Landslide Susceptibility
Breifne Area (North West Ireland) Landslides Susceptibility Mapping

The implementation of a methodology for systematic production of landslide susceptibility maps hasn’t been embraced yet in Ireland. The Bréifne Area, located in North West Ireland and covering an area of 2129.7km2 within the Republic, was chosen as a suitable test area for this purpose. landslide susceptibility





















Map 1. Landslide susceptibility map for bedrock slides


Several digital datasets available were tested: orthophotography, Landsat ETM+ imagery, and Digital Elevation Models (DEM) at 20 and 50m resolution. Colour and black and white (B&W) orthophotography was deemed to be the most suitable due to their high spatial resolution. The B&W orthophotography was combined with a DEM at 20m resolution using ãFledermaus software to identify and classify the landslide occurrences in the area. Digital stereophotography was also employed for certain areas.

Landslides mentioned in the literature were analysed using 3D orthophotography to recognize its signature and to identify elements that would assist the classification of similar events. The work was focused in areas with slopes greater than 15° and the region systematically surveyed to identify, classify, locate, orientate and measure landslide events. ArcGIS software was used to digitise the head of the event as a point feature. A line from this point to the toe of the landslide was digitised to recognize the total length of the event and its orientation. A total of 694 landslides were identified and classified into 17 subgroups. These were subsequently grouped into 4 main landslide types: bedrock slides, peat slides, falls and flows.

A total of 52 landslides were visited and described during fieldwork carried out to verify the methodology. The following parameters were recorded when available: morphological setting, dimensions of the event, geological setting, impact on the landscape and the infrastructure, remedial measures taken and climatic conditions during event. The remote sensing based classification was revised and contrasted with fieldwork descriptions. Following the fieldwork 30% of bedrock slides and 50% of peat slides were reclassified. The fieldwork was essential to properly categorize and catalogue land movements previously identified using remote sensing techniques.

The literature review and the fieldwork observations reveal that conditioning factors triggering landslides differ depending on the landslide type. Each of the landslide types mentioned above was assumed to encompass similar conditions to landsliding. Six datasets were developed as thematic maps for the area and considered as triggering factors: bedrock type, soil parent material, land cover, slope, aspect and elevation. Statistical analysis was used to determine the influence of each triggering factor on each landslide type. The percentage of a landslide type occurring in a class within a triggering factor was assigned as the weight controlling the susceptibility to landsliding in areas covered by that class (Fig. 3).

Four susceptibility maps were produced, one for each landslide type. The triggering factors expressed as raster thematic maps were reclassified according to the weights mentioned above (e.g. bog & heath was reclassified to 56 for bedrock slides, to 82 for peat slides, to 38 for flows, and to 42 for falls - Fig. 3). The reclassified pixel values of the six thematic maps were summed using the Spatial Analyst extension of ArcGIS. On the resulting susceptibility maps high pixel values indicate high susceptibility to landsliding and low pixel values represent low susceptibility. Maps were divided into 7 levels of susceptibility to landsliding ranging from extremely high to extremely low values.

Error assessment was also performed to analyse the correlation between landslides mapped and susceptibility values where these events occur (Table 1). An example of the landslide susceptibility map for bedrock slides is displayed in Map 1. Integration of additional conditioning factors such as rainfall or structural geology might improve this correlation. The methodology used during this project allows the incorporation of new datasets (triggering factors) to derive more accurate landslide susceptibility map outputs. 

Percentage of landslides by landcover type
Figure 3. Percentage of landslides by landcover type

Susceptibility

% Bedrock Slides

% Peatslides

% Flows

%Falls

Extremely High

8.82

19.01

8.97

9.22

Very High

18.63

15.29

19.23

4.96

High

32.35

36.36

23.08

35.46

Medium

36.27

29.34

41.03

47.52

Low

3.92

0.00

7.69

2.48

Very Low

0.00

0.00

0.00

0.00

Extremely Low

0.00

0.00

0.00

0.35

Table 1. Error assessment. Percentage of events mapped contained within each susceptibility category

 

Back to the top