This guide is designed to help you startup, explore, and evaluate the Tracking and Analysis Framework (TAF) model. It assumes no prior experience with Analytica®, the modeling environment in which TAF has been implemented.

Sections in this guide explain how to begin using the TAF module, how to navigate the TAF model, how to select policy scenarios, and how to evaluate nodes within the model.

What is TAF?

Welcome to the Tracking and Analysis Framework (TAF) Model. It was designed, generally, to inform discussions on environment and policy and, specifically, to support the National Acid Precipitation Assessment Program (NAPAP) during the 1996 assessment of the impacts of the 1990 Clean Air Act Amendments. TAF is an example of an integrated assessment, incorporating information and expertise from across the United States into a single model. Structurally, it is composed of several modules (smaller models), each of which was developed by experts in their respective fields. By incorporating the work of experts across many disciplines, integrated assessment aims to create a tool that informs the policy process with the best and most recent science of the day. Because integrated assessment has not been attempted before on this scale, TAF is as much a demonstration of the power of integrated assessment as it is a policy tool.

TAF is implemented in Lumina Decision Systems' Analytica® Decision Modeling Software. Both a Macintosh (preferably a Power Macintosh) and a copy of the Analytica Decision Modeling software are required for using and viewing the TAF Model.

Getting Started

System Requirements

Macintosh Systems

• PowerPC processor
• System 7 and up
• A minimum of 4 MB of RAM

Windows/Intel Systems

• At least a 486 processor
• Windows 95, 98, or NT 4.0
• A minimum of 8 MB of RAM

Figure 1: TAF Model Top Level Influence Diagram

The TAF Model uses influence diagrams to graphically display model form and flow. Each variable is represented by a node. Arrows between nodes illustrate functional relationships and give you a sense of how information flows through the model.

The Policy Selector module is the primary input for the TAF Model. Since it determines which of the 18 emissions projections will be evaluated, it directly influences the Emissions Projections module and the Compliance Cost module. Similarly, the Atmospheric Pathways module takes the emissions trajectories of the Emissions module and calculates how the atmosphere will redistribute the emissions species across North America. The Effects modules, Soils-Aquatics, Health, and Visibility modules use the deposition and species concentration outputs of the transport module to estimate the effects. Finally, the Benefits Valuation module puts a dollar value on each of those effects. Thus, you can see how different emissions policy decisions might affect the environment and human health.

TAF is designed in Analytica using hierarchical influence diagrams. The bold rectangular nodes with rounded corners represent modules, which contain influence diagrams of their own. To see an outline form of the module hierarchy, click on the Outline Button of the toolbar. Like folders containing files and other folders on a disk, a module contains nodes and other modules. Clicking on the triangle to the left of a module title causes the contents of the module to be listed beneath the module title (see Benefits Valuation below). Clicking again on the triangle causes the listing to disappear. Also, by clicking in the box next to Modules Only, you can view the hierarchy with only modules shown, all other nodes are excluded.


Figure 2: TAF Model Outline Window

To close any windows in Analytica, either select Close from the File pulldown menu or click in the small square box in the upper left-hand corner of the window. Close the Outline window.

Module Diagram

To view a module's influence diagram double-click on the module's node. For example, you can open the Atmospheric Pathways module by double-clicking on its node.

Figure 3: Atmospheric Pathways Module Influence Diagram

Four of the nodes most commonly used in TAF are displayed in Figure 3. The non-bold, rounded-corner, rectangular nodes, such as Seasonal SO2 Emissions by Source Region, represent deterministic variables. These variables usually contain data tables or equations that are functions of other variables. The Atmospheric Transport node is yet another module, and the oval-shaped nodes are stochastic variables. The parallelograms at the bottom of the diagram are indexes used in defining the dimensions of other variables in the module.
This influence diagram is typical of the diagrams in each module of the TAF Model in that it lists all module inputs on the left, encapsulate the "guts" of the module in the center, and lists the module outputs on the right.

Object Window

To open a node's Object window, you can double-click on the node, or click on the node once and then click on the Object Button of the toolbar. Double- clicking on a module node opens its Diagram window instead. The object window in Figure four is displayed by double-clicking the node Seasonal Average Ambient Species Concentration.


Figure 4: Seasonal Average Ambient Species Concentration's Object Window

The Object window provides the following detailed information about a variable.

Class:

Is the variable certain or uncertain, an index for other variables, or an objective quantity to be minimized or maximized?

Identifier:

This is the abbreviated name that is used by Analytica to refer to this node.

Units:

Such as $ millions, tons/year, micrograms/cubic meter.

Title:

Up to 36 chars. Comprehensible, mixed upper/lowercase.

Name:

Up to 20 chars. Must be unique, without spaces or punctuation, starting with a letter. Should be reasonably comprehensible.

Description:

A text description of what this variable represents, so that it is unambiguous.

Definition:

A single equation (like a spreadsheet cell entry), or a table containing data or equations.

Inputs:

A listing of the variables (represented by nodes in Analytica) that are used to compute the value of the current node.

Outputs:

A listing of the variables (represented by nodes in Analytica) that incorporate the outputs of the current node.

By clicking on the Edit Table button you can open an Edit Table window and display the contents of the table. You can close the Seasonal Average Ambient Species Concentration Object window by single-clicking the box on the left of the window's menu bar. You may then return to the top level TAF Model Diagram by either closing the Atmospheric Pathways Diagram window or clicking on the Parent Diagram Button of the toolbar.

Balloon Help

To view a node's description field in a Diagram window, activate Balloon Help by selecting Show Balloons in the Balloon pulldown menu, located in the upper right-hand corner of the screen. After activating Balloon Help, a balloon containing the contents of a node's description field will appear whenever the mouse pointer is placed over a node. To de-activate Balloon Help, you can select Hide Balloons from the Balloon pulldown menu.

Public Index Library

The Public Index Library is an example of a library node; it contains the indexes used by variables in multiple modules in the TAF Model. To see the various indexes of the Public Index Library, you can double-click on its node in the TAF top-level diagram.

Figure 5: Public Index Library Diagram Window

Each of the parallelogram-shaped nodes represent an index of labels or numbers used in the TAF Model. You may view the definition of an index (or any other node) by clicking once on the index node and then clicking on the Definition Button of the toolbar.

Policy Selector

The Policy Selector is the chief input node for the TAF Model. Double-click on the Policy Selector module to open its Diagram window. The Policy Selector asks you to select two emissions policies: a baseline emissions trajectory and a scenario emissions trajectory. In the course of evaluating the model, the two policies will be compared against one another. Double-click on the Baseline Policy Selector module to choose a baseline emissions trajectory. You will find that the Baseline Policy Selector Diagram consists of 18 separate buttons, corresponding to the eighteen emissions projection options. Each of the lower 16 buttons corresponds to an emissions trajectory based on three variables:

Policy Option (No Title IV, Title IV with trading, Title IV without trading,

Beyond Title IV -- Title IV refers the emissions-relevant section of the 1990 Clean Air Act Amendments),

Growth Rate (Low or High Growth in electricity demand),

Average Power Plant Retirement Age (40- or 60-years).

For more information about the specific definitions of these variables, see the Emissions Module Documentation in the TAF Modellers Guide.

The EPA Emissions projection is based on the emissions projections of the Environmental Protection Agency. The User-Defined Emissions projection permits the user to choose one of two methods for generating a unique emissions trajectory.

Figure 6: Baseline Policy Selector Diagram Window

To choose an emissions trajectory, simply click on one of the buttons while in Browse Mode (i.e., using the index-finger mouse pointer, which you select from the floating toolbar).

User-Defined Emissions

The TAF Model was designed to be flexible in its emissions projections. If none of the proposed seventeen emissions projections are acceptable, you may opt to set your own emissions using the User-Defined Emissions button. First, you should click on the User-Defined Emissions button of the Baseline Policy Selector, then close the Baseline Policy Selector and the Policy Selector Diagram windows to return to the TAF Model top level diagram. Finally, double-click on the Emissions Projections module.

Figure 7: Emissions Projections Module Influence Diagram

In Figure 7 you can see the User Emissions Settings module. Double-click on it to open the user interface for creating a user-defined emissions trajectory.

Figure 8: User Interface for Setting Custom Emissions Projection

The TAF Model offers two methods for generating a customized emissions projection: Base Plus Growth and Population Driven. The first method, Base Plus Growth, uses 1990 emissions data as the base year emissions. From there, the model calculates emissions for subsequent years based on annual percentage growth rates in emissions. The growth rates may be set at several levels of specificity. The Average Growth Rate input box enables you to set an average, general growth rate that is applied to all emissions in all states in every year. The Half-Decade Growth Rate Delta enables you to make an additive adjustment the Average Growth Rate at five year time intervals. The NOx and SOx Delta allows for yet another additive adjustment to the growth rate(s) depending on emissions species. Finally, the State Growth Rate Delta give you a matrix of states, years, and emissions species in which to enter additive adjustments. The choice of specificity is left to the user. None or all of the adjustments may be used since their effects are additive. To reset all of the deltas back to zero, click on the Reset Deltas to 0 button located below the State Growth Rate Delta Edit Button.

The Base Plus Growth method for defining a customized emissions projection is selected in the method choice pulldown menu (see Figure 7). Use the pulldown menu to select the second method for defining an emissions projection, Population Driven. This method allows you to make predictions about energy consumption per person and then lets census bureau population growth predictions determine the emissions trajectory. The 1990 emissions per capita data is used as a reference. Change in Energy/person and Change in Emissions/Energy are both additive effects that allow you to modify emissions per capita in five year time intervals. To reset the change matrices back to zero, click on the Reset Changes to 0 button.

After defining either growth rates or emissions per capita parameters, you may wish to see how your emissions trajectory looks. To do this, close the User Emissions Settings window and open the User-Selected Emissions module in the Emissions Projections module.

Figure 9: User-Selected Emissions Module Influence Diagram

To see a summary of your user-defined emissions scenario, click once on Scenario Emissions and then click on the Results button of the toolbar.

Figure 10: Scenario Emissions Results Window

Any variable in the TAF Model may be evaluated using the procedure described above.

Sample Size

Close the Scenario Emissions Results Window, the User- Selected Emissions module, and the Emissions Projections module to return to the TAF Model top level diagram. An important contribution and strength of the TAF Model is in the incorporation of uncertainty. A Median Latin Hypercube method is used to generate and model the uncertainty in TAF. The default sample size is 25. If you wish to raise or lower the sample size, then select Uncertainty Options from the Result pulldown menu.

Figure 11: Uncertainty Options Window

Click in the Sample Size box and change the number to whatever sample size you wish. Click on Set Default. If you do not see a field for sample size when you select Uncertainty Options from the Result pulldown menu, make sure that the Analysis option pulldown menu inside the Uncertainty setup dialog box is set to Uncertainty Sample.

Helpful Hints

If you find that you are unable to get a result for a variable due to lack of memory, you may not have set a high enough memory allocation for Analytica. Exit Analytica by selecting Quit from the File pulldown menu and find the Analytica application icon on your hard drive. Click one on it and select Get Info from the File pulldown menu. Change the Preferred size setting to 16000 (if you have 16 megabytes free RAM in your system).

Figure 12: Analytica Info Window

If you require assistance operating Analytica, you can contact Lumina Decision Systems Support at

Lumina Decision Systems, Inc.

59 North Santa Cruz, Suite Q

Los Gatos, California 95030

USA

Telephone: (408) 354-1841

Electronic Mail: support@lumina.com

World Wide Web: http://www.lumina.com

If you have any questions about specific models, then contact the individual modellers at their respective institutions:

Argonne National Laboratories

9700 S. Cass Ave. Bldg. 900

Argonne, IL 60439-4812

Foster Wheeler Environmental

Northwoods II

8101 N. High St., Suite 260

Columbus, OH 43235

Resources for the Future

1616 P Street, N.W.

Washington, DC 20036

Carnegie Mellon University

Department of Engineering and Public Policy

Pittsburgh, PA 15213-3890