The history of seismology at HVO begins in 1912 with the completion of the Whitney vault. After two decades of pioneering, Jaggar was amply aware that seismometry on an active volcano is quite different from that at a station that studies mainly distant earthquakes.
Instrumentation provided a challenge from the beginning. The story of the evolution of seismic instruments at HVO is documented in Curtis (1913), Jaggar and Romberg (1918), Apple (1978), and various issues of HVO's own publications, the Monthly Bulletin and Volcano Letter. The following account is synthesized and condensed from these sources.
Seismologist H.O. Wood reported for duty June 13, 1912, and he had the crated instruments from Japan installed on their "concrete tables" by July 2--installed, but not necessarily performing properly. These instruments were: (1) a sensitive Omori tromometer, designed for the registration of earthquake motion proceeding from a distant origin; (2) a less-sensitive Omori tromometer, designed for the registration of weak or moderate shocks of local origin; and (3) an ordinary Omori seismometer that started itself upon detecting a strong shock and that registered three components of motion, east-west, north-south, and vertical. The more sensitive tromometer arrived missing the north-south component; it was therefore mounted in such a way as to measure and register the component of earth motion in the east-west direction. Both tromometers had smoked drums and timing capability with marks every one minute. The ordinary Omori was set up so that when earth motion was strong enough to activate its starting device, the smoked drum revolved rapidly and marks were made every one-half second. This instrument could also be set to ring an alarm bell when it started.
Heavy horizontal pendulums detected horizontal earth motions, and a "floating" weight suspended by two balanced helical springs detected, in theory at least, the component of earth motion in the vertical direction. Tune marks were made by an electromagnet operated by the brief closing of an electric circuit. Wood had all three instruments operating by August 1912, but all had, sometimes separately and sometimes simultaneously over the years, periods of nonoperation when work was done in the Whitney vault or on the instruments.
Of these three seismometers, the first to leave the Whitney vault was the less sensitive Omori. It was moved to the Technology Station at Halemaumau in February 1913 and later was apparently dismantled for parts. The three-component self-starter never really worked properly, even after it was rebuilt in 1918, renamed "Domesticus," put on full-time duty, and its vertical detection apparatus removed.
The most sensitive Omori seismometer was rebuilt into an optical recorder and given a corner mount in 1918 in the hope that it would detect both north-south and east-west teleseismic motion with its single boom. In 1923 it was retired. After retirement, this and all seismometers were cannibalized for parts to make or repair other instruments in the HVO machine shop.
During their service, the seismometers were modified and changed by the HVO staff to try to make them more suitable for use on the rim of Kilauea caldera--to be more responsive to local short-period earthquakes, volcanic tremor, and the extraordinary ground tiltings. These were the three distinctively Hawaiian geophysical phenomena Jaggar had identified by 1918. Jaggar noted that the mechanical imperfections of standard instruments, designed for cheapness and convenience of operating and ill-adapted to such special problems, combined to yield only mediocre results. Dr. Arnold Romberg was hired by HVRA for the summer of 1918 to rebuild and otherwise improve the two original seismometers still in the Whitney vault and to do some work on the Bosch-Omori seismometer that had been acquired in 1913. Romberg also made at least two new one-component seismometers for installation in cellars to be established by HVRA in Hilo and Kealakekua. As Romberg worked, Jaggar noted with satisfaction that small improvements had been made over experiments by Milne, Galitzin, and others and that already the difficulties of friction, magnification, damping, opening the record, and time marking had been partially overcome.
The two-component Bosch-Omori seismometer, purchased from the Whitney Fund, had finally arrived from Germany by April 1913. Wood noted that in order to permit the construction of foundations and the work of installing the Bosch-Omori, a 100-kg tromometer, the seismographs in operation had to be partially dismounted temporarily. The new instrument was mechanically and dynamically superior to those hitherto in use there, and it was expected that a larger proportion of teleseismic records would be written, as well as more precise records of local shocks.
Trial runs of the new Bosch-Omori's east-west and north-south components were made on April 30, 1913; a feeble teleseism was registered, as well as strong volcanic vibrations. Adjustments were completed on May 8, 1913, for the final trial runs. From that time on through 1953, the Bosch-Omori seismometer was the main reliable, basic geophysical instrument of HVO. Fine tunings of the Bosch-Omori never ended and were required after almost every unhinging caused by a strong earthquake. Modifications and upgradings were intermittent: steel wire replaced silk fibers to suspend the weights; oil damping baths were added, adjusted, redesigned, and readjusted; friction reduction was tried; new hinges were designed and manufactured; magnifications were lowered and raised; different levers, recording pens, and pins were tried; timing devices were improved; better drum-smoking devices were found; Dr. Romberg modified linkages and levers in 1918 so that one smoked drum recorded what two drums did originally; and one of the piers was rotated 7.5º to permit its boom to swing closer to a true east-west line.
In 1948, the Bosch-Omori seismograph in the Whitney Laboratory had a period of 7.7 seconds and a magnification of about 115 times the earth movement. It gave satisfactory records of short-period local earthquakes and of the ground tilt. This standard was the goal always attempted and sometimes met over the decades. Credit goes to the HVO staff for the daily attention they gave to the Bosch-Omori. One or more of the staff came day or night, storm or sunshine, workday or weekend, whenever a felt earthquake or home alarm signaled a potential dismantling.
Whitney vault's Bosch-Omori was the workhorse of the first four decades of the Hawaiian Volcano Observatory, 1913 to 1953. HVO staff habit, respect, and momentum kept the Bosch-Omori seismometer operating in the Whitney Laboratory of Seismology for ten years after mechanical seismometers became technologically obsolete. The Bosch-Omori's last official smoked seismogram was removed, shellacked, and read on February 1, 1963, ending a daily series begun 50 years before. Since the standard commercial seismometers had been tried and found wanting for the detection of short-period earthquakes, volcanic tremor, and ground tiltings, HMO turned its talents to designing and building special-purpose instruments. The most successful of these was the Hawaiian-type seismograph, in service worldwide by 1928. Briefly, it consisted of two horizontal components at right angles, each having a mass as much as about 100 kg (225 lb) consisting of a 69-cm (27-in.) length of pipe of 20-cm (8-in.) diameter, filled with sand. It was hinged above and below with short links of piano wire, had a free period of about 7 seconds, and was damped with light aluminum vanes dipping in oil near the outer end of the boom, giving it a lever magnification of about 130. Time was marked by an electromagnet recording on a smoked drum with paper speed of about 3.2 cm (1.25 in.) per minute. Both components recorded on the same drum, and a drum would run 24 hours without need of a change. The pair was designed for wall mounting, one on each side of a corner, with the mounting so arranged that the assemblies at the ends of the booms could record on the single drum. The only pier needed for the Hawaiian-type seismometer was one to hold the recording drum.
Pairs of Hawaiian-type seismometers were built in the HVO machine shop and supplied to Kodiak, Alaska (1927); Hilo, Hawaii (1927); the U.S. Coast and Geodetic Survey for testing and eventual placement at Sitka, Alaska (1927); Kealakekua, Hawaii (1928); Lassen National Park, California (1928); and Dutch Harbor, Alaska (1928). One design specification met by the Hawaiian-type seismograph was convenience of operation for the semiprofessionals hired by HVRA, USGS, or the Coast and Geodetic Survey at the more remote stations.
By 1929, although so far defeated in the search for a mechanical seismometer that would record the vertical component of seismic motion, staff at HVO continued experimenting. A vertical-component instrument to match the two-component Hawaiian-type already in use was especially desired. HVO seismologists wanted to measure the angle of emergence of earthquake waves at the station and believed this vertical angle would help furnish data on the depth of origin of the quake. By 1930, the HVO machine shop had built a vertical-component seismograph, with the heavy mass hung on spiral springs and temperature compensation provided by small springs. This instrument evolved after a variety of spring arrangements were tried, and it was installed in the Whitney vault. Its performance was never up to expectation, and it was the only seismometer installed under Building 41 in 1941. When that vault was abandoned by HVO in 1948, this instrument appears to have joined those in the bin to be cannibalized.
Recognizing that seismometers took the continuing care of specialists, Jaggar searched for instruments of simple enough design to be "put in the hands of amateurs" (Jaggar, 1932, p. 3). His first "shock recorder," designed in 1928, did record separate east-west and north-south motions, but it "also recorded the motion of rats, kittens, chickens, cockroaches and spiders" (Jaggar, 1931). This was redesigned so that it consisted of two wall-mounted inverted pendulums to be set up in a corner of a room and did not require a pier. Three pairs of these were set up around Halemaumau. The first field test in the hands of "amateurs" came during the earthquake swarm in 1929 under Hualalai Volcano in Kona. A shock recorder pair was installed at Puu Waawaa Ranch on the slopes of Hualalai. Mr. Hinds the ranch owner and his family learned how to twirl the two daily recording discs over the smoking chimney of a kerosene lantern until each disc was coated an even brown. They also wound the clocks used in making the time marks.
It was this model of shock recorder that Jaggar took with him on the U.S. Navy expedition to Niuafo'ou Island (Tin Can Mail Island) in 1930 and left there with the Tongans. When earthquakes suddenly became a problem in 1930 in New Zealand, Jaggar hurriedly dispatched one such instrument in response to a plea from the New Zealand government. HVRA promised to supply eight more as soon as they could be manufactured, while in New Zealand machinists copied the one Jaggar sent. By 1937, Jaggar had a new version of his shock recorder. What it gained in simplicity, it lost in sensitivity. It no longer separated arrival times of earthquake waves but registered only maximum intensity. Each instrument depended on a small mercury cup and was called an annunciator-type shock recorder. It showed a number from 1 to 6, depending on the intensity of the earthquake. Once a number became visible, the annunciator had to be reset.
Only two of the new design were on hand when an urgent request was made in 1938 to ship as many as possible on a Coast Guard cutter leaving almost immediately to supply colonists sponsored by the U.S. Government on tiny South Pacific Islands. H.H. Waesche, a geologist at HVO, went aboard the cutter Roger B. Taney to install the two which were portable and could be set up in a few minutes with only a screwdriver and a pair of pliers. No smoked paper was needed; each instrument needed only a solid base and checking every hour or so. The colonists, all young Hawaiian men, would do the checking and report observations through the mail pouch. One recorder was set to read in an east-west direction on Canton Island, and the other to read north-south motions on Jarvis Island.
It was not necessary to wait for a natural earthquake to arrive to test seismometers and shock recorders built in the HVO machine shop; artificial earthquakes of known intensity and duration were created. HVO's earthquake machine, an oscillating table, was designed and built in 1928 in the machine shop by R.M. Wilson, HVO topographic engineer. It consisted of a massive, 365-km (800-lb) concrete slab resting on steel rollers. Instruments to be tested were mounted on the slab. Levers connected the rollers to the chuck of a nearby lathe; amplitude was governed by varying the radius of a crank in the lever train, and a range of periods could be provided by changing the speed of the lathe. Seismologists believed the oscillating table could simulate in part the ground motions of a variety of local earthquakes.
While seismometers were still being tailormade at Kilauea for local earthquakes, HVO acquired a shiny, new, compact, three-component Imamura seismometer from Japan for display to the visiting public in the building at Uwekahuna. All three components recorded on a single smoked drum. Although cars in the Uwekahuna parking lot, crowds moving about the museum floor, people clapping in the projection hall, and the range of daily temperatures in the building all confused the record, Park Naturalists believed that a working seismometer, no matter how befuddled its record, was a worthwhile exhibit and fit in with their lecture content. The Imamura instrument, with its magnification of as much as 50 times, was abandoned when HVO occupied the Uwekahuna building in 1948.
Seismographs are still the principal geophysical instruments at HVO. The electronic amplification and radio-telemetered signals of the current models would be startling to Jaggar's experiences, but not to his dreams.
A principal goal of seismological studies at HVO from their beginning has been to determine where earthquakes occur beneath the volcanic edifices. For this purpose the inference of traveltimes for earthquake waves is critical, and much has been learned from the successive attempts to interpret those traveltimes in Hawaii. The story of this evolution has been told in Jones (1935) and in various issues of HVO's Weekly Report and Volcano Letter. The following account has been gleaned from these sources.
In May 1915, HVO seismologist H.O. Wood reported that he had determined the distances of origin for 411 of the 604 local earthquakes registered in the first two and one-half years of operation of the Whitney Laboratory of Seismology. Wood used a table by Zeissig to determine these distances; for clock time he used a chronometer lent by the Territorial Surveyor, corrected by solar observations with a transit borrowed from the College of Hawaii.
The tables of Zeissig, in use then at most seismographic observatories, were designed for large earthquakes many hundreds of miles away. In 1925 HVO Seismologist R.H. Finch reported that HVO had found other tables by Omori, published in the Bulletin of the Imperial Earthquake Investigation Committee of Japan, to be more satisfactory. Omori's values came in fractions of a mile and thus helped in determining earthquake origins only a few kilometers (2-3 mi) away. That Omori's tables were better than the Zeissig table for very local earthquakes was shown by the good agreement between computed and actual distances of felt earthquakes whose approximate locations were known.
By 1925 Finch had in the Whitney cellar a Howard pendulum clock, corrected by radio signals originating from Pearl Harbor. There was a flurry of speculation in 1927 when it was noted that this clock, after being consistently late by 0.08 second per day (s/d), went to a consistent 0.66 s/d and then to a consistent 1.71 s/d. The speculation was that its pendulum might be sensibly indicating changes in the value of gravity; elevations had changed by amounts on the order of 1 m in 1924, and magma movement below might be involved.
In 1931, HVO Seismologist A.E. Jones found that the Omori traveltimes did not completely agree with the seismic facts in Hawaii. He developed a traveltime-distance graph for Hawaii, and by using S-P-wave intervals he showed "how to locate Hawaiian local earthquakes without the necessity of accurate time checks on all the master clocks of the several seismograph stations on Hawaii" (Jones, 1935, p. 59).
Trying to keep the scattered stations of the seismic network on the same time was given up by 1937. Most attention had gone to the station at Kealakekua. The line between Kealakekua and Kilauea passed practically through the middle of the great mass of Mauna Loa, and the stations were about equidistant from its center. In 1926, HVO's R.M. Wilson complained that, while the time at the Whitney vault was corrected by radio, the instrument tender at Kealakekua got his time--often more than a minute in error--over the telephone from the Kona switchboard operator. A radio set was supplied to the Kealakekua station in 1926 to receive time signals from the U.S. Naval radio station NPM at Pearl Harbor. A key beside the set, wired into the station clock, was depressed the instant the time signal was received; this recorded a mark on the current seismogram along with the one-minute marks made by the station clock. A measurement between the two marks on the seismogram revealed the correction necessary. It was believed this system recorded the time to within one-half second or better.
By 1939 H.H. Waesche, HVO geologist, had tied together for time control the seismograph stations around Kilauea summit, as well as the one on Mauna Loa at the end of the truck road. The Howard clock in the Whitney vault sent minute and hour signals of sufficient strength to operate relays controlling the electromagnets that lifted the recording needles on the drums. The connecting links between the stations and the clock were the existing National Park telephone trunklines, not connected to the commercial system. Waesche worked out all the circuitry and procured or made the parts to make the system operate. After some troubleshooting, the time signals did not interfere with telephone service in the Park.
Today, seismograph time signals accurate to a few thousandths of a second arrive by satellite. Tom Jaggar would be pleased. HVO's R.M. Wilson used Volcano Letter no. 124 in May 1927 to describe the Rossi-Forel scale and explain how HVO graded its earthquakes in terms of acceleration from "feeble" through "alarming.' Jaggar took most of Volcano Letter no. 223 in April 1929 to explain the Rossi-Forel, Cancani, and Mercalli scales, the last of which he abridged to show human effects. Jaggar concluded: "In the writer's tests, and Professor Mercalli among Italian peasants evidently had similar experience, ignorant persons in city or country can understand a human scale. They usually distort or exaggerate physical effects on masonry, water, trees, chimneys, etc., because they have no training in measured or judicial statement" (Jaggar, 1929). Because he believed in the use of a human scale to grade earthquakes, Jaggar borrowed from Japan its postcard reporting system. Preaddressed and franked, with a form to be filled in and signed on the back, postcards were widely supplied to residents on all parts of the Island of Hawaii starting in 1932. People were to fill out and mail cards whenever they experienced a felt earthquake.
In keeping with the seismicity concept brought to Hawaii by Perret, charts and graphs of seismicity curves over various periods and for various purposes were prepared as early as 1920 and regularly by the 1930's. Seismologists at HVO developed for their own use in Hawaii the Hawaiian Volcano Observatory Scale, whose grades were "Tremor, Very Feeble, Feeble, Slight, Moderate, and Strong" (Jones, 1932). It was based on the double amplitude of motion on the Bosch-Omori seismograph and carried a description of the noninstrumental effects ("a human scale"). Its grades were eventually cross-referenced to the Rossi-Forel and modified Mercalli scales.
"Just as four pounds of sugar ought to be four times as big as one pound, so a number four earthquake should be four times as big as number one. Herein lies the difficulty" (Jaggar, 1929).
When Perret lived in his cabin near Halemaumau for six weeks in 191l, part of his working time was spent with his seismoscope. His views through its eyepiece showed the ground to be continually in motion. Accordingly Perret devised a scale based on the seismicity for the day; five on a scale of 10 was an average day.
H.O. Wood as charter HVO seismologist in 1912 used the Rossi-Forel scale of intensity, but also used phrases such as "this [earthquake] manifested an intensity of little more than 1/20th of that of the minimum shock perceptible to the senses." In his report for the week ending December 19, 1912, Jaggar adopted for HVO the Cancani scale, which soon appeared on the back page of the printed weekly bulletins and stayed there through 1920. Inside, however, the text often used the Rossi-Forel scale by name, and a Cancani reference is hard to find. By the 1920's, R.H. Finch was using the Rossi-Forel to rate an earthquake recorded in the Hilea (Kau) cellar, and was rating the same quake as recorded in the Whitney vault as "slight" or "feeble."
Coastal areas in the Hawaiian Islands are vulnerable to seismic seawaves (tsunamis) generated by major earthquakes anywhere in the Pacific Ocean. The tsunami of 1933 demonstrated that seismology could be used to predict the advent of such a wave and therefore to give warning to people in threatened areas. The story of that occasion, in which HVO staff prevented possible loss of life, has been told in the Honolulu Advertiser for March 14, 1933, and in Jaggar (1933).
Capt. Robert V. Woods, a retired ship's master who was the HVRA employee at Kealakekua, Kona, operated the seismograph vault there in the cellar of his house. Woods was more than a technician who changed drums and did routine maintenance H.s.t. his instruments. He had learned from the HVO seismologists how to determine the distance and origin of an earthquake from the arrival times of its different waves.
Woods was in his basement vault at 0700 H.s.t. March 2, 1933, getting ready to change drums, when he noticed his seismometer begin to record the arrival of a distant earthquake. The time between arrivals of the first and second different waves was about ten minutes. Woods calculated the distance at about 6,320 km (3,950 mi) and put the origin at the west edge of the Tuscarora Deep, east of Japan. Meanwhile, in the Whitney vault on the other side of the Island, HVO seismologist Austin E. Jones saw the earthquake arrive on the Bosch-Omori instrument. He also calculated an origin point off Japan. Over the telephone Woods and Jones agreed that such a quake could generate a tsunami in the Pacific Ocean that would travel at about 720 km/h (450 mi/h) and arrive at the Island of Hawaii about 8.5 h after the earthquake.
Jones notified the Hilo harbormaster at about 1000 to look for a tsunami at about 1530 local time; Woods also notified American factors at the small Kona ports of Kailua and Napoopoo.
Noon radio newscasts in Hawaii featured a Japanese disaster in Iwate Province from both an earthquake and a tsunami, confirming the potential of a tsunami reaching the Island of Hawaii.
People at Napoopoo, recalling tsunamis of 1893 and 1923, removed all cargo stored on the wharf and adjacent warehouses. Other Kona coastal settlements evacuated people but not property to any great extent. Tidal surges in Kona began at about 1520 and continued for hours. They began with a recession that left the sea bottom bare at Napoopoo, Kaihia, Keauhou, and even at Kaalualu near Ka Lae (South Point).
On the seventh surge at Napoopoo, water receded vertically 2.5 m (8 ft) and then rose 3 m (9.5 ft). Up and down the Kona coast, rock walls were knocked over and scattered in low-lying areas. Boats were unmoored and capsized, houses flooded, lumber displaced and scattered, and property was washed out to sea. Cars left parked were flooded and some of their motors damaged by sand.
At Hilo, the sampan fleet moved out to anchorages in the harbor. Waves began to arrive at 1536, with an up-and-down motion of as much as 1 m (3 ft). There was no property damage.
No lives were lost in Hawaii during the 1933 tsunami, thanks to the warnings issued by the Hawaiian Volcano Observatory. This appears to have been the first time that a tsunami was predicted through interpretation of seismograms--a forerunner by decades of the Pacific-wide tsunami-alert system now in place.
ABSTRACT
INTRODUCTION
JAGGER AND THURSTON: BACKGROUND
ACKNOWLEDGMENTS
BUILDINGS AND FACILITIES
TECHNOLOGY STATION (1911-1918)
INSTRUMENT HOUSES
WHITNEY LABORATORY OF SEISMOLOGY (1912-PRESENT)
OTHER FACILITIES
BUILDING 41 (1940-PRESENT)
BUILDING 131 AT UWEKAHUNA (1927-PRESENT)
NEW BUILDING AT UWEKAHUNA (1985-PRESENT)
MAUNA LOA
ACCESS ROUTES AND FACILITIES
THE 1926 ERUPTION
CONTROLLING LAVA FLOWS
THE OHIKI AND OTHER EXPERIMENTS
