Predicting Ephemeral Gully Location and Length Using Topographic Index Models

Abstract

ABSTRACT. Ephemeral gullies (EGs

Evaluating functional connectivity in a small agricultural catchment under contrasting flood events by using UAV

“…The absence of elements interrupting the connectivity of the water and sediment flows favoured the good performance of the predictions, as illustrated in the photograph in Figure 5-I. However, as pointed out by Parker (2007) and Daggupati et al (2013), anthropic features acting as barriers such as (physical) boundaries, dry-stone walls, tracks and roads or channel levees affected IC predictions. In some cases, these elements impeded sediment transfer, but not the hydrological connectivity and the erosive capacity of water flows (e.g.…”

“…Functional connectivity in Can Revull catchment and the performance of the IC in representing it were assessed by the method of Daggupati et al (2013), which comparesin terms of occurrence and lengththe evidence of erosion (e.g. EG), with the potential trajectories delineated by establishing various critical thresholds in morphometric indices.…”

“…To evaluate the ability of the IC to predict the activation of different processes during events with contrasting characteristics, the kappa index, κ (Congalton, 1991) was applied to the error matrix, as an occurrence statistic that measures the relationship between beyond-chance agreement and expected disagreement (Daggupati et al, 2013), defined as:…”

“…This is why predicting the location of erosion trajectories by applying topographic index models from the combination of primary topographic attributes, in order to calculate an index value at each point (or pixel) of a catchment, is one of the main research insights of catchment geomorphology (e.g. Moore et al, 1988;Vandekerckhove et al, 1998;Poesen et al, 2003;Parker et al, 2007;Daggupati et al, 2013;Momm et al, 2013). Based on these models, the aligned points where the index exceeds a certain threshold value are assumed to define potential trajectories for the development of concentrated erosion (Montgomery and Dietrich, 1994).…”

“…These results indicated that a greater number of parameters adds more uncertainties to the model, which does not necessarily result in better predictive accuracy. In addition, the performance of different topographic index models can be significantly affected by the quality of the input datasets, depending on the resolution of digital terrain models (DTMs), the inclusion/exclusion of land-use factors or the choice of thresholds (Daggupati et al, 2013). For each specific index, a critical threshold value appropriate for a given catchment may not be useful for another catchment with different characteristics (area, slope, lithology or land uses).…”