The urban growth management initiative: Confronting the expected doubling of the size of cities in the developing countries in the next thirty years - Methods and preliminary results

Procedures are overviewed and preliminary results are presented for a project addressing urban growth mapping for a sample of 120 metropolitan areas from a universe of 2,719 cities having populations in excess of 100,000 in the year 2000. A modified unsupervised classification approach applied to Landsat imagery serves as the basis for urban cover mapping. At the time of the preparation of this paper, work had been completed for 54 cities, and another 37 were in various stages of completion. The average annual urban growth for a sample of 31 metropolitan areas was 5.1 percent, based on an average temporal difference between T ₁ and T ₂. Accuracy for a sample of seven of the classifications for the completed metropolitan areas yielded an overall mapping accuracy of 88.3 percent for both for T ₁ and T ₂. It is expected that classification and accuracy assessment for T ₁ and T ₂ Landsat data for all cities will be completed by Spring 2005. The urban classification maps developed in this project, as well as summary data, will be made available to the scientific community via the Internet.