Evolution of soil surface roughness and flowpath connectivity in overland flow experiments

During rainfall events, surface roughness affects runoff generation by providing water surface storage in the depressions and altering the flow direction on the surface. The process of runoff initiation, or triggering, involves the gradual filling of individual depressions and the connection of those overflowing depressions toward the outflow boundary. Although studies have been conducted to relate surface roughness to total depression storage, little work has been done in quantifying the roughness effects on runoff initiation. This study examines the role of surface roughness on overland flow triggering in interrill areas. Laboratory experiments were conducted on a 2.4-m x 2.4-m soil box exposed to a sequence of four rainfall events with treatments including two levels of initial roughness and two slope gradients. Surface microtopography was digitised by a laser scanner before and after each rainfall event. Soil roughness was analysed by the variogram method and its role in runoff triggering was evaluated using a numerical model that gradually fills depressions with a conditioned-walker method. Experimental variograms showed a gradual lowering of semivariances in a homothetic way after each additional rainfall, indicating that all roughness scales are affected similarly. All variograms showed a distinct topographic correlation length that can be related to depressional storage capacity. The linear relationship between these two variables also has a threshold roughness term below which the storage capacity tends to zero, Analyses of the runoff triggering showed that a small modification of micro-topographic structure had a major effect on runoff initiation. Even if the storage capacity is an important parameter of the runoff characteristics, large differences are observed between the initial stages of each experiment and final stages. We attributed these differences to the creation of preferential connections between topography depressions due to the material redistribution. Since the variogram analysis may not be sensitive toward the development of preferential flow path in a local scale, additional methodologies, such as the conditioned-walker analysis used in this study, need to be incorporated in quantifying the role of surface microtopography an the dynamics of runoff generation. (C) 2001 Elsevier Science B.V. All rights reserved.