RoadNet > Communications 

Navigating Urban Disasters:
The Importance of Critical Roads in Tacloban City’s Evacuation Network

Jaderick P. Pabico
Institute of Computer Science
University of the Philippines Los Baños

5th International Conference on Poverty Alleviation and Sustainable Development
Summit Hotel, Tacloban City, Philippines
16-17 October 2024

Abstract

Critical road segments in a city's road network are essential pathways that, if disrupted, can significantly increase travel distances for residents. This issue becomes particularly concerning during emergencies, as evacuation centers are not always easily accessible, especially for vulnerable populations, including those living in poverty or urban areas with severed routes. To explore this, we utilized data from OpenStreetMap to map the walkable urban road network of Tacloban City, representing it as an undirected graph, G. In this graph, intersections and dead-ends are vertices, and road segments are edges. We included private and public pathways accessible to pedestrians as part of the walkable roads. Using the Philippine Institute of Volcanology and Seismology's (Phivolcs) geotagged database, we pinpointed the locations of evacuation centers in the city and added them as sink vertices in G. The rest of the vertices served as source vertices, representing the population's starting points during an orderly evacuation along the network's shortest paths. Our study employed a computational data science approach to assess how failures in critical road segments could affect evacuation times, particularly focusing on the elderly as the slowest-moving group. The outcome was a series of isochrone maps that visually display areas where travel times to evacuation centers are increased. These maps provide valuable insights for disaster planning and decision-making, aiming to enhance the city's resilience and inclusivity in disaster response efforts.

Keywords: local bridge, safe space, travel time, isochrone, maps

 Rationale

• Poor communities are more at risk during disasters because they often live in areas that are unsafe (e.g., flood zones) (UNDRR, 2015; Wisner, et al., 2004).

• Evacuation centers may be far or hard to reach for people living in poverty, especially in rural or remote areas (UNDRR, 2015; Wisner, et al., 2004).

• Transportation is a big issue — those in poverty may not have access to cars or money for transportation during an evacuation (Wisner, et al., 2004).

• Information is limited for poor populations, making it harder for them to prepare or evacuate in time (Wisner, et al., 2004).

• Experts recommend more inclusive disaster planning that ensures all groups, especially the vulnerable, can access evacuation centers easily (UNDRR, 2015).

Objectives

• Map the Walkable Urban Road Network: To create a detailed map of walkable roads in Tacloban City representing the map as an undirected graph;

• Identify Critical Road Segments: To determine the key road segments in the city that, if blocked, would significantly increase evacuation times for the population;

• Analyze Evacuation Accessibility: To assess the impact of road blockages on the population’s ability to timely reach evacuation centers, with a focus on vulnerable groups like the elderly;

• Develop Isochrone Maps: To create visual maps that show the areas in the city where travel times to evacuation centers increase due to critical road failures; And

• Support Disaster Planning and Decision-Making: To provide tools and insights that can improve disaster response planning, making it more inclusive and resilient for all residents.

Methodology

1. We mapped the Tacloban City walkable roads using OpenStreetMap (OSM) (OpenStreetMap contributors, n.d.) and abstractly treated it as a graph G(V, E), where V is the set of intersections and dead-ends, while E is the set of road segments.

2. Using data science methods, we identified the critical road segments by applying the concept of local bridges from graph theory and calculating the betweenness and closeness centralities of G.

3. We conducted accessibility analyses to evacuation centers: 

   • We identified the respective locations of evacuation centers in the map using Phivolcs’ Database of Geotagged Evacuation Centers (Alejandria, et al., n.d.) and considered these locations as sinks in G.

   • We considered all intersections in V as proxies for sources in G (i.e., where individuals flowed from a source to a sink via the shortest path during an orderly evacuation).

   • We conducted flow simulations using the average walking rate of the slowest age group (Fidel, et al., 2021) in the population (i.e., the elderly) and identified the time from sources to sinks.

4. To see how blocked critical roads affect evacuation times, we repeated the flow simulations by randomly blocking 10%, 20%, 30%, and 40% of these roads. The results are visualized as isochrone maps.

Results

Figure 1 shows the zoomed-in section of Tacloban City’s Walkable Map (TCWM). Overlaid on this OSM map are the V and E in blue. OSM can identify the parts of the road system that are accessible to pedestrians. The walkable road network parameters are as follows:

• • Number of intersections and dead-ends |V|: 7,724

  • Number of road segments |E|: 18,694

Efficient computational techniques for computing the local bridges in graphs provide a set of critical road segments. Centrality metrics, such as the betweenness and closeness centralities,  provide objective measures that rank the important critical road segments. Figure 2 below shows the nodal betweenness centrality of TCWM. The betweenness centrality values range from 0 (dark blue in the figure) to 0.35 (red in the figure). A betweenness centrality of 0.35 means that 35% of shortest paths between all pairs of intersections pass through that node. That is to say that if one is to walk between any pair of nodes within TCWM, 35% of the time that node will be along the shortest path that will be chosen. Note that the road segments with the highest betweenness centrality values are along the Pan-Philippine Highway (Asian Highway 26).

Figure 3 below shows the nodal closeness centrality of TCWM.  The nodal closeness centrality of a node n is the reciprocal of the sum of the length of the shortest paths between n and all other nodes in the graph. This means that the more central n is, the closer it is to all other nodes. Dark red nodes have the highest nodal closeness centrality, while blue nodes have the lowest. Note that the intersections within the Tacloban City's downtown area have the highest nodal closeness centrality.

Figures 4 to 7 below show the respective isochrone maps of TCWM when 10%, 20%, 30% and 40% of critical road segments have been deemed unpassable to pedestrians. On the OSM maps, green edges are still passable road segments, red edges are blocked critical road segments, colored nodes represent intersection travel times to the nearest evacuation center: blue (30 minutes), purple (25 minutes), magenta (20 minutes), light red (15 minutes), orange (10 minutes), and yellow (5 minutes). Intersections without colored nodes mean that the travel time to the evacuation center is more than 30 minutes for the elderly group. The evacuation centers are located within the yellow nodes (nodes that are colored in white but are not clearly seen in these visualizations).

Implications

The following are some of the many practical useful insights that we can obtain from correctly reading the maps:

• Identify Vulnerable Areas: The maps can highlight which neighborhoods or zones experience the greatest increases in evacuation times as more roads are blocked. This allows planners to identify and prioritize areas where intervention is most needed.

• Optimize Evacuation Routes: By visualizing how road blockages affect travel times, authorities can design alternative evacuation routes and contingency plans to ensure people can still reach safety in case of widespread road failures.

• Enhance Public Awareness and Preparedness: These maps can be used for public awareness campaigns, educating residents about which roads are most critical and the importance of having alternative evacuation plans in place.

• Support Policy-Making: The insights from the maps can inform disaster response policies, encouraging the LGU to improve transportation networks, build more evacuation centers, and ensure that evacuation plans are equitable and accessible to all.

References

• HCM Alejandria, MJN Bernardo, KRC Peña, HPT Abalos, JVOQ Samorano, YPM Aquino, CAM Favis, MT Cahulogan & RJU Solidum. n.d. Geotagging and Multi-hazard Assessment of Evacuation Centers in the Philippines. Philippine Institute of Volcanology and Seismology (Phivolcs), 15pp.

• MBCC Fidel, CB Gonzales-Suarez, AR dela Cruz, EA Roxas, MR Fernandez & CG Cruz. 2021. Spatiotemporal parameters of gait in Filipino adults using the 3-d motion capture system. Journal of Medicine University of Sto. Tomas 5(2):744-754 (DOI: 10.35460/2546-1621.2021-0019)

• OpenStreetMap contributors. (n.d.). OpenStreetMap. Retrieved from https://www.openstreetmap.org.

• United Nations Office for Disaster Risk Reduction (UNDRR). 2015. Global Assessment Report on Disaster Risk Reduction. United Nations, 316pp (ISSN: 9789211320428).

• B Wisner, P Blaikie, T Cannon & I Davis. 2004. At Risk: Natural Hazards, People's Vulnerability and Disasters. Routledge: London, 496pp (ISBN: 9780203714775).

Acknowledgement

This research endeavor is funded by the Institute of Computer Science (ICS) core-research fund under University of the Philippines Los Baños (UPLB) fund code 2326103.

Citation

JP Pabico. 2024. Navigating urban disasters: The importance of critical roads in Tacloban City’s evacuation network. 5th International Conference on Poverty Alleviation and Sustainable Development (iPOVCON 2024), Summit Hotel, Tacloban City, Philippines, 16-17 October 2024.