Correlated timing errors and terrain dependent uncertainty in synthetic slope microseismic localisation

Abstract Surface microseismic arrays can support monitoring of subsurface fracture activity, but increasing receiver density does not necessarily improve the reliability of reported locations. We present a reproducible synthetic benchmark that evaluates terrain coordinates, velocity calibration and timing correlation together. Six surface geometries share a 34.5 m horizontal envelope, with 16–576 receivers assigned to four timing groups. An independently implemented two-layer ray solver generates arrivals for 200 paired test events; a separate set of 12 known sources calibrates layer velocities. Generalised least squares is compared with diagonal weighting, homogeneous propagation and flattened receiver coordinates. At 64 receivers and a 0.15 ms array-shared clock standard deviation, covariance-aware inversion gives 90th-percentile location errors of 0.78–0.82 m and empirical 95% ellipsoid coverage of 96.5–97.5%. Diagonal weighting gives errors of 2.06–2.12 m and coverage of 33.0–34.5%. With 576 receivers, diagonal coverage falls to 3.5–4.0%, whereas propagating calibration uncertainty improves covariance-aware coverage to 94.0–96.0%. Additional waveform tests use unknown event times, noise-only detection calibration and automatic arrival picking. These tests expose sensitivity to coherent transients and reference contamination. A weak lateral slowness anomaly also produces severe undercoverage despite modest location errors. The results identify conditional accuracy and uncertainty-calibration requirements for slope-array design, without establishing field performance or an early-warning threshold.