Landslide Ratio Research Articles

Distribution and formation of soil balls under heavy rainstorm conditions in the northern Loess Plateau

The occurrence of a peculiar phenomenon called “soil balls” was observed in Dingbian County, northern Loess Plateau, due to an extremely heavy rainstorm on July 10, 2022. Determining the distribution characteristics and formation of these soil balls is crucial for enriching the knowledge and the precise prevention and control of soil erosion on the Loess Plateau. This research is particularly significant in studying the off-site effects of soil erosion caused by rainstorm and flooding disasters. This study conducted a comprehensive analysis through field investigation and drone aerial photography in the three dam-controlled watersheds selected. The results showed that the soil balls originated from the soil clods generated by landslides and collapses that occurred within the watersheds during heavy rainstorms. Under the action of hyperconcentrated flows, the soil clods were transported downstream. The size of the soil balls is mainly in the range of 20–60 cm in diameter, accounting for 81% of the total soil balls in the dam lands, with significant differences in the morphology at different locations within a dam land region. The average area density and volume density of soil balls in the dam lands of the watersheds are 0.13–0.25 m2 m−2 and 0.05–0.11 m3 m−2, respectively. In each watershed, the area density and volume density of the soil balls exhibit the following order: one-third of the dam land nearest from the dam wall > the middle one-third of the dam land > one-third of the dam land furthest the dam wall. The total erosion intensity of landslides and collapses in the three watersheds was 18,730, 5,213, and 9,879 m3 km−2, respectively. The sediment transport ratio of landslides and collapses expressed by the soil balls in the dam land is at least 4–10%. The study also indicated that the volume of landslides and collapses, the shape coefficient of the watershed, the density of the confluence paths, and the peak discharge had a significant impact on the volume density distribution of the soil balls in the dam lands. We recommend implementing certain measures in similar watersheds, including converting steep slopes into terraced fields or grasslands, constructing water storage and drainage facilities, building energy dissipation measures in channels, and reinforcing check dams, to reduce the risk of soil ball events occurring under heavy rainstorm conditions, and protect the ecological environment on the Loess Plateau.

Landslide susceptibility mapping in Three Gorges Reservoir area based on GIS and boosting decision tree model

As one of the most destructive geological disasters, a myriad of landslides has revived and developed in the Three Gorges Reservoir area under the combined action of various detrimental factors. Therefore, the pertinently regional landslide susceptibility mapping (LSM) is of great significance for disaster prevention and mitigation. In this study, LSM is prepared by using a boosting-C5.0 decision tree model. Under the landslide verification of on-site investigations, the study area is divided into accumulation and rock areas, and a total of 12 impact factors are selected. TOL and VIF are employed to determine the multicollinearity among the impact factors. The independent training (80%) and validation (20%) datasets are constructed by random sampling for LSM. ANN, C5.0, and SVM are selected for comparative analysis. The results show that there is no rigorous multicollinearity among the impact factors proposed in this paper. The landslide susceptibility in the study area is divided into low, moderate, high, and very high. The highest susceptibility area distributes along the riverside where the landslide ratio is 37.05% in boosting-C5.0 model. Then the ROCs are expropriated to infer the accuracy of each model. The boosting-C5.0 performs the best with the largest area under the curve in both accumulation and rock areas, reaching at 0.991 and 0.990 in the validation sets, respectively. Finally, the composite modification of the 5 validation sets shows that the uncertainty of boosting-C5.0 is concentrated in the intermediate probability areas of susceptibility. This study reveals the feasibility of machine learning in landslide susceptibility assessment, which could provide a basis for the risk management and control of geological disasters.

Introducing SlideforMAP: a probabilistic finite slope approach for modelling shallow-landslide probability in forested situations

Abstract. Shallow landslides pose a risk to infrastructure and residential areas. Therefore, we developed SlideforMAP, a probabilistic model that allows for a regional assessment of shallow-landslide probability while considering the effect of different scenarios of forest cover, forest management and rainfall intensity. SlideforMAP uses a probabilistic approach by distributing hypothetical landslides to uniformly randomized coordinates in a 2D space. The surface areas for these hypothetical landslides are derived from a distribution function calibrated on observed events. For each generated landslide, SlideforMAP calculates a factor of safety using the limit equilibrium approach. Relevant soil parameters are assigned to the generated landslides from log-normal distributions based on mean and standard deviation values representative of the study area. The computation of the degree of soil saturation is implemented using a stationary flow approach and the topographic wetness index. The root reinforcement is computed by root proximity and root strength derived from single-tree-detection data. The ratio of unstable landslides to the number of generated landslides, per raster cell, is calculated and used as an index for landslide probability. We performed a calibration of SlideforMAP for three test areas in Switzerland with a reliable landslide inventory by randomly generating 1000 combinations of model parameters and then maximizing the area under the curve (AUC) of the receiver operation curve. The test areas are located in mountainous areas ranging from 0.5–7.5 km2 with mean slope gradients from 18–28∘. The density of inventoried historical landslides varies from 5–59 slides km−2. AUC values between 0.64 and 0.93 with the implementation of single-tree detection indicated a good model performance. A qualitative sensitivity analysis indicated that the most relevant parameters for accurate modelling of shallow-landslide probability are the soil thickness, soil cohesion and the precipitation intensity / transmissivity ratio. Furthermore, we show that the inclusion of single-tree detection improves overall model performance compared to assumptions of uniform vegetation. In conclusion, our study shows that the approach used in SlideforMAP can reproduce observed shallow-landslide occurrence at a catchment scale.

Long-term hillslope erosion and landslide–channel coupling in the area of the catastrophic Wenchuan earthquake

Approximately 200,000 landslides were triggered by the 2008 Wenchuan earthquake and subsequent strong rainfall events, and the amount of landslide sediment delivered to the channels within catchments depended on the amount of landslide debris generated by the earthquake and the amount of post-earthquake rainfall. To better understand long-term hillslope erosion and the long-term changes in landslide–channel coupling, this study quantifies the coupling degree of landslide–channel by defining a landslide-channel coupling index (LCCI) and estimates the long-term hillslope erosion driven by landslides within 12 debris flow basins. We track landslide activity using multiyear imagery from 2008 to 2019, and the landslide ratio, landslide new regeneration ratio, and landslide reactivated ratio indicate that the overall landslide activity decayed following the Wenchuan earthquake. The calculation results of the coupling index illustrate that the coupling degree of landslide–channel in the study area follows a power law trend of decay over time. However, due to the influence of landslide activity, the coupling index decreased significantly in the first 5 years after the Wenchuan earthquake and then decreased slowly. The hillslope erosion driven by landslides is estimated using a conservative statistical model. The result shows that the landslide-driven hillslope erosion has decayed following a power law, and elevated levels of erosion may last for an unexpectedly long period. In addition, the landslide–channel coupling degree affects the movement of landslide sediment to channels. While the coupling degree of landslide–channel and the amount of landslide sediment supplied to the channels have decreased significantly in recent years, abundant landslide sediment could still be delivered to the channels when strong rainfall occurs in this region. Furthermore, a considerable amount of sediment appears to remain within the tributary channels and trunk channels, which means the activity of debris flow may last longer period than hillslope erosion.

The Effect of Temporal Characteristics on Developing a Practical Rainfall-Induced Landslide Potential Evaluation Model Using Random Forest Method

With the unique rainfall patterns of typhoons, plum rains, and short-term heavy rainfalls, the frequent landslide and debris flow disasters have caused severe loss to people in Taiwan. In the studies of landslide susceptibility, the information of factors used for analysis was usually annual-based content, and it was assumed that the same elements from different years were independent between each year. However, the occurrence of landslides was usually not simply due to the changes within a year. Instead, landslides were triggered because the factors that affected the potential of landslides reached critical conditions after a cumulative change with time. Therefore, this study had well evaluated the influence of temporal characteristics and the ratios of antecedent landslide areas in the past five years in the landslide potential evaluation model. The analysis was conducted through the random forest (RF) algorithm. Additional rainfall events of 2017 were used to test the proposed model’s performance to understand its practicality. The analysis results show that in the study area, the RF model had considerably acceptable performance. The results have also demonstrated that the antecedent landslide ratios in the past five years were essential to describe the significance of cumulative change with time when conducting potential landslide evaluation.

Slow surface subsidence and its impact on shallow loess landslides in a coal mining area

Surface subsidence caused by underground coal mining affects the hillslope stability conditions. However, few studies have focused on the coupling relationship between slow surface subsidence and landslide occurrences. A detailed landslide and fissure inventory in a coal mining area in Shaanxi Province, China, was produced based on interpretation of multitemporal satellite images and unmanned aerial vehicle (UAV) surveys. We used the interferometric synthetic aperture radar (InSAR) technique and landslide and fissure spatiotemporal statistics to investigate the spreading process of the slow subsidence caused by underground mining and examined its impact on the occurrence of shallow landslides. The InSAR results indicate that the actual extent of the subsidence zone is larger than the range of underground mining, which formed a subsidence basin along the coal mining panels. The subsidence curves go through initial, accelerative, and slow subsidence stages and characterized by S-shaped, which can be adequately fitted with logistic regression. Moreover, subsidence does not cease after the end of coal exploitation. Logistic models predicted that the duration of residual subsidence reached about 2–3 years. Subsidence significantly increased the likelihood of landslide occurrences. The spatial pattern of landslides is associated with the actual coal mining. We also investigated the clustering phenomenon of landslides and fissures under the impacts of subsidence. The frequency ratio of landslides and fissures increased with the cumulative subsidence. Finally, we propose a schematic view for landslides caused by coal mining and precipitation. This study will be helpful for elucidating the spatial–temporal evolution of slow subsidence and its impact on loess landslides in coal mining area.

Improving Geospatial Agreement by Hybrid Optimization in Logistic Regression-Based Landslide Susceptibility Modelling

This study aims to develop a logistic regression model of landslide susceptibility based on GeoDetector for dominant-factor screening and 10-fold cross validation for training sample optimization. First, Fengjie county, a typical mountainous area, was selected as the study area since it experienced 1,522 landslides from 2001 to 2016. Second, 22 factors were selected as the initial conditioning factors, and a geospatial database was established with a grid of 30 m precision. Factor detection of the geographic detector and the stepwise regression method included in logistic regression were used to screen out the dominant factors from the database. Then, based on the sample dataset with a 1:10 ratio of landslides and nonlandslides, 10-fold cross validation was used to select the optimized sample to train the logistic regression model of landslide susceptibility in the study area. Finally, the accuracy and efficiency of the two models before and after screening out the dominant factors were evaluated and compared. The results showed that the total accuracy of the two models was both more than 0.9, and the area under the curve value of the receiver operating characteristic curve was more than 0.8, indicating that the models before and after screening factor both had high reliability and good prediction ability. Besides, the screened factors had an active leading role in the geospatial distribution of the historical landslide, indicating that the screened dominant factors have individual rationality. Improving the geospatial agreement between landslide susceptibility and actual landslide-prone by the screening of dominant factors and the optimization of the training samples, a simple, efficient, and reliable logistic-regression–based landslide susceptibility model can be constructed.

Spatiotemporal Hotspots and Decadal Evolution of Extreme Rainfall-Induced Landslides: Case Studies in Southern Taiwan

The 2009 Typhoon Morakot triggered numerous landslides in southern Taiwan, and the landslide ratios in the Ailiao and Tamali river watershed were 7.6% and 10.7%, respectively. The sediment yields from the numerous landslides that were deposited in the gullies and narrow reaches upstream of Ailiao and Tamali river watersheds dominated the landslide recovery and evolution from 2010 to 2015. Rainfall records and annual landslide inventories from 2005 to 2015 were used to analyze the landslide evolution and identify the landslide hotspots. The landslide recovery time in the Ailiao and Tamali river watershed after 2009 Typhoon Morakot was estimated as 5 years after 2009 Typhoon Morakot. The landslide was easily induced, enlarged, or difficult to recover during the oscillating period, particularly in the sub-watersheds, with a landslide ratio > 4.4%. The return period threshold of rainfall-induced landslides during the landslide recovery period was <2 years, and the landslide types of the new or enlarged landslide were the bank-erosion landslide, headwater landslide, and the reoccurrence of old landslide. The landslide hotspot areas in the Ailiao and Tamali river watershed were 2.67–2.88 times larger after the 2009 Typhoon Morakot using the emerging hot spot analysis, and most of the new or enlarged landslide cases were identified into the oscillating or sporadic or consecutive landslide hotspots. The results can contribute to developing strategies of watershed management in watersheds with a dense landslide.

Application of High-Resolution Radar Rain Data to the Predictive Analysis of Landslide Susceptibility under Climate Change in the Laonong Watershed, Taiwan

Extreme rainfall has caused severe road damage and landslide disasters in mountainous areas. Rainfall forecasting derived from remote sensing data has been widely adopted for disaster prevention and early warning as a trend in recent years. By integrating high-resolution radar rain data, for example, the QPESUMS (quantitative precipitation estimation and segregation using multiple sensors) system provides a great opportunity to establish the extreme climate-based landslide susceptibility model, which would be helpful in the prevention of hillslope disasters under climate change. QPESUMS was adopted to obtain spatio-temporal rainfall patterns, and further, multi-temporal landslide inventories (2003–2018) would integrate with other explanatory factors and therefore, we can establish the logistic regression method for prediction of landslide susceptibility sites in the Laonong River watershed, which was devastated by Typhoon Morakot in 2009. Simulations of landslide susceptibility under the critical rainfall (300, 600, and 900 mm) were designed to verify the model’s sensitivity. Due to the orographic effect, rainfall was concentrated at the low mountainous and middle elevation areas in the southern Laonong River watershed. Landslide change analysis indicates that the landslide ratio increased from 1.5% to 7.0% after Typhoon Morakot in 2009. Subsequently, the landslide ratio fluctuated between 3.5% and 4.5% after 2012, which indicates that the recovery of landslide areas is still in progress. The validation results showed that the calibrated model of 2005 is preferred in the general period, with an accuracy of 78%. For extreme rainfall typhoons, the calibrated model of 2009 would perform better (72%). This study presented that the integration of multi-temporal landslide inventories in a logistic regression model is capable of predicting rainfall-triggered landslide risk under climate change.

A random forest model of landslide susceptibility mapping based on hyperparameter optimization using Bayes algorithm

The choice of model parameters in landslide susceptibility mapping makes a major determinant of model accuracy. The purpose of this study is to optimize the hyperparameters based on a Bayesian optimization algorithm, and to obtain a high accuracy random forest landslide susceptibility evaluation model. The research steps are detailed as follows. Firstly, taking a typical landslide prone mountainous area as an example, 16 conditioning factors, such as elevation, annual average rainfall, distance from roads, distance from buildings and so on, were preliminarily selected as the conditioning factors of landslide susceptibility. Combined with 1520 historical landslide events, a geospatial database was established with 30 m resolution. Secondly, the geospatial data sample set was constructed by random sampling according to ratio of historical landslides and non-landslides of 1:10. Based on the whole sample set, the random forest model adopted the Bayesian optimization algorithm to optimize the hyperparameters. Next, the optimal hyperparameters were selected to be trained to get the evaluation model of landslide susceptibility. In addition, they were carried out the analysis of landslide susceptibility mapping for the whole study area. After that, the recursive feature elimination method was used to screen out the dominant conditioning factors that can explain the degree of landslide susceptibility. The results indicated that the area under curve (AUC) values of receiver operating characteristic (ROC) curve in training data set, verification data set and regional simulation were 0.95, 0.87 and 0.93, respectively. 65% of the historical landslides fell between the high susceptibility and very high susceptibility regions, which made up <20% of the research area. The model was in good agreement to the distribution characteristics of historical landslides in the study area. We noted that all the three recent landslides with impact on the study area occurred at the locations predicted by the model to have high or very high susceptibility in terms of typical landslides in the near future. As for conditioning factors, the contribution related to human activities accounted for a large proportion. In conclusion, an evaluation model with high precision for random forest landslide susceptibility can be built based on hyperparameter optimization with Bayesian optimization algorithm. Simultaneously, using recursive feature elimination method, a random forest landslide susceptibility model with fewer dominant conditioning factors and guaranteed evaluation accuracy can also be built to save the running time and input data resources of the model.

Risk models for assessing the derived disasters caused by watershed landslides using environmental indicators

Lands reserved for indigenous people in Taiwan are mostly located at junctions between forestry compartments and hillside lands. After heavy rains, sediment disasters can easily occur in such locations. Therefore, identifying primary and derived disaster landslide areas of watersheds in lands reserved for indigenous people is imperative. This study used landslide-prone slopes to correct the vulnerability variable in the model proposed by Lin et al. (2017), and recalculated the landslide risks of watersheds. In addition, the study used protected objects as study subjects, and considered the likelihood of landslide disaster occurrence and its scale. Risk models for assessing the derived disasters caused by watershed landslides were constructed according to the spatial distribution of sediment delivery ratio. Comparisons between the present model and the original model regarding the relationship between landslide risks and landslide ratios in watershed subdivisions reveal that the corrected model shows significant positive correlation (R 2 increased from 0.59 to 0.91). Risk models for assessing the derived disasters caused by watershed landslides indicate that the risk and the disaster ratio of the watershed subdivisions show significant and positive correlation (R 2=0.84), indicating that the model developed in this study can effectively estimate the likelihood of protected objects encountering sediment disasters.

Estimating landslide occurrence via small watershed method with relevance vector machine

The mechanism of landslide occurrence is complicated due to the dependency and nonlinear relationship of various physical factors. A promising prediction model, which can be used to locate the high-risk regions and a corresponding prevention strategy can be prepared to reduce the slide occurrence and its consequence, is therefore desired. To perform the landslide assessment for a large-scale slope, this study proposes to use the method of small watershed that is integrated with Relevance Vector Machine (RVM) to enhance the prediction accuracy. Effect of physiographic and hydrological factors such as slope steepness, dip slope ratio, landslide ratio, and cumulative rainfall are investigated. To estimate the occurrence of landslide, RVM first maps the aforementioned factors into a feature space using Gaussian radial basis function. A linear boundary, distinguishing landslide or not, is then obtained through the search of the optimal weights. To find these weights, a Bayesian theory-based optimization problem is formulated and solved by iteratively reweighted least squares algorithm and the Laplace approximation procedure. The proposed model is validated by the data collected from Kaoping River Basin. Results indicate that the proposed RVM-based small watershed approach possesses a prediction accuracy of 87.5%, which is better than those of using Support Vector Machine (SVM), Least-Square Support Vector Machine (LS-SVM), and logistic regression, providing authorities in their hazard alert system to minimize the life or property losses caused by the landslide.

Controls of preferential orientation of earthquake‐ and rainfall‐triggered landslides in Taiwan's orogenic mountain belt

AbstractLandsliding usually occurs on specific hillslope aspect, which may reflect the control of specific geo‐environmental factors, triggering factors, or their interaction. To explore this notion, this study used island‐wide landslide inventories of the Chi‐Chi earthquake in 1999 (MW = 7.6) and Typhoon Morakot in 2009 in Taiwan to investigate the preferential orientation of landslides and the controls of landslide triggers and geological settings. The results showed two patterns. The orientations of earthquake‐triggered landslides were toward the aspect facing away from the epicenter in areas with peak ground acceleration (PGA) ≥ 0.6 g and landslide ratio ≥ 1%, suggesting that the orientations were controlled by seismic wave propagation. Rainfall‐triggered landslides tended to occur on dip slopes, instead of the windward slopes, suggesting that geological settings were a more effective control of the mass wasting processes on hillslope scale than the rainfall condition. This study highlights the importance of the endogenic processes, namely seismic wave and geological settings, on the predesigned orientation of landslides triggered by either earthquake or rainfall, which can in turn improve our knowledge of landscape evolution and landslide prediction. © 2019 John Wiley &amp; Sons, Ltd.

The impact of physiographic factors upon the probability of slides occurrence: a case study from the Kaoping River Basin, Taiwan

ABSTRACTThis study has investigated the influence of physiographic factors upon the probability of slide occurrence in the Kaoping River Basin. According to previous literature, statistical tests, and physical mechanisms, three physiographic factors (slope steepness, the dip slope ratio, and the time-dependent landslide ratio) were significantly related to slide occurrence. These were combined with hydrological factors (cumulative rainfall) to establish an assessment model for estimating the probability of slide occurrence using logistic regression. The model’s overall accuracy in the training and validation stages was about 81%. Overall, 20 randomly selected historical rainfall events were employed for verification, including 10 events each with and without slide occurrence. The results showed that the model accuracy was approximately 80%, if the probability threshold Psh is set to be 0.5.This assessment model can assist in prediction of slide occurrence and the proposed subsequent engineering measures or vegetation restoration is often able to reduce the landslide ratio and the probability of slide occurrence. After landform changes, the model’s physiographic factors can be updated to adjust the rainfall threshold for slide occurrence. Given an appropriate cumulative rainfall and an acceptable risk of slide occurrence, the model can identify priority regions for slide prevention.

THE PERFORMANCE OF LAND USE CHANGE CAUSATIVE FACTOR ON LANDSLIDE SUSCEPTIBILITY MAP IN UPPER UJUNG-LOE WATERSHEDS SOUTH SULAWESI, INDONESIA

The study aims to develop and apply land use change (LUC) performance on landslide susceptibility map using frequency ratio (FR), and Logistic regression (LR) method in a geographic information system. In the study area, Upper Ujung-loe Watersheds area of Indonesia, landslides were detected using field survey and air photography from time series data image of Google Earth Pro from 2012 to 2016 and LUC from 2004 to 2011. Landslide susceptibility map (LSM) was constructed using FR and LR with nine causative factors. The result indicated that LUC affect the production of LSM. Validation of landslide susceptibility was carried out in this study at both with and without LUC causative factors. First, performances of each landslide model were tested using AUC curve for success and predictive rate. The highest value of predictive rate at with LUC in both FR and LR method were 83.4 % and 85.2 %, respectively. In the second stage, the ratio of landslides falling on high to a very high class of susceptibility was obtained, which indicates the level of accuracy of the method.LR method with LUC had the highest accuracy of 80.24 %. Taken together, the results suggested that changing the vegetation to another landscape causes slopes unstable and increases probability to landslide occurrence.

Comparing the performance of TRIGRS and TiVaSS in spatial and temporal prediction of rainfall-induced shallow landslides

This study compares the performance of transient rainfall infiltration and grid-based regional slope stability (TRIGRS) model and time-variant slope stability (TiVaSS) model in the prediction of rainfall-induced shallow landslides. TRIGRS employs one-dimensional (1-D) subsurface flow to simulate the infiltration rate, whereas a three-dimensional (3-D) model is utilized in TiVaSS. The former has been widely used in landslide modeling, while the latter was developed only recently. Both programs are used for the spatiotemporal prediction of shallow landslides caused by rainfall. This study uses the July 2011 landslide event that occurred in Mt. Umyeon, Seoul, Korea, for validation. The performance of the two programs is evaluated by comparison with data of the actual landslides in both location and timing by using a landslide ratio for each factor of safety class ( $${\text{LR}}_{\text{class}}$$ index), which was developed for addressing point-like landslide locations. Moreover, the influence of surface flow on landslide initiation is assessed. The results show that the shallow landslides predicted by the two models are highly consistent with those of the observed sliding sites, although the performance of TiVaSS is slightly better. Overland flow affects the buildup of the pressure head and reduces the slope stability, although this influence was not significant in this case. A slight increase in the predicted unstable area from 19.30 to 19.93% was recorded when the overland flow was considered. It is concluded that both models are suitable for application in the study area. However, although it is a well-established model requiring less input data and shorter run times, TRIGRS produces less accurate results.

Optimal Subdivision for Treatment and Management of Catastrophic Landslides in a Watershed Using Topographic Factors.

Recent extreme rainfall events led to many landslides due to climate changes in Taiwan. How to effectively promote post-disaster treatment and/or management works in a watershed/drainage basin is a crucial issue. Regarding the processes of watershed treatment and/or management works, disaster hotspot scanning and treatment priority setup should be carried out in advance. A scanning method using landslide ratio to determine the appropriate outlet of an interested watershed, and an optimal subdivision system with better homogeneity and accuracy in landslide ratio estimation were developed to help efficient executions of treatment and/or management works. Topography is a key factor affecting watershed landslide ratio. Considering the complexity and uncertainty of the natural phenomenon, multivariate analysis was applied to understand the relationship between topographic factors and landslide ratio in the interested watershed. The concept of species-area curve, which is usually adopted at on-site vegetation investigation to determinate the suitable quadrate size, was used to derive the optimal threshold in subdivisions. Results show that three main component axes including factors of scale, network and shape extracted from Digital Terrain Model coupled with areas of landslide can effectively explain the characteristics of landslide ratio in the interested watershed, and a relation curve obtained from the accuracy of landslide ratio classification and number of subdivisions could be established to derive optimal subdivision of the watershed. The subdivision method promoted in this study could be further used for priority rank and benefit assessment of landslide treatment in a watershed.

Distributions of landslides, vegetation, and related sediment yields during typhoon events in northwestern Taiwan

This study examined landslides caused by typhoon events in the upstream area of the Tao-Cheng River in northwestern Taiwan during a 13-year period, 1996 to 2008. Data from six typhoon events were compiled to analyze the relevant characteristics of landslides, vegetation, and changes in sediment discharge with particulate carbon in rivers in this area. The landslide ratio, the ratio of the landslide area to the total study area, was determined to be between 0.60% and 1.29%, and a high new generation ratio and low reactivated ratio of landslides were noted. Analysis of different bands of satellite images taken over a 10-year period disclosed that the Normalized Difference Vegetation Index (NDVI) in the upstream area of the Tao-Cheng River ranged from 0.47 to 0.63 before typhoon events and from 0.38 to 0.46 after typhoon events. The low landslide ratio and high new generation ratio in this area indicated that sporadic landslides tend to occur in new areas and do not have a great impact on vegetation conditions. Thus, the decrease in NDVI after typhoon events was caused by the seasonal effect with withering of vegetation. On the other hand, landslides in this area tend to occur on nonforest land because of the interventions of external forces, such as human developments and reclamation.Geological materials produced by landslides and vegetation flushed by heavy rainfall during typhoon periods have a great effect on the amount of sediment discharge and particulate carbon discharge in the rivers. Furthermore, during typhoon periods, the amount of accumulated rainfall has a greater impact than the peak rainfall intensity on the increase of sediment and particulate carbon in rivers and on the decrease in the NDVI. Analysis of the particulate carbon from the rivers in this area revealed that in the rainy season, high strength and accumulated rainfall will lead to more sediment discharge and particulate carbon discharge in rivers. However, the particulate carbon is limited with further increases in rainfall because of the limited supply of particulate carbon in the upstream of the Tao-Cheng River.

Landslide susceptibility mapping by using landslide ratio-based logistic regression: A case study in the southern Taiwan

The object of the research is to compare the model performance and explain the error source of original logistic regression landslide susceptibility model (abbreviated as or-LRLSM) and landslide ratio-based logistic regression landslide susceptibility model (abbreviated as lr-LRLSM) in the Chishan watershed with a serious landslide disaster after 2009 Typhoon Morakot. The landslide inventory induced by 2009 Typhoon Morakot in South Taiwan is the main research material, while the Chishan watershed is the research area. Six variables, including elevation, slope, aspect, geological formation, accumulated rainfall, and bank erosion, were included in the two models. The performance of lr-LRLSM is better than that of or-LRLSM. The Cox & Snell R2, Nagelkerke R2 value, and the area under the relative operating characteristic curve (abbreviated as AUC) of lr-LRLSM is larger than those of or-LRLSM, and the average correct ratio for the lr-LRLSM to predict landslide or non-landslide is larger than that of or-LRLSM by 5.0%. The increase of the average correct ratio (abbreviated as ACR) difference from or-LRLSM to lr-LRLSM shows in slope, revised accumulated rainfall, aspect, geological formation and bank erosion variables, and only light decreases in elevation variable. The error sources of continuous variables in building the or-LRLSM is the dissimilarity between the distribution of landslide ratio and production of coefficient and characteristic values, while those of categorical variables is due to low correlation of landslide ratio and the coefficient value of each parameter. Using the classification of landslide ratio as the database to build logistic regression landslide susceptibility model (abbreviated as LRLSM) can revise the errors. The comparison of or-LRLSM and lr-LRLSM in the Chishan watershed also shows that building the landslide susceptibility model (abbreviated as LSM) by using lr-LRLSMis practical and of better performance than that by using the or-LRLSM.

Effects of landslide and other physiographic factors on the occurrence probability of debris flows in central Taiwan

This study focuses on the effects of physiographic factors, such as landslide area, on the occurrence of debris flows in central Taiwan. Four physiographic factors were selected for their significance to the occurrence of debris flows, including landslide ratio (the ratio of landslide area over watershed area), average steepness of the streambed, effective watershed area, and form factor. Two quantifying methods of factors were performed and compared: one using genuine values of factors and another one using values converted by degree of membership from fuzzy theory. Then the logistic regression method was applied for building a model to assess the occurrence probability of debris flows from five variables: four physiographic factors and one hydrologic factor. The model is consistent with the mechanism of debris flow occurrence, with all physiographic and hydrologic factors positively correlated with the occurrence probability. In addition, the accuracy of the model was validated with randomly selected historical events and demonstrated fairly satisfactory validity, ranging from 70 to 80 %. It was found that adopting the degree of membership made the model more stable and more reliable. In addition, the model also shows that reducing the landslide area can significantly reduce the occurrence probability of debris flows. The results show that the model built in this study has the potential to be well applied and fully integrated into current or future warning systems.