USGS, Science for a Changing World, Earthquake Hazards Program, Ground Motion

Configuration of the Seattle Urban Seismic Array for the February 28, 2001, M6.8, Nisqually Earthquake and its Aftershocks


David Carver1, A. D. Frankel1,W. T. Bice1, B. Norris2, and N. Dickman1

Open-File Report 01-344

1 Golden, Colorado
2 Seattle, Washington

thumbnail of the Seattle Urban Seismic Array mapThis is a description of the configuration of the Seattle Urban Seismic Array (click on thumbnail map to see full size map) at the time of the Nisqually earthquake and its aftershocks (from April 28 through March 3, 2001). The array is operated by the Geologic Hazards Team of the U. S Geological Survey. It has been recording small local earthquakes since its initial installation immediately after the Duvall earthquake in May of 1996. Since then, we have been able to convert the array to all Kinemetrics K2s after retiring the aging DR-200s in June of 1998. The array stations have been moved around to suit our research interests. We had 26 instruments in urban Seattle at the time of the Nisqually earthquake.

The primary research goal of the array is to record earthquake ground motion at sites scattered around the city that represent the major types of surficial geologic materials found in the city. With this information we plan to draw seismic hazard maps that will show the expected variability of ground shaking throughout the city. Our strategy to achieve our goal includes:

1)    sampling ground motion at sites that represent the major geologic materials including man-made fill and alluvial surficial materials,

2)    recording high-resolution ground motions from a tripartite array (SDN, SDS, and SDW) with about 200 meter spacing located on man-made fill in the SoDo (South of Downtown) industrial district,

3)    recording ground motion in areas with man-made fill and alluvial surficial materials which have a high potential for liquifaction,

4)    sampling ground motion at two sites on Capitol Hill (THO, and SEU) that are located near drill holes that have been logged for seismic velocity,

5)    Providing long-term continuity at a few sites for comparison and reference, especially SEW and ALK, which are located on Tertiary bedrock.

6)    recording ground motion at sites with similar surficial geology both inside and outside the Seattle Basin

The configuration of the array is summarized in the accompanying table which is further described below.

Station Locations

The K2 has a roof-mounted GPS (Global Positioning System) antenna and an internal GPS board that continuously calculates the location of the instrument and averages the results to increase the accuracy of the location. In the table, we report latitude and longitude determinations that are the average of at least 1500 locations from the GPS. These averaged locations have error ellipses that are less than 20 meters in diameter. The locations refer to the WGS84 datum. GPS errors are large in the vertical dimension so elevation of the recording sites is taken from the 7.5 minute USGS topographic quadrangle maps.

Most of the stations are sited on private property. We have not asked owners for permission to release their names or addresses.

Sensor Orientation

Sensor orientations are listed for each station in the configuration table using the SEED convention. The SEED convention refers to dip and azimuth. For example, " -90, 0" indicates a vertical orientation with ground motion up yielding a positive voltage from the sensor. An orientation of  "0, 180" indicates a horizontal sensor with ground motion south generating a positive voltage. 

An error has been found in the orientation of the sensors at ALK. The channels were misidentified but the error has now been corrected in the table.

Timing Quality

The K2 GPS system also provides precise UTC time to compare against the K2s internal Real Time Clock (RTC). The K2 continuously synchronizes the RTC with UTC time from the GPS. When the RTC drifts more than 3 milliseconds from UTC, the clock is reset. The result is that the K2 timing is always within 3 ms when the GPS is acquiring a satellite time signal. If there is no GPS signal, the RTC drifts without control and there is no way to determine the clock correction between UTC and the RTC. In that case, the timing is probably acceptable for some non-critical uses but the correct time is unknown. The only instrument that was not acquiring a GPS signal during the Nisqually earthquake sequence was MAR. This is noted in the configuration table with a “U”. All of the other instruments were receiving a GPS time signal at the time of the Nisqually earthquake. Therefore, the timing quality is designated as excellent (“E”). The K2 sample rate is given in the table as samples per second. All channels are sampled simultaneously.


All of the Seattle instruments are Kinemetrics K2 ( ) with internal Force Balance Accelerometers (fba_23) or the newer Episensor accelerometers. They are all triaxial 2 G sensors. Some of the K2s record 3 channels and others record 6 channels. The six-channel instruments record acceleration on the first 3 channels and velocity on the last 3 channels. All of the K2 gain boards are set to unity gain. The external sensors are all velocity transducers, either Sprengnether S-6000s or Mark Products L-22s.  Sensor response data for all of the transducers are given in the poles and zeros file that accompanies the data served up by the database at the web site (see Data Availability below).

All of the K2s have locking nuts installed on their leveling feet to make them rigid. The feet are epoxied to a concrete slab at the site.  All of the instruments were checked to make sure that they were firmly attached to the concrete slab both before and after the Nisqually earthquake.

Soil Type

Soil-type classification is taken from Waldron and others, 1962, “Preliminary geologic map of Seattle and vicinity”, USGS, Miscellaneous geologic investigations map, I-354.

The array has two rock sites (SEW and ALK) for use as references in site response studies. SEW is installed in a small one story brick building in Seward Park. The entire south end of the park’s peninsula is mapped as Tertiary sandstone and there is an outcrop about 500 meters west of the recording site. The building itself is built on colluvium that has washed down the hillside and no bedrock is visible at the site. The accelerometers at SEW are installed on the concrete slab foundation of the building. The velocity sensors are buried about 1 meter deep in the colluvium 6 meters up the slope from the building. ALK is probably the best rock site. There is a Tertiary sandstone outcrop about 20 meters away and rock is only a few centimeters deep under the building. Station HAR is located on the northeast corner of Harbor Island. During the Nisqually earthquake, Bob Norris witnessed a large sand boil very near the recording site.

Williams et al. (see presents near-surface P- and S-wave velocity measurements that they have made at several of the array sites.

Structure Type

Our goal is to find small one-story structures for recording sites. However, sometimes compromises are made to get a site in a specific location. The compromise at KDK was putting the instrument in a basement so we could get a location very close to the King Dome. MAR is located in downtown Seattle in a 3-story concrete post and beam structure at the same level as the alley. The front of the building is one story higher. Station NOR is located in the basement of a 1-story brick structure. The concrete basement floor has subsided and was probably not firmly attached to the soil before the Nisqually earthquake. The tripartite array, SDN, SDS, and SDW, is located in an industrial district where there are few acceptable choices given the required 200 meter spacing. SDN and SDS are located at opposite ends of a long and narrow 1-story metal warehouse building. The thick concrete pad is about 1 meter above grade. UNR and UNK are on the roof and in the basement (respectively) of a 10-story steel International style office building. Plans may be available for that structure.

Data Availability

Data for all of the earthquakes that our Seattle array has recorded up to October of 2004 may be found at

All of our stations have Kinemetrics K2 2g accelerographs. The data headers indicate whether the sensor is the older FBA-24 or the Episensor. Some stations also have Mark L-22 velocity sensors on channels 4, 5, and 6 (as indicated in the event headers.)

The event data is available in ASCII, the original evt, and SAC formats. The file, k2autoed.cfg, contains station configuration data that has been used to correct event data headers (such as orientation) for each of the formats. You will also find files that give ANSS Catalog Locations for all of the events for which we have data. The AssociatedEvent list can be used to easily determine the accuracy of timing. The column after ".evt" gives the number of hours since the last GPS clock correction. Generally speaking, if the GPS at a station has not locked within 12 hours we do not trust the accuracy of the timing for the data. For the Nisqually event, station MAR was not receiving GPS timing signals and so the clock correction is unknown. All of the other station's clocks were accurate to within .002 seconds for Nisqually.

Any use of trade names is for descriptive purposes only and does not imply endorsement by the USGS. logo  Take Pride in America button