MSc  Abstracts 1999

Martin C Ellis

New developments in dynamic spatial decision support systems: An evaluation of current proposals

The literature on Dynamic Spatial Modelling often contains statements like "…GIS is not suitable for dynamic spatial decision support". The questions which then arise, are: "Why not?" and if it is true, "What should be done to improve the situation?". This report examines these questions and describes the development of an evaluation framework for the comparison of two new dynamic spatial decision support systems (SDSS) on trial in the Netherlands.

The study begins with a review of the requirements and the problem areas of planners and decision–makers. Also, a landuse-planning case is introduced as a theme to be used throughout the study. The review and case study are used to develop an initial set of criteria for the evaluation of decision-making support systems. These criteria are extended during the rest of the study.

The next chapter discusses where and how information technology can support the planning processes. This chapter pays special attention to the specific aspect of the dynamic spatial modelling engine needed for a dynamic SDSS. Therefore, this section includes a thorough review of current dynamic spatial modelling methods. These include time-series multi-criteria potential overlay, fractals, raster map algebra, cellular automata and multi-agent methods.

The following chapter compares two new SDSS prototypes with the criteria developed during the previous sections. The two prototypes which were specifically designed for landuse change forecasting are:

• the "Landuse-Planner", a co-operative information system for national landuse planning developed by RIVM, LEI-DLO,SC-DLO,RPD,VU_Amsterdam & GEODAN,
• and "RAMCO", a decision support system for regional resource planning developed by the RIKS and the Dutch Institute for Coastal and Marine Management.

Finally a proposal is made for a conceptual design of a space-time decision support system (STDSS) which integrates the ideas reviewed and discussed.

Andrew Fenton

Spatially-Enabling the Database

This information age is upon us and organisations are investing in information systems (IS) to increase competitive advantage and profit. The typical development stages of an IS within an organisation may be modelled and a common goal for such systems is to be open in terms of the systems themselves and the data upon which they rely. Geographic information systems (GIS), being a subset of IS, should be no different in terms of this goal. However, the complex demands of the operations which GIS are expected to perform have lead to the popular hybrid architecture which has resulted in GIS-islands and vendor lock-in. Consequently, it is proposed that GIS do not progress through the stages of growth model as would be expected and that instead, they tend to falter at the stand-alone stage.

By changing the architecture from the hybrid model to one which is database centric, GIS may progress more easily beyond this stand-alone stage and thus become more integrated with other IS within the organisation. This should in turn allow the organisation to use the GIS technology more strategically and share the data throughout the enterprise. A spatially enabled database which supports this database-centric approach, allows both the spatial and attribute data to be stored openly in the same database. This study examines the strength of this proposition in order to support detailing how a database may be spatially enabled and two applications have been developed to illustrate this.

Michael Fisher

Project Management for GIS: Towards a method for selecting appropriate GIS development methodologies

In referring to GIS projects, Reeve (1995) stated that "it is only when technology is successfully married with appropriate organisational structures that any real benefit accrues".

This dissertation examines the organisational issues that regularly beset modern information systems (IS). Cartography has traditionally been controlled through the use of Critical Path Analysis (CPA), but modern Cartography/Geographical Information Systems (GIS) are too complex for simple CPA. Mainstream computer projects are notorious for running late and being over budget, with a fair percentage failing to meet user requirements, which suggests that something very important to the survival of a project is missing. The failure to acknowledge the organisational and staffing issues is thought to be the cause of some of the failures.

Project management, which is characterised by new ways of doing things, used methodologies, to combat these failures. This study initially reviews the subjects of project management, types of organisational groups and systems theory, in order to maintain a totally unbiased view of project management. This is important so as not to distort the views relative to GIS during the study.

Given the complexities of the IS/IT (Information Technology) field it was difficult to define GIS’ place in the business field. GIS is examined stage by stage, separately from project management to compare it’s needs relative to IS/IT and MIS (Management Information Systems). It was found that GIS’ general requirements were not too far removed from those of MIS, and therefore could utilise similar methodologies for it’s management. Several lifecycle models were examined from which a best course for the management of a GIS was established, but nowhere was there to be found a perfect methodology. This was because the complexity of organisations coupled with the variety of projects and systems made it highly unlikely that a perfect match could be found for any system. Therefore the best course of action ws to ensure that stakeholders were knowledgeable of the issues involved with these type of projects, and GIS in particular. In the final section a review of the study summarises GIS’ position, suggesting a best way forward for the management of GIS projects, followed by a few suggestions towards a methodology for GIS.

William Gallagher

A Strategic Planning Methodology for Introducing GIS into Large Organisations: A Case Study of the Renfrewshire GIS Project

Stuart Gill

GIS and Crime Analysis Within the Metropolitan Police Service

Hans Hainje

An assessment of the problems of planning and implementing GIS in a (complex) multi-partner organisation

Organisations are not stable; they are forced to react on many changes. As a reaction on interaction with their environment or due to cost sharing, co-operation is an important organisational instrument. Merges, joint ventures or other forms of co-operation occur regularly.

Another organisational instrument is information technology (IT). Because many organisations produce and use a large amount of geographically referenced data (especially in the public sector), many organisations embrace Geographical Information Systems (GIS) as a specific information technology. Whereas GIS is growing to maturity and becomes a general accepted instrument on the one side and co-operation on the other side more often occurs, the (desired) use of GIS in multi-partner organisations will be a returning issue. For single-user organisations GIS implementation projects are complex projects. The implementation must be seen as a process of organisational change. The implementation of GIS is a (complex) multi-partner organisation is not the same as implementation in single user organisation.

A business case (six co-operating Friesian Water Boards) confirms that a GIS implementation project in a multi-user organisation is more complex than implementation projects in single user organisations. The co-operation of the Friesian Water Boards is therefore not only for economical reasons. The filling in of an integrated water management approach requires far-reaching co-operations on several terrains. This enlarges possibilities for the use in the multi-partner organisation of the Water Boards. Due to ambition, difference in tempo, the existence of actual (spatial oriented) bottlenecks and differences in available resources, the points of departure are not the same.

During a business study is concluded that GIS co-operation demands of: data-management, standardisation, a well designed support organisation, communication for the creation of awareness on several management levels. Beside (simple) pilot projects (for each organisation) with early deliverables, a new pilot is introduced during the business study: a joint exploration project. The exploration project delivers greater insight into future organisational aspects, and implementation strategies and may produce some potential GIS-champions.

The design of a support organisation in a multi-partner environment is complex. Informal communication is being inferior to a single user organisation. For this reason short communication channels between specialists is recommended. Furthermore it is recommended to draw a distinction between central and de-central tasks on one hand and to draw a distinction between data management, technical management and application management on the other.

Johan Hallgren

Towards Open GIS : The Database Component Evaluated

Paul Hopgood

The use of Remote Sensing and GIS to Ascertain the Spatial/Temporal Change in Vegetation Composition on the Southern Coastline of the Yangtze Estuary, Peoples Republic of China

Bassem Katlan

Agricultural Land Use Evaluation at National Scale Case Study Syria

This dissertation is an investigation into the use of Agro-ecological zone methodology to make a quantitative land suitability assessment for the strategic crop (wheat) under rainfed conditions using geographic information system GIS techniques as modelling tool. This allows us to answer the following questions:

• Is there sufficient land to meet future food and agricultural needs?
• Where are the potential utilisable areas and what is the range of their potential?
• What is the risk of land degradation?
• What are the limitations to production increases?

Agro-ecological zone projects are suitable to assess production potential of developing land resources and provide the physical database necessary for future planning of agricultural activities especially for areas in dry lands and developing countries. (Report on the AEZ projects FAO).

 

Dry and semi-dry environments. Wheat as one of the most important crops in Syria, could be undergone to some debate of land suitability for a specific crop within the Agro-ecological zone (AEZ) methodology.

Machtelt Kusters

A study on metadata creating an independent company metadata set

GIS is a tool useful in supporting the planner with spatial planning problems. A GIS converts geospatial data to meaningful information. Generating and managing geospatial data are cost and time consuming. One way to reduce the costs is re-using existing geospatial data.

When searching for geospatial data as input for GIS, it is quickly identified that it is of great importance to have the geospatial data organised. The data should be organised in such a way that the user can easily identify whether the data is suitable geospatial data to perform a specific GIS task.

A helpful tool in here can be a system or method that collects organises, selects and most of all will find this suitable geospatial data. Before finding the suitable geospatial data, characteristics of the geospatial data have to be defined. These characteristics should be organised in a uniform way so that it is simple to retrieve, to establish the quality and provide geospatial data. Systematic recording of all kinds of characteristics of geospatial data is the basis of a metadata system.

Planners working at the regional government of Limburg are dealing with spatial planning problems situated within the Netherlands, Belgium and Germany. Question arising when using GIS and searching for geospatial data are amongst: What data are available? Does the data cover the target area? How is the quality of the data? Are the data usable for the project? The metadata of geospatial data have to meet the requirements of content, quality and availability.

Of great importance of the regional government is to give answers to describe geospatial data by metadata in such a way that it meets the requirements of supply and demand: the search for geospatial data as available in - and outside the organisation and to get access to this data.

Based on a theoretical evaluation, recommendations are made to meet the demands of the regional government of Limburg to support on the one hand the possibilities of registrations of geospatial data and on the other hand to facilitate access to this data.

There are several initiatives for the standardised registration of geospatial metadata. CEN/TC 287 is a standard developed for registration of this geospatial metadata at European level. CSDGM in the United States and ISO/TC 211 at worldwide level. Each geospatial metadata standard describes the characteristics of the geospatial dataset with help of metadata elements.

Clearinghouses are initiatives for searching and tracing geospatial datasets with help of the electronic highway; the Internet. NCGI, the Dutch clearinghouse gives the regional government the opportunity to search for geospatial data outside the organisation. In the future the search will be extended outside the country. The NCGI works according to the European geospatial metadata standard, CEN/TC 287. This is the format in which the organisation of the NCGI will accept the requests and descriptions of geospatial data.

When investigating the geospatial metadata standards, it can be concluded that there are similarities in the metadata elements as used. An important item is that CEN/TC 287 will be dissolved in the future by ISO/TC 211. The questions remaining are when will this take place, how will the adapted ISO?TC 211 standard look like and, consequently at what point in time, will European clearinghouses migrate from CEN- to the ISO standard?

An implementation of the CEN/TC standard within the regional government of Limburg at this point of time implies taking the risk of a short-term solution. A solution showing more foresight could be implementing a geospatial metadata system consisting of core metadata elements with additional specific metadata elements. The core metadata consists of the mandatory metadata elements of all important geospatial metadata standards of this time; CEN, ISO, CSDGM added with those elements as used by clearinghouses to search for geospatial data.

The additional metadata are company own metadata elements necessary to record for the companies own data housekeeping and may not be relevant to other organisations.

Implementing a core metadata extended with additional specific metadata gives the company the possibility of an independent geospatial metadata standard, the company geospatial metadata standard. Furthermore, by the suggested translation tables it is possible to translate the company geospatial metadata elements into every desirable format of geospatial metadata standard.

In doing so, the implemented company’s geospatial metadata standard company can be independent and will maintain stable to changes made elsewhere in the field of geospatial metadata standard.

Bernhard Maier

GIS-based estimation of forest growth on the basis of forest vegetation maps and topographical parameters in a spatially continuous manner 
(A pilot study in a mountainous region of Austria)

Estimating forest productivity, expressed as yield class or site index, is an important requirement for forestry planning and management. While foresters usually obtain yield class directly in the course of establishing forest inventories, an indirect approach is based on the explanatory value of different ecological variables. Variables which describe the moisture and nitrogen regime of the soil were derived from forest vegetation maps representing potential natural forest communities (PNFC). Combined with topographical parameters derived from a digital elevation model (DEM) these ecological variables were used to fit a multiple regression model to predict yield class. In the course of various investigative analyses of the relations between explanatory variables and yield class a theory-based pre-selection of explanatory variables was made in order to avoid over-parameterisation of the model and minimize multicollinearity. The model accounted for 60% of yield-class variation. Nitrogen figure, altitude, aspect (transformed) and the type of parent material proved to be the best predictive variables.

Because yield class source data were positively autocorrelated, a continuous thematic surface was regarded as a better representation than the conventional representation of non-overlapping polygons. Since all variables were available in a spatially extensive manner, a grid-based yield-class surface was calculated using the regression model equation. In a comparison with a geostatistically estimated reference surface, by means of Ordinary Kriging, it was found that the regression model underestimated yield class in areas with maximum growth, in particular.

Although the predictive value of the regression model is too low to warrant immediate practical application, the study shows how the complexity of forest vegetation maps can be simplified and how this integrative information can be used for forest growth prediction. Finally, certain modifications are outlined which might help to improve the results and allow a wider applicability on a regional scale.

Derek O'Brien

The Changing Nature of Business and the Role of GIS. A Case Study of the Valuation Office (Ireland)

Olutola Odedina

Building the Right System: Using Aspects of SSADM to Develop a Geographic Information System within the Environment Agency

Caroline Pennekamp

Cartographic rules for visualisation of time geographical data sets

To visualise regular 2-D map elements one can use the Cartographic Grammar as set up by Bertin (1973). The guidelines he made for visualisation of qualitative, ordinal and qualitative data sets are used by cartographers all over the world and can be used by GIS users who like to produce effective maps.

GIS software and hardware possibilities however, made it possible to make new kinds of maps from non-regular data sets. For instance by connecting a database to spatial features it is possible to make time sequences of each object. But also the new technical possibilities as multimedia, quick regenerating red/green 3D images generates new output products. These new possibilities have a great new impulse in the world of cartography. The possibility of representing time in maps is one of those new research items.

Research on this new aspect in Cartography has started. In 1994 MacEachren stated "Animation provides a new cartographic variable that can increase the possibilities of exploratory data analysis". He also suggested that "It should be useful to consider this new variable in relation to Bertin’s work".

Subject of interest is to set up new easy to use rules, just like Bertin’s grammar, which can be used by GIS users as a guideline for choosing the most effective representation to express their particular time variant.

In this thesis these issues are explored in theory. An overview is given of Bertin’s theory and the characteristics of time related geographical data sets. An evaluation of the extensions made by fellow cartographers to the theory of Bertin in order to incorporate the factor time in his scheme is made. Finally, a personal proposal of the representation of time related geographical data sets is given, based on the theory models available and extended with new selection criteria. A selection scheme will help a mapmaker in his creating process based on the kind of data, goal and technical possibilities.

Otmar Pilgerstorfer

Scoping into BPR - A Model Assisting the Application of BPR in GIS

Arend Ruizendaal

Modelling hydrological change in Flevo Lacus caused by a Roman canal and dam near Nijmegen (Noviomagus) in the first century A.D

Flevo Lacus was a lake in the Netherlands during the Roman age and a predecessor of the present Lake Ijessel. According to the classical geography of Flevo Lacus, that is the geometry of Flevo Lacus based on available literature, this lake was characterised by open water conditions directly before the beginning of the Roman Age. In the Netherlands, the Roman Age commenced when the Roman general Drusus conquered the present day Netherlands and Germania territory north of the river Rhine. Therefore, he constructed a canal from the river Rhine to Flevo Lacus and filled it with water by damming up the river Waal.

In Microsoft Excel 97, using spreadsheet functions as the main modelling language, a hudrological model of Flevo Lacus was built. The cells of a spreadsheet were used as 1 by 1 kilometre georeferenced rasterpoints. Four rasterpoints were given special roles, two of them were assumed as the main inlets of Flevo Lacus (the rivers Ijessel and Vecht), and the other two as the outlets (the Oer-IJ and the Vlie). These inlets and outlets were assigned some hydrological characteristics, like discharge and fluid velocity, as expected directly before and after the Drususevent (the construction of the Drususcanal and Drususdam).

Using these boundary conditions, the spreadsheet model of potential flow was used to calculate the fluid velocity, in all rasterpoints of Flevo Lacus, directly before and after the Drususevent. Based on these results, it was concluded that, due to the formation of peat, no open water conditions could be found in most parts of Flevo Lacus before the Drususevent. Directly after the Drususevent, Flevo Lacus developed its open water conditions by erosion, a process reported by the Romans.

Roger Scott

The Application of a Simple GIS, with GPS Input, as an Aid to Fieldwork

Peter B Shilson

Just Getting By: A pragmatic approach to GIS training

When a new or revised method of information handling is introduced into an organisation its users have to learn to use it, a process generally supported by training. This Paper defines the context within which any organisation selects training, the market place within which GIS training resources are offered and the choices that can be made or the reaction between the context and marketplace.

The context is identified at three levels. Firstly, at the broadest level, is the ignorance that may exist of the nature and likely impact of a GIS, the competing demands for resources and the extended timescale over which GIS training may need to be delivered. Secondly, other factors concern the nature of an organisation particularly the degree to which it is a learning organisation. These are not necessarily under the control of those who have to plan GIS training. Thirdly, several detailed considerations are discussed, particularly the roles of self-development and syllabus.

Two surveys were carried out. A survey of possible academic and commercial providers of GIS education and training identified several problems of cost, timeliness and course content. A case study of a major local authority town planning department identified many points relating to objectives, budget and training practices although previous experience of GIS could help its training strategy.

The Paper agrees with other commentators that it is the ability of an organisation to learn, adapt and move on that will provide it with a usable GIS and a staff capable of using it to good effect. It is important to move in a direction rather than towards a defined end-point. This Paper concludes, from its definition of the context and from the surveys, that the method of training delivery most likely to be effective in supporting such open-ended implementation will have to be Just-get-by, a pragmatic blend of Just-in-case academic education, Just-in-time short course and commercial training and Just-for-you self-teaching

John Sweeney

Diffusion of Geographic Information Systems in Local Government in the Republic of Ireland

Freek Vos

"GIS Tools" For a Landscape Architect

As no other the Landscape Architect makes use of spatial data. The spatial data are collected by the Landscape Architect himself, on several manners. But at present a lot of digital files and maps are available. A disadvantage of these digital files and maps is, that they are not composed by Landscape Architects. But with some adaptations, these maps and files are suitable for the tasks of the landscape architect. A GIS provides information about the world, presented in maps, graphs, and tables, based on which one may perform actions. The landscape is built by a number of landscape elements. These elements have a number of attributes. At first, it must be possible to insert the landscape (elements) into a GIS. Then and with help of various analysis and visualisation tools, a GIS is a helpful tool for the Landscape Architect. But in practice the use of a GIS by the Landscape Architect, lags behind. This is in contrast with other disciplines. A serious barrier between the Landscape Architect and GIS forms the language and the definition of the spatial entity. The language of the Landscape Architect is an emotional one; the language of a GIS is not. In order to unite these disciplines it is necessary to identify a common language. GIS requires definition in absolute terms. GIS requires that its subjects are classified, so it is necessary to identify common features in the landscape and then describe them in absolute terms. Another problem is that the described features can only partially inserted into a GIS. In the case this problem is solved by a dual approach between the visual and functional features. But other options and methods are possible. But when these problems are solved, or partly solved, a GIS will be a very useful tool for the activities of the Landscape Architect.

[ Return to top ]