Spatio-temporal evolution of landslides along transportation corridors of Muzaffarabad, Northern Pakistan

Landslide spatio-temporal distribution is an effective approach to understand the landslide mechanism and triggering factors. However, the quantitative characterization of the spatio-temporal distribution of landslides along with their causative factors remains a critical challenge for geoscientists due to limited historical landslide records. This study presents the landside spatio-temporal distribution analysis by developing landslide inventories from World view-3, SPOT-5, Quick Bird and Google Earth imageries and verified in the field through extensive field visits along main road corridors (i.e., Neelum Road, Jhelum valley road and Kohala road) of the Muzaffarabad district. Past landside records of 15 years were collected, and temporal inventories were prepared for 2005, 2007, 2012, 2015, and 2019 years. Based on the research activities, the landslide spatial variations were traced and analyzed to classify them into fall and slide types. The temporal analysis of the landslides was then compared with eleven causative factors, i.e., slope, aspect, surface relief, curvature, lithology, distance to roads, faults and streams, land use, Topographic Wetness Index (TWI) and Normalized Differential Vegetation Index (NDVI). The spatio-temporal analysis demonstrates that the total number of landslides along the selected road corridors are 107, 164, 169, 92, and 182 during the years 2005, 2007, 2012, 2015, and 2019, respectively. The analysis of the study area reveals that the landslide area and events depict an abrupt increase in 2005, 2007, 2012, and 2019 whereas a significant decrease in landslide area and events were recorded during 2015. The present research concluded that the combined effect of topographic factors (slope, aspect, elevation and curvature), lithology and distance to roads have found significant influence on the landslide phenomenon. This variation in landslide areas and events indicates the influence of causative factors with respect to time. The present work will be helpful to understand the spatial patterns, trends over the years and landslide triggering mechanisms.