This page will explain how to perform an archaeological geophysics survey using modern magnetometry equipment. It assumes that you have already read the page on Setting Up A Grid. Magnetometry tends to show up a wider range of archaeological features better than resistivity and is a lot faster, but is limited on some geologies and is badly affected by modern metal junk. Resistivity will show up walls better than magnetometry. Magnetometers are more complex pieces of equipment than resistance meters, due to the sensitivity needed for them to perform well, so they tend to be a lot more expensive.

The equipment dealt with on this page is the fluxgate gradiometer, which is the type of magnetometer used in most archaeological surveys. The older proton precession magnetometers are too slow to be useful, and the alkaline vapour magnetometers are too expensive. Fluxgate magnetometers are fast and (relatively) cheap. A fluxgate sensor consists of wire wound around a metal ring. Passing a current through the wire will give different results based on the strength, and direction of the magnetic field. Because the sensors are directional, in fact they can be used to make an electronic compass, they are used in what is known as a gradiometer setup. This involves two sensors, one some distance vertically above the other. The top sensor is less affected by what is under the grounds, so it can be used to cancel out the effects of the Earth's magnetic field. This requires the two sensors to be perfectly aligned, which is the downside compared to sensors that measure the total field strength, like alkaline vapour magnetometers. The magnetic field is measured in nano-Teslas (nT), and archaeological features can be very weak magnetically, often a fraction of a nano-Tesla.

Unlike resistivity, magnetometry is not affected by the level of ground water, so surveys can be performed all year round and get consistant results. If the conditions under foot make walking with a steady and unimpeded gait a problem, the results can be affected. By how much is dependent on how well the machine is balanced. Whilst you may be only interested in the archaeology, modern metal can have a negative effect on the results. Metal pipes, fences, bits of broken farm machinery, land drains and rubbish dumped on the survey area will all show up much stronger than any archaeology hidden below.

The underlying geology itself will play a big part in the results that you see. Magnetometers are good at picking up ditches and other cuts in the underlying geology, but some geologies, such as chalk or clay, will not show these cuts unless they are to some extent filled with occupation debris. Other geologies, such as sandstone, give a much better response to magnetometry.

Most importantly, the operator of the device should we wearing no metal. That means, no metal buttons, zips, watches, glasses or belts. Particular attention should be paid to the footwear, which may have metal pins in the heel. Items of clothing can be checked by setting the device to scan mode, and moving them closer to one of the sensors, seeing if the reading changes. The presence of metal is only a problem when taking readings with the machine, or when balancing it at the 'Zero point' (described below).

Because of the directional sensitivity to the Earth's magnetic field, fluxgate gradiometers need to be balanced, so that both sensors in the gradiometer column provide an equal response to the ambient field, whichever direction they are facing, otherwise you could get a different reading at the same spot simply by turning slightly. This balancing should result in a background reading of zero nT. This state is achieved by finding a magnetically quiet spot in the survey area, to use as a 'Zero point'. It is important that this spot is quiet magnetically, as the device will not be able to be balanced correctly if there is a significant magnetic signal in the ground in addition to the Earth's magnetic field. Such a spot is found by wandering around with the device in scanning mode and finding a spot where the reading doesn't change significantly over a small area.

Once such a spot is found, the device can be balanced. This involved using a compass to set up non-magnetic pegs at the four cardinal points, and going through the balancing process used by the individual machine. This process differs by manufacturer, and will be described fully in the manual. Older machines will have a manual balancing process that involves turning knobs to adjust the orientation. This can be difficult, and the resulting setup is not usually very stable, as knocking the machine can have an affect on the balance. More modern machines have fixed sensors, and balance electronically, which is a lot simpler and a lot more stable.

Once the machine is balanced, the zero point should be left in place until the survey is done, as rebalancing the machine during the survey is often needed. Rebalancing may be needed due to the machine being knocked, thermal drift, where the temperature of the sensor changes and gived a different reading, or because the survey is being conducted over several days. It is worthwhile keeping an eye on the readings whilst surveying, which should hover around zero nT. If you survey a line where the readings hover around 1 or 2 nT, for example, then return to the zero point at the end of the grid, and rebalance it.

Whilst it is possible to take individual readings manually with a magnetometer, for a survey of any size, that would be quite slow. Genarally, the method of collection involves the machine taking readings at a constant rate over time. For these readings to be correctly assigned their proper place in the survey results, the machine needs to know where it is. For any given line in a survey grid, the process works like this. Firstly, before the survey is started, the machine is told how big the survey grids are. Both how many lines, and how long the lines are. Whilst resistivity grids tend to be 20x20m in size, magnetometry grids tend to be 30x30m or 40x40m, as you don't have to worry about the twin probe cable. Once the machine is set up to know how big the grids should be, the survey can begin. A button is pressed to start a line, and the machine will start taking readings. It will also beep at a constant rate. Once when the line starts after the button is pressed, and again for each metre along the line you are travelling, until it gets to the expected end of the line give the line size it was told. The operators job is to match the beeps to the real distance travelled along the ground, usually by following metre marks on an adjecant string laid across the grid. Once a line is finished, the next line is started, usually in a zig-zag pattern rather than starting at the beginning of the line again, until the grid is finished. Whilst the machine will only beep once per metre, that is not when it is taking its readings, the beeps are purely for your benefit. Magnetometers can take several readings per metre, with four being pretty standard, though this is usually adjustable, so in between each been it will be taking a number of readings. For most surveys, the lines are 1 metre apart. You need to remember to take readings in the middle of the line, so for 1 metre spaced lines, the sensor column should half a metre (horizontally) from the string as it travels along it.

It would be nice, if in the course of a grid, no obstacles were ever encountered, but in the real world, there are things like trees and funny shaped partial grids. Readings can be stopped at any time along the length of a line. You can enter dummy readings, just like with resistivity. You can fill the rest of the line with dummy readings if you have come to a fence, or you can enter a few to get past a tree, and then resume the survey. When entering just a few readings and resuming the survey of a line, you should remember how many readings per metre the machine is set to record. For example, if the machine is recording 4 readings per metre and you want to move on 3 metres to get past a tree, you need to enter 12 dummy readings.