Earthquakes are notorious for what appear to be selective assaults. Why do some buildings seem untouched while others crumble to rubble? The United States Geological Survey (USGS) is using GIS to study the phenomenon of soil liquefaction to understand urban neighborhoods' risk levels for earthquake destruction.

Liquefaction models show where earthquakes could cause the greatest structural damage.

One of the world's most infamous fault lines is the San Andreas Fault. Its reputation is due to the havoc it has wreaked on some of the nation's most well-known cities, such as its demolishing of San Francisco in 1906. A wealth of infrastructure has been built on or adjacent to the fault, such as interstate highways, bridges, and pipelines. The amount of destruction an earthquake inflicts on urban infrastructure is related not only to the shaking intensity of the quake but also to the liquefaction.

Recently, USGS designed a method for assigning hazardous risk ratings of liquefaction to urban areas. By using sensors to determine subsurface soil behavior and GIS technology to calculate and display data, probabilistic scenarios can be created for a given area. Based on geologic data collected on earthquakes in the past, geologists use the technology to model how that earthquake phenomena would affect the exact same area today.

Ground with high liquefaction response is composed of newly formed sediments, such as that on the margins of San Francisco Bay. These young sediments will slow the velocity of the earthquake waves, trapping the energy and causing large amplitudes of shaking. Older compacted materials that over time have densified or transitioned from sand to sandstone and cemented together are less responsive to liquefaction and less hazardous. Ground made up of newer sediments near water will often contain water-filled pore spaces. As an earthquake's energy cyclically compresses these sediments, the pore pressure in the pore spaces increases to a level greater than the confining pressure of the soils above them, causing small geyser-like sand boils that spew water and sand and result in ground subsidence. If buildings and other structures sit atop these hazardous areas, they have a high risk of damage or collapse.

To determine liquefaction probability, USGS scientists have created a liquefaction potential index that indicates what the impact of a given size earthquake from a specific fault will be on a certain geographic area. Liquefaction data is captured from fieldwork using a cone penetrometer with a conical tip, which penetrates the ground at a constant rate. During the penetration, the forces on the cone and the friction sleeve are measured and recorded. The probe is pushed 30 meters into the earth to capture data about the terrain's subsurface properties, such as the relative strength of the soil. During the test, small shear (shock) waves are periodically generated, simulating earthquake energy. The wave forms and their travel times are recorded, providing geologists with data about the relative density of the soil and how fast seismic waves will travel thorough it.

The liquefaction that affects the human-built environments is mostly limited to the upper 15 meters of soil, but USGS performs testing down to 30 meters so that it can accurately determine the velocity of the material through which the waves pass. Approximately 30 of these tests, or soundings, are taken in each type of geologic environment within the study area. The data is recorded and stored in the liquefaction database and input into the USGS's ArcGIS for processing, calculating, modeling, and creating 2D and 3D visualizations.

U.S. Geological Survey truck uses cone penetrometer to record soil liquefaction characteristics.

The first part of the model calculates the nearest distance from the fault to each 50-meter grid point within the study area and populates a table with that information. The grid is attributed with geologic data, including subsurface velocity values, water table depth, and the liquefaction characteristics for each type of geologic unit. Then the model factors earthquake magnitude levels into the equation and calculates the peak ground acceleration at each grid point. Finally, the GIS model combines all the aforementioned data and calculations to calculate the liquefaction probability at each 50-meter grid cell. The results are valuable to city planners and transportation engineers, who can factor in the index rating when choosing site locations for buildings, roads, and bridges. The model has also attracted the attention of the Nuclear Regulatory Commission. It is interested in using the liquefaction potential index for planning nuclear generator site locations. Nuclear generators are often built near water on potentially saturated, young, liquefiable deposits.

USGS geologist Tom Noce explains, "We use GIS software to automate our velocity map processes, allowing us to create more detailed models. We are also moving toward creating a new type of earthquake response map, the LiqueMap, which displays near real-time projections within 20 minutes of an earthquake. After a disaster, first responders will be able to view data via the Web and immediately anticipate damage to vulnerable infrastructure, such as pipelines and bridges."

Cities continue to change, but the threat of earthquake and liquefaction remains. What if the 1989 Loma Prieta earthquake that brought down part of the San Francisco Bay Bridge happened today? What damage to the city could be expected? GIS temporal models bring together the past and the present. USGS scientists join liquefaction data captured from the Loma Prieta earthquake with current liquefaction data and overlay as-built urban structure data to show the impact that the earthquake could have if it occurred today. The combination of historical data and probabilistic modeling displays a high level of correlation, indicating that the models should indeed provide a good estimation of the effects of similar earthquakes.

Scientists continue to add greater depth and complexity to their studies of earthquakes and liquefaction in the New Madrid, Missouri, seismic region of the United States. During the series of earthquakes of 1811 and 1812, there was liquefaction over very large areas that are now rapidly becoming more urbanized. The USGS team is also creating these temporal what-if maps by inputting 1906 San Francisco earthquake data into the model and creating comparisons with today's downtown Oakland, California, data to see what the effects would be on various areas of the city. Next on the itinerary is the 1906 data's effect on California's Santa Clara Valley, along with scenario maps for anticipated earthquakes on the nearby Hayward and Calaveras faults.

Ultimately, USGS' goal is to provide the public and state and local officials with the information necessary to mitigate liquefaction hazard zones prior to an earthquake and be able to respond to possible liquefaction after an earthquake.

More Information

For more information, contact Tom Noce, USGS geologist/earthquake hazards team (e-mail: tnoce@usgs.gov).