J. Tablazon1,3, A. M. F. Lagmay1,2, C. Ladiero1, J.V. Puno1, J.K. Suarez1, and J. Santiago1,4
1Nationwide Operational Assessment of Hazards, U.P. Diliman, Quezon City, Philippines
2National Institute of Geological Sciences, University of the Philippines, C.P. Garcia corner Velasquez street, Diliman, Quezon City, Philippines
3Institute of Environmental Science and Meteorology, University of the Philippines Diliman, Quezon City, Philippines
4School of Urban and Regional Planning, University of the Philippines Diliman, Quezon City, Philippines
Typhoon Hagupit was the strongest tropical cyclone of 2014, tied with Typhoon Vongfong (Japan Meteorological Agency, 2014). It formed on 01 December as a tropical depression and due to the favorable low vertical wind shear and excellent radial outflow, it developed into a super typhoon on 03 December. It entered the Philippine Area of Responsibility on 04 December and was given the local name Ruby. It reached peak intensity with a 10 min maximum sustained winds of 213 kph. However, when it made landfall over Dolores, Eastern Samar, the typhoon weakened with a 10 min maximum sustained winds of 167 kph (Figure 1).
The Nationwide Operational Assessment of Hazards under the Department of Science and Technology (DOST-Project NOAH) started simulating storm surge for Typhoon Hagupit on 03 December 09:00 UTC. On 06 December 06:00 UTC, the municipality of Daram in the province of Samar was forecasted to experience storm surge of 2.6-3.6 meters. The conflicting tracks of Typhoon Hagupit from the different meteorological agencies made the storm surge forecasting complicated, since the storm surge model is strongly dependent on the accuracy of the meteorological input.
Given the timely storm surge warnings, residents of Bicol, Southern Luzon, and Western Visayas had time to take precautionary measures against possible storm surge. Pre-emptive actions undertaken include class and business suspensions, cancellation of transport operations, and evacuation of residents in low-lying areas. Despite all of this, the National Disaster Risk Reduction Management Council accounted 18 fatalities and 916 injured. Washed out houses by storm surge were reported in Bicol and Western Visayas, but there were no casualties.
The municipality of Daram reported that the typhoon affected 99 percent of the total number of households and completely damaged 1,664 houses (mainly attributed to storm surge) and partially damaged 5,773 houses (due to winds and storm surge) amounting to USD 5.5 million. Despite the massive damages, the zero-casualty after Typhoon Hagupit was made possible through lesson learned during Typhoon Haiyan and collaboration of various government agencies and local government unit of Daram in disaster preparedness, monitoring, and information dissemination.
It is important to note that Daram implements the purok system. It is a system of organizing residents at the purok level – the smallest geographical unit in the Philippines. The purok system is a social organization operating on the concept of cooperation and highlights the importance of working together in achieving a common goal. This is also a way of empowering communities and equipping them with knowledge on disaster preparedness. When an impending hazard is about to hit Daram, warnings from the local government unit is sent to the heads of the barangays. These warnings will then be communicated to the purok representatives, which will then facilitate evacuation.
The DOST-Project NOAH served as a leading source of timely and reliable storm surge forecasts. Moreover, storm surge hazard maps made available at their website contributed to the risk knowledge of the communities through identification of safe and unsafe areas. In order to provide accurate data in the storm surge hazard mapping process, the results of the storm surge inundation model must be validated using observed water levels in the coastal areas. The objective of this study is to validate the storm surge inundation simulation results of Typhoon Hagupit done by DOST-Project NOAH.
The fieldwork was conducted to validate the inundation map of the flooding from storm surge for Typhoon Hagupit using the FLO-2D software – a two-dimensional flood routing model. It is a GIS-integrated software that facilitates the creation of the flood model grid system and is a simple volume conservation model composed of a processor program that facilitates graphical editing and mapping of flood details. The continuity equation and the dynamic wave momentum equation are the governing equations of the model (FLO-2D Software, 2015).
The input data used in FLO-2D were 20-meter resolution digital elevation model generated from Interferometric Synthetic Aperture Radar (IfSAR) and storm tide time series derived from the results of the Japan Meteorological Agency (JMA) storm surge model combined with the tide data from WXTide. The JMA storm surge model is a numerical model developed by the Japan Meteorological Agency to simulate and predict storm surge mainly caused by tropical cyclones (Higaki, 2006). The input for the storm surge simulation includes the track, maximum sustained wind speed, and central atmospheric pressure of Typhoon Hagupit. On the other hand, the WXTide is a software containing archives/catalogues of world-wide astronomical tides (WXTide32, 2015).
To validate the model results, interviews were conducted in the eight barangays of Daram. Residents of the coastal communities were interviewed to collect data on the extent and height of the storm surge inundation during Typhoon Hagupit. Flood depth and inundation extent data with their corresponding geographic coordinates were collected using a measuring tape and a hand-held GPS device. Photographs on-site were also taken for reference.
Data validation method was done by comparing the simulated and observed flood depths and calculation of the root mean square error (RMSE). RMSE has been used as a standard statistical metric to measure model performance in meteorology, air quality, and climate research studies.
Interviews were conducted in eight barangays of Daram, particularly in barangays Casab-ahan, Jacopon, Losa, Mabini, Mayabay, Mongolbongol, Rizal, and Ubo. A total of 114 respondents living near the coast and were in the vicinity when Typhoon Hagupit struck were interviewed (Table 1). Accounts of the residents from the eight barangays will also be discussed in detail.
Flooding due to storm surge from Typhoon Hagupit was recorded to validate the results of the storm surge inundation model. The flood heights were divided into five categories: no flood, extent of flood (from floor up to ankle), 0.11-0.60 meter (from ankle up to knee), 0.61-1.00 meter (from knee up to waist), and 1.01 meters and above (above waist level).
Fifty three houses, seawall, and fishing boats were destroyed due to the strong waves brought by Typhoon Hagupit (Figure 2). Moreover, rock boulders were carried inland by the waves. Residents evacuated the day before the onslaught of Typhoon Hagupit upon the instruction of the barangay officials. As seen in Table 2, most respondents reported extent of flood, which only reached the elevated pavement in the barangay. The highest reported inundation in Barangay Casab-ahan is 0.6 meter.
Barangay Jacopon has a rocky sloping beach. Observations were mostly flood extent, except in the rocky part of the barangay shoreline, where barangay residents reported 1.0 meter flood depth (Table 3). The remains of a destroyed seawall and house are testaments to the destruction brought to the barangay (Figure 3).
Typhoon Hagupit washed out and damaged houses in Barangay Losa (Figure 4). Statistics from the local government unit of Daram revealed that 74 houses were totally damaged due to strong waves. Barangay officials related that they got advice from the Municipal Disaster Risk Reduction and Management Office (MDRRMO) to evacuate, thus the residents along the shore relocated to higher ground. 70% of the respondents reported extent of flood (Table 4), but the highest reported flood depth is 1.1 meters.
Sixty seven houses, the barangay’s basketball court, and port were destroyed due to the strong waves of Typhoon Hagupit (Figure 5). High flood depths were also observed (Table 5) with 1.5 meters as the highest reported inundation. One resident located three meters from the shore reported sand deposition of 1.5 meters high.
Thirty six houses were completely damaged due to the strong waves of Typhoon Hagupit (Figure 6). Table 6 shows the frequency of reported flood depths in Barangay Mayabay. The highest flood depth reported by the residents is 1.1 meters.
Strong waves of Typhoon Hagupit damaged the seawall and washed out houses in Barangay Mongolbongol (Figure 7). Additionally, strong waves carried boulders ashore. Barangay officials were in charge of the evacuation as they received the advice from the MDRRMO. The residents evacuated immediately upon the instruction of the barangay officials. Data collected in the barangay were mostly flood extent accounting for 66.7% of the total respondents (Table 7). The highest reported inundation is only 0.25 meter.
In Barangay Rizal, the damaged seawall, basketball court, and health center proved the strength of the storm surge event during Typhoon Hagupit (Figure 8). A barangay councilor reported that the typhoon washed out 46 houses. The residents evacuated in the morning of 06 December upon the instruction of their barangay officials. The barangay officials received constant text messages from the MDRRMO that they evacuate to avoid casualties. They imposed forced evacuation and evacuated the residents in the school in their barangay. The waves were reportedly approximately 4.0 meters high, reaching the height of the bridge in Figure 9. Most of the respondents (39.29% of the total) reported greater than one meter flood depths due to storm surge during Typhoon Hagupit with 3.0 meters as the highest reported inundation (Table 8). Whereas, residents along the main road traversing the barangay reported only of flood extent or no flood. One resident along the shore recently rebuilt their house on cemented stilts as a precautionary measure. In the left end of the barangay, facing the sea, residents located five meters away from the shore also built an elevated pavement that also serves as seawall.
Barangay Ubo was mildly affected by Typhoon Hagupit. Table 9 shows the frequency of reported flood depths during Typhoon Hagupit. The highest reported inundation is 0.75 meter. Residents evacuated in houses at high elevation, since their school that is typically used as evacuation center is located near the river. Strong waves also caused the damage to the walkway to the port of the barangay (Figure 10).
The results of the storm surge model were compared with the field validation observation using the root mean square error (RMSE). Barangays Casab-ahan, Jacopon, and Rizal generated an RMSE of more than 1.0 meter.
In Barangay Casab-ahan, the computed error is 1.783 because the simulation results extend further inland than the observed (Figure 11). The flood only reached the elevated pavement where it served as barrier.
Barangay Jacopon returned an RMSE of 1.115 because observations indicate that flood only reached 0.2 meter while the model shows flood reaching 2.0 meters (Figure 12). The model did not account for the presence of rocks and boulders along the shore which may serve as wave barrier.
Barangay Losa field validation returned an RMSE of 0.721 meter. Figure 13 shows that the simulated flood extent and flood depth closely agree with the field observations. The error may be explained by the barangay’s modified shoreline which prevented the flood from reaching higher than the paved road. The impact of the waves were absorbed by the structures along the shore as proven by damage reports such as destroyed kitchen or totally washed out houses.
The field validation in Barangay Mabini reported an RMSE of 0.848. In Figure 14, the simulation results indicate that the northern part was not inundated while interviews indicated flood reaching 1.0 to 1.5 meters. Meanwhile, in the southern portion field observation and survey agree with each other.
The calculated RMSE for Barangay Mayabay is 0.460. The field survey reports that the flood extend inland in the southern portion while the model simulated no flood in the said sampling points (Figure 15). Meanwhile, the neck of the spit received a fair amount of inundation but the sampling points were not enough to draw a conclusion.
Barangay Mongolbongol returned the lowest RMSE for Typhoon Hagupit with 0.224 meter. This is due to the low number of observations in Barangay Mongolbongol, since the number of collected observation points for flood extent isexcluded from the RMSE computation. However, the flood extent fairly indicates that some portion is flooded as expected (Figure 16).
Barangay Rizal garnered an RMSE of 1.365 meters. The residents were forced to evacuate the area and reports along the shore were approximations only. Flood overtopping their roofs were estimated at 3.0 meters, which possibly included the wave height. In Figure 17 and 18, the built seawall prevented further inundation inland. The houses along the shore are not considered in the model and may have served barriers increasing the bed roughness in the area thus decreasing the flow of seawater.
In Barangay Ubo, the calculated RMSE is 0.342. The survey data are generally on flood extent that is further inland than the model results (Figure 19). Though the sampling points for no flood were the same for the model and field survey, flood observation in locations not flooded in the model increase the error.
Typhoon Hagupit was one of the strongest tropical cyclone of 2014. DOST-Project NOAH released storm surge forecasts to alert the public of the impending danger. Storm surge hazard maps were also made available in the website to notify the public which areas will be affected and which areas are safe for evacuation. Storm surge from Typhoon Hagupit brought considerable damages in the municipality of Daram. Moreover, given the strength of Typhoon Hagupit, there were no casualties in the municipality. Many people could have died if there were no warnings given to the public. This can be attributed to the effective disaster preparedness of the municipality, including its effective information dissemination made possible by the purok system.
Interviews were conducted in eight coastal barangays of Daram to validate the results of the storm surge inundation model. The observed and simulated flood depths were compared and the RMSE was computed to test the performance of the model. Most respondents from the interviews reported extent of flooding (from floor to ankle flood depth) during Typhoon Hagupit. Moreover, the highest reported inundation is in Barangay Rizal with 3.0 meters storm surge. These results were compared with the simulated results of the FLO-2D. Results show varying RMSE with the different barangays. The RMSE values for barangays Mongolbongol, Ubo, Mayabay, Losa, and Mabini are in the acceptable range of less than 1.0 meter. However, high RMSE with more than 1.0 meter is computed for barangays Casab-ahan, Jacopon, and Rizal. One of the reasons for the discrepancies between the observed and simulated data can be attributed to the presence of structures such as seawalls and other coastal structures that protect the coastal areas, since these structures serve as defense against storm surge and were not taken into account in the FLO-2D.
The results of the validation proved that the storm surge inundation maps of DOST-Project NOAH are reliable given the low RMSE values for the five out of eight barangays surveyed. These maps are important in helping local government units to locate safe areas for people to evacuate when storm surge is predicted to hit their locality. Furthermore, this will improve existing evacuation plans and delineate safe and unsafe areas to prevent loss of lives, injuries, and damages to properties.
FLO-2D Software: www.FLO-2D.com, last access: 06 December 2015.
Higaki, M.: A Guide to JMA Storm Surge Model, in: Fourth Regional Workshop on Storm Surge and Wave Forecasting, Office of Marine Prediction, Global Environment and Marine Department, Japan Meteorological Agency, Tokyo, 2006.
Japan Meteorological Agency: Western North Pacific Typhoon Best Track File, available at: http://www.jma.go.jp/jma/jma-eng/jma-center/rsmc-hp-pub-eg/besttrack.html, last access: 15 December 2015.
WXTide32: Tides and Currents for Win9x/NT, available at: www.wxtide32.com, last access: 04 December 2015.