GIS Analyses

The concepts and skills involved in these analyses can be found in full description and with illustration in our GIS Concepts section.

Annual Economic Loss for Counties in California:

The county_california.shp shapefile was uploaded into ArcMap. It was projected from the GCS North American 1983 to North American Lambert Conformal Conic coordinate system.No additional shapefiles were needed for this map because the analysis was merely classifying each county according to the amount of money lost in millions. By right clicking on the shapefile we accessed the attribute table.

Once opened, we were able to navigate to the options menu at the bottom and click "Add Field." We labeled this new field E-loss to represent the attribute we were representing. Each county was designated a number (from 1-9) based on the expected amount lost due to an earthquake. Number one represents 0-1 million while 9 represents 500-900 million. The information for each county was provided by the California Geologic Survey.The map was color coded for visual efficiency.

Humboldt County Analyses:

To start, we needed first shapefilse of only Humboldt County, fault lines within the county, and city polygons within the county. Humboldt County layer was obtained by selecting Humbolt within the California Counties layer in ArcMap, then right-clicking on the California Counties layer, selecting 'Data' then finally 'Export Data'. With this new Humboldt County layer we could then clip both the California fault lines layer and the California cities layer to the extent of Humboldt County.

For the slope instability and liquefaction zoning we had three separate shapefiles from the Humboldt County Community Development Services website. For each of these we separated out each type of slope instability (1, 2 and 3) and whether the land had potential for liquefaction. A new layer was created for each zone from each inital shapefile. Then we Merged (Toolbox/Data Management/General/Merge) the layers together to result in a new layer of each type of zone. The original shapefiles and intermediate layers were then removed.

Then we clipped the cities to the extent of each zone to characterize the city land. From this clip we could get the summary statistics on the resulting land area: the total land area in each risk zone, and also the percentage of city land in each risk zone (in Excel).

The final part of the Humboldt County risk analyses was to look at where community points were within the county in relation to the zones and fault lines. These community points consisted of government building, airports, hospitals, schools, parks and golf courses primarily. First we did a Near calculation (Toolbox/Analysis Tools/Proximity/Near) to determine the distance from each point to the nearest fault line. The results of this show up in a new field in the attribute table for the input layer (community points in this case). The results were then converted to miles and averaged by type of community point (in Excel).

We employed another, though uncommon, use of the Near function in determining which risk zones the community points exist. By computed a Near distance with the community points and a given zone layer, if the resulting ouput distance was "0" in the table, then that point was in that specifc zone. We summarized how many points were in each layer in a table along with the fault distances (again, in Excel).

Bay Area Assessment:

The eight counties designated as the Bay Area counties were isolated from the California shapefile that had all counties listed and copied to a new data file. Using the options key a new field was added named "CGS_to_per" the represented the total percentage of acres that resided within the California Geologic Survey zone. The actual data used for calculations was uploaded into excel and the calculations were performed using the following equation:

(acres within CGS zone)/ (total county acres) * 100

The results from these calculations were uploaded into the Bay Area attribute table using the editor toolbar. In order to portray the results in a manner that captured the audience different colors were used to represent different percentages using the Symbology tab.

Central California 2004 Earthquake Shakemap analyses:

Initially we created new layers for the counties that the shakemap overlapped from the California Counties shapefile, in a similar manner to how the Humboldt County layer was created. Additionally, we created clips of the fault lines, California Cities and even a county layer clipped only to the extent of the Shakemap (a square, depending on its projection).

First for the main Shakemap investigation: the Peak Ground Acceleration (PGA) assessment, we first classified the data into three breaks, as indicated by the USGS and other literature. These classes are: 0-10%g, 10.1-60%g and 60-max%g. The maximum for this particular earthquake was 226%gravity. We created a new layer of each of these classes by selected the polygons within each class in the attribute table, and exporting the data to a new layer. We then clipped the Central California County cities layer to the extent of each of these classes. From these clips we could get summary statistics on total city area. This data was then used in Excel to determine percentages of total area. Note though that this could easily be done in ArcMap by creating and new field (as described in the Economic Loss section), then using the field calculator to figure the percentage for you.

For the sake of risk illustration side by side with the PGA map, the shakemap layer of Peak Ground Velocity (PGV) was classified into two breaks : 0-10cm/sec and 10.1cm/sec and up. As discussed in our Results sections, 10cm/sec is the velocity threshold at which liquefaction may occur in the event of an earthquake. No further analyses were performed on this data.

A final minor analysis of this area involved Housing Unit data from the US Census Bureau. The housing unit (HU) data (essentially number of housing structures in a given county in a given year) was in a separate California county shapefile then our base county shapefile. We first created a new field in this HU attribute table, then used the field calculator to determine the change between HU in 2005 and 2003 (simple subtraction). Then we imported this data to our base California County attribute table by using a join. We joined the tables based on county name. From there we created a clip of the county shapefile to only include the eight counties affected by the 2004 earthquake, then classified the housing unit changes into four natural breaks.

Last Modified December 7, 2010.