LANDSLIDE HAZARDS ZONATION OF CHITRAL AREA, NW PAKISTAN USING GEOSPATIAL TOOLS

Abstract

The Chitral region of Hindu Kush Karakoram mountainous terrain is more prone to natural disasters i.e. landslide, flash floods, glacial lake outburst floods and debris flow, which alter the landscape and damage the property. This study has been carried out in two section (i) initial mapping using Geographic Information System (GIS) and (ii) detection of unstable (slow moving) hillslope using Interferometric Synthetic Aperture Radar (InSAR) techniques to map landslides in Chitral. Initially landslides were mapped in the field, followed by inventory mapping. In first phase two approaches were implemented i.e. Analytical Hierarchy Process and Frequency Ratio (AHP and FR) to construct landslide susceptibility maps via (GIS). Analytical Hierarchy Process is one of the most comprehensive system is measured to make decisions with several criteria because this technique gives to express the problem as a hierarchical and believe a mixture of quantitative and qualitative criteria as well. Wheras, FR is an approach that can display association between the distribution of landslides and different causing factors. Nine main contributing factors were considered to generate landslide susceptibility map of whole Chitral which includes, topographical parameters (slope, aspect and landcover), geological parameters in term of (lithology, distance to fault) hydrological parameters (rainfall and distance to stream), distance to roads and seismological parameter. Results suggested that majority of the landslides triggering in the study area is the result of steep slope of mountains, followed by the precipitation, lithology and earthquake. AHP weights with consistency ratio 0.04 were used in the preparation of final susceptibility map and classified into very high susceptibility zone, high, moderate, low and very low susceptible classes. Furthermore one case study was carried out in Reshun area located in the central Chitral, which is frequently subjected to cataclysmic landslides. Two techniques AHP coupled with FR techniques were also carried out in comparative system. Seven landslide causing factors, (i.e., slope, lithology, land cover, aspect, and distance to stream, road, and fault) were investigated for the susceptibility assessment of Reshun area. Using the AHP and FR methods the historical landslides were linked to the causing factors to compute their importance on landslide susceptibility. A susceptibility map was extracted from each method, classified into five classes, and validated by area under curve (AUC) metric. Results revealed that FR with AUC=0.84 performed better than AHP with AUC=0.78 in zoning the study areas with respect to historical landslides. The FR method classified 18% of the area as having very high susceptibility to landslides, 27% as high susceptibility, 28% as medium susceptibility, 18% as low susceptibility, and 9% as very low susceptibility. These maps would support more informed land-use planning in the region. In addition, geoscientific community has recently investigated viii the effects of climate change in high mountainous regions has mostly looked at the receding process of glaciers, permafrost thawing and snowmelt which initiate landslide hazards. In addition, aside from the inventory mapping, a valuable contribution to understand this complex system could come from the use of Interferometric Synthetic Aperture Radar (InSAR) techniques. In this work, more focus is given to investigate an appalling situation where an extremely large landslide (~20 km2) is slowly moving, while hosting a village on the surface of it body and overlooking another village right next to its toe. Our InSAR analyses span over the last six years and allowed us to retrieve the spatio-temporal pattern of downslope deformation for both ascending and descending orbits of Sentinel-1 data. The results highlight a worrying situation where the crown of the large landslide is moving fast. Sackung type of features and other signs of instability there are diffused all over the crown. As for the toe of the large landslide, a western sliding sector offers a different mechanism compared to its eastern counterpart where the deformation appears to being accumulating through time. Furthermore the study area was extended to Karimabad valley where Karimabad village is also exposed to hillslope moving threats. For this section, this work analyzed the Sentinel-1 data in two sets (Ascending and Descending) for the period of three years. Results from the data revealed that the hillslopes are moving with an average velocity of -25 to +28 mm/yr. Seven deformation zones were identified over the deformation map which reflects an alarming situation for the Karimabad valley especially for Beshgram village. Finally the deformation time series were then correlated with rainfall, earthquake and snowfall data which reveals that the area experiencing a variated deformation due to the combination of several causative factors acting together, predominantly, topography/steep slopes, freezing and thawing of snow and seismic events. This brief description has two implications. One from the most practical perspective, as it calls for further studies and attention from local administrations. And, it also scientifically highlight the strength of InSAR when it comes to unveil slow deformation regimes, which may be invisible through the eye of other techniques, although they may still lead to catastrophic failures.