176
Decision Support for Ecosystem Management and Ecological Assessments
In contrast to goal-driven systems, INFORMS
(Perisho et aI., 1995; Williams et aI., 1995) is a
data-driven EM-DSS. Data-driven systems do not
require the existence of an explicit goal hierarchy.
Indeed, it is often the case that the only existing
goals are implicit goals that reside in the private
knowledge base of the manager. Data-driven systems begin with a list of actions that the user wants
to explore and search the existing landscape conditions, as reflected in the system database, to find
possible locations where these management actions
can be implemented.
Both approaches have their strengths and weaknesses. Goal-driven systems tend to be rather prescriptive. They require the user to follow a certain
sequence of events and force the user to make certain critical decisions in order to follow a predefined ecosystem management process. Data-driven
systems allow users more freedom to craft their
own process in an ad-hoc fashion; this provides
great freedom of action, but places all the burden
of knowing what to do and why to do it on the user.
Goal-driven systems, by definition, tend to ensure
that management actions move the landscape toward the specified desired future conditions by
committing to a particular ecosystem management
decision process. This reduces the utility of the
EM-DSS to that set of decision makers who wish
to use this particular decision process. On the other
hand, data-driven systems offer no guarantee that
the results of the sum of the actions have any
resemblance to the desired future conditions as defined by the strategic objectives. Data-driven systems, however, do allow competent and knowledgeable decision makers maximum flexibility in
the analyses that they perform and how they put
them together to arrive at a decision. Hybrid goaland data-driven systems may offer users the advantages of both approaches.
12.5.2 Functional Service Modules
The full-service EM-DSSs rely on specialized software service modules to add a broad range of capabilities (Table 12.1). Tools to support group negotiation in the decision process are both extremely
important and generally unavailable and underutilized. ARJGIS (Faber et aI., 1997) is the most fully
developed software available for this function.
IBIS, another group negotiation tool, is an issuebased information system that implements argumentation logic (the logic of questions and answers) to help users to formally state problems,
understand them, clearly communicate them, and
explore alternative solutions (Conklin and Begeman, 1987; Hashim, 1990). Vegetation dynamics
simulation models, both at the stand and at the landscape scale, provide EM-DSSs with the ability to
forecast the consequences of proposed management actions. Disturbance models simulate the effects of catastrophic events, such as fire, insect defoliation, disease outbreaks, and wind damage.
Models that simulate direct and indirect human disturbances on ecosystems are not widely available.
Although models that simulate timber harvesting
activities exist, they provide little, if any, ecological impact analyses, such as the effect of extraction on soil compaction, on damage to remaining
trees, or on the growth response of the remaining
tree and understory vegetation. Models that simulate the impact of foot traffic, mountain bikes, and
horseback riding on high-use areas are largely
missing. Models that simulate climate change, nutrient cycling processes, acid-deposition impacts,
and other indirect responses to human disturbance
exist, but are rarely practical for extensive forest
analyses. Stand- and landscape-level visualization
tools have improved dramatically in the last few
years. It is now possible, with relatively little effort, to link to and provide data for threedimensional stand-level models such as SVS (McGaughey, 1997) and landscape-level models such
as UVIEW (Ager, 1997) and SMARTFOREST
(Orland, 1995).
12.6 Interoperability in Ecosystem
Management DSS
Existing EM-DSSs (Table 12.1), with few exceptions, are islands of automation unable to easily
communicate with each other. They have been written in different software languages, they reside on
different hardware platforms, and they have different data access mechanisms and different component-module interfaces. For example, nongeographical databases may be written in Oracle,
geographical information system (GIS) databases
in ARCIINFO, knowledge bases in Prolog, and a
simulation model in C or Fortran. Some execute
only on a UNIX platform, others only in a Microsoft Windows environment. As a group, they
have poorly developed mechanisms for achieving
integrated operations with (1) other existing fullservice EM-DSS; (2) the many available functional-service modules (Schuster et aI., 1993;
Jorgensen et aI., 1996); (3) readily available, highquality commercial software; or (4) new software
modules that independent development groups are
continually producing in their efforts to support
ecosystem management.
Decision Support for Ecosystem Management and Ecological Assessments
In contrast to goal-driven systems, INFORMS
(Perisho et aI., 1995; Williams et aI., 1995) is a
data-driven EM-DSS. Data-driven systems do not
require the existence of an explicit goal hierarchy.
Indeed, it is often the case that the only existing
goals are implicit goals that reside in the private
knowledge base of the manager. Data-driven systems begin with a list of actions that the user wants
to explore and search the existing landscape conditions, as reflected in the system database, to find
possible locations where these management actions
can be implemented.
Both approaches have their strengths and weaknesses. Goal-driven systems tend to be rather prescriptive. They require the user to follow a certain
sequence of events and force the user to make certain critical decisions in order to follow a predefined ecosystem management process. Data-driven
systems allow users more freedom to craft their
own process in an ad-hoc fashion; this provides
great freedom of action, but places all the burden
of knowing what to do and why to do it on the user.
Goal-driven systems, by definition, tend to ensure
that management actions move the landscape toward the specified desired future conditions by
committing to a particular ecosystem management
decision process. This reduces the utility of the
EM-DSS to that set of decision makers who wish
to use this particular decision process. On the other
hand, data-driven systems offer no guarantee that
the results of the sum of the actions have any
resemblance to the desired future conditions as defined by the strategic objectives. Data-driven systems, however, do allow competent and knowledgeable decision makers maximum flexibility in
the analyses that they perform and how they put
them together to arrive at a decision. Hybrid goaland data-driven systems may offer users the advantages of both approaches.
12.5.2 Functional Service Modules
The full-service EM-DSSs rely on specialized software service modules to add a broad range of capabilities (Table 12.1). Tools to support group negotiation in the decision process are both extremely
important and generally unavailable and underutilized. ARJGIS (Faber et aI., 1997) is the most fully
developed software available for this function.
IBIS, another group negotiation tool, is an issuebased information system that implements argumentation logic (the logic of questions and answers) to help users to formally state problems,
understand them, clearly communicate them, and
explore alternative solutions (Conklin and Begeman, 1987; Hashim, 1990). Vegetation dynamics
simulation models, both at the stand and at the landscape scale, provide EM-DSSs with the ability to
forecast the consequences of proposed management actions. Disturbance models simulate the effects of catastrophic events, such as fire, insect defoliation, disease outbreaks, and wind damage.
Models that simulate direct and indirect human disturbances on ecosystems are not widely available.
Although models that simulate timber harvesting
activities exist, they provide little, if any, ecological impact analyses, such as the effect of extraction on soil compaction, on damage to remaining
trees, or on the growth response of the remaining
tree and understory vegetation. Models that simulate the impact of foot traffic, mountain bikes, and
horseback riding on high-use areas are largely
missing. Models that simulate climate change, nutrient cycling processes, acid-deposition impacts,
and other indirect responses to human disturbance
exist, but are rarely practical for extensive forest
analyses. Stand- and landscape-level visualization
tools have improved dramatically in the last few
years. It is now possible, with relatively little effort, to link to and provide data for threedimensional stand-level models such as SVS (McGaughey, 1997) and landscape-level models such
as UVIEW (Ager, 1997) and SMARTFOREST
(Orland, 1995).
12.6 Interoperability in Ecosystem
Management DSS
Existing EM-DSSs (Table 12.1), with few exceptions, are islands of automation unable to easily
communicate with each other. They have been written in different software languages, they reside on
different hardware platforms, and they have different data access mechanisms and different component-module interfaces. For example, nongeographical databases may be written in Oracle,
geographical information system (GIS) databases
in ARCIINFO, knowledge bases in Prolog, and a
simulation model in C or Fortran. Some execute
only on a UNIX platform, others only in a Microsoft Windows environment. As a group, they
have poorly developed mechanisms for achieving
integrated operations with (1) other existing fullservice EM-DSS; (2) the many available functional-service modules (Schuster et aI., 1993;
Jorgensen et aI., 1996); (3) readily available, highquality commercial software; or (4) new software
modules that independent development groups are
continually producing in their efforts to support
ecosystem management.