Amira Hankin

From BenningtonWiki
Jump to: navigation, search

return to projects list

FInal Project

Introduction

On October 17th, 1989 at 5 :04 pm a 7.1 magnitude earthquake originating from the San Andreas fault hit the Bay Area. This disastrous event killed 63 people throughout the area, injured over 3,500 and left almost 12,000 people homeless. The earthquake became known as the Loma Prieta, named after its epicenter near the Loma Prieta lockout, one of the most devastating natural disasters to hit the bay area in recent years (1).

Relatively soon after the earthquake, several reports came out in an effort to highlight the underlying cause of the majority of the damage. A report in 1994 (4) detailed that the majority of the damage from the earthquake was the result of liquefaction, the phenomenon where low density soil patches suddenly lose their stiffness and form in response to stress, causing the areas to behave like a liquid. The shaking experienced in a particular region during an earthquake depends very much on the type of deposits found close to the surface of that area (2). Areas that are composed of softer soils and mud shake much harder during and earthquake than areas composed of bedrock (4). One area that was hit particularly hard during the Loma Prieta quake was San Francisco’s Marina district, an area built on filled land (a mixture comprised of sand, rubble, waste, dirt, and other materials that contain a high percentage of groundwater). Ironically, some of this material was left over rubble resulting from the 1906 San Francisco earthquake mixed with other debris laid down in preparation for Panama-Pacific International Exposition in 1915, a celebration that highlighted San Francisco’s resilience to the catastrophe in 1906. During the Loma Prieta quake this material underwent liquefaction, allowing the seismic activity to affect the ground more severely. In total, seven buildings in the district collapsed completely and 63 were deemed to dangerous for habitation (4).

Another area affected by liquefaction was the ground around the San Francisco-Oakland bay Bridge, and as a result the bridge itself. As the soft, muddy soil that the bridge was constructed on shook more strongly than the surrounding regions of stronger ground, a freeway approach to the bridge, as well as a section of the bridge itself, collapsed completely (4).

In 1994, the Northridge earthquake in southern California also damaged many areas due to the same phenomenon, liquefaction. The difference in this earthquake however, was that the majority of the damage was caused by liquefaction induced landslides, triggered by the quake. The Northridge earthquake triggered over 11,000 landslides in an area about 10,000 km squared, the majority of which occurred in areas susceptible to liquefaction and in mountainous, sloped regions. These areas underwent a transformation during liquefaction, causing severe landslide damage to the surrounding developed areas (3).

Could San Francisco also be susceptible to liquefaction induced landslides triggered by the next big quake? San Francisco is a city known for its rolling hills, and highly sloped landscape. It is also clear that history will surely repeat itself during the next large Bay Area earthquake in terms of the resulting liquefaction in areas built upon softer soils. What areas in San Francisco could be susceptible to landslides after an earthquake due to liquefaction? Are these areas inhabited? How developed are these areas and to what extent could they damaged in a future earthquake? These are the questions that drove my research and analysis for this project. I wanted to map the potential high risk areas in San Francisco that would most likely be susceptible to landslides triggered by liquefaction in the event of a big earthquake.

Methods

In order to answer the question of what areas in San Francisco could be susceptible to landslides after an earthquake due to liquefaction, I looked at regions that were already considered liquefaction risk areas in combination with the steepness of the slope in that area.

To begin, I used a compiled dataset (6) of liquefaction risk areas in San Francisco. This dataset was the result of over a decades worth of work by geologists form the US Geological Survey and the California Geological Survey. In this dataset, the researchers used the location and presence of quaternary deposits, where ground sediments are loose and not consolidated, to determine the liquefaction risk potential and susceptibility of areas in San Francisco. The dataset was comprised of seven risk level polygons on a scale from “very low” liquefaction susceptibility to “very high” liquefaction susceptibility. My first step was to re-categorize these polygons into three risk categories “low risk”, “medium risk”, “high risk”, for simplicity. In my analysis, I made the decision that medium and high risk categories were both important for determining general risk areas, regardless of the extent to which they could potentially cause damage, therefore in my final intersection of polygons, I used both high risk and medium risk polygons as general liquefaction risk areas (see figure 1).

The next dataset incorporated into my research was a digital elevation model (DEM) of the San Francisco Bay Area sourced from USGS Earth Explorer (see figure 2). I used this DEM to create a raster file of slope values. From this dataset I categorized the resulting slope values into three categories, using the natural angle of repose of the substrate as defined by Kamp et al. (5) as the basis of these categorizations. The ranges of slope values were simplified into three categories, low slopes with an angle of 0 degrees to 10 degrees, medium slopes with an angle of 10 degrees to 25 degrees, and high slopes with an angle of 25 degrees and above. The slopes in each of these categories was then polygonized accordingly (see figure 3).

The liquefaction susceptibility polygons of high and medium risk level were then intersected with the high and medium slope value polygons respectively (see figure 4), creating new polygons that represented areas of high landslide risk and medium landslide risk (see figure 5).

As a final step, these polygons representing high risk areas were overlaid on top of a vector file of polygons representing categorized neighborhood values: high values of 903k-1.15 M, average values of 493k-903k, and low values of 493k and below. These polygons representing categorized neighborhood values were created by inputting found values for each San Francisco neighborhood into a found neighborhood boundary dataset (see figure 6).

Results

The resulting risk polygons indicating ares of high landslide susceptibility are scattered throughout San Francisco, and are not confined to any one neighborhood in particular.There are fewer high slope, high liquefaction susceptibility landslide hazard polygons than there are medium slope, medium liquefaction susceptibility polygons. Their presence in general, however, indicates that there is a high or medium landslide risk in these areas of San Francisco.

The resulting dataset created from the intersection of these risk areas with neighborhood values highlights does not show a correlation between landslide risk and neighborhood value, however we can glean some information about the potential damage these landslides in these polygons would cause. For example, there are many risk polygons present in the Presidio neighborhood, where there are few inhabited buildings and developed areas, but many large trees, and older historical buildings, which could be threatened during a landslide. There is also a high landslide risk potential in the downtown area of San Francisco, where there are skyscrapers, cable cars that wind themselves up the steep SF slopes, and a high concentration of people every day. These are all connections I would like to explore further.

Conclusion

There is an apparent landslide risk potential throughout San Francisco, scattered throughout neighborhoods and districts. In order to better assess areas that could potentially be damaged during an earthquake induced landslide, or liquefaction more generally it would be helpful to look at other datasets as well.

There are several datasets that I would like to work with in addition to these risk polygons. The first dataset that would be helpful in better understanding where these potential risk areas are in San Francisco could be the location of gas lines that run throughout the city. During the 1989 earthquake a natural gas main ruptured at the intersection of Beach and Divisadero Streets, causing a major fire in the area. During this event, the fire department was unable to run their hoses from a nearby water source because the quake had also disrupted that hydrant system (4). Because of this fact, in the future I would also like to look at the location of San Francisco’s water mains and water pipes to assess any potential damage that could occur to them in the event of an earthquake or earthquake-induced landslide. Additionally, I would like to look at the relative ages of building in San Francisco, and buildings with ground floor garages, as both of these aspects of a building can be important in determining the hazard potential for that structure during an earthquake (4).

Another thing I would like to pursue in the continuation of this project, would be the creation of landslide risk polygons that were representative of a more gradual gradient of landslide risk potential. For example, instead of creating “high risk” “medium risk” and “low risk” polygons, I would like to examine and categorize these risk polygons based on all of the interactions between different slopes and different liquefaction susceptibilities. I could then, perhaps, more accurately see where the highest risk areas were located compared to the medium and low risk areas. More work could be done, using the model I have already built as a starting point, to create a detailed map of the highest risk areas of San Francisco in order to better prepare the people of the city for the next big earthquake.

Figures

References

1. Eberhart-Phillips JE, Saunders TM, Robinson AL, Hatch DL, Parrish RG (June 1994). Profile of mortality from the 1989 Loma Prieta earthquake using coroner and medical examiner reports. Disasters 18 (2): 160–70. doi:10.1111/j.1467-7717.1994.tb00298.x.PMID 8076160

2. Perkins JB. (2001). The Real Dirt on Liquefaction: A Guide to Liquefaction Hazard in Future Earthquakes Affecting the San Francisco Bay Area. Association of Bay Area Governments (ABAG). ABAG Publication Number: P01001EQ.

3. Harp EL. & Jibson RW. (1996) Landslides triggered by the 1994 Northridge, California, earthquake. Bulletin of the Seismological Society of America. Vol. 86 no. 1B S319-S332.

4.Sims JD & Garvin CD (1994) Recurrent liquefaction induced by the 1989 Loma Prieta earthquake and 1990 and 1991 aftershocks: Implications for paleoseismicity studies. Bulletin of the Seismological Society of America February vol. 85 no. 1 51-65

5. Kamp U, Growley BJ, Khattak GA, Owen LO (2008). GIS-based landslide susceptibility mapping for the 2005 Kashmir earthquake region. Geomorphology 101 631-642

6. Liquefaction Susceptibility Map: http://gis.abag.ca.gov/Website/LiquefactionSusceptibility/

Initial Questions

  • How much do Vermont’s watersheds contribute to its water supply systems? How much does groundwater account for in the system? Where are various reservoirs/dams/wells? I’m also curious about pollution levels in fresh bodies of water. Where are these contaminated areas in relation to bodies of fresh water?
    • Some good potential in working with watershed mapping to think about surface water amounts, potential supplies. Dealing with ground-water would get complicated very quickly. Now think in terms of step-wise putting of things together -- what kinds of data sets would you need, how would they be linked in amodel developed to (for example) help a local town manager think about where to put their water intake...Kwoods 00:01, 3 October 2011 (UTC)
  • How closely can you map the historic seas/bodies of water that covered the United States billions of years ago based on current maps of known salt fields/mineral deposits or even fossil beds? Is there a way to predict the areas that these seas covered using topography and historic sea level information? (This question wouldn’t necessarily apply to Vermont…would it?)
  • I’m interested in mapping under bodies of water in order to compare the underwater terrain with organism diversity (would need to narrow this down). I’d love to look at coral reefs, but this could also extend to local bodies of fresh water.
    • These last two are interesting conceptually but VERY broad (both conceptually and spatially), AND probably requiring development of new geospatial data-sets that would be very difficult to acquire. Think in terms of sharper, more constrained targets? Kwoods 00:01, 3 October 2011 (UTC)

Ahankin 02:31, 20 September 2011 (UTC)


Project Proposal

I would like to map earthquake damage susceptibility due to liquefaction-induced landslides in the Bay Area. To keep it short and sweet, my question is essentially:

What areas in San Francisco could be susceptible to landslides after an earthquake due to liquefaction?

  • I'm working on writing up a brief "why this question matters" piece to post on here, to give a little more background about the Bay Area's earthquake damage history.

In order to answer this question (and come up with some sort of shapefile that will allow me to see these hazard areas) these are the datasets that I need/steps to be taken:

Potential datasets to include:

1. Fault lines in Bay Area/areas that are prone to extreme shaking

2. Earthquake liquefaction susceptibility (which I have as a PDF from a study on liquefaction, and just need to georeference at this point!!).

  • This felt like an incredible dataset to find, one of those things you hope exists but think you have little hope actually finding it. I was considering trying to make my own liquefaction susceptibility dataset (kind of where this larger project originated) and thought that I'd need to combine datasets involving soil type, bedrock type, etc. so I'm curious to know more about the step by step process through which the researchers doing this study. These particular maps are the result of over a decade of work and collaboration between geologists from the consulting firm William Lettis & Associates, the U.S. Geological Survey, and the California Geological Survey. As I find out more about this research, I'll continue to post details!
    • Upon further investigation, it appears as though a key factor in determining liquefaction susceptibility is the presence of quaternary deposits....more info about those coming.

3. Topography of the Bay Area/ elevation maps from which I can determine the slope (and thereby categorize various areas using a steep→gradual incline scale). Still trying to figure out a way to do this. Any suggestions?

(this next one I would add as an interesting sort of society-type experiment if I had the time, but it’s not directly related to the sets above)

4. A map of houses/apartments and their market values, which I would overlay with these “landslide danger zones”. I’ve heard rumors that some of the most expensive properties/houses in San Francisco are located in extreme danger zones, and I’d love to look into this.

  • Maybe an addition to this project would be creating an ideal housing/building map for San Francisco based on the areas that are the most stable in terms of their susceptibility to liquefaction-induced disasters.


  • I like this new project idea! All the data sets seem like they will be easily found, probably from the CA clearing house. You might want to also use aerial photos, if that seems like it would be visually helpful to overlay shapefiles on. Edarham2 13:16, 16 October 2011 (UTC)
    • I really like the idea of using aerial photos as a ay of determining elevation (and slope perhaps). In my research I've actually found several papers that mention this as a way of determining the resulting slope of an area after a landslide! I would also still like to be able to determine slope just from a digital elevation map...I know this is something you are also working on....where's a good place to start? Ahankin 02:26, 18 October 2011 (UTC)
      • Photos make useful backgrounds for subjective interpretation; they would be very difficult to use for slope models -- it would tke a lot of photointerp using stereopairs -- it's EASY to get slope directly from DEM in QGIS, however... Kwoods 12:55, 18 October 2011 (UTC)


Annotated Bibliography

1. Perkins, Jeanne B. (2001). The Real Dirt on Liquefaction: A Guide to Liquefaction Hazard in Future Earthquakes Affecting the San Francisco Bay Area. Association of Bay Area Governments (ABAG). ABAG Publication Number: P01001EQ.


This report is good for a more conceptual understanding of the project I’m taking on. The main purpose of the report is to serve as a means for understanding the liquefaction hazard in the San Francisco Bay Area. In this respect, it provides critical background information (mainly historical) for making my project relevant and meaningful.


2. Harp, Edwin L. & Jibson, Randall W. (1996) Landslides triggered by the 1994 Northridge, California, earthquake. Bulletin of the Seismological Society of America. Vol. 86 no. 1B S319-S332.

Harp & Jibson’s paper first gives real-world context to their project, talking about the 1994 Northridge, California earthquake. This was helpful to see in order to better understand some of the context for my own project (although the area they look at is in Southern California and I’m focusing mainly on Northern California). Something very interesting about this particular earthquake was an after effect best explained by the resulting landslides in the area. This unusual effect was an outbreak of Valley Fever (disease caused by inhaling airborne spores), which was thought to be carried in the large clouds of dust created by the landslides resulting from the quake. This paper was also helpful in looking at Harp & Jibson’s methods. In order to determine the scale of the landslides that occurred, they used GIS software to digitize and map aerial photographs that were taken the morning after the earthquake. They used this to determine the volume of sediment that was carried down in the slide as well as the resulting slope of the land. This seemed like a really interesting approach to determining the results of what happened to the land after the landslide. Although Harp & Jibson used their GIS mapping skills to determine other details as well, I’m interested in the connection between the aerial photographs and determining the resulting slope of the land after the slides.

sf neighborhoods: http://gispubweb.sfgov.org/website/sfshare/accept.asp?gisID=772900Menon690041&email=sandeep.menon@ideateinc.com&urlExtract=noUrlExtract

sf neighborhood values: http://www.trulia.com/home_prices/California/San_Francisco-heat_map/