GISdevelopment.net ---> AARS ---> ACRS 1996 ---> GIS

Towards a Geographic Information Infrastructure for Sri Lanka

Peter Bitter
Dept. of Geography
University of Zurich - Irchel
Switzerland

Sarath Jayatilaka
Survey Department
Sri Lanka


Abstract
Many GIS installations in developing countries are realised within the framework of development cooperation, i.e. they are tied to the life cycles of 'projects' which typically last a few years only. Quite often, little of the efforts remains operations after the projects end. As a counter position to this, the concept of a national geographic information infrastructure (GIIS) is discussed in this paper, and it will be pointed out why it is important that future GIS projects can operate within the environment of such an infrastructure. Institutional as well as technical components of the GIIS are identified. The we go on to present two types of initiatives that are being proposed in Sri Lanka to create this infrastructure: 'institutional initiatives' are taken to create the necessary bodies to set standards and define policies in the fields, while 'production initiatives are undertaken to provide the data itself.

1.The concept of a geographic information infrastructure
It is common ground nowadays that 'data are the fuel of GIS'. In other words, any GIS project can meet its objectives only if the appropriate data is available. Standard computer hardware and GIS software can help us to produce customized products, even it we don't have a deep understanding of geographic data of cartographic processes. However, they will help us only if we feed the correct data.

Moreover, data are usually the logiest-living part of any GIS implementation: Computer generations are typically outdated in two or three years, software vendors keep on asking us to by 'new' versions every couple of months, but geographic data are there to stay. This, of curse, doesn't mean that no updating will take place, but this will typically affect far less then 5% of the items in a database.

On the other hand, the initial acquisition of the necessary geographic data commonly accounts for about 80% of the total cost of an average GIS project. Maintenance of the data (updating) will later add to the cost. Therefore it seems quite later add to that cost. Therefore it seems quite logical to try to cut on the cost of data acquisition, but, as we said, if that means that the necessary data is not available at all or at least not in any useful form (scale, accuracy, etc.), the success of the GIS project will be endangered.

The solution to the dilemma lies in the keyword 'infrastructure'. Basically, infrastructures are things tht can be used by anybody without having to play the full price of it's establishment. The price can be measured interms of money, but also in terms of waiting time. We use transport and communication infrastructures every day, but no one could afford to found a road or a telephone network personally, and no one realizing a journey. So what we really need to ralsise GIS projects efficiently is a "Geographic Information Infrastructure".

We think it is helpful to consider geographic data as an important part of this infrastructure that allows GID to function efficiently in terms of cost and time. However, the data itself is only one part of the infrastructure. It is the part which corresponds to the physical aspects of common infrastructures, e.g. the roads or the telephone lines. Quite obviously, there are other aspects which make up the infrastructure: for telephones to be useful, rules about signal transmission, numbering schemes etc. are needed - in short, technical standards are absolutely required to avoid chaos. In the days where access to most telephone networks was controlled by a single Government corporation, setting and enforcing standards was relatively easy to accomplish. However, the regard to the proliferation of geographic data, we clearly live in a world with many different providers and users with different needs. The conclusion from that is that is that is that common standards (e.g. common vocational reference, common attribute definitions, data exchange standards, data documentation standards) are required for users to be above to efficiently use data from different sources.

In addition to the just mentioned technical parts of the Geographic information infrastructure (he data itself and the data standards), institutional or organizational elements are required too. These are:
  • A consistent copyright legislation which also includes digital data. It is quite obvious hat no potential provider, exp. From private sector, is interested in producing and selling digital geographic data if he has n means to reasonably protect his investment. Since thee are no practical possibilities to protect data through technical measures (such as software keys), the only way is to protect data by means of contracts. The legislation also has to protect the interests of the creator of original documents, e.g. Survey Department Topographic maps. A survey of the situation in Europe can be found in [EUROG].
  • Next to copyright, a consistent pricing scheme at lest for data provided by Government agencies is not an absolute must, but very helpful for potential users outside the administration. Also the question whether commercial and non commercial users should be rated alike has to be resolved. In either case, the price must not be prohibitive - otherwise, potential data users will find it cheaper to capture the data by themselves (and thus duplicating efforts), to use other, inadequate data (e.g. of too small scale), or take the risk of illegal copying. Thus the geo information community will never develop. The extreme example is that of the US who consider data produced by Federal Agencies as public property which must be available at the nominal cost of reproduction.
  • Co-ordinate efforts for education, research and development are needed to build up a critical mass of people with various GIS capabilities, and to Crete certain capacities for applied research. With constrained research budgets, the top priority will not be on fundamental research to create new data structures and algorithms, but rather on solving some specific application problem.
  • Last, but not least part of the Geographic information infrastructure are something which could be called "Entry points to the GIS worlds": This includes for instance Metadata such as data catalogues and data dictionaries, but also a responsive customer serve by the providers. Quite often, the first steps of potential would - be users of without much experience, they have to specify and find what they exactly need, and they have to fight their way through a maze of officers, rule & regulations which they have to fight their way through a maze of offices, rule & regulations which they barely understand to eventually get what they want. Quite likely, they will give up their project somewhere. Halfway, or find it cheaper to re-gigitize the data they need. (this situation is by no means specific to Sri Lanka - it is a very common situation also in industrialized countries). The contrary would be situation where the available dataset are well documented in publish data catalogues, and inexperienced users can get assistance. Experience from Switzerland and suggests that users are quite prepared to pay a price four such advice, it is competent and responsive.
Further reference to such institutional aspect can be found in [PSP], [FOX], [YEH].

To conclude this part of our paper, we would like to draw the attention on some thoughts on who could create such a geographic information infrastructure :

It is quite clear that multitude of players will take part in the game: it is usually Survey Departments who provide the locational reference and the topographic base data, Department of Census who provide statically data, perhaps Geological Departments to provide geological maps in digital form, international satellite image distribution, private companies providing value added cartographic and statistical data, and, last but not least, the users (who are often at the same time producers of certain datasets) with their won expectations. The conclusion from that is that an infrastructure which is useful for (almost) all players has to created under the leadership of a broad - based committee with a certain powers to issue guidelines and directives which are compelling at least for Government Departments. In other words, backing by Governments. In other words, backing by Government or specific legislation is essential.

The second through is on the question, whether a geographic information intrastate can be crated through foreign-funded aid project

It is obvious that the creation of this infrastructure is a long-term goal. Coordination can be extremely tiring and time-consuming, and the risk that the final compromise will not meet ones own expectations is rather high. Moreover, building the infrastructure should be perceived as a continuous process which will not end after a certain number of dataset have been digitized. Existing datasets have to be updated, new demands will arise, technical possibilities will change. Thus what is required to set this process in motion and to keep it alive is unlimited political and finical commitment to it-unlimited meaning here 'not temporally limited".

Contrarily, most foreign funded aid projects have a clearly limited life-cycle of typically two or three years, sometimes with options for extension. In their lifetime, projects have to achieve clear-cut results. Building a digital geographic database for the project area may well be such a result, disbursing certain amounts on GIS hard-and software may be another expected result. However, the project will inevitably end some day, and the chances that the counterpart can (or even wants to ) maintain" his" GIS past that day are rather small. There are numerous examples of that kind in Sri Lanka. Moreover, donors like their counterparts often have a habit of 'fencing in their garden' and not tetting anybody trespassing, which of course inhibits coordination within the framework of a common infrastructure. This means that aid projects works under conditions fundamentally different from what is required to create a coordinated geo-graphic information infrastructure.

For the stated reasons it seems unwise to expect a foreign-funded project to create a national geographic information infrastructure. This doesn't mean that external projects can't help in the process, e.g. by supplying consultants, training, also hard- and software, and supporting certain specific tasks such as the creation of specific datasets. However, this must always be compatible with the directives and technical standards of the national geographic information infrastructure. Otherwise it will be pretty useless for the country as whole, through it might still be useful for the limited purpose of the particular project at hand.

2.Institutional Initiatives in Sri Lanka

2.1 GIS society

A GIS Society has been initiated by the staff of the Department of Geography at the University of Colombo and the Forest and Land Use Mapping Office in the Mahaweli Board in Kandy. The Society consists of members who are mostly professional in GIS, but excludes the vendors.

This society has not yet been fully operation. The constitution is being prepared at the moment, but however, the Society held a Technical Seminar on GIS which enabled the GIS users to interact.

2.2 Inter-Agency Committee
On a national seminar on policies on spatial data which was attended by the then Hon. Minister and a large number of top Government officers and University staff, the idea to create an Inter Agency Committee was first discussed [PSP]. It took quite some time to clarify which Department was to take the lead in this, but finally a paper on it has been submitted to the Cabinet this year. The Cabinet has approved the formation of the Inter-Agency-Committee and gave it a mandate to coordinate the GIS activities under the Ministry of Agriculture, Lands and Forestry. Thus top-level support can be taken for granted. The Committee is now in the formative stage. It will certainly draw on the experience of the professionals in the GIS society also.

2.3 Initiatives to Strengthen Education and R&D capacities
the University of Colombo has been equipped with a small GIS lab in 1992. Gradually , GIS training is being built. Into the curriculum of geography students. In additions to that, the Department recently offered a two-week intensive course on GIS for professional outside the University. The University of Peradenyia too has taken steps in the establishment of a GIS section. However, the authors of the paper feel that more concentric efforts in training at University level should be undertaken in GIS in all the University.

2.4Initiatives to improve accessibility of geographic data.
Several authors have complied catalogues of GIS activities: Among others, [WEERAKMARA] contains a survey of GIS user in Sri Lanka; [DENT] lists of providers of maps and other information related to land resources, and [CRS] contains a detailed to land resources, and [CRS] contains a detailed data dictionary of a pilot GIS that has been set up to support decentralized development planning. The Center for Remote Sensing is also, in cooperation with Zurich University, establishing a digital satellite image catalogue.

The common denominator of these publications is that they are essentially piecemeal, and they are not kept up to date. What is clearly lacking is a regularly updated comprehensive catalogue of available GIS data set. This information oon each dataset in the catalogue should comply with some standard, e.g. the FGDC Metadata standard [FGDC]. On the log run, it is also desirable to make this catalogue accessible through the Internet. The authors feel that to maintain such a catalogue should be one of the tasks of the Inter -Agency-Committee.

3.Production initiatives

3.1 Topographic Data Base project

the Survey Department is planning to establish a topographic vector database of the whole country database of the whole country. There are three main objective behind this move:
  • to be able to supply digital topographic data of homogenous quality to GIS users.
  • To issue the 1:10'000 sheets by computer plotting on demand
  • To facilitate easier updating of the printed 1:50'000 topographic sheets
It is quite clear that the database will have to be established at two different scales (10'000 and 50'000). The 1:10'000 data is already being collected during stereo plotting, but not cleaned and organized into a database. The 1:50'000 data organized into a database. The 1:50'000 data is not available yet, the existing sheets could be canned and victories as an interim solution. A team of officers of the Survey Department then went to investigate the various possibilities and to establish a preliminary data base design for a pilot project.

A first meeting with potential users held in March 1996 showed that needs n terms of digital data seem to be roads, land cove, administrative boundaries, elevation. The scales in demand are both 1:10'000 and 1:50'000. However, one has to be aware that users will -understandably - possible scale for the largest possible scale for the least amount of only", if they are being asked in such a way.

The Survey Department team is at the moment studying the various options and also trying to build a small database, and they will again meet the users to discuss the database structure.

3.2 Land use mapping 1:50'000
The Center for Remote Sensing at the Survey Department is creating a new series of land use maps at 1:50'000 scale. The legend (classification) ahs been discussed with the main users and is in accordance with standard land use classification schemes. The base map is taken from the normal topographic sheets, but has so far not been digitized database. The series will not be printed, but will be available as digital data or plotted sheets on request. The landuse mapping is almost completed for the Southern Province now.

3.3 Other production initiatives: FORLUMP, FD, NBRO
There is or was a number of projects which have create geographic database, some of them crated in the country is at the Environmental Management and sustained Development Management and sustained Developed Project in the Upper Mahaweli Catechment (ENDEV, former Forest and Land use Mapping Project) at the Mahaweli Authority [TURNER]. They have at the moment digital data based on updated Survey Department 1:10'000 maps covering most of the Central Province. Others are the Forest Management Project at the Forest Department [LEGS], the Landslide Hazard Mapping Project at the National Buildings Research Organisation, to name but a few. Further examples can be found in [MORAPAY] or [YAPA 1988]. With varying success some of the institutions try to make their geographic databases which were , of course, created as an infrastructure for the respective projects, self -sustainable over the end of the foreign aid. However in all the cases it seems presently unclear how these databases will fit into a larger geographic information infrastructure at a national level.

To be fair, one has to mention that standards didn't really exist when thee projects started their GIS activities, and there was little awareness on the long -term benefits of a coordinated sense, these projects are certainly important lessons in an ongoing process.

4.Conclusion
We have discussed only a few production and institution initiatives, there are certainly more, and there need to be more activities. However, the crucial point is that these activities fit together and form a comprehensive Geographic Information Infrastructure, on which future GIS projects can base upon. Only then GIS will become an efficient tool in planning, resource management, business, and other applications of spatial information. Otherwise, it will be like having a car, but not roads and no filling stations.

References
  • CRS Center for Remote Sensing (1995) : Report on the Pilot GIS Weeraketya & Bulatsinhala Divisional Secretary Areas
    Survey Department of Sri Lanka, Colombo
  • CHRISTIANSEN, T (1995): Geographical Information Systems for Regional Rural Development Projects in Developing Countries.
    Ph D thesis, University of Giessen, Germany
  • DENT, D L & L K P Goonewardene (1993): Resource Assessment and Land Use Planning in Sri Lanka: A Case Study
    International Institute for Environment and Development, London
  • FGDC Federal Geographic Data Committee (1994): Content Standards for Digital Geospatial Metadata
  • FOX, J M (1991) : Spatila information for resource management in Asia: A review of institution issues
    International Journal of Geographic Information System 1/1991
  • GEGG, C & N Jewell (1992): A new Forest Map of Sri Lanka
    Forest Department, Colombo
  • MORAPAYA, R B (1990) : regional Planning and Information Systems in Sir Lanka
    in: Critical Success Factors - key to the use of Information Systems Technology in Local / Regional Planning
    UN Centre for Regional Development, Nagoya, Japan
  • MORRISON, J L (1993): Tomorrow's National Spatial Data Infrastructure -Federal Digital Data Today
    In : ASCM Bulletin Nov. /Dec. 1993.
  • EUROGI (1996): Legal protection of geographical information
    European Umbrella Organisation for geographical Information.
  • PSP Sri Lanka/Swiss Planning Support Project (1994): Report on the Policy Seminar on Spatial Data.
    Survey Department of Sri Lanka
  • TURNER, S T R White (1994): Geographical Information & Systems for natural resource management in South East Asia.
    Environment and Forest Conservation Division, Mahaweli Authority of Sri Lanka
  • WEERAKUMARA, (1995) : Geographical information Systems in Sri Lanka: Problems -Objectives (Preliminary Report)
    National Aquatic Resource Agency, Colombo
  • YAPA,LS (1988): Computer -aided regional planning : a study in rural Sri Lanka Environment and Planning B
  • YAPA, L S (1991): Is GIS appropriate technology?
    International Journal of Geographic Information System 1/1991
  • YEH, A G (1991): The development and application of geographic information systems for urban and regional planning in the development countries
    International Journal of Geographic Information Systems 1/1991