Digital mapping of surface sealing ratios using landsat 5 TM satelite images: an example of space-time analysis in Almada council, Portugal.

Abstract

The objective of the present work is to present the results of an expedite methodology, in Geographical Information Systems (GIS), designed to locally estimate the impervious surface ratio for several dates where satellite thematic images and complete and detailed land use cartography are available. Firstly, in the proposed methodology, superficial sealed areas identified by vector land use cartography and the land use obtained by a supervised classification of a homologous thematic image (training image) are correlated making use of a regular grid of square polygons with several sizes multiples of the satellite image resolution. Further on, and after choosing the grid size that better represents the compromise between correlation and spatial resolution, it is performed a linear regression of the superficial sealed ratio for the remaining images posterior or ulterior of the training image. This methodology is illustrated with a case study developed within the area of Almada council, Setúbal Peninsula, Portugal. Results show changes in land use between 1987 and 2001, highlighting the dramatic increase of impervious surface ration due to growing settlements and transport infrastructure. Keywords: remote sensing; Landsat 5 TM; supervised classification; surface sealing ratio; space-time analysis.