Volume 8, Issue 16 (2-2018)                   jwmr 2018, 8(16): 170-177 | Back to browse issues page

XML Persian Abstract Print

Abstract:   (701 Views)
Knowing the number, area, and frequency of landslides occurred in each area has a prominent role in the long-term evolution of area dominated by landslides and can be used for analyzing of susceptibility, hazard, and risk. In this regard, the current research is trying to consider identified landslides size probability in the Pivejan Watershed, Razavi Khorasan Province. In the first step, landslides inventory map was created using Google Earth images and extensive field surveys. In the next step, area of each landslide was determined using ArcGIS software and Xtools Extension. Subsequently, probability of landslides size identified were calculated in R statistical software using Double Pareto (DP), Double Pareto Simplified (DPS), and Inverse Gamma (IG) probability density functions in the study area. Also, in the present study for optimization of parameters coefficients used of two non-parametric probability density function namely Histogram Density Estimation (HDE) and Kernel Density Estimation (KDE) and a parametric Maximum Likelihood (ML) estimation. The results showed that non-parametric estimation methods (i.e., HDE and KDE) provided accurate results for all the landslides; whereas, ML failed to provide a good result. Also, results of landslide occurrence probability showed a good similarity between DPS and IG with different optimization techniques, meanwhile the DP model had under estimation results and can’t presented a correct calculation for probability of landslides in the study area.
Full-Text [PDF 1746 kb]   (216 Downloads)    
Type of Study: Research | Subject: Special
Received: 2018/01/30 | Revised: 2018/02/24 | Accepted: 2018/01/30 | Published: 2018/01/30