Tectono-Climatic Coupling Across the Foreland-to-Rift Transition Controls Vertical Source-Rock Heterogeneity in the Triassic–Jurassic Succession of the Northern Kuqa Depression, Tarim Basin, NW China

Vertical heterogeneity is the main obstacle to predicting effective source rocks in lacustrine superimposed basins; yet, its origin remains poorly resolved because most studies examine a single interval or a single control. Five Triassic–Jurassic source-rock units (Huangshanjie, Taliqike, Yangxia, Kezilenuer and Qiakemake formations) of the northern Kuqa Depression were therefore analyzed within one organic–inorganic framework combining Rock-Eval pyrolysis, organic petrology, molecular biomarkers, and major- and trace-element geochemistry (redox proxies, Sr/Ba paleosalinity and the chemical index of alteration, CIA). Weathering, Rock-Eval matrix, and thermal-maturity effects were screened explicitly and do not control the between-horizon contrasts. Stable Al2O3–TiO2, Ti/Al, and SiO2/Al2O3 relations show that provenance was constant, so the contrasts between horizons record environmental change. Across the foreland-to-rift transition marked by the Triassic–Jurassic angular unconformity, lake hydrology and hydrochemistry evolved from suboxic brackish–fresh stratified lakes, through freshwater swamps and oxic deltaic systems, to a brackish, dysoxic semi-deep lake; organic-matter (OM) sources co-evolved from aquatic, through mixed terrigenous–aquatic and higher-plant plus algal, to terrigenous and, finally, phytoplankton dominance. Source-rock quality was set not by any single factor but by the shifting balance among productivity, preservation, and clastic dilution, whose relative weighting was reset when the basin changed type: the TOC maximum (29.1 wt %) and the hydrogen-index maximum (775 mg HC/g TOC) fall in different units and record opposite environmental configurations. Because no single parameter ranks the horizons correctly, a depositional model calibrated on one interval cannot be extrapolated across a change in basin type. The resulting coupled model explains why chemically distinct intervals stack vertically and yields a transferable prediction workflow for analogous basins.