GEOG 566






         Advanced spatial statistics and GIScience

May 22, 2017

Tutorial 2: Automated Autocorrelation of Tsunami Inundation Time Series at Various Points of Interest

Filed under: Tutorial 2 2017 @ 12:11 pm

Question

Tsunami inundate shorelines in a chaotic manner, resulting in varying degrees of flooding from location to location. The spatial variability of tsunami hazards has important implications for the tsunami-resistant design of structures and evacuation plans because different safety margins may be required to achieve consistent reliability levels. Currently, tsunami mitigation strategies are primarily informed prediction models due to the lack of nearshore field data from tsunami disasters. Thus, to investigate the spatial variability of tsunami hazards, we analyze the relationship between different test locations or points of interest (POIs) in our domain. The question is:

“How much does tsunami inundation vary spatially?”

Essentially saying if one observes a certain flooding depth at their location, could they assume that a similar flood level is occurring elsewhere? And how does this change with time? And how to certain coastal features affect this?

Tools

To answer this question, I examined the autocorrelation of the time series of inundation at various locations as well as the difference in inundation levels between these locations. In this case, 12 locations were selected and thus the process to analyze all of these locations was automated. All of this was implemented in R.

Methods

As before, the tsunami inundation data is stored in a large NetCDF file. Thus we use the same procedure as mentioned in Tutorial 1 to read the NetCDF file into R. This process requires the “ncdf4” package to be installed.

For this analysis, we begin by selecting the points of interest along the coast near Port Hueneme, California. These points were arbitrarily selected but were also ensured to be somewhat well distributed along the coastline and were within the domain of the dataset. Figure 1 shows the locations of the points of interest (POIs) along the coast of Port Hueneme, CA. A total of 12 points were selected and were all meant to represent some building or port location. Their names and geographical coordinates were stored in a .csv file as shown in Figure 2.

 

Figure 1 – POIs where time series of inundation were extracted along coast near Port Hueneme, CA

Figure 2 – CSV file storing locations and their geographic coordinates

The .csv file was read into R by using the “read.csv” function and this table was subsequently converted to a data frame as shown by the code in Figure 3. Next, the “which.min” function from the “stats” package was used in a for loop to find the indexes. Next a loop is used to extract all the different time series and plot them accordingly. Figure 4 shows the time series of inundation for each of the locations stored in the CSV file. Note that the first 3 locations are in the ocean and the rest are land locations.

The time series data was stored in a separate data frame and then the “acf” function from the “stats” package in R was run in loop form to obtain the autocorrelation values. THis was followed by using another loop with a nested loop to take the difference in inundation value between each of the locations and the autocorrelation function was applied to these as well. The code for these procedures is shown in Figure 3.

Figure 3 – R code used to read in .csv file and procedures to perform autocorrelation analysis

Figure 4 – Time series of inundation for each location shown in Figure 1

Results

Figure 5 shows the autocorrelation plot for each of the inundation time series shown in Figure 4. These plots show that there are indeed variations between the different locations. The autocorrelation plots also appear to vary more as the POIs get further away from one another.

Figure 5 – Autocorrelation plots of the inundation time series from Figure 4

Figure 6-8 show some examples of the autocorrelation plots of differences in inundation level between two locations. Figure 6 shows the autocorrelation of the difference in inundation level between Port Hueneme and all of the other locations. A placeholder image is shown in place of the difference between itself (which doesn’t exist). Figure 7 shows the same plot but for Del Monte Produce. And Figure 8 shows this for the naval base Navsea 1388.

There are 9 more of these plots that were created by this process, but in the interest of not inundating this report, only 3 are shown. Each of these figures shows rather interesting differences for the different locations.

Figure 6 – Autocorrelation plots between difference in inundation levels for all locations relative to Port Hueneme

Figure 7 – Autocorrelation plots between difference in inundation levels for all locations relative to Del Monte Fresh Produce

Figure 8 – Autocorrelation plots between difference in inundation levels for all locations relative to Navsea 1388

Critique of Method

The method provided a very convenient way of viewing the different locations and their associated autocorrelations with the touch of a button. In the end, POIs can be easily added or removed by modifying the CSV file and the source code should be able to adjust accordingly without any modification to the code. This feature has already been tested and shown to be quite robust for handling any number of locations with coordinates within the data domain. This shows that this procedure is efficient and effective. Overall, I was very pleased with the operational capability of this method.

Observing the autocorrelations for each of the  locations and for the inundation differences between the regions was quite interesting. From a glance, I could immediately tell that there were differences between the POIs. I think this information will indeed be quite useful, though my experience with interpreting these types of autocorrelation plots will require some additional thought.

Print Friendly, PDF & Email


No Comments

No comments yet.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

© 2019 GEOG 566   Powered by WordPress MU    Hosted by blogs.oregonstate.edu