DISCUSSION

Three modelling techniques considered show relatively good accuracy with validation dataset. But Geomod gives a conclusive output considering the better validation results and proposed city development plan. But Geomod has a capability of using only two land use classes. Comparatively using 4 land use classes LCM provides better visual interpretation since it uses agents to derive the growth. CA-Markov is dependent on neighborhood and rules and shows an exaggerated output of further aggregation of already urbanised city.

The predicted land use reveals of similar patterns of urbanisation of last decade. The main concentration will be mainly in the vicinity of arterial roads and proposed outer ring roads. Predicted land use also indicate of densification of urban utilities near the Bangalore international airport limited (BIAL) and surroundings. Further an exuberant increase in the urban paved surface growth due to IT Hubs in south east and north east. The results also indicated the growth of suburban towns such as Yelahanka, Hesaragatta, Hoskote and Attibele with urban intensification at the core area in almost all modelling techniques used. The results indicate that the urban area would cover close to 50 to 60 % of the total land use in and surrounding Bangalore. Thus providing insights to relevant information. Further modelling can be improved using nature and bio inspired techniques.