Data and Method

Temporal remote sensing data of Landsat TM and ETM+ downloaded from GLCF were preprocessed to correct geometrical and radiometrical accuracy USGS (http://www.usgs.gov). This was further used to analyse and model LULC changes. Remote sensing data were supplemented with the Survey of India topographic maps (of 1:50000 and 1:250000 scale), which were used to generate base layers of the administrative boundary, drainage network, Road network etc. Slope map was extracted using ASTER data (30 m) downloaded from USGS (www.usgs.gov). Ground control points (GCPs) and training data were collected using pre calibrated Global Positioning System (GPS) and virtual online spatial maps such as Bhuvan and Google Earth. GCPs were useful in geometric correction of remote sensing data. Census data (1991, 2001 and 2011) was used to capture population dynamics.

Modelling of urbanization and sprawl involved: i)  Remote Sensing data acquisition, geometric correction, field data collection, ii) Classification of remote sensing data and accuracy assessment using GRASS, iii) Land use analysis, iv) Identification of agents  and development of attribute information, v) Prediction: Designing scenarios of  urban growth and calibrating the model to find out the best weights based on the influence on the neighborhood pixels vi) Accuracy assessment and validation of the model, vii) Prediction of future growth based on validated data.

Land use analysis was carried out using supervised pattern classifier - Gaussian maximum likelihood algorithm based on probability and cost functions (Duda et al., 2000). Land use with gradient analysis results were further used in Modelling. These results can be accessed in previous working literatures (Ramachandra et al., 2014a, 2014b, 2014c, 2014d, Chandan et al., 2014). These data was used in Modelling and visualizing the growth of these cities.

Using the combination of Fuzzy Logic, Analytical Hierarchical Process (AHP), Multi Criteria Evaluation (MCE), Markov chains and Cellular Automata (CA). Agents of urbanisation such as roads, industries, educational institutions, bus stands, railway stations, metro, population, etc. were normalized. Conservation regions as per city development plan (CDP) water bodies were considered as constraints. The fuzzy based analysis is used to normalize the contributing factors between 0 and 255, where 255 showing the maximum probability of change and 0 indicating no change, for different land uses. The normalized agents were taken as input to AHP to determine the weights of driving factors using pair wise comparisons ith weights as Eigen vectors. The weights analysed and calibrated through AHP is verified using measured consistency ratio (CR). CR below 0.1, the model is consistent and used for subsequent processes.
These weights along with the factors of growth are combined along with the constraints to obtain site suitability maps for different land uses using equation below
LC =                  ……. (1)
Where LC is the linear combination of weights, n is the number of factors, D decision factor, W is the weight of the factor.
The Markov chains are used to determine the change probability between two historical datasets to derive the growth in the future scenarios based on different criteria’s. The Markovian transition matrix indicates the probability of the particular land use being converted to other land uses on single time step.

Table 1: Criteria’s for simulating and predicting urban sprawl

The cellular automata based on the site suitability and the transition matrix is used to spatially predict the changes in land use based on current land use at every single time step, based on the neighbouring pixels. Two scenarios were designed to predict the land use changes as shown in table 1.
Validation of the simulated datasets of were performed with classified datasets through kappa indices, as a measure of agreement. Once these data and agents are trained and validated, data is used to model and simulate for the year 2030 (ten years) with definite time steps.

Geo-visualisation of urbanisation of five tier I cities are depicted in fig.1 to fig.5 and results are provided in tables 2 to 6 respectively. The cities on an average would grow by 1.5 to over 2 times the current state in next decade. By 2025, it is predicted that built up area in these cities and surroundings, grows over 57% (Delhi), 27% (Mumbai), 45.8% (Chennai), 50% (Pune) and 37% (Coimbatore) respectively. The various drivers of growth for different cities are as in annexure 1. In all these cases, spatially it could be understood that the CDP if implemented properly would play a major role in curtailing the unsustainable growth of the city in its limits, while some growth still takes place at the outskirts. Prime factors of growth include the transportation network, industrialisation, and educational sector.