A study on the spatial scaling properties of topographic index for China

Bin Yong,Wanchang Zhang,Dengzhong Zhao
DOI: https://doi.org/10.1109/IGARSS.2007.4423865
2007-01-01
Abstract:Topographic Index (TI), ln(alpha/tanbeta) (alpha= upslope contributing area per unit contour; tanbeta= local slope angle), as an important parameter approximately characterizing the likely spatial distribution of soil water states within a watershed, has been widely applied in some topography-based land-surface schemes in recent decades. As TI concept and its computation was originally developed from hill slope hydrological process study and was most frequently utilized within small to middle sized watershed simulations, persistent disputation therefore has been concentrated on whether the TI computed from the coarse resolution DEM can be used to characterize the likely spatial distribution of moisture states in extensive areas effectively. The objective of this study was thus focused on investigating the statistical properties of TI computed from different resolution DEMs for clarifying the dynamic relations among them in whole territory of China. Statistical analyses were conducted to the TIs computed with an improved MFD algorithm from DEMs in different spatial resolution of 90 x 90 m, 1000 x 1000 m. Surprisingly, perfect linear relationships in almost the same form were found existed among the mean values of TI extensively sampled from these two different resolution DEMs under 0.1deg x 0.1deg, 0.5deg x 0.5deg and 1deg x 1deg block conditions respectively, which implies that accurate TI values can be obtained with coarse resolution DEMs usually employed in GCM or RCM simulations with an unique linear downscaling equation. Systematic analyses further point out that the algorithm for TI computation distinguishingly affects the form of this equation, while the size, location, and number of selected TI samples has minor affections.
What problem does this paper attempt to address?