Lincoln Count Tsunami Evacuation
INTRODUCTION
Imagine a 100 foot wall of water rising out of the deep ocean in less than a few minutes notice and crashing down on you with even less warning, smashing buildings, hurling debris, and dragging unfortunates back out to see, still struggling for shore. These are tsunamis. The Pacific is a particular hotspot for both locally and distantly generated tsunamis as 59% of all tsunamis have been generated in the Pacific Ocean (National Oceanic and Atmospheric Administration [NOAA]). Distantly generated tsunamis can often be detected hours or even days before their impact. Local tsunamis can occur in a matter of minutes, though, often without time for proper warning or evacuation (NOAA). Oregon’s location on the Cascadia subduction zone where the Juan de Fuca oceanic plate is being forced beneath the continental North American plate puts Oregon at imminent risk of distant and locally generated tsunamis (United States Geological Survey). The interplate thrust generated from the stress and grind of these plates has the potential to initiate a tsunami, which many scientists say is inevitable (Colella). In this event, often referred to as the Cascadia Megathrust, it is of utmost importance Oregon is prepared with a tsunami evacuation plan to insure peoples’ safety. However, this is easier said than done. Many variables such as wave speed, coastal depth, wind, and even temperature can affect the magnitude of a tsunami (Bureau of Meteorology).
To prepare for the different scenarios possible in a tsunami event, scientists use GIS MCE modeling to plan evacuation centers. MCE is a Multi-Criterion Evaluation system employed by GIS (Geographic Information Systems) scientists that allows them to take multiple variables into account and the comparative importance of each variable in relationship to the others when predicting an event. In tsunami events, distance inland and elevation covered by water, and proximity to roads becomes extremely important. Our model seeks to account for these different scenarios and ensure civilian safety by proposing multiple tsunami evacuation center locations in Lincoln County, Oregon, which lies at low elevation with multiple inlets in a high-risk zone.
METHODS
In order to examine where evacuation centers should be placed, Lincoln County was spatially isolated, including all of its roads, rivers, streams, and other waterbodies. Apart from these objects, land elevation of the features also had to be taken into account using a Digital Elevation Model (DEM) or Oregon clipped to the Lincoln County perimeter. Within Lincoln County, I identified areas that would be impacted by a tsunami by creating a two-mile radius around water features to reflect how far inland a tsunami can travel (NOAA). I also created a one-mile area from water features that reflects how far a person can run in the event of local tsunami before being overtaken by the water. This one-mile radius also represents the area from which suitable locations for tsunami evacuation centers should be drawn so people have a feasible chance to reach them. Next, the least costly path from rivers, streams, and waterbodies were calculated that takes into account horizontal and vertical distance to reflect the actual land surface. While it is important locations are unsubmerged, it is also critical that they are accessible. The least costly path distance from roads was also calculated, but reversed to reflect that distances closer to roads would actually be safer than distance away (opposite of when considering waterbodies). For people who are disabled, elderly, or young, travelling over non-road surfaces may be very difficult. I set a maximum distance of half a mile people could travel to a road. Apart from linear distance, this analysis also took into account elevation that would increase the difficulty (cost) of travel. In order to draw relationships between rivers, streams, waterbodies, and roads, values were standardized (mean of 0 and standard deviation of 1). It was particularly important to reclassify slope values to exclude elevation below 100 feet that is considered in the tsunami hazard zone (NOAA).
The collection of rasters now created provides the information and tools to assess where evacuation centers can be placed. However, not all factors are of equal importance. To account for this, attributes can be assigned different weights to determine how heavily one attribute will be considered in relation to the others. In my primary model, proximity to roads and elevation were considered more heavily than inundation of streams, rivers, and waterbodies. Roads were given a weight of 30; elevation 25; rivers 16; waterbodies 16; and streams 13 [Figure 1]. (I waited streams less because they are smaller than rivers and other waterbodies so I believe would have less of an impact.) There are many different factors and scenarios, though, under which a tsunami can occur. So, I made additional models that weighted variables differently to represent different scenarios. In my second model the effects of water were considered most heavily. Roads were given a weight of 15 and elevation 15, while flooding of waterbodies, streams, and rivers were considered more heavily with weights of 25, 20, and 25 respectively [Figure 2]. In a third model, I considered the possibility of people being able to travel more easily on foot over non-road terrain and weighted slope and water variables much more heavily than proximity to roads. Slope and waterbodies were given a weight of 25, streams and rivers were weighted at 20, and roads were only given a weighted influence of 10 [Figure 3]. Below is a diagram of the methodological framework used to establish these weights and their results:
Input attribute (roads, rivers, waterbodies, streams, and elevation) were reclassified to have common values, then assigned individual weights to provide the most suitable areas for evacuation centers.
After incorporating this information into my map, I visually analyzed the suitable areas shown by each model to identify the best locations for evacuation centers. Once they were identified, I looked at satellite imagery using Google Maps to examine the terrain of the area and positions of roads.
RESULTS
Images of suitable evacuation areas based on differently weighted models are shown below. Weights of the roads, elevation, rivers, waterbodies, and streams are given in the upper left corners.
Figure 1
Figure 2
Figure 3
Together, these three maps show the impact of weighting variables differently. Depending on the type of coastal terrain (steady vs changing elevation, easily traversable vs non-traversable ground surface), each could be considered more applicable than the other. For Lincoln County, the first map in Figure 1 with an emphasis on elevation and proximity to roads will most likely be the most helpful when deciding where to locate evacuation centers because the terrain is relatively unapproachable on foot. Figure 4, below, depicts where the best locations centers should be placed based on the weighting scheme of the first model, best suited to Lincoln County.
Figure 4
These models provide realistic location of potential evacuation centers. However, it is still important to look at real-world imagery. Satellite Imagery from GoogleMaps exposes some problems of relying solely on MCE modeling for placement of evacuation centers.
Satellite imagery also allows us to observe the terrain of an area to conclude how accessible it would be on foot and to look for any additional roads not included in the Oregon database the “roads” information was taken from, but still usuable during an evacuation.
It also allows us to see potential problems unlisted by a model using purely numbers.
Overall, using the maps generated from weighted models is an informed and useful way to begin constructing evacuation plans of events yet to occur. Weighting variables differently can also help us imagine the variability of tsunami events in order to be prepared for multiple scenarios. However, other resources should be incorporated with these models to increase their accuracy. Apart from problems I spotted from satellite imagery, like roads on the wrong sides of rivers, there are other elements that could be added to this analysis to increase its usefulness. Including locations of pre-existing emergency centers like hospitals would help narrow down the best locations for evacuation centers. Population density is another factor that would contribute greatly to where there should be a concentration of evacuation centers and/or larger centers to accommodate residents.
CONCLUSION
In the scenario of a locally generated tsunami, there are locations for at least a dozen evacuation centers in Lincoln County that are within a half mile of road access, above 100 feet elevation, and two miles outside water areas that could flood. These are only the best sites, though, and additional sites could be established at slightly less suitable areas that would still be beneficial to have in the event of a tsunami. As shown with the variously-weighted models, it is difficult to predict how a tsunami will hit and depending on which variables have the most weight the evacuation areas will vary. Being able to observe the corresponding outputs of differently weighted scenarios is vital to prepare for every event and insure safety of citizens. The tsunami evacuation maps currently provided by the Oregon Department of Geology and Mineral Industries only provide the worse scenario evacuation plan. While it is good to prepare for the worse, considering smaller tsunami events is also important. In a scenario with less flooding and inundation evacuation centers could be placed closer to the coast, rivers, streams, and other waterbodies to reduce the walking distance of people to relief. In the Pacific Northwest, located on a known fault line, it is crucially important to consider what our evacuation plans will be in the event of any tsunami, large or small, and where our corresponding relief centers should be located to minimize loss of life and maximize safety.
REFERENCES:
Colella, Harmony V., James H. Dietreich, and Keith Richards-Dinger. “Spatial and Temporal Patterns of Simulated Slow Slip Events on the Cascadia Megathrust – Colella – 2013 – Geophysical Research Letters – Wiley Online Library.” Spatial and Temporal Patterns of Simulated Slow Slip Events on the Cascadia Megathrust – Colella – 2013 – Geophysical Research Letters – Wiley Online Library. Geophysical Research Letters, 2013. Web. 15 Nov. 2015.
National Oceanic and Atmospheric Administration. “Lincoln County.” Google Maps. Google, 2015. Web. 15 Nov. 2015.
“NOAA Tsunami Website.” NOAA Tsunami Website. National Oceanic and Atmospheric Administration, 2015. Web. 15 Nov. 2015.
United States Geological Survey. “Oregon Earthquake Information.” Oregon Earthquake Information. United States Geological Survey, 18 Apr. 2014. Web. 15 Nov. 2015.
Bureau of Meteorology. “Tsunami Frequently Asked Questions.” Tsunami Frequently Asked Questions. Bureau of Meteorology, 2015. Web. 15 Nov. 2015.