Study area: Opalton opal field, Central-West Queensland (142.5–143.0°E, 23.5–24.0°S; active claim ~142.735°E, −23.748°S). Host rock: Winton Formation (Late Cretaceous sandstone/siltstone/mudstone/ironstone). Mining method is excavation (bulldozer/excavator), not detection.
Summary. The good news: the two most important layers — surface geology showing the Winton Formation host, and the mining-tenement/legal overlay — are both live, cover Opalton, and are Queensland Government open data. Geoscience Australia’s national geophysics WMS (magnetics + radiometrics) is verified working over Opalton with a sibling app already using it. AWS terrarium tiles give online terrain. The gaps are honest ones: there is no LiDAR-grade DEM over remote Opalton (SRTM ~30 m is the ceiling), the BoM Geofabric’s old WFS was decommissioned (use the ArcGIS FeatureServer instead), GA’s palaeovalley mapping does not cover Queensland, and cultural heritage data is legally restricted — advisory text only, never ingest.
Every endpoint below was hit during this research. Where a request returned a valid image or feature, it is marked verified. Where I could only confirm existence but not a specific layer name, it says so.
| # | Source | Covers Opalton? | Licence | v0.1-ready? | Notes |
|---|---|---|---|---|---|
| 1 | Winton Fm geology — GSQ GeologyDetailed (1:100k) | ✅ Yes (verified) — returns “Winton Formation, Kw” at the claim | CC-BY 4.0 (© State of QLD, Dept of Resources) | ✅ Yes | ArcGIS REST; query→GeoJSON. Best single layer. |
| 1b | GA Surface Geology 1:1M/2.5M | ✅ Yes | CC-BY 4.0 (© Commonwealth/GA) | ⚠️ Fallback only | Too coarse to resolve one formation; national backstop. |
| 2 | Aeromagnetics — GA magmap_v7_2019_RTP WMS |
✅ Yes (GetMap verified) | CC-BY 4.0 (© Commonwealth/GA) | ✅ Yes | Direct WMS image overlay. Sibling app already uses it. No WCS on this endpoint. |
| 3 | Radiometrics — GA radmap_v4_2019_* WMS |
✅ Yes (K GetMap verified) | CC-BY 4.0 (© Commonwealth/GA) | ✅ Yes | Same WMS. K/Th/U/ternary layer names confirmed verbatim. |
| 4 | DEM/terrain — AWS terrarium tiles | ✅ Yes (tile verified) | CC-BY / ODbL (mixed, SRTM-derived here) | ✅ Yes (online) | Online-only; slope/drainage. SRTM ~30 m. |
| 4b | GA SRTM 1-Second (DEM-H) via ELVIS | ✅ Yes | CC-BY 4.0 | ⚠️ Needs processing | Download+reproject for offline. National 5 m LiDAR does not cover Opalton. |
| 5 | Hydrography — BoM Geofabric v3 (ArcGIS FeatureServer) | ✅ Yes (extent verified) | CC-BY 4.0 (© Commonwealth/BoM) | ⚠️ Needs processing | Old WFS decommissioned Nov 2023; use the FeatureServer. |
| 5b | GA palaeovalley/palaeochannel | ❌ No | CC-BY 4.0 | ❌ No | GA mapping is WA/SA/NT only. Genuine gap for QLD. |
| 6 | Tenements / legal — QLD MinesPermitsCurrent | ✅ Yes (extent verified) | CC-BY 4.0 (© State of QLD) | ✅ Yes | Granular ML/MC/EPM granted layers. THE legal overlay. |
| 6b | QLD Designated Fossicking Land (Opalton) | ✅ Yes (Opalton IS a DFL) | CC-BY 4.0 (assumed — verify) | ⚠️ Find exact layer | Layer exists in QLD Globe/QSpatial; exact WMS name unconfirmed. |
| 7 | Faults — GSQ GeologyDetailed/State faults | ✅ Yes | CC-BY 4.0 (© State of QLD) | ✅ Yes | Vector faults + folds + lineaments. |
| 8 | Cultural heritage — ACHRIS/DATSIP | (restricted) | Restricted — NOT open | ❌ Never ingest | Advisory text only. Duty of care. |
1a. GSQ Detailed Surface Geology (1:100k) — PRIMARY
https://spatial-gis.information.qld.gov.au/arcgis/rest/services/GeoscientificInformation/GeologyDetailed/MapServer
.../MapServer/WMSServer).copyrightText: “© State of Queensland (Department of Natural Resources and Mines, Manufacturing and Regional and Rural Development)”. Attribute: “© State of Queensland (Department of Resources), CC-BY 4.0”.query (bbox over Opalton, outFields=*, f=geojson) → clip to study area → bundle GeoJSON. ~small over one field. Alternatively consume the WMS as an image overlay (loses attributes).1b. Geoscience Australia — Surface Geology of Australia (1:1M / 1:2.5M)
https://services.ga.gov.au/gis/rest/services/GA_Surface_Geology/MapServer
AUS_GA_1M_GUPoly_Lithostratigraphy (1:1M polygons).AUS_GA_2500k_GUPoly_* (1:2.5M).https://services.ga.gov.au/gis/services/GA_Surface_Geology/MapServer/WMSServer"© Commonwealth of Australia (Geoscience Australia) 2016. …Creative Commons Attribution 4.0 International Licence."https://services.ga.gov.au/gis/geophysical-grids/wms
geophys:magmap_v7_2019_RTP (TMI-RTP, 2019 v7). Also present and verified in capabilities: geophys:magmap_v7_2019_TMI, ..._TMI_40m, ..._1VD, and a full VRTP derivative suite (..._VRTP_AS analytic signal, ..._VRTP_1VD, upward-continuation series, etc.).GetMap for geophys:magmap_v7_2019_RTP over bbox -23.9,142.6,-23.6,142.9 (EPSG:4326) returned a valid PNG. National grid.https://services.ga.gov.au/gis/geophysical-grids/wms. Layer names confirmed verbatim from GetCapabilities:
geophys:radmap_v4_2019_filtered_pctkgeophys:radmap_v4_2019_filtered_ppmthgeophys:radmap_v4_2019_filtered_ppmugeophys:radmap_v4_2019_filtered_ternary_imagegeophys:radmap_v4_2019_filtered_dose; ratios ..._ratio_tk, ..._ratio_uk, ..._ratio_ut, ..._ratio_u2t. (Older v3 2015 series also present.)GetMap for geophys:radmap_v4_2019_filtered_pctk over the Opalton bbox returned a valid PNG. National.https://s3.amazonaws.com/elevation-tiles-prod/terrarium/{z}/{x}/{y}.png. Public S3, no auth. Tile over QLD (z10/906/615) verified returns a valid PNG. Registry of Open Data on AWS lists it as current (no deprecation). This is what the sibling app uses.https://elevation.fsdf.org.au/ (clip-and-ship download UI; GA product page: https://www.ga.gov.au/scientific-topics/national-location-information/digital-elevation-data). GA also exposes SRTM through ArcGIS image services, but I could not confirm a specific public WCS URL that resolves — treat SRTM as a download, not a live service.5a. BoM Geofabric (Surface Hydrology)
geofabric.bom.gov.au was DECOMMISSIONED 30 Nov 2023. The live, working route is the ArcGIS FeatureServer (BoM-published, Digital Atlas):
https://services-ap1.arcgis.com/ypkPEy1AmwPKGNNv/arcgis/rest/services/Geofabric_Surface_Hydrology_Network/FeatureServer
Layers include: 4 = AHGF Network Stream — Major, Named; 5 = Network Stream — All; 1 = Catchment; 6/7 = Waterbody; 8 = Water Storages.copyrightText: “© Commonwealth of Australia (Bureau of Meteorology) 2022”./4/query or /5/query, bbox, f=geojson) → clip → bundle GeoJSON. Straightforward.geofabric.bom.gov.au WFS — it’s dead. Use the FeatureServer.5b. Palaeovalley / palaeochannel (GA)
https://www.ga.gov.au/about/projects/water/palaeovalley-groundwater-resources (data on data.gov.au).https://spatial-gis.information.qld.gov.au/arcgis/rest/services/Economy/MinesPermitsCurrent/MapServer — updated nightly. Confirmed layers: 44 ML granted / 45 ML surface granted / 46 ML access granted; 36 MC granted / 37 MC access granted; 3 EPM granted (plus application variants 40–42, 33–34, 2, etc.).https://spatial-gis.information.qld.gov.au/arcgis/rest/services/Economy/MineralTenement/MapServer — single MineralTenement polygon layer (GDA2020) if you want all tenure in one call and filter by type attribute./WMSServer.copyrightText: “© State of Queensland (Department of Natural Resources and Mines, Manufacturing and Regional and Rural Development) 2024” / “© State of Queensland (Department of Resources) 2024”. Attribute: “© State of Queensland, CC-BY 4.0”.6b. QLD Designated Fossicking Land (Opalton)
https://qldspatial.information.qld.gov.au/catalogue/ before wiring it. Licence is almost certainly CC-BY 4.0 (QLD default) but confirm on the catalogue record..../GeoscientificInformation/GeologyDetailed/MapServer — Layer 4 = “Detailed faults and shear zones”, Layer 5 = folds, Layer 8 = lineaments, Layer 6 = dykes/veins/sills, Layer 3 = geological boundaries. All polyline vectors..../GeoscientificInformation/GeologyState/MapServer — Layer 3 = “State Faults”, Layer 4 = “State Folds”, Layer 2 = boundaries.magmap_v7_2019_RTP/_VRTP_1VD (§2) if needed.https://www.culturalheritage.qld.gov.au/achris/ (register look-up + application-for-advice). There is no open WMS/WFS/REST feed of heritage site locations, and there must not be.These four are verified over Opalton, correctly licensed (CC-BY 4.0), and need little or no processing:
GeologyDetailed MapServer, layer 15) — the Winton Formation (Kw) favourability base. Query→GeoJSON, clip to bbox, bundle. Verified: returns “Winton Formation” at the claim.Economy/MinesPermitsCurrent MapServer, ML/MC/EPM granted layers) — the legal + historical-working overlay. Verified: covers Opalton.geophys:magmap_v7_2019_RTP) and GA radiometrics WMS (geophys:radmap_v4_2019_filtered_ternary_image / _pctk) — direct WMS image overlays, ironstone + weathering. Verified: GetMap returns imagery over Opalton.Add GSQ faults (GeologyDetailed layer 4) and the BoM Geofabric FeatureServer streams as easy fifth/sixth layers — both covered, both CC-BY, light processing.
magmap_v7_2019_RTP LatLon bounding box from capabilities (XML was truncated) — but the GetMap over Opalton succeeded, so coverage is empirically confirmed even though I couldn’t read the bbox element.Everything above (§1–8) is passive, open, basin-to-regional data you pull from a government server. This section is different: it’s the active geophysics you (or a contractor) run ON Steven’s 15 ha claim to try to map the thin, shallow (0.3–8 m) ironstone/clay profile before the dozer goes in. The target is thin (ironstone bands are 10s of cm each) and shallow (a few metres) — but the stratigraphic order is what decides which technique wins, and it’s the opposite of what a geophysicist would fear.
The clay is BASAL — it sits UNDER the ironstone target, not over it (per Steven’s ground truth). In most of Steven’s dig areas the profile is: pure sandstone from near-surface → ironstone layers (the target) → basal clay at the very bottom. Only some areas carry a capping clay above the sandstone; the majority don’t. That order flips the usual “clay kills radar” verdict:
Net: GPR is the top pick for typical (uncapped-sandstone) Opalton ground; ERT is best-in-class where a capping clay is present or suspected.
One honest framing up front: no surface geophysics finds opal. Opal itself is a tiny, erratic target with no reliable geophysical signature. What these techniques can do is map the host architecture — the depth to the clay “floor”, the ironstone levels sitting on it, and faults/slumps that pond silica-rich water — so Steven digs the right ground instead of the whole paddock. That’s the realistic prize.
How it works. A transmit antenna fires a short EM pulse into the ground; reflections come back off boundaries where the dielectric constant changes (e.g. dry sandstone → ironstone → clay). Two-way travel time × velocity = depth. You drag/roll it along lines and stack the traces into a vertical “radargram” cross-section — exactly the kind of layer-cake picture Steven wants.
Depth vs resolution trade-off for THIS target. Antenna frequency sets both: low frequency = deep but blurry, high frequency = shallow but sharp. As a rule of thumb in favourable (dry, resistive) ground: ~100–250 MHz reaches roughly 5–10 m with layer resolution of tens of cm; 400+ MHz gives crisp sub-10-cm resolution but often only 1–3 m depth. For a 5–8 m target with 3–4 thin ironstone bands, 200–250 MHz is the sweet spot — if the ground cooperates.
The clay question — and why the basal position saves GPR here. The received wisdom is “clay kills GPR”: wet/conductive clay attenuates the signal at ~50–300+ dB/m and can drop penetration to under 1 m. That’s true — but it only matters if the clay is between the antenna and your target. At Opalton the clay is basal, sitting under the ironstone target. GPR fires through clean resistive sandstone (near-ideal for radar), returns crisp reflections off the ironstone bands (the target), and only then hits the basal clay and dies — by which point it has already imaged the layer-cake Steven wants. The clay is a floor, not a ceiling. The QOMA (Queensland Opal Miners Association) note that GPR “reflects more from the clay interface, which can sometimes mask a boulder level” actually confirms GPR resolves the interface at the base — you read the levels above it and stop at the clay. The one failure mode is the minority stratigraphy where clay CAPS the sandstone above the target — there the pulse dies before reaching the ironstone, and you switch to ERT. Verdict for Opalton: GPR is the top pick for the typical uncapped-sandstone dig area — high resolution at shallow depth, cheap per site, walk-behind or ute-dragged, and it produces the direct 2-D/3-D subsurface image a miner can read. Carry ERT (below) as the answer for capped or unknown-cap ground.
Portability. All three form factors exist: hand-carried/handheld (small shielded antenna, walk it), hand-cart / push-cart (survey wheel, most common for systematic grids), and ute-towed (antenna in a sled dragged behind a vehicle for fast coverage). For a 15 ha claim a push-cart is the practical unit; ute-tow only helps if the ground is smooth enough.
Commercial units (brand + model). GSSI SIR-4000 / SIR-30 controllers with 200 MHz & 350–400 MHz antennas; Sensors & Software pulseEKKO (100/250 MHz) and Noggin carts; Guideline Geo / MALÅ (GX, ProEx, Ground Explorer HDR); IDS GeoRadar / Leica (Opera Duo, dual-frequency). All reputable; antenna frequency choice matters more than brand.
Published use for opal / shallow regolith. Yes — GPR is named repeatedly for Winton/Opalton boulder opal (“opal is found by gamma-ray logging, ground-penetrating radar, magnetic surveying…”), and QOMA describes it detecting “faults, slips and slides and sub-strata layers and sometimes even the boulder levels.” So it’s a real, used method here — just an unreliable one because of clay.
Cost (AUD, ballpark). DIY rental of a GPR system in Australia is roughly $300–$800/day plus bond (unverified — varies by hire house and antenna set). Contractor GPR is commonly quoted at ~$150–$450/hr or ~$1,200–$3,600/day including a basic report (Australian market figures; treat as indicative, not a quote). Buying a mid-range dual-frequency system new is well into tens of thousands of AUD — not sensible for a single claim.
Steven recalls Rio Tinto “dragging something behind a ute for density mapping.” Here’s the honest disambiguation, because two different things are getting conflated:
What “dragged behind a ute” almost certainly is: a towed ground EM/TEM conductivity array — NOT a Rio Tinto density tool. There’s a whole real class of systems where a transmit/receive coil frame is towed behind a quad, ute or ATV and logs near-surface electrical conductivity continuously as you drive. These are genuinely “dragged behind a vehicle.” Confirmed examples:
That family maps conductivity contrasts — which is exactly the clay-vs-ironstone contrast Steven cares about — so it’s the useful thing to chase (see §6 TDEM below).
What Rio Tinto actually uses for “density” — and why it’s probably NOT a towable consumer tool. Rio Tinto’s public orebody-density / grade tech in the Pilbara is a different animal:
So: I could NOT confirm any specific Rio Tinto “ute-dragged density mapper,” and there is no evidence of a scaled-down consumer version of one. My honest read (moderate confidence): Steven is remembering the towed-EM-conductivity-array concept (which is real, is dragged behind a vehicle, and is buy/hire-able) and has attached the “Rio Tinto / density” label to it from separate mining-news memory. The genuinely available small-operator version of the useful idea is an operator-carried or towed TEM system — see the Australian Loupe system in §6, which is the closest real-world thing to “walk/drive it over the claim and get a conductivity-depth section.” Do not treat “Rio Tinto’s ute tool” as a purchasable product — it isn’t a confirmed single thing.
How it works. You lay out a line of steel electrodes at fixed spacing, inject current through pairs and measure voltage through others; a switching resistivity meter cycles through hundreds of quadripoles and inverts them into a 2-D vertical resistivity cross-section (a “pseudosection”, then an inverted model). Clay = conductive (low resistivity); ironstone/silcrete = resistive (high). That is a textbook strong contrast — ERT should light up the clay floor and the ironstone bands sitting on it better than any other surface method here.
Array types & depth for ~8 m. Common arrays: Wenner (best vertical resolution / signal-to-noise, good for flat-lying layers like this), dipole-dipole (better lateral/structural resolution, noisier), Schlumberger (a compromise). Depth of investigation ≈ ⅓–½ of the total line length, and roughly the electrode spacing sets the shallow detail. For an 8 m target: ~1–2 m electrode spacing over a ~40–60 m line gets you there with metre-scale layer resolution. Very achievable on a 15 ha claim.
Portability. Ute-portable, not vehicle-towed. You carry the meter (a box), a couple of cable reels, and a bag of electrodes; two people lay and move lines. No towing, no smooth-ground requirement — works on rough claim country.
Commercial units. AGI SuperSting R8 (8-channel, the workhorse), ABEM Terrameter LS / LS2 (Guideline Geo), IRIS Instruments Syscal Pro / Kid, PASI / GF Instruments for budget end. All Australian-serviceable.
Published use for shallow regolith / opal. Resistivity is explicitly listed among “geophysical surveys using magnetic, resistivity, shallow seismic and ground penetrating radar [that] have been used with varying success” in the opal fields — and it’s the one whose physics best matches a clay-dam target. Widely used generally for clay/regolith/palaeochannel mapping.
Cost (AUD). Contractor ERT in Australia runs roughly $2,000–$4,000/day for field acquisition, plus ~half a day processing per field day (indicative market figures, not a quote). Rental of a resistivity system is possible but less common than GPR hire — unverified day-rate; expect it to sit above GPR rental because the kit is dearer. A new SuperSting-class system is a large five-figure AUD purchase — contractor-per-day is the sane path for one claim.
How it works. A single 3-component sensor sits and records ambient ground vibration (wind, distant traffic, ocean microseisms). The horizontal-to-vertical spectral ratio (HVSR) peaks at a resonance frequency set by the impedance contrast and thickness of a soft layer over a stiff one; f₀ ≈ Vs / 4H lets you back out depth-to-hard-layer. One person, one puck, a few minutes per station — cheapest option here.
Portability. Hand-carried, single station. The Tromino (MoHo/Italy) is the classic — pocket-sized, battery, no cables, no electrodes.
Realistic resolution at 5–8 m (be honest). Probably too coarse. HVSR shines from a few tens of metres to hundreds of metres of sediment; getting a clean, reliable resonance for a 5–8 m layer needs a strong impedance contrast and pushes the method to the shallow edge of its useful range (high f₀, easily lost in cultural noise, systematic depth underestimation of ~10–15%). Some studies do resolve <10–40 m sediment thickness, so it’s not impossible — but for mapping 3–4 thin ironstone bands at a few metres it will at best give a fuzzy depth-to-clay, not layer detail. Verdict: a cheap reconnaissance “how deep is the hard floor here?” tool, not a layer-mapper. Don’t expect it to resolve the ironstone levels.
Cost (AUD). A Tromino is a five-figure AUD purchase (unverified exact figure); occasional rental/contract HVSR exists but pricing is unverified. Cheapest per-station of any method if you already have the instrument.
How it works. You put energy into the ground (hammer plate, weight-drop, small vibrator) and record reflections on a spread of geophones; standard for imaging hundreds of metres to kilometres deep (oil/gas, coal, deep structure).
Why it’s wrong for 5–8 m. At a few metres depth the reflections arrive almost on top of the source pulse — you’re inside the near-field/ground-roll mess, and the vertical resolution you’d need (sub-metre) demands very high frequencies that don’t propagate cleanly. It’s expensive, crew-and-gear-heavy, and mis-scaled for a shallow claim. (Shallow refraction or MASW surface-wave seismic is more appropriate at shallow depth than reflection, but still coarser and pricier than ERT for this job.) Verdict: not fit for purpose — noted only to rule it out. Contractor reflection seismic is a many-thousands-to-tens-of-thousands AUD mobilisation — wildly disproportionate to a 15 ha opal claim.
How it works. A transmit loop energises the ground; the current is switched off sharply and the decaying secondary field (eddy currents) is measured over time. Later time = deeper. It returns a conductivity-vs-depth sounding — and, like ERT, it keys directly off the clay (conductive) vs ironstone/sand (resistive) contrast, so it’s a strong physical match for a clay-dam target. Unlike ERT it needs no electrodes in the ground (no ground contact), so it’s faster to cover area.
Depth range & the shallow catch. TDEM’s strength is tens to hundreds of metres; the very shallow (<5–8 m) end is where it’s weakest, because the earliest time-gates (which carry the near-surface signal) are the hardest to record cleanly. Purpose-built near-surface systems push this down. This is the class Steven’s “dragged behind a ute” memory really belongs to (§2).
Portable / towed units (this is the practical bit).
Published use. EM/conductivity methods are standard for clay, salinity, regolith and palaeochannel mapping; not as specifically cited for opal as GPR/gamma, but the physics fit is arguably better than GPR’s for the clay floor.
Cost (AUD). Contractor ground-TEM day-rates are unverified but broadly comparable-to-dearer than ERT (specialist kit + processing). Loupe/AgTEM purchase and hire are contact-for-quote / unverified. Treat this as “get a quote from Loupe Geophysics (Brisbane) and one ground-EM contractor” rather than a known number.
| Technique | Depth fit for 5–8 m | Resolution | Clay-dam signal match | Portability | DIY/rental (AUD) | Contractor (AUD/day) | Verdict |
|---|---|---|---|---|---|---|---|
| GPR (200–250 MHz) | ✅ Yes through sandstone → ironstone | 🟢 High (10s cm), direct image | 🟢 Good — basal clay is UNDER the target; images it, then stops | Handheld / cart / ute-tow | ~$300–800/day (unverified) | ~$1,200–3,600/day incl. report (indic.) | Top pick for typical uncapped-sandstone areas ⚠️ fails only where clay CAPS the target |
| ERT (Wenner, ~1–2 m spacing) | ✅ Ideal | 🟢 Metre-scale layers | 🟢🟢 Excellent — clay conductive, ironstone resistive | Ute-portable (carry + lay) | Rental uncommon (unverified) | ~$2,000–4,000/day + processing (indic.) | Best-in-class where clay CAPS or cap unknown; the fallback to GPR |
| Passive seismic HVSR (Tromino) | ⚠️ Shallow edge / marginal | 🔴 Coarse (depth-to-floor only) | 🟡 Impedance, not layers | Hand-carried single station | Instrument = 5-fig buy (unverified) | HVSR contract pricing unverified | Cheap recon “how deep is hard floor”; won’t map ironstone bands |
| Active reflection seismic | 🔴 Overkill / too deep-focused | 🔴 Poor at few m | 🟡 N/A here | Crew + spread, heavy | n/a | Many $1,000s–$10,000s mob (indic.) | Ruled out — wrong scale |
| TDEM / towed EM (Loupe, tTEM, GEM-2) | ⚠️ Weakest at very shallow; near-surface rigs help | 🟡 Moderate | 🟢 Strong — conductivity contrast | Backpack-carried (Loupe) / ATV-towed (tTEM) | Purchase/hire = contact-for-quote (unverified) | Unverified; ≈ ERT+ | Strong match; “the ute-dragged thing”; get Loupe (Brisbane) quote |
Bottom line for the claim: GPR is the top pick for Steven’s typical dig conditions — where the profile is uncapped sandstone → ironstone → basal clay, GPR fires through the sandstone, images the ironstone target in high-resolution 2-D/3-D, and only dies at the clay beneath the target. It’s higher-resolution at shallow depth than anything else, cheapest per site, and walk-behind or ute-dragged. ERT is the best-in-class fallback for the minority of areas where clay CAPS the sandstone (or the cap is unknown) — its clay-vs-ironstone physics doesn’t care about layer order, at ~$2,000–4,000/day. TDEM/Loupe is the honest home of the “dragged behind a ute” idea and worth a quote for fast conductivity reconnaissance. Passive seismic is a cheap “how deep is the hard floor” extra; reflection seismic is out. Practical plan: budget a GPR trial first; keep ERT in the back pocket for capped ground.
Steven’s scepticism about the free government geophysics layers is correct, and this section says why — so the app doesn’t oversell them.
Two independent reasons the national mag grid (GA magmap_v7_2019_*) is a weak signal at Opalton:
Even a weak field can reveal faults, lineaments and basement trends, and those structures matter because they control where the basal clay slumps, folds and ponds silica-rich water into clay-dams. You don’t read structure off raw TMI — you apply enhancement filters:
geophys:magmap_v7_2019_RTP — already the app’s default mag layer (§2).geophys:magmap_v7_2019_1VD (and VRTP_1VD) as a separate WMS layer — the app could add a 1VD toggle and it’d be the better structural view than RTP for shallow trends...._VRTP_AS).So the app’s honest move on mag is: keep it, but relabel it a “structure/lineament” layer, not an “ironstone detector,” and add the 1VD layer (already published, one WMS call) as the sharper shallow-structure view.
Gamma-ray spectrometry (GA radmap_v4_2019_*, §3) reads the top ~30–40 cm of the surface — it’s a weathering/alteration map, and weathering is exactly what makes opal country:
GA publishes K (..._pctk), Th (..._ppmth), U (..._ppmu), ternary image and the ratio grids (incl. ..._ratio_tk) on the same WMS already wired in §3 — so a K/Th ratio or ternary layer is a one-call add and is more defensible than mag over Opalton.
Over Opalton: airborne mag = weak signal (too coarse, wrong iron chemistry — keep it only as a structural/lineament layer via RTP/1VD); radiometrics = moderate signal (real weathering/alteration proxy, with a genuine opal-gamma precedent — prefer the K/Th ratio or ternary); the real geophysical prize is run over the claim itself — GPR for the typical uncapped-sandstone dig areas (images the ironstone target above the basal clay), ERT where a capping clay is present or suspected.
Practical app take: add the GA 1VD mag layer and a K/Th-ratio (or ternary) radiometrics layer, label both honestly as regional context — and make clear the actual pre-dig decision comes from on-claim ERT/GPR, not the free basin data.
Future-planning only — not v0.1. Recorded so a later “where else could Steven look” question starts from evidence, not vibes.
The opal-bearing Winton Formation crops out in a ~1,000 km belt down the eastern margin of the Eromanga Basin: Winton–Opalton in the north, through Jundah/Yowah/Koroit, to Quilpie in the south. Historical opal mining sits almost entirely on this eastern arc. Why?
Why mining concentrates on the EASTERN margin (weighing the causes):
Published surveys of the western side? I found no dedicated published opal survey of the western Eromanga margin. The prospectivity modelling is basin-wide (so it “covers” the west mathematically) but the west scores low, and there’s no field campaign literature I could locate specifically assessing western-margin opal ground. Flag as a genuine gap / unverified — worth a proper GSQ literature check before acting.
Is the west a real opportunity, or dead ground? — honest lean. On the evidence, the western margin is more likely genuinely less favourable than merely under-explored — because the controlling factor is shallow weathered Winton Formation, and that thins/deepens westward under younger basin cover. The concentration of mining in the east looks mostly geological, not just historical accident. However, confidence is moderate, not high: the “it’s just under-explored” case isn’t zero (access bias is real, and the prospectivity model does extend prospectivity into “northern and southern sectors” beyond the classic fields), and I could not find a study that directly tested western-margin ground. Lean: east is favoured for sound geological reasons; the west is probably deeper/less-weathered dead-ish ground rather than a hidden field — but treat this as a hypothesis to test against GSQ mapping (depth-to-Winton, regolith/weathering-depth maps), not a settled conclusion.