How pixel size affects a sediment connectivity index in central Belgium

Connectivity has become an increasingly used concept in hydrological and sediment research. In order to quantify it, various indices have been proposed since the start of the 21st century including the index of connectivity developed by Borselli et al. (2008). This index is based on a limited number of factors, the most important one being topography. Sediment connectivity indices values are likely to depend on the digital elevation model (DEM) resolution. The aim of this study was, first, to compare the effect of DEM pixel size (between 0.25 and 10 m, using an UAV) in the Belgian loess belt, a lowland area. We show that the index values were lower when the pixel size decreased (a difference of about 20 % in value between 0.25 and 10 m). In addition, the impact of linear features in the watershed (e.g., grass strip, bank and road) was lower with the largest pixel sizes, and the connectivity pattern was affected with a pixel size of 5 m or more. At lower pixel sizes (1 m or below), some more disconnected regions appeared. These corresponded to zones where there had been water stagnation during and after rainfalls, and was corroborated by field observations. This confirmed the need for a proper resolution according to the objectives of the study. The second aim of this study was to deduce a minimum pixel size for connectivity study, helping local erosion or sedimentation location and consequent land management decisions. In our context, 1 m stands as the optimum DEM resolution. This pixel size permitted to locate all “key areas” in terms of erosion. Very high resolutions (<0.5 m) did not generate much more information, and their calculation time was far greater.