Predicting gully initiation: comparing data mining techniques,analytical hierarchy processes and the topographic threshold

Predicting gully initiation at catchment scale was done previously by integrating a geographical information system (GIS) with physically based models, statistical procedures or with knowledge‐based expert systems. However, the reliability and validity of applying these procedures are still questionable. In this work, a data mining (DM) procedure based on decision trees was applied to identify areas of gully initiation risk. Performance was compared with the analytic hierarchy process (AHP) expert system and with the commonly used topographic threshold (TT) technique. A spatial database was used to test the models, composed of a target variable (presence or absence of initial points) and ten independent environmental, climatic and human‐induced variables. The following findings emerged: using the same input layers, DM provided better predictive ability of gully initiation points than the application of both AHP and TT. The main difference between DM and TT was the very high overestimation inherent in TT. In addition, the minimum slope observed for soil detachment was 2°, whereas in other studies it is 3°. This could be explained by soil resistance, which is substantially lower in agricultural fields, while most studies test unploughed soil. Finally, rainfall intensity events >62.2 mm h^‐1 (for a period of 30 min) were found to have a significant effect on gully initiation. Copyright © 2012 John Wiley & Sons, Ltd.     

On the dynamics and morphology of extensive tidal mudflats: Integrating remote sensing data with an inundation model of Cook Inlet, Alaska

A new method of integrating satellite remote sensing data and inundation models allows the mapping of extensive tidal mudflats in a sub-Arctic estuary, Cook Inlet (CI), Alaska. The rapid movement of the shorelines in CI due to the large tides (~10 m range) is detected from a series of Landsat imagery taken at different tidal stages, whereas GIS tools are used to identify the water coverage in each satellite image and to extract the coordinates of the shoreline. Then, water level along the shoreline for each satellite image is calculated from the observed water level at Anchorage and the statistics of an inundation model. Several applications of the analysis are demonstrated: 1. studying the dynamics of a tidal bore and the flood/ebb processes, 2. identifying climatic changes in mudflats morphology, and 3. mapping previously unobserved mudflat topographies in order to improve inundation models. The method can be used in other regions to evaluate models and improve predictions of catastrophic floods such as those associated with hurricane storm surges and tsunamis.     

Effects of vegetation on channel morphodynamics: results and insights from laboratory experiments

A series of laboratory experiments demonstrates that riparian vegetation can cause a braided channel to self‐organize to, and maintain, a dynamic, single‐thread channel. The initial condition for the experiments was steady‐state braiding in non‐cohesive sand under uniform discharge. From here, an experiment consisted of repeated cycles alternating a short duration high flow with a long duration low flow, and uniform dispersal of alfalfa seeds over the bed at the end of each high flow. Plants established on freshly deposited bars and areas of braidplain that were unoccupied during low flow. The presence of the plants had the effect of progressively focusing the high flow so that a single dominant channel developed. The single‐thread channel self‐adjusted to carry the high flow. Vegetation also slowed the rate of bank erosion. Matching of deposition along the point bar with erosion along the outer bend enabled the channel to develop sinuosity and migrate laterally while suppressing channel splitting and the creation of new channel width. The experimental channels spontaneously reproduced many of the mechanisms by which natural meandering channels migrate and maintain a single dominant channel, in particular bend growth and channel cutoff. In contrast with the braided system, where channel switching is a nearly continuous process, vegetation maintained a coherent channel until wholesale diversion of flow via cutoff and/or avulsion occurred, by which point the previous channel tended to be highly unfavorable for flow. Thus vegetation discouraged the coexistence of multiple channels. Varying discharge was key to allowing expression of feedbacks between the plants and the flow and promoting the transition from braiding to a single‐thread channel that was then dynamically maintained. Copyright © 2010 John Wiley & Sons, Ltd.     

The KC method: Numerical investigation of a new analysis method for reinforced soil walls

Reinforced soil walls commonly include facing elements which affect the mechanical behavior of the system. However, the design procedures involved in the existing codes and manuals (e.g. FHWA, BS8006, AASHTO, etc.) do not consider the structural contribution of the facing to the wall stability. Recently, a new computer based method for the analysis of reinforced soil walls which takes into account the interaction between the facing and the soil reinforcement layers was presented [Klar A and Sas T. Rational approach for the analysis of segmental reinforced soil walls based on kinematic constraints. Geotextiles and Geomembranes 2009;27:332-340]. This method demands full compatibility between the reinforcement layers and the deforming wall, and is solved as an optimization problem on this constraint. This kinematic compatibility (KC) method entails several assumptions regarding the interaction between the three components of the system (soil, wall, and reinforcement). This paper compares the KC method to a more rigorous continuum analysis. Results show that the KC method is capable of replicating the behavior of the more rigorous system, with a good agreement on both the value of maximum tensile forces in the reinforcement and shear and bending moment distributions along the wall. The KC method has a certain advantage over continuum methods, such as finite element or finite difference, since it requires limited input data that can easily be obtained from field tests.