Exploration & geophysics

SpecificationUSGS Earth MRI practice on a 2022 western Arkansas block: north-south flight lines at 300 m spacing, east-west tie lines every 3,000 m, median terrain clearance 170 m above topography, 31,006 line-km flown over a 91.55 x 92.15 km area, magnetometer in a tail stinger. A USGS fixed-wing contract specification for a Colorado survey (USGS Open-File Report 2006-1327, appendix A) called for 400 m traverse spacing on a 090/270 true heading, 3,200 m tie lines, an optically pumped caesium-vapour sensor sampling at 20 Hz with 0.001 nT resolution and 0.005 nT sensitivity, and survey noise not to exceed +/-0.1 nT over 1,000 m. Ground magnetometers reach 0.1 nT sensitivity and optically pumped sensors cycle in as little as 0.04 s (EPA, archived Environmental Geophysics guidance on magnetic methods).

Where it falls downMagnetics sees only lateral susceptibility contrast, not vertical contrast, and the anomaly-source problem is formally ambiguous — an infinite number of source bodies can produce a given anomaly. The measurement responds only to the field component along the Earth's main field, so anomalous fields roughly perpendicular to it are undetectable. Non-magnetic ore (most sediment-hosted zinc, many gold systems, bauxite) produces no direct response. False anomalies come from pipelines, powerlines, railroads, vehicles, buried steel, near-surface magnetite in soil and boulders, and from magnetic storms — USGS-contracted practice is to stop surveying when diurnal variation exceeds 5 nT in any five-minute period (USGS Open-File Report 2006-1327, appendix A).

SpecificationUSGS describes gravity as usable at 'all' depths in principle (Open-File Report 95-831, table 1). EPA guidance, from EPA's archived Environmental Geophysics pages, which are no longer maintained and which the current EPA gravity-method page does not reproduce: modern gravimeters have a least-significant scale reading of 0.01 micrometre per second squared, an irreducible reading error around 0.05, and electronic meters repeat to 0.02-0.03; under adverse field conditions expect 0.08-0.1. Station spacing for detailed work is typically 1 to 10 m, with 1-3 m intervals needed to map a 10 m-wide anomaly. Horizontal position must be good to 50-100 cm or 10 percent of station spacing, whichever is smaller, and elevation to +/-1 to 2 cm. Corrections applied in order: instrument drift (from base-station reoccupation), Earth tide, latitude (International Gravity Formula), free-air, Bouguer slab, and terrain.

Where it falls downResolution is geometry-limited: 'a body must be almost as big as it is deep' — a depth-to-diameter ratio greater than about 2.0 falls below the practical noise level, so a 10 m sphere is detectable only to roughly 20 m. USGS notes small shallow features such as underground workings are hard to detect unless very shallow. The inversion is ambiguous — a shallow small mass mimics a deep large one — so density contrast or geometry must be constrained externally. Elevation error of 1-1.5 cm alone injects 30-40 nm/s2 of error; a +/-10 percent uncertainty in near-surface density injects about +/-60 nm/s2 for every 1 m of topographic variation (the archived EPA page also prints a figure of +/-0.6 mm/s2 for a 10 m site, which is inconsistent with its own per-metre value by a factor of 1,000 and is not used here). Regional-residual separation is cited as one of the most critical and most subjective processing steps.

SpecificationFrequency-domain helicopter system flown by USGS in Alaska: six coil pairs from about 400 Hz to 140,000 Hz (378, 1,843, 3,260, 8,180, 40,650 and 128,510 Hz), bird flown at a nominal 30 m above ground and suspended about 30 m below the helicopter, measurements roughly every 3 m along line, flight-line spacing 350 m. Depth of investigation ranges from a few metres at the highest frequency to on the order of 80 m at 400 Hz. Time-domain systems go deeper: a USGS fixed-wing time-domain survey of the Mississippi Alluvial Plain flew 6 km-spaced lines at 120 m nominal terrain clearance for 24,030 line-km and maps structure to depths up to 300 m depending on subsurface resistivity; a helicopter time-domain survey in northeast Wisconsin flew 800 m line spacing at 30-40 m above terrain for 3,170 line-km, giving resistivity models with a typical depth of investigation of about 300 m and 1-2 m near-surface resolution.

Where it falls downAirborne EM screens large areas but USGS notes it always requires ground follow-up. It does not work well in very resistive ground, and a thick conductive overburden (clay, saline water, conductive regolith) shields anything beneath it. Late-time signal decays roughly as t to the power -5/2 — a factor of about 300 for a factor of 10 increase in time — so signal-to-noise collapses at late times and therefore at depth. False conductors come from graphite, serpentinite, clay, saline groundwater, powerlines (which give an oscillating rather than decaying response), pipelines and fences. Induced-polarisation effects in clay soils can suppress or even reverse the sign of the transient. 3D interpretation material is still described as under development.

SpecificationFigures below are from EPA's archived Environmental Geophysics guidance, which is no longer maintained. Transmitter is usually a single-turn square loop whose side length is approximately the desired exploration depth — for shallow work under 40 m in resistive ground, loops as small as 5 to 10 m a side are used, and a 10 m by 10 m loop is quoted for a 20 m investigation. Receivers typically use about 20 narrow time gates opening sequentially, spanning from a few microseconds to tens or even hundreds of milliseconds depending on target depth. Signal amplitude falls from many thousands of nV/m2 at early times to less than 0.1 nV/m2 at late times. Measured voltage varies as conductivity to the power 3/2, making TDEM intrinsically more sensitive to small conductivity variations than DC resistivity. Lateral resolution is better than DC methods because the array width is much less than the exploration depth, whereas a DC array is typically about three times the exploration depth. A sounding can be completed in a matter of minutes.

Where it falls downDoes not work well in very resistive material; there is no galvanic contact problem, but there is also little signal. Local thunderstorm activity and 50/60 Hz powerline fields produce large signals that can overload the receiver. Buried pipes, fences and powerlines generate spurious responses. Induced polarisation in clayey soils can reduce the transient amplitude or drive it negative. Equipment is more complex and costly than DC resistivity, and 3D interpretation remains limited.

SpecificationUSGS gives a depth of investigation of about 2 km for both DC resistivity and IP, and states IP can detect as little as 0.5 volume percent total metallic-lustre sulphide minerals. Units: chargeability from the area under the decay curve (reported in mV/V, or as mV-s from the raw integration); percent frequency effect as the normalised difference in resistivity measured at two frequencies, typically 0.1 Hz and 10 Hz; phase in milliradians. Spectral IP work is done at frequencies below 1,000 Hz, and below 100 Hz where metallic sources dominate. Dipole-dipole is used almost exclusively for profiling; Schlumberger, Wenner and pole-dipole arrays are also used, described by an electrode spacing 'a' and a spacing multiplier 'n'. Typical practice, not a published standard: mineral-exploration dipole-dipole IP is run with a-spacings of roughly 25-100 m and n up to 6-10 to reach a few hundred metres of depth.

Where it falls downUSGS calls IP inefficient as an areal reconnaissance tool and notes it works best where the target location is already approximately known. Bulk resistivity is controlled by pore-water content and dissolved ions rather than by metallic minerals, so a resistivity low is not by itself evidence of sulphide. Classic IP false positives are graphite, clay (montmorillonite gives membrane polarisation through ion exchange), magnetite, pyrrhotite and barren pyrite — IP cannot distinguish economic sulphide from barren sulphide. Electromagnetic coupling between current and potential wires corrupts readings at large separations. Anomalies from local sulphide oxidation may be low-amplitude and hard to resolve. Resolution falls off sharply with depth, and 2D inversion assumes no variation perpendicular to the line.

SpecificationUSGS gives a depth-of-investigation hierarchy for electromagnetic methods: about 10 m for VLF, about 100 m for controlled-source methods, intermediate depth of about 1 km for audiomagnetotellurics (AMT), and deep — about 10 km — for magnetotellurics (MT). The physical control is skin depth, which increases as the square root of resistivity and of period, so lower frequency means greater depth. Measured quantity is apparent resistivity in ohm-metres plus impedance phase.

Where it falls downNatural-source amplitude is weak and unevenly distributed in frequency — the 'AMT dead band' around 1-5 kHz is a well-known signal minimum. Cultural noise from powerlines, electric fences, railways, pipelines and cathodic protection can render whole survey areas unusable. Resolution is inherently diffusive and degrades rapidly with depth; MT cannot resolve thin resistive layers and suffers static shift from near-surface heterogeneity. Like all EM, it maps conductivity, not metal, so graphite and saline fluid look like mineralisation. Station spacing and recording duration are survey-specific and are not given here.

SpecificationUSGS-flown systems use sodium iodide detectors with large crystal volumes of 33 to 50 litres. Nominal survey elevation 122 m above ground; detailed surveys flown on east-west lines at about 400 m spacing, regional surveys at about 5 km spacing. Data are corrected by the contractor for cosmic and aircraft background, Compton scattering (stripping), airborne bismuth-214 and altitude variation. Calibration used fixed concrete pad facilities and a calibration test strip. Results are reported as percent K and as equivalent uranium (eU) and equivalent thorium (eTh) in ppm because radioactive equilibrium cannot be assumed. Depth of investigation is extremely shallow: USGS gives the upper 50 cm (Open-File Report 95-831, table 1).

Where it falls downIt is a surface method only — anything under a metre of transported cover, water, snow or dense vegetation is invisible. Soil moisture attenuates the signal, and measurements made after rain are suppressed. Because uranium is measured through its bismuth-214 daughter, radon escape and disequilibrium make eU an unreliable proxy for uranium grade; thorium is the most immobile of the three radioelements and is the most reliable lithological mapper. Anomalies over ploughed fields, fertiliser application, mine waste and building aggregate are common false positives.

SpecificationUSGS lists seismic as usable at 'all' depths in principle (Open-File Report 95-831, table 1); the practical figures that follow are from the ITRC surface-geophysics guidance cited. Practical figures: refraction is generally limited to mapping features at depths less than about 100 ft (30 m), with an array length approximately four to five times the depth to target, at least seven shots per spread and often 13 shots per spread for high-resolution work. Reflection can define sequential stratigraphy to great depth (greater than 1,000 m / 3,281 ft) but is difficult shallower than about 50 ft because ground roll and air blast swamp the reflections. Both can resolve layers down to about 1 m (3 ft) thickness under optimal conditions. Production is slow: a three-person reflection crew may manage about 400 shots in a day, equating to roughly 900 ft (274 m) of line, the cited source printing 300 m, which is a bad conversion. Multichannel analysis of surface waves, a related technique, gives shear-wave velocity to typically less than 150 ft, with receiver spacing set between 30 percent of the investigation depth and the investigation depth itself.

Where it falls downRefraction fails outright in a velocity inversion — where a higher-velocity layer such as stiff clay overlies a lower-velocity sand or gravel, it returns an incorrect answer rather than no answer. A thin layer can also be missed entirely (the hidden-layer problem). Lateral resolution degrades as array length grows. USGS (Open-File Report 95-831) singles out the relatively high cost and the difficulty of acquiring and interpreting seismic data in strongly faulted and altered igneous terranes — exactly the setting of most hardrock ore bodies, where impedance contrasts are weak and scattering is severe. Reflection requires expert processing and is expensive.

SpecificationDepth of investigation is highly variable and frequency-dependent: typically 100 ft or less. A 100 MHz antenna can penetrate up to 60 ft in clean, dry sand; a 1,000 MHz antenna penetrates only about 6 ft. Resolution trades against depth in the same way — a 100 MHz antenna can identify a 2 ft pipe at 20 ft, while a 1,000 MHz antenna can resolve shallow wire mesh and a 3/16-inch hose at 6 ft. Survey design normally uses two perpendicular transect sets.

Where it falls downConductivity is the killer. Signal penetration is best in dry, sandy or rocky soils and worst in moist, clayey and conductive soils; in conductive material attenuation can limit penetration to 3 ft (1 m) or less. The same point applies in minerals work — GPR signals are attenuated by wet or clay soil and the method is inefficient wherever those conditions exist in the surficial layer. That rules GPR out over most weathered, clay-rich or saline terrain, and over sulphide-bearing ground with conductive alteration haloes.

SpecificationUSGS national soil survey design (Scientific Investigations Report 2017-5118): one site per 1,600 square kilometres, 4,857 sites across the conterminous United States, three samples per site (0-5 cm depth, a composite of the A horizon, and the C horizon or 80-100 cm where the C horizon was deeper than 1 m), the less-than-2 mm fraction analysed for a suite of 45 major and trace elements, with quantitative X-ray diffraction using Rietveld refinement on the A and C horizons. No stream-sediment sampling protocol is given here: USGS stream-sediment and heavy-mineral-concentrate procedures (size fraction, sieve mesh, heavy-liquid separation, analytical suite) vary from survey to survey and should be taken from the report for the specific survey being used.

Where it falls downStream sediment only works where there is active drainage and clastic transport — it fails in glaciated terrain with disrupted drainage, in arid basins with no through-going streams, in karst and in areas of thick transported cover. A catchment anomaly tells you the metal is somewhere upstream, not where. Soil geochemistry is defeated by transported overburden, by leaching in tropical profiles, and by hydromorphic dispersion that moves an anomaly off the source. Contamination is a routine false positive: mine waste, road metal, fertiliser, smelter fallout, treated timber and shotgun pellets. Nugget effect in gold makes single low-density samples unrepeatable, and analytical noise near the detection limit generates spurious highs.

SpecificationUSGS greenhouse evaluation of big sagebrush (Artemisia tridentata subsp. tridentata) as a medium for gold exploration in the Great Basin combined leaves, twigs and stems and analysed them by instrumental neutron activation analysis. Plants grown on ore-bearing soil did not accumulate significantly more gold than controls, but did accumulate significantly more arsenic and antimony. The conclusion was that sagebrush would be a good prospecting medium for detecting concealed Carlin-type deposits if arsenic and antimony are used as the pathfinder elements. The work was a 1995 greenhouse evaluation (Stewart and McKown), not a field survey, so it establishes what sagebrush takes up rather than a field sampling protocol.

Where it falls downUptake is species-specific, organ-specific and season-specific, so sampling must be rigidly standardised or the results are not comparable. The target element may simply not be taken up — the USGS study found gold itself was not concentrated. Airborne (aeolian) dust contamination is the dominant false positive in arid country, and the USGS study states directly that true gold anomalies in sagebrush will be harder to separate from aeolian contamination than arsenic and antimony anomalies. Rooting depth is rarely known, so the depth being sampled is uncertain. Ashing introduces its own variability and volatile losses.

SpecificationUSGS (Geology, Geophysics and Geochemistry Science Center spectroscopy programme) has produced regional mineral maps covering over 380,000 square kilometres in California and Nevada at 15 m spatial resolution, and works at three scales: laboratory spectra of hand specimens and drill core, ground scanning of outcrops and mine walls, and airborne imaging. The NASA airborne imaging spectrometer used for much of this work, AVIRIS-Classic, records 224 contiguous spectral channels with a 10 nm nominal channel bandwidth calibrated to within 1 nm, and a 1 milliradian instantaneous field of view calibrated to within 0.1 mrad, using silicon detectors in the visible, InGaAs in the near-infrared and InSb in the shortwave infrared.

Where it falls downIt sees only what light touches. USGS classifies remote sensing as 'surface only' — vegetation canopy, lichen, desert varnish, snow, soil crust and transported cover all block the signal. It maps mineralogy, not grade or depth, so an alteration footprint says nothing about whether metal is present below. Atmospheric correction errors, shadowing on steep terrain and mixed pixels at 15 m resolution all generate spurious mineral identifications, and spectrally similar minerals are routinely confused.