Abstract:

After the declaration of Ankara as the capital city of Turkey in 1923, the size of the city was identified to be insufficient to cope with the developmental and spatial needs of the city. In this study, the analysis and detection of land cover changes were conducted for the last three decades with ten-year time interval by using remotely sensed satellite data in Ankara to monitor the change in land cover, and growth and development of the city. Four classes; manmade area, land area, green area, and water area were created for each year images to assess change in land cover in central neighborhoods of Ankara. Maximum Likelihood Classifier (MLC) and Random Forest (RF) algorithms were performed and classification results were compared. Overall classification accuracy and overall kappa statistics computed as 85%-92% and between 0.78-0.87 for MLC algorithm, respectively. Comparing with MLC algorithm, RF algorithm’s performance was unsatisfied. As a second step of this study, administrative data of Ankara such as population, land use types, number of buildings and flats, and spatial development relationships were analyzed in integration with remote sensing data results to analyses land development in Ankara.

Keywords: