Measuring spatio-temporal dynamics of impervious surface in Guangzhou, China, from 1988 to 2015, using time-series Landsat imagery

Abstract

This study evaluated the spatio-temporal change characteristics of urban development at different scales with time-series impervious surface fractions. Landsat-5 Thematic Mapper (TM) and Landsat-8 Operational Land Imager (OLI) images were used to extract impervious surface fractions using a modified linear spectral mixture analysis method in Guangzhou from 1988 to 2015. The results indicated that the impervious surface area has substantially increased, from 70.3 km2 in 1988 to 580.5 km2 in 2015. In 2015, the impervious surfaces were distributed almost throughout the whole region of the study area, except in the forest region. Next, impervious surface weighted mean centre (ISWMC) and the standard deviational ellipse (SDE) methods were used to systematically analyse the principle orientation, direction, spatio-temporal expansion trends, and the distribution differences of impervious surfaces at the whole and local region scales from 1988 to 2015. The spatio-temporal dynamics of ISWMC exhibited different expansion directions and intensities of impervious surfaces at the whole and local region scales. On a whole region scale, the principle expansion direction of impervious surfaces was northward. However, the expansion trend of impervious surfaces in the different districts was significantly different from other trends at the local region scale. The parameters of SDE were used to investigate the orientation and the clustering or dispersion degree of impervious surface at different scales. The results from SDE analysis indicated that the impervious surfaces exhibited uncertainty in the expansion direction at the whole region scale; in contrast, they had a distinct preferred orientation and expansion direction at the local region scale. The analysis revealed that urban expansion exhibited different change characteristics in various directions at the local region scale. In summary, the results at the local region scale can better reflect the change trajectory of spatio-temporal dynamics of urban development and its fine spatial structure than at the whole region scale.