Landslide Susceptibility Mapping Using Machine Learning in Himalayan Region: A Review

Landslides, among all, are hazardous natural disaster events. It is caused due to the creation of instability in the slopes of the terrain. In the past 20 to 30 years’ time span it has been observed that the frequency of landslides has increased tremendously due to urbanization and climate change. The landslide susceptibility maps could be used for planners to mitigate the destruction being caused by landslides beforehand. By this review article we want to put emphasis on use of new and advanced learning algorithms for preparation of landslide susceptibility maps. Few machine learning techniques like Random Forest (RF) artificial neural network (ANN) and Support vector machine (SVM) etc. have been giving very accurate maps for prediction of future natural hazard events. Himalayan region being topographically unstable has many complexities in doing study for a region. But based on the reviews of different studies it has been found that machine learning algorithms are very accurate in addressing the challenges being faced and gives vulnerable areas of landslide events through susceptibility maps. Ensemble techniques of different methodologies were found to be improving the accuracy of the landslide susceptibility assessment, but not many researches have done it. One of the most challenging tasks is to find the inventories of the landslide events. For which many organizations are working hard with the help of researchers to make a database for it. Still few researches are successful in prediction of temporal and spatial events of future landslides. Machine learning can act as a cutting edge in prevention of landslide vents ahead of time. This article tries to present the recent advancement of machine learning in the field of landslide susceptibility mapping. It also suggests a future perspective of more machine learning techniques like XGBoost that can be implemented in rough and complex topographic terrains of Himalayas.