Prediction of landslide susceptibility using rare events logistic regression: A case-study in the Flemish Ardennes (Belgium)

Eeckhaut, Miet Van Den; Vanwalleghem, Tom; Poesen, Jean; Govers, Gérard; Verstraeten, Gert; Vandekerckhove, Liesbeth

doi:10.1016/j.geomorph.2005.12.003

Cited by 359 publications

(209 citation statements)

References 29 publications

Supporting

Mentioning

194

Contrasting

Unclassified

Order By: Relevance

“…In a landslide susceptibility prediction in W Belgium, Van den Eeckhaut et al (2006) used recently a "rare event logistic regression" model to circumvent the problem of small datasets. Although they also concluded that this multivariate model is a useful tool to These cumulative plots indicate how many "most hazardous" percents of the study area have to be taken into account to predict a given percentage of the future landslides according to each model and the adopted cross-validation process.…”

Section: Sample Size and Prediction Ratementioning

confidence: 99%

“…(1). The logistic model in Van den Eeckhaut et al (2006) and the estimator used in Eq. (3) here are two different estimation procedures based on two different assumptions (Chung, 2006).…”

Section: Sample Size and Prediction Ratementioning

confidence: 99%

“…(3) here are two different estimation procedures based on two different assumptions (Chung, 2006). In order to interpret the estimate of the logistic model directly as a probability, Van den Eeckhaut et al (2006) proposed several modifications in the estimation procedure. We do state that a direct probabilistic interpretation of the likelihood ratio function for the occurrences of future landslides is indeed not possible without several specific assumptions on the landslides.…”

Section: Sample Size and Prediction Ratementioning

confidence: 99%

See 2 more Smart Citations

Mapping landslide susceptibility from small datasets: A case study in the Pays de Herve (E Belgium)

Demoulin

Chung

2007

Geomorphology

View full text Add to dashboard Cite

A landslide susceptibility map is proposed for the Pays de Herve (E Belgium), where large landslides affect Cretaceous clay outcrop areas. Based on a Bayesian approach, this GIS-supported probabilistic map identifies the areas most susceptible to deep landslides. The database is comprised of the source areas of ten pre-existing landslides (i.e. a sample of 154 grid cells) and of six environmental data layers, namely lithology, proximity to active faults, slope angle and aspect, elevation and distance to the nearest valley-floor. A 30-m-resolution DEM from the Belgian National Geographical Institute is used for the analysis. Owing to the small size of the sample, a special cross-validation procedure of the susceptibility map is performed, which uses in an iterative way each of the landslides to test the predictive power of the map derived from the other landslides. Four different sets of variables are used to produce four susceptibility maps, whose prediction curves are compared. While the prediction rates associated with the models not involving the "proximity to active fault" criterion are comparable to those of the models considering this variable, strong weaknesses inherent in the fault data on which the latter rely suggest that the final susceptibility map should be based on a model that excludes any reference to fault. This highlights the difference between a triggering factor and determining factors, and in the same time broadens the scope of the produced map. A single reactivated slide is also used to test the possibility of predicting future reactivation of existing landslides in the area. Finally, the need for geomorphological control over the mathematical treatment is underlined in order to obtain realistic prediction maps.

show abstract

Section: Sample Size and Prediction Ratementioning

confidence: 99%

“…(1). The logistic model in Van den Eeckhaut et al (2006) and the estimator used in Eq. (3) here are two different estimation procedures based on two different assumptions (Chung, 2006).…”

Section: Sample Size and Prediction Ratementioning

confidence: 99%

Section: Sample Size and Prediction Ratementioning

confidence: 99%

See 1 more Smart Citation

Mapping landslide susceptibility from small datasets: A case study in the Pays de Herve (E Belgium)

Demoulin

Chung

2007

Geomorphology

View full text Add to dashboard Cite

show abstract

“…The method adopted in the literature to divide the susceptibility histogram into different categories is in many cases based on expert criteria (Dai and Lee, 2002;Ohlmacher and Davis, 2003;Van Den Eeckhaut et al, 2006) and does not take into account the real underlying data. It is necessary to explore data and obtain knowledge of their statistical distribution before applying any method of classification (Foote and Crum, 2014).…”

Section: Data Source and Landslide Susceptibility Mapmentioning

confidence: 99%

“…This index is now widely used in literature (Guzzetti et al 2006; Van den Eeckhaut, 2006;Thiery et al, 2007 andSterlacchini et al, 2008) for the comparison of susceptibility models. In Table 3a most kappa (K) values below 0.4 suggest that the LSMs are not as similar as revealed by Spearman's coefficients, with some exceptions.…”

Section: Unweighted and Linear Weighted Kappa Indexmentioning

confidence: 99%

Statistical and spatial analysis of landslide susceptibility maps with different classification systems

Baeza¹,

Lantada

Amorim

2016

Environ Earth Sci

View full text Add to dashboard Cite

A landslide susceptibility map is an essential tool for land-use spatial planning and management in mountain areas. However, a classification system used for readability determines the final appearance of the map and may therefore influence the decision-making tasks adopted. The present paper addresses the spatial comparison and the accuracy assessment of some well-known classification methods applied to a susceptibility map that was based on a discriminant statistical model in an area in the Eastern Pyrenees. A number of statistical approaches (Spearman's correlation, Kappa index, Factorial and Cluster analyses and Landslide density index) for map comparison were performed to quantify the information provided by the usual image analysis.The results showed the reliability and consistency of the Kappa index against Spearman's correlation as accuracy measures to assess the spatial agreement between maps. Inferential tests between unweighted and linear weighted Kappa results showed that all the maps were more reliable in classifying areas of highest susceptibility and less reliable in classifying areas of low to moderate susceptibility. The spatial variability detected and quantified by Factorial and Cluster analyses showed that the maps classified by Quantile and Natural Break methods were the closest whereas those classified by Landslide Percentage and Equal Interval methods displayed the greatest differences. The difference image analysis showed that the five classified maps only matched 9% of the area. This area corresponded to the steeper slopes and the steeper watershed angle with forestless and sunny slopes at low altitudes. This means that the five maps coincide in identifying and classifying the most dangerous areas. The Equal Interval map overestimated the susceptibility of the study area and the Landslide Percentage map was considered to be a very optimistic model. The spatial pattern of the Quantile and Natural Break maps was very similar but the latter was more consistent and predicted potential landslides more efficiently and reliably in the study area.

show abstract

An advanced process-based distributed model for the investigation of rainfall-induced landslides: The effect of process representation and boundary conditions

2015

View full text Add to dashboard Cite

Extreme rainfall events are the major driver of shallow landslide occurrences in mountainous and steep terrain regions around the world. Subsurface hydrology has a dominant role on the initiation of rainfall-induced shallow landslides, since changes in the soil water content affect significantly the soil shear strength. Rainfall infiltration produces an increase of soil water potential, which is followed by a rapid drop in apparent cohesion. Especially on steep slopes of shallow soils, this loss of shear strength can lead to failure even in unsaturated conditions before positive water pressures are developed. We present HYDROlisthisis, a process-based model, fully distributed in space with fine time resolution, in order to investigate the interactions between surface and subsurface hydrology and shallow landslides initiation. Fundamental elements of the approach are the dependence of shear strength on the three-dimensional (3-D) field of soil water potential, as well as the temporal evolution of soil water potential during the wetting and drying phases. Specifically, 3-D variably saturated flow conditions, including soil hydraulic hysteresis and preferential flow phenomena, are simulated for the subsurface flow, coupled with a surface runoff routine based on the kinematic wave approximation. The geotechnical component of the model is based on a multidimensional limit equilibrium analysis, which takes into account the basic principles of unsaturated soil mechanics. A series of numerical simulations were carried out with various boundary conditions and using different hydrological and geotechnical components. Boundary conditions in terms of distributed soil depth were generated using both empirical and process-based models. The effect of including preferential flow and soil hydraulic hysteresis was tested together with the replacement of the infinite slope assumption with the multidimensional limit equilibrium analysis. The results show that boundary conditions play a crucial role in the model performance and that the introduced hydrological (preferential flow and soil hydraulic hysteresis) and geotechnical components (multidimensional limit equilibrium analysis) significantly improve predictive capabilities in the presented case study.

show abstract

Prediction of landslide susceptibility using rare events logistic regression: A case-study in the Flemish Ardennes (Belgium)

Cited by 359 publications

References 29 publications

Mapping landslide susceptibility from small datasets: A case study in the Pays de Herve (E Belgium)

Mapping landslide susceptibility from small datasets: A case study in the Pays de Herve (E Belgium)

Statistical and spatial analysis of landslide susceptibility maps with different classification systems

An advanced process-based distributed model for the investigation of rainfall-induced landslides: The effect of process representation and boundary conditions

Contact Info

Product

Resources

About