Landslides are one of the deadliest and most frequent dangers in hilly areas. They frequently occur without warning and cause damage and human casualties. The local topography and the influence of geological and geomorphological processes determine the frequency of landslides. However, external forces including excessive rainfall, earthquakes, flooding, snow melting, stream erosion, changes in ground water level, or any combination of these natural phenomena can cause landslides on unstable slopes [1–5]. Previous research has also shown that human activity, including the built-up areas, deforestation, forest-fire, shifting agriculture, and shoddy road construction, increases the frequency and magnitude of landslides in many hilly or mountainous areas [6–10]. Therefore, reducing the likelihood of landslide occurrences on unstable slopes and evaluating the short- and long-term harmful effects of landslides on the natural landscape require an interdisciplinary approach.
The Himalaya, western Ghats, and the eastern Ghats are among the hilly regions of India that are most vulnerable to landslides, accounting for over 12.6% of the country's land area [11]. Landslides commonly occur in the youngest mountain chain of the world i.e. the Indian Himalayan Region (IHR) as a result of internal elements such surface drainage groundwater, high relief, steep slopes, higher altitude, unstable soil, and lithology setting. But prior research also highlights that landslides were further exacerbated in many potentially unstable slope areas by external factors like intense rain, earthquakes, and human activities like deforestation, shifting agriculture, road construction, and agricultural expansion [12–16].
The landslide susceptibility map is a useful technique for managing the risk of landslides. Under the presumption that landslides would continue to occur under the same circumstances as they have in the past, these maps can be produced using spatial prediction of landslides [17]. Thus, landslide susceptibility can be evaluated by examining the spatial relationship between a group of causative factors and previous landslide incidents. Recently, a large number of landslide susceptibility maps have been developed in various parts of the world using Geographic Information Systems (GIS). Landslide vulnerability is heightened by a combination of factors such as geological conditions, land use patterns, and climate change impacts, posing significant risks to communities worldwide (UNDRR, 2019; Guzzetti et al., 2020). Understanding these dynamics in vulnerable regions is crucial for effective risk mitigation strategies and risk assessment. Currently, the most often used method for assessing landslide vulnerability is the statistical technique, which is a subjective method. This strategy has been used in a variety of techniques, including logistic regression [20], weights of evidence [19], the information value method [23], frequency ratio [18]. The frequency ratio (FR) approach is one of these techniques that is most frequently used for assessing landslide vulnerability and has good performance [17].
The present work aims to create a landslide susceptibility and risk probability map for Garhwal Himalaya, Uttarakhand, India through FR model. The performance of the FR model was evaluated using the ROC-AUC curve.
The study on landslide risk probability assessment is very limited in the Garhwal Himalaya. Therefore present research work is carried out using Normalized Difference Built-up Index (NDBI) and census data of 2011 (recent data of 2021 is awaited).