Environmental noise, generally high in urban and low in rural areas, governs the frequency selection particularly at the lower end of the GEM spectrum. The GEM data stream includes the amplitude of a specified powerline frequency (60 Hz or 50 Hz) in a unit of milli-Gauss. To minimize noise induced by powerline harmonics, the GEM-2 uses frequencies that are "odd multiples of half the powerline frequency." In 60-Hz areas (U.S., Canada), these are 90, 150, 210, 270 Hz, and so on. In a 50-Hz areas (Europe, Japan, Australia), these are 75, 125, 175, 225Hz, and so on. These frequencies have minimal powerline interferences. In fact, the concept of "base period" of operation (1/30 second at 60Hz areas, 1/25 second in 50-Hz areas; for more, read the "How it works" page) is based on minimizing the powerline noise problem.
As a general guide, first select a frequency below 1 kHz., say 750 Hz, assuming that you are in a 60-Hz area. For the rest, multiply an odd number, say 5, each time: 750Hz, 3,750Hz, and 18,750Hz. Try not to use more than 3-4 frequencies at a time because the total transmitter energy is shared equally among all frequencies and, therefore, you may encounter reduced signal-to-noise ratio as you add more frequencies.
For specialists, the GEM-2 allows a "passive mode" in which the unit, with the transmitter turned off, collects a "time-series" over one "base period" (1/30 second at 60Hz areas, 1/25 second in 50-Hz areas). The time-series, consisted of several thousand points, can be Fourier-transformed to obtain a precise environmental noise spectrum of the survey area, which may guide the user to select operating frequencies only in spectrally quiet bands.
This is a very complex question because the answer depends on many factors, particularly on ground conductivity, target volume, and ambient electromagnetic noise. Based on many analyses and field data, we estimate the GEM-2 should be able to see about 20-30 m in resistive areas (>1000ohm-m) and about 10-20 m in conductive areas (<100 ohm-m). This figure assumes an ambient noise level of 5 ppm. The noise level is generally high in urban areas and low in rural areas. For typical applications, we do not recommend the GEM-2 for depths deeper than 30 m. For more discussion on this subject, consult the “Skin-depth Nomogram” in “GEM-2 Principle of Operation” on our webpage.
Depth of exploration for a given earth medium is determined by the operating frequency. Therefore, measuring the earth response at multiple frequencies is equivalent to measuring the earth response from multiple depths. Hence, such data can be used to image a 3-D distribution of subsurface objects. Results from several environmental sites indicate that the multi-frequency data from GEM-2 is far superior in characterizing buried, metallic and non-metallic targets to data from conventional single-frequency sensors. For handheld sensors, the transmitter-receiver coil separation is inherently small and has little to do with the depth of exploration.
For detailed technical discussions, please read an article in GEOPHYSICS by Huang, Depth of investigation for small broadband electromagnetic sensors, Geophysics, v. 70, n. 6, pp. G135-G142, 2005. Very simply speaking, the "practical" depth of exploration for a GEM-2 type sensor is approximately the square-root of the skin depth expressed in meter.
Most simple GEM surveys (geological, geotechnical, and environmental) do not require navigational aids, such as GPS. The operator steadily walks across a survey area between two opposite rectangular boundaries in a zigzag fashion at a predetermined line spacing. We call it a “dead-reckoning” survey. Please see the manual or visit our website for graphic description.
For dead-reckoning data, WinGEM assigns coordinates to each data point and generates Excel (csv) files that contains X, Y, I1, Q1, I2, Q2, etc.,where X and Y are the data coordinates and the rest are the inphase and quadrature data for frequency 1, 2, 3, and so forth. To make 2-D maps, the user must provide other (commercial) software such as SURFER, Geosoft, etc. SURFER is the most commonly available 2-D plotter. For such a dead-reckoning survey, the operator can view the whole data on contoured maps within minutes after the survey.
How Do You Merge GEM and GPS Data?
To integrate GPS or other data, we use a PC that has two ports or more. Many PCs these days have two PCMCIA slots that can take two ports each, providing up to four ports to merge into a single data stream. All commercial GPS units (Trimble, Novatel, etc.) have standard data output that can be merged with the GEM output in this way. For a long survey, the WinGEM software allows "remote mode" where the GEM data go directly to the PC, bypassing the internal data logger. If necessary, order an extra long download cable (typically 20-ft long) that can connect the ski to the remote PC from the ski platform to a PC in a car, for instance. For merging the GEM and GPS data, most users opt to write their own code for a particular PC environment, which writes all channels into a single data stream. One also can use commercial software such as Geometrics’ Maglog NT that does the merging.
The GEM-2 is designed to minimize the temporal drift in a quite different way from all other instruments. The GEM-2 ski contains three coils that are precisely maintained in their relative separations amongst each other. Any small changes in the relative separations can cause a shift (or drift) in the signal level. The GEM-2 coils are permanently entombed in a “ski” structure made of synthetic materials that has a low thermal expansion coefficient and, therefore, their relative locations are firmly fixed. The sensor is designed to “linearly” expand or contract following the ambient temperature. This linear expansion precisely maintains the relative positions and, therefore, the bucking condition. This feature is quite unique for this design. Owing to this careful design and manufacturing, the GEM-2 has not shown any appreciable drift, as would be testified by numerous users in the last five years of its usage.
The other important feature is the way the console (which is a big chunk of metal) is designed and mounted on the ski. The location of the console is where the field gradient is minimum so that its slight displacement (due to not tightening the mounting screw, for instance) would cause little shift in the signal. There are no moving parts in the console; for instance, the battery is internal and not removable. If you think the sensor maybe drifting, tighten the screws that mount the console to the ski; please do not over tighten. Diurnal temperature changes should not be a significant cause.
How Do You Plot the Data on a Rough Terrain?
The GEM as a handheld unit can go wherever the operator can go. For data-plotting purposes, if the surface grade isn't steep or the slope is more or less constant, the interpreted results may be vertically shifted in parallel to the surface. For unusual surfaces (e.g., across a cliff), the user must come up with his or her own codes. Note that the primary purpose of the GEM is to find anomalies first, and then worry about what they mean. Most of the time, the answers are obvious to the people who are familiar with the site.
How Do You Convert the GEM-2 Data to Apparent Conductivity and Susceptibility?
This is a standard feature for the GEM-2 in the “locate” menu in WinGEM operating system. On WinGEM, the GEM-2 ppm data can, at a mouse stroke, be converted to apparent conductivity at each frequency and susceptibility at the lowest frequency used for the survey. For theory and practice, see the article by Huang and Won (2001), entitled “Conductivity and Susceptibility Mapping Using Broadband Electromagnetic Sensors,” published in Journal of Environmental and Engineering Geophysics. The article may be downloaded from the Geophex website.
How Do You Calibrate the GEM-2?
There are two complex (i.e., inphase and quadrature) calibration sets, each as a function of frequency. Since they are done at Geophex and stored in your GEM software, you shouldn't need any further calibration. The two calibration sets are:
Amplitude calibration - this is done using a "Q coil" with known radius, number of turns, resistance, and inductance. It mainly sets the amplitude scale.
Free-air calibration - The sensor output must approach zero when you move it away from any conductor, which is of course hard to do on earth. For airborne sensors, it is done by flying the sensor high. The needed height is typically 5-10 times of the coil spacing. For GEM-2, we raise the sensor to about 6-10m - we pull it up to a pine tree in our backyard - and we call the sensor response there the "free-air values" that are also stored in the sensor software. As you notice, this calibration does depend on the ground conductivity, but is simply a DC offset. A ferrite rod - permeable but not conductive - cannot be used for either of the two calibrations described above. We sometimes use it, but not necessary, for a quick calibration check since it produces only the inphase response. Theoretically, a ferrite rod produces a constant inphase and zero quadrature at all frequencies. Our older manual described it for users to try, but we deleted it since we decided that it is not necessary.
One can buy ferrite rods locally at any "radio shacks" where they sell electronic components, since it is commonly used in circuits to enhance coil inductance. Most radios have them in their antenna tuning coils. Usually, they are about 1cm or less in diameter and 2-3cm long. The size doesn't matter since all you want to check is that its inphase response is constant over the bandwidth.
Both calibration factors are either a constant multiplier - for amplitudes - or a DC offset - for the free air. In other words, they do not affect the appearance of your data in a map or profile. The "bumps" are always there and the calibration affects only the scale and offset. We suggest that you do not change the amplitude calibration.
The offset calibration can be a problem over a very conductive area, where one wants an absolute conductivity map. In this case, if you know the background conductivity from other measurements (DC resistivity, for instance), you can simply add or subtract a constant from the entire dataset so that it fits the background. Still, the process does not change the map appearance.
The program is available for the time being as Beta Version free of charge to all GEM owners. It is downloadable from our website. We request users to work with our Geophex scientists to improve it. The primary author is Haoping Huang at Geophex. The code makes a 2-D cross-section based on continuous 1-D interpretation. Also there are several commercial software packages that invert the GEM data into a multi-layered model.
Geophex website, www.geophex.com, provides ample information about the GEM sensors, design and operating principles, survey tips, manuals, and down-loadable operating software. The site also carries many, many journal articles related to the GEM sensors, which can be downloaded or printed. For more information and questions, please contact Geophex.