True colours of urban green spaces : identifying and assessing the qualities of green spaces in Kuala Lumpur, Malaysia

This thesis starts from the proposition that the ingrained perception of urban green space as being synonymous only with fairly well maintained amenity parkland is too narrow and generally overlooks the many environmental and social benefits that other types of green space and their natural habitats bestow on urban residents and wildlife. A critical review of the literature on the benefits which different kinds of green space confer on urban residents in environmental, social, health and well being and economic terms confirms the need for a more holistic approach to the study of green spaces in cities and also highlights the need to develop and realise a more comprehensive "ontology" of urban green space in tropical countries, a fundamental task which is a main concern of the present thesis. From reviewing the classification schemes or typologies used in different countries to formally recognise and to distinguish different types of green space, the author develops a new, expanded typology for urban green space adapted to Malaysian conditions, aiming to use this as far as possible as a framework to categorise the green spaces of Kuala Lumpur (KL). KL provides a particularly interesting case study as a rapidly growing city in a developing country with a tropical climate, a context where there has been relatively little research on urban green space, despite shade being particularly appreciated in very hot climates. Also KL has experienced much loss of green space in recent decades: on its periphery from urban expansion; and around the city centre from the drive, fuelled by economic growth, to use central land more intensively. The main empirical analysis in the thesis uses data obtained from remotely sensed satellite images of high resolution (from the IKONOS satellite) to try to identify all vegetated forms of land cover in KL and to discern their nature, primarily whether trees, shrubs or grass, regardless of their location, using object oriented software to process the IKONOS data. The degree to which the different types and functions of green space can be identified from IKONOS imagery using both semi-automated and manual methods of visual interpretation is then compared. The results show that, using high resolution IKONOS imagery, it is not possible to identify unambiguously all the types of green land use or green land cover that are found in the proposed, new typology of green space, either by using semi-automatic classification or by visual interpretation, although the latter enables more types of green space to be distinguished with confidence. A key result of the preceding analysis, nevertheless, is to produce maps of green space showing the foregoing 3 classes of vegetation (plus water, bare ground and built up areas) for the entire city in very fine detail using first a semi-automated classification followed by selective manual revision. This produces a more complete picture of the geography of these 3 basic types of green space across the whole city than the typical picture purely or mainly of public parks generated from the typologies used by city governments in developing countries, including KL, simply reflecting their traditional concerns being largely restricted to the latter kinds of green space. These finely detailed maps showing the complex mosaic of green space are, in some respects, the most important result of the thesis. These maps of green space produced from satellite data are linked in a geographic information system (GIS) with data on land use for small land parcels and, using dasymetric methods, with data on population from the census to produce a range of alternative, illuminating perspectives on the nature and extent of green space across the whole city, often at a very fine geographical scale, and including an analysis of the relative provision (or lack thereof) of green space over the whole city; this also yields insight into the role of particular green spaces in the wider urban system. Subsequently, the use of GIS operations enables officially recognised green spaces and the even more extensive and diverse areas of green space not officially recognised to be mapped and examined separately, possibly for the first time in KL. A social survey designed mainly for urban planners and landscape architects in KL was carried out mainly to learn and study their views on the nature, roles and benefits of urban green space, on the new expanded typology, on the problems of protecting urban green space in KL and on what attributes of green spaces they considered should be seen as most important when considering how much priority a particular green space should be given for preservation. From some 38 environmental and social criteria the 41 respondents considered very important, 31 criteria (13 environmental and 18 social) were chosen as attributes to use in evaluating 17 different green spaces of various types in different parts of the city through assessment on site by a small team of trained assessors. A smaller subset of 4 environmental and 3 "social" (actually all accessibility) criteria, selected from the foregoing 31 criteria, was identified which could be estimated "remotely" by "desk based" methods i.e. by using the satellite data and the population data held in our GIS, as well as by direct field survey. It was then possible to compare the 3 sets of evaluations for the 17 green areas in the form of overall rankings in turn on the environmental and then accessibility criteria: firstly the ranks of the sites on all 13 environmental criteria, then on the subset of 4 environmental criteria (both of the latter from field assessment) and finally on the same subset of 4 criteria estimated "remotely". The equivalent overall rankings for the 18 social amenity criteria, then the subset of 3 accessibility indicators from field observation and lastly the same subset of 3 but estimated remotely were then compared. The results showed clear similarities and strong correlations between the three sets of evaluations for the 4 environmental criteria measuring aspects of vegetative cover and "green connectivity" but less consistent similarity for the social and accessibility measures, with only weak correlations between rankings on the field and remote estimates for the 3 accessibility indices. The main conclusion is therefore that "remote" evaluation could potentially have a useful role, complementary to ground surveys, in monitoring and assessing green spaces as regards some key environmental criteria and, more debatably, may also be able to provide useful measures of accessibility, which are difficult to estimate from field visits. However, observation on site is necessary for assessment of nearly all the social criteria relevant to evaluating urban green spaces.