LONG VALLEY OBSERVATORY QUARTERLY REPORT

JULY-SEPTEMBER and OCTOBER-DECEMBER 2002

AND ANNUAL SUMMARY FOR 2002

Long Valley Observatory

U.S. Geological Survey

Volcano Hazards Program, MS 910

345 Middlefield Rd., Menlo Park, CA 94025

http://lvo.wr.usgs.gov

This report is a preliminary description of unrest in Long Valley caldera and Mono-Inyo Craters region of eastern California. Information contained in this report should be regarded as preliminary and is not be cited for publication without approval by the Scientist in Charge of the Long Valley Observatory. The views and conclusions contained in this document do not necessarily represent the official policies, either express or implied, of the U.S. Government.

LONG VALLEY OBSERVATORY QUARTERLY REPORT

July-September and October-December 2002

CONTENTS

EARTHQUAKES

CALDERA ACTIVITY

SIERRA NEVADA ACTIVITY

REGIONAL ACTIVITY

DEFORMATION

SUMMARY OF EDM AND GPS MEASUREMENTS

CONTINUOUS BOREHOLE AND STRAIN MEASUREMENTS

Instrumentation

Highlights

TILT MEASUREMENTS

Instrumentation

Data

MAGNETIC MEASUREMENTS

BACKGROUND

DATA

HIGHLIGHTS

CO₂ STUDIES

HYDROLOGIC MONITORING

GROUND WATER LEVEL MONITORING

SURFACE WATER MONITORING

THERMAL WATER DISCHARGE ESTIMATES

REVIEW OF 2002

CALDERA UNREST

REGIONAL EARTHQUAKE ACTIVITY

SUMMARY FOR JULY-DECEMBER 2002

The relative quiescence in Long Valley caldera that began in the spring of 1998 continued through the last half of 2002. The resurgent dome, which essentially stopped inflating in early 1998 and showed minor subsidence (of about 1 cm) through 2001, began renewed inflation earlier this year at a rate of 1 to 2 cm/year. The center of the resurgent dome still stands roughly 80 cm higher than prior to 1980. Seismic activity within the caldera, which has typically included fewer than five small earthquakes per day since 1999, also showed a slight increase through the last six months of 2002. Diffuse emission of carbon dioxide (CO₂) in the tree-kill areas around the flanks of Mammoth Mountain continue at the relatively high levels that have persisted since 1996. Perhaps the most intriguing activity in the last six months involved a burst of some 80 small earthquakes beneath the south flank of Mammoth Mountain beginning 17 minutes after the magnitude M=7.9 Denali Fault earthquake in Alaska on November 3 and an earthquake swarm in the south moat the next day that included a M=3.0 earthquake. Both of these sequences appear to have been triggered by the energetic seismic waves generated by the Denali Fault earthquake.

Up-to-date plots for most of the data summarized here are available on the Long Valley Observatory web pages (http://lvo.wr.usgs.gov).

EARTHQUAKES (D.P. Hill and A.M. Pitt)

CALDERA ACTIVITY:

Earthquake activity within the caldera increased slightly during the last six months of 2002 with respect to relatively low levels that persisted from 1999 through mid-2002. This increase follows the onset of renewed uplift of the resurgent dome that began in early 2002. Notable activity included a flurry of some 17 small earthquakes (all with magnitudes M < 2.0) centered beneath the south moat just south of the Highway 203-395 junction on July 3-4. We detected a number of small (M<2) earthquakes with shallow focal depths (< 2 km) beneath the southern section of the resurgent dome (in the vicinity of Fumarole Valley) during the last half of August. Many of these events have the character of long-period earthquakes similar to those that occurred in the same area during October 1997. Seismograms from the 2-km-deep borehole seismometer in LVEW that was operating during the fall of 1997 showed that actual number of shallow, resurgent dome events in October 1997 was at least 10 times that recorded on the nearby surface seismic stations.

On November 3 a burst of over 80 small earthquakes (all with magnitudes less than M=1.0) began as the surface waves from the magnitude M=7.9 Denali Fault earthquake in Alaska passed through the area at 3:29 PM (PST). These earthquakes were all located beneath the south flank of Mammoth Mountain, and they coincided with a strain change on the POPA, MCX, and BSP borehole dilatometers (see the section of dilatational strain and tilt under DEFORMATION below). This remotely triggered response of Mammoth Mountain to the Denali Fault earthquake was similar to that for the magnitude 7.2 Hector Mine earthquake of October 16, 1999, which was located in the Mojave Desert 400 km to the south. In the case of the Hector Mine earthquake, however, the burst of small earthquakes was centered beneath the north flank of Mammoth Mountain. On November 4, a day following the Denali Fault earthquake, the most energetic earthquake swarm since 1998 developed in the south moat of the caldera. This swarm included some ten earthquakes with magnitudes greater than M=2.0, the largest of which was a M=3.0 earthquake at 8:08 PM on the 4^th. These earthquakes had unusually shallow focal depths (all less than 4 km), and they were located in the relatively aseismic, central section of the south moat between the east and west lobes where most of the swarm activity has been concentrated since 1980. The earthquake epicenters for this swarm define a northwest striking lineation similar to lineations in the caldera seismicity triggered by the M=7.4 Landers earthquake of June 28, 1992. On November 12, a swarm of small earthquakes began beneath the Shady Rest area just east of Mammoth Lakes that persisted through the 17^th. This swarm included over 40 earthquakes, five of which had magnitudes of M=2.0 or greater. The largest was a M=2.5 earthquake at 5:29 AM on the 15^th.

REGIONAL ACTIVITY

The level of earthquake activity in the region surrounding Long Valley Caldera for the last six months of 2002 showed no significant change from the first half of the year or from the activity levels in 2000 and 2001. Most of the activity continues to be concentrated in the aftershock zone of the three M5 earthquakes of June and July of 1998 and May 1999, which extends from the southeastern margin of the caldera for some 20 km to the south-southwest into the Sierra Nevada (Figure S-1 through S-8). The most active clusters in this zone during the last half of 2002 were midway along this lineation (2 miles south-south east of McGee Mountain) and near the southern end of the lineation in the vicinity of Grinnell Lake. The McGee Mountain cluster produced largest of these earthquakes with a M=3.8 event at 12:47 AM (PDT) on October 6, which followed a M=2.9 foreshock by one minute. A M=3.0 earthquake in the same cluster occurred at 4:56 PM on October 22. The Grinnell Lake cluster produced a M=2.9 earthquake at 5:08 PM on October 27 and M=3.0 and 3.3 earthquakes on December 7 and 23, respectively. Elsewhere, a M=3.7 earthquake at 1:18 PM on July 15 was located just 2 miles north-northwest of Bishop. M=2.9 and M=3.5 earthquakes at 9:26 PM and 10:58 PM on December 12, respectively, were located beneath the Volcanic Tableland 12 miles north-northwest of Bishop.

DEFORMATION

SUMMARY OF EDM AND GPS MEASUREMENTS

John Langbein, Stuart Wilkinson, Elliot Endo, Eugene Iwatsubo, and Jerry

Svarc

Over the past 6 years, 18 GPS (Global Position System) receivers have been installed within and near the Long Valley Caldera. Of these, 14 were installed by Elliot Endo of the Cascades Volcano Observatory. The locations of the 12 receivers within the caldera are shown in Figure G-1. It is intended that data from these receivers and a few more additional installations will take over the long-term monitoring supplied by the two-color EDM (Figure G-2). The site at CASA now has two receivers; one operating since 1994 and the second one, CA99, installed this past summer.

The travel-time measurements from each receiver is processed daily to produce a position in a reference frame with North America fixed. Additional processing involves removing a temporal, common-mode signal from each time-series of displacements as well as the gross outliers. To re-adjust the data to a more local reference frame, a rate is removed from each time series. This rate is the average displacement rate from 1996 to the present of the 2 Sierra Nevada stations, CMBB and MUSB. In the plots, to show any deviation from a constant rate, the local rate is also removed and that rate is posted next to the trace of the residual displacements. These preliminary GPS data indicate inflation of the resurgent dome by just over 1 cm since the beginning of the year.

The most significant change for year 2002 is a reversal of the steady contraction across the resurgent dome that had started in mid-1998 (Figure G-3). Since the Spring of 2002, we now are detecting extension across the resurgent dome; the rate for the Casa-Krak baseline is 3 cm/yr. Modeling of a combination of continuous GPS and two-color EDM data suggests that the extension is due to inflation beneath the resurgent dome. A Mogi point-source model is consistent with these data; the depth is 9KM, and the center of inflation is at 37.688^oN and -118.899^oW. This location is deeper by 2 to 3 KM and translated about 2 KM east relative to the sources inferred from the geodetic data from 1989/92 and 1997/98 episodes of inflation beneath the resurgent dome. The volume is 0.02 km^3/yr over the 8 month interval since May 2002. A few other inflation models have been tested against the data and although these models, the prolate and ellipsoidal point sources, have slight better fits to the data, there is not enough signal in the data to justify these more complex models.

The vectors representing the displacements from continuous GPS since May, 2002 are shown in Figure G-1. This is a summary of the GPS data used in the results presented above. The signal from inflation stands out clearly with the vectors directed radially outward from the resurgent dome.

Figure G-1. Horizontal displacement vectors in mm/year for continuous GPS sites in and around Long Valley Caldera from May 2002 through January 2003.

There are now 5 baselines from the frequently measured, two-color EDM network that are also measured by continuous GPS. The location of the EDM and GPS networks is shown in Figures G-1 and G-2. A comparison of the length-changes derived from GPS and those measured by the EDM are shown in Figure G-3 for the last three years. Although for the Casa-Krak and Casa-Knolls baseline the comparison could be extend back in time, this comparison includes only the GPS data from the sites installed by USGS. There are now two GPS stations at CASA (Casa and Ca99), and two stations at KRAK (Krak and Krac). Since the USGS installations use more modern receivers, they have better day-to-day repeatability than the JPL operated receivers. Finally, it should be noted that the EDM measurements on the Casa-Hot baseline have more scatter than desirable; this is because the optics have deteriorated on the Hot

reflector.

More plots of both the GPS and EDM data can be found at:

http://lvo.wr.usgs.gov/monitoring/index.html#deformation

Figure G-2 Map showing 2-color EDM baselines

The measurements of length changes shown in Figure G-3 for the frequently measured EDM baselines (together with the associated GPS data) show that the gradual contraction that began in early 1999 appears to have stopped in mid-2000. These two-color data indicate that the baselines spanning the resurgent dome began another episode of extension in early 2002. Based on the relation between leveling and 2-color data, the center of the resurgent dome remains about 80 cm higher than in the late 1970’s prior to the onset of caldera unrest.

Figure G-3. Line-length changes for the EDM baselines (crosses) measured from CASA for the period January 1, 1999 through January 4, 2003 compared with continuous GPS data for the same lines (circles).

CONTINUOUS BOREHOLE STRAIN MEASUREMENTS (Malcolm Johnston, Doug Myren, Bob Mueller and Stan Silverman)

Instrumentation

Dilational strain measurements are being recorded continuously at the Devil's Postpile (POP), Motorcross (MX) near the western moat boundary in the south moat, Big Springs (BS) just outside the norhtern caldera boundary, and at Phillips (PLV1), just to the north of the town of Mammoth Lakes. The site locations are shown in Figure D1. The instruments are Sacks-Evertson

dilational strain meters and consist of stainless steel cylinders filled with silicon oil that are cemented in the ground at a depth of about 200m. Changes in volumetric strain in the ground are translated into displacement and voltage by an expansion bellows attached to a linear voltage displacement transducer. This instrument is described in detail by Sacks et al.(Papers

Meteol. Geophys.,22,195,1971).

.

Figure D-1. Locations of dilatometers and tiltmeters.

Data from the strainmeters are transmitted using satellite telemetry every 10 minutes to a host computer in Menlo Park. The data are also transmitted with 24-bit seismic telemetry together with 3-component seismic data to Menlo park.

Highlights

The data this half-year has been relatively quiet at all sites strainmeters. Raw data are shown in Figures D-2.  The only unusual events of interest were strain transients at all sites (with associated seismicity) that were triggered by the M7.9 Denali Fault earthquake on November 3.

Fig D-3a shows the strain and cumulative seismicity that was triggered under the south side of Mammoth Mountain. Fig D-3b shows the triggered strain signal both before and after earth tides are removed.

Figure D-2. Dilatational strain for July-December 2002

In order to learn what aspect of the seismic wave train triggered the strain event and seismicity, we have analyzed the strain seismograms for the earthquake. Fig D-3c shows these seismograms

and Fig D-3d show the strain signals after the seismograms were (mostly) removed with a 20 sec low-pass filter. It appears that the triggering occurred during the passage of the S-wave train but not necessarily when the seismic strain amplitude was greatest. Since strain amplitudes greater than this do occur for smaller local earthquakes that do not trigger seismicity in the region, it is possible that the triggering is frequency dependent. Most of the power in the strain seismogram appears to be at periods between 20-30 seconds (as shown in Fig D-3e) so it is possible that low-frequency seismic waves are more effective in the triggering process. These data, together with the tiltmeter data from Motor Cross and Big Springs were used to model the source of this deformation event as aseismic slip on an east-dipping normal fault beneath the east side of Mammoth Mountain (see Figure D-4).

Figure D-4. A source model for the strain transient triggered by the M=7.9 Delnali Fault earthquake of 3 November 2002 that is consistent with the dilatometer data (Fig. D-3) and the tilt data from BSP and MXP. The model involves a southeast-dipping normal fault centered at a depth of 7 km beneath the east flank of Mammoth Mountain.

TILT MEASUREMENTS (Mal Johnston, Vince Keller, Bob Mueller and Doug Myren)

Instrumentation

Instruments recording crustal tilt in the Long Valley caldera are of two types - 1) a long-base (LB) instrument in which fluid level is measured in fluid reservoirs separated by about 500 m and connected by pipes, which was constructed by Roger Bilham of the University of Colorado, and 2) borehole tiltmeters that measure the position of a bubble trapped under a concave lens.(All Others). For tiltmeter locations, see Figure D-1. Real time plots of the data from these instruments can be viewed at http://quake.wr.usgs.gov/QUAKE/longv.html.

All data are transmitted by satellite to the USGS headquarters in Menlo Park, Ca. Data samples are taken every 10 minutes. Plots of the changes in tilt as recorded on each of these tiltmeters are shown. Removal of re-zeros, offsets, problems with telemetry and identification of instrument failures is difficult, tedious and time-consuming task. In order to have a relatively up-to-date file of data computer algorithms have been written that accomplish most of these tasks most of the time. Detailed discussion or detailed analysis usually requires hand checking of the data.  Flat sections in the data usually denote a failure in the telemetry Gaps denote missing data.

All instruments are scaled using tidally generated scale factors.

Highlights

The data from the long base tiltmeter is shown in Figure T1. Very little of geophysical interest occurred this year and the data are generally uneventful, and show a general tilt down to the south

consistent with increased inflation under the resurgent dome. The North Component was repaired in late July.

Figure T-1. Longbase tiltmeter data for July-December 2002. Increasing values indicate tilt down to the south and east, respectively.

Data for the shallow (1-m deep) borehole tiltmeters Escape, Fossil, Little Antelope, Casa, Sherwin, and Valentine are shown in Figures T-2a,b. The only event of interest occurred on Valatine (VA). While hints of this event may have been detected on the closest tiltmeter Sherwin (SH), single recordings like this indicate something of local origin and are generally discounted.

The deep (100-m deep) borehole tiltmeters, Motor Cross and Big Springs, are shown in Figure T-2b. These instruments were cemented in the fall and the transients shown at this time reflect the thermal curing of the expansive grout. Since being grouted in the data quality has improved and the drift rate has decreased.

Figure T-2a. Borehole tiltmeters CASA, ESCAPE, FOSSIL, and LANT for July-December 2002. Increasing values indicate tilt down to the east and north, respectively. See Figure D-1 for locations.

Figure T-2b. Borehole tiltmeters SHERWIN, VALT, BG, and MX for July-December 2002. Increasing values indicate tilt down to the east and north, respectively. See figure D-1 for locations. Note that the tilt transients on BG and MX beginning in mid-September reflect cementing and subsequent thermal curing of the boreholes.

MAGNETIC MEASUREMENTS (M.J.S. Johnston)

BACKGROUND

Local magnetic fields at 12 sites in the Long Valley Caldera are transmitted via satellite telemetry to Menlo Park every 10 minutes. These and other data provide continuous 'real-time'

monitoring in this region through the low-frequency data system. The location of these sites is shown on Figure M1. Temporal changes in local magnetic field are isolated using

simple differencing techniques.

Figure M1. Locations of differential magnetic field stations within Long Valley caldera. The reference station MGS (not shown) is located along Highway 395 approximately 20 km southeast of the caldera.

DATA

Plots of daily averaged data from the telemetered magnetometer stations

in and near the caldera are shown in Figure M-2.

HIGHLIGHTS

[1] While there are some unusual changes at individual stations the only event of real interest during this period was a decrease in magnetic field of several nanotesla at stations in the south moat (SH and VA) starting in early October. See Fig M2.

[2] Detailed analysis of all magnetic field data indicates the source of most of the annual variation in the magnetic difference data results from local demagnetization and remagnetization effects. Correction of these effects is now possible in Long Valley and on other volcanoes around the world. This result will be presented by Hashimoto et al. at the IUGG meeting in June 2003.

[3] Tests of a new natural magnetic amplifier effect to increase the sensitivity to stress changes in Long Valley are being conducted at the PL site. This effect relies firstly on the fact that stress generated magnetic changes are larger at depth than at the surface and secondly, that these larger changes can be piped to the surface using deep borehole casing to trap the fields. Using this effect, we have discovered hints of a magnetic signature with the same timescale as the strain transient triggered by the 1999 Hector Mine earthquake. This result will also be reported at the IUGG. 

CO₂ STUDIES (Ken McGee, Terry Gerlach, and Mike Doukas, Cascades Volcano Observatory Vancouver, WA)

The GOES-telemetered carbon dioxide monitoring network (Figure C-1) in the Mammoth Lakes area continued to transmit data on soil gas carbon dioxide concentrations throughout the report period. Station HS1 is located near the central portion of the Horseshoe Lake tree kill in an area of high CO₂ ground flux while HS2 is located in a lower flux area near the margin of the tree kill and HS3 is outside the tree-kill zone in the group campground area. Stations located away from Horseshoe Lake include SKI, located near Chair 19 in the Mammoth Mountain Ski Area, SRC, located at Shady Rest Campground adjacent to the USFS Visitor Center in Mammoth Lakes, and EQF, located near Earthquake Fault. At all sites, CO₂ collection chambers are buried in the soil. Air from these collection chambers is pumped to nearby carbon dioxide sensors housed in USFS structures or culverts. Local barometric pressure is also measured at HS1 using a Vaisala Pressure Transducer. Data are collected from the sensors every hour and are telemetered every three hours via GOES satellite. The GOES transmitting antennas, mounted inside the USFS structures, continue to produce strong signals to the satellite even after significant snow buildup on the roofs of the structures. All monitoring sites have backup data loggers that also record ambient temperature. Snow data are obtained from a U.S. Bureau of Reclamation monitoring station at Mammoth Pass

Data for 2002 for most of the telemetered monitoring stations are shown in Figure C-2 along with snow depth (SWE) at Mammoth Pass. [Note: all dates and times in UT. Data not corrected for pressure and temperature.] There are several gaps in data in 2002 caused mainly by power problems. The obvious breaks in the data in August at HS1 result from annual servicing of the monitoring stations. Station HS2 ceased operation in February and was not repaired until August due to budget constraints. Likewise. Station EQF ceased operation in late May and was not put back into service until August. Unfortunately and unbeknownst to us in advance, the USFS tore down and moved the building housing our monitoring station at EQF in September. This occurred after our annual servicing trip to Long Valley and thus we will not be able to rebuild the station at EQF until next summer. On the last day of September, the company that operates the campground facilities in the region for the USFS had the power shut off to all the facilities at Horseshoe Lake. Despite our pleas and phone calls, the power was not turned back on for a few days causing all three of our monitoring stations at Horseshoe Lake to exhaust their standby battery supply and cease operation. As we were not able to travel during the first week of the new fiscal year, we shipped our GOES programmer to Stuart Wilkinson who was able to bail us out once again and restart all of the stations. This abrupt power shut down at the end of the summer season has been an intermittent problem for us for some years.

The typical annual buildup of CO₂ in the core area of the anomaly at Horseshoe Lake can be seen during the first few months of 2002 in the plots for HS1A, HS1B, and to some degree at HS3 (Figure C-2). The peaks at HS1, HS2, and HS3 in November are also likely a response to the early snowfall this season. Beyond a small unusual event at SKI in early April, there were few CO₂ degassing events of any significance during 2002.

During the annual monitoring station servicing trip to Long Valley in August, CVO gas project personnel also conducted another soil CO₂ efflux survey at Horseshoe Lake. The results, shown in Figure A-4 under the REVIEW OF 2002 section, do not differ significantly from earlier surveys. The long term degassing rate average is about 100 tons of CO₂ per day at Horseshoe Lake.

Figure C-1 Map showing locations of the continuous CO₂ -monitoring stations.

Figure C-2. Carbon dioxide (CO₂) concentrations for the monitoring stations in Figure C1 for 2002. CAUTION: Raw Data - not corrected for pressure or temperature.

HYDROLOGIC MONITORING (Chris Farrar, Jim Howle, and Michelle Sneed: U.S. Geological Survey, Carnelian Bay and Sacramento, CA).

Hydrologic data collected for the USGS Volcanic Hazards Program in this report include ground-water level data from five wells; stream flow, water temperature, and specific conductance from one site on Hot Creek; and estimated thermal water discharge in Hot Creek Gorge (figure H1). Additional data are available upon request – contact: Chris Farrar or Jim Howle at Carnelian Bay 530.546.0187.

BACKGROUND

Ground-water levels in wells and the discharge of springs can change in response to strain in the Earth’s crust. The network of five wells and one surface water station provides hydrologic data that contributes to monitoring deformation and other changes caused from magmatic intrusions and earthquakes in Long Valley Caldera.

GROUND-WATER LEVEL MONITORING

Ground-water levels are measured continuously in five wells, LKT, LVEW, SF, CW-3, and CH-10B (figure H1), using pressure transducers that are either submerged below the water surface or placed above ground and sense back-pressure in a nitrogen-filled tube extending below the water surface. Barometric pressure is also measured at each site using pressure transducers. The data are recorded by on-site data loggers and telemetered on a three-hour transmit cycle using the GOES satellite and receivers at Menlo Park and Sacramento. All sites are visited monthly to collect data from on-site recorders and to check instrument calibrations.

Data processing is done in the Sacramento Office. Records of barometric pressure are used in combination with the water-level records to determine aquifer properties from the observed water-level response to atmospheric loading and earth tides. The influences of barometric pressure changes and earth tides are removed from the water-level records. The result yields the filtered water-level record that may contain other hydraulic and crustal deformation signals. Filtered data for wells LKT, CW-3, and CH-10B are given in figures H2, H5, and H6. The steep pressure drops recorded during late 1997 in all three wells probably are mostly caused by the high rate of crustal extension in the central part of Long Valley Caldera during that same period. Analysis of the records from LVEW and SF to provide filtered data is not yet complete; therefore raw data are presented for these two sites (figures H3 and H4).

Water levels in CW3 are affected by pumping at the Casa Diablo geothermal field. Examples of these effects include the large pressure drop in 1991 and the distinct peak in 2000.

Water levels in all five monitor-wells declined steadily or with only minor periods of rise during 2002. This pattern is similar to that recorded for 2001. By the end of 2002, water levels in LKT reached the lowest level since 1996, the lowest level in SF since the beginning of record (1996), the lowest level in CW3 since 1995, and the lowest level in CH10B since 1986. The low water levels are due to in part to below average annual precipitation in 1999 and 2000 followed by approximately average precipitation in 2002.

SURFACE WATER MONITORING

Site HCF is located downstream from the thermal springs in Hot Creek Gorge (figure H1). Stage, water temperature, and specific conductance (figure H7) are recorded every 15-minutes. The data are recorded by an on-site data logger and telemetered every three hours. Specific conductance is a measure of total dissolved ionized constituents. Water at HCF is a mixture of thermal water from springs along Hot Creek and non-thermal water from the Mammoth Creek basin. Changes in specific conductance are related to changes in the mixing ratio of thermal and non-thermal components of stream flow. Water temperatures change in response to ambient temperatures and the mixing ratio.

THERMAL WATER DISCHARGE ESTIMATE

Estimates of total thermal water discharge (figure H8) are computed from monthly measurements of discharge, and boron and chloride concentrations collected at a non-recording site (HCA) located upstream of the Hot Creek gorge thermal area and at site HCF downstream. The quantity of thermal water discharged to Hot Creek is known to vary in response to seasonal variations in precipitation, snow-melt, earthquakes, and other processes. It is believed that spring discharge may change in response to crustal strain. Thermal water discharge in Hot Creek gorge was lower in total during 2002 than in any year since 1994.

Ground-Water Level Responses to the Nov. 3, 2002 Alaska Earthquake

The water levels in all five of the key Long Valley volcanic hazards monitoring wells responded to the magnitude 7.9 Alaskan earthquake on the Denali Fault, November 3, 2002 (figs H9 and H10). Note units are in feet and the scale varies for each site to accentuate the response.

The two wells on the resurgent dome (SF and LVEW) both showed a rapid rise in water level ranging from about 0.1 to 0.3 feet (3 to 9 cm). The three wells around the perimeter of the resurgent dome (LKT, CW-3, and CH10B) all responded with a rapid drop in water level, ranging from about 0.3 to 1.0 feet (9 to 30 cm). Water levels returned to pre-earthquake trends almost immediately in most wells and within 5 days or less for LVEW.

REVIEW OF 2002

Activity in 2002 was dominated by the onset of renewed inflation of the resurgent dome following nearly three years of gradual subsidence. Earthquake activity within the caldera, which remained low through the first half of the year, showed a slight increase through the second half. Of particular note was the response of the caldera to the shear and surface waves generated by the M=7.9 Denali Fault earthquake of 3 November 2002 in the form of a burst of some 60 small earthquakes beneath the south flank of Mammoth Mountain, a coincident strain transient consistent with asiesmic slip on a normal fault beneath the east flank of the mountain, and an earthquake swarm the following day in the south moat that included the first M=3.0 earthquake since 1999. This is the third time we have a well-documented response of the caldera to large, distant earthquakes, the first two being with the M=7.4 Landers earthquake of 28 June 1992 and the M=7.2 Hector Mine earthquake of 16 October 1999. No other significant changes occurred within the caldera during the year. Both the carbon dioxide flux from the flanks of Mammoth Mountain and the rate deep long-period (LP) volcanic earthquakes beneath Mammoth Mountain showed little change from previous years. We detected no very-long-period (VLP) earthquakes during 2002.

CALDERA UNREST

Beginning around the first of the year, both the 2-color EDM and continuous GPS data for the baselines radiating from the CASA monument turned from gradual contraction to renewed extension that persisted through the year at rate of 2.5 to 3.0 cm/year (see Figure G-2). This rate is comparable to extension rates that prevailed through the mid-1990’s (Figure A-1).

Figure A-1. Line-length changes across the resurgent dome with respect to the monument CASA for 1984 through 2002 based on the 2-color EDM measurements (red crosses) and continuous GPS data (black circles). See Figure G-1 for locations.

Cumulative uplift of the center of the resurgent dome associated with this extension has returned to its 1999 value of roughly 80 cm with respect to the late 1970’s.

Earthquake activity within the caldera remained low through the first half of the year averaging fewer than five earthquakes per day, most with magnitudes less than M=2.0 (Figures A-1 through A-4). The largest event within the caldera during this period was a M=2.8 earthquake at 12:20 PM (PST) on 15 March located in the west lobe of the south moat seismic zone 1 mile south of the 203-395 Highway junction. Activity increased slightly in mid-June beginning with a cluster of small earthquakes beneath the west flank of Mammoth Mountain on June 26 that included four M~2 earthquakes. A number of small (M<2) events with the appearance of LP earthquakes occurred at shallow depths (< 2 km) beneath the southern section of the resurgent dome during the last half of August.

Figure A-2. Earthquake epicenters in the Long Valley caldera region for 2002. The plotted epicenters all have CUSP locations.

The most notable activity began with a burst of over 60 small M<1 earthquakes beneath the south flank of Mammoth Mountain as the surface waves generated by the M=7.9 Denali Fault, Alaska, earthquake of 3 November 2002 just 17 minutes after the 22:12:41 UTC mainshock rupture. At the same time, the borehole dilatometers detected a 0.1-microstrain strain transient that is consistent with slow (aseismic) slip on a normal fault at a depth of about 7 km beneath the west flank of Mammoth Mountain. As with the caldera activity remotely triggered by the M=7.4 Landers earthquake of 28 June 1992 and the M=7.2 Hector Mine earthquake of 16 October 1999, this strain transient is much larger than can be explained by cumulative slip for the 60 or so M<1 earthquakes triggered by the Denali Fault earthquake. The following day, the largest earthquake swarm in the south moat of the caldera since 1998 developed as a sequence that included six M>2 earthquakes and M=3.0 earthquake, the latter at 9:08 PM (PST) on the 4^th (Figure A-2). This south-moat swarm was unusual in that (1) it occurred in a relatively aseismic section of the south moat, (2) focal depths of the swarm earthquakes were unusually shallow (z < 4 km), and (3) the NNW lineations of the swarm epicenters cuts across the prevailing WNW-trend of the usual south-moat swarm activity. The latter was also true for the swarm activity triggered by the M=7.4 Landers earthquake of 1992. This south-moat earthquake swarm was not accompanied by detectable strain changes.

Midcrustal long-period (LP) earthquakes have continued at depths of 10 to 25 km beneath Mammoth Mountain at a fairly steady rate over the past three years. Occasional bursts of activity include 12 to 15 events per week (Figure A-4).

Figure A-3. Temporal occurrence of earthquakes in the Sierra Nevada block south of the caldera (left) and within the caldera (right) for 2002 plotted in Figure A-2. Top: number of M>1 earthquakes per day. Bottom: earthquake magnitudes with time with magnitudes proportional to line lengths and corresponding focal depths in km keyed to symbols at the right.

Figure A-4. History of deep long-period earthquakes beneath Mammoth Mountain: 1989-2002.

Diffuse emission of carbon dioxide from the flanks of Mammoth Mountain showed little change from previous years. Emission rates estimated for the Horseshoe Lake tree-kill area continue to fluctuate between 50 to 150 tones of CO₂ per day with an average flux of 100 tones per day since 1995 (Figure A-4). The Horseshoe Lake area produces roughly one third of the total CO₂ flux from the flanks of Mammoth Mountain.

Values for the helium isotope ratio He³/He⁴ from samples taken in early- and mid-2002 from the Mammoth Mountain Fumarole (MMF) located at the 3,000 m level some 300 m east of the Chair 3 ski lift average 5.5, or essentially the same as the 2001 values (Figure A-5). These values are significantly higher than the 1999 value of 3.0. The increase with respect to 1999 is consistent with an increase in the magmatic component in the gas emissions from the fumarole. Whether the elevated values for 2002-2002 are related to the occurrence of the very-long-period (VLP) volcanic earthquakes that occurred at a depth of 3 km beneath the summit of Mammoth Mountain in July and August of 2000 remains to be seen.

Figure A-5. Results of soil CO₂ efflux surveys at the Horseshoe Lake tree-kill area by the CVO Gas Project team since 1995. Results from the surveys in the early years were corrected for the effects of surface preparation as described by Gerlach and others [Chemical Geology, 177, 101-116, 2001]. The long term average is about 100 tons of CO₂ per day. The dashed line is the emission rate trend based on linear regression.

Figure A-6. Helium isotope ratios in fumarole MMF on the north side of Mammoth Mountain together with the occurrences and magnitudes of deep long-period (LP) earthquakes (see Figure A-4) beneath the west flank of Mammoth Mountain and large regional earthquakes for the period 1989 through mid-2002. The two very-long-period earthquakes of July and August 2000 are indicated by the double arrows labeled “VLP Events”.

REGIONAL EARTHQUAKE ACTIVITY

Seismic activity in the region surrounding Long Valley caldera continued to be dominated by earthquakes in the SSW-trending aftershock zone of the June and July 1998 M=5.1 earthquakes and the May 1999 M=5.6 earthquake in the Sierra Nevada south of the caldera (Figure A-2). Activity within this aftershock zone included a cluster of earthquakes near the southern end of the zone centered just east of Grinnell Lake that began on June 6 and persisted through the end of June (Figure A-3). This cluster included three M>3 earthquakes, the largest of which was a M=3.3 earthquake at 2:37 PM (PDT) on June 2. The largest earthquake of the year was a M=3.8 event at 12:47 AM (PDT) on October 6 located midway along this zone (2 miles SSE of McGee Mountain), which followed a M=2.9 foreshock by one minute. A M=3.0 earthquake in the same area occurred at 4:56 PM on October 22. Additional activity in the Grinnell Lake cluster produced a M=2.9 earthquake at 5:08 PM on October 27 and M=3.0 and 3.3 earthquakes on December 7 and 23, respectively. A M=3.7 earthquake at 11:59 AM (PST) on March 10 was located 2 miles SSW of Mount Morrison

Elsewhere, a M=3.7 earthquake at 1:18 PM on July 15 just 2 miles north-northwest of Bishop produced felt shaking throughout the Bishop area. M=2.9 and M=3.5 earthquakes at 9:26 PM and 10:58 PM on December 12, respectively, were located beneath the Volcanic Tableland 12 miles north-northwest of Bishop.