Multi-scaled identification of landscape character types and areas in Lushan National Park and its fringes, China

Abstract

China’s national parks adopt a resource-oriented protection and planning approach that cannot restrain the continuous landscape fragmentation and deterioration, whereas, we propose to characterise the landscape in order to protect its integrity. This paper described a hierarchical identification of landscape character types and areas in Lushan National Park and its fringes according to a refined combination of the parametric and the holistic methods in a multiscalar approach. In terms of the functional hierarchy of landscape character, we decided to order the available data sources in ‘downscaling’. At the broad scale, landscape typologies were delimited by raster datasets of four natural attributes: land cover, soil, vegetation, and altitude. At the intermediate scale, landscape typologies were determined by raster datasets of six natural and cultural attributes: aspect, slope, relief amplitude, heritage density, geology and land use. At these two scales, we adopted the principal component analysis (PCA) and two-step cluster analysis in SPSS software to visualise landscape types, to modify and integrate the results obtained in the eCognition software, as well as to rectify the visualisation with manual identifications. At the detailed scale, landscape typologies were demarcated by two raster and one vector datasets of cultural attributes: building density, visual influence and time depth. We performed the visualisation and integration with a similar method except for the PCA step. This multi-scaled identification will provide a nested framework facilitating the integration of the broad Lushan region in both spatial and administrative dimensions.