user the ability to
build, specify and simulate Bayesian models connecting options and measures. Again, the
foundation for the technical analysis is the PL/R functionality that connects the web interface,
the underlying database, and the computations and visualizations in R.
3.7 Advantages of Design Using Open Standards
There are several advantages of designing DASEES using Open Source Software (OSS)
standards. Use of OSS is advantageous in terms of software licensing and cost, and in terms of
the philosophy that underlies the Open Source community. The Open Source philosophy is
aimed at sharing information at all levels, gathering and responding to feedback for continuous
improvement, and encouraging users to supply functionality and content. For DASEES, this
consists of sharing content and all resources (including code), operating a continuous feedback
option, and encouraging users to submit case studies that can be shared with the DASEES
community of users.
The benefits of following the Open Source approach to the development of DASEES include:
• Potential users are not encumbered with software licensing issues.
• DASEES will allow users to engage at various levels of complexity depending on their
interest. Different levels of complexity for each component might include, for each
component: an overview, access to supporting information, details of mathematical
methods (for example, environmental modeling, risk assessment, statistics, economic
modeling, and decision analysis), computer code (for example, HTML, JavaScript,
PHP, R), and case study examples.
11
-------
The holistic approach facilitates defensibility, traceability and transparency for each
application.
The approach promotes Quality Assurance QA through internal testing and through
user-supplied feedback. The open nature of the JavaScript, PHP, and R code allows
each aspect of DASEES to be reviewed thoroughly.
As a web-based system, DASEES can be updated as new tools, technologies, and
approaches become available without requiring users to install any additional
software.
4 Software Development and Implementation
4.1 Development Process
4.1.1 Module Skeleton
DASEES developers will create an initial navigational, scripting, database, and technical analysis
structure that reflects the system design as depicted in Figure 4. The user interface skeleton will
be developed in HTML and JavaScript, the Model/Controller skeleton will be developed in PHP,
the Database skeleton will be developed using SQL and PL/R, and the Technical Analysis
skeleton will be developed in R. These skeleton components will provide containers for
implementing specific functionality required by individual tools.
4.1.2 Module Interfaces
Once skeleton modules have been created, interfaces will be developed to allow communication
among the modules. The User Interface will communicate with the Model/Controller via AJAX
requests; the Model/Controller will communicate with the Database via direct SQL queries and
calls to PL/R stored procedures; and the Database will communicate with the Technical Analysis
module via R code embedded in PL/R stored procedures. These interfaces will be placeholders
that will be replaced by task-specific communication mechanisms as individual tools are
developed.
4.1.3 Tool Development
Once the skeleton infrastructure is implemented, tool development can begin. Each tool will
have a specification for its user interface and functionality. This will involve coding in multiple
modules. While some tools will not use all modules, no tool will be confined to a single module.
When a new tool is integrated into the overall framework all existing tool test plans (see Section
5) will be run to ensure that the new code introduced to support the new tool does not
compromise existing functionality.
12
-------
4.2 Deployment
The DASEES system will undergo a three-phase deployment process: development, beta-testing,
and deployment. During development phase, each of the new functions and components will be
continuously tested as outlined in Section 5. At the end of the development phase, the new
functions and components will be deployed on a test server. In the beta-testing phase, an internal
review of the content and function of the new DASEES functions and components is conducted.
After that review is complete and identified concerns have been addressed, a further review of
those new functions and components is conducted by members of the DSF team. The DSF team
reviewers provide the peer reviews needed for the EPA clearance process. The contractor
(Neptune and Company, Inc.) then will address DSF team comments. The EPA will complete
the clearance process. Once EPA approves the modified system in the beta-testing phase, the
running system is deployed onto the deployment server and made available at www.dasees.org.
4.3 Change Control and Archiving
DASEES will be managed under Subversion (subversion.tigris.org). Subversion is an open
source version control system that keeps track of changes to files and folders in a project
repository. It will act as a central repository of the most up-to-date project files and will allow
users, contributors, and developers to share them conveniently. All major releases of the
DASEES system, source code, and documentation will be tagged and archived in Subversion. In
addition, the entire Subversion system will be backed up on a daily basis and an off-site backup
is performed weekly.
5 Quality Assurance (QA) and Review Activities for Validation,
Verification, and Testing
5.1 Review Cycles
Implementing the DASEES decision analysis steps in a web-based user interface requires a
complex system of information, databases, analysis tools, and decision analysis integration tools.
It is essential that each DASEES component receive appropriate review prior to being deployed
on the public DASEES website. These reviews will be performed for each of the DASEES
components (guidance, databases, technical analysis tools, and user interface). Sections 5.2 and
5.3 provide details on the specific review for each DASEES component.
5.2 Testing DASEES Components
This section outlines the QA testing plan that is applied to each of the components of DASEES.
Review questions to promote critical thinking will be completed by using a QA user protocol.
The review questions are comprised of a series of yes/no questions for each of the four review
categories. These review questions are outlined in the sections below. Technical QA review will
be performed by members of the DSF team.
13
-------
5.2.1 Technical Guidance
The quality objective for the technical guidance is to provide appropriate guidance using
consistent and standard terminology. Affirmative answers to the following table of questions
should ensure that quality objectives are met.
• Does the guidance use consistent and standard technical terminology?
• Does the guidance present appropriate interpretation of technical concepts?
• Are technical terms explained adequately?
• Are technical concepts explained adequately for the purposes of the software system?
• Is the guidance user friendly?
• Will the guidance be understood by the user?
5.2.2 Databases
The quality objective for the database searches is to produce accurate and anticipated results.
The quality expectations of the databases will be assessed by answering each of the following
questions:
• Are the database query results appropriate and accurate given the user input?
• Is data entered by the user accurately stored in the database?
• Are queries completed in a timely manner?
5.2.3 Technical Analysis
The quality objective for the technical computations is to produce accurate and anticipated
results.
Technical computations will be performed in R. In general, the QA provided by R developers
and the R user community provide a high level of accountability for the D ASEES technical
engine. In most cases, computer codes that need to be compiled include documentation with
example datasets and results. Once a DASEES component has been programmed, these
examples will be run and compared to the documented results. Each DASEES function will
include an example dataset with known results. Comparison to known results will confirm that
the functions are performing as expected. The quality expectations for the technical analysis will
be assessed by answering each of the following questions:
• Do the technical analysis results compare favorably with the known results from the
example datasets?
• Has the technical tool or method had sufficient QA to provide adequate defense of the
data analysis results?
14
-------
• Are assumptions and limitations described in such a way that the end user can make
informed decisions regarding use of the method?
• Are examples of appropriate use given?
5.2.4 User Interface
The quality objective for the user interface is that it be intuitive, user-friendly and that it
facilitates use of and interaction with technical analysis tools. The interface will be designed to
intuitively elicit user responses and provide for easy navigation. User responses are collected
through user interfaces and then are translated using JavaScript, PHP, and R to database search
or technical analysis appropriate inputs.
The quality expectations of the user interface will be assessed by answering each of the
following questions:
• Does the user interface use standard terminology and definitions throughout?
• Can the user easily navigate between different sections of the website?
• Are inputs validated (error-checked) and does the user interface provide feedback to
the user on appropriate input for the tool if errors are found?
• Does the interface have guidance to aid the user?
5.3 DASEES Peer Review
Prior to the deployment of the DASEES website a Letter Peer Review will be conducted by
EPA. EPA will select a peer group from the identified stakeholders, develop review questions,
and mail these questions out to the peer group for review. The peer group will then review the
DASEES website based on these review questions.
6 Documentation, Maintenance, and User Support
6.1 Documentation
DASEES electronic documentation provides information on:
• How to use DASEES
• The technical components and methods DASEES employs
• General guidance
• The Quality Assurance Proj ect Plan (QAPP)
The documentation will be updated regularly to reflect functional changes in DASEES. The
most recent EPA approved version of the documentation will be made available on the public
15
-------
(deployed) DASEES website. In addition, since DASEES is open source software the underlying
software code will be available for download.
6.2 Maintenance and User Support
As a web-based application, DASEES has the advantage of centralized maintenance. If defects in
the software are discovered, fixes can be applied directly on the public site without the need to
distribute software patches to users. If a defect is identified, either by the DASEES team or by
reviewers or users, the procedure described in the flowchart below (Fig. 5) will be followed to
identify the problem, implement a fix, and eventually deploy the fix on the public site.
6.3 Security
DASEES is developed on Windows, Linux, and Macintosh machines located at Neptune's Los
Alamos, New Mexico and Lakewood, Colorado offices. Both offices are secured by software
firewalls, anti-virus software, and a combination of two to three Anti-Spyware programs for all
Windows servers and desktops. Anti-Virus definitions are updated daily on all machines. Anti-
Virus scans are run daily on servers and weekly on desktops. As the deployment site, the Los
Alamos office has an added layer of security in the form of a hardware firewall. All servers and
desktops at Los Alamos are located behind the hardware firewall.
The servers used for DASEES development, beta-testing and deployment and the server that
hosts the Subversion source control system are Linux servers running the Fedora operating
system. Linux was chosen as the operating system for deployment and testing because it offers
superior performance to Windows and is more secure because the majority of viruses and
malicious software is written specifically to target Windows. Thus, Linux is immune to these
threats. Updates on Linux are automatically downloaded and installed nightly. The servers also
implement a RAID-5 configuration to protect from hard drive failure and use Uninterrupted
Power Supplies. All major releases of the DASEES system, source code, and documentation are
tagged and archived in Subversion. In addition, the entire Subversion system is backed up on a
daily basis and an off-site backup is performed weekly.
7 Reporting
The final deliverable for this project is an operation DSF. The operational DSF is now called
DASEES and will be housed under www.dasees.org. Additionally, the DSF team members will
prepare several peer reviewed journal articles based on DASEES, case study development, and
interactions with stakeholders and decision-makers. DASEES will be Open Source Software and
all code and documentation will be provided through a version control web site that will allow
anonymous checkout of all the of DASEES' components.
16
-------
Yes
ew information
available?
Gather more information
Defect Reported
Log or update issue in tracking system
Assign developer to reproduce
reported behavior
r— No —
-------
8 References
National Research Council. 1996. Understanding Risk: Informing Decisions in a Democratic
Society. P. C. Stern and H. V. Fineberg, eds., National Research Council, Washington, DC.
U. S. Environmental Protection Agency, 2000.Toward Integrated Environmental Decision-
Making. EPA-SAB-EC-00-011. ed,. Office of Research and Development, Washington, DC.
U. S. Environmental Protection Agency, 2009. U. S. EPA Office of Research and Development
Ecosystem Services Research Program (ESRP): Decision Support Framework (DSF) Team
Research Implementation Plan. EPA/600/R-09/104. Cincinnati, OH.
UNEP, 2007. Global Environmental Outlook GE04. United Nations Environment Programme,
Nairobi
United States. 1997. The Presidential/Congressional Commission on Risk Assessment and Risk
Management Framework for Environmental Health Risk Management. Final Report. V. 1, ed.,
United States.
Balint, P.J., Stewart, R.E., Desai, A. and Walters, L.C., 2011. Wicked Environmental Problems:
Managing Uncertainty and Conflict. Island Press, Washington, D.C.
Bayley, C. and French, S., 2008. Designing a Participatory Process for Stakeholder Involvement
in a Societal Decision. Group Decision and Negotiation, 17:195-210.
Black, P. and Stockton, T., 2009. Basic Steps for the Development of Decision Support Systems
in Decision Support Systems for Risk-Based Management of Contaminated Sites, A. Marcomini,
G. W. Suter II, A. Critto eds. Springer.
Dozier, J. and Gail, W., 2009. The Emerging Science of Environmental Applications in The
Fourth Paradigm: Data-Intensive Scientific Discovery, T. Hey, S. Tansley, K. Tolle, eds.
Microsoft Research, Redmond, WA.
French, S., Maule, J. and Papamichail, N., 2009. Decision Behavior, Analysis, and Support, 1st
ed., Cambridge University Press, New York.
Gregory, R.S. and Keeney, R.L., 2002. MAKING SMARTER ENVIRONMENTAL
MANAGEMENT DECISIONS 1. JAWRA Journal of the American Water Resources
Association, 38:1601-1612.
18
-------
Panagiotopoulos, P., Gionis, G., Psarras, J. and Askounis, D., 2011. Supporting public decision
making in policy deliberations: an ontological approach. Operational Research, 11:281-298.
Rios Insua, D., Kersten, G., Rios, J. and Grima, C., 2008. Towards decision support for
participatory democracy. Information Systems and E-Business Management, 6:161-191.
Schultz, P.W., 2011. Conservation Means Behavior. Conservation Biology, 25:1080-1083.
19
-------
9 Glossary of Web Application Development Terminology
AJAX
Apache
Asynchronous JavaScript And XML contains a consortium of client-side web
application development methods to allow for asynchronous sending and
receiving of data form a server.
Apache HTTP Server is open source web server software usable by a wide
variety of operating systems.
API
Application programming interfaces are specified source code to interface
between software components.
ArcGIS Server ESRI server for supplying web-based GIS mapping and spatial analysis
services.
ArcIMS
Arc Internet Map Server is an ESRI web map server.
Canvas
An element of HTML5 allowing dynamic, scriptable renderings of two-
dimensional shapes and bitmap images.
CRUD
Create, read, update, delete are the four basic functions of persistent storage in
relational database applications.
CSS
Cascading Style Sheets are a simple method for adding style (fonts, color,
etc.) to web documents.
ECMAScript
Ext JS
Flex
ECMAScript is the name of the scripting language standardized in ECMA-
262. It is used for client-side scripting for the web with JavaScript being a
well-known version.
A JavaScript Rich Internet Application (RIA) Framework has a library of
tools supporting development of cross-browser interactive web applications
using techniques such as AJAX.
An Adobe software development kit for developing cross-platform web
applications using Adobe Flash.
20
-------
FGDC
Federal Geographic Data Committee is an interagency entity promotes the
national coordinated development, use, and sharing of geospatial data.
GeoExt A project that combines ExtJS with OpenLayers and provides a suite of
customizable widgets and data support for making geospatial data
applications.
GeoJSON An open source format for encoding geographic data structures. It is based on
JSON (JavaScript Object Notation).
GeoServer Open source Java-based server for editing and sharing geospatial data.
Designed for interoperability, it can be used to connect existing information
to web-based maps in OpenLayers and GoogleMaps.
HTML Hypertext Markup Language uses elements which are the basic building
blocks of webpage design.
HTML5 The fifth revision of the HTML standard. It is intended to improve HTML to
support multimedia while maintaining human and machine-readable formats.
HTML5 should include JavaScript.
JavaScript A recognized version of the ECMAScript language standard used mainly for
client-side development of website user interfaces.
JSON
KML
JavaScript Object Notation is a text-based open standard using JavaScripting
language for transmitting data over a network connection. It can serve as an
alternative to XML.
Uses XML syntax to make visual and geographic annotation to web-based
maps.
MVC
Model-View-Controller is a software engineering architecture that permits the
independent development of different aspects of the application logic.
.NET
A Microsoft Windows-based software framework that facilitates web-based
application development through software language interoperability.
OpenLayers An open source JavaScript library and lays the foundation for displaying
maps in web browsers via application programming interfaces (API).
21
-------
PHP
Open source server side scripting language suitable for web application
development and can be embedded in HTML.
PL/R
PL/R is a PostgreSQL Procedural Language for R that allows R functions to
be stored and run in the PostgreSQL database using SQL.
PostGIS
Open source software supporting geographic objects in the PostgreSQL
object-relational database structure.
PostgreSQL A highly scalable, Structured Query Language (SQL) compliant open source
object-relational database management system.
R
Open source version of S, a statistical computing, programming, and
graphics language
RESTful REST is a software architecture style exemplified by the internet structure of
origin servers, gateways, proxies, and clients. RESTful implies conformity
with the existing REST architecture.
S A statistical programming language for data analysis and graphics.
SQL
Structured Query Language is a programming language for managing data in
a relational database management system.
S-PLUS
Commercial version of the S programming language by Tibco Software.
XML
WMS
Extensible Markup Language defines internet coding formats that are both
human and machine-readable. Its open source specification follows the
World Wide Web Consortium (W3C) international standards.
Web Map Service is an internet standard for geo-referenced images generated
via map server with GIS data.
22
------- |