Do not cite, quote or distribute. Please refer to the published version of this article, which appears in: Archaeological Prospection, Volume 12, Issue 3, 2005. Pages 191-197, and may be dowloaded in PDF format.

ABSTRACT

Recent studies of known lightning impacts and laboratory testing of samples from a suspected lightning-induced anomaly appearing in magnetic survey data allow us to characterize these formerly enigmatic responses. Remanent magnetization associated with the current path of a lightning discharge produces strong anomalies that can be recognized in magnetic survey data, and can be positively identified using laboratory methods.

LIGHTNING-INDUCED MAGNETIC ANOMALIES ON ARCHAEOLOGICAL SITES

Geoffrey Jones and David L. Maki
Archaeo-Physics, LLC

Key Words:
Magnetic; gradiometer; lightning; remanence; lightning-induced remanent magnetism (LIRM)

INTRODUCTION

Lightning-induced remanent magnetism (LIRM) is a phenomenon that appears to be commonly encountered on archaeological sites, but rarely recognized in magnetic survey data. Strong bipolar anomalies of linear, radial, or dendritic form appearing in magnetic data plots have sometimes been interpreted as ferrous metal or igneous intrusions. A number of these enigmatic anomalies have appeared in magnetic field gradient surveys conducted by the authors as well as by other investigators. Several of these were investigated by coring and hand excavation, but no apparent sources for these anomalies could be found. The lack of visible anomaly sources and polarities not aligned to the geomagnetic field (either past or present) suggested lightning strikes as a possible cause of magnetization. Fortunately, soil samples from one of these (Figure 1), the 30-30 Winchester Site (48CA3030), located in northeastern Wyoming, were retained by the archaeological investigator, and were found to contain still-consolidated lumps that could be subjected to laboratory analysis.

Figure 1. Magnetic survey data from the 3030 Winchester Site. A Lightning induced anomaly appears at N20/E20. Smaller circular anomalies in the southern portion of the data plot are caused by Prehistoric hearth features (Jones, 2001; Munson, 2002).

LABORATORY TEST RESULTS

Consolidated soil samples from the suspected LIRM anomaly were analyzed using paleomagnetic techniques. The orientations of the samples were not recorded as they were collected. However, this was not a factor, as the laboratory analysis did not measure inclination or declination, but rather focused on intensity of magnetization and the magnetic mineral concentration, grain size, and mineralogy of the magnetized soil. The analysis used criteria for identifying LIRM previously presented by Dunlop et al. (1984) and Verrier and Rochette (1999, 2002). A complete discussion of this rather involved topic is the subject of a separate article (Maki, in press), however the results will be briefly summarized below.

The laboratory analysis determined that samples collected from near the center of the LIRM anomaly possessed an unusually intense remanent magnetization. The natural remanent magnetization (NRM) of these samples was more than 20% larger than a laboratory imparted saturation isothermal remanent magnetization (SIRM). The ratio of NRM to the magnetization induced in the presence of the earth's magnetic field (Königsberger ratio, Qn) was greater than 10. Such a large Qn ratio cannot be explained by other more common sources of remanence such as thermoremanent magnetization (TRM) or chemical remanent magnetization (CRM). Alternating field (AF) demagnetization of the LIRM sample resulted in a relatively hard curve shape, indicating that high coercivity minerals had been magnetized. Finally, the ratio known as REM' should peak at greater than 0.1 in materials magnetized by high-field isothermal processes (Verrier and Rochette, 2002; Gattacceca and Rochette, in press). REM' is the derivative of the AF demagnetization curve (dNRM/dAF) normalized by the derivative of the AF demagnetization curve of a sample given a laboratory imparted SIRM (dSIRM/dAF). Soil samples from the suspected LIRM possessed peak REM' values of approximately 1.5, well in excess of the 0.1 criteria for LIRM acquisition suggested by Verrier and Rochette (2002).

Figure 2. Suspected LIRM anomaly from site AR-03-04-06-838 CNF, in Coconino National Forest, Arizona. Blank values were recorded where the gradient exceeded the instrument's dynamic range of +/- 2047.5 nT. Excavation revealed no visible anomaly source (Somers et al., n.d.).

In addition to the above mentioned measures of magnetic remanence, several tests were conducted in an effort to determine whether the magnetic mineral concentration, grain size, or mineralogy from near the center of the LIRM anomaly was different than that found in samples obtained from surrounding undisturbed soil profiles. These tests included measurement of magnetic susceptibility, anhysteretic remanent magnetization (ARM), and the S-Ratio (a complete description of these environmental magnetism parameters can be found in Evans and Heller (2003)). No significant variation in any of these parameters was observed. The lack of a susceptibility contrast shows that induced magnetization could not have been responsible for the observed anomaly. The lack of contrast in magnetic mineral concentration, grain size, or mineralogy indicates that thermally induced mineralogical transformations did not occur in the vicinity of the lightning strike. This supports the previously discussed interpretation that TRM was not a factor in the observed signal response, as temperatures above the Curie temperature of most magnetic minerals would also result in some high-temperature mineralogical transformations. These findings are consistent with Cox (1961) and Sakai et al. (1998), who conclude that TRM is, at best, a relatively minor component of LIRM anomalies.

Figure 3. Suspected LIRM anomalies from the Elusive Porcupine site (32RB300), southeastern Montana (Jones, 2003).

DISCUSSION

Recent studies have mapped known lightning-induced anomalies in soils (Verrier and Rochette, 1999, 2002; Sakai et al., 1998). Although these maps are schematic in nature, based on laboratory analysis of collected samples, they are consistent with the models proposed in this article. While these studies are mainly concerned with magnetization directly associated with the point of impact, earlier studies (Cox, 1961; Graham, 1961) documented horizontal paths of lightning conduction in the near surface of bedrock outcrops. Bevan (1995) calculates that the anomaly resulting from a vertical current path would be bipolar with the maxima of the poles in east-west alignment. Vertical current paths have been observed on sandy seashores, and may be expected where more conductive materials occur beneath the surface.

Figure 4. Two suspected linear LIRM anomalies appear at N55/E15 and N23/E83 in this data plot from the Odessa Yates Site (34BV100) in western Oklahoma. These anomalies have no correlates in resistance data and extensive coring revealed no visible anomaly source. The source of the magnetic high at N18/E67 is known to be a Prehistoric pit house (Brosowske et al., 2000).

Several suspected LIRM anomalies appear in Figures 1-5. Three of these data sets (Figures 1, 3, and 4) were collected by the authors on sites on the Great Plains of the United States. The majority of the suspected LIRM anomalies observed by the authors (typically several each year) have occurred in this region. This may represent a regional bias in our work, although the American Plains do experience frequent thunderstorms. Figures 2 and 5 illustrate examples collected by other investigators in the Southwest and Midwest (respectively) of the United States.

Figure 5. Suspected LIRM anomaly from the Great Hopewell Road, Ohio (Simpson and Kvamme, 2001)

These anomalies are either linear or radial in form, the radial anomalies having a number of linear components radiating from a common center. Each linear component is bipolar, with a polarity that is perpendicular to the long axis of the anomaly. As lightning often strikes trees, it is tempting to imagine that anomalies of radial pattern result from the conduction of current by tree roots, although there is no evidence of trees being associated with the examples given here. Other examples showing a simple linear anomaly are presumably the result of the conducted current following a single path of low resistance.

Over 90% of lightning strikes worldwide are negatively charged. The polarity of the lightning discharge is expressed by the configuration of positive and negative components of the LIRM field. IRM associated with a current path will be concentric, as illustrated in Figure 6. The direction of magnetic field follows the "right hand rule." The observed gradient anomaly would be positive where the direction of magnetization adds to the geomagnetic field, and negative where it subtracts. All of the radial anomalies illustrated here (Figures 1, 2, 5, and one of the anomalies in Figure 3) appear to result from negative discharges, with the current flow toward the point of impact. Where isolated linear anomalies occur (Figure 4), the polarity of the discharge may not be apparent if the point of impact is not known.

Figure 6. Model showing the direction of magnetization of earth surrounding a lightning discharge path.

Anomaly strengths will depend on a number of factors that have been studied extensively by lightning researchers, but are of little interest in an archaeological context. Observed magnetic gradient anomalies have ranged from beyond the maximum range of the instrument (+/-2047.5 nT) to a few tenths of a nT, only faintly perceptible in even the quietest data set. We suggest that lightning-induced anomalies may initially display a very high remanence, but may diminish with time. As individual sediment grains are re-oriented by turbation processes, vector subtraction of increasingly randomised magnetic moments will eventually reduce the net remanence to zero. Unlike lightning-induced anomalies, those due to thermal features can remain detectable indefinitely because of enhanced magnetic susceptibility.

CONCLUSION

Although the examples presented here are of North American provenance, Lightning and LIRM occur throughout the world. Presumably LIRM effects may be observed in magnetic surveys elsewhere. It is the purpose of this report to familiarize archaeological geophysicists with the characteristics of LIRM anomalies, so that they may be identified and differentiated from anomalies originating in the archaeological record. Although laboratory testing of soil samples may be required to conclusively identify LIRM as an anomaly source, LIRM may be suspected based on survey and excavation data. The strength and geometry of magnetic anomalies may suggest LIRM; upon excavation, a lack of a visible anomaly source (or the presence of fulgurites) would tend to support this identification. Ongoing lightning research by geophysicists and atmospheric scientists will undoubtedly contribute to a better understanding of the expression of LIRM in magnetic survey data.

ACKNOWLEDGMENTS

The authors would like to thank Kenneth L. Kvamme and Lewis Somers for contributing data sets used as examples, and Pierre Rochette, Violaine Verrier, and Bruce Bevan for their invaluable insights and assistance in developing our understanding of this phenomenon.

REFERENCES

Bevan B.
1995. Magnetic Surveys and Lightning. Near Surface Views (newsletter of the Near Surface Geophysics section of the Society of Exploration Geophysics. October 1995, p.7-8.

Brosowske SD., Drass RR, Maki DL.
2000. The Odessa Yates Project (OYP): Investigation of a Plains Village Period Settlement in the Oklahoma Panhandle. Paper presented at the Joint Midwest Archaeological/Plains Anthropological Conference. St. Paul, Minnesota.

Cox A.
1961. Anomalous Remanent Magnetization of Basalt. U.S. Geological Survey Bulletin 1038-E, p. 131-160.

Dunlop DJ, Schutt LD, Hale CJ.
1984. Paleomagnetism of Archaean rocks from northwestern Ontario: III. Rock Magnetism of the Shelley Lake Granite, Quetico Subprovince. Canadian Journal of Earth Science, Vol. 21, p. 879-886.

Gattacceca J, Rochette P. n.d.
Toward a robust normalized magnetic paleointensity method applied to meteorites. Article in press.

Graham KWT.
1961. The Re-magnetization of a Surface Outcrop by Lightning Currents. Geophys. J. Roy. Astron Soc., 6, p.85-102.

Jones G.
2001. A Geophysical Investigation at 48CA3030: A Prehistoric Archaeological Site in Campbell County, Wyoming. Archaeo-Physics report of investigation # 34. Archaeo-Physics LLC, Minneapolis, Minnesota.

Jones G. 2003.
Geophysical Survey Of Five Archaeological Sites On The Western Energy Company Rosebud Mine Area C. Archaeo-Physics report of investigation # 64. Archaeo-Physics LLC, Minneapolis, Minnesota.

Maki DL.
n.d. Lightning Strikes and Prehistoric Ovens: Determining the source of Magnetic Anomalies Using Techniques of Environmental Magnetism. Article in press.

Munson G.
2002. Excavation of 3030 Winchester 48CA3030. Prepared for Antelope Coal Company Gillette, Wyoming. GCM Services, Inc., Butte Montana.

Sakai HS, Sunada, S, Sakurano H.
1998. Study of Lightning Current by Remanent Magnetization. Electrical Engineering in Japan, Vol. 123, No. 4. p.41-47.

Simpson D, Kvamme KL.
2001. Final Report of Geophysical Investigations at the Great Hopewell Road, Cynthia Street Park, Heath, Ohio. Archeo-Imaging Lab, Department of Anthropology, University of Arkansas, Fayetteville, Arkansas. Submitted to the Ohio Historical Society, Columbus, Ohio.

Somers LE, Wegener RM, Van West CR.
n.d. Foragers and Farmers of the Middle Verde Valley: Data Recovery Along State Route 89A-the Lower Oak Creek Archaeological Project, Volume 2: Analytical Studies compiled and edited by Rein Vanderpot and Carla R. Van West. Forthcoming report. Statistical Research, Tucson Arizona.

Verrier V, Rochette P.
1999 Expertise Magnétique des Matériaux Foudroyés. Paper presented at the 10th Conference of the European Union of Geoscience, 28 March - 1 April 1999, Strasbourg, France.

Verrier V, Rochette P.
2002 Estimating Peak Currents at Ground Lightning Impacts Using Remanent Magnetization. Geophysical Research Letters, Vol. 29, No. 18, p. 14-1-4.