(C.J. Hodge)

(R.D. Dubois)

(A. Holt)

(C.J. Hodge)



What is Magnetometry? | Data Collection and Processing | Interpretation | References Cited

What is Magnetometry?

Magnetometry is the study of the earth's magnetic field and how it interacts with local magnetic surroundings. In archaeology this method is used to detect archaeological sites in a non-invasive manner. The magnetometer used is a Cesium vapor magnetometer that works by analyzing the effect of ambient magnetic fields on the Cesium atoms (Breiner 1973). Basically the Magnetometer measures magnets and how strong they are. Magnets have both a north pole and a south pole, and coming from these poles is the magnet's field. The field is usually pictured as a set of lines that come out of the south pole and circle around into the north pole (Figure 1). These lines represent vectors of force. They show the direction that a tiny magnet would line up with if it was placed in the magnet's field. You can see these lines if you place iron filings around an ordinary bar magnet.

The earth itself acts like a huge magnet. Its magnetic field varies in intensity and in the direction that it makes other magnets line up with it. It appears that its strength was around 52,000 nano-tesla in New Bedford according to our readings. Nano-tesla are the measurements used for the strength of the magnet. The earth's field affects magnets at a 70 to 75 degree angle in New Bedford. The earth's magnetic field interacts with local features in a number of ways. On the most basic level it induces magnetization in objects, making them into magnets themselves. Some kinds of objects are more susceptible to magnetization than others. Metals in particular can be made into very strong magnets that are picked up easily by the magnetometer. An object's magnetic field will line up with the angle of the earth's magnetic field, which as I have said is around 70 to 75 degrees (Figure 1).

Figure 1. Visual depiction of magnetic field as it exists around an object in the ground and as recorded by a magnetometer.

Some objects already have their own magnetic field, so the earth's magnetic field adds to it. Examples of real magnets that aren't made by the earth's field are heated clays. The atoms in clay, when it is heated to a high temperature are free to line up with the magnetic field of the earth. When the clay cools they can't move anymore and are frozen all pointing the same direction creating a magnet. This makes it possible to pick up, using the magnetometer, pottery, brick, pottery kilns, and tiles. This is very useful to archaeologists since all of these things were used by many ancient cultures.

Magnetometer readings can be affected by electromagnetic interference in the survey environment. Slight electromagnetic fluctuations may occur naturally and can reach, during a magnetic storm, hundreds of nano-tesla. However, because the survey only involved a 15 to 20 minute sweep of the area with the magnetometer, I was not worried about such fluctuations since the intensity of the earth's magnetic field usually stays fairly stable for such a short period of time. Archaeological remains themselves can generate magnetic fields that may interfere with adjacent readings. Additionally, magnetometry data can be affected by modern cultural features, especially electric currents running through phone and power lines.

Data Collection and Processing

We were invited to New Bedford to survey the backyard of the Polly and Nathan Johnson house. Since the backyard was "L" shaped we set up two rectangles, called grids, that we planned to survey with the magnetometer. We covered this area by walking in lines a meter apart going from south to north. In Grid B however, we were not able to cover as much area as the radar since the magnetometer was too long to fit right up against the wooden fence. For this reason the magnetometer grid starts one meter to the north of the wooden fence (Figure 2). Since induced magnets are oriented north/south this would allow us to pick up the whole magnet, not just one pole or the other (Figure 3).

Figure 2. Plan of the Johnson back yard showing general measurements, testing grids, and notable features (©2000 cjhodge).

Figure 3. Idealized 2D contour map of a whole magnet oriented north, showing positive pole (north) and negative pole (south).

The magnetometer we used is made by Geometrics™. It is equipped with two Cesium vapor sensors, typically placed one on top of the other, so that the analyst can see the variation between measurements at different heights. Each Cesium atom acts like a little magnet, just like the iron filings around the bar magnet. They line up with the earth's magnetic field, or if there is a local disturbance like an archaeological feature it will line up a little differently showing you approximately where that feature is and how strong its magnetic field is. Before actually starting I had to take off any metal I was wearing because even a very small piece of metal, like an earring, can effect the field measurements. For the same reason the magnetometer is made primarily out of plastic, or aluminum, that does not conduct electricity, and so cannot become an induced magnet. During the survey I had to keep the sensors a steady distance from the ground and walk along the lines slowly so that the machine could get enough readings. Finally we downloaded the measurements from the memory of the magnetometer to a laptop computer where we could display it in the 2D contour maps that I have included in this report. These maps were generated using Mag Mapper™ software. North is always up on these images.


The measurements we collected were difficult to analyze because there were so many modern disturbances of the magnetic field. There were power lines running over the yard, and the yard was fenced in on all sides. Since the fencing was either made with metal or metal nails, both of which make very good induced magnets, they show up well in the magnetometer's measurements. When we mapped our measurements we saw a number of very large anomalies. We could see a strong high, which looks like a red and orange rectangle for sensor 1 (the top sensor) at the top of Grid B (Figure 4). For sensor 2 of Grid B we see a low area, which looks like a blue patch at the top of the map (Figure 5).

Figure 4. Grid B 2D contour map of readings from sensor 1 only.

Figure 5. Grid B 2D contour map of readings from sensor 2 only.

Both of these run parallel next to a metal fence, and so we concluded that they showed the magnetic field for this fence. You can also see a significant anomaly, or unusual reading, next to the house on the right of Grid A and next to the fence in Grid A in the upper left-hand corner (Figures 6 and 7). The single most damaging modern feature probably were the power lines overhead, since a current of electricity was running through them and electric currents can induce magnets similar to the way the earth's field does. These wires are probably responsible for inducing such large magnets as we see here. One reason why this seems likely is that the higher of the two sensors showed more anomalous measurements, which would only make sense if the wires running above it were affecting it.

Figure 6. Grid A 2D contour map of readings from sensor 1 only.

Figure 7. Grid A 2D contour map of readings from sensor 2 only.

As far as we can tell no natural anomaly could be as large as most of the anomalies in my data set. All of these anomalies represent modern additions to the landscape. Since we were looking for archaeological features, they did not help us. In fact these modern anomalies were so big they probably covered any smaller archaeological anomaly. Using the gradiometer function of the magnetometer that subtracts the readings of one sensor from another, we can see that the differences in readings between the two sensors were often over several thousand nano-tesla in strength (Figures 8 and 9).

Figure 8. Grid A 2D contour map of pseudo gradient, sensor 1 readings subtracted from sensor 2 readings.

Figure 9. Grid B 2D contour map of pseudo gradient, sensor 1 readings subtracted from sensor 2 readings.

This would obscure any archeological features since even several pounds of metal, the equivalent of an anvil, would only make an anomaly of a few nano-tesla. When analyzing the data I tried to mathematically filter out the unusually large anomalies in hopes to see some more archaeological features, but the data was already so flawed that the filters only confused the picture more (Figure 10). The other figures in this report are therefor products of raw unfiltered data.

Figure 10. Grid B 2D contour map of pseudo gradient using data filtered to remove high and low anomalous readings, sensor 1 readings subtracted from sensor 2.

Another problem that we had with our survey was that the area surveyed was so small that no single magnet was visible. An entire single magnet will look approximately like Figure 3. We were only able to see parts of the magnetic fields of various objects. This is unfortunate because if the entire magnetic field of an object can be read then depth estimates can be made. One interesting thing we noted was that the fence in Grid B appeared as a positive anomaly for sensor 1 and a negative anomaly for sensor 2. At first we did not understand why the polarization of the magnetic field would change so much between the lower sensor, about a foot of the ground, and the upper sensor, about three feet of the ground. Then we realized that it was because the metal fence was a vertical object. The higher sensor was placed above the fence at its magnetic north pole, placing that sensor in a very strong positive field whereas the lower sensor was right in the middle of the field putting it in a strong negative field. This is expressed best by Figure 11. It shows that the higher sensor is at a place where the earth's magnetic field is running in the same direction to the field of the magnet making it positive, and the lower sensor is at a place where the magnet's field runs opposite to the earth's field and subtracts from the earth's field. This makes the lower sensor read a negative anomaly.

Figure 11. Visual depiction of the relationship between a vertical object's magnetic field and the two magnetometer sensors. The upper sensor reads strongly positive while the lower reads strongly negative.

The one possible archaeological feature is a small anomaly in the lower right hand corner of the map of Grid A, sensor 2. This anomaly also shows up in both the ground-penetrating radar and conductivity, but it does not correspond to any modern object. Because of this we think that it represents an archaeological feature, perhaps the remains of one of the possible outbuildings for the house.

This case study shows that, while magnetometry can be a useful archaeological tool, it is not the best form of remote sensing for a small urban setting where modern metallic objects are abundant. Unlike the magnetometer and the conductivity meter, metals do not disproportionately affect data obtained from ground penetrating radar units. Radar is therefore the most promising means of remote sensing at the Johnson site or similar sites.

References Cited

Breiner, S.
1973 Applications Manual for Portable Magnetometers. San Jose: Geometrics, Inc.

[images and text ©2001 aholt unless otherwise noted]