Kwekwe reef prospect

A prospect the holder identified from an old geological map as a "long red oxide reef", ~4 km from both the Kwekwe partners block and Kwekwe block 2. No pegs exist yet (pegging planned 2026-07-15): the boundary shown is a nominal 1500 × 1300 m analysis window (~195 ha) around the box the holder highlighted on a GPS-app screenshot, not a claim. Three waypoints legible on that screenshot pin the reef trace itself — a ~1.1 km NE–SW line from 780734 E / 7910650 N to 779900 E / 7909989 N (UTM 35S). 87 known gold mines fall within 40 km, and 54 within 25 km — the richest local calibration set of any block in this study. This is a where-to-walk assessment on free satellite data, not a claim that gold is present.

Peg the box, but weight it southwest. As a whole the window reads at the 40th percentile of local background (fused P(gold) median 0.228 at 10 m) — background-level, comparable to Kwekwe block 3. The signal is in the reef trace itself: P(gold) climbs steadily NE→SW along the line (0.26 → 0.29 → 0.32 at the waypoints), peaks at 0.47 at 779948 E / 7910025 N near the SW waypoint, and the window's three best spots cluster 100–200 m beyond that same SW terminus, led by 780022 E / 7909903 N (P(gold) 0.59). If the pegging has flexibility, extend the block south to ~7909800 N to capture them. Walk-first ground is the ~300 m arc from 779948 E / 7910025 N through 780022 E / 7909903 N to 780442 E / 7909833 N — the reef's SW terminus and the two model maxima in one traverse.

One honest caution: current Sentinel-2 shows no elevated red-oxide staining along the mapped reef — the trace reads at the 13th percentile in ferric ratio, on fully exposed ground at 0% cloud. That does not refute the old map (the district's 54 known mines carry no ferric signal either, +0.07σ; laterite and farmed soil swamp it, and the ground may simply be cultivated over), but the satellite cannot confirm the reef as a surface red-oxide feature. The terrain arm is what backs this ground: 79th percentile against the strongest terrain model in the set (LOO AUC 0.937 on 54 mines).

The best-ground spots (UTM 35S). Spots 1–3 are the SW terminus cluster (purple); spots 5 and 6 sit on the reef trace itself (red); spot 4 is the northern control point (grey):

The reef trace, sampled

The three legible waypoints from the holder's screenshot, densified to 50 m spacing (24 samples over 1065 m) and read from the 10 m fused P(gold) grid:

Along the reef (NE → SW) P(gold) Percentile vs window
NE waypoint 780734 E / 7910650 N 0.256 50th
mid waypoint 780348 E / 7910324 N 0.288 61st
SW waypoint 779900 E / 7909989 N 0.320 71st
hottest point on the line, 779948 E / 7910025 N 0.470 —

The gradient is monotonic: whatever the model is reading strengthens toward the SW end and continues just past it into the spots 1–3 cluster. The ferric ratio along the same line has median 1.407 — the 13th percentile of the window.

The five checks behind that

Each is scored over the window against the ground immediately around it.

Check This window What it means
Terrain match 79th percentile scored against the terrain signature of 54 known gold mines within 25 km — the largest and most skilful local terrain ensemble in the set (LOO AUC 0.937, spatial CV at 8 km holdout 0.872)
Red-oxide (ferric) 33rd percentile (13th on the reef trace) below background on fully bare ground (S2 2025-09-27, 0% cloud); the mine calibration shows ferric is uninformative in this district anyway (54 mines at +0.07 sigma separability — laterite and farmed red soil dominate)
Structure / lineaments 51st percentile background; dominant trend 120–135 degrees; farmed flat terrain caveat applies
ASTER alteration (SWIR) Al-OH 49th, carbonate/Mg-OH 49th percentile archive day scene 2004-07-30 at 0% cloud; flat, and the 54 local mines carry no alteration signal beyond chance (Al-OH −0.04 sigma) — consistent with the sub-pixel-selvage finding across this district
Regional P(gold) 0.228 whole-window (40th percentile); SW cluster 0.55–0.59 background-level overall, with the SW reef terminus as the sole above-average ground

Compared to your other blocks

Same estimate, run over each block:

Block P(gold) median Percentile vs local background
Belingwe-West 0.394 n/a
Silobela site 2 (window) 0.358 71st
Silobela site 1 (window) 0.331 63rd
Zvishavane 0.282 67th
Kwekwe block 3 0.285 36th
Kwekwe block 2 0.254 45th
Kwekwe reef prospect (this), whole window 0.228 40th
Kwekwe reef prospect (this), SW cluster 0.55–0.59 n/a
Kwekwe block 1 0.129 8th

Whole-window this is background ground, at the bottom of the Kwekwe set apart from block 1. What distinguishes it is that the above-average patch is not an arbitrary corner: it coincides with, and extends, the SW end of an independently mapped reef. An old-map structure and a satellite model agreeing on the same 300 m is a better reason to walk than either alone.

Data used

The single most useful next step is ground truth from the SW terminus arc, 779948 E / 7910025 N → 780022 E / 7909903 N → 780442 E / 7909833 N (UTM 35S): a handheld pXRF reading on outcrop or soil, or pan-and-grab sampling, on that one ~300 m traverse. It tests the old map's reef and the model's hottest ground in a single visit — and if the reef is real but cultivated over, that is where it should still be findable at surface.