Saint Vincent & Grenadines - Landslide Susceptibility - CHARIM

The aim of this study was to generate a national‐scale landslide susceptibility map for the island of Saint Vincent, and not for the other islands (Grenadines) that are part of Saint Vincent and the Grenadines (SVG). As the available data turned out to be insufficient to generate reliable results, we decided to generate several new data layers and significantly improved some of the existing data. We generated a new database of hydrometeorological disaster events for Saint Vincent, for the period 1718 to 2014, making use of many different sources.

This is the most complete inventory up to our knowledge. It is quite clear from this database that the disaster reporting became more frequent in recent years, although limited information on landslides is available for Saint Vincent, even more when going back in time, whereas the data on tropical storms and hurricanes seems to be much more constant over time. The underreporting of landslides is a big problem in trying to evaluation landslide frequency/magnitude relations, which are required to convert susceptibility maps into hazard maps. We also compiled all available landslide data from different sources. Unfortunately there was only one inventory available, generated in 1988 by DeGraff which contained 554 landslides polygons, which cannot be linked to a particular triggering event.

We generated a completely new landslide inventory using multi‐temporal visual image interpretation using high resolution satellite images from 2014 and historical images from Google Earth, and generated an extensive landslide database for Saint Vincent. The resulting landslide database contains 1647 landslides. We also used digital image interpretation for generating a Geomorphological map for Saint Vincent. We analyzed the triggering conditions for landslides as far as was possible given the available data, and generated rainfall magnitude‐frequency relations. However, there were not enough data (both in terms of landslide dates and date‐related inventories) to be able to calculate magnitude‐frequency relations for landslides, in terms of the number or density of landslide for different frequencies.