Sprawl is perceived as one of the potential threats for development as the biggest confront is to ensure adequate housing with basic infrastructure and amenities including health, sanitation, etc. Currently, sprawl regions are devoid of any infrastructure, since planners are unable to visualize this type of growth patterns. This growth is normally not accounted in all government surveys (even in national population census), as these pockets are grouped either under urban or rural. The investigation of patterns of this kind of growth is very crucial from regional planning point of view to provide basic amenities (Ramachandra, et al., 2004). Prior visualizing of the trends and patterns of growth enable the planning machineries to plan for appropriate basic infrastructure facilities (water, electricity, sanitation, etc.). Also, by 2050 over 6 billion people, two thirds of humanity, will be living in towns and cities. This necessitates understanding the type, extent and nature of sprawl taking place in a region. In this regard, temporal remote sensing data aids in capturing the spatial growth patterns and process. Remote sensing coupled with geospatial analysis aid greatly in monitoring and management of the urbanisation process. The spectral pattern present within the remote sensing data for each pixel is used to perform the classification and, indeed, is used as the numerical basis for categorization of various spatial features (Lillesand, et al., 2002). Identifying, delineating and mapping urban areas on temporal scale provide an opportunity to monitor the changes, which is important for natural resource management and sustainable planning activities. Geospatial system helps in the integration of remotely sensed classified data and ancillary information from various other sources (population, natural resources, etc.) to arrive at decisions related to development of urban growth.

In this work, Landsat data of 1973 (of 79 m spatial resolution), 1992 and 2000 (30 m), IRS LISS-3 data of 1999 and 2006 (23.5 m) and MODIS data of 2002 and 2007 (with 250 m to 500 m spatial resolution) are used with supervised pattern classifiers based on maximum likelihood (ML) estimation followed by a Bayesian statistical approach. This technique quantifies the tradeoffs between various classification decisions using probability and costs that accompany such decisions (Duda, et al., 2000). It makes assumptions that the decision problem is posed in probabilistic terms, and that all of the relevant probability values are known with a number of design samples or training data collected from field that are particular representatives of the patterns to be classified. The mean and covariance are computed using maximum likelihood estimation with the best estimates that maximizes the probability of the pixels falling into one of the classes. Also, an attempt is made to map land surface temperatures across various land cover types to understand heat island effect. Urban Heat Island (UHI) studies have traditionally been conducted for isolated locations and with in situ measurements of air temperatures. The advent of satellite remote sensing technology has made it possible to study UHI both remotely and on continental or global scales (Streutker, 2002).