J.A.M. Galang*1,2, J.J Sulapas1,2, C.M. Escape1,2, K.R. Montalbo1,2, R.N. Eco1,2
1Nationwide Operation Assessment of Hazards, U.P. NIGS, C.P. Garcia Ave., U.P. Diliman, Quezon City, 1101 Philippines
2National Institute of Geological Sciences, University of the Philippines, C.P. Garcia corner Velasquez street, U.P. Diliman, Quezon City. 1101 Philippines
*Corresponding author: Email address: firstname.lastname@example.org
Typhoon Ineng (international name Goni), entered the Philippine area of responsibility (PAR) on August 18, 2015 pouring on heavy rain in northwest Luzon. Bakun, a municipality in the mountainous Benguet province, was one of the most affected areas receiving rainfall amounts of 513 mm in a span of four days. Five days after the Typhoon entered PAR, a landslide broke in Sitio Buagi, Barangay Poblacion, Bakun transporting soils, rocks and also houses downslope. The generation of post-landslide digital elevation model and availability of pre-existing IfSAR DTM enabled a volume estimation of 140,000 m3. The landslide is classified as a deep-seated soil slump sliding on a rotational surface. Field investigations confirmed that the thick soil layer slid down when the supporting underlying mudstone layer broke. This event yields a threshold rainfall value for landslides in Bakun, Benguet.
On 18 August 2015, Typhoon Ineng (international name Goni) entered the northern part of the Philippine Area of Responsibility (PAR). It did not make landfall but it delivered heavy rains in northwest Luzon by enhancing the southwest monsoon. Ensuing floods and landslides left 33 people dead and approximately 4.4B Php worth of damages to infrastructure and agriculture (NDRRMC, 2015).
Three years prior to this event, on 28 July 2012, Typhoon Gener (international name Saola) entered the Philippine Area of Responsibility. The tropical cyclone also did not make landfall, similar to Goni, but it brought about 769 mm of heavy rainfall over the span of 4 days which has caused landslides in Sitio Buagi, Poblacion, Bakun damaging seven residential structures. The Mines and Geosciences Bureau (MGB) reported the landslide as a progressive creep where the downward movement is imperceptibly slow and at a certain time during the typhoon, reached a point of failure where it progressed to translational sliding MGB (2012). The said landslide occurred approximately 200 meters northwest from the new landslide. Tension cracks with 30 cm vertical displacements and 5 cm openings on average formed in the area is now the site of the 2015 landslide area (Fig. 2).
Description of Failure
The failure site is a residential area in Sitio Buagi where a large mass of land started sliding down and sinking up to several meters below its original position according to respondents Celestino Bautista and Maximo Macasling. Figure 3 shows an aerial photograph of the damage done by the landslide which includes the displacement of a few houses and soil-covered rice terraces. Drainage has been naturally carved at the outline of the landslide.
The source area includes the 130-meter wide crown opening up to the southwest and the 20 meter high exposed scarp. The slope on the scarp is about 40⁰, gently decreasing away from the scarp and eventually attens to mark the end of the body. The foot of the landslide covered rice terraces and was bordered by the owing river. In total, the failure site is 150 meters high from scarp to toe, and approximately 500 meters in length.
To better describe the landslide, a rough 3D model and a post-landslide digital elevation model (DEM) was produced using aerial photographs and videos from OCD-CAR. A rough landslide volume estimate of 140,000 m3 of transported material was computed by overlaying the new elevation data on the pre-existing IfSAR digital terrain model (DTM) of the area, and subtracting their difference in morphology. The cross section of the landslide seen on Figure 4 was based on the said 3D model.
Sitio Buagi sits atop a tilted sedimentary sequence of conglomerate, mudstone, and limestone layers (Fig. 5). This conglomerate layer dipping 36⁰, was seen as an outcrop near the landslide area conformably overlain by the mudstone layer. The limestone is assumed to overlie the mudstone as it is the rock to which the fallen houses were built on and it is a major component of the transported mass. The sliding surface is likely to be parallel to the tilted beds.
The Bakun landslide is a deep-seated soil slump based on the classification of Hungr et al. (2014) and Varnes (1978), and is a rotational slide with mostly soil on a rupture surface. This kind of landslide motion occurs in water saturated homogeneous soils with low permeability, such as the mudstone underlying Sitio Buagi. Landslide movement lasted for 2 minutes, which is classified as a slow moving landslide (Varnes, 1978), as was recorded in video accounts. This soil slump resulted in a high degree of strength loss, and thus exposed the land to future slides at or near the present scarp.
The displaced materials were mostly soil and pebble- to boulder-sized mudstone and limestone fragments. An outcrop a few hundred meters away from the slide exposed a conglomerate bed trending N 40⁰ W, and dipping 36⁰ SW. The mudstone, conformably overlies this conglomeratic bedrock. A thick layer of unconsolidated materials of soil and limestone fragments overlies this bedrock. The planar orientation of this outcrop is similar to the direction of the exposed scarp in the head, taking with it the underlying siltstone that was also part of the transported debris.
Trigger of Slope Failure
Northwestern Luzon received intense rains during Typhoon Ineng. In Bakun alone, 59 landslides along the road have been plotted. It is likely that more of these landslides throughout Bakun and the surrounding municipalities. A 4-day rainfall contour (Fig. 7) shows the specific areas receiving large amounts of rainfall; this includes Bakun, along with Baguio and Abra, as one of the areas with more than 400 mm of accumulated rainfall.
The landslide was triggered by the large amounts of accumulated rainfall in Bakun. Automated rain gauge (ARG) data from the nearby town of Bugias (Fig. 6) recorded continuous rain for more than 84 hours. Rainfall amounts were highest on August 21 where 200 mm of rain was accumulated in a 24-hours. By noon of August 23, the accumulated rainfall starting August 20 has reached 513.2mm. The 24-hour rainfall contour shows the accumulated rains less than 100mm. This means that the antecedent rainfall from Aug 20 to 22 (approximately 400mm) was the biggest factor in triggering the landslide.
The continuous and heavy rainfall contributed to the instability of the slope as rainwater percolated and saturated the subsurface. The underlying mudstone also highly saturated as the water table is close to the surface. This water table was intersected by the landslide, as manifested by the spring in the landslide scarp. The area was more susceptible to collapse due to its thick and homogeneous soil cover. Moreover, the conglomerate bedrock tilts 36⁰ to the southwest, the same direction as the present soil slump, making it a viable plane of weakness for any sliding event.
Five houses were displaced by 100 meters during the landslide event. Residents were able to react to the slow moving soil slump, and were evacuated during the event. A day before the actual landslide, there were accounts from the residents of a strong explosive-like sound that resulted to the seepage of black water under the eventual landslide area. This black water is now observed in the drainages that outline the present landslide. Aside from the few houses, approximately 60,000 m2 of agricultural lands below the scarp were completely covered by soil, organic matter, and rocks.
Tension cracks were also observed before the crown as seen on Figure 8, towards the other residential area. These were located at or near residential houses, whose expressions are also pronounced even at fences, concrete pavements, and walls.
Future movements may involve areas above the present landslide as there is enough evidence to infer that the landslide is retrogressing. This means there will be further erosion towards the direction of the rest of the community (Northwest). As proven by this large landslide event, the identification and observation of tension cracks in this area should not be taken lightly. The 2015 landslide is the area where tension cracks were reported on proximity of the 2012 landslide.
It is important to monitor tension cracks in any landslide prone area, more so with this community, as the tension cracks observed in 2012 resulted in the 2015 landslide.
The same kind of cracks were found in houses and canals, which should be considered as markers of a landslide susceptible area. Rates of ground motion must not be disregarded as these are vital precursors to the more rapid land movements. Lastly, as a safety measure, drainage system, may be improved to prevent water infiltration that may cause soils to be oversaturated.
Constant communication with the Disaster Risk Reduction (DRR) officer of Bakun, or the Regional DRR in Baguio should be practiced. When internet connection is available, monitoring rainfall could also be possible through the utilization of the Project NOAH website (beta.noah.dost.gov.ph). Bakun has an automatic rain gauge (ARG), but it has not been serviceable since May 2015, DOST-ASTI has committed to address this problem. Rainfall contours are useful in connecting the rainfall measurements from nearby Bugias ARG and Mankayan ARG. The present threshold value for Northern Benguet municipalities of Bakun, Bugias, Mankayan and Kibungan is 513.2 mm of accumulated rain. Any rainfall event bringing more accumulated rains than the threshold value is expected to trigger landslides.
Hungr, O., Leroueil, S., and Picarelli, L., 2014. The Varnes classification of landslide types, an update. Landslides, 11, 2, 167-194.
MGB, 2012. Landslide Report on Sitio Buagi, Barangay Poblacion, Bakun, Benguet. Technical report, Mines and Geosciences Bureau.
NDRRMC, 2015. Situation Report no. 23. Technical report, NDRRMC, NDRRMC Center, Camp Aguinaldo, Quezon City.
Varnes, D. J., 1978. Slope movement types and processes. Transportation Research Board Special Report, 176.