Landslide susceptibility mapping of Rangamati District of Bangladesh using statistical and machine intelligence model

A review on landslide susceptibility mapping research in Bangladesh

“…Due to severe landslides in 2017, Rangamati metropolitan area got attention for studying landslide susceptibility [ 6 , 18 , [45] , [46] , [47] ]. Two papers were found studying the whole Rangamati district [ 6 , 48 ] and one paper was found studying the whole Chattogram district [ 49 ] for preparing the landslide susceptibility map ( Fig. 7 ).…”

“…Ahmed et al (2020a) [ 51 ] did not mention the inventory data source. Sifa et al (2020) [ 18 ] used SAR offset tracking technique, Khatun et al (2022) [ 6 ] used Google Earth images with sphere identification and authentication and Hafsa et al (2022) [ 48 ] used both google image interpretation and field survey. Rabby and Li (2019) [ 20 ] prepared a landslide inventory database for the whole hilly areas using used both google image interpretation and field survey.…”

“… The most frequently used ratio is 75% training and 25% testing data [ 18 , 45 , 46 , 48 ] followed by 80% training and 20% testing data [ 8 , 47 , 51 ] and 70% training and 30% testing data [ [55] , [56] , [57] ]. Adnan et al (2020) [ 54 ] used 60%/40% and Rahman et al (2017) [ 58 ] used 60%/40% data for training and testing purposes, respectively ( Table 1 ).…”

“…12 ). In some papers, instances of model integration was found such as frequency ratio integrated with logistic regression [ 8 , 48 ] and analytical hierarchy process [ 8 ] and Dempster-Shafer method integrated with the weight of evidence method [ 7 ].

Fig.

…”

Modelling hydrological factors from DEM using GIS

“…Researchers across the world measured a good agreement of landslides and DEM processed topographical, morphological and hydrological factors [ [7] , [8] – 9 ] such as slope aspect, slope angle, elevation, plan curvature, profile curvature, general curvature, topographic wetness index (TWI), stream power index (SPI), sediment transport index (STI), topographic roughness index (TRI), topographic position index (TPI), distance to stream density etc. In landslide susceptibility mapping and assessment, these factors are used as independent variables whereas landslides are used as dependant variables [ 10 , 11 ]. The practice of the use of these factors has increased due to the availability of open-source DEM data of the globe.…”