One section, four competing explanations. Shallow geometry is shared; the deep structure and the driver change with the model you pick. Public sources only — no proprietary well or seismic data.
Regional cross-section with four working models
WHERE NW Dangerous Grounds → SE Sulu SeaLENGTH ~830 km along a dog-leg lineSCALE upper 15 km ~17× vertical · lithosphere ~3×MODELS 4 · one is a new hypothesis, untested
◈ Locator · section A–A′
Line and stations are schematic. Axis line drawn approximately along the Sabah Trough.
◈ Cross-section A–A′ · pick a working model
How to read this. The shallow geometry (basins, granite, unconformities) is the same in every model. Press a button to switch what sits under the Deep Regional Unconformity and below ~15 km: A rift-axis intrusions?, B overpressured mud canopy, C underthrust Dangerous Grounds, D slab remnant + gravity tectonics. Model A is the author's untested working hypothesis — drawn at equal weight so it can be argued with, not because it is established. These explanations address different depths and processes and can coexist; they are not four mutually exclusive whole-Earth models. What is drawn and what is inferred. One extension mechanism is drawn: delta-load extension at the shelf (shallow, amber). Post-subduction crustal thinning is inferred from petrology, thermal history and receiver functions — it is not imaged on this section, so it is not drawn as a horizontal glyph. Moho: the dashed line is schematic. The published constraint is ~60 km beneath the Crocker Range and ~24 km in central Sabah (Gilligan et al. 2026); the Crocker root is shown as an unattached dashed constraint, not fitted to the schematic Moho, and the 230–540 km stretch has no drawn Moho at all. That gap is the honest state of the data, not an omission to be papered over. What Model D does show deep is the vertical mantle driver above the slab. Time slice: the rock geometry is present-day; the ages on the event bar below are history, not a claim that extension is active now.
◂ ▸ Swipe the section horizontally — it is 1,100 px wide; the whole transect does not fit a phone screen.
◈ Model signatures at a glance
What each model predicts, in one ring each
What would decide it · discriminating observations
✓ predicted · ✗ contradicted / not expected · ○ silent. “Observed” means published; “open” means the decisive data exist in industry but are not public. Three rules for reading this table near the magnetic equator: (1) mafic bodies skew to magnetic lows here and weak-susceptibility intrusions can fall below detection — so the magnetic row can support a model but can never kill one alone; (2) amplitude and detectability depend on body strike at low latitudes — a near N–S strike is nearly invisible in the induced field, and remanence can flip an anomaly's sign entirely — so absence of an anomaly on a near N–S strike is INCONCLUSIVE, not evidence of absence; (3) heat-flow and magnetics results are local — domains pass and fail independently along the transect, no single result decides the section.
◈ Event strip · 35 Ma to present
Shared events (grey) and the selected model's events (colour)
Two evidence lines, one set of surfaces. All four regional unconformities — BMU/TCU, DRU, UIU, SRU — match global sequence boundaries in the Haq & Ogg (2024) chart (NgSB1, NgSB8, NgSB11, NgSB13; Δ ≤ 0.2 Ma in GEOX's correlation). Tectonics alone does not produce that coincidence: these surfaces carry a eustatic signal as well as a tectonic one. That is why "the DRU records the collision" is a claim to test, not a given — GEOX rule: DRU ≠ collision onset without structural evidence. Provenance: Morley (2024) tectonics × Haq & Ogg (2024) eustasy, relayed by GEOX geox://earth-system/sabah-correlation.
Sources
Read this honestly. Topography and bathymetry are schematic profiles from published values (Mt Kinabalu 4,095 m; Sabah Trough ~2.9 km). Bodies are interpretive schematics, not seismic picks; depths of deep crust and Moho are approximate. Dips on the upper panel are exaggerated (about 17x vertical scale): a true 5 degree dip is drawn near 56 degrees - read shapes, not angles. Present-day heat flow is not a record of past heating: an old intrusion can leave a maturity anomaly after its heat anomaly is gone. Lineage: this figure is an interpretive schematic hand-plotted from published values - it is not machine-derived from survey data or a canonical transect pipeline. Model A is a new working hypothesis by Arif Fazil and is untested: it is drawn at the same weight as the published models so that it can be argued with, not because it is established. ◂ BACK TO DOSSIER · SEVEN-MODEL OBSERVATORY ↗ · DOWNLOAD PDF ↓ · LAUNCH GEOX ↗