University of Twente Student Theses


The Legacy Effect of the 2005 Kahsmir Earthquake on Post-Seismic Landslide Susceptibility during the Western Monsoon Season in Northern Pakistan

Hielkema, Laura Louise (2021) The Legacy Effect of the 2005 Kahsmir Earthquake on Post-Seismic Landslide Susceptibility during the Western Monsoon Season in Northern Pakistan.

Link to full-text:
(only accessible for UT students and staff)
Abstract:Landslides are frequently occurring hazards and pose a signifcant threat to people and property. Especially in mountainous regions, landslide risk is prominent. Both earthquakes and rainfall are important triggers of landslides. This research combines these two triggers by assessing whether dynamic landslide susceptibility analysis based on a combination of post-seismic and rainfall-induced increases in landslide susceptibility in northern Pakistan can inform early action to reduce the impact of those landslides. More specifcally, this study examines whether it would have been possible to utilise the time between the 2005 Kashmir earthquake and the first western monsoon season after the earthquake to have predicted where the landslides occurred. Therefore, this study explores a new approach to landslide risk reduction by assessing the possibility of using earthquake parameters to predict the spatial variability of landslides in hopes of reducing the impacts of landslides. This is done using Bayesian versions of a Generalised AdditiveModel (GAM) to estimate landslide susceptibility before, during and after the 2005 Kashmir earthquake. To do so, a pre-seismic, a co-seismic and a post-seismic landslide inventory were developed using ASTER satellite images. These were used to assess whether including ground motion parameters in the GAMs show elevated landslide susceptibility in areas and could potentially be applied for early action in the study area to reduce landslide risk during upcoming western monsoon seasons. Contrary to regular landslide predictive modelling, no separate training and validation data sets were used to assess the predictive capacity of the landslide models. Instead, the explanatory models are based on the same parameters per slope unit. For the post-seismic landslide models, the only difference is the inclusion or exclusion of the ground motion parameters. This was done to merely examine the possible benefits of earthquake information on the identification of locations prone to post-seismic landslides. Three post-seismic GAMs were conducted, one including earthquake ground motion parameters, one excluding them and one with merely the earthquake parameters. The last one was done to evaluate to what extent the earthquake parameters could accurately predict land sliding. For each of the GAMs, the influence of the landslide controlling parameters on landslide occurrence was analysed and compared to assess whether the effects of the earthquake parameters could potentially inform early action to reduce landslide risk. Surprisingly, the fixed and random effects of the earthquake parameters in the GAMs showed no signifcant influence on landslide occurrence. The post-seismic GAMs including and excluding the earthquake parameters were very similar, and both were similarly accurate. The resulting susceptibility maps for the study area showed only minor differences in susceptibility. Because of minor susceptibility differences, this approach has not been adequate to develop effective early action strategies. Therefore, the focus should be on other landslide risk reduction strategies to reduce the impacts of landslide risk in the study area and elsewhere.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Spatial Engineering MSc (60962)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page