Summary

This exercise is intended to show the potential of GIS in hazard zonation of landslides triggered by earthquakes. This exercise is based on data from Manizales (Colombia) which were collected in the framework of the ITC-UNESCO project on the use of GIS for mountain hazard mapping in the Andean environment.
Since this exercise is only intended to show the potential of using GIS in seismic landslide hazard zonation and not to prepare an accurate hazard map to be used for planning purposes, the geomorphologic and seismological setting are highly simplified.

Theoretical background

Geological studies have shown that a seismotectonic zone exists approximately 15 km north of Manizales capable of generating earthquakes with depths of 10 to 13 km and magnitudes over 6 on the Richter scale (Valencia, 1988). Landslide types that are triggered by earthquakes can roughly be classified into three groups (Keefer, 1984):

1. Rock falls and disrupted soil slides
2. Coherent deep-seated slumps
3. Lateral spreads

Given the geomorphologic setting of Manizales, landslides of type I and type II are most likely to be triggered by earthquakes. Landslides of type I are mainly found in closely jointed or weakly cemented rocks (rock falls) or on steep slopes in loose, residual or colluvial materials with low cohesion (disrupted soil slides). Coherent deep-seated, mostly rotational soil slumps (landslides of type II) on the other hand, are often found in relatively flat zones in deposits with significant cohesion.

Arias intensity is defined as the sum of all the squared acceleration values from seismic strong motion records. Arias intensity has been found to be a fairly reliable parameter to describe earthquake shaking necessary to trigger landslides. For the calculation of the spatial distribution of Arias intensity around Manizales the attenuation law of Wilson and Keefer (1985) is used:

where:
Ia	=	Arias intensity	[m/s]
Mw	=	Moment magnitude	[-]
d	=	Closest distance to surface projection of fault rupture	[km]
h	=	Focal depth of earthquake	[km]
P	=	Probability of exceedance	[-]

Harp and Wilson (1995) found a minimum Arias intensity of 0.11 m/s for the initiation of landslides of type I. The same authors reported a minimum Arias intensity of 0.32 m/s required for the initiation of landslides of type II. Larger Arias intensities indicate that stronger and longer-duration shaking is required to trigger landslides of type II.

Methods

In order to prepare a simple regional hazard zonation map which indicates where an earthquake with a certain magnitude could trigger landslides of type I or type II, the following three steps have to be carried out (see also the flowchart):
Step 1: Creating a map that shows the minimum Arias intensity required to initiate a landslide. This is a function of the geomorphologic terrain conditions, that decide whether the terrain is susceptible to landslides of type I or type II.
Step 2: Creating a map which shows the Arias intensity that can be expected at a specific site, with a given earthquake magnitude, earthquake depth and distance from earthquake's epicenter.
Step 3: Combining the maps that were created at step 1 and 2. At those sites where the expected Arias intensity exceeds the minimum required Arias intensity, landslides may be triggered.

Step 1: Rasterizing and reclassifying the geomorphologic map

The geomorphologic map of Manizales combines information on rock types, soil types and slopes. This map has to be reclassified into two classes indicating the susceptibility of the terrain to landslides of either type I or type II. The units characterized by steep slopes and thin soils are mainly susceptible to landslides of type I, while more gentle slopes with a thick soil cover are susceptible to deep-seated landslides of type II. The regional geomorphologic map is a polygon map that has to be rasterized. First a new georeference is created for the output raster map. After rasterization, a new table with three classes is created; this table is used to reclassify the geomorphologic map into a landslide susceptibility map.

Step 2: Preparing and classifying the Arias intensity map

For the preparation of the Arias intensity map, a point map with the location of the epicenter is required. After this the point map is rasterized with a user-defined georeference. With this raster map the distance to the epicenter for each pixel is calculated. The distance map is much larger than the rasterized and reclassified geomorphologic map and therefore a submap (left figure) is created using the georeference of the geomorphologic map. With some mapcalc formulas, the attenuation law is applied to obtain the spatial distribution of Arias intensity for an earthquake with magnitude 6.3. Finally the Arias intensity map is classified into classes of minimum Arias intensity (right figure).

Step 3: Preparing seismic landslide hazard zonation map

The seismic landslide susceptibility map (result of step 1) and the classified Arias intensity map (result of step 2) are combined in order to prepare a map which shows the zones that can be affected by landslides of type I and type II during an earthquake with a magnitude of 6.3 north of Manizales. In ILWIS, two maps can be combined in several ways. In this application a cross operation is used. The result of this cross operation is a cross map (shown above) and a cross table. A new column is added to the cross table and this column is used to reclassify the output cross map into a Hazard map (see below), that consists of the following three zones.

1. Zones where there is no landslide hazard either because the terrain is not susceptible to landslides or because the Arias intensity is not large enough to trigger landslides.
2. Zones with a hazard for landslides of type I.
3. Zones with a hazard for landslides of type II.

Conclusions

This simple hazard map for earthquake triggered landslides can serve as a base map for a more detailed study in which additional maps such as a slope map, a hydrogeological map and a surface deposits map have to be used. In order to create a risk map also data about infrastructure, buildings and economic activity of the region have to be included.

References

• Harp. E.L. and R.C. Wilson. (1995). Shaking intensity thresholds for rock falls and slides: Evidence from the 1987 Whittier Narrows and Superstition Hills earthquake strong motion records. Bulletin of the Seismological Society of America, 85 (6): 1739-1757.
• Keefer, D.K. (1984). Landslides caused by earthquakes. Geol. Soc. Am. Bull., 95: 406-421.
• Valencia, C.E. (1988). Geotectó regional del antiguo Caldas con é en la aplicació a la ingenierí sí. Unpublished Msc. Thesis, Universidad de los Andes, Bogotá, Colombia, 54 pp.
• Wilson, R.C. and Keefer, D.K. (1985). Predicting aerial limits of earthquake-induced landsliding. In, J.I. Ziony (ed.), Evaluating earthquake hazards in the Los Angeles region - An Earth-Science perspective, USGS Professional paper 1360, pp. 316-345.

For more information on this case study, contact:

C.J. van Westen / M.T. J. Terlien
Department of Earth Systems Analysis,
International Institute for Geo-Information Science and Earth Observation (ITC),
P.O. Box 6, 7500 AA Enschede, The Netherlands.
Tel: +31 53 4874263, Fax: +31 53 4874336, e-mail: westen@itc.nl