การพัฒนาแผนที่เสี่ยงภัยดินถล่มเชิงพลวัตที่เกิดจากการกระตุ้นของปริมาณน้ำฝนบนลาดภูเขาในภาคเหนือของประเทศไทย

Abstract

This objective of study was to develop a landslide risk estimate model and develop dynamic rainfall-induced landslide risk map on mountain slopes in Northern Thailand. Two study areas were selected for landslide risk mapping: Phu Chi Fa, Tub Tao Subdistrict, Thoeng District and Doi Mae Salong, Mae Salong Nok Subdistrict, Mae Fah Luang District in Chiang Rai which is an area of important tourist attractions on mountain slopes and risk to landslides. This study was to create and develop a risk estimate model using the artificial neural network (ANN) technique for landslides by collecting field data on past landslides in the study areas in Chiang Rai and Chiang Mai Provinces. The variables for forecasting were 1) types of land cover, 2) physiographic, 3) slope angle, and 4) five-day cumulative rainfall. Two hidden layers were used to create the model. The number of nodes in the first and next hidden layers were five and one respectively, which were derived from a total of 25 trials, and the highest accuracy was 96.74%. After that, the model was applied together with Rainfall Depth – Duration – Frequency Curve (DDF Curve) which was obtained from the analysis. This resulted in the output data as a numerical landslide risk value of the study area, which can be used to generate landslide risk maps at return period 2, 5, 10, 20, 50 and 100 years, respectively. To be a database for dissemination to relevant local authorities and people who live in the study area can use the information to monitor disasters and plan preventive measures from landslides that will happen in the future.