Slope spectrum is a statistic model of slope distribution in a certain area. Previous researches display a potential importance of the slope spectrum in geomorphological studies. To quantitatively depict the slope spectrum, three indices (H, Td, S) were proposed, which can appropriately depict quantity features of slope distribution, but are difficult in depicting spatial structure of slope distribution. Hence, this paper suggests slope-landscape TUPU to quantitatively depict the spatial structure of slope distribution. The slope-landscape TUPU take each test area as an independent
landscape unit, and the slope class as patches constituting the landscape. So, theory and methodology of landscape ecology are applied to describe the spatial structure of slope distribution directly. Results show that the slope-landscape TUPU is capable of depicting spatial structure of slope spectrum. A continuous changes of the slope-landscape TUPU from south to north of the Loess Plateau shows an obvious spatial variation of surface roughness in the area, which is proved to be of great significance in describing the surface roughness. This research also suggests relationship between
digital terrain analysis and landscape ecology.