Processing of the 2010 LiDAR data at the Pos Selim landslide and assessment of the uncertainty of survey information

Abstract

Contract of Simpang Pulai-Lojing Highway was commenced in 1997. Since then, the unstable hillside at a site at CH23+800 in the north-western Cameron Highlands Malaysia was cut to ridge top but instability still observed. The cutting was mainly in schist and was about 250m high. In August – September 2003, a landslide about 190m high and between 200m and 430m wide was formed at the site as the hillside was cut into for roadworks. The work reported here shows that the overall velocity of movement of the displaced mass of the landslide is declining slightly year on year and the head of landslide had moved more than 30m by August 2010 since September 2003. In 2010 the top of the main body of the landslide was moving at about 0.5m per year. Points on the landslide continue to move 2005-2010 in generally the same direction as 2003-2005. The writer was provided with survey data at the landslide and adjoining area obtained in 2003, 2005 and 2010 by different methodologies. The survey in 2003 was a composite as-built survey; 2005 was an aerial survey and 2010 was a LiDAR survey. The 2010 LiDAR data has been used as a basis to give both the 2003 as-built survey plan and 2005 orthophoto a common horizontal reference frame. Both the 2003 as-built survey plan and 2005 orthophoto have been adjusted to fit the 2010 LiDAR data, which is expressed in terms of old MRSO grid coordinates system after transformation. The adjustments are affine transformation (2003 and 2005). The 2010 LiDAR data has been reduced to the vertical datum used in both 2003 as-built survey and 2005 aerial survey. A number of recognisable points within the landslide have been identified in all three surveys. The differences in coordinates of the points on the dates of the three surveys provide two sets of movement vectors (2003-5 and 2005-10). There are 150 such vectors. Thirty-three check points have been identified in the 3 surveys outside and adjacent to the landslide at places where movements are not expected. The apparent movement of these points is an indication of the uncertainty in the displacement data. Total displacements of 12 total station survey markers within the landslide have been plotted with time. Ground points very close to these markers are plotted for comparison. The latter are greater than the former by about two metres suggesting that the landslide had started moving before the start of deformation monitoring.