Generating landslide inventory by participatory mapping: an example in Purwosari Area, Yogyakarta, Java

Abstract

This paper proposes an approach for landslide inventory mapping considering actual conditions in Indonesia. No satisfactory landslide database exists. What exists is inadequate, focusing, on data response, rather than on pre-disaster preparedness and planning. The humid tropical climate also leads a rapid vegetation growth so past landslides signatures are covered by vegetation or dismantled by erosion process. Generating landslide inventory using standard techniques still seems difficult. A catalog of disasters from local government (village level) was used as a basis of participatory landslide inventory mapping. Eyewitnesses or landslide disaster victims were asked to participate in the reconstruction of past landslides. Field investigation focusing on active participation from communities with the use of an innovative technology was used to verify the landslide events recorded in the disaster catalog. Statistical analysis was also used to obtain the necessary relationships between geometric measurements, including the height of the slope and length of run out, area and volume of displaced materials, the probability distributions of landslide area and volume, and mobilization rate. The result shows that run out distance is proportional to the height of the slope. The frequency distribution calculated by using non-cumulative distribution empirically exhibits a power law (fractal statistic) even though rollover can also be found in the dataset. This cannot be the result of the censoring effect or incompleteness of the data because the landslide inventory dataset can be classified as having complete data or nearly complete data. The so-called participatory landslide inventory mapping method is expected to solve the difficulties of landslide inventory mapping and can be applied to support pre-disaster planning and preparedness action to reduce the landslide disaster risk in Indonesia. It may also supplement the usually incomplete data in a typical landslide inventory.