The following words are defined according to how they are used within CRAFT. Our definitions may differ somewhat from how others use them. Many definitions are consistent with how they are used in DP 2.0.

Acceptable limit: The range of variability that can be tolerated within a specified spatial and temporal context.

Action alternative: The set of projects or activities that are undertaken to achieve a particular measurable end. See No-action alternative.

Active management: A management philosophy that emphasizes active engagement in ecosystem processes to achieve desired ends. See passive management.

Activity: Specific project actions that are undertaken to achieve a particular measurable end.

Acyclic graph: A graphical model of cause and effect in which return pathways, or loops, do not exist. Most influence diagrams and belief networks must be acyclic in order to calculate probabilities. For more information see the Building Belief Networks webpage.

Adaptive management: A type of environmental management in which decisions are made through an on-going process of implementation, monitoring and evaluation.

Adverse effect: An outcome of an action alternative or the no-action alternative that is inconsistent with one or more measurable objectives.

Alternative action: The individual or set of management activities that are being proposed or implemented within a project.

Alternative design: The process of creating and combining activities into alternative actions to accomplish objectives.

Alternative refinement: Reconsideration of the alternatives under consideration following preliminary effects analysis.

Ancillary models: Secondary models that can be used in conjunction with a belief network. A large number of ancillary vegetation and fire models are available.

Bayesian belief network (BBN): A graphical and mathematical representation of causal pathways and relationships for conditional probability analysis. They are represented using nodes and uni-directional arrows. Belief networks are similar to influence diagrams, but technically speaking, only influence diagrams have decision nodes. In CRAFT, we refer to both as belief networks. For more information see the Building Belief Networks webpage.

Cause and effect: In the context of an influence diagram or BBN, the direction of influence between two or more nodes using a uni-directional arrow. Cause and effect may be direct or indirect.

Commonalities: In a belief net, similar projected effects that result from alternative actions. For more information, see the Background Synthesis Material webpage.

Comparative risk assessment: A formal evaluation process that compares the relative risks of a range of alternatives and the no-action alternative. The process may compare risks over a range of spatial and temporal scales.

Conceptual model: A mental or graphical representation of a modeled system. In CRAFT, the different components and cause and effect relationships of conceptual models are reflected in influence diagrams or belief networks. For more information, see the Conceptual Modeling webpage.

Conditional probability: When two or more factors or causal variables affect another (i.e., a daughter node) within a belief network , the condition of the daughter node is contingent on the values of both causal nodes.

Conditional probability table: Within a belief network node, the supportive table or matrix that includes all possible combinations of categorical values from two or more nodes. For more information, see the Building Belief Networks webpage.

Consequences: The end result or outcome of an Action or No-action alternative.

Cumulative effect: An environmental effect which results from the additive or interactive influence of multiple activities. For more information, see the Cumulative Effects webpage.

Decision: A choice among alternative actions for solving a problem or leading to desirable change.

Decision analysis: A collection of techniques for eliciting and integrating uncertainty and choosing among alternatives. CRAFT emphasizes uses of Bayesian belief networks. For more information see the The Building Blocks of CRAFT webpage.

Decision process: The decision process includes setting objectives, specifying alternatives, analyzing effects and choosing a preferred alternative.

Decision protocol: The procedure, process or steps followed to make a decision. CRAFT's decision protocol is modeled after Decision Protocol 2.0. For more information see The Building Blocks of CRAFT webpage.

Decision support: Tools and techniques for making improved decisions. For example, CRAFT is a decision support system that helps clarify objectives and predict consequences.

Desired future conditions: A comprehensive set of objectives for a place that describe and integrate preferred future attributes.

Disturbance: An episodic phenomenon that causes measurable change in an environmental system, such as a forest fire, flood, wind storm or logging activity. For more information see the Vegetation Change webpage.

Duration: The length of time that a process or cumulative effect will last. Understanding duration is critical for analysis of tradeoffs over multiple scales.

Ecological Risk Assessment: A process to assess risks that relate to the environment. For more information see The Building Blocks of CRAFT webpage.

Effect: The difference between current and/or expected future values of a measure in light of the action and no-action alternatives. See cumulative effects.

Endpoint: In CRAFT, the terminal node in an effects assessment that tiers to lower-level objectives in the objectives hierarchy. The effects of alternatives are measured based on how this node changes.

Evaluation: The means by which effects are weighed or compared.

Exogenous factor/variable/node: A phenomenon that originates outside the system and has an effect on assessment endpoints. For example, fire weather results from hemispheric or global-scale climate processes.

Expert judgment/opinion: The estimation or prediction of a measure through the informed opinion of one or more specialists.

Feasibility: The ability of an action to be accomplished based on technical, financial, logistic, legal or other barriers to implementation.

Fundamental objectives: Upper level goals for a location in an objectives hierarchy that are achieved through specific means objectives.

Hazard: The chance of an event or an outcome occurring. For example, the hazard of severe fire increases as fuel accumulates and/or the weather gets dryer. See risk.

Implementation: The process of carrying out the action alternative on the ground.

Independent driver node: Within a belief net, chance parent nodes that are independent of the influence of other nodes in the network.

Indicator: A measurable attribute that represents some larger feature of interest. For example, individual "indicator species" might be used to measure "forest health." See measure.

Influence diagram: A graphical representation of cause and effect relationships consisting of boxes and uni-directional arrows. A probabilistic influence diagram (PID) provides a mathematical representation of the conditional strength of relationships. Influence diagrams are similar to belief networks, except belief networks need not have a decision node. For more information, see the Building Belief Networks webpage.

Interdisciplinary team (ID team): An analysis team composed of members from different disciplinary specialties.

Late Successional Reserve (LSR): An area of old growth set aside by the Northwest Forest Plan in California, Oregon or Washington.

Likelihood: A relative measure of probability, but unlike probability, it is not restricted to a 0 to 1 scale. For example, the probability of a crown fire in a stand within the next five years might be 0.1, but because of high fuel conditions, crown fire might have a greater likelihood than surface fire.

Magnitude: The relative scope or extent of the outcome of a proposed activity or natural disturbance.

Means objectives: Lower level goals or objectives in an objectives hierarchy that are designed to achieve one or more broader fundamental objectives. These may tier to or constitute the activities within an action alternative.

Measure: The quantifiable scale used to describe an environmental attribute. See endpoint.

Model: A conceptual, graphical, or mathematical representation of social or environmental relationships. For example: an objective hierarchy provides a structured model of community or agency goals. A fire-successional model consists of qualitative or quantitative assumptions of how vegetation changes over time. Environmental relationships are modeled in CRAFT using influence diagrams or BBN.

Monitoring: The systematic collection, tracking and interpretation of the effects of action and no-action alternatives. Monitoring is an essential part of adaptive management.

Natural variability: The inherent change or variation of a phenomenon that can be measured or described across temporal or spatial scales. In CRAFT, this variability is captured using probability distributions.

No-action alternative: The management alternative that involves no change in the level or nature of activity, also known as the status quo. See action alternative.

Node: Within an influence diagram or belief net, a variable that affects or is affected by other variables. There are three types of nodes: decision nodes, chance nodes and value (or utility) nodes. For more information, see the Belief Network Modeling webpage.

Objectives: The general and specific goals established for a particular management unit. Place-specific objectives can include sustaining existing environmental attributes or achieving desired future conditions. Project-specific objectives tier to measurable objectives within the objectives hierarchy.

Objectives hierarchy: A hierarchical structuring of place-specific objectives so that precise measurable objectives are embedded under more general overarching objectives. For more information see the Objectives Hierarchy Tutorial webpage

Passive management: A management philosophy that emphasizes that human intervention is unnecessary to achieve desired ends. See active management.

Preferred alternative: Among action and no action alternatives, the management option that best achieves measurable objectives, given the effects analysis and utility valuation.

Problem: The condition occurring when one or more realized or likely outcomes are unsatisfactory.

Probability: The degree of belief that a proposition is true measured from 0 to 1. For more information see the Primer on Risk and Uncertainty webpage.

Process: CRAFT's models environmental processes using belief networks. The decision making process includes specifying objectives, designing alternatives, analyzing effects and synthesizing results.

Reference conditions: A range of conditions against which the effects of past and future actions can be compared. Forest managers often define reference conditions as the ecosystem state before fire exclusion or the arrival of an invasive species. They may or may not be desired future conditions. For more information, see the Reference Conditions and Background Levels webpage.

Reference values: Precise measures against which the effects of actions can be compared. They may or may not be reference conditions. For more information, see the Indicators, Measures and Reference Values webpage.

Risk: The expected loss based on the probability that an event or outcome will occur and its value. See hazard. For more information, see the Primer on Risk and Uncertainty webpage.

Risk assessment: A formal process of evaluating risk within a specified spatial and temporal scale. See comparative risk assessment.

Risk-averse: A decision-maker's preference for minimizing uncertainties that result from a lack of knowledge and minimizing undesirable consequences. See risk-neutral and risk-taking.

Risk-neutral: A decision-maker's preference for using the long-term probabilistic average of expected outcomes. See risk-averse and risk-taking.

Risk-taking: A decision-maker's preference for the most desirable outcome even when other alternatives may be better based on the long-term probabilistic average. See risk-averse and risk-neutral.

Scale: The extent and time frame of modeling within CRAFT. Long-term and broad-scale analyses are typically under-analyzed in analyses.

Scenario: In belief network modeling, the different states of a chance node that are independent of the alternative action. For more information, see Background Synthesis Material webpage.

Sensitivity: The responsiveness of a node in a belief network to change from one or more parent nodes.

Sensitivity analysis: In CRAFT, the process of testing the belief network model to determine the conditional probability of different outcomes that result from different action and no action alternatives and the different assumptions of chance nodes.

Stakeholders: Individuals or groups with interest in the place being managed and/or the action and no action alternatives.

Synthesis: The final step in CRAFT's four-stage process that distills the results of the prior steps and presentation of results into a coherent summary. For more information, read the Synthesis webpage.

Threshold value: The value beyond which the consequences of the action or no-action alternative is sharply altered or unacceptable. For more information, see the discussion of Thresholds of concern in the Cumulative Effects webpage.

Tradeoff analysis: Within a belief net, formal comparison of how different action and no action alternatives will affect different measurable values or objectives of concern.

Uncertainty: The degree to which something is unknown or not understood. This results from both variability and ignorance. Phenomenon that are inherently variable, such as future weather, are not readily predictable. Uncertainty that results from ignorance can be more readily reduced through the acquisition of new knowledge. For more information, see the Primer on Risk and Uncertainty webpage.

Utility: Within a belief net, utility values allow modelers to compare different measurable objectives. It allows for comparison of disparate things such as treatment costs and wildlife habitat. For more information, see the discussion of Utility in the Building Belief Networks webpage.

Value: The worth, usefulness or importance of an object or service. specific level of a measurable attribute. For more information, see the Primer on Risk and Uncertainty webpage.

Back to Top

Home ~ Getting Started ~ Resources ~ Wizard ~ Site Map
I. Specifying Objectives ~ II. Designing Alternatives ~III. Modeling Effects ~ IV. Synthesis

Version 1.0
Last updated: July 14, 2005