Generating Evacuation Route for Tsunami Evacuation Based on Megathrust Scenario Hazard Model in Palabuhanratu Village, Sukabumi, West Java

. Abstract Palabuhanratu Village is one of the villages in Sukabumi, West Java, that is susceptible to earthquake and tsunami risks. This research intends to revise the tsunami hazard map, undertake a spatial analysis of the distribution of evacuation sites, and identify optimal tsunami evacuation routes. The tsunami hazard map was updated using tsunami modeling with COMCOT based on the worst-case scenario of potential magnitude moment 8.8 for the Megathrust segment in the south of West Java from PuSGeN. This modeling was used to predict the worst probable tsunami impact. On the basis of field survey data regarding the location of evacuation sites, evaluation of the distribution of evacuation sites was conducted. In addition, service area analysis is utilized to assess the service area of the present evacuation site in relation to each hamlet in Palabuhanratu village. Approximately 57.33 percent of the town could be affected by a tsunami, according to the findings of this study. The greatest tsunami height along the coast is expected to be between 18 and 22 meters, and the arrival time is 22 minutes. From a total of 35 hamlets, we determined that two hamlets in the Palabuhanratu village area were not harmed by the tsunami. Because not everyone can reach the evacuation location in time, the findings of this study show the need for an additional vertical evacuation site.

Based on the 2017 Pusat Studi Gempa Nasional (PUSGEN), the maximum potential magnitude in Southwest Java reaches Mw 8.8, caused by the megathrust of the Sunda-Banten Strait segment. This scenario has been updated from the previous one in 2010 with a maximum potential magnitude of Mw 8.5. A big earthquake can cause a tsunami, such as the Pangandaran tsunami caused by an earthquake with a magnitude of 7.7 (Fujii & Satake, 2006). The tsunami hazard map for Palabuhanratu needs to be revised using the latest worst-case scenario.
Making a tsunami hazard map requires modeling a tsunami source to determine the level of exposure to a tsunami disaster. Exposure in tourist areas will increase as the number of tourists increases over time (Marfai et al., 2021).
According to the Central Bureau of Statistics for Sukabumi Regency (2020), Palabuhanratu sub-district consists of ten villages. Because it is located in a coastal area and is a tourist destination (Juliandri et al., 2020), tourists and residents in the area should have more knowledge about the evacuation zones and safe routes for tsunami evacuation (Takabatake et al., 2017). A tsunami evacuation is an effective step to save yourself from a near-distant tsunami disaster (Z. Wang & Jia, 2021). Evacuation is the process of protecting human beings from a dangerous area to an evacuation site to reduce vulnerability in terms of life and health. In a disaster response, evacuation must be carried out quickly and adequately (Usman et al., 2017). Based on this background, this study aims to revise the Tsunami Hazard Map using the latest scenario from PUSGEN, identify the optimal tsunami evacuation route for the people in Palabuhanratu Village and evaluate the distribution of evacuation sites.

Study Site
Palabuhanratu is a village in the Palabuhanratu sub-district, Sukabumi Regency, West Java province.
Palabuhanratu village has an area of approximately 821.62 hectares and is located in the Western part of the regency, precisely at coordinates 6° 58' 40.8" S -7° 01' 01.2" S and 106° 32' 16.8" E -106° 34' 40.8" E. The population in the Palabuhanratu village reaches 34,443 people. Palabuhanratu village consists of 133 Neighbourhoods and 35 hamlets, according to data from the Central Bureau of Statistics. Its location facing the Indian Ocean has made Palabuhanratu Village a potential tourism sector in West Java. However, this causes this area also to hold the potential threat of a tsunami hazard.

Data Collection
The first step taken is data collection. This study generally uses two kinds of data, namely primary and secondary data. To perform the tsunami modeling, we need input data such as bathymetry and topography. This study uses topographic data from DEMNAS and bathymetry data from BATNAS provided by the Geospatial Information Agency. In addition, earthquake source parameter data is also needed, which will be explained in point 2.3. Land cover data is also necessary to determine the friction coefficient used in tsunami modeling and obtained from the Ministry of Environment and Forestry. Then, to carry out the Network Analysis, road network data needed to be obtained from the University of Indonesia database. Then, we get the location of existing evacuation points/assembly points and evacuation signs from BPBD Sukabumi Regency and field survey data. Furthermore, the evacuation route's starting point is determined using hamlet administrative boundary data from the Central Bureau of Statistics.

Identification of Earthquake Source
Earthquake source parameters are obtained using scientific references and mathematical calculations, which will later be used as input in tsunami modeling. The detailed parameters used as modeling input are shown in Table 1 and Fig. 1. The earthquake source parameters used are as follows:  The magnitude used is based on a study from PUSGEN (2017) where the Sunda Strait Megathrust is estimated to have a maximum potential magnitude of Mw 8.8;  The rigidity used in the model is 30 GPa for a source depth of around 30 km (Bilek & Lay, 1999 ;Tanioka et al., 2017);  Parameters of depth, strike, and source dip using the subduction model for the Sumatra-Java segment of the USGS (Hayes et al., 2018).
The length and width parameters of the fracture are obtained by using the Wells & Coppersmith, Kevin (1994) equation which shows the relationship between the magnitude and the length and width of the fracture. The equation used is as follows: where is the length of the fault, is the width of the fault, dan is the moment magnitude of the earthquake.
Meanwhile, to get the value of vertical dislocations using the equation of the relationship between magnitude and seismic moment from Hanks & Kanamori (1979). Here is the equation used: where is the moment magnitude of the earthquake, is the earthquake's seismic moment, µ is the rigidity, is the vertical dislocation of the fault, is the fault length, and is fault width.

Tsunami Modeling
The next step is to do tsunami modeling with numerical modeling using the COMCOT (Cornell Multigrid Couple Tsunami) software (X. Wang & Power, 2011), which can be used to reconstruct tsunami wave propagation from the source (Syamsidik et al., 2019). COMCOT is a numerical solver for Non-linear Shallow Water Equations. The equation in question is the Non-linear Shallow Water Equations using the spherical coordinate system as follows: where λ and θ are latitude and longitude, η is water level, M and N are discharge fluxes in λ and θ directions, t is time, g is gravitational acceleration, h is water depth, R is earth radius, n is coefficient manning, D is the total depth (= η + h).

Field Survey
The field survey was carried out to identify and geotag the location of the existing evacuation site/assembly point, find out the evacuation route, and take samples of the travel time for evacuating from the assembly point to the evacuation site will later be used as evaluation material. The survey was conducted five days from 29 October to 2 November 2022. The field data were collected using the Avenza application.

Network Analysis -Shortest Path
The shortest path is analyzed to determine the shortest distance to the evacuation site. We conduct the shortest path analysis using input data such as a starting point. This study uses a starting point or assembly point recommendations determined using the travel time criterion from the hamlet centroid point ≤ 5 minutes. The theoretical speed assumption used in this study is 0.751 m/s (the average walking speed of older people), which is intended to ensure that all people can evacuate on time. Then, the endpoint used is the existing evacuation/assembly point obtained from the field survey. To perform the analysis, we use Quantum GIS software that can be calculate the shortest path between two points on any line layer and plots this path over the road network (Ilayaraja, 2013) .

Assembly Point
An assembly point is a designated location where individuals or groups gather after they have evacuated a building or area in an emergency. It is typically a safe and easily accessible location (Sianturi et al., 2021), away from the building or area that was evacuated and identified by signs or markings. In the case of a tsunami disaster, there are two types of assembly points, namely assembly points located outside the house or building as a location for coordination and checking of family members or building occupants before evacuating to a safe point.
Furthermore, an assembly point often refers to a temporary/final evacuation site. In the case study in Palabuhanratu, the local government uses the term assembly point as the location of the temporary/final evacuation site. As mentioned in the Data Collection section, we obtained the temporary/final evacuation site location from BPBD Sukabumi. However, we also need to recommend an assembly point as a meeting point location for residents before heading to a temporary or final evacuation site. So, in this study, the location of the gathering point is divided into two, namely, the assembly point as the final evacuation location and the recommended assembly point used as the initial evacuation location.
This research also uses assembly points as the starting point for evacuation. Some previous studies (Habibi & Khakim, 2016;Rumondor et al., 2019) determined the starting point of evacuation can use the location of the assembly point. This is because after a tsunami disaster, according to the evacuation procedure, everyone will first gather at the gathering point before evacuating to a safe temporary or final evacuation point/assembly point.

Network Analysis -Service Area
Service area analysis aims to determine the coverage of the area that can be reached from the existing evacuation site/assembly point using the evacuation time parameter ( Almost the entire Palabuhanratu area was inundated by a tsunami with a reasonably extreme height because this study used the worst scenario, namely an earthquake with a moment magnitude of 8.8. The earthquake's magnitude significantly affects the size of the tsunami on land. In general, the greater value of moment magnitude and the width of the fault area of the earthquake influences the tsunami height caused by the earthquake (Nainitania & Darmawan, 2020). However, there are also rare earthquake events where the earthquake is not of very large magnitude but can generate high tsunami waves, such as tsunami earthquakes (Bilek & Lay, 2002). This potential big earthquake shows that a thorough evacuation is needed to save people.
Another factor that affected the tsunami height was bathymetry conditions. The height of the tsunami will increase in bathymetry conditions of shallow sea waters or located near bays. Besides, the speed of the tsunami decreases significantly as it approaches the coast, indicating that the bathymetry of the coastal area can influence the behavior of tsunami waves propagating toward the shore. Bathymetric conditions around the bay and Pelabuhanratu District have diverse morphologies. Starting from the shallow coastal morphology, although relatively few, then directly bordering the deep sea (Haryadi et al., 2022). These conditions will cause the rate or speed of propagation of tsunami waves around Pelabuhanratu Bay, especially those heading towards the coast in Pelabuhanratu Subdistrict, to be very fast.
The steepness and gentle morphology of the coast largely determine the inundation of the tsunami to land. The tsunami will not reach far inland on a steep coast because it is retained and reflected back by the coastal cliffs.
Meanwhile, a tsunami could hit up to several kilometers inland on a sloping beach. Based on the results of the tsunami modeling in this study, where the distance of the tsunami reaches 2 km, we can see that the coast of Palabuhanratu has a gentle slope. This condition is also consistent with the results of research from Oktariadi (2009), where the coastal area of Palabuhanratu has a less sensitive or low level of sensitivity to the coastal slopes.

Tsunami Evacuation Route for People in Palabuhanratu.
Evacuating people from tsunami-affected areas to safe places is a necessary and immediate action to save the population. Information on the shortest route and minimum travel time for evacuation play a critical role in the safety of the people to be evacuated (Sudarsana et al., 2013). Therefore, in this study, the determination of the optimal evacuation route was carried out using the Network Analysis -Shortest Path method at 34 points spread across 33 hamlets and 1 Pier in Palabuhanratu Village. Determination of 34 starting points or recommendation assembly points is carried out using the criterion of travel time ≤ 5 minutes from the hamlets centroid point. The location chosen must also be on the main road and crossroads. We select these criteria to represent the conditions where the community generally spends more time around their homes, and the general public meets at main roads and crossroads during evacuation. Each recommendation assembly point will generate the shortest evacuation route/path to one of the nine available evacuation locations/assembly points, along with an estimated time needed to get to the evacuation location. The estimated travel time required to reach an evacuation site assumes that the elderly people walking speed is 0.751 m/s (Purbani et al., 2022) choose the speed value because it is an assumed worst speed value, so it can be used as a reference to ensure the success of the whole community in evacuating.
Based on data processing using the Geographic Information System (GIS), the shortest evacuation route results are obtained as shown on the map, which is shown in Fig. 3. Details of the estimated distance and travel time required by each recommendation assembly point to the nearest evacuation location are shown in Table 2.   Based on the map and table, it can be seen that the evacuation locations from each hamlet are not evenly distributed among the nine existing evacuation sites/ assembly points. This condition could increase the number of people and cause congestion in evacuation sites such as the Taman Tenjo Evacuation Site and the RSUD Assembly Point. In addition, we found that two existing evacuation sites/assembly points were not selected as evacuation destinations, namely the DPRD Assembly Point and SMAN 1 Assembly Point, because their locations were too far from each hamlet.

Distribution of Evacuation Sites
Locations of evacuation sites/ assembly points were determined using field surveys and BPBD Sukabumi Regency data. Then the data is used to determine whether the evacuation location can be reached by the tsunami-affected community in Palabuhanratu Village using service area analysis. This service area analysis is one method for finding service areas around evacuation sites on the network or roads (ESRI, 2019). Service area analysis in this study was based on the evacuation time (ET) parameter using the empirical formula from Post et al. (2009). As which was obtained from the results of tsunami modeling. The technical time for natural warnings was obtained from the sum of the decision-making time from the institution, namely the BMKG, which was 4 minutes, and the notification time from the institution, which was from the BPBD, was around 3 minutes. At the same time, the reaction time is determined by 7 minutes (Yunarto & Sari, 2018). So, the evacuation time parameter (Evacuation Time/ET) in this study is 8 minutes.
The service area analysis results show that some areas were affected by the tsunami but were not covered by the existing evacuation points/assembly points. When the government issues an evacuation order, people tend to evacuate to avoid the coastal area as far as possible (Imamura et al., 2012). This means that people will try to get to higher ground or away from the coastal area as quickly as possible, creating the potential for congestion on evacuation routes. Therefore, this study proposes using vertical evacuation sites so that all people can be safe from a tsunami threat. One solution is to use existing tall buildings in the area as vertical evacuation sites. This can reduce construction costs and land availability. The results of the service area analysis are shown on the map in Fig. 4.

Validation
Validation of travel time in this research was carried out by comparing the evacuation travel time based on mathematical calculations and the result of the field survey. The evacuation travel time was calculated following the optimal evacuation route obtained in this research. We selected three sample locations to collect the data: Pier, Hamlet 27, and Hamlet 33. Then, the result of the time difference between the calculation and field survey data was around 2 minutes in Pier, 7 minutes in Hamlet 27, and 2 minutes in Hamlet 33. The detail of travel time is shown in Table 3. The different conditions of each evacuation route may cause the travel time difference obtained from each location. On average, the travel time difference from the comparison of calculation data and field survey data is around 3 minutes and 40 seconds.

Conclusions
The Palabuhanratu village can be affected by the tsunami, especially from the source of the Sunda Strait megathrust earthquake, which can reach at least a magnitude of 8.8. Based on the analysis of the level of tsunami hazard in Palabuhanratu Village, we found that around 57.33 percent of the village was affected by the tsunami.
The maximum tsunami height reaches between 18 and 22 m, the tsunami arrival time is 22 minutes, and the maximum inundation is as far as 2 km.
Using the shortest path feature in network analysis, we successfully determined 34 optimal evacuation routes to the existing evacuation site/assembly point. The evacuation travel time ranges between 2 and 38 minutes, while the distance varies between 87 and 1638 meters.
We also estimate the evacuation time is around 8 minutes. Based on service area analysis we found that not everyone can reach the evacuation location in time, the finding of this study show the need for an additional vertical evacuation site.