Home Articles Creating a spatial database for the Mumbai fire service by using GIS...

Creating a spatial database for the Mumbai fire service by using GIS / RS techniques: A case of the CBD, Mumbai

C. B. Sunil
School of geography, UNSW, 2052, NSW, Australia.
Tel: (61) (02) 9385 5537. Fax: (61) (02) 9313 7878.
[email protected]

Introduction and Issues

figure 1

Urban fires are often devastating resulting in the loss of property worth millions and the lives of many. There is a worldwide understanding of the importance of creating a database on the land (urban or rural) for the assessment and analysis of fire hazards / risks. However, in some major cities, this process is in a stalemate phase and databases are not at all existent. The Mumbai fire service is a case in point. In this congested multicultural cosmopolitan city, which is characterised by dense population, frequent traffic jams, presence of many slums, the effects of a major fire breakout will be disastrous and the consequent damage will run into heavy loss of lives and property running into millions of rupees. The Mumbai fire service adopts a crude methodology of combating fire, where a short-term methodology is followed, using the local knowledge of the staff, and a rough calculation is made for resource allocation and management. The location of fire stations has been the same from the time they were incepted, and have not considered the phenomenal exponential burst of population and the economic boom in Mumbai over the last decade. There are various regions that are not serviced by the fire stations within a reasonable response time and yet there are some regions serviced by two stations! The lack of proper database has been the bane of fire service and has gone unnoticed by the authorities and bureaucrats alike. This methodology followed by the fire service is beset with disastrous results, and very often generates remarks about the overall operational capability of the fire service department. The availability of spatial database is more acute here than anywhere else.

Two major issues need to be addressed in the case of Mumbai Fire Service:

  1. There is a need for building of a database that gives an understanding of the topology of the region, which could also be in the form a generalised land use map (figure no 1)
  2. Secondly, to map the various features (manmade and natural) in space that are directly concerned to the fire service for hazard and risk assessment, and integrate cartographic information available in unprojected maps, it is necessary to tackle disasters by comprehensively taking into account the intricate variables like population, infrastructure, topography etc. These variables are a mixture of textual and spatial features or in other words discrete like petrol pumps, buildings etc and continous features like the land cover, water bodies etc.

To create such a database in a very short time is essential since emergency situations can happen anytime without notice. This is possible by integrating variables, that are considered important for the respective emergency organisations (Chart 1). These variables are indeed a mixture of spatial and textual details. The technologies of GIS and Remote Sensing, can help in generating maps and reports, that will prove extremely desirable to support the intricate nature of operations by these emergency organisations.

Study area

Vector Overlay on Raster CBD Area Mumbai

The port city of Mumbai is located on the western coast of India and has been the traditional trade and commerce hub of India since long time. The study area (approximately 7 square kms) is a section of the southern tip of Mumbai, that encompasses the CBD also.

A map to the scale of 1:25,000 was to be generated en route the study of networks and generation of shortest path for the Mumbai fire service, as a mere demonstration of the efficiency of a “Geographical Information System”. The emphasis was to display each and every building and street so as to create a database, which could be used by the Mumbai fire service to plan for an emergency.

The Mumbai Corporation planning department generates maps to the large scale of 1:2,500. These maps are generated as part of the development plan of Mumbai for 2000 and beyond. These maps are priced at $15 each, and are very comprehensive and detailed. They carry information about every street, building and some adjacent manmade features. However, these maps are not georeferenced, and are therefore not fit for ready analysis on a vector based GIS system. These maps were georeferenced with the survey of India (SOI) maps, and projected to generate an “error free coverage”, where the root mean square (rmse) is within acceptable limits, using ArcInfo’s PC version 3.5.1. The whole to part concept was used to digitise the five maps together in a manner that their streets were properly aligned The base maps were ready for digitisation and feeding into the computer spatial database creation in GIS. Tics (registration points) were marked on the maps, exactly matching the common points to which spherical coordinates were assigned. Once all the maps were fed with the registration points, it was digitised by keeping the part 1 and part 2 maps together on the digitising table`s active area, in a manner that 70% of part 1 and 30% of part 2 were a single entity. The remaining three parts were digitised similarly, and joined to make a geo referenced coverage. The whole to part concept enabled a creation of a perfectly joined map layout of registration points Mumbai CBD.

A blank coverage was created that stored the registration points, and the coverage was created onto these. During digitisation, additional care was taken to avoid editing as far as possible. This was achieved by employing ‘overshooting’ and setting the snap distance of the arcs. Once the whole map was digitised, it was cleaned to create a polygon topology and build command was issued to merge intersections. The weed tolerance was carefully specified along with the clean command. Once cleaned, labels were created by the “createla ” command which helped in displaying camouflaged slivers and duplicate arcs. Simple macro languages (sml) were extensively used throughout the inputting of the data.

The final land use map (fig 1) was quite detailed accurate and comprehensive. This landuse map shows the parcel of land in the CBD area of Mumbai and is derived from the planning maps which are to the scale 1: 2500. It is generalised and is actually a starting point to the creation of a database that will help in the process of hazard / risk assessment for emergency services particularly fire services. This map gives an idea for priority zoning so that allocation of resources and the concentration of weights can be planned. It is noteworthy to mention here that by itself the resolution of the IRS pan image as well as a hybrid with Pan IRS 1C and Liss III (figure no 2) is not enough for generating details required by the fire services, therefore it is essential to generate information by integrating aerospace data with GIS, that consists information on each and every feature on the land parcel. This database will help in hazard / risk assessment, location / allocation modelling, network analysis, spatial analysis etc, which will further improve the planning of dynamic and static resources. The overlay of such maps on temporal aerospace data (figure no 3) will give currency to the maps and enable landuse monitoring and change detection.

Integration of GIS and Aerospace Data

Scope and discussions
The Mumbai fire service needs access to a strong GIS database that is very essential to combat fire hazard. It is interesting to note that a huge number of maps containing cartographic information are incapable of being analysed, since they do not have a proper projection, but with the above methodology such maps can be georeferenced and their information can be used for creating, at first a landuse map and then a space use map (map that stores information about the various objects like buildings and other structures), on which spatial integration and analysis for emergency services can be carried out. High resolution remotely sensed data are very useful for lending repeativity and currency. In India, where emphasis at the moment is on creating a large database on a national scale, non-availability of data data is a serious impediment to emergency services. But this technique is effective in utilising the vast reservoir of cartographic information lying moribund in many maps. The high resolution (5.6 metres to a pixel) of the Indian Remote Sensing Satellite (IRS 1C) can be used in conjunction with the vector GIS maps for maintaining the currency of databases.