¹Energy & Wetlands Research Group, Centre for Ecological Sciences, Indian Institute of Science, Bangalore. cestvr@ces.iisc.ernet.in
² Research Associate at Indian Institute of Science (IISc). bharath@ces.iisc.ernet.in
³research scholar at Energy and Wetlands Research Group (EWRG), Centre for Ecological Sciences (CES), Indian Institute of Science (IISc). vinay@ces.iisc.ernet.in
⁴Faculty at Centre for Ecological Sciences (CES), Indian Institute of Science (IISc), Bangalore
⁵currently Post-doctoral fellow at NASA AMES Research Centre, USA
⁶Group Head, Urban Studies & Geoinformatics, National Remote Sensing Centre (NRSC), Indian

Introduction

Urbanisation gained momentum with the globalization and consequent market relaxations in developing nations. Spatial patterns of urbanisation is the most studied issue in the current times across the globe [1-2]. Rapid urbanization and urban sprawl are associated with the developments that are complex, unsustainable and unstable, leading to increasing social, economic and environmental related problems [3-4]. Space Research Organisation, Hyderabad, India. venu_koppaka@nrsc.gov.in

Rapid urbanization in India, at the metropolitans since early 2000’s has captured the global attention due to the availability of resources (human or land) for investors and industries, social, political and cultural aspects. Favorable environment gave impetus to the uncontrolled and unprecedented growth in the core and fringes [5]. The process of urban sprawl due its uncontrolled dynamics has a devastating effect on the natural resources [6] such as vegetation, quality and quantity of surface and sub-surface water resources [7], climatic factors [8] such as rainfall, temperature [9], etc.  In order to monitor the dynamics of urbanization and other associated land use changes remote sensing [10] technology is being used since it has the ability to cover large areas, cost effective and faster compared to ground based surveying techniques. The remotely sensed data is classified into different land uses based on different classification techniques to understand the distribution of heterogeneous features [11] Gaussian maximum likelihood classifier algorithm was used to classify land uses from the remote sensing data. Multi temporal land use analysis helps in understanding the factors that have caused the change, potentially allowing for management strategies targeted toward cause rather than simply the symptoms of the cause [12-13]. The temporal information helps to measure, quantify and monitor changes, which are necessary to model and simulate the likely changes in spatial patterns [14]. Different modeling techniques have been used in order to predict the land use changes, such as simple rule based models or agent based models, which includes Cellular Automata, CA-Markov, Geomod, AHP-CA-Markov, LCM, MCE, Regression, Bayesian [15-16], etc.

Cellular Automata Markov Chain integrated model is a rule based model where in multiple rules are used in order to simulate the future scenario though historical data sets. The Markov chains (MC) are stochastic non-spatial processes [17] that shows how likely does one state change to another, and is used to develop transition probability matrix controls temporal change among the land use types [18] using the two map sets i.e., current map and historical map set [19] to define the site suitability’s of the land use. The cellular automata (CA) process introduced by Von Neumann and Ulam [20] is a systematic process consisting sequence of cells carrying 0 or 1 arranged in a matrix, the value of each cell evolves deterministically based on the rules and involving the neighboring pixels [20]. Cellular Automata algorithms define the state of the cell based on the previous state of the cells within a neighborhood, using a set of transition rules. Coupled MC and CA eliminates the short comings of CA and MC respectively, MC quantifies future changes based on past changes, thereby serving as a constraint for CA, which addresses spatial allocation and the location of change [21], although CA-Markov gives promising results, it fails to achieve accurate results since the driving forces are not accounted in this model [22].
Analytical Hierarchical Process based modeling on the other hand is an agent based modeling which accounts the agents of change those include drivers or constraints, socio-economical and infrastructural activities, human actions [23] in order to model and simulate the land use dynamics, providing ample opportunities and challenges which complement or extend to other approaches [24]. ABM’s weigh/rank the growth factors and constraints as reflected by the real world scenarios [23] to develop site suitability maps in order to model the land use. The site suitability maps provide the transitional areas describing where the particular land use has the probability to change or retain its state. The site suitability maps area combined with the CA-Markov in order to simulate and predict the land use dynamics. ABM emerges as a promising approach in understanding the complex urban processes, such as urban land use changes since it accounts the social, infrastructural and other human based aspects [23]. Objective of current analysis is to model urban growth pattern of 4 major cities of India using Agent Based Modeling.