Urban remote sensing for a fast-growing megacity: Delhi, India
Urbanization has become a global phenomenon. In developing nations, however, it is characterized by a set of social and economic conditions that differ from those of the industralized world, as well as by explosive population growth. India, for example, is currently experiencing mass migration as people move from the countryside to the cities, and from smaller urban pockets to bigger, metropolitan centers like Delhi.
The population of Delhi is currently around 14 million and is estimated to rise to more than 22 million by 2021. In 2001, the city's population density was 14,387 persons/km2 in urban areas and 1627 persons/km2 in rural areas. In 1901, 47.34% of Delhi's population lived in rural areas. That figure showed a gradual decline to 17.60% in 1951, then 6.99% in 2001.1 Indeed, the rural area is shrinking: it fell from 1158km2 in 1961 to 592km2 in 2001. On the other hand, the urban area increased from 182km2 in the 1970s to more than 750km2 in 2001.
In assessing urban environmental issues, planners and administrators would benefit greatly from up-to-date information about the dynamic processes in and around the city. To this end, data generated by remote sensing, with its repetitive and synoptic viewing and multispectral capabilities, could constitute a powerful tool for mapping and monitoring emerging changes in the city's urban core, as well as in peripheral areas.
For such purposes, various types of information, such as satellite imagery and planning and revenue maps, can be layered over a given landscape using geographic information systems (GIS) and image-processing software. These superimposed maps can reveal an urban environment's complexity and provide valuable insights for planning and land management. But given the widespread mixed land-use patterns in Delhi and other Indian cities, classification can be a very difficult task, and unfortunately, no single source provides all the needed data in a format in that is accessible to every user. Thus, the real challenge in using GIS is to integrate different types of information spatially and thematically. Ultimately, combining such results with socioeconomic data could inform land use planning and policy. Here, using varied data, land cover and surface temperature are investigated as markers of urban change.
In India, data gathered by the LISS-III multispectral sensor and the PAN digital panchromatic sensor aboard the Indian Remote Sensing satellite IRS-1C has been very useful for urban analysis and urban land use/cover mapping.2,3 For this study, maps were produced from the Landsat satellite's thematic mapper sensor for 1992, and from LISS-III images for 2004. Expansion of the city of Delhi from 1992 to 2004 was also mapped. The results show how GIS and remote-sensing technology can provide critical physical input and intelligence regarding the terrain for preparing base maps, formulating planning proposals, and monitoring implementation.
Out of Delhi's total 148,312ha (hectares) of geographical area, agriculture constituted 65,114ha in 1992 and 54,153ha by 2004, a decline of 12% (see Figure 1). During this same 12-year period, highly dense residential areas more than doubled, mainly at the cost of fertile agricultural land. Similar land transformations occurred around Delhi's fringe areas, especially in the city's eastern, southwestern, and northern districts. Medium- and low-density residential areas also decreased. Moreover, the Delhi ridge, once considered to be the lungs of the city, is fast degrading, having diminished from 6.69% of the city's total area in 1992 to 5.52% in 2004. This reduction is attributed to continued illegal tree cutting, quarrying, and construction activity in the southeastern part of the city.
To assess the surface temperature of Delhi, data sets from ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer) taken in 2001 and 2005 were used to analyze the spatial thermal structure of the urban environment. This study included analysis of `hot spots,’ or high-temperature surfaces, as they relate to the physical characteristics and uses of urban surfaces. The method involved deriving parameters governing surface heat fluxes, constructing statistics of ASTER thermal infrared imagery, and using GIS to combine this information with land use/cover information derived from ASTER data sets, as well as with data from in situ measurements. The average images reveal spatial and temporal variations of surface temperature and distinct microclimatic patterns (see Figure 2). The combined interpretation of the thermal images and the land use/ cover classification clearly shows the effects and consequences of physical surface properties.
The estimated surface temperature ranged from 26.93 to 38.88°C with a mean temperature of 32.66°C. Central and eastern areas, characterized by dense development, had higher surface temperatures, whereas parts of the northwestern, northeastern, and extreme southern areas, corresponding to agricultural cropland, wasteland and bare soil, and fallow land, were cooler. Temperatures over water bodies ranged between 33.01 and 36.00°C during nighttime due to high thermal capacity. Nighttime values over high-density built-up areas were generally highest (34.82–36.41°C), followed by water bodies (33.85–36.15°C), commercial/industrial areas (32.70–34.99°C), then low-density built-up lands (31.89–34.65°C). Overall, the nighttime thermal gradient decreased from high-density built-up areas to fallow land, owing to the lower thermal capacity of this latter coverage.
In summary, the satellite images enabled detection of land cover characteristics such as informal settlements, fast-growing areas, open spaces, and green spaces. The metropolitan region of Delhi is developing very rapidly, mainly in areas that were previously predominantly agricultural lands. The study also clearly demonstrates a sharp increase in the amount of impervious land (e.g. paved surfaces), a decrease in open spaces, reductions in vegetation cover, a decrease in water bodies, and shrinkage of agricultural lands, all of which threaten to adversely affect the urban health of this capital city of India.
This type of data was previously scarce, but the combination of remote sensing and applied GIS has made it readily available. The data sets will be widely used by local government bodies, such as the Municipal Corporation of Delhi, the New Delhi Municipal Corporation, and the Delhi Development Authority, in preparing developmental plans and establishing new residential colonies.4
