We prepared a landslide susceptibility map for the Sarkhoon watershed, Chaharmahal-w-bakhtiari, Iran, using novel ensemble artificial intelligence approaches. A classifier of support vector machine ...(SVM) was employed as a base classifier, and four Meta/ensemble classifiers, including Adaboost (AB), bagging (BA), rotation forest (RF), and random subspace (RS), were used to construct new ensemble models. SVM has been used previously to spatially predict landslides, but not together with its ensembles. We selected 20 conditioning factors and randomly portioned 98 landslide locations into training (70%) and validating (30%) groups. Several statistical metrics, including sensitivity, specificity, accuracy, kappa, root mean square error (RMSE), and area under the receiver operatic characteristic curve (AUC), were used for model comparison and validation. Using the One-R Attribute Evaluation (ORAE) technique, we found that all 20 conditioning factors were significant in identifying landslide locations, but “distance to road” was found to be the most important. The RS (AUC = 0.837) and RF (AUC = 0.834) significantly improved the goodness-of-fit and prediction accuracy of the SVM (AUC = 0.810), whereas the BA (AUC = 0.807) and AB (AUC = 0.779) did not. The random subspace based support vector machine (RSSVM) model is a promising technique for helping to better manage land in landslide-prone areas.
We used a novel hybrid functional machine learning algorithm to predict the spatial distribution of landslides in the Sarkhoon watershed, Iran. We developed a new ensemble model which is a ...combination of a functional algorithm, stochastic gradient descent (SGD) and an AdaBoost (AB) Meta classifier namely ABSGD model to predict the landslides. The model incorporates 20 landslide conditioning factors, which we ranked using the least-square support vector machine (LSSVM) technique. For the modeling, we considered 98 landslide locations, of which 70% (79) were used for training and 30% (19) for validation processes. Model validation was performed using sensitivity, specificity, accuracy, the root mean square error (RMSE) and the area under the receiver operatic characteristic (AUC) curve. We also used soft computing benchmark models, including SGD, logistic regression (LR), logistic model tree (LMT) and functional tree (FT) algorithms for model validation and comparison. The selected conditioning factors were significant in landslide occurrence but distance to road was found to be the most important factor. The ABSGD model (AUC= 0.860) outperformed the LR (0.797), SGD (0.776), LMT (0.740) and FT (0.734) models. Our results confirm that the combined use of a functional algorithm and a Meta classifier prevents over-fitting, reduces noise and enhances the power prediction of the individual SGD algorithm for the spatial prediction of landslides.
Mapping flood-prone areas is a key activity in flood disaster management. In this paper, we propose a new flood susceptibility mapping technique. We employ new ensemble models based on bagging as a ...meta-classifier and K-Nearest Neighbor (KNN) coarse, cosine, cubic, and weighted base classifiers to spatially forecast flooding in the Haraz watershed in northern Iran. We identified flood-prone areas using data from Sentinel-1 sensor. We then selected 10 conditioning factors to spatially predict floods and assess their predictive power using the Relief Attribute Evaluation (RFAE) method. Model validation was performed using two statistical error indices and the area under the curve (AUC). Our results show that the Bagging–Cubic–KNN ensemble model outperformed other ensemble models. It decreased the overfitting and variance problems in the training dataset and enhanced the prediction accuracy of the Cubic–KNN model (AUC=0.660). We therefore recommend that the Bagging–Cubic–KNN model be more widely applied for the sustainable management of flood-prone areas.
Shallow landslides damage buildings and other infrastructure, disrupt agriculture practices, and can cause social upheaval and loss of life. As a result, many scientists study the phenomenon, and ...some of them have focused on producing landslide susceptibility maps that can be used by land-use managers to reduce injury and damage. This paper contributes to this effort by comparing the power and effectiveness of five machine learning, benchmark algorithms-Logistic Model Tree, Logistic Regression, Naïve Bayes Tree, Artificial Neural Network, and Support Vector Machine-in creating a reliable shallow landslide susceptibility map for Bijar City in Kurdistan province, Iran. Twenty conditioning factors were applied to 111 shallow landslides and tested using the One-R attribute evaluation (ORAE) technique for modeling and validation processes. The performance of the models was assessed by statistical-based indexes including sensitivity, specificity, accuracy, mean absolute error (MAE), root mean square error (RMSE), and area under the receiver operatic characteristic curve (AUC). Results indicate that all the five machine learning models performed well for shallow landslide susceptibility assessment, but the Logistic Model Tree model (AUC = 0.932) had the highest goodness-of-fit and prediction accuracy, followed by the Logistic Regression (AUC = 0.932), Naïve Bayes Tree (AUC = 0.864), ANN (AUC = 0.860), and Support Vector Machine (AUC = 0.834) models. Therefore, we recommend the use of the Logistic Model Tree model in shallow landslide mapping programs in semi-arid regions to help decision makers, planners, land-use managers, and government agencies mitigate the hazard and risk.
We used three state-of-the-art machine learning techniques (boosted regression tree, random forest, and support vector machine) to produce a multi-hazard (MHR) map illustrating areas susceptible to ...flooding, gully erosion, forest fires, and earthquakes in Kohgiluyeh and Boyer-Ahmad Province, Iran. The earthquake hazard map was derived from a probabilistic seismic hazard analysis. The mean decrease Gini (MDG) method was implemented to determine the relative importance of effective factors on the spatial occurrence of each of the four hazards. Area under the curve (AUC) plots, based on a validation dataset, were created for the maps generated using the three algorithms to compare the results. The random forest model had the highest predictive accuracy, with AUC values of 0.994, 0.982, and 0.885 for gully erosion, flooding, and forest fires, respectively. Approximately 41%, 40%, 28%, and 3% of the study area are at risk of forest fires, earthquakes, floods, and gully erosion, respectively.
Glacial lake outburst floods (GLOFs) are highly mobile mixtures of water and sediment that occur suddenly and are capable of traveling tens to hundreds of kilometers with peak discharges and volumes ...several orders of magnitude larger than those of normal floods. They travel along existing river channels, in some instances into populated downstream regions, and thus pose a risk to people and infrastructure. Many recent events involve process chains, such as mass movements impacting glacial lakes and triggering dam breaches with subsequent outburst floods. A concern is that effects of climate change and associated increased instability of high mountain slopes may exacerbate such process chains and associated extreme flows. Modeling tools can be used to assess the hazard of potential future GLOFs, and process modeling can provide insights into complex processes that are difficult to observe in nature. A number of numerical models have been developed and applied to simulate different types of extreme flows, but such modeling faces challenges stemming from a lack of process understanding and difficulties in measuring extreme flows for calibration purposes. Here we review the state of knowledge of key aspects of modeling GLOFs, with a focus on process cascades. Analysis and simulation of the onset, propagation, and potential impact of GLOFs are based on illustrative case studies. Numerical models are presently available for simulating impact waves in lakes, dam failures, and flow propagation but have been used only to a limited extent for integrated simulations of process cascades. We present a spectrum of case studies from Patagonia, the European Alps, central Asia, and the Himalayas in which we simulate single processes and process chains of past and potential future events. We conclude that process understanding and process chain modeling need to be strengthened and that research efforts should focus on a more integrative treatment of processes in numerical models.
This review examines interpretive issues relating to catastrophic, long-runout landslides in the context of large numbers of recently discovered late Quaternary events. It links relevant research in ...landslide science, including some novel or hitherto-ignored complexities in the nature and role of these events, to broader concerns of mountain geomorphology. Attention is drawn to mountain ranges known to have large concentrations of events. In particular, discoveries in three regions are singled out; the Karakoram Himalaya, the coastal mountains of northwestern North America, and the Southern Alps of New Zealand. In each region, many new events, or previously unrecognized complexities, have been identified in the past decade or two. Research on the sedimentology and geomorphology of prehistoric, eroded deposits has been critical to identifying rock avalanches, including many that were formerly attributed to other processes. Discoveries of rock avalanches in the ancient stratigraphic record have helped with the field recognition of rock-avalanche materials and in developing facies models of deposits with complex emplacement histories. The stratigraphic record also provides insights into interactions of streaming rock debris with deformable substrates. Such interactions are responsible for “landslide-tectonized” forms and transformation of rock avalanches into debris flows. Of special interest are runout geometries involving the interactions of rock avalanches with topography or substrate materials, and travel over glaciers. Other emerging issues relate to reconstruction of detachment-zone geometries, and slow, deep-seated slope movements that may trigger catastrophic failure. Most previous landslide studies have focused on individual events or general models, whereas the questions addressed here arise from a comparative approach emphasizing common and contrasting features among events in sets and in different regions. The scale and frequency of landslides in the regions of interest mean they have an important role in denudation, regional landform development, watershed evolution, and Quaternary environmental change. A major developmental factor, largely neglected, is persistent disturbance of high mountain fluvial systems by many successive landslides. Damming of streams and subsequent breaching of landslide barriers strongly influence inter-montane sedimentation and denudation, with particular significance in post-, para-, and inter-glacial contexts. Although an individual landslide appears as a “catastrophe” lasting only a minute or two, its legacy can persist as a morphogenetic influence for millennia or tens of millennia through disturbance of other processes. The influence is permanently felt; in effect, multiple events make the event a “normal” one in regions such as the three considered here.
Loess covers approximately 6.6% of China and forms thick extensive deposits in the northern and northwestern parts of the country. Natural erosional processes and human modification of thick loess ...deposits have produced abundant, potentially unstable steep slopes in this region. Slope deformation monitoring aimed at evaluating the mechanical behavior of a loess slope has shown a cyclic pattern of contraction and expansion. Such cyclic strain change on the slope materials can damage the loess and contribute to slope instability. The site showing this behavior is a 70-m high loess slope near Yan'an city in Shanxi Province, northwest China. A Ground-Based Synthetic Aperture Radar (GB-SAR) sensor and a displacement meter were used to monitor this cyclic deformation of the slope over a one-year period from September 2018 to August 2019. It is postulated that this cyclic behavior corresponds to thermal and moisture fluctuations, following energy conservation laws. To investigate the validity of this mechanism, physical models of soil temperature and moisture measured by hygrothermographs were used to simulate the observed cyclic deformations. We found good correlations between the models based on the proposed mechanism and the exhibited daily and annual cyclic contraction and expansion. The slope absorbed energy from the time of maximum contraction to the time of maximum expansion, and released energy from the time of maximum expansion to the time of maximum contraction. Recoverable cyclic deformations suggest stresses in the loess are within the elastic range, and non-recoverable cyclic deformations suggest damage of the loess material (breakage of bonds between soil grains), which could lead to instability. Based on these observations and the models, we developed a quantitative relationship between weather cycles and thermal deformation of the slope. Given the current climate change projections of temperature increases of up to 3.5 °C by 2100, the model estimates the loess slope to expand about 0.35 mm in average, which would be in addition to the current cyclic "breathing" behavior experienced by the slope.
Episodic failures of ice-dammed lakes have produced some of the largest floods in history, with disastrous consequences for communities in high mountains
. Yet, estimating changes in the activity of ...ice-dam failures through time remains controversial because of inconsistent regional flood databases. Here, by collating 1,569 ice-dam failures in six major mountain regions, we systematically assess trends in peak discharge, volume, annual timing and source elevation between 1900 and 2021. We show that extreme peak flows and volumes (10 per cent highest) have declined by about an order of magnitude over this period in five of the six regions, whereas median flood discharges have fallen less or have remained unchanged. Ice-dam floods worldwide today originate at higher elevations and happen about six weeks earlier in the year than in 1900. Individual ice-dammed lakes with repeated outbursts show similar negative trends in magnitude and earlier occurrence, although with only moderate correlation to glacier thinning
. We anticipate that ice dams will continue to fail in the near future, even as glaciers thin and recede. Yet widespread deglaciation, projected for nearly all regions by the end of the twenty-first century
, may bring most outburst activity to a halt.