GIS-based spatial analysis and modeling for landslide hazard assessment in steeplands, southern Honduras

Landslides are a serious concern in steepland areas where intensive storms are common. The level of landslide hazard is not only dependent on storm and land physical attributes, human activities, especially agriculture practices, are also important considerations. Hurricane Mitch was one of the strongest hurricanes of the 20th century; it generated intense rain that triggered widespread landslides and flooding throughout Honduras and Nicaragua. The objective of this study was to develop and test a heuristic approach for predicting spatial distribution of landslide hazard. Four variables, slope, aspect, stream proximity, and land cover type, generated from available topography and remote sensing data, were used in the model. This model was developed and verified in southern Honduras based on data from the Namasigue watershed and validated in the adjacent El Triunfo watershed. Results of GIS analyses showed that the likelihood of landslide was significantly influenced by slope and land cover. As slope increased, the percentage of land affected by landslides, when the soils were saturated, increased sharply on land used for crop production, indicating that agricultural activity and the associated removal of deep-rooted permanent vegetation increased the landslide hazard on steep sites. Consequently, sites covered by shrub fallow and forests had relatively low incidence of landslides regardless of the topographic features. The model was effective in predicting landslide hazard and generating landslide hazard maps; this model has been found to be a useful aid for soil and water conservation organizations in the region to target conservation investment.