Urban sprawl is also referred as irresponsible, and often poorly planned development that destroys green space, increases traffic, contributes to air pollution, leads to congestion with crowding and does not contribute significantly to revenue, a major concern. Increasingly, the impact of population growth on urban sprawl has become a topic of discussion and debate. Typically conditions in environmental systems with gross measures of urbanisation are correlated such as population density with built-up area (Smart Growth America, 2000; The Regionalist, 1997; Berry, 1990). The relation of population growth and urban sprawl is that the population growth is a key driver of urban sprawl. The population growth variable was responsible for about 31 percent of the growth in land area leading to suburban sprawl over the course of the 1980s in 282 metropolitan areas of United States of America. The population increased from 95 million to 140 million (47 percent) while urbanised land increased from 25,000 square miles to 51,000 square miles (107 percent) among 213 urbanised areas in USA between 1960 and 1990 (The Regionalist, 1997). This implied that density per square mile decreased by 28%.

The study on urban sprawl (The Regionalist, 1997; Sierra Club, 1998) was attempted in the developed countries (Batty et al., 1999; Torrens and Alberti, 2000; Barnes et al., 2001, Hurd et al., 2001; Epstein et al., 2002) and recently in developing countries such as China (Yeh and Li, 2001; Cheng and Masser, 2003) and India (Jothimani, 1997 and Lata et al., 2001).  In India alone currently 25.73% of the population (Census of India, 2001) live in the urban centres, while it is projected that in the next fifteen years about 33 % would be living in the urban centres. The spatial patterns of urban sprawl over different time periods, can be systematically mapped, monitored and accurately assessed from satellite data along with conventional ground data (Lata, et al., 2001). The physical expressions and patterns of sprawl on landscapes can be detected, mapped, and analysed using remote sensing and geographical information system (GIS) technologies (Barnes et al., 2001). The patterns of sprawl are being described using a variety of metrics, through visual interpretation techniques, all with the aid of software and other application programs. The earth scientists with the Northeast Applications of Useable Technology In Land Use Planning for Urban Sprawl (NAUTILUS) program are using techniques of statistical software to characterise urbanising landscapes over time and to calculate spatial indices that measure dimensions such as contagion, the patchiness of landscapes, fractal dimension, and patch shape complexity (Hurd et al., 2001; Nautilus 2001). Hurd et al, (2001) focused on a method to generate images depicting the pattern of forest fragmentation and urban development from the derived classifications of satellite imagery.

The built-up is generally considered as the parameter of quantifying urban sprawl (Torrens and Alberti, 2000; Barnes et al., 2001; Epstein et al., 2002). It is quantified by considering the impervious or the built-up as the key feature of sprawl, which is delineated using toposheets or through the data acquired remotely. Yeh and Li (2001) use Shannon's entropy, which reflects the concentration of dispersion of spatial variable in a specified area, to measure and differentiate types of sprawl. This measure is based on the notion that landscape entropy, or disorganisation, increases with sprawl. The urban land uses are viewed as interrupting and fragmenting previously homogenous rural landscapes, thereby increasing landscape disorganisation. Lata et al. (2001) employed a similar approach of characterising urban sprawl for Hyderabad City, India, in terms of Shannon's entropy. In an attempt to map the sprawling trends and changes in the urban core Jothimani (1997) used Landsat-MSS and IRS LISS-II data through visual interpretation techniques for analysis and identified the trends of emergence of sprawl along transportation network for Surat and Ahmedabad cities.

The impacts of urban patterns on ecosystem dynamics should focus on how patterns of urban development alter ecological conditions (e.g. species composition) through physical changes (e.g. patch structure) on an urban to rural gradient. The use of gradient analysis for studying urban-to-rural gradient of land-use intensity to explain the continuum of forest change from city centre to non-urban areas might help to explore ecosystem effects of different urban configurations, but current applications do not differentiate among alternative urban patterns (Alberti et al., 1999). Most studies of the impacts of urbanisation do not differentiate among various urban patterns. Planners need this ecological knowledge, so that their decisions can minimise impacts of inevitable urban growth. Decisions by urban dwellers, businesses, developers, and governments all influence patterns. Spatial pattern is one (of very few) such environmental variable, which can be controlled to some extent by land-use planning. Design strategies for reducing urban ecological impacts will remain poorly understood and ineffectual if spatial pattern issues are not addressed in ecological studies of urban areas.

The convergence of GIS, remote sensing and database management systems has helped in quantifying, monitoring, modeling and subsequently prediction of the process. At the landscape level, GIS aids in calculating the fragmentation, patchiness, porosity, patch density, interspersion and juxtaposition, relative richness, diversity, and dominance in order to characterise landscape properties in terms of structure, function, and change (ICIMOD, 1999; Civco et al., 2002). Modeling the spatial and temporal dimensions has been an intense subject of discussion and study for philosophy, mathematics, geography and cognitive science (Claramunt and Jiang, 2001). Mostly modeling of the spatial dynamics rests with the land cover / land use change studies (Lo and Yang, 2002) or urban growth studies. In order to predict the scenarios of land use change in the Ipswich watershed, USA over a period of two decades Pontius et al. (2000) predict the future land use changes in the Ipswich watershed based on the model validated for 1971, 1985 and 1991.

In urban growth modeling studies, the spatial phenomenon is simulated geometrically using techniques of cellular automata (CA). The CA technique is used extensively in the urban growth models (Clarke, et al., 1996) and in urban simulation (Torrens and Sullivan, 2001; Waddell, 2002). The inadequacy in some of these is that the models fail to interact with the causal factors driving the sprawl such as the population growth, availability of land and proximity to city centres and highway. Cheng and Masser (2003) report the spatial logistic regression technique used for analysing the urban growth pattern and subsequently model the same for a city in China. Their study also includes extensive exploratory data analyses considering the causal factors. The inadequacies in these were to accurately pinpoint spatially where the sprawl would occur. This problem could be effectively addressed when neural network is applied to the remote sensing data especially for classification and thematic representation (Foody, 2001). The neural spatial interaction models would relieve the model user of the need to specify exactly a model that includes all necessary terms to model the true spatial interaction function (Fischer, 2002).