Craig Onafrychuk

Topographically driven soil saturation has been a key area of scientific inquiry in geographical and hydrological analysis and modelling. Current approaches in computing the secondary DEM attribute, the topographic wetness index (TWI), as an indicator of surface saturation, tend to rely on digital elevations models (DEMs) computed with inverse distance weighted (IDW) algorithms and without hydrological correction or stream enforcement such as that available in the ANUDEM procedure. With these approaches, it is common for the TWI to be calculated using the flow direction model of 0' Callaghan and Mark (1984), which has two main restrictions: (i) flow originating over a two-dimensional pixel is treated as a point source (non-dimensional) and is projected downslope by a line (one dimensional) (Moore et aI., 1991), and (ii) the flow direction in each pixel is restricted to eight possibilities (Costa-Cabral and Burges, 1994; Tarboton, 1997). This paper presents a cumulative approach to wetness index modelling emphasizing the interpolation of a hydrologically correct DEM and the use of a distributed flow direction algorithm to improve the overall distribution of the TWI. It was found that the TWI provided a good predictive classification of surface water distribution including areal features such as wetlands and open aquatic areas (i.e., lakes and ponds) as well as linear features including ephemeral and permanent streams. With this cumulative approach, the secondary DEM attribute, W = in (As / tan j3) can be used as an automated means of mapping the distribution of surface water and what is described as a surface's topographical hydrologic connectivity index (THCI). The index is proposed as a tool for site restoration planning and watershed conservation management.