The legacy of the 1811-1812 New Madrid, Central United States, earthquakes is one of tremendous enigma. We are left with just enough contemporary information to provide a measure of constraint on the isoseismal contours and therefore magnitudes of the three principle events (Nuttli, ; Street, ; Johnston, ; Hough et al., 2000), yet given the sparse population density and limited documentation of effects, our interpretations will always be plagued by a significant degree of uncertainty. Although the magnitudes of the three principal New Madrid mainshocks will likely never be established with precision, all contemporary analyses [see above references] obtain magnitudes upwards of 7 for all three events--large enough to produce perceptible ground motions as far away as the Atlantic seaboard.
One enduring and interesting bit of folklore concerning the New Madrid earthquakes of 1811-1812 involves the tale of a volcanic eruption in North Carolina at the time of the first mainshock on December 16, 1811. The site of this supposed volcano was approximately 750 km east of the mainshock and 40 km northwest of Asheville, North Carolina, at an area known in the early 1800s as "the springs" or "the warm springs." I will henceforth refer to the location by the name of the town that exists there now: Hot Springs. The tale has merited a brief mention in some modern treatises on the New Madrid sequence. For example, Penick  mentions the 1812 letter by Asheville, North Carolina resident John Clarke Edwards that described the supposed eruption. Penick  goes on to note that the letter was quickly discredited as a hoax.
However, there turns out to be more to this particular tale than has yet been told. The primary purpose of this note is to provide a more thorough account of the tall tale than has been available previously in the modern literature. This includes presentation of another letter evidently written by the same Mr. Edwards, one that predated the notorious hoax letter and that has not previously been included in modern compilations of accounts of the New Madrid events. The secondary purpose of this note is to suggest that the first letter may have been a reliable account, including observations that clearly sowed the seeds from which the outlandish volcano tale later sprung. I will discuss the possibility that the first letter may indeed relate a credible account of the phenomenon of "earthquake lights."
However, the infamous volcano letter was in fact the second published letter signed by a John Clarke Edwards. A first letter, dated 12/19/1811 and signed John C. Edwards, was published in the Chillicothe Fredonian in February of 1812 and is reproduced in Appendix 1. Whereas the second letter is striking in its outlandishness, the first is generally striking in its sobriety.
First, if one can assume that the date of the letter is not a fabrication, it was very likely written before news of the first New Madrid earthquake could have reached an observer in North Carolina. The New Madrid sequence predated the invention of the telegraph, so news traveled at the speed of boats and horses (to Asheville, presumably the latter).
Second, the second-hand description of the lights and noises near the warm springs does not fit simple, conventional notions of volcanic eruptions. Instead, it presents observations that would have almost certainly been highly perplexing to observers in the early 1800's. It therefore appears unlikely that the initial account was written with the specific intent of creating a hoax.
Finally, one can look to the rest of the letter for insights into its possible veracity. According to Edwards' first-hand accounts, many residents of Asheville heard a rumbling noise before they felt shaking. Given the efficiency of high-frequency wave propagation in the eastern United States [e.g., Frankel et al., 1990], it is not uncommon for the P wave to be heard rather than felt. And since the S-P time would have been on the order of 1-2 minutes at that epicentral distance (~800 km), it is reasonable to conclude that the first shaking perceived by observers at their doorways was the S wave. The description of the sudden and stronger shaking that began quickly after the first jolt is consistent with a (generally stronger) Lg arrival following the S wave by 20-30 sec. Even the total duration of the shaking--about three minutes--is remarkably consistent with descriptions at comparable epicentral distance elsewhere.
The first Edwards letter makes another observation that has the ring of truth to the modern seismologist, "It is somewhat strange that its effects were more violent in the vallies than on the mountains." In making and recording this observation, Edwards became one of a half-dozen individuals to explicitly document the effects of site response at sedimentary sites during the New Madrid sequence [see Hough et al., 2000]. Again, it defies logic to suppose that Edwards fabricated an observation that he himself found puzzling.
The first Edward letter does include a couple of remarks that perhaps sow seeds of suspicion. First is the recounting of noises heard at the warm springs over the course of several hours. Such an observation would defy a modern understanding--which of course does not make it necessarily wrong--but could be a simple mistake in the relaying of a second-hand account. I do note that there is another contemporary account from the southeastern United States that describes rumbling heard for an hour prior to the December 16, 1811 event, from Charleston, South Carolina (see Mitchill, 1815; http:pasadena.wr.usgs.gov/office/hough/mitchill.html). Again, though, if the P wave were heard rather than felt at Hot Springs, the noises would have lasted for several minutes.
A more problematic aspect of the letter is its reference to the fall of Painted Rock on the road from Asheville to Knoxville. This report would eventually be discredited along with the volcano story. However, from the letter it is unclear if Edwards is claiming first-hand observation of Painted Rock or not.
On balance, I argue that the first letter from Edwards is as credible as a number of other contemporary accounts from the New Madrid sequence that have been used in contemporary analyses. The assignment of shaking levels based on individual, uncorroborated accounts is, moreover, standard practice in the interpretation of felt reports from the New Madrid sequence [Nuttli, 1973; Street, 1982; Hough et al., 2000]. Although clearly not ideal, the paucity of accounts precludes any other option.
The early morning shock described by Mr. Nelson was undoubtedly the large aftershock that occurred at approximately 7:15 a.m. local time on the morning of 12/16/1811 [see Street, 1982]. Although inferred to have been smaller than the mainshock, the aftershock was almost as widely felt because it occurred during waking hours. Here again, the description of shaking--the agitation of trees and the duration of shaking--are very much consistent with observations elsewhere. Although unusual, the descriptions of the ground water phenomena appear credible.
There is no question that naturally-occurring warm springs have existed for some time at Hot Springs. As discussed by O'Hara et al., , the springs are thought to result from unusually deep groundwater circulation within the Rector Branch thrust (RBT) fault. The RBT is one of the primary faults associated with the southern Appalachian Mountains in North Carolina. Fluid inclusion studies at Hot Springs also indicate high salinity (CaCl2 - MgCl2) levels [O'Hara et al., 1995].
Because of the unique geology and hydrology of Hot Springs, it is reasonable to consider sonoluminescence as a plausible explanation for lights associated with earthquake waves (so-called ``earthquake lights''). As discussed by Johnston (1991), sonoluminescence is a well-documented phenomenon whereby visible lights are generated from molecular reactions in water that has been strongly shaken by compressional waves. Sonoluminescence occurs in the laboratory at compressional wave energy densities of 10-20 erg/cm**3; Johnston (1991) shows that strong earthquake shaking can produce energy densities 20-100 times as high. Although sonoluminescence in pure water produces blue to bluish-white light, the color can shift to yellow and red in the presence of dissolved salts, which could account for the "flashes of fire."
I consider plausible ground motion parameters at Asheville based on the descriptions of shaking provided in the first Edwards letter. The earthquake shaking and rumbling appears to have awakened most, or at least many, residents of Asheville. Although Edwards says that the effects were too numerous mention, the only effect that is described is a modest one: tanning vats (which one imagines were likely to have been top-heavy and unstable) being knocked around. These vats were moreover apparently located in a valley, where stronger shaking was experienced than in the hills. This sort of inconsistency--a dramatic overall description of shaking but minimal reported effects--is extremely common in accounts of the New Madrid earthquakes at distances of several hundred kilometers from the source (Hough et al., 2000). I do not consider it evidence of incredible reporting but rather an understandable reaction to the circumstances. That is, the ground motions from Mw7+ earthquakes at regional distances can seem surprisingly powerful by virtue of their long-period energy and not insubstantial amplitudes, yet often will cause only minimal damage. Overall, the description of shaking and effects at Asheville is similar to many others from comparable distances (see Figure 1) and is consistent with a Modified Mercalli Intensity (MMI) value of V (see Richter, 1958; Stover and Coffman, 1993).
A number of studies have shown that the MMI scale can be reasonably well calibrated with peak ground acceleration values (e.g., Trifunac and Brady, 1975; Wald et al., 1999). According to conventional interpretations, an MMI of V corresponds to peak shaking of 1.5-4% g. A more recent study by Wald et al. (1999) finds that a wide range of peak accelerations corresponds to MMI V, from roughly 1-30% g. Thus, the lower estimate of a few percent g is probably, if anything, conservative.
One can proceed to consider the frequency content of the ground motions by virtue of the inference that the P wave was audible. Frequencies as low as 20 Hz are within the audible range of sound for humans. Because attenuation of ground motions will increase with increasing frequency, we will assume a ground motion frequency of 20 Hz (i.e., and no higher). A peak ground acceleration of 1.5-4% g with a period of 0.05 s implies a peak ground displacement of ~0.04-0.10 cm.
Using these ground motion parameters in the relation for energy density e induced by one cycle of an advancing wavefront,
e = [2*(pi)^2*rho][(A/tau)]^2
one obtains an energy density of 14-38 erg/cm**3. I also note that, given the long duration of P-wave shaking, this value may well represent a lower bound, as fluids would have been shaken by a large number of cycles. Even as a lower bound, however, the value is high enough to generate sonoluminescence effects. The high salinity content of the water is, moreover, consistent with a yellow-red color, which would have imparted a fire-like appearance to the flashes of light.
The loud snaps described in conjunction with the lights cannot be explained as a sonoluminescent effect and remain somewhat enigmatic. As a speculation, they could, however, have been produced by a mechanical consequence of the significant shaking and/or ground water effects. Recall the description of water having been "thrown up" into the air during the 7:15 a.m. aftershock. Although the generation of geyser effects is unusual, earthquakes are known to generate significant hydrological disturbances at large epicentral distance, often involving increased stream and spring flow [e.g. Rojstaczer et al., 1995]. If the 2:15 a.m. mainshock generated similar effects, loud noises could have resulted either from hydrofracture or as a consequence of rock falls.
I conclude that the second-hand account of the effects at Hot Springs was also most likely credible, and I have shown that the observations are generally consistent with the phenomenon known as earthquake lights. Given the unusual local geology at Hot Springs and the prolonged high-frequency shaking, earthquake lights can be plausibly explained as a sonoluminescence effect.
Unfortunately, given that the phenomena at Hot Springs may have only been witnessed by two individuals, it appears unlikely that any additional information exists to be gleaned from original sources regarding this particular observation. The discussions presented in this paper are, therefore, speculative in nature. However, it is possible that accounts of past and future earthquakes will provide a test of the conclusion that, under the right circumstances, earthquake lights can be generated at epicentral distances of several hundred kilometers, especially at naturally occurring springs. It is also possible that an archival search would uncover additional reports that at least corroborate the shaking effects in Asheville, which would be of value in establishing the credibility of Edwards' first letter. (The Raleigh Star was read and found to contain no reports from Asheville, but this is not surprising considering the ~350 km distance and lack of a waterway between the two towns).
In any case, it is hoped that this note will serve as a more complete,
and, perhaps, interesting, account of an intriguing tale than has
not previously been available in the modern literature. The New Madrid
earthquakes generated a number of interesting stories and legends,
few quite as colorful as the one concerning the (alleged) volcano
in Hot Springs, North Carolina.
Acknowledgments. I thank Ken Hudnut, Lisa Wald, and John Ebel for helpful reviews and discussions. Figure 1 was generated using GMT software Wessel and Smith, 1991].
Frankel, A., A. McGarr, J. Bicknell, J. Mori, L. Seeber, and E. Cranswick (2000). Attenuation of high-frequency shear-waves in the crust - Measurements from New York State, South Africa, and Southern California, J. Geophys. Res. 95, 17441-17457.
Hough, S.E., J.G. Armbruster, L. Seeber, and J.F. Hough (2000). On the Modified Mercalli Intensities and Magnitudes of the 1811-1812 New Madrid, Central United States earthquakes. J. Geophys. Res., 2000.
Johnston, A.C. (1991). Light from Seismic-waves, Nature 354, 361.
Johnston, A.C. (1996). Seismic moment assessment of earthquakes in stable continental regions--III. New Madrid 1811-1812, Charleston 1886, and Lisbon 1755, Geophys. J. Int. 126, 314-344.
Mitchill, S.L. (1815). A detailed narrative of the earthquakes which occurred on the 16th day of December, 1811, Trans. Lit. Philos. Soc. N. Y. 1, 281-307.
Nuttli, O.W. (1973). The Mississippi Valley earthquakes of 1811 and 1812: Intensities, ground motion, and magnitudes, Bull. Seism. Soc. Am. 63, 227-248.
O'Hara, K.~D., D.~L. Kirschner, and D.~P. Moecher (1995). Petrological constraints on the source of fluid during mylonization in the Blue Ridge Province, North Carolina and Virginia, USA, J. Geodynamics 19, 271-287.
Penick, J.L., Jr. (1981). The New Madrid Earthquakes, revised edition, University of Missouri Press, Columbia and London.
Richter, C.F. (1958). Elementary Seismology, W.H. Freeman and Co., San Francisco.
Rojstaczer, S., S. Wolf, and R. Michel (1995). Permeability enhancement in the shallow crust as a cause of earthquake-induced hydrological changes, Nature 373, 237-239.
Stover, C.W. and J.L. Coffman (1993). Seismicity of the United States, 1568-1989 (revised), U.S. Geol. Surv. Prof. Pap., 1527.
Trifunac, M.D. and A.G. Brady (1975). On the correlation of seismic intensity scales with the peaks of recorded ground motion, Bull. Seism. Soc. Am. 65, 139-162.
Street, R. (1982). A contribution to the documentation of the 1811-1812 Mississippi Valley earthquake sequence, Earthq. Notes 53, 39-52.
Wald, D.J., V. Quitoriano, T.H. Heaton, and H. Kanamori (1999). Relationships between peak ground acceleration, peak ground velocity, and modified Mercalli intensity in California, Earthq. Spectra 15, 557-564.
Wessel, P., and W.H.F. Smith (1991). Free software helps map and display data, Eos Trans. Am. Geophys. U. 72, 441-446.