EP53B-1011
Scaling Characteristics of Rill Networks
Abstract:
Sediment transport in overland flow interacts dynamically with the soil surface morphology. Often, sediment transport models assume a simplified and static morphology. This assumption, although it limits the predictive capacity of models, is reasonable since the evolution of morphology is difficult to quantify, particularly when rill networks form. Such networks evolve due to local features of the surface, which are difficult to identify even in well controlled laboratory experiments. Instead of attempting to predict details of rill networks, we hypothesize that their statistical properties can (i) be measured reliably and (ii) that under reasonable background conditions they exhibit scale invariance in space.We report initial results of laboratory experiments to test these hypotheses. An agricultural soil was placed in a 5 m × 2 m flume with a 5% slope to which a uniform rainfall was applied. Prior to the rainfall, the top 10 cm of the soil was ploughed and smoothed. Rill networks are generated in three 3-h experiments using different precipitation rates of 30, 45 and 60 mm h-1. The surface morphology was measured using a laser scanning every 30 min (rainfall was halted to permit scanning). For the measured Digital Elevation Models, the exceedance probabilities and corresponding scaling factors for the drainage area, upstream length of the network were calculated.
The results showed that, similar to river networks, there is a power law relation in the exceedance probabilities for the parts of the network in which rill erosion is dominant. However, contrary to large scale river networks, the scaling exponents were found to be dependent on rainfall intensity.