Data	Year	Purpose
Landsat Series Multispectral sensor (57.5m)	1975	Land cover and Land use analysis
Landsat Series Thematic mapper (28.5m)	1990, 2000, 2010
Survey of India (SOI) toposheets of 1:50000 and 1:250000 scales		To generate boundary and base layer maps.

Figure 2: Procedure outline

Pre-processing:

The remote sensing data obtained were geo-referenced, geo-corrected and cropped pertaining to the study area – Ahmedabad administrative area with 10 km buffer. Landsat satellite data were resampled to 30 m in order to maintain uniformity in spatial resolution across all temporal data.

Land Cover analysis:

Land cover analysis was performed to understand the changes in the vegetation cover during the study period. Normalised Difference Vegetation Index (NDVI) was computed for assessing the extent of vegetation cover. NDVI values range from -1 to +1. Very low values of NDVI (-0.1 and below) correspond to soil. Zero indicates the water cover. Moderate values represent low density vegetation (0.1 to 0.3), while high values indicate thick canopy vegetation (0.6 to 0.8).

Land use analysis:

False colour composite (FCC) of remote sensing data (bands – green, red and NIR), was generated to visualise the heterogeneous patches in the landscape. Training data were digitized and then loaded to pre-calibrated GPS. The training data were chosen so as to cover at least 10% of the study region and uniformly distributed throughout the study region. Field investigations were carried out to collect the attribute information of these training polygons. The signatures were also digitized as polygons with the help of Google Earth. Land use categories (in table 2) were classified using supervised classifier based on Gaussian Maximum Likelihood Classifier (GMLC) algorithm with the help of training data (60% data were used for classification).

 

Table 2: Land use categories

 

Land use class Land uses included in the class Urban This category includes buildings and all paved surfaces and also mixed pixels having built up. Water bodies Tanks, lakes, reservoirs, canals. Vegetation Forest, cropland, nurseries. Others Rocks, quarry pits, open ground, un-metaled roads. Table 2: Land use categories GMLC is considered as one of the superior classifiers as it uses various classification decisions using probability and cost functions (Duda et al., 2000). Land use was computed through open source program GRASS - Geographic Resource Analysis Support System (http://wgbis.ces.iisc.ernet.in/grass). Out of the total generated signatures 60% were used in classification and balance, 40% were used for validation and accuracy assessment of the classified data. Classes of the resulting image were recoded to form four land-use classes. Accuracy assessment to evaluate the performance of classifier (Ramachandra et al., 2012a) was done through a confusion matrix and comparison of kappa coefficients (Congalton et al., 1983; 1991; 2009).

Zonal Analysis:

As most of the definitions of a city or its growth are defined in directions, hence it was considered appropriate to divide the study region into four zones [as Northeast (NE), Southwest (SW), Northwest (NW) and Southeast (SE)] based on directions considering the central pixel (Central Business district).

Gradient Analysis:

Each zone was divided into concentric circles of incrementing radius of 1 km radius from the centre of the city. This analysis helps in visualising the urbanisation process at local levels and understanding the agents responsible for changes. 3.7 Shannon’s Entropy (Hn): Shannon’s entropy given in equation 1 was computed for each zone to understand the growth of the urban area in specific zones and to understand if the urban area is compact or divergent. Shannon’s entropy explains the urban process by characterising the growth either as concentrated / aggregated or sprawl (Sudhira et al., 2004; Ramachandra et al., 2012c). ….. (1) Where Pi is the proportion of the built-up in the ith concentric circle and Hn ranges from 0 to log n. If the distribution is maximally concentrated, the lowest value zero will be obtained. Conversely, if it is an even distribution among the concentric circles, then the value will be maximum of log n.

3.8 Computation of spatial metrics: Spatial metrics are helpful to quantify spatial characteristics of urbanising landscape. FRAGSTATS (McGarigal and Marks,1995) was used to compute metrics (details of the metrics are in Bharath et al., 2012b), which include Number of patches (Built-up) (NP), Patch Density (PD), Largest patch Index (Built-up) (LPI), Normalised landscape shape Index (NLSI), Area-Weighted Mean Shape Index (AWLSI), Edge Density (ED), Clumpiness Index (CLUMPY) and Aggregation Index (AI).