Omar F. Althuwaynee

Application of an evidential belief function model in landslide susceptibility mapping

The objective of this paper is to exploit the potential application of an evidential belief function model to landslide susceptibility mapping at Kuala Lumpur city and surrounding areas using geographic information system (GIS). At first, a landslide inventory map was prepared using aerial photographs, high resolution satellite images and field survey. A total 220 landslides were mapped and an inventory map was prepared. Then the landslide inventory was randomly split into a testing dataset 70% (153 landslides) and remaining 30% (67 landslides) data was used for validation purpose. Fourteen landslide conditioning factors such as slope, aspect, curvature, altitude, surface roughness, lithology, distance from faults, ndvi (normalized difference vegetation index), land cover, distance from drainage, distance from road, spi (stream power index), soil type, precipitation, were used as thematic layers in the analysis. The Dempster-Shafer theory of evidence model was applied to prepare the landslide susceptibility maps. The validation of the resultant susceptibility maps were performed using receiver operating characteristics (ROC) and area under the curve (AUC). The validation results show that the area under the curve for the evidential belief function (the belief map) model is 0.82 (82%) with prediction accuracy 0.75 (75%). The results of this study indicated that the EBF model can be effectively used in preparation of landslide susceptibility maps.

A novel integrated model for assessing landslide susceptibility mapping using CHAID and AHP pair-wise comparison

ABSTRACT This article uses an integrated methodology based on a chi-squared automatic interaction detection (CHAID) model combined with analytic hierarchy process (AHP) for pair-wise comparison to assess medium-scale landslide susceptibility in a catchment in the Inje region of South Korea. An inventory of 3596 landslide locations was collected using remote sensing, and a random sample comprising 30% of these was used to validate the model. The remaining portion (70%) was processed by the nearest-neighbour index (NNI) technique and used for extracting the cluster patterns at each location. These data were used for model training purposes. Ten landslide-conditioning factors (independent variables) representing four main domains, namely (1) topology, (2) geology, (3) hydrology, and (4) land cover, were used to produce two landslide-susceptibility maps. The first landslide-susceptibility map (LSM1) was produced by overlaying the terminal nodes of the CHAID result tree. The second landslide-susceptibility map (LSM2) was produced using the overlay result of AHP pair-wise comparisons of CHAID terminal nodes. The prediction rate curve results were better with LSM2 (area under the prediction curve (AUC) = 0.80) than with LSM1 (AUC = 0.76). The results confirmed that the integrated hybrid model has superior prediction performance and reliability, and it is recommended for future use in medium-scale landslide-susceptibility mapping.

Semi-quantitative landslide risk assessment using GIS-based exposure analysis in Kuala Lumpur City

ABSTRACT A semi-quantitative landslide-risk assessment method, which would provide a spatial estimate of future landslide risks in a densely populated area in Kuala Lumpur City, was presented in this study. This work focused on detail risk assessment by identifying the number of elements at risk. A medium-scale analysis was performed using geospatial based techniques. The estimation of rainfall threshold and the landslide hazard map used in the current work are obtained from the previous literature published by the same authors. Subsequently, the vulnerability value was generalized, and then a valid integration between elements at risk and the hazard map was conducted to determine the expected number of elements that would likely be under direct risk. Results showed that the approximate number of predicted affected elements per pixel, as a percentage of the settlement unit, is nearly 50% in residential areas, 35% in commercial buildings, 31% in industrial buildings, 31% in utility areas, and 18% in densely populated areas. Similarly, a significant percentage of predicted losses (27%) were found for the road network. The results showed the capability of the method to approximately predict the number of infrastructure elements and the population density under landslide risk in data-scarce environments.

An Alternative Technique for Landslide Inventory Modeling Based on Spatial Pattern Characterization

The present study analyses the spatial patterns of historical/present landslide inventory in the Kuala Lumpur and vicinity areas. The main objective is to statistically test the spatial nature pattern of landslide inventory, i.e. to determine whether it rejects the independency of spatial pattern or not (i.e. random or cluster distribution). For that purpose, the nearest neighbor index (NNI) was applied to measure and test the randomness. First, we tested the spatial patterns of 153 landslides. The results showed a percentage of clustered to dispersed was 85 % (130 events) to 15 % (23 events), indicating landslides have a cluster pattern tendency. Then, the spatial relationship between the cluster landslides and conditioning factors were analyzed using evidential belief function (EBF) model. Additionally, the susceptible map produced by an earlier study was used to compare the results of the inventory selection. Finally, two landslide susceptible maps (LSMs) were validated by using prediction rate curve techniques. Prediction accuracy of the cluster data LSM2 was 0.80 (80 %), whereas the random data produced LSM1 showed 0.75 (75 %) prediction accuracy. From the results obtained in this study, one can infer that the spatial nature pattern of landslide inventory follows a cluster patterns. Secondly, clustered data can be used as training data instead of random selection technique. As a conclusion, the same technique can be replicated elsewhere.

Landslide susceptibility mapping using decision-tree based CHi-squared automatic interaction detection (CHAID) and Logistic regression (LR) integration

This article uses methodology based on chi-squared automatic interaction detection (CHAID), as a multivariate method that has an automatic classification capacity to analyse large numbers of landslide conditioning factors. This new algorithm was developed to overcome the subjectivity of the manual categorization of scale data of landslide conditioning factors, and to predict rainfall-induced susceptibility map in Kuala Lumpur city and surrounding areas using geographic information system (GIS). The main objective of this article is to use CHi-squared automatic interaction detection (CHAID) method to perform the best classification fit for each conditioning factor, then, combining it with logistic regression (LR). LR model was used to find the corresponding coefficients of best fitting function that assess the optimal terminal nodes. A cluster pattern of landslide locations was extracted in previous study using nearest neighbor index (NNI), which were then used to identify the clustered landslide locations range. Clustered locations were used as model training data with 14 landslide conditioning factors such as; topographic derived parameters, lithology, NDVI, land use and land cover maps. Pearson chi-squared value was used to find the best classification fit between the dependent variable and conditioning factors. Finally the relationship between conditioning factors were assessed and the landslide susceptibility map (LSM) was produced. An area under the curve (AUC) was used to test the model reliability and prediction capability with the training and validation landslide locations respectively. This study proved the efficiency and reliability of decision tree (DT) model in landslide susceptibility mapping. Also it provided a valuable scientific basis for spatial decision making in planning and urban management studies.