Review of Information Technologies for Consideration in Comprehensive Resource Assessments of Forests

Kim Malafant and Stuart Davey, Bureau of Resource Sciences
February 1996

Table of Contents

Executive Summary
Terms of Reference
Introduction
The CRA Process and Integration Project
- CRA Integration Assessment
- Product Specification
Classification of Technologies Considered in the Review
Models, Methods and Concepts
- ANUSPLIN
- ANUDEM
- ANUCLIM
- CODA
- IHACRES
- Environmental Domain Analysis
- PVA Analysis
- Irreplaceability
- CAARNET
Management Support Systems
- LUPIS
- DIVERSITY
- FORUM
- MONASH
- FORPLAN/IFPS
- TRACS
- ArcForest
Integration Framework Systems
- GAMS
- Calyx
- Facet
- Whatif?
- Other Systems
Other Tools
- GIS Tools
- Image Processing Tools
- Visualisation Tools
- Mathematical and Statistical Tools
- Database
References
- Cited References
- Other References
Appendix 1. CRA Integration Project Review Proforma

Executive Summary

This review provides some of the information to facilitate the decisions of the Commonwealth Integration Technical Working Group (ITWG) on an appropriate framework for Comprehensive Regional Assessment (CRA). The integration of the technologies required for a CRA should provide a modular computing structure which can be applied across a number of agencies, with great flexibility.

This review considered a number of methods, concepts, algorithms, programs and packages currently under development or use by State and Commonwealth agencies. Most of the systems reviewed form building-blocks of a larger integrated framework, where these building-blocks or components can be changed for one of their competitors or for a new component as research develops. Most of these technologies were not complete systems for the CRA process.

Terms of Reference

The Commonwealth's Integration Technical Working Group (ITWG) agreed at its meeting on the 19th of January to commence a review of the technologies that could be incorporated into the integration model:

• the review was to be undertaken by BRS (Kim Malafant) with the support of the ITWG members and commonwealth technical agencies
• the review was to be completed in five weeks (21 February 1996)

The scope of the review was to include internationally and nationally developed technologies encompassing individual assessments (environmental, economic, resource etc.) and technologies associated with integrating assessments, including trade-off, landuse allocation, reserve selection, visualisation, scenario and optimisation modelling.

Technologies were to be assessed for:

• purpose (conceptual framework, method, functionality)
• compatibility with the UNIX platform
• ease of importing data and input data requirements
• output information and compatibility with other systems
• reliability and transparency of technology
• suitability for incorporation into a modular architecture
• availability of documentation, critiques and reviews
• costs associated with and time required for implementation
• fitness for purpose for use in CRAs

The assessment criteria were divided into five broad functional areas:

• Purpose of the technology
• System description
• Data description
• Integration requirements
• Use in CRAs (fitness for purpose)

A series of question or requirements against each of these five broad areas was developed for use in discussions with interested groups. This proforma is presented in Appendix 1. The ITWG assessed the use of technologies within the context of CRAs.

Introduction

"The complexity of the environment, the time and spatial scales involved, and the diversity of environmental effects are such that the implications of decisions affecting the environment are not always readily apparent. This complexity and the number of stakeholders involved makes it difficult to arrive at decisions which accommodate all wants and needs" (Qureshi et al. 1995).

This quote holds not only for making ecological decisions, but also for the reviewers of the technologies involved. It is not possible to provide an answer that is acceptable to all. So it is with this review. So first a few disclaimers:

1. Given the short timescale for the review it is unreasonable to expect that ALL possible technologies can be reviewed. What has been reviewed are the technologies that are generally known, either by the reviewers or from members of the taskforce and their colleagues. As such the technologies will show a bias, but hopefully many of the technologies currently being used will have been included in the review.
2. We have provided considerably more information on the integration tools than on a number of other technologies. This is in the belief that these tools or toolkits are currently not well known or understood in Australia. Such tools are becoming used for large ecological and management systems in other countries and especially in particular areas or sectors eg. Defence Departments.
3. The TOR included international technologies. This has been attempted in the short time-scale but is in no way complete. The review has had to draw on personal knowledge of reviewers to identify groups, companies and individuals that may have appropriate technologies.
4. Like all computer-based technologies, applications and systems change rapidly. Some disappear or are modified dramatically. Others appear and have little initial documentation or available information. We have drawn on the best information for the chosen technologies available - some are incomplete, while others have changed or have been modified from the available information.
5. The short timescale for the review as well as the unavailability of certain parties has precluded discussions with all interested parties. Those groups that have been visited have either been relatively local, or have been available. Some groups not personally visited have provided background information. This is especially true for the international material for which we have had to rely on a number of international contacts, E-mail and the World Wide Web. A list and summary of discussions with groups visited is presented in Appendix 2.
6. Where a technology has been the subject of a recent review we have summarised the review results and indicated the reference(s) rather than rewriting the review into this document.

This review is seen as an initial attempt at identifying immediate ways forward and areas that require further development and research for integrating CRA assessments.

This review aims to provide some of the information to facilitate the decisions of the Commonwealth Technical Working Group on an appropriate framework for achieving the integration of the technologies under a modular computing structure which can be applied across a number of agencies, with great flexibility.

The CRA Process and Integration Project

Comprehensive Regional Assessments (CRAs) will be undertaken in state forested regions. These assessments will lead to regional forest agreements between the States and the Commonwealth. The CRA process involves the application of two main streams of assessment: environment and heritage, and economic and social. Currently, there is no appropriate mechanism to integrate these assessments to enable the recognition and consideration of all forest values in forest use decisions. Without such integration, all values may not be considered in scenario development and decision making, which may lead to preferred options not being considered.

A key problem with the current non-integrated approach is the amount of time required for agencies to give technical advice on revisions to forest use scenarios. There are additional problems with proof of transparency and consistency in their advice. The scope of decision makers to interactively explore forest use options is limited.

The aim of the integration project is to determine the best options to integrate the environmental, resource, ecological sustainable management, heritage, economic and social assessments in forest planning, conservation and industry development. A specific objective of the project is to develop an information technology framework which:

• integrates the outcomes of the assessments being developed by the different technical agencies for the CRA assessments;
• is portable across the CRA regions (within and between States) to enhance consistency in assessments;
• permits greater transparency and, hence, adds credibility to the assessments;
• is flexible to changes in approach which may be adopted by agencies in their assessments;
• is capable of comparing real world spatial environmental, resource, economic and social implications of different resource use options; and
• includes a user friendly interface for the interactive examination of and experimentation with a range of forest use scenarios.

CRA integration assessment

The integration assessment will consist of a data gathering, storage and reporting system for all available data necessary to evaluate resource use options, including the development of spatial visualisation tools and information technology systems. The assessment will adopt a land use planning approach in which a number of forest land units across a region will be represented as consisting of a range of economic (social and reserved) and environmental/heritage attributes. For each forest unit there are a number of allowable land use and management options permitting the calculation of environmental values and sustainable yields under different silvicultural regimes. Assessments will be undertaken in an ESFM context.

For each forest use system (reserve system plus off-reserve management system) there are a number of values. Among these are biodiversity, wilderness, recreation and forestry values. Depending on the availability of information on the economic and environmental attributes of forests, various decision models may be used to explicitly compare alternative options. Where it is feasible to quantify other forest use and non-use values these may be directly incorporated. The decision models chosen will also be related to the important issues and concerns of specified CRAs. That is to say that the integration model will need flexibility to cater for variation across CRAs.

The information technology framework proposed by the ITWG uses a modular computing approach which could potentially be run across a number of computer systems, hence, providing great flexibility in the system. Reflecting the Commonwealth's delineation of CRA tasks, each agency will be responsible for the development of its own assessment methods. The modules which implement these methods would use a common GIS which contains the forest attributes required by the various modules and necessary for graphical representation of forest uses.

Product Specification

The aims of the development of an integration model (or system) for forest management should include the provision of a real time and reliable tool to enable regional level trade-offs, optimisation and scenario modelling, and have the capacity to consider off-reserve management, resource use and industry development options. Successful development of the model will have implications for the assessment of options in terms of forest sustainability, reserve design and industry structure. Credibility, transparency and consistency across States will be valued elements of the model.

Product outputs sought include:

• Data presentation in an easy to understand format, incorporating visual aids such as graphs and maps.
• Quick response time, allowing interactive exploration of 'what-if' questions.
• A system which will allow constraints on forest use to be included.
• Marketable product in a policy sense.

Data outputs from the model to include but not be constrained to:

• Overall area reserved.
• Biodiversity indicators.
• Old Growth reserved.
• Wilderness indicators.
• Forest and forest industry employment.
• Changes to regional and state income.
• Economic and social impact.
• Forest and forest product production.
• Environmental Impact.
• Water catchment.
• National Estate.
• Industry Development Options.
• Ecological Sustainable Forest Management (ESFM) options

A more complete outline of the CRA project with special reference to the Integration Project and its guidelines is found in the accreditation document.

Classification of Technologies Considered in the Review

The technologies considered by this review can be divided into five broad categories:

1. Models, methods or concepts which are technologies or systems that present either a concept, method or algorithm for the production of one, or a small set, of analyses. These may be available in a "coded" form or simply a mathematical description of the algorithm. These form the basic building blocks or modules of more complicated analyses and modelling frameworks.
2. Management support systems which integrate a mix of models and concepts, generally to solve a specific set of requirements or tasks. These may consist of a mix of the modelling category and framework systems, or may in fact be developed "from scratch". Integration framework systems which have few inherent models, but facilitate the integration of various models or analysis tools into a single coherent framework. They provide the "glue" to form complex modelling systems from many diverse sources. They may have their own modelling paradigm which may be very dissimilar to the paradigms used in the other categories.
3. Sustainable yield models or calculations.
4. Other Tools which include Geographic Information Systems (GIS), Databases, Image Processing systems, visualisation and mathematical and/or statistical analysis tools.

Models, Methods and Concepts

1. ANUSPLIN

• Purpose: ANUSPLIN is a FORTRAN based package for fitting surfaces to noisy data as a function of one or more variables. The package contains methods for interrogating the fitted surfaces in both point and grid form. It also includes methods for calculating standard error surfaces. The package contains 11 routines of which two, SPLINA and SPLINB, and their variants are the major application routines. These routines fit partial thin plate smoothing splines to multivariate noisy data. A brief overview of the basic theory and some applications is presented in Hutchinson (1991), with a more extensive discussion of the algorithms and the statistical analysis associated with them given in Hutchinson and Gessler (1994) and Hutchinson (1993, 1995). This package has been primarily developed for the fitting of climatic surfaces where we generally have latitude and longitude and perhaps elevation. More detail and description of the package can be obtained from http://cres.anu.edu.au.
• System Description: ANUSPLIN is normally distributed as binary executables for either the SUN UNIX or MS Windows based systems. For conversion to other platforms then (possibly) the FORTRAN code can be made available for conversion. The package uses the LINPACK library (Dongarra et al. 1979) for linear algebra routines. The current limits on data size vary between 600-2000 for UNIX versions and 350-1000 for DOS versions.
• Data Description: Input is accepted from the terminal or a command file and is routine dependent. Output is to the standard terminal or to an output file and is again routine dependent. A set of test command files with expected output files are included for testing the package after installation.
• Integration Requirements: This package is a general purpose spline smoothing package which could be used as a basic building block when needed. To achieve this use however on systems other than the supported both the source code and some conversion effort would be required. To any "normal" UNIX system this effort is likely to be minimal, since the SUN FORTRAN compiler implements the FORTRAN 1977 standard. The software is supplied under license from CRES, ANU and costs $600 for Academic users and $1000 for Standard users. This is for a single site license and the license agreement stipulates non-transferability of the package.
• Use in CRAs: Climatic surfaces will be important inputs into resource and biodiversity models.

2. ANUDEM

• Purpose: ANUDEM produces accurate digital elevation models with sensible drainage patterns from small sets of elevation and stream lines. The program calculates values on a regular grid of a discretised smooth surface fitted to large numbers of irregularly spaced elevation data points, contour line data and stream line data. The program imposes a global drainage condition which removes sinks where possible. The algorithm implemented by the program interpolates elevation data onto a regular grid by minimising a roughness penalty function on the fitted grid values and by simultaneously imposing constraints which ensure connected drainage structure and the proper representation of ridges and streams deduced from input contour line data.
  Imposing these global drainage constraints can significantly increase the accuracy, especially from the drainage property point-of-view, of digital elevation models from small or sparse sets of surface data. The size of data sets can be smaller and reduces the expense of capturing primary elevation data. It also reduces the need for detailed manual editing of interpolated grids. More detailed descriptions of the methods can be found in Hutchinson (1988, 1989) and Hutchinson and Dowling (1991). Further details on the program and methods cans also be found on http://cres.anu.edu.au.
• System Description: ANUDEM is normally distributed as binary executables for either the SUN and HP UNIX or MS Windows based systems. For conversion to other platforms (possibly) the FORTRAN code can be made available for conversion. The program uses two non-standard FORTRAN constructs: the inclusion of messages after the FORTRAN stop statement and the use of LOGICAL*1 for short integers. These constructs may need to be modified for some systems. The package implements dynamic memory allocation using the malloc/free procedures support on SUN and HP UNIX systems. Thus limits on "size" for the package will be dependent on the individual configuration (virtual memory) of the UNIX system. The package allows linkages with ARC/Info for input/output and it is planned to provide this linkage for IDRISI and GRASS systems. Future enhancements include the ability of handling breaklines and cliffs, the implementation of plan curvature and the rigorous treatment of geographic - latitude and longitude - coordinates.
• Data Description: Input is accepted from the terminal or a command and output is also to the standard terminal or to an output file. The input data files may be one of six types:
  1. point elevation data: ordered x,y,z triples which give the position and elevation of each data point.
  2. sink point data: ordered x,y,z triples which give the position and height of sink points.
  3. stream line data: strings of x,y coordinate pairs arranged in descending order and identified by a header containing the number of x,y pairs in each streamline.
  4. boundary polygons: strings of x,y coordinates identified by a header containing the number of pairs.
  5. contour lines: strings of x,y coordinate pairs identifying the contour line and preceded by a header containing the number of pairs and the contour line elevation.
  6. lake boundaries: a polygon file in the same format as the boundary polygons file.
  Data in standard ARC/Info format can be read directly by ANUDEM without modification. Output files can be written in lattice or x,y,z format, ASCII or binary grid formats for ARC/Info are also supported. When ARC/Info format is selected for output diagnostic files are also written in this format.
• Integration Requirements: This package is a general purpose digital elevation model package which could be used as a basic building block when needed. To achieve this use however on systems other than those supported both the source code and some conversion effort would be required. To any "normal" UNIX system this effort is likely to be minimal, since the SUN FORTRAN compiler implements the FORTRAN 1977 standard. The availability of direct import/export of data and output files in ARC/Info (and IDRISI?) formats allows for a rapid generation and integration of DEM(s) into a GIS for further analysis. ANUDEM has been incorporated into the ARC/Info package as the TOPOGRID command. The software is supplied under license from CRES, ANU and costs $600 for Academic users and $1000 for Standard users. This is for a single site license and the license agreement stipulates non-transferability of the package.
• Use in CRAs: DEMs are an integral component of the planning base and will provide input into models across a number of assessments.

3. ANUCLIM

• Purpose: The ANUCLIM package is the combination of the three packages ESOCLIM, BIOCLIM and BIOMAP. They have been combined by using a common front-end and common subroutines. Also included in ANUCLIM is the GROCLIM program, an extension of the GROWEST package (Nix et al. 1977). ANUCLIM writes a command file in response to user input, which is then used as the control or input file to the other packages. This package simplifies the generation of the input for ESOCLIM, BIOCLIM and GROCLIM.
  Although the programs have been combined a quick description of each is of value:
  a. The BIOCLIM package consists of two programs, BIOCLIM and BIOMAP. BIOCLIM is a bioclimatic prediction system which uses bioclimatic parameters derived from mean monthly climate estimates to approximate energy and water balances at a given location (Nix 1986). An outline of the system and its use are provided in Busby (1986).
  BIOMAP is used primarily as a grid matching program using as input two separate runs of BIOCLIM. The program tests if cell values of the two grids are within defined statistical bounds of each other.
  b. ESOCLIM provides an estimate of climate and can interpolate mean monthly climate from surface coefficient files for a maximum of 16 climatic variables.
  c. GROCLIM is an extension of the GROWEST package, a simple generalised model of crop response to light, thermal and water regimes. GROWEST calculated weekly indices of light, temperature, moisture and growth for up to four different plant types based on supplied climate data. GROCLIM produces the same indices based on estimated monthly mean climate data.
  Further detailed information, including example runs, on the ANUCLIM package can be found in McMahon et al. (1995) and from http://cres.anu.edu.au.
• System Description: All programs in the ANUCLIM package are normally distributed as binary executables for either the SUN UNIX or MS Windows based systems.
  A windowed menu system version is planned for the future.
  ESOCLIM, BIOCLIM and GROCLIM all require climate information based on surface coefficient files produced by the ANUSPLIN package (Hutchinson 1989). They also require a digital elevation model which can be provided by the ANUDEM package (Hutchinson 1988, 1989, 1991).
• Data Description: Input is accepted from the terminal or a command file and is routine dependent. Output is to the standard terminal or to an output file and is again routine dependent. A set of test command files with expected output files are included for testing the package after installation.
• Integration Requirements: This package is a general purpose command file builder for the ESOCLIM, BIOCLIM and GROCLIM packages. The ANUCLIM program is of limited use by itself, however, the other parts of the package could be used as a basic building blocks when needed. The software has been used in land capability studies for plantations (Nix et al. 1992).
  To achieve this use however on systems other than those supported both the source code and some conversion effort would be required. To any "normal" UNIX system this effort is likely to be minimal, since the SUN FORTRAN compiler implements the FORTRAN 1977 standard. The software is supplied under license from CRES, ANU. Costs of each package are unknown.
• Use in CRAs: Climatic surfaces and growth response models will be important inputs into resource and biodiversity models.

4. CODA

• Purpose: The CODA system - Conservation Options and Decision Analysis - is designed to provide a framework and a set of tools to aid in the regional conservation planning process. It allows for the building and comparison of alternative conservation proposals. The system uses a simple stepwise procedure - the iterative or minimum set algorithm - to attempt to identify the smallest or least costly set of selections that fulfills the user requirements. The CODA procedure is an explicit method of exploring the flexibility of networks, allowing the display of alternatives to selected sites and the calculation of the degree to which targets for features have been changed (Pressey et al. 1993).
  The system produces a network of areas which meet specified conservation objectives. The user may modify the network produced and "iterate" to an acceptable solution. A more comprehensive description of CODA is provided in Bedward et al. (1992).
• System Requirements: Unknown
• Data Requirements: Information for the CODA system is divided into two main categories:
  a. Data sets which consist of the following four layers:
  1. Selection units, which are land parcels that will be used in the comparison of conservation alternatives. Examples include grid cells, stream catchments, remnant vegetation and tenure.
  2. Conservation features which consist of the components of biodiversity to be included in the analysis. These could be vegetation types, environmental domains, land systems or point data such as rare plant localities.
  3. Land suitability units which divide the region into classes of suitability for conservation. These could include land use, vegetative cover or extent of pollution.
  4. Costs including acquisition costs or opportunity costs of conservation to other land uses. Examples include the number of selected areas, purchase price of areas, timber volume and tonnages of mineral ore.
  Generally this data would be available in a GIS, but the system at present only accepts the data in ASCII format.
  b. Conservation objectives, where the user defines explicit target values for each conservation feature to be sampled by the network of protected areas. Generally these targets would be largely arbitrary such as 20% of each land system or 3 localities for each rare plant species.
• Integration Requirements: Unknown
• Use in CRAs: One of a suite of reserve selection algorithms that requires consideration.

5. IHACRES

• Purpose: The conceptual model used extends the unit hydrograph theory by assuming a linear relationship between effective rainfall, quick flow and other identifiable hydrograph properties. The model consists of a linear rainfall loss module which converts observed rainfall and a linear module which converts the excess rainfall into observed streamflow. This facilitates the assessment of natural variability of runoff, especially for purposes such as water supply evaluation (Ye et al. 1995).
• System Requirements: Unknown
• Data Requirements: Requires long-term daily records of rainfall and streamflow.
• Integration Requirements: Unknown
• Use in CRAs: Input to environmental models where behaviour of water and runoff are required at a catchment or regional scale.

6. Environmental Domain Analysis

• Purpose: "Environmental domain analysis (EDA) uses existing data on environmental attributes (altitude, terrain, climate and geology, for example) together with classification techniques to identify and map the different types of environments that occur in a region; these types are called 'environmental domains'. Knowledge of the distribution of these environmental domains, together with some biological distribution data, may be used to predict the distribution of plants and animals in the region and to assess the design systems of nature reserves. The main advantage of EDA analysis is that it can greatly reduce the need for costly direct surveying of plants and animals, especially for large scale analyses. The technique has the additional advantage of using data that are readily available and relatively constant over time. The principal disadvantage is that it can indicate only the potential for the occurrence of particular species or communities; there is no certainty that a community is present even if the physical and climatic conditions are appropriate" (Commonwealth of Australia 1995; p 46).
• System/Data Description: The analysis technique requires environmental data and a classification package (eg. PATN or Splus). The basic procedure is embedded in the Australian Environmental Geographic Information System.
• Integration Requirements: The analysis approach can be simulated using existing GIS and classification systems.
• Use in CRAs: Deriving environmental diversity values for conservation and heritage assessments.

7. PVA Analysis

• Purpose: PVA analysis is the ".. quantitative evaluation of all factors and their interactions that act on populations and contribute to their risk of short- and long-term decline or extinction" (Lindenmayer et al. 1995). The technique is recommended as a tool for quantifying land use impacts by estimating extinction probabilities. A more comprehensive discussion of PVA methods, their uses and risk assessment of natural populations is found in Burgman et al. (1993).
• System Description: A comprehensive review of three generic PVA programs, ALEX, RAMAS/space and VORTEX, is provided in Lindenmayer et al. (1995). To quickly summarise there findings:
  1. All packages are only available on MS-DOS (PC based) systems.
  2. All are copyright and cost between $0 and $40 with source code generally unavailable.
  3. Two are written in Turbo-Pascal and one in C.
  4. Maximum population sizes range between 1000 (VORTEX) and 2 billion (RAMAS/space). With VORTEX it is noted that complex models with more than 1000 individuals could take several days to run on a 386 machine.
  5. Other limits are between 20 and 160 populations, 1000 to 32000 replications. Limited adjustment (or none) is provided for the random number generators used.
• Integration Requirements: If any of these programs were to be included in the integration project some form of conversion would be necessary. Cost and timing of this conversion is unknown - if possible. DCNR is planning to include a PVA module in their Integrated Forest Planning System (IFPS).
• Use in CRAs: Consideration should be given to the use of these algorithms in conservation planning.

8. Irreplaceability

• Purpose: Irreplaceability of areas can be defined in two ways (Pressey et al. 1994): the potential contribution of any area to a reservation goal and the extent to which the ability to find a representative reserve system is lost if a given area is lost. Irreplaceability is a measure of the conservation importance of an area as well as a basis for negotiating individual protected areas or networks of protected areas. Pressey et al. (1994) indicate that "The full potential of irreplaceability in conservation planning can be realised by applying it interactively, i.e. with a graphical system on a microcomputer that allows the user to make all the decision about protection measures and provides information on the screen to guide each decision".
• System Description: The system links an external program for predicting irreplaceability to a geographic information system (E-RMS) which can then query an external database management system. Other GIS packages that can link to an external database could replace E-RMS, for example ARC/Info. The system is PC-based using MicroSoft Windows.
• Data Description: Digital data describing the areas for consideration.
• Integration Requirements: The irreplaceability program is external to the GIS and database and could be integrated especially under one of the integrating toolkits. Availability and cost of source code to provide this functionality is unknown, as is the effort required.
• Use in CRAs: Consideration of the method should be given in respect of its use in conservation planning.

9. CAARNET

• Purpose: CAARNET (Comprehensive Adequate and Representative system for Reserve NETwork design) is a current development aimed at the design and assessment of reserve networks. The system uses a recently developed theory of information management for biodiversity (Davey et al. 1995) to translate the reserve selection problem into a visualisation problem which can then use the methods of image processing for solution (Stockwell 1996).
• System/Data Description: The system has been developed under a UNIX operating system (LINUX) and utilises the suite of "pmbtools" which is an image processing software set which is available at no cost on many UNIX systems with an X11 environment. The system uses HTML/CGI to produce output suitable for use on a World Wide Web.
• Integration Requirements: Unknown but unlikely to be difficult given the development environment.
• Use in CRAs: Further evaluation for possible use in reserve selection should be considered.

Management Support Systems

1. LUPIS

• Purpose: LUPIS is a PC-based spatial decision support system designed for the task of identifying a preferred land use or management regime from a number of competing management uses (Ive 1992). The aim is to allocate to ".. competing uses, land resources that are both finite and spatially heterogeneous so as to maximally meet the many and varied values of all interest groups. This task is the essence of land use planning" (Ive 1992).
  LUPIS was developed as a spatial decision support system to encourage the use of two related methods of land-use planning:
  1. SIRO-PLAN for the allocation of land under the control of a single agency.
  2. SIRO-MED for the allocation of land with multiple interest groups and agencies (Cocks and Ive 1996).
  LUPIS ranks candidates for land-use in accordance with the extent to which the allocation addresses the issues given by the allocation guidelines and the relative political importance attached to each guideline. The approach is one of adaptive learning rather than a more rigid approach (Ive 1992). A more mathematical treatment of the allocation of a land-use on any mapping unit is provided in Ive (1992).
• System/Data Description: The system has always been targeted at the "... lower end of the computer technology spectrum" (Ive 1992) and currently runs on MS-DOS platforms with a minimum of 640K of memory. The program is coded in QuickBasic version 4.0 and has been converted to provide a Windows interface. The program consists of approximately 500000 lines of code with the major limitations being the allocation of memory. It has been tested with up to 50000 spatial units. It has not been adapted to either a client/server environment or UNIX.
  Linkages between LUPIS and other systems including GIS systems such as SPANS and ARC/Info is "seamless", but relies on the import/export capabilities of LUPIS. That is, linking between LUPIS and most other systems of interest is not dynamic and requires some manual intervention of the user.
• Integration Requirements: It is not clear that LUPIS could be easily integrated into a modular architecture other than by passing data to/from the system in a flat file manner. This could be achieved under some of the integration tools or through UNIX shell scripts. Conversion of the program to run on a UNIX platform appears to be quite a challenge. No cost or time estimates can be made for this conversion, which would almost certainly be necessary to handle the large number of units expected to be involved in each CRA. On the transparency issue, the program has a high level of traceability and user interaction/intervention with the methods well published.
  An exhaustive example of the use of LUPIS is provided in Cocks et al. (1995).
• Use in CRAs: LUPIS has been used in demonstration and allocation studies for developing land use options and in the consideration of landuse planning issues.

2. DIVERSITY

• Purpose: DIVERSITY is built on the idea that "Biodiversity and other criteria, such as forestry suitabilities, can be explored regionally by varying the relative weightings on the criteria and determining how these affect trade-offs and consequent total net benefits" (Faith 1995). The system selects or evaluates sets of areas using surrogate forms of information for biodiversity. The system consists of three programs:
  1. Environmental diversity (ED) "... describes applications in which environmental data, sometimes in combination with biotic data, are used to place the areas in a pattern interpreted as an environmental space" (Faith and Walker 1995). This together with costs can be used to achieve a balance of coverage which represents the pattern at minimum cost.
  2. Phylogenetic diversity (PD) uses the same data as ED, but represents the areas as a hierarchical pattern.
  3. Extinction and expected diversity (XD) is an extension of PD which incorporates the vulnerability or possible extinction into the assessments of expected biodiversity.
  In the forestry case the program can be used to provide an optimal balance or trade-off between regional biodiversity protection and forest opportunities. The area allocation will protect an area if and only if the complementarity-based contribution of the area exceeds the opportunity cost (Faith 1996). For a discussion on the principle of complimentarity see Pressey et al. (1993).
• System/Data Description: DIVERSITY is written in Fortran and consists of approximately 12000 lines of code. It has been designed for use on a PC-based system running DOS. It does not have a point and click interface, is not client/server in nature and is not available in a Windows environment. The major limitation would appear to be the available memory with up to 1000 sites or locations being somewhere near the upper limit of the current system. The program appears to do a complete search of the environmental space for each iteration which could be computationally intensive for large data sets.
  Data input is via ASCII files defining the required inputs. An export facility is available from the package to save output or results. Linking to other systems such as SPANS and ARC/Info is not dynamic, although the write-up for PD seems to indicate a dynamic link to GIS and the image processing system IDRISI, and is facilitated through export from these systems to ASCII files. These files generally have a free-format.
  The authors of DIVERSITY consider that the future development of the package should include the inclusion of adjacency constraints allowing spatial interdependencies. Also future development would make the system more interactive with a higher level of user integration into the concepts to facilitate conflict resolution and mediation.
• Integration Requirements: DIVERSITY may be able to be integrated into a modular architecture by passing data to/from the system in a flat file manner. This could be achieved under some of the integration tools or through UNIX shell scripts. Conversion of the program to run on a UNIX platform would appear to be more constructive although no cost or time estimates can be made for this conversion. This conversion would almost certainly be necessary to handle the large number of units expected to be involved in each CRA as would research into algorithms to reduce the need for a complete search of the environmental space thereby improving processing efficiency.
  On the transparency issue, the program has a high level of traceability and user interaction/intervention with the methods well published.
  An exhaustive set of examples of the use of DIVERSITY is provided in Faith and Walker (1995).
  CSIRO Division of Wildlife and Ecology has submitted a proposal to provide two components for the Integration project:
  1. A completed case study demonstrating and evaluating the application of LUPIS/DIVERSITY to biodiversity assessment and environmental-socioeconomic trade-offs, tailored to the CAR criteria and CRA requirements.
  2. Generic procedures and tools on a UNIX platform for carrying out trade-offs analyses in the context of any of the CRAs/RFAs.
  This process would require three scientific staff at 2 days per week plus technical support for 9 to 12 months. The cost for the case study would be $275000 and the cost for the UNIX conversion is currently unknown. DCNR has estimated that is would take 4-6 weeks to integrate DIVERSITY into its IFPS system.
• Use in CRAs: A package that could be useful in conservation assessment, reserve system planning and determination of resource options.

3. FORUM

• Purpose: FORUM is a regional linear programming model of forest production which incorporates the variables and relationships of an operating forest system. The model can be used to simulate interactions between regional forest resources, forest industries and product markets. The model uses spatially disaggregated data and is being developed by ABARE (Hansard et al. 1996).
  The model estimates a set of control variables which maximises forest return subject to a set of constraints. The methods used in FORUM include discounted cash flows, residual pricing and linear programming. The system is not intended to be a land optimisation model but rather a tool to evaluate the effects of alternative forest processing options (Hansard et al. 1996).
• System Description: FORUM is written using the GAMS system.
• Data Description: Cost and price data are required as is data on forest management areas, including yield and transport cost data.
• Integration Requirements: No automatic link between FORUM and GIS exists although this is planned. Also under development is the link between FORUM and MONASH, as well as a planned link between LUPIS/DIVERSITY to allow consideration of the land-use allocation problem. Discussions with the authors of LUPIS/DIVERSITY suggest that this link would provide the "costs" of protection for a defined set of protected areas, allowing the extension of the trade-offs algorithm in DIVERSITY and the combining of FORUM and LUPIS costings. Other packages such as DCNR's variant of FORPLAN could complement the use of FORUM.
  ABARE is developing/integrating models such as FORUM into more integrated products using the network concepts or approach typified in the visualisation product AVS. This integration will allow sophisticated visual appreciation of the model(s) and their outputs as well as increasing user interaction.
• Use in CRAs: Package to be used in economic assessment and the analysis of industry development options and economic impacts.

4. MONASH

• Purpose: MONASH is a general equilibrium model (Hansard et al. 1996) which is based on the Orani model with extensions. Extensive documentation for MONASH is not available but it does build on the Orani model.
• System Description: Unknown
• Data Description: Unknown
• Integration Requirements: ABARE is currently building an interface between FORUM and MONASH.
• Use in CRAs: Package to be used in socio-economic assessments to investigate economic options.

5. FORPLAN/IFPS

• Purpose: FORPLAN is a linear programming (LP) model originally developed in the late 1970's in the United States by the USDA Forest Service. The original formulation consisted of a matrix generator that converted the input into a mathematically solvable form, a commercial LP software solution system and a report generator.
  FORPLAN itself has been the subject of a number of reviews, criticisms and modifications. These are all well outlined in the literature and summarised in Turner and Church (1995). One extra comment on FORPLAN from an international colleague involved in modelling systems is worth mentioning - ".. these models are not appropriate for longer term dynamics as forest dynamic systems are non-linear in the constraint sets."
  The Integrated Forest Planning System (IFPS) developed by the Department of Conservation and Natural Resources (DCNR) of Victoria integrates FORPLAN with other facilities. The Victorian system integrates four main components:
  1. Geographic Information System - ARC/Info
  2. Growth and Yield module
  3. Optimisation module - FORPLAN
  4. Monitoring module
  Turner and Church (1995) consider that one of the major differences between the USA implementations of FORPLAN and IFPS is the GIS link, which in the Victorian systems provides the basis for planning and reduces the dependency on FORPLAN as a planning tool. As these authors comment, "The use of a GIS allows a spatial representation of FORPLAN output, and most importantly, a greater understanding of the spatial constraints involved."
  A special front-end for IFPS was also written using the menu facilities of SIR, a commercial database.
  A separate review of FORPLAN with respect to wildlife information and planning was also commissioned by DCNR. This review found that "FORPLAN is useful as a tool to focus questions and research, but it has serious limitations when used as a primary planning tool. An array of tools is required for forest planning including FORPLAN, modelling environments of other kinds, iterative research and public participation" (Burgman 1995).
• System Description: The IFPS system currently runs on a SUN UNIX environment incorporating the SIR database system for the interface, GIS using ARC/Info, FORPLAN and the LINDO LP solver. The system used is a 4 processor Sparc 20 system with runs taking up to one hour.
  The ERMapper Image Analysis system is used by DCNR to check the accuracy and precision of analysis areas (AA's) against the existing forest (see Lau et al. 1995).
  DCNR is currently looking at linking the IFPS with additional systems and other modifications including:
  1. The LUPIS/Diversity programs to provide trade-offs in FORPLAN.
  2. A "tail" module for PVA analysis as suggested by Burgman (1995) is under development.
  3. Links to a visualisation system are being considered.
  4. The conversion of SPECTRUM - essentially a new version of FORPLAN - from PC to UNIX. This would remove two constraints including the current 999 analysis area limit (becomes unlimited) and the restriction to 20 periods (increased to 99), allowing finer time steps.
  Currently the system does not handle socioeconomics but it can consider other forest products.
• Data Requirements: Although the data has been suggested as the limiting factor, the more important problem appears to the long and tedious process of manipulating the data to form the Analysis Areas (AA's). Certainly digital data will be required, but this is required for any of the methods. Changing the AA boundaries, or more importantly their number, requires considerable validation time before running the model. The development of analysis areas is a critical component in the final model output quality (Lau et al. 1995). Care needs to be exercised in choosing the themes defining the AAs as the number of themes is limited to 8.
• Integration Requirements: The adaptability of the system is probably quite good, but a base of skilled people in FORPLAN or IFPS does not exist outside of the four staff in DCNR. The IFPS system is a very in-house development with many specialised components and linkages which would need to be more fully documented for transportation to another location. FORPLAN is only one component of the system and only one or two individuals in DCNR have the expertise to run this efficiently (Turner and Church 1995).
  Conversion or transporting the system to another computer installation would require the expertise/advice of the DCNR staff, the conversion of the SIR interface (or purchase of SIR), purchase and installation of the LINDO LP solver and the possibility of conversion of the FORPLAN code if moved to a non SOLARIS system. The overall conversion cost could be quite high, but the technology transfer and knowledge gains could be quite substantial.
  An indicative costing for DCNR staff to undertake the initial case study using their existing in-house systems was given as:
  1. To build a new model including the preparation of the input from already collected data: 2 to 4 weeks.
  2. To debug and validate the model: 4 to 6 weeks.
  3. Run model and build reports: 2 weeks.
  A total of 8 to 12 weeks, full-time for all four staff. The integration of other systems such as DIVERSITY would take an extra 4 to 6 weeks. Use of other inexperienced staff would/could extend the time period required. This timing assumes that the staff were available - they all currently have full work schedules - and assumes no cost is incurred for computing resources - an unlikely event given the likely significant impact the running of the model would have on the DCNR facility.
  Using a conservative figure of $2000 per staff week, the cost of developing the pilot study using the IFPS system (excluding Diversity) would be between $64000 and $98000 for staffing costs plus any additional costs. Further discussion with Bruce Kilgour of DCNR (21/2/1996) have refined the time estimates to be 3-4 months for the equivalent of 3 full-time staff (2 full-time, 2 half-time). This modifies the staff costings to between $72000 and $96000.
• Use in CRAs: A system that has merit for considering forest management options, sustainability and landuse planning.

6. TRACS

• Purpose: TRACS aims to identify land and marine regions with current or potential conflicting multiple land-uses and determines using the available thematic spatial data and the TRACS system methodology, possible trade offs in multiple land use.
  The TRACS system allows the user to communicate graphically different interpretations of data classification, data survey methods and thematic importance with respect to other user's interpretations. These interpretations may be scientifically, socially or economically defensible interpretations of the data or they may be relatively subjective or emotive contentious perspectives reflecting a wide range of viewpoints associated with land use conflict and debate.
  The TRACS methodology uses the premise that, in theory, any scenario description can be generated by a series of Boolean operations (OR/AND) on spatial and thematic attributes. That is, pre-determined land use scenarios can be explored through the addition (AND) of relevant thematic spatial databases and the reclassification (OR) of thematic data and spatial attributes, given the availability of relevant data. This idea is based on the ASSESS methodology used for site selection for low-level radioactive waste disposal (Bowyer and Veitch 1995).
• System Description: TRACS uses a graphical interface based on the ARC/Info software. A menu-based set of functions allows the user to select natural resource and land use themes and to then vary their interpreted values as required. Results of analyses are presented in maps, graphics and tables.
  The system requires a UNIX system capable of running the ARC/Info package. Current development is done on SUN workstation equipment.
• Data Description: TRACS requires a range of GIS based datasets or themes. Currently TRACS uses national datasets. Import/Export as well as other graphical facilities are identical to those provided by ARC/Info.
• Integration Requirements: TRACS has been developed as a standalone system working at the national scale. It can be adapted easily to regional scales. The integration of the methodology and datasets could be achieved through a integrating toolset with strong links to ARC/Info. Products such as Calyx and Facet may provide this functionality.
• Use in CRAs: A toolbox and methodology for considering resource use options.

7. ArcForest

• Purpose: ArcForest is an integrated forest management decision support system developed under a joint venture between The Ontario Ministry of Natural Resources (MNR) and ESRI-Canada. The system which uses ARC/Info and the ORACLE relational database management system, is designed to help the forestry industry handle current and future forest management as the industry moves towards more ecosystem-based, sustainable resource management practices.
  The system includes a number of tools to assist users in forest management planning. These tools include the ability to consider a variety of forest allocation options and resource use conflicts.
• System Description: The system uses both ARC/Info and ORACLE running on UNIX servers. For the MNR system, ESRI-Canada ported the product to the DEC Alpha workstation platform.
• Data Description: Unknown
• Integration Requirements: Currently BRS has a demonstration copy of ArcForest for evaluation. This evaluation has only just commenced and details are not currently available. The system has generated interest from the Malaysian Government and has been introduced into the curriculum of at least one Ontario College School of Natural Resources.
  Future development directions include the linkage of ArcForest and ARC/View, research on new data modelling approaches and enhanced methods for inclusion/integration of external models. Approximate cost for the system is unknown.
• Use in CRAs: A system that could be useful in the consideration of forest management and ESM issues.

Integration Framework Systems

1. GAMS

• Purpose: The General Algebraic Modelling System (GAMS) is designed for modeling linear, nonlinear and mixed integer optimisation problems. GAMS is able to formulate models in many different types of problem classes. Currently GAMS supports the following model types:
  1. Linear programming (LP)
  2. Non-linear programming (NLP)
  3. Mixed-integer programming
  4. Mixed-integer nonlinear programming
  5. Mixed complimentarity problems
• System Description: The GAMS language is similar to commonly used programming languages and is available on a range of PC's, workstations, mainframes and supercomputers. The models are fully portable from one platform to another and will allow for dynamic models involving time sequences, lags and leads. The GAMS system uses commercially available solution software for each of the model types. Further information including worked examples can be found on the GAMS homepage (www.gams.com).
• Integration Requirements: It is not clear whether this system would be useful for the CRA process, but it is used by many groups including ABARE to provide a generic LP/NLP solution system. Integration could be achieved by using UNIX shell scripts or by using one of the toolkits.
• Use in CRAs: A toolbox for developing and coding models to evaluate optimisation solutions.

2. CALYX

• Purpose: The Calyx family of products from ESSA Software in Canada are PC-based software tools that allow the prediction and examination of environmental impacts and their consequences. Of interest for the Integration project are two products:
  1. Calyx GIS, which is a decision support framework which uses expert systems and ArcView to provide analysis of project and environmental information. The system applies a consistent methodology based on a rulebase to determine the impacts and produce optimal decision plans.
  2. Calyx EA enables users to identify and quantify an entire spectrum of potential environmental impacts and produce reports on them. This product addresses both physical and biological effects and looks beyond primary impacts by determining secondary impacts to commercial, visual and social environments.
• System Description: Calyx currently runs on PC-based systems and uses ARC/View for display and spatial data handling. It provides direct data exchange capabilities to other Windows-based software applications.
• Data Description: Calyx can directly use any spatial data generated by ARC/Info.
• Integration Requirements: Calyx is at the PC-end of the technologies, rather than being supported on a UNIX system. However, much of the spatial data that any Calyx system would use could be generated from either ARC/Info or ARC/View running on a UNIX server. Calyx uses a rule-based approach or knowledge base to determine the possible effect or impacts and any methods of mitigation available for single or multiple spatial layers. This rulebase must be developed by "experts" and input to the Calyx framework.
  Linkage to other Windows-based software applications is via DDE (direct data exchange). Many of the other technologies considered here do not have this capability. Calyx has been used within the Commonwealth to develop prototype environmental management systems for the Defence forces (NRIC/BRS).
  Calyx was a component of the Prototype Environmental Management Information System developed for the Department of Defence by BRS/NRIC and ESSA. Calyx was used to determine the environmental impacts of a pre-defined set of activities within a set of environmental variables and to define the remediation effort required. The activities were the movement of vehicles along tracks, roads or across country and the environmental variables were a set of constraint maps developed in ARC/Info. Rule bases relating vehicle usage to environmental impacts and remediations were incorporated into the Calyx knowledge base, the output a map indicating the impacts for vehicles. The constraint maps were exported from ARC/Info as ARC/View shape files. Calyx used the ARC/View files to carry out the analysis.
• Use in CRAs: A toolbox that enables artificial intelligence to be built into the integration system. It enables expert knowledge or rules to be modelled.

3. FACET

• Purpose: The Facet Spatial Spreadsheet system is described as a "complete spatial decision tool, not a GIS" by its manufacturers. It is a modelling and management decision making tool which includes a range of interactive analysis, data fusion and visualisation tools. Models can combine economic, geographic, environmental and natural resource concerns.
  Facet has been used quite extensively in a number of areas overseas. The product was originally developed as an object oriented environment for research in digital mapping, image processing and machine vision. The initial research was jointly funded by the Canadian Government and MacDonald Dettweiler. However, now the system has been expanded to provide a multi-disciplinary environment for problem solving. The product has been used in strategic planning, exploration, habitat modelling, coastal zone management, stock market analysis and atmospheric research. Descriptions of the use of Facet in land use conflict mitigation can be found in Melzer et al. (1993) and Ayers et al. (1993).
• System Description: Facet currently only runs on the SUN Solaris environment and requires a machine with 32 Mbytes of memory and approximately 75 Mbytes of storage for the Facet system and demonstrations. The more memory available the better the product will perform, especially when integrating image processing tasks.
  The system utilises quite sophisticated computing technologies including:
  1. Object Oriented technology including object classes, instances and inheritance. This leads to high levels of reuse of "code" (see Wilkie, 1993 for a discussion of object oriented software engineering).
  Parallel processing for multi-processor systems, Facet will use separate processors for each cell calculation when possible.
  3. Distributed processing, Facet allows for linking of models across the network to form one distributed model.
  Facet has a large number of built-in functions which cover the following areas:
  1. Maps and Map Analysis
  2. Image Analysis
  3. Trigonometry
  4. Finance
  5. Mathematical functions, exponents and logs
  6. Database manipulations
  7. Basic Statistics
  8. Map mathematics
  9. Numeric and constant conversions
  10. Comparisons and logical operators
  11. Data Import
  12. UNIX connectivity
  The area of connectivity to other programs is well covered by Facet from the simple ability to link to UNIX shell scripts, add functions and link to function libraries via the spreadsheet, through to dynamic linking of external algorithms to a Facet server.
  Facet provides two facilities for linking to other systems or algorithms. These are the:
  1. Dynamic Linker which allows systems written in C++, C, Fortran or Pascal to be linked into the Facet function library. After linking the user has complete user-level access to these new library functions.
  2. Pipe Option allows a user to pipe data to an external program, run the program and return the results as though they were from a Facet function.
• Data Description: Facet can import many of the standard graphics formats as well as a number of more specialised data formats. Direct import of data from ARC/Info is supported as well as from standard ASCII files and Dbase III and IV. Image formats such as DXF, TIFF, GIF, JPEG and PICT are directly supported as well as "public domain" formats such as PNM, PPM, PGM and PBM.
  The system supports a number of visualisation facilities including layout facilities for cartographic map production, interactive statistical, business and technical graphs, animation facilities and a user interface builder.
• Integration Requirements: Facet appears to be a very adaptable toolkit which allows the integration of many different aspects of environment modelling and visualisation. The ability to link to other specialised systems increases the flexibility of the product. The cost of the system is $23500 (Can.) for a development license and $2000 (Can.) for a runtime license. However, the major difficulty with Facet is the lack of experience/expertise with the product in the Commonwealth and States, although the product has been used extensively by the Canadian Forest Service and the reuse and adaptation of their framework is a distinct possibility which should be examined further.
  Facet developed a "Forest Ecosystem Network" for the Canadian Forestry Service which included the analysis of compliance rules about wildlife corridors, rare sites, preserving forest diversity and old growth connectedness as well as other ecological requirements. This framework as well as a Forestry Landscape Design System developed for the British Columbia Ministry of Forests is available for reuse in the CRAs. The typical cost to implement these types of systems is $10000 to $20000 (Can.).
• Use in CRAs: A system that could be developed to provide the modular architecture, flexibility and compatibility with other tools required of the integration system.

4. Whatif?

• Purpose: Whatif? is an object-oriented scenario and modelling package providing a structured set of tools for groups of experts to interact, express their ideas and apply concepts to achieve resolutions to debate(s) of economically and ecologically sustainable resource issues.
  The system uses the "design" approach to modelling (Gault et al. 1987) which means that a system is modelled by firstly identifying what components, as objects, exist in a system and then drawing linkages of these to each other in a functional sense as a framework. A model framework incorporates input and output objects as well as other objects which may modify these inputs and outputs.
  In this approach the user is tightly integrated with the models and simulation framework and provides novelty and change to the system by the specification of control variables over the simulation timeframe. The approach aims to explore, rather than predict the future, and is not oriented towards global optimisation or equilibrium conditions. This lack of global optimisation or equilibrium constraints can lead to results which are inconsistent or socially unacceptable - tensions. The user can then resolve these tensions by either exploring alternative scenarios or the user may exercise choice and accept scenarios where tensions still exist (Veitch et al. 1993). Further informatiom can be found at http://www.robbert.com
• System Description: Whatif? consists of three main components: TOOL (Tool On Object Language), an interactive coding language for manipulating data objects; SAMM (Scenario And Model Manager), provides the interface linking sets of models; and Documenter, a text and graphics system for preparing structural and relationship diagrams and for the generation of framework documentation.
  Whatif? has been designed as a client/server architecture. The Whatif? software running on a UNIX server provides services to multiple Macintosh clients using other software components of Whatif?. For example, the design diagrams for a framework are created on the Macintosh by a Macintosh application but must be stored on a UNIX server for use. Once stored on the UNIX server the SAMM application on the Macintosh can be used to load and run a framework. The client/server environment allows multiple Macintosh clients to access simulation frameworks held on a common UNIX server.
  The UNIX server provides the computational engine for running the framework, with the Macintosh client providing the interaction for the user and the display of output from the framework. Currently, the server is supported by UNIX systems from IBM, SUN, HP and Apple. The client software only runs on a Macintosh.
  Whatif? provides an extensive toolkit of built-in functions covering:
  1. Mathematical tools from simple functions such as addition of objects to more complex functions such as solving systems of linear differential equations.
  2. Accounting tools which evolve stocks of people, things or biological systems such as forests.
  3. Statistical tools including linear interpolation or trend analysis.
  4. Financial tools to perform a range of functions such as present value and depreciation calculations.
  5. Utility tools to convert data between multiple formats.
  6. Shaping tools for generating or changing the shape of objects.
  7. Display tools for displaying objects in graphical, tabular or cartographic form.
  Using the developer's toolkit linkages to other algorithms or systems can be provided. This toolkit allows any algorithm coded in C or Fortran to send and receive objects from Whatif?. This method has been used to provide a direct link to ARC/Info from Whatif? for example. It has also been used to include specialised simulation models (eg. hydrological models) as functions within the Whatif? software. The simplest method of providing a link to specialised models requires modification to the source code for the external model. Alternatively, the new Whatif? function may simply convert to/from the data structures of Whatif? and the external model and "call" the model as an external program.
• Data Description: Whatif? objects may be one of a number of types:
  - arrays - integer or real
  - set
  - sequence
  - category
  - constant
  - polygon
  - line
  - point
  These objects may be filled with data using a table application, MicroSoft Excel, direct assignment to a constant or from another object and from a graphical application which allows the user to "digitise" the response or object data directly on the Macintosh. Alternatively, data may be read from white space separated ASCII files from within the Whatif? code.
  Once objects are filled with data they may be exported in Whatif? format for later reloading or via MS Excel for reuse elsewhere. Output objects can be displayed graphically, including a cartographic display using polygons or can be tabulated. Links to other visualisation systems can be done using the developer's toolkit or by exporting the output object(s).
  Once a simulation framework is developed, view scripts can be written which restrict the user to being able to view and/or change only those objects allowed by the framework developers.
• Integration Requirements: Whatif? is a very adaptable toolkit which allows the integration of many different aspects of environmental and socio-economic modelling. The ability to link to other specialised systems increases the flexibility of the product. Whatif? does lack the strong mapping and image processing capabilities of Facet and Calyx, but the direction of the developers is to provide maximum modelling capability and to provide easy links to GIS and visualisation systems, rather than developing their own. The other disadvantage is the existence of only Macintosh client software. This environment can be emulated on some UNIX systems (SUN and HP) quite successfully however. The client software is currently being rewritten to support MS WINDOWS 95.
  The cost of the system is $10000 (Can.) for a development license for the first client and server. Second and subsequent license costs vary depending on the need for the development functionality. With this functionality subsequent licenses cost $7500 (Can.) and without $2500 (Can.). Unlike Facet however, there is some experience/expertise with the product in the Commonwealth on long-term ecological and environmental modelling (see Fordham and Malafant, 1995 and Malafant and Fordham, 1995).
• Use in CRAs; A sophisticated scenario modelling toolkit which could provide the facilities for scenario modelling of resource use and sustainability of decisions.

5. Other Systems

Other integration tools exist which may provide some integration facilities, especially on Mac and PC platforms. Systems which can be considered in this category include:
1. STELLA: This is a MacIntosh based scenario modelling tool (now called Ithink) which can be used to code sophisticated models and interactions. Arguably the best known application of STELLA is the World3/91 model of Meadows et al. (1992).
2. Extend: A PC based system which allows general purpose models to be coded and run. ABARE has considerable expertise with the system.
3. @RISK: A PC based risk assessment system used by many groups including BRS.
4. ModelMaker: A PC based system that uses conceptual or model diagrams to construct models. The system has a number of modelling tools, including optimisation, standard mathematical functions, statistics, sensitivity analysis, interpolation and differential equation tools. Currently being evaluated by NRIC/BRS.
5. Queensberry: "Queensberry is a computer software system which offers facilities that are useful for setting up or analysing rational debates, especially debates about planning" (Macpherson and Grant, 1992). A PC based system using the Smalltalk language.

Other Tools

1. GIS Tools

The dominant GIS systems being used by agencies are the ESRI products ARC/Info and ARC/View. Some mention of SPANS, MapInfo, E-RMS and MIPS is also made but in the larger agencies and the Commonwealth the use of ARC/Info has dominated. It is not relevant to describe these products in any detail other than to say that considerable expertise in their use and a large body of additional software that interacts with the products exists.
On the issue of GIS, almost all agencies are using the ESRI ARC/Info and/or ARC/View products and no other reasonable choice is available given the extensive linkages this software suite has with other systems. The adoption of other GIS systems in CRAs would require considerable modification to possible solutions and require an extensive investment in software, time, training and conversion of spatial data.

2. Image Processing Tools

No one package appears to dominate this area. The agencies involved in the CRAs use a small number of products including IDRISI (PC based), ERMapper and Imagine. Of these ERMapper is the most used in larger agencies, both state (eg. Victoria) and Commonwealth (eg. BRS, AGSO). The Imagine product integrates more seamlessly with the ARC products by providing hot links between the two systems. The cost of this product is reasonably high (approx. $27000) per license.
A number of agencies are using the Australian developed product ERMapper (DCNR, BRS, AGSO), while others use the PC-based IDRISI package. ERMapper does run well on UNIX systems and there is considerable experience with the product in both States and Commonwealth agencies, however, it does not link as seamlessly with ARC/Info as does ERDAS/Imagine. Given that a single Imagine license costs ~$27000 it does not seem sensible to adopt it in preference to ERMapper.

3. Visualisation Tools

The only agencies that appear to be using sophisticated UNIX-based visualisation tools are ABARE and BRS. Both groups use the AVS product which is considered to be the market leader in this software area. AVS is currently developing a link with ARC/Info which is being beta tested by BRS/NRIC and ABARE. However, there is an extensive body of public domain software which can be added to AVS available from the Universities and other groups via the World Wide Web.
The network editor in AVS can be used to develop distributed analysis and visualisation networks and has some potential for building or integrating applications. However, AVS does not contain the broad range of mathematical, analysis and control tools that many of the other integration toolkits contains. It lacks intelligence (control functionality) and extensibility except for a restricted set of AVS objects and types. The cost of individual licenses for AVS are again quite expensive, averaging somewhere around $20000 for each user. The expertise and experience in using the product does exist in the Universities and in ABARE and BRS. The software/hardware configurations for AVS are far from inexpensive so change would be costly. The system has broad market acceptance, a large suite of public domain utilities and a link to ARC/Info (currently in beta test with BRS/NRIC and ABARE).

4. Mathematical and Statistical Tools

Two mathematical/statistical tools currently being used by agencies allow for integration into complex frameworks. These are: Splus and Mathematica. Both tools provide a broad range of mathematical and statistical functionality, coupled with Exploratory Data Analysis, Experimental Design, Time Series Analysis and Graphics displays. Both products allow the addition of external algorithms or code.
Splus has a code development environment which allows it to be called and commands executed from within a program coded in C. This functionality has be used to provide direct linkage between ARC/Info and Splus with data and results being passed from ARC/Info and returned from Splus quite seamlessly - a distinct advantage in providing and integrated framework.
Mathematica is a fine tool especially for symbolic manipulation, but the seamless link that Splus has with ARC/Info places it at the top of the list. Both products are available for UNIX based systems. Both systems also run on PC's. Their costs are: Splus ~$1200, Splus for ARC/Info~$2000 and Mathematica ~$1200.
The SAS system is also used by some agencies, especially ABARE, to handle most types of statistical analyses as well as operations research, links to database and graphs.

5. Database

Most systems considered in this review have no affinity or preference for a particular database system. The IFPS system from DCNR does use the SIR database system for its user interface - which could be redeveloped. SIR is not a mainstream database when compared to ORACLE, INGRES, INFORMIX or SYBASE, and one of these four should be adopted, probably ORACLE or INGRES.

References

Cited References

Ayers R., C. Murray, M. Prisco, S. Akenhead and N. Melzer. 1993. Live Impact Analysis in Land Use Planning Workshops. Facet Decision Systems Inc.

Bedward M., R.L. Pressey and D.A. Keith. 1992. A new approach for selecting fully representative reserve networks: addressing efficiency, reserve design and land suitability with an iterative analysis. Biological Conservation. 62, pp. 115-125.

Bowyer J.K. and S.M. Veitch. 1995. ASSESS: A system for selecting suitable sites for a landuse. Proceedings of OZRI 95, Hobart, Australia.

Burgman M.A., S. Ferson and H.R. Akcakaya. 1993. Risk Assessment in Conservation Biology. Chapman and Hall, London.

Burgman M.A. 1995. Wildlife information and forest planning: a review with reference to FORPLAN. Final Report to Department of Conservation and Natural Resources, Victoria.

Busby J.R. 1986. Bioclimatic Prediction System (BIOCLIM). User's Manual Version 2.0. Australian Biological Resources Study Leaflet.

Cocks K.D., J.R. Ive and J.L. Clark (Eds.). 1995. Forest Issues. Processes and Tools for Inventory, Evaluation, Mediation and Allocation. Report on a case-study of the Batemans Bay area, New South Wales, Australia. CSIRO, Australia.

Cocks K.D. and J.R. Ive. 1996. Mediation Support for Forest Land Allocation: The SIRO-MED System. Environmental Management , 20(1), pp. 41-52.

Commonwealth of Australia. 1995. National Forest Conservation Reserves Commonwealth Proposed Criteria: a position paper. AGPS, Canberra.

Davey S.M., D.R. Stockwell and D.G. Peters. 1995. Managing biological diversity with intelligent systems. AI Applications, 9:2, pp. 62-89.

Dongarra J.J., J.R. Bunch, C.B. Moler and G.W. Stewart. 1979. LINPACK Users' Guide, SIAM, Philadelphia.

Faith D.P. 1995. Biodiversity and Regional Sustainability Analysis. CSIRO, Australia.

Faith D.P. and P. Walker. 1995. DIVERSITY: a software package for sampling phylogenetic and environmental diversity. User's Guide. Chapters 1-4, Vol. 3, BIORAP tools for assessing priority areas. CSIRO, Australia.

Faith D.P. 1996. Biodiversity assessment and opportunity costs. Final consultancy report to OECD. Paper to be presented at OECD International Conference on Biodiversity Incentive Measures, Cairns, 25-28 March.

Fordham D.P. and K.W.J. Malafant. 1995. Biophysical, Agricultural Production and Socioeconomic futures in Irrigation Regions: A twenty year profile. In: Proceedings of Agricultural and Biological Engineering - New Horizons, New Challenges. Newcastle Upon Tyne, 20-23 September.

Gault F.D., K.E. Hamilton, R.B. Hoffman and B.C. McInnis. 1987. The design approach to socioeconomic modelling. Futures, February, pp. 3-25.

Hansard A., T. Dann, M. Stephens and J. Clark. 1996. An Economic Model for Comprehensive Regional Forest Assessments: A case study - some issues and considerations. 40th Annual Conference of the Agricultural and Resource Economics Society, Melbourne, 13-15 February.

Hutchinson M.F. 1988. Calculation of hydrologically sound digital elevation models. Proc. Third International Symposium on Spatial Data Handling, August 17-19, Sydney. International Geological Union, Ohio.

Hutchinson M.F. 1989. A new method for gridding elevation and stream line data with automatic removal of pits. Journal of Hydrology 106: 211-232.

Hutchinson M.F. 1991. The application of thin plate smoothing splines to continent-wide data assimilation. In: J.D. Jasper (ed), BMRC Research Report No. 27, Data Assimilation Systems, Bureau of Meteorology, Melbourne, 104-113.

Hutchinson M.F. and T.I. Dowling. 1991. A continental hydrological assessment of a new grid-based digital elevation model. Hydrological Processes 5: 45-58.

Hutchinson M.F. and P.E. Gessler. 1994. Splines - more than just a smooth interpolator. Geoderma 62: 45-67.

Hutchinson M.F. 1995. Interpolating mean rainfall using thin plate smoothing splines. International Journal of GIS 9: 305-403.

Ive J.R. 1992. LUPIS: Computer Assistance for Land Use Allocation. Resource Technology 92 Taipei: Information Technology for Environmental Management, November 16-18, Taipei, Taiwan, R.O.C.

Lau J.A., W.G. Vandenberg, R.U. Willig, O. Chikumbo and A. Siora. 1995. Visual and spatial approach to large scale forest modelling in multiple-use forest planning with Australian (Victorian) Experience. Large-Scale Forests Scenario Models: Experience and Requirements. Joensuu, Finland, 15-22 June. Proceedings of the European Forest Institute.

Lindenmayer D.B., M.A. Burgman, H.R. Akcakaya, R.C. Lavy and H.P. Possingham. 1995. A review of the generic computer programs ALEX, RAMAS/space and VORTEX for modelling the viability of wildlife metapopulations. Ecological Modelling, 82, pp. 161-174.

Macpherson D. and I. Grant. (1992). Queensberry: the debate handling system. Users' Manual. CSIRO, Division of Water Resources.

Malafant K.W.J. and D.P. Fordham. 1995. Decision support systems and visualisation tools for modelling biophysical, production and socioeconomic futures in irrigation regions. Proceedings International Congress on Modelling and Simulation, MODSIM 95, November 27-30, University of Newcastle. Modelling and Simulation Society of Australia, 3, pp. 236-244.

McMahon J.P., M.F. Hutchinson, H.A. Nix and K.D. Ord. 1995. ANUCLIM Users's Guide, CRES, ANU, Australia.

Meadows D.H., D.L. Meadows and J. Randers. 1992. Beyond the Limits. Earthscan Publications Limited, London.

Melzer N.C., D. Hawkins and S.A. Akenhead. 1993. Building Concensus: New Approaches for Land Use Conflict Mitigation. Facet Decision Systems Inc.

Nix H.A., J. McMahon and D. Mackenzie. 1977. Potential areas of production and the future of Pigeon Pea and other grain legumes in Australia. In: The Potential for Pigeon Pea in Australia. Proceedings of Pigeon Pea (Cajanus cajan (L.) Millsp.) Field Day. Wallis. E.S. and Whiteman. P.C. (eds.). Dept. Agriculture, Uni. Qld. Australia.

Nix H.A. 1986. A biogeographic analysis of Australian Elapid Snakes. In: Atlas of Elapid Snakes of Australia. R. Longmore (ed.), pp 4-15. Australian Flora and Fauna Series No. 7. Australian Government Publishing Service, Canberra.

Nix H.A., J.A. Stein and J.L. Stein. 1992. Developing an environmental geographic information system for Tasmania: An application to assessing the potential for hardwood plantation forestry. Consultant's Report to Land Resources Division and Bureau of Rural Resources, Department of Primary Industries and Energy.

Pressey R.L., C.J. Humphries, C.R. Margules, R.I. Vane-Wright and P.H. Williams. 1993. Beyond opportunism: Key principles for systematic reserve selection. Trends in Ecological Evolution, 8, pp. 124-128.

Pressey R.L., I.R. Johnson and P.D. Wilson. 1994. Shades of irreplaceability: towards a measure of the contribution of sites to a reservation system. Biodiversity and Conservation, 3, pp. 242-262.

Pressey R.L., S. Ferrier, C.D. Hutchinson, D.P. Silvertsen and G. Manion. 1995. Planning for negotiation: using an interactive geographic information system to explore alternative protected area networks. Nature Conservation 4: the Role of Networks. D.A. Saunders, J.L. Craig and E.M. Mattiske (Eds.). Surrey Beatty and Sons, Sydney.

Qureshi M.E., P.F. Greenfield, F. Kingham and A.A. Krol. 199?. Use of Multi-criteria analysis in decision support for ecologically sustainable development. Proceedings of the Inaugural Ecological Economics Conference, Coffs Harbour, 1995, pp. 181-186.

Stockwell D.R.. 1996. Iterative reserve system development using image processing and the WWWeb. In press.

Turner B.J. and R. Church. 1995. Review of the Use of the FORPLAN (FORest PLANning) Model. Final Report to Department of Conservation and Natural Resources, Victoria.

Veitch S.M., D.P. Fordham and K.W.J. Malafant. 1993. WHATIF? Scenarios to explore Land Management Options. Applications of Advanced Information Technologies: Effective Management of Resource, Spokane, Washington, pp. 351-360.

Wilkie G. 1993. Object-oriented Software Engineering. The Professional Developer's Guide. The Institute of Software Engineering. Addison-Wesley, Workingham, England.

Ye W., D.P. Hansen, A.J. Jakeman, P. Sharma and R. Cooke. 1995. Assessing the Natural Variability of Runoff: Clarence Basin Catchments, N.S.W., Australia. Proceedings International Congress on Modelling and Simulation, MODSIM 95, November 27-30, University of Newcastle. Modelling and Simulation Society of Australia, 1, pp. 188-193.

Other References

Anderssen R.S. and J.R. Ive. 1982. Exploiting structure in linear-programming formulations for land-use planning. Environment and Planning B: Planning and Design, Vol. 9, pp. 331-339.

Anderssen R.S., K.D. Cocks and J.R. Ive. 1983. Implications of the generalised upper bounding structure in land-use allocation. Environment and Planning B: Planning and Design, Vol. 10, pp. 207-217.

Anon. 1994. Centre for Resource and Environmental Studies, 1994 Annual Report. CRES, ANU.

Anon. 1994. Forest Research at the Centre for Resource and Environmental Studies, December 1994. CRES, ANU.

Biggs P.H. and R.D. Spencer. 1990. New approaches to extensive forest inventory in Western Australia using large-scale aerial photography. Australian Forestry, Vol. 53, No. 3, pp. 182-193.

Bren L.J. 1995. Aspects of the geometry of riparian buffer strips and its significance to forestry operations. Forest Ecology and Management, 75, pp. 1-10.

Bren L. and C. Greenwood (Eds.). 1995. Applications of New Technologies in Forestry. Proceedings of the Institute of Foresters of Australia 16th Biennial Conference, Ballarat, Victoria, 18-21 April.

Cocks K.D., J.R. Ive, J.R. Davis and I.A. Baird. 1983. SIRO-PLAN and LUPLAN: an Australian approach to land-use planning. 1. The SIRO-PLAN land-use planning method. Environment and Planning B: Planning and Design, Vol. 10, pp. 331-345.

Cocks K.D. and J.R. Ive. 1988. Evaluating a Computer Package for Planning Public Lands in New South Wales. Journal of Environmental Management, 26, pp. 249-260.

Cocks K.D. and I.A. Baird. 1991. The Role of Geographic Information Systems in the Collection, Extrapolation and Use of Survey Data. In: Nature Conservation: Cost Effective Biological Surveys and Data Analysis. C.R. Margules and M.P. Austin (Eds.), CSIRO Australia.

Cocks K.D., J.R. Ive and M.P. Austin. 1991. SIRO-MED and LUPIS: Some Background Notes. CSIRO, Division of Wildlife and Ecology.

Common M. 199?. What is Ecological Economics?. Ecological Economics X, pp. 1-16.

Coveney P. and R. Highfield. 1995. Frontiers of Complexity: The search for order in a chaotic world. Ballantine Books, New York.

Davey S.M., R. Prinsley and D.H. White. 1991. Forestry and Agroforestry. Agricultural Systems and Information Technology Newsletter. Vol. 3, No. 3, BRS.

Davey S.M., J.R.L. Hoare, S.G. Wayth and J.C. Sutton. 1995. Forest and Wood Products Compendium of Australian Research in Progress, No. 2. Forest and Wood Products Research and Development Corporation and Bureau of Resource Sciences, Canberra.

Ive J.R. and K.D. Cocks. 1983. SIRO-PLAN and LUPLAN: an Australian approach to land-use planning. 1. The LUPLAN land-use planning package. Environment and Planning B: Planning and Design, Vol. 10, pp. 347-355.

Ive J.R. and K.D. Cocks. 1987. The Value of adding Searching and Profiling Capabilities to a Land Use Planning Package. Soil Survey and Land Evaluation, 7, pp. 87-94.

Ive J.R. and K.D. Cocks. 1989. Incorporating multiparty preferences into land-use planning. Environment and Planning B: Planning and Design, Vol. 16, pp. 99-109.

Ive J.R. 1996. Spatially Quantified Guidelines for Assessing Park Values: A Decision Support Approach. Selection and Delimitation of Parks and Protected Areas, International Seminar and Workshop, 23-25 January.

Lunney D., B. Pressey, M. Archer, S. Hand, H. Godthelp and A. Curtin. 199?. Integrating Ecology and Economics: a conflict of Space and Time. Ecological Economics, X, pp. 17-26.

Maser C. 1990. The Redesigned Forest. Stoddart Publishing, Toronto, Canada.

Pressey R.L., I.R. Johnson and P.D. Wilson. 1994. Shades of irreplaceability: towards a measure of the contribution of sites to a reservation goal. Biodiversity and Conservation 3, pp. 242-262.

Rogers M.F. and T. DeLacy. 199?. What's New about BCA and Risk Analysis: Its Application to The Precautionary Principle. Ecological Economics X, pp. 513-520.

Spencer R.D. and L.O. Jellinek. 1995. Public concerns about pine plantations in Victoria. Australian Forestry, Vol. 58, No. 3, pp. 99-106.

Spencer R.D. 1996. Multi-criteria analysis as a decision tool. (in press).

Toohey B. 1994. Tumbling Dice: The story of modern economic policy. William Heinemann Australia as part of Reed Books, Australia.

Appendices

Appendix 1. CRA Integration Project Review Proforma

CRA Integration Project Review

Technology Questions:

• Purpose of "technology", including the conceptual framework, methods and functionality. References to published material describing the technology, methods and uses.
• System description including:
  - Platform availability, especially UNIX and coding language and some indication of the "size" of the system.
  - System requirements in terms of hardware and/or software.
  - Compatibility with other systems and linkages that exist or are planned with other systems.
  - "Missing" pieces of the system, or a "wish-list" of things that it should/could do.
• Data description of the system:
  - Input data requirements - how much, how often, what sort?
  - Facilities for importing/exporting data to and from other systems.
  - Output facilities and requirements - visualisation needs, GIS etc.
• Integration requirements:
  - Adaptability/suitability for incorporation into a modular architecture.
  - How reliable and/or transparent is the technology/method/system.
  - Cost and time required to implement the system

Top of page