Lea Tien Tay

Modeling and Testing Landslide Hazard Using Decision Tree

This paper proposes a decision tree model for specifying the importance of 21 factors causing the landslides in a wide area of Penang Island, Malaysia. These factors are vegetation cover, distance from the fault line, slope angle, cross curvature, slope aspect, distance from road, geology, diagonal length, longitude curvature, rugosity, plan curvature, elevation, rain perception, soil texture, surface area, distance from drainage, roughness, land cover, general curvature, tangent curvature, and profile curvature. Decision tree models are used for prediction, classification, and factors importance and are usually represented by an easy to interpret tree like structure. Four models were created using Chi-square Automatic Interaction Detector (CHAID), Exhaustive CHAID, Classification and Regression Tree (CRT), and Quick-Unbiased-Efficient Statistical Tree (QUEST). Twenty-one factors were extracted using digital elevation models (DEMs) and then used as input variables for the models. A data set of 137570 samples was selected for each variable in the analysis, where 68786 samples represent landslides and 68786 samples represent no landslides. 10-fold cross-validation was employed for testing the models. The highest accuracy was achieved using Exhaustive CHAID (82.0%) compared to CHAID (81.9%), CRT (75.6%), and QUEST (74.0%) model. Across the four models, five factors were identified as most important factors which are slope angle, distance from drainage, surface area, slope aspect, and cross curvature.

Analysis of geophysical networks derived from multiscale digital elevation models: a morphological approach

We provide a simple and elegant framework based on morphological transformations to generate multiscale digital elevation models (DEMs) and to extract topologically significant multiscale geophysical networks. These terrain features at multiple scales are collectively useful in deriving scaling laws, which exhibit several significant terrain characteristics. We present results derived from a part of Cameron Highlands DEM.

Granulometric analyses of basin-wise DEMs: a comparative study

Digital elevation models (DEMs) are very useful for terrain characterization. We apply a morphological approach to characterize 14 sub‐basins decomposed from interferometrically generated DEMs of Cameron Highlands and Petaling regions of Peninsular Malaysia. Physiographically, these two regions possess a distinct geomorphologic set‐up as they belong to region with higher and lower altitudes, respectively. Fourteen sub‐basins are extracted from the DEMs, and pattern spectra by opening and closing of these sub‐basins relative to flat discrete binary patterns (square, octagon and rhombus) are computed. Pattern spectra are used to compute probability size distribution functions of both protrusions and intrusions that are conspicuous in topography, based on which shape‐size complexity measures of these sub‐basins are estimated by means of average roughness and size. Furthermore, fractal dimensions of channel networks derived from these 14 basins are computed by applying the box‐counting method. Comparisons between shape‐size complexity measures and fractal dimension are carried out.

Derivation of terrain roughness indicators via granulometries

Digital elevation models (DEMs) provide rich clues about various geophysical and geomorphologic processes. These clues include conspicuous protrusions and intrusions of foreground and background portions that testify the presence of channels and ridges in DEMs. We show an application of greyscale granulometries to characterize DEMs through shape–size complexity measures relative to symmetric rhombus, octagon and square templates. We first compute pattern spectra that measure the size distributions of protrusions and intrusions in a DEM. We then employ pattern spectra to compute probability size distribution functions of protrusions and intrusions relative to three templates. We finally compute shape–size complexity measures of DEM by employing these probability functions. To illustrate the implementation of granulometric approach to compute these measures of both background and foreground, we consider an interferometrically generated DEM of a part of Cameron Highlands of Malaysia. Hierarchical watersheds that could be decomposed from DEMs can be better classified via these measures.

Determination of Important Topographic Factors for Landslide Mapping Analysis Using MLP Network

Landslide is one of the natural disasters that occur in Malaysia. Topographic factors such as elevation, slope angle, slope aspect, general curvature, plan curvature, and profile curvature are considered as the main causes of landslides. In order to determine the dominant topographic factors in landslide mapping analysis, a study was conducted and presented in this paper. There are three main stages involved in this study. The first stage is the extraction of extra topographic factors. Previous landslide studies had identified mainly six topographic factors. Seven new additional factors have been proposed in this study. They are longitude curvature, tangential curvature, cross section curvature, surface area, diagonal line length, surface roughness, and rugosity. The second stage is the specification of the weight of each factor using two methods. The methods are multilayer perceptron (MLP) network classification accuracy and Zhous algorithm. At the third stage, the factors with higher weights were used to improve the MLP performance. Out of the thirteen factors, eight factors were considered as important factors, which are surface area, longitude curvature, diagonal length, slope angle, elevation, slope aspect, rugosity, and profile curvature. The classification accuracy of multilayer perceptron neural network has increased by 3% after the elimination of five less important factors.

Landslide Hazard Mapping Using a Poisson Distribution: A Case Study in Penang Island, Malaysia

Landsliding is one of the most destructive natural geohazards in Malaysia. Landslide hazard maps are very useful for urban development planning. This paper presents landslide hazard mapping using a new approach, i.e. a Poisson distribution, and compares the result with previous probabilistic approaches, i.e. frequency ratio (FR), statistical index (SI) and landslide nominal susceptibility factor (LNSF). These approaches were implemented in Penang Island to produce landslide hazard maps. The landslide causative factors considered are elevation, slope gradient, slope aspect, curvature, land cover, vegetation cover, distance from nearest road, distance from nearest stream, distance from nearest fault line, geology, soil texture and precipitation. Landslide hazard maps were assessed using the Receiver Operating Characteristics (ROC) method. Accuracy obtained for FR, SI and LNSF are 78.52 %, 78.12 % and 72.93 % respectively. Poisson distribution approach gives high accuracy of 78.51 % as FR.

Morphological convexity measures for terrestrial basins derived from digital elevation models

Abstract Geophysical basins of terrestrial surfaces have been quantitatively characterized through a host of indices such as topological quantities (e.g. channel bifurcation and length ratios), allometric scaling exponents (e.g. fractal dimensions), and other geomorphometric parameters (channel density, Hacks and Hurst exponents). Channel density, estimated by taking the ratio between the length of channel network ( L ) and the area of basin ( A ) in planar form, provides a quantitative index that has hitherto been related to various geomorphologically significant processes. This index, computed by taking the planar forms of channel network and its corresponding basin, is a kind of convexity measure in the two-dimensional case. Such a measure – estimated in general as a function of basin area and channel network length, where the important elevation values of the topological region within a basin and channel network are ignored – fails to capture the spatial variability between homotopic basins possessing different altitude-ranges. Two types of convexity measures that have potential to capture the terrain elevation variability are defined as the ratio of (i) length of channel network function and area of basin function and (ii) areas of basin and its convex hull functions. These two convexity measures are estimated in three data sets that include (a) synthetic basin functions, (b) fractal basin functions, and (c) realistic digital elevation models (DEMs) of two regions of peninsular Malaysia. It is proven that the proposed convexity measures are altitude-dependent and that they could capture the spatial variability across the homotopic basins of different altitudes. It is also demonstrated on terrestrial DEMs that these convexity measures possess relationships with other quantitative indexes such as fractal dimensions and complexity measures (roughness indexes).

Landslide Occurrence Prediction Using Trainable Cascade Forward Network and Multilayer Perceptron

Landslides are one of the dangerous natural phenomena that hinder the development in Penang Island, Malaysia. Therefore, finding the reliable method to predict the occurrence of landslides is still the research of interest. In this paper, two models of artificial neural network, namely, Multilayer Perceptron (MLP) and Cascade Forward Neural Network (CFNN), are introduced to predict the landslide hazard map of Penang Island. These two models were tested and compared using eleven machine learning algorithms, that is, Levenberg Marquardt, Broyden Fletcher Goldfarb, Resilient Back Propagation, Scaled Conjugate Gradient, Conjugate Gradient with Beale, Conjugate Gradient with Fletcher Reeves updates, Conjugate Gradient with Polakribiere updates, One Step Secant, Gradient Descent, Gradient Descent with Momentum and Adaptive Learning Rate, and Gradient Descent with Momentum algorithm. Often, the performance of the landslide prediction depends on the input factors beside the prediction method. In this research work, 14 input factors were used. The prediction accuracies of networks were verified using the Area under the Curve method for the Receiver Operating Characteristics. The results indicated that the best prediction accuracy of 82.89% was achieved using the CFNN network with the Levenberg Marquardt learning algorithm for the training data set and 81.62% for the testing data set.

Landslide hazard mapping of Penang Island using dominant factors

Landslide is one of the natural disasters in Malaysia and precipitation is the triggering factors for landslide in Malaysia. Besides rainfall factors, topographical factors also play key role in the susceptibility analysis of landslide. Since there are many available landslide-causative factors involved, selection of dominant factors is a crucial steps in landslide susceptibility analysis. This paper reports the landslide hazard mapping using Frequency Ratio (FR) approach with selected dominant factors in the area of Penang Island of Malaysia. Landslide hazard map of Penang Island is first generated by taking into account of twenty-two (22) landslide-causative factors. Among these twenty-two (22) factors, fourteen (14) factors are topographic factors. They are elevation, slope gradient, slope aspect, plan curvature, profile curvature, general curvature, tangential curvature, longitudinal curvature, cross section curvature, total curvature, diagonal length, surface area, surface roughness and rugosity. The other eight (8) non-topographic factors considered are land cover, vegetation cover, distance from road, distance from stream, distance from fault line, geology, soil texture and rainfall precipitation. After considering all twenty-two factors for landslide hazard mapping, the analysis is repeated by removing one factor at one time to identify the dominant landslide-causative factors. Twelve dominant factors are selected from the twenty-two factors. Landslide hazard map was segregated into four categories of risks, i.e. Highly hazardous area, Hazardous area, Moderately hazardous area and Not hazardous area. The maps was assessed using ROC (Receiver Operating Characteristic) based on the area under the curve method (AUC). Landslide hazard map produced by including all 22 factors has an accuracy of 77.76%. By removing 10 irrelevant factors and employing only 12 dominant factors, the generated hazard map achieves better performance with accuracy of 79.14%.

Some New Findings on Gauss–Seidel Technique for Load Flow Analysis

In this paper, research is conducted using conventional Gauss–Seidel (GS), GS with acceleration factor (AF), first iteration of Newton–Raphson (NR) and followed by Gauss–Seidel (with and without acceleration factor) for load flow (LF) solution on three IEEE test systems, i.e. IEEE-30 bus system, IEEE-57 bus system and IEEE-118 bus system. Besides, GS with and without acceleration factor were also tested on different loadings of these IEEE bus systems to ascertain the robustness of the acceleration factors. The results are compared in terms of number of iterations and computation time taken for convergence.