Estimating rainfall threshold and temporal probability for landslide occurrences in Darjeeling Himalayas

Abstract

The Indian Himalayan region has been severely affected by landslides causing an immense loss in terms of human lives and economic loss. The landslides are usually induced by rainfall which can be slow and continuous or heavy downpour. The incidences of landslide events in Indian Himalayas have been further aggravated due to the rapid increase in urbanization and thus its increasing impact on socio-economic aspects. There is a dire need for understanding landslide phenomena, estimating its occurrence potential and formulating strategies to minimize the damage caused by them. One of the most affected area is Kalimpong of Darjeeling Himalayas where significant studies have been conducted on zonation, threshold estimation and other related aspects. However, a comprehensive study in terms of temporal prediction for this region remains unattended. The paper deals with assessing landslide hazard using a rainfall threshold model involving daily and cumulative antecedent rainfall values for landslide events. The threshold values were determined using daily rainfall and antecedent rainfall using precipitation and landslide records for 2010–2016. The results show that 20-day antecedent rainfall provides the best fit for landslide occurrences in the region. The rainfall thresholds were further validated using rainfall and landslide data of 2017, which was not considered for threshold estimation. Finally, the results were used to determine the temporal probability for landslide incidence using a Poisson probability model. The validated results suggest that the model has the potential to be used as a preliminary early warning system.