Frequency Ratio Method Research Articles

Landslide susceptibility mapping using GIS and FR method along the Cagayan de Oro-Bukidnon-Davao City route corridor, Philippines

Landslides along highways in Mindanao including the Cagayan de Oro (CDO)-Bukidnon-Davao City (Buda) highway pose high risk to many motorists and passengers. It also hampers the flow of transportation affecting the economy of the region not counting the cost of road maintenance and rehabilitation. This study aimed to evaluate the casual factors of rainfall-induced landslide along the study area using Geographic Information System (GIS) and Frequency Ratio (FR) approach. Global Positioning System (GPS) was used to determine the coordinates of 78 landslide locations recorded. Map layers of different parameters with FR values were overlaid to produce the rainfall-induced Landslide Susceptibility Index (LSI) map. The generated LSI map was classified into low, moderately low, moderate, moderately high and high risk. Results showed that rainfall coupled with soil type and slopes were the main triggering parameters of based on their corresponding FR values. Majority of high risk areas for rainfall-induced landslides were located along the southern portion belonging to Davao City and Kitao-tao and Arakan Municipalities. Validation using Receiver Operator Characteristic (ROC) revealed a model accuracy rate of 76.4% at 0.05 level. Hence, the LSI map generated can be a useful input in planning out projects for road repairs and maintenance along landslide prone areas so that cost, disturbances and exposure to such hazards may be minimized.

Landslide susceptibility mapping based on landslide history and analytic hierarchy process (AHP)

Landslide phenomena are responsible for significant economic damages, loss of human lives and irreversible changes in natural environment. In this manuscript, in a landslide-prone site of Cyprus, a landslide susceptibility model was developed by coupling the popular analytic hierarchy process and the frequency ratio method in a GIS environment. Initially, 10 triggering factors and their modified Landslide Relative Frequencies (mLRF) were determined. Then, AHP was employed so as to determine the relative weights of each landslide instability factor and to omit those that were not essential for the model. Land cover was the most important parameter in the manifestation of landslides, while a 0.00404 consistency ratio index value revealed that the weights of the landslide contributing factors were well assessed. By combining the landslide instability agents along with their weights, a landslide susceptibility map emerged illustrating that almost 64.1 % of the study area was expanded on high-risk zones. The ROC curves and La.M.I. index statistic measures were generated by utilizing a binary file of 977,500 pixels, which had incorporated equal areas of landslide cells and randomly selected non-landslide ones. The above measures indicated that the overall model accuracy in the detection of landslide phenomena was 73.9 % and the model’s performance was enhanced when the parameters with the smallest weights were omitted from the model. This information could be significant in undertaking the appropriate series of measures according to landslide susceptibility zones in order to mitigate the landslide risk in the area.

Quantitative assessment of landslide susceptibility along the Xianshuihe fault zone, Tibetan Plateau, China

The Xianshuihe fault (XSF) zone is a sinistral-slip fault system on the eastern margin of the Tibetan Plateau, with high Quaternary activity and frequent historic earthquakes. Large landslides along the fault zone have caused substantial damage and are the second-most important regional hazard after earthquakes. A landslide inventory in a 7339-km2 area along the XSF zone documented 415 landslides and evaluated landslide distribution characteristics using frequency ratio (FR) and weight-of-evidence (WOE) models to map landslide susceptibility. A total of 11 variables were analyzed as input variables: slope angle, slope aspect, altitude, planform curvature, topographic wetness index, distance from active faults, lithology, annual rainfall, distance from rivers, distance from roads, and the NDVI of the study area. Among these factors, the distance from active faults, altitude, slope angle, aspect, lithology, and rainfall were the dominant influencing factors. Assessment of six susceptibility mapping schemes with FR and WOE models using the ROC method showed that, while all six of the model schemes produced good results, on the whole, the FR model performed better than the WOE model. The landslide susceptibility map of the FR model with six variables performed as well as the WOE model with the full 11 variables. Inclusion of more variables did not necessarily translate into a better predictive capability. Several factors typically associated with susceptibility to rainfall-triggered landslides—topographic convergence, rainfall, and topographic wetness index—either did not follow expected patterns (rainfall) or were not good general predictors of landslide locations (topographic convergence and wetness index). In contrast, distance to active faults, slope aspect, and topographic divergence were the best predictors as measured by the FR method. This combination points to earthquakes rather than rainfall as the dominant landslide-triggering mechanism in the study area.

Factors controlling vegetation fires in protected and non-protected areas of myanmar.

Fire is an important disturbance agent in Myanmar impacting several ecosystems. In this study, we quantify the factors impacting vegetation fires in protected and non-protected areas of Myanmar. Satellite datasets in conjunction with biophysical and anthropogenic factors were used in a spatial framework to map the causative factors of fires. Specifically, we used the frequency ratio method to assess the contribution of each causative factor to overall fire susceptibility at a 1km scale. Results suggested the mean fire density in non-protected areas was two times higher than the protected areas. Fire-land cover partition analysis suggested dominant fire occurrences in the savannas (protected areas) and woody savannas (non-protected areas). The five major fire causative factors in protected areas in descending order include population density, land cover, tree cover percent, travel time from nearest city and temperature. In contrast, the causative factors in non-protected areas were population density, tree cover percent, travel time from nearest city, temperature and elevation. The fire susceptibility analysis showed distinct spatial patterns with central Myanmar as a hot spot of vegetation fires. Results from propensity score matching suggested that forests within protected areas have 11% less fires than non-protected areas. Overall, our results identify important causative factors of fire useful to address broad scale fire risk concerns at a landscape scale in Myanmar.

Landslide susceptibility zonation in part of Tehri reservoir region using frequency ratio, fuzzy logic and GIS

A comprehensive study for the identification of landslide susceptible zones using landslide frequency ratio and fuzzy logic in GIS environment is presented for Tehri reservoir rim region (Uttarakhand, India). Temporal remote sensing data was used to prepare important landslide causative factor layers and landslide inventory. Primary and secondary topographic attributes namely slope, aspect, relative relief, profile curvature, topographic wetness index, and stream power index, were derived from digital elevation model. Landslide frequency ratio technique was adopted to correlate factors with landslides. Further, fuzzy logic method was applied for the integration of factors (causative factor) to map landslide susceptible zones. Normalized landslide frequency ratio value was used for the fuzzy membership function and different fuzzy operators were considered for the preparation of landslide susceptibility/hazard index map. The factors considered in this study were found to be carrying a wide range of information. Accordingly, a methodology was evolved to integrate the factors using combined fuzzy gamma and fuzzy OR operation. Fuzzy gamma integration was performed for six different gamma values (range: 0–1). Gamma value of 0.95 was selected for the preparation of final susceptibility map. Landslide susceptibility index map was divided into the following five hazard zones – very low, low, moderate, high, and very high – on the basis of natural break classification. Validation of the model was performed by using cumulative percentage curve technique. Area under curve value of cumulative percentage curve of proposed landslide susceptibility map (gamma = 0.95) was found to be 0.834 and it can be said that 83.4% accuracy was achieved by applying combined fuzzy logic and landslide frequency ratio method.

Landslide hazard zonation mapping using frequency ratio and fuzzy logic approach, a case study of Lachung Valley, Sikkim

Abstract Background Sikkim Himalaya is under consistent distress due to landslides. Abrupt thrust on infrastructure development in the valley regions of Sikkim Himalaya has led to a need for a prior planning to face landslide hazard. A comprehensive study for the identification of landslide hazard zones using landslide frequency ratio and fuzzy logic in GIS environment has been presented for the Lachung valley, Sikkim, India, where a number of hydroelectric projects are proposed. Temporal remote sensing data was used to generate significant landslide causative factors in addition to landslide inventory. Primary topographic attributes namely slope, aspect and relative relief were derived from digital elevation model. Landslide frequency ratio approach was adopted to correlate landslide causal factors with landslide incidences. Further, fuzzy logic method was used for the integration of landslide causative factors in order to delineate the landslide hazard zones. Fuzzy memberships were derived from the landslide frequency ratio values. Different gamma values were used in fuzzy gamma integration process, which resulted different landslide hazard index maps. Receiver operating characteristic curves were prepared to analyze consistency of the resulting landslide hazard index maps. Results Landslide frequency ratio values have emphasised the importance of factors/classes in landsliding. High slope angle (35°-45°), very high slope category (>45°), High and very high relative relief categories; south, southeast and southwest aspects; drainage and lineaments buffer range of 0-50m, 50-100m and 100 to 150m; quartzite/garnet schist and migmatite type of lithology; Sandy loam and Rock/loam classes of soils; fallow land and sparse vegetation classes of land use/land cover were found to be associated with landsliding. Five landslide hazard zonation maps with each comprising five relative landslide hazard zones namely; very low, low, moderate, high and very high hazard zones were prepared by using five fuzzy gamma operators. Maps indicated that steep talus slopes, close proximity to drainages, ridges and spars fall under high hazard zones. Settlement areas were observed in low to moderate hazard zones. Very high hazard zones were observed in steep slopes, cliffs and cut slopes excavated for the roads. Low hazard zones were observed in agricultural terraces and permafrost areas. Conclusions Hence it can be concluded that landslide causative factor’s integration using fuzzy logic has yielded good results for Lachung valley. Frequency ratio method for determination of fuzzy membership value has reduced subjectivity in the model. The final LHZ map (γ = 0.92) can be used for the planning of future infrastructure, settlement and ecological development in Lachung region.

Landslide susceptibility mapping and factor effect analysis using frequency ratio in a catchment scale: a case study from Garuwa sub-basin, East Nepal

For assessing landslide susceptibility, the spatial distribution of landslides in the field is essential. The landslide inventory map is prepared on the basis of historical information of individual landslide events from different sources such as previously published reports, satellite imageries, aerial photographs and interview with local inhabitants. Then, the distribution of landslides in the study area is verified with field surveys. However, the selection of contributing factors for modelling landslide susceptibility is an inhibit task. The previous studies show that the factors are chosen as per availability of data. This paper documents the landslide susceptibility mapping in the Garuwa sub-basin, East Nepal using frequency ratio method. Nine different contributing factors are considered: slope aspect, slope angle, slope shape, relative relief, geology, distance from faults, land use, distance from drainage and annual rainfall. To analyse the effect of contributing factors, the landslide susceptibility index maps are generated four times using (a) topographical factors and geological factors, (b) topographical factors, geological factors and land use, (c) topographical factors, geological factors, land use and drainage and (d) all nine causative factors. By comparing with the pre-existing landslides, the fourth case (considering all nine causative factors) yields the best success rate accuracy, i.e. 81.19 %, which is then used to produce the final landslide susceptibility zonation map. Then, the final landslide susceptibility map is validated through chi-square test. The standard chi-square value with 3 degrees of freedom at the 0.001 significance level is 16.3, whereas the calculated chi-square value is 7,125.79. Since the calculated chi-square value is greater than the standard chi-square value, it can be concluded that the landslide susceptibility map is considered as statistically significant. Moreover, the results show that the predicted susceptibility levels are found to be in good agreement with the past landslide occurrences.

Remote sensing and GIS-based landslide susceptibility mapping using frequency ratio, logistic regression, and fuzzy logic methods at the central Zab basin, Iran

A remote sensing and geographic information system-based study has been carried out to map areas susceptible to landslides using three statistical models, frequency ratio (FR), logistic regression (LR), and fuzzy logic at the central Zab basin in the mountainsides in the southwest West Azerbaijan province in Iran. Ten factors such as slope, aspect, elevation, lithology, normalized difference vegetation index (NDVI), land cover, precipitation, distance to fault, distance to drainage, and distance to road were considered. Landsat ETM+ images were used for NDVI and land cover maps. A landslide inventory map of the study area was identified by a SPOT 5 satellite after which fuzzy algebraic operators were applied to the fuzzy membership values of landslide susceptibility mapping. In addition, FR and LR models were applied to determine the landslide susceptibility. The three models are validated using the receiver operating characteristic and the area under which curve values were calculated. The validation results showed that the LR model (accuracy is 96 %) has better prediction than fuzzy logic (accuracy is 95 %) and FR (accuracy is 94 %) models. Also, among the fuzzy operators, the gamma operator (λ = 0.975) showed the best accuracy (94.64 %) while the fuzzy OR operator when applied showed the worst accuracy (85.11 %).

Flood susceptibility analysis and its verification using a novel ensemble support vector machine and frequency ratio method

Flood is one of the most commonly occurred natural hazards worldwide. Severe flood occurrences in Kelantan, Malaysia cause damage to both life and property every year. Due to the huge losses in this area, development of appropriate flood modeling is required for the government. Remote sensing and geographic information system techniques can support overall flood management as they can produce rapid data collection and analysis for hydrological studies. The existing models for flood mapping have some weak points that may improve through more sophisticated and ensemble methods. The current research aimed to propose a novel ensemble method by integrating support vector machine (SVM) and frequency ratio (FR) to produce spatial modeling in flood susceptibility assessment. In the literature, mostly statistical and machine learning methods are used individually; however, their integration can enhance the final output. The FR model can perform bivariate statistical analysis and evaluate the correlation between the flooding and classes of each conditioning factors. The weights achieved by FR can be assigned to each conditioning factor and the resulted factors can be used in SVM analysis. In order to examine the efficiency of the proposed ensemble method and to show the proficiency of SVM, another machine learning algorithm such as decision tree (DT) was applied and the results were compared. To perform the methods, the upper catchment of the Kelantan basin in Malaysia was chosen. First, a flood inventory map with a total of 155 flood locations were extracted from various sources over the study area. The flood inventory map was randomly divided into two dataset; 70 % (115 flood locations) for the purpose of training and the remaining 30 % (40 flood locations) was used for validation. The spatial database included digital elevation model, curvature, geology, river, stream power index, rainfall, land use/cover, soil type, topographic wetness index and slope. For model validation, area under curve method was used and both success and prediction rate curves were calculated. The validation results for ensemble method showed 88.71 and 85.21 % for success rate and prediction rate respectively. The DT model showed 87.00 and 82.00 % for the success rate and prediction rate respectively. It is evident that the accuracies were increased using the ensemble method. The acquired results proved the efficiency of the proposed ensemble method as rapid, accurate and reasonable in flood susceptibility assessment.

Frequency Ratio: a method for dealing with missing values within nearest neighbour search

In this paper we introduce the Frequency Ratio (FR) method for dealing with values within nearest neighbour search. We test the FR method on known medical datasets from the UCI machine learning repository. We compare the accuracy of the FR method with five commonly used methods (three imputation and two bypassing methods) for dealing with values that are missing completely at random (MCAR) for the purpose of classification. We discovered that in most cases, the FR method outperforms the other methods. We conclude that the FR method is a strong addition to the commonly used methods for dealing with values within the nearest neighbour method.

Sinkhole Susceptibility Mapping Using a Frequency Ratio Method and GIS Technology Near Karapınar, Konya-Turkey

The purpose of this study is to produce a sinkhole susceptibility map of the Obruk Plateau in central Turkey, based on a frequency ratio method within a Geographic Information System environment. Using field surveys, the locations of the 182 sinkholes were determined in the study area. In this study, 30 sinkhole-related factors were determined and used in the analysis. The values of the frequency ratio were estimated using frequency ratio analysis and were used to calculate the sinkhole potential for the entire study area. For model validation, the constructed sinkhole occurrence potential map was overlaid onto the 182 observed sinkholes, showing that 169 sinkholes (93% of all sinkholes) were determined to be in the high and moderate sinkhole occurrence potential areas. In addition, the predictive capabilities of the models were determined by the area under the relative operating characteristic (ROC) curve. The area value under the ROC curve was 0.953. These results indicate that the model is a good estimator of the sinkhole potential in the study area. The sinkhole potential map shows that the areas of very low or no, low, moderate and high sinkhole potential classes are 712.31 km2; 641.15 km2, 572.45 km2; and 438.01 km2, respectively. The interpretations of the potential map showed that geology (lithology), elevation, aspect, fault density, distance to fault, and decline of the groundwater level play major roles in sinkhole occurrence and distribution in the Obruk Plateau. The resulting sinkhole potential map can aid planners and engineers in development plans and land-use planning.

Application of logistic regression and frequency ratio in the spatial distribution of debris-rockslides: Precordillera of San Juan, Argentina

This study employs statistical modeling techniques and geomorphological mapping to analyze the distribution of debris-rockslides in relation to altitude, aspect, slope angle, lithology and distance to lineaments using optical remote sensing techniques with GIS. The study area includes a portion of the Central Precordillera of San Juan around 32°S, 69°W, where few geomorphological studies have been conducted. A debris-rockslides inventory map was prepared using debris-rockslides locations identified from satellite imageries, a total of 35 debris-rockslides were identified in an area of 288.13 km2 (with a total slide area of 6.36 km2). All of the debris-rockslides were rotational slides, affecting Upper Ordovician and Upper Devonian deposits. The volumes of debris-rockslides were estimated using the sloping local base level technique that allows computing the possible pre-rockslide topography and the bedrock–deposits interface. The relationship between the variables and the slide distribution was analyzed using the frequency ratio method and logistic regression models. The analytical results show that lithology and slope exert major influences. Lithology is found to be the most important aspect in the study area.

Susceptibility Assessment of Shallow Landslides in Hulu Kelang Area, Kuala Lumpur, Malaysia Using Analytical Hierarchy Process and Frequency Ratio

Hulu Kelang is known as one of the most landslide susceptible areas in Malaysia. From 1990 to 2011, a total of 28 landslide events had been reported in this area. This paper compares two models as Analytical Hierarchy Process (AHP) and probability–frequency ratio (FR) methods for recognizing landslide susceptibility regions in the Hulu Kelang area. Eleven landslide influencing factors were considered to form the probability–FR and AHP matrix, i.e. lithology-weathering, land cover, curvature, slope inclination, slope aspect, drainage density, elevation, distance to lake and stream, distance to road and trenches, the Stream Power Index and the Topographic Wetness Index. The accuracy of the maps produced from the two models were verified using a receiver operating characteristics. The verification results indicated that the probability–FR model based on probabilistic analysis of spatial distribution of historical landslide events was capable of producing a more reliable landslide susceptibility map in this study area compared to AHP model. About 89 % of the landslide locations have been predicted accurately by using the FR map.

An assessment of multivariate and bivariate approaches in landslide susceptibility mapping: a case study of Duzkoy district

Landslide susceptibility maps are valuable sources for disaster mitigation works and future investments of local authorities in unstable hazard-prone areas. However, there are limitations and uncertainties inherent in landslide susceptibility assessment. For this purpose, many methods have been suggested and applied in the literature, which are generally categorized as bivariate and multivariate. Here, in this paper, the most popular and widely used multivariate [support vector regression (SVR), logistic regression (LR) and decision tree (DT)] and bivariate methods [frequency ratio (FR), weight of evidence (WOE) and statistical index (SI)] were compared with respect to their performances in landslide susceptibility modeling problem. Duzkoy district of Trabzon Province was selected due to its unique topographical and lithological characteristics, magnifying shallow landslide risk potential. Slope, lithology, land cover, aspect, normalized difference vegetation index, soil thickness, drainage density, topographical wetness index and elevation were employed as landslide occurrence factors. Accuracy measures based on confusion matrix (i.e., overall accuracy and Kappa coefficient) and receiver operating characteristic (ROC) curve were employed to compare the performances of the methods. Furthermore, McNemar’s test was employed to analyze the statistical significance of differences in method performances. The results indicated that multivariate approaches (i.e., SVR, LR and DT) outperformed the bivariate methods (i.e., FR, SI and WOE) by about 13 %. Within the multivariate approaches, SVR method performed the best with the highest accuracy, while FR method was the most effective and accurate bivariate method. Interpretation of AUC values and the McNemar’s statistical test results revealed that the SVR method was superior in modeling landslide susceptibility compared with the other multivariate and bivariate methods.

Landslide susceptibility mapping by combining the three methods Fuzzy Logic, Frequency Ratio and Analytical Hierarchy Process in Dozain basin

Abstract. Landslides are among the most important natural hazards that lead to modification of the environment. Therefore, studying of this phenomenon is so important in many areas. Because of the climate conditions, geologic, and geomorphologic characteristics of the region, the purpose of this study was landslide hazard assessment using Fuzzy Logic, frequency ratio and Analytical Hierarchy Process method in Dozein basin, Iran. At first, landslides occurred in Dozein basin were identified using aerial photos and field studies. The influenced landslide parameters that were used in this study including slope, aspect, elevation, lithology, precipitation, land cover, distance from fault, distance from road and distance from river were obtained from different sources and maps. Using these factors and the identified landslide, the fuzzy membership values were calculated by frequency ratio. Then to account for the importance of each of the factors in the landslide susceptibility, weights of each factor were determined based on questionnaire and AHP method. Finally, fuzzy map of each factor was multiplied to its weight that obtained using AHP method. At the end, for computing prediction accuracy, the produced map was verified by comparing to existing landslide locations. These results indicate that the combining the three methods Fuzzy Logic, Frequency Ratio and Analytical Hierarchy Process method are relatively good estimators of landslide susceptibility in the study area. According to landslide susceptibility map about 51% of the occurred landslide fall into the high and very high susceptibility zones of the landslide susceptibility map, but approximately 26 % of them indeed located in the low and very low susceptibility zones.

Landslide susceptibility mapping using ensemble bivariate and multivariate statistical models in Fayfa area, Saudi Arabia

A comprehensive landslide susceptibility mapping (LSM) should be produced to reduce damages to individuals and infrastructures. In the international landslide literature, various statistical methods such as logistic regression (LR) and frequency ratio (FR) have been widely used individually for LSM. However, these methods have some weaknesses. This study aims to propose an ensemble method of FR and LR to overcome their weak points. This ensemble method was applied on Fayfa area which is located in the southwest of Saudi Arabia (SA) in Jazan region. The study area covers an area of ~260 km2, and is located between latitudes 17°11′46″N to 17°21′16′′N and longitudes 43°0′30′′E to 43°11′38′′E. For LSM, an inventory map with 86 landslide occurrences was extracted from various sources. Then, the landslide inventory was randomly divided into two datasets: 70 % for training the models and 30 % for validation. The landslide conditioning factors used in the LSL include altitude, curvature, distance from wadis, distance from road, distance from fault, stream power index, topographic wetness index, soil type, geology, slope, and aspect. The produced susceptibility maps were validated by an area under the curve technique. The success rate curves of the models measure the goodness of fit of the training data, while the prediction rate curves provide the validation of the susceptibility maps, measuring the prediction capability. The validation results showed 65, 69, and 79 % success rates for FR, LR, and ensemble models, respectively. Furthermore, the prediction rates were 58, 77, and 82 % for FR, LR, and ensemble models, respectively. The proposed ensemble method can be replicated in other natural hazard studies because it can produce accurate assessment for disaster management and decision making.

Landslide hazard mapping considering rainfall probability in Inje, Korea

This study evaluated the landslide hazard at Inje, Korea, using a geographic information system (GIS) and rainfall probabilistic data. The locations of landslides were identified in the study area by aerial photograph interpretation and field surveys. Data about rainfall probability, topography, and geology were collected, processed, and compiled in a spatial database using GIS. Then, the probability of landslides in the study area in recurrence interval years in the future was calculated assuming that landslides are triggered by a daily rainfall of 202 mm or a three-day cumulative rainfall of 449 mm. Twelve factors that influence landslide occurrence were chosen from a database of topography, soil, and forest cover. Landslide susceptibilities were analysed and mapped according to these landslide-occurrence factors, employing the frequency ratio method. Of the total landslide locations, 50% were used for hazard analysis and the remaining 50% were used for model validation. Validation results for the daily rainfall of 202 mm and three-day cumulative rainfall of 449 mm for recurrence interval years were from 89.22% to 91.80% and from 89.38% to 93.80%, respectively. This analysis of landslide hazards took rainfall probability into account. Rainfall, including heavy rainfall, is expected to increase in the future.

Using Frequency Ratio Method for Spatial Landslide Prediction

Numerous landslides have occurred in the study area and they damage to agriculture and pasturelands. Since the study area do not have any landslide inventory and landslide predicted maps, landslide inventory produced based on field research (GPS) and satellite image (Geoeye and Ikonos). Frequency ratio technique is a statistical approach to simulation environmental conditions. It also uses to take the factors related to dependent variable. Frequency technique considered for generating landslide susceptibility map. Pixel landsliding and non-landsliding calculated in eight factors related-landslide. Landslide susceptibility map produce in five insensitive to very high sensitive classes based on natural breaks method. Receiver Operating Characteristics (ROC) graph implement to evaluate of the frequency ratio method. In particular, the model will be able to predict landslide area occurrence in future completely (sensitivity = 1). Although, model identify insensitive area with 17% errors (specificity = 0.83).

Earthquake induced landslide susceptibility mapping using an integrated ensemble frequency ratio and logistic regression models in West Sumatera Province, Indonesia

An 8 Richter Scale (RS) earthquake struck West Sumatra on Wednesday, 30 September 2009, at 17.16pm which led to huge number of landslides. Hence a comprehensive landslide susceptibility mapping (LSM) should be produced in order to reduce the damages to people and infrastructures. In the international landslide literature, various statistical methods such as frequency ratio (FR) and logistic regression (LR) have been widely used individually for LSM, but they have some weaknesses. FR which is able to perform bivariate statistical analysis (BSA) assesses the influence of classes of each conditioning factor on landslide occurrence. However, the correlation between the factors is mostly neglected. On the other hand, LR is able to analyze the relationship among the factors while it is not capable to evaluate the classes of each landslide conditioning factor. This paper aims to propose an ensemble method of FR and LR in order to overcome their weak points. For LSM, a landslide inventory map with a total of 87 landslide locations was extracted from various sources. Then the landslide inventory was randomly divided into two datasets 70% for training the models and the remaining 30% was used for validation purpose. The landslide conditioning factors consist of: altitude, curvature, river, SPI, rainfall, soil type, soil texture, land use/cover (LULC), peak ground acceleration (PGA), geology, slope, aspect, lineament and topographic wetness index (TWI). Four PGA of 7.5, 8, 8.6 and 9 were acquired and PGA 8 which was related to the 2009 earthquake was used to generate the model. Finally, the produced landslide susceptibility maps were validated using an area under the (ROC) curve method. For the model which was derived by PGA 8, the validation results showed 84% and 78% success and prediction rates respectively. Furthermore, the prediction rates for the models made by PGA 7.2, 8.6 and 9 are 79%, 78% and 81% respectively. The result proved the reasonable efficiency of the proposed method for earthquake induced landslide susceptibility mapping. Also the proposed ensemble method can be used in other hazard studies as it is capable to produce rapid and accurate assessment for disaster management and decision making.

Landslide susceptibility mapping by frequency ratio and logistic regression methods: an example from Niksar–Resadiye (Tokat, Turkey)

Turkey confronts loss of life and large economic losses due to natural disasters caused by its morphologic structure, geographical placement, and climate characteristics. The Kuzulu (Koyulhisar) landslide, which caused loss of life and property on 17th March 2005, occurred in an area near the country’s most important active fault, the North Anatolian Fault Zone. To mitigate and prevent landslide damages, prediction of landslide susceptibility areas based on probabilistic methods has a great importance. The purpose of this study was to produce a landslide susceptibility map by the logistic regression and frequency ratio methodologies for a 733-km2 area near the North Anatolian Fault Zone from the southeast of Niksar to Resadiye in Tokat province. Conditioning parameters, such as elevation, slope gradient, slope aspect, distance to streams, roads, and faults, drainage density, and fault density, were used in the analysis. Before susceptibility analysis, the landslides observed in the area were separated into two groups for use in analysis and verification, respectively. The susceptibility maps produced had five different susceptibility classes such as very low, low, moderate, high, and very high. To test the performance of the susceptibility maps, area under curve (AUC) approach was used. For the logistic regression method, the AUC value was 0.708; while for the frequency rate method, this value was 0.744. According to these AUC values, it could be concluded that the two landslide susceptibility maps obtained were successful.