Content uploaded by Apostolos Papageorgiou
All content in this area was uploaded by Apostolos Papageorgiou on Oct 12, 2014
Content may be subject to copyright.
Determination of Site Response in Anchorage,
Alaska, on the Basis of Spectral Ratio Methods
U. Dutta, N. Biswas,
and R. Combellick
This paper deals with the site response (SR) in the Anchorage basin in south-
central Alaska. The investigation is based on the analysis of seismograms of 114
earthquakes recorded by 22 weak-motion stations and 46 earthquakes recorded by
19 strong-motion stations in the study area. We have computed SR for 41 sites,
using standard spectral ratio and horizontal to vertical spectral ratio methods in
the frequency range from 0.5 to 11 Hz. Based on these results, we have calculated
band-average site response values in two frequency ranges: low frequency (from
0.5 to 2.5 Hz) and high frequency (from 3 to 7 Hz). There is a good correlation
between SR values and surficial geology of the Anchorage area in the low
frequency range. SR values increase by a factor of three from the foothills of
Chugach Mts. in the east to the west towards the deeper part of the basin. The
highest site response values (SR > 2.5) in the same frequency range are observed
in the westcentral part of the city, which is underlain by cohesive facies of the
Bootlegger Cove formation. The SR has a good correlation with the uppermost 30
m time-average shear wave velocity with a correlation coefficient of 0.82.
Moreover, the low frequency SR values are close to the NEHRP site coefficients
for 1 sec. However, high frequency SR values lack correlation with 30 m average
shear wave velocity and short-period NEHRP site coefficients.
Anchorage lies within the highly active seismogenic zone of south-central Alaska. The
first-order feature of the seismotectonic setting of this section of Alaska consists of the
underthrusting of the Pacific plate beneath the North American plate. Geologically, most
parts of the study area lie in a sedimentary basin, with the sedimentary formations abutting
metamorphic bedrocks exposed in the Chugach Mts. to the east (Figure 1a). The formations
of the basin dip towards the west as shown by an east-west cross-section (Figure 1b) along
AA′ in Figure 1a. An important Quaternary unit of the Anchorage basin is called the
Bootlegger Cove formation (BCF); it consists of glacial deposits of interbedded clay, silt and
sand. This formation underlies lowland areas in westcentral Anchorage, including the
downtown area, in a north-south oriented zone several kilometers wide (Updike and Ulery
1986, Combellick 1999). The eastern extent of about 10 m thick BCF deposits is shown in
Figure 1a as a broken line (Combellick 1999). Updike and Ulery subdivided BCF into two
facies - cohesive and noncohesive. The cohesive facies, consisting of silty clay and/or clayey
silt, are susceptible to failure. They suffered extensive ground failure during the 1964 Prince
William Sound earthquake (M
Following the catastrophic results in and around Anchorage from the 1964 earthquake
(AM, UD, NB) Geophysical Institute, University of Alaska Fairbanks, AK 99775-7320
(AP) State University of New York at Buffalo, NY 14260-4300
(RC) Alaska Division of Geological and Geophysical Surveys, 794 University av., Fairbanks, AK 99709
-150.05 -149.95 -149.85 -149.75 -149.65
1 2 3 km
Figure 1. (a) Surficial geology of Anchorage. Thick broken line is approximate 10 m isopach of the
Bootlegger Cove formation, which dips from east to west; (b) Geological cross-section along AA',
shown in Figure 1a (reproduced from Selkregg et al. 1972).
and the vigorous socio-economic growth of the area during the past two decades, seismic
hazard has become a major concern to the inhabitants of the area. At the same time, there is
a severe lack of information on the characteristics of strong ground motion. In order to
address this problem, we implemented a seismic microzonation project for the metropolitan
area of Anchorage. The determination of S-wave site response (SR) constitutes an important
part of this project. Herein we report the SR results obtained for this area.
NETWORK AND DATABASE
Two seismological databases, namely, weak-motion (WM) and strong-motion (SM) have
been used to investigate the site response in the study area. Figure 2 shows the locations of
WM and SM stations, and their coordinates are given in Table A1 of Appendix A. A
summary of the data used in this study is given in Table 1.
-150.0 -149.9 -149.8 -149.7 -149.6
1 2 3km
Gulf of Alaska
Figure 2. The inset shows the location of the study area with respect to the rest of the Alaska-
Aleutian region. The locations of Anchorage weak-motion (circles) and strong-motion (triangles)
stations are shown on the background of main roads (solid lines).
The WM network was operated temporarily for about six months with 22 IRIS-
PASSCAL instruments; they were all installed in the free-field. Each station consisted of a 3-
component seismometer (Mark product L-28) connected to a Reftek recorder with a GPS
receiver; data were recorded at a 50 Hz sampling rate. The stations were located in areas of
various geological units. During six months of operation of the network, 114 earthquakes in
the local magnitude (ML) range from 1.5 to 5.5 were recorded with good signal-to-noise
ratio. The average number of stations that recorded an event simultaneously was about 19.
The epicentral locations of the 114 events selected for the study are shown in Figure 3 as
circles, which were noted from the catalog of the Alaska Earthquake Information Center.
Table 1. Summary of the Anchorage weak-motion and strong-motion networks and databases
used in this study.
Weak-motion network Strong-motion network
Operational period July – December 1996 Since November 1995
IRIS-PASSCAL short period
sensors and Reftek recorders
Kinemetrics Altus K2
Sampling rate (Hz) 50 (continuous mode) 200 (trigger mode)
Number of stations 22 19
Number of events 114 46
Number of records 2209 512
Magnitude range (ML) 1.5 – 5.5 3.1 – 6.3
Depth range (km) 1 – 160 10 – 140
-154 -150 -146 -142
G u l f o f A l a s k a
50 100 150km
Figure 3. Epicenter map of earthquakes used in this study. Weak-motion events are shown as circles
and strong-motion events as triangles. Symbol size is proportional to the event magnitude. Large
black circle indicates the network location.
The SM network (22 stations) has been operational since 1995 in the metropolitan area of
Anchorage. The stations were installed on the concrete floors of small buildings, such as the
city fire station or similar structure, which are not expected to significantly influence our
results below 10 Hz. Each station consisted of a Kinemetrics Altus K2 accelerograph. Of
these, two stations (K2-17 and K2-18) are located outside Anchorage proper to obtain a
distance coverage of about 60 km in the NE direction. Moreover, the station K2-19 has
recorded very few events, and data from this station were not considered in this study. We
selected 46 earthquakes recorded by the SM network for this study; their epicentral locations
are shown in Figure 3 as triangles. The largest event in this group is ML 6.3. However, the
maximum peak ground acceleration (75 gal) has been recorded for an event of ML 5.1,
having a focal depth of 36 km and located 50 km from the center of the network.
The spatial coverage (about 16 x 16 km) obtained by the WM and SM networks is
approximately the same with average station separation of about 2.5 km. Two stations - An01
(WM) and K2-16 (SM) - have been installed along the Chugach Mts., where metamorphic
rocks are exposed, and they have been considered as reference sites for this study. There
were four sites with collocated WM and SM instruments, with station-to-station distances
less than 150m; the others could not be collocated due to logistic difficulties. Despite the fact
that the operational periods of the two networks entirely overlapped, there were only 5 WM
events, which triggered some of the SM stations. The azimuthal coverage and focal depth
distribution in the WM and SM datasets are comparable, while the average epicentral
distance of the WM events is relatively greater, that is, 115 km compared to 80 km for the
The ground motion at a given site is a convolution of earthquake source, path and site
factors. In this study, we are concerned with the site response, and for its determination we
used two spectral ratio methods, namely, standard spectral ratio (SSR) and horizontal to
vertical spectral ratio (HVR). The SSR method was introduced by Borcherdt (1970) and
productively used later in site response studies. It requires simultaneous recordings of
earthquakes at rock and soil sites. Assuming that the rock site has negligible site response,
the amplitude spectral ratio of the record at the soil site to that at the rock site provides an
estimate of site response at the soil site under the assumption that the source and path effects
are common to the rock and soil sites. However, this assumption may not be true over the
entire frequency range of interest. The other technique is the HVR method, first used by
Nakamura (1989) for the interpretation of data of microtremor measurements. Lermo and
Chavez-Garcia (1993) showed correlation of HVR results with those obtained by the SSR
method. However, Lachet et al. (1996) and Bonilla et al. (1997) among others showed that
the HVR method provides an estimate of the dominant frequency, but not the actual level of
the site response for a given site.
In processing the data of the present study, we first compared in the time domain the
amplitudes between the S-wave envelope (S) and pre-event trace (B). If the condition S ≥ 3B
was satisfied, the seismogram was selected, otherwise it was rejected for further analysis. The
selected seismograms were corrected for instrument response and bandpass-filtered with
bandwidth from 0.1 to 16 Hz. The record lengths considered in the WM and SM
seismograms were 20 and 7 sec, respectively, with the Hanning taper. The pre-event data
were processed in the same manner and used to correct the signal spectrum for each record.
The corrected amplitude spectra were smoothed using a triangular window of 0.5 Hz half-
width. The smoothed spectral values were interpolated using a spline interpolator at 0.05 Hz
interval from 0.5 to 15 Hz frequency range. Next, the root mean square of the amplitude
spectra were computed from:
where f is the frequency, H
(f) and H
(f) represent smoothed amplitude spectra of the
windowed and filtered data of the N-S and E-W components. The H(f) values were used to
obtain spectral ratios at different sites.
RESULTS AND DISCUSSION
Since the SR values can be approximated by a lognormal distribution (Field and Jacob
1995), we computed the logarithmic event-average of individual spectral ratios obtained from
the WM and SM network data; the results are shown in Figures 4 and 5. The results show
considerable variations of SR in the basin. At some sites (An02, An15, K2-13, K2-21, K2-
08), SR tends to approach a value of 6, while others (An06, An17, An18, K2-09, K2-22) are
associated with relatively low values (1-3) of SR. Moreover, some sites (An02, K2-01) show
sharp peaks of SR while others (An07, An12, K2-06) show broadly distributed highs as a
function of frequency. Spatially, SR values in general increase from the foothills of Chugach
Mts. in the east towards the west in the deeper part of the basin.
Figure 4. Site response as a function of frequency (from 0.5 to 11 Hz) for the weak-motion sites. The
solid and dashed lines represent the results based on the SSR and HVR methods, respectively.
The mean of the logarithmic standard deviations of the SR values obtained by the SSR
and HVR methods are 0.32 and 0.23, respectively. In Mexico, Lermo and Chavez-Garcia
(1993) also observed less scatter in HVR results compared to those for SSR. Moreover, we
obtained a decrease in standard deviation with the increase in focal depths of earthquakes.
This relationship is shown in Figure 6 for the SSR case. A similar result was obtained for the
HVR case. The decrease is by a factor of about two for earthquakes with hypocenter depths
greater than 50 km compared to those shallower than 50 km.
Figure 5. Site response as a function of frequency (from 0.5 to 11 Hz) for the strong-motion sites.
The solid and dashed lines represent the results based on the SSR and HVR methods, respectively.
The comparison of site responses yielded from the WM and SM datasets is of particular
interest. Earlier, we mentioned that besides the reference site, there were three more sites
with collocated WM and SM stations: K2-01 and An15 (distance – 100 m), K2-12 and An05
(distance – 130 m), and K2-06 and An12 (distance – 145 m). With the exception of one case
(K2-06 and An12), the SR variations in Figure 4 and 5 show that site responses of collocated
stations are more or less consistent, especially between HVR results. An12 shows
considerably higher amplifications above 6 Hz compare to the collocated site K2-06 in both
SSR and HVR results. This result is more than likely caused by variations of local geology at
short distances. Similarly, the SSR results for An15 in northern Anchorage exceeds one for
the collocated SM site K2-01 by a factor of 1.3-1.4 in a wide frequency range from 0.5 to 9
Hz. At the reference site, the two HVR-based estimations are in accord, gradually increasing
with the frequency up to a factor of 3 at 10 Hz. This trend indicates that the reference site has
site responses at higher frequencies. We particularly analyzed five events, simultaneously
recorded by both WM and SM stations at collocated sites, in order to compare the spectral
ratios yielded from the velocity and acceleration records of the same event. However, the
results show large scatter, preventing any specific conclusion.
0 20 40 60 80 100 120 140 160
Logarithmic standard deviation of SR
Figure 6. Relationship between logarithmic standard deviation of SR values and focal depths of
Furthermore, we computed network-average site responses from the WM and SM
datasets, assigning equal weight to each individual spectral ratio from both datasets. The
results are shown in Figures 7a and 7b. It may be noted that the differences lie within the ±σ
range, where σ refers to the standard deviation of SR values. Similar comparison of SR
values obtained by the SSR and HVR methods from the data of both networks is displayed in
Figures 8a and 8b. The results show that below 6 to 7 Hz, HVR values are higher by factors
of about 1.3 and 1.7 than SSR values for WM and SM cases, respectively. For higher
frequencies, these differences increase slightly. The source of the difference between the WM
and SM cases is not clear at this stage.
Since the HVR method provides a good estimation of predominant frequency (f
) of the
sedimentary column at the site of measurement (Theodulidis and Bard 1995, Field and Jacob
1995, Riepl et al. 1998), we determined f
from the HVR results for most of the WM and SM
sites, where the peaks with maximum amplitudes are well defined (dashed curves in Figures
4 and 5). These values are shown in Figure 9, where in most cases f
decreases from about 3
to 4 Hz along the foothills of Chugach Mts. to a minimum of 0.7 Hz on the western side of
the basin. This result is in accord with the variation of the thickness of the Quaternary
sediments (Figure 1b), which increases from the east to the western side of the basin where
the Quaternary sediments are several hundreds meters in thickness. However, predominant
frequencies shown in Figure 9 correspond to the ground motion level less than 0.1 g. At
higher levels of ground motions these values may change considerably due to nonlinear
The network average SR values obtained from both WM and SM data by the SSR method
show deamplification above 7 to 8 Hz (Figure 7a). A computation of the normalized response
spectra with the SM data showed a cross-over of the rock site spectrum by those of the soil
sites around 7 Hz. Aki (1988) has pointed out that SR of soil sites is about 2 to 3 times higher
than that of rock sites below 5 Hz. This relationship reverses at higher frequencies (>5 Hz).
This may be due to higher attenuation at the soil sites with respect to the rock sites. The
overall effect of this reversal will be exhibited by the deamplification of SR values yielded by
the SSR method at high frequencies.
Figure 7. Network-average site responses; (a) - by SSR method and (b) - by HVR method. Thick gray
lines correspond to WM results, thin black lines to SM results. Shaded areas represent one standard
deviation limits of SR values.
Figure 8. Network-average site responses; (a) - from WM records and (b) - from SM records. Thick
gray lines correspond to HVR method, thin black lines to SSR method. Shaded areas represent one
standard deviation limits of SR values.
Recently, the drilling of a special purpose borehole of 9 m depth located close to the
reference sites (An01 and K2-16; Figures 4 and 5) showed the presence of about 4 m thick
weathered and fractured metamorphic rock. We anticipate that the weathered layer is the
result of repeated freezing and thawing of the insitu formations. It appears that this weathered
layer, whose spatial extent is not known, may play a part in the amplification of the ground
motion at reference sites at higher frequencies (Steidl et al. 1996).
In addition to the factors referred so far, source directivity and topographic effects may be
thought of as potential sources of the above observed phenomenon (Kawase and Aki 1990;
Pedersen et al. 1994; Zhang et al. 1998). However, as shown in Figure 3, the earthquakes
considered for this study are distributed in different azimuths and epicentral distances with
respect to the networks. Therefore, the effects of these two factors on the event-average site
response are assumed to be minimal. Nevertheless, for further discussion of SR values in the
Anchorage area, we restricted them to 7 Hz for the reasons mentioned above.
-150.1 -150.0 -149.9 -149.8 -149.7 -149.6
Figure 9. Predominant frequencies (in Hz) for some of the WM and SM sites, estimated from HVR
SPATIAL DISTRIBUTION OF BAND-AVERAGE SR VALUES
In order to facilitate usage of the SR by practicing engineers, we computed frequency
band-averages of the SR values. For this purpose, a low frequency (LF) range from 0.5 to 2.5
Hz and a high frequency (HF) range from 3.0 to 7.0 Hz with approximate geometric center
frequencies of about 1 and 5 Hz, respectively, were selected. Moreover, Dutta et al. (2000b)
obtained SR values by the generalized inversion method (GI) for the Anchorage basin, and
these values were found similar to those obtained by the SSR method of the present study.
We, thus, considered the three sets (SSR, HVR, and GI) of SR values and assigned weights
of 0.45, 0.10 and 0.45, respectively, to obtain weighted average values. The relatively low
weight given to the HVR results is due to their likely overestimation of SR values over the
frequency range considered, as mentioned earlier.
Figures 10a and 10b show the spatial distribution of SR values at 0.5-unit interval
contours over a geological background (Combellick 1999) of the study area. The contours are
based on grid values of dimensions 0.6 x 0.6 km, which were obtained by the Kriging method
(Stein 1999). At 1 Hz, the SR values increase from 1.5 from the eastern side of Anchorage to
3 and slightly above in the westcentral and northwestern parts of the city. At 5 Hz, the trend
of the spatial distribution of SR changes, where the areas of SR ≈ 3 is concentrated in the
south and southeastern parts of the city adjoining the Turnagain Arm. Moreover, an area with
SR values greater than 2 is located near the center of the city.
In the LF range, the areas with SR ≤ 1.5 are characterized by deposits grouped as older
glacial drift (GD), consisting of heterogeneous undifferentiated till, stony glaciomarine,
glaciofluvial deposits, etc. GD is the dominant geological unit in the eastern side of
Anchorage. The areas with SR > 2.0 are underlain by the Bootlegger Cove formation (BCF),
which predominates in the western side of the city.
-150.05 -149.95 -149.85 -149.75 -149.65
A - Pt. Woronzof
B - Pt. Campbell
C - International Airport
-150.05 -149.95 -149.85 -149.75 -149.65
A - Pt. Woronzof
B - Pt. Campbell
C - International Airport
Figure 10. Contour maps of band-average SR values at 1 Hz (a) and 5 Hz (b) are shown on the
background of Anchorage surficial geology.
The 10 m isopach of BCF (Figure 1a), approximately follows the SR ≈ 2. As mentioned
earlier, Updike and Ulery (1986) have subdivided BCF into cohesive (CBCF) and
noncohesive (NBCF) facies. The western edge of the city (Pt. Woronzof and Pt. Campbell),
including the western part of the International Airport, is underlain by NBCF with SR ≈ 2.5.
Areas in westcentral and northwestern parts of the city are underlain by CBCF with SR
values of 2.5 to 3.5. This correlation of CBCF and relatively high SR values is further
illustrated in Figure 11 along a N-S cross-section (CC′, Figure 10a). Note that values of SR >
2 are associated with the thick deposits of CBCF in the areas at the two ends of the cross-
In the HF range, there is a lack of correlation between the spatial distribution of SR
values and geological trend. According to the quarter wavelength estimation (Kramer 1996),
SR values around 5 Hz are mostly controlled by the physical properties of the shallower 10 to
20 m thick soil column (assuming 200 to 400 m/s shear wave velocity) which are highly
variable across the Anchorage basin. This variability is interpreted as the primary source of
the lack of correlation at 5 Hz.
Figure 11. (a) Site response along CC′ as shown in Figure 10a; (b) Geological cross-section along
RELATIONSHIP BETWEEN SR AND UPPERMOST 30 M AVERAGE SHEAR WAVE
In the Anchorage area, the distribution of the uppermost 30 m time-average shear wave
) was obtained at 36 sites (including 15 SM sites) by surface measurements in
cooperation with the Vibration Instrument Company of Tokyo, Japan and Ensole Corporation
of Raleigh, North Carolina. The details of the method have been given by Rodriguez-
Ordonez (1994), Nath et al. (1997), and Dutta et al. (2000a) and are not repeated here.
The results of these measurements are shown in Figure 12, where the β
from 220 to 600 m/s. On the basis of the NEHRP site classification (Building Seismic Safety
Council 1997), it appears that the Anchorage basin consists of site class C (360 < β
m/s) and D (180 < β
≤ 360 m/s). Moreover, though there is no velocity data coverage along
the Chugach Mts., we anticipate site class B (760 < β
≤ 1500 m/s) in this area from rock
exposures and samples obtained from the special purpose 9 m borehole mentioned earlier.
Following the known subsurface lithologic variation from the foothills of the Chugach
Mts. toward the west side of Anchorage, it is necessary to introduce a transition zone (C/D)
between the areas of site classes C and D. The β
in C/D ranges from 320 to 410 m/s and this
zone extends in the N-S direction with an eastward bent around downtown Anchorage
(Figure 12). SR values in this zone approximately lie between 1.5 and 2.0 in the LF range
(Figure 10a) and this zone includes the 10 m isopach of BCF. The band-average SR values,
and site classes obtained for the WM and SM station sites are listed in Table A2 of
The relationships between band-average SR values and β
at 1 and 5 Hz, respectively,
are shown in Figures 13a and 13b. At 1 Hz, the two parameters correlate with correlation
coefficient of 0.82. A power law fit to the data (Figure 13a) yielded the relation:
Dutta et al. (2000b) also obtained a similar relation as in (2) for the Anchorage basin,
using weak-motion data only. However, at 5 Hz there is a lack of correlation between band-
average SR values and β
. The extensive lateral variation of the geologic formations in the
uppermost 10 to 20 m, as mentioned earlier, may be the source of this lack of correlation at
-150.10 -150.00 -149.90 -149.80 -149.70
Figure 12. The shear wave velocity measurement sites and
values in m/s. Two contour lines,
which represent velocities 320 and 410 m/s, separate site classes C, C/D and D. Broken line is
approximate 10 m isopach of deposits of BCF (Combellick 1999).
200 250 300 350 400 450 500 550 600
30m average shear wave velocity (m/s)
Site Class C/D
Site Class D
Site Class C
200 250 300 350 400 450 500 550 600
30m average shear wave velocity (m/s)
Site Class C/D
Site Class D
Site Class C
Figure 13. Relationship between band-average site response values and uppermost 30 m average
shear wave velocities for the LF (a) and HF (b) ranges. The velocity values for sites with no direct
velocity measurements were interpolated.
COMPARISON WITH THE NEHRP SITE COEFFICIENTS
The two frequency intervals for the band-average SR values were selected such that the
average values can be directly related to the NEHRP site coefficients (Building Seismic
Safety Council 1997). It may be noted that the NEHRP site coefficients were derived from
both Fourier and response spectra (Borcherdt 1994, Dobry et al. 2000). Although our band-
average SR values were obtained from Fourier spectra, we have also computed SR from the
response spectra of SM records for 5% damping, using the SSR method. The results closely
agree with those obtained from Fourier spectra, as shown in Figure 14 for two SM sites,
although the response spectra-based estimations in general are smoother, especially at higher
frequencies. Furthermore, a good correspondence between band-average SR values is
illustrated in Figure 15 for both LF and HF intervals.
In Table 2, the characteristics of site classes obtained from this study are compared with
those given by the NEHRP provisions of 1997 for a ground motion level less than 0.1 g. In
the LF range, we have obtained class-average SR values of 1.31, 1.84 and 2.67 for site
classes C, C/D, and D, respectively. With respect to these values, the NEHRP recommended
site coefficients at 1 sec are higher for site class C (F
=1.7) and lower for site class D
=2.4). These differences are attributed to the additional subdivision between soil classes C
and D used here; the site class C/D occupies the lower SR values, compared to the NEHRP-
specified site class D (i.e. sites with β
between 320 and 360 m/s) and the higher SR values
from site class C (360 ≤
< 410 m/s, see Figure 13a). At 5 Hz, the lack of spatial
correlation between β
and SR seems to yield values of SR close to each other for the three
site classes -1.68, 1.66, and 1.70 for C, C/D, and D, respectively.
Figure 14. Site responses for two SM sites: K2-01 (a) and K2-02 (b). Black lines are based on
Fourier spectra, gray lines are based on response spectra.
Band-average site amplification
(from response spectra)
Band-average site amplification
(from Fourier spectra)
Figure 15. Comparison of band-average SR values, obtained from Fourier and response spectra.
Filled and open triangles correspond to LF and HF ranges, respectively.
Table 2. Site class characteristics; N is the number of sites in a given site class used in this study.
Site class C (N = 18) C/D (N = 6) D (N = 19)
interval (m/s) (present study)
410 - 760 320 - 410 180 - 320
360 - 760 - 180 - 360
Band-average SR value at 1 Hz (present study)
1.31 ± 0.30 1.84 ± 0.33 2.67 ± 0.53
NEHRP site coefficient at 1 sec (F
) 1.7 - 2.4
Band-average SR value at 5 Hz (present study)
1.68 ± 0.50 1.66 ± 0.53 1.70 ± 0.55
NEHRP short-period site coefficient (F
) 1.2 - 1.6
The sedimentary section of the Anchorage metropolitan area exhibits considerable spatial
variations of seismic site response (SR). Band-average SR values in the low frequency (LF)
range from 0.5 to 2.5 Hz increase by a factor of three from the foothills of Chugach Mts. in
the east to the west towards the deeper part of the basin. The areas of relatively higher site
response values (SR > 2.5) in the LF range are observed in westcentral and northwestern
parts of the city underlain by the cohesive facies of the Bootlegger Cove formation.
The SR values yielded by the weak-motion and strong-motion datasets are comparable
and lie within the one standard deviation limits. The results obtained by the standard spectral
ratio (SSR) method are more consistent than those yielded by the horizontal to vertical
spectral ratio (HVR) method, which tends to overestimates SR values. However, there is less
scatter in HVR results with respect to that of SSR. Also, the logarithmic standard deviation of
the SR values decrease by a factor of about two for earthquakes with focal depths greater
than 50 km compared to those shallower than 50 km. The HVR method provides a good
estimation of predominant frequency of the sedimentary column, which in general decreases
from about 3 Hz to 4 Hz along the foothills of Chugach Mts. to a minimum of 0.7 Hz on the
western side of the basin. This is in accord with the variation of the thickness of Quaternary
The site responses computed by the SSR method from the Fourier spectra closely agree
with those obtained from the response spectra for 5 percent damping.
Following the known subsurface lithologic variation and the uppermost 30 m time-
average shear wave velocity (β
) in the area, it is necessary to introduce a transition zone of
C/D between areas of site classes C and D in the Anchorage basin. The width of the zone
C/D is about 2-3 km and β
in this zone varies from 320 to 410 m/s.
The frequency band-average SR values at 1 Hz have good correlation with β
correlation coefficient of 0.82. SR values at this frequency (1 Hz) are consistent with the
NEHRP specified site coefficients. However, frequency band-average SR values at 5 Hz lack
correlation with the trend of β
The authors are grateful to Toshifumi Kono for operation and initial data reduction of the
weak-motion data and maintenance of the Anchorage strong-motion network, and David
Cole of Dowl Engineers of Anchorage for useful discussion. We also thank Mitch Robinson
and Kent Lindquist of the Geophysical Institute for assisting in computer related problems
and locations of earthquakes used in this study. We are grateful to three anonymous
reviewers and the editor for their comments and suggestions. This work was supported in part
by the Alaska Science and Technology Foundation Project #91-2-125 and #97-3-131 and in
part by the Geophysical Institute, University of Alaska Fairbanks.
Table A1. Anchorage strong-motion and weak-motion station coordinates. Three SM stations, K2-04,
K2-11, and K2-14 have been relocated to nearby locations after a few years of operation. The old
locations are listed at the end of the table, marked with star.
K2-01 61.235 149.869 An01 61.098 149.687
K2-02 61.224 149.822 An02 61.075 149.807
K2-03 61.219 149.718 An03 61.114 149.820
K2-04 61.177 150.010 An04 61.101 149.863
K2-05 61.200 149.911 An05 61.156 149.794
K2-06 61.191 149.822 An06 61.155 150.052
K2-07 61.160 150.001 An07 61.127 149.933
K2-08 61.177 149.919 An08 61.157 149.987
K2-09 61.185 149.744 An09 61.153 149.929
K2-10 61.130 149.928 An10 61.189 150.015
K2-11 61.149 149.855 An11 61.186 149.875
K2-12 61.156 149.792 An12 61.191 149.824
K2-13 61.113 149.856 An13 61.181 149.720
K2-14 61.124 149.766 An14 61.210 149.909
K2-15 61.087 149.750 An15 61.235 149.870
K2-16 61.099 149.685 An16 61.249 149.818
K2-20 61.155 150.053 An17 61.223 149.726
K2-21 61.153 149.949 An18 61.206 149.787
K2-22 61.088 149.834 An19 61.174 149.846
K2-04* 61.178 150.015 An20 61.217 149.849
K2-11* 61.157 149.869 An21 61.187 149.938
K2-14* 61.140 149.781 An22 61.137 149.889
Table A2. Characteristics of the SM and WM sites. Site classes for the sites with no direct velocity
measurements are defined according to their positions relative to the velocity contour lines 320 and
K2-01 238 D 2.58 1.28 An01 B Ref. Ref.
K2-02 366 C/D 1.94 0.86 An02 C 0.80 3.56
K2-03 474 C 1.34 1.46 An03 C 1.95 2.93
K2-04 - D 1.96 1.38 An04 C/D 2.51 1.84
K2-05 284 D 2.84 1.75 An05 C 1.14 1.21
K2-06 491 C 1.26 2.53 An06 D 2.31 1.43
K2-07 270 D 2.07 1.41 An07 D 3.58 3.57
K2-08 274 D 2.58 1.45 An08 D 3.19 2.90
K2-09 582 C 1.19 1.01 An09 D 3.92 2.21
K2-10 269 D 2.24 1.61 An10 D 2.86 1.80
K2-11 - C 1.50 1.82 An11 C 1.48 1.29
K2-12 514 C 0.90 1.33 An12 C 1.44 3.15
K2-13 354 C/D 1.65 1.85 An13 C 1.02 2.22
K2-14 - C 1.38 1.34 An14 D 3.01 1.50
K2-15 412 C 1.28 2.16 An15 D 3.49 1.85
K2-16 - B Ref Ref An16 D 2.67 1.04
K2-20 - D 2.06 1.87 An17 C 1.55 1.02
K2-21 - D 2.71 2.20 An18 C 1.68 0.90
K2-22 - C 1.62 1.64 An19 C 1.60 1.56
K2-04* 309 D 2.17 1.39 An20 D 2.36 1.01
K2-11* 394 C/D 1.75 2.07 An21 D 3.20 2.45
K2-14* 524 C 1.14 1.75 An22 C/D 1.87 1.89
Aki, K., 1988, Local site effects on strong ground motion, Proceedings of Earthquake Engineering &
Soil Dynamics II GT Div/ASCE, Park City, Utah, June 27-30, 1988.
Bonilla, F. L., Steidl, J. H., Lindley, G. T., Tumarkin, A. G., and Archuleta, R. J., 1997, Site
amplification in the San Fernando Valley, CA: variability of site effect estimation using the S-
wave, coda and H/V methods, Bull. Seism. Soc. Am., 87, 710-730.
Borcherdt, R. D., 1970, Effects of local geology on ground motion near San Francisco Bay, Bull.
Seism. Soc. Am., 60, 29-61.
Borcherdt, R. D., 1994, Estimates of site-dependent response spectra for design (methodology and
justification), Earthquake Spectra, 10, 617-653.
Building Seismic Safety Council (BSSC), 1998, 1997 Edition NEHRP Recommended Provisions for
Seismic Regulations for New Buildings and Other Structures, FEMA 302/303, Part I (Provisions)
and Part II (Commentary), Washington, DC.
Combellick, R. A., 1999, Simplified geologic map and cross sections of central and east Anchorage,
Alaska: Alaska Div. Geol. Geophys. Surveys Preliminary interpretive report 1999-1, Fairbanks,
Dobry, R., Borcherdt, R. D., Crouse, C. B., Idriss, I. M., Joyner, W. B., Martin, G. R., Power, M. S.,
Rinne, E. E., and Seed, R. B., 2000, New site coefficients and site classification system used in
recent building seismic code provisions, Earthquake Spectra, 16, 41-67.
Dutta, U., Biswas, N. N., Martirosyan, A. H., Nath, S., Dravinski, M., Papageorgiou, A., and
Combellick, R., 2000a, Delineation of spatial variation of shear wave velocity with high-
frequency Rayleigh waves in Anchorage, Alaska, GJI, 143, 1-16.
Dutta, U., Martirosyan, A. H., Biswas, N. N., Papageorgiou, A., Dravinski, M., and Combellick, R.,
2000b, Estimation of S-wave Site Response in Anchorage, Alaska from Weak Motion Data Using
Generalized Inversion Method, Bull. Seism. Soc. Am.,91, 335-346.
Field, E. H. and Jacob, K. H., 1995, A comparison and test of various site-response estimation
techniques, including three that are not reference-site dependent, Bull. Seism. Soc. Am., 85, 1127-
Kawase, H. and Aki, K., 1990, Topography effect at the critical SV-wave incidence: Possible
explanation of damage pattern by the Whittier Narrows, California, earthquake of 1 October
1987, Bull. Seism. Soc. Am., 80, 1-22.
Kramer, S. L., 1996, Geotechnical Earthquake Engineering, Prentice Hall, Upper Saddle River, NJ.
Lachet, C., Hatzfeld, D., Bard, P-Y., Theodulidis, N., Papaioannou, C., and Savvaidis, A., 1996, Site
effect and microzonation in the city of Thessaloniki (Greece). Comparison of different
approaches, Bull. Seism. Soc. Am., 86, 1692-1703.
Lermo, J. and Chavez-Garcia, F. J., 1993, Site effect evaluation using spectral ratios with only one
station, Bull. Seism. Soc. Am., 83, 1574-1594.
Nakamura, Y., 1989, A method of dynamic characteristics estimation of subsurface using
microtremor on the ground surface, QR RTRI, 30, 25-33.
Nath, S., Chatterjee, D., Biswas, N., Dravinski, M., Cole, D., Papageorgiou, A., Rodriguez, J., and
Poran, C., 1997, Correlation study of shear wave velocity in near surface geological formation in
Anchorage, Alaska, Earthquake Spectra, 13, 55-76.
Pedersen, H. A., Sanchez-Sesma, F. J., and Campillo, M., 1994, Three-Dimensional Scattering by
Two-Dimensional Topographies, Bull. Seism. Soc. Am., 84, 1169-1183.
Riepl, J., Bard, P-Y., Hatzfeld, D., Papaioannou, C., and Nechtschein, S., 1998, Detailed Evaluation
of Site Response Estimation Methods across and along the Sedimentary Valley of Volvi (Euro-
Seistest), Bull. Seism. Soc. Am., 88, 488-502.
Rodriguez-Ordonez, J. A., 1994, A new method for interpretation of surface wave measurements in
soils, PhD dissertation, North Carolina State University, Raleigh, NC.
Schmoll, H. R. and Dobrovolny, E., 1972, Generalized geologic map of Anchorage and vicinity,
Alaska: U.S. Geological Survey Miscellaneous Geologic Investigations Map I-787-A, 1 sheet,
Selkregg, L. L., Buck, E. H., Buffler, R. T., Cote, O. E., Evans, C. D., and Fisk, S. G., 1972,
Environmental atlas of the greater Anchorage area borough, Alaska, Arctic Environmental
Information and Data Center, University of Alaska, Anchorage.
Stein, M. L., 1999, Interpolation of Spatial Data: Some Theory for Kriging, Springer Series in
Statistics, Springer Verlag, NY.
Steidl, J. H., Tumarkin, A. G., and Archuleta, R. J., 1996, What is a Reference Site?, Bull. Seism. Soc.
Am., 86, 1733-1748.
Theodulidis, N. and Bard P-Y., 1995, Horizontal to vertical ratio and geological conditions: an
analysis of strong motion data from Greece and Taiwan (SMART-1), Soil Dyn. Earthquake Eng,
Updike, R. G. and Ulery, C. A., 1986, Engineering Geologic Map of Southwest Anchorage, Alaska:
Alaska Div. Geol. Geophys. Surveys Professional Report 89, 1 sheet, scale 1:15,840.
Zhang, B., Papageorgiou, A. S., and Tassoulas, J. L., 1998, A Hybrid Numerical Technique,
Combining the Finite Element and Boundary Element Methods, for Modeling the 3D Response of
2D Scatterers, Bull. Seism. Soc. Am., 88, 1036-1050.