The Yellow River, China's second-longest river and the largest sediment source to the Bohai Sea, only became a single through-going river draining to the sea when it incised the Sanmen Gorge — the last gorge along its course, at the exit from the Fenwei Basins toward the North China Plain. Before that event, the middle-reach basins hosted internally drained lakes and the river's reach to the ocean did not exist in its modern form. Proposed ages for this drainage integration span more than an order of magnitude — from late Miocene–early Pliocene, through ~1.8 Ma, ~1.25–1.5 Ma, ~0.94 Ma, and mid-Pleistocene terrace sequences (ca. 621–245 ka), down to ca. 150 ka — and the dominant forcing (Tibetan Plateau and North China Craton tectonics versus Mid-Pleistocene climatic/eustatic change versus headward erosion breaching upstream paleolakes) is equally contested. The problem is to determine when the Sanmen Gorge was incised and the exorheic Yellow River was established, identify the dominant forcing, and account for the conflicting terrestrial, drill-core, loess, and offshore constraints.

Background

The Yellow River (Huang He) flows ~5464 km from the Tibetan Plateau through a chain of Cenozoic basins to the Bohai Sea. During the late Cenozoic, much of its middle course was not a single river: basins such as the Fenwei Basin (which contains the Weihe and Sanmen sub-basins) held internally drained (endorheic) lakes, isolated from the sea, until they were captured and integrated into one exorheic (externally draining, ocean-reaching) system. A key bottleneck is the Sanmen Gorge, cut through bedrock near the modern Sanmenxia ('Three Gate Gorge') between the Loess Plateau margin and the North China Plain: breaching this gorge connected the upstream basins with the plain and the Pacific, an event often called the birth of the modern Yellow River. Determining when this happened matters because the integration age anchors the interpretation of the thick Mangshan loess sequence downwind of the gorge (whose anomalous thickness has been attributed to sediment supplied by the newly connected river), of the basin fills of the North China Plain, and of Bohai Sea margin stratigraphy, and it bears directly on the debate over whether large-river drainage integration in East Asia is driven mainly by tectonic uplift of the Tibetan Plateau and extension of the North China Craton, or by global climate change.

Dating the incision uses several independent techniques. Fluvial terraces (abandoned former floodplain levels) within and downstream of the gorge can be dated by optically stimulated luminescence (OSL), which determines when quartz or feldspar grains were last exposed to sunlight before burial, and by cosmogenic nuclide burial dating, which uses the decay of in-situ-produced radionuclides such as 10^{10}Be and 26^{26}Al in buried sediment to estimate burial age. Magnetostratigraphy ties sedimentary polarity reversals in drill cores to the geomagnetic polarity timescale (the dated standard sequence of Earth's magnetic field reversals), and detrital zircon U–Pb provenance fingerprinting (uranium–lead ages of zircon grains in river sediment, matched to source-terrane age spectra) identifies when sediment of unmistakably upstream derivation (e.g., from the Ordos Block — the stable crustal block in the river's middle reaches, whose sediment carries a distinctive, well-characterized zircon age signature) first appeared downstream of the gorge.

The published constraints disagree sharply. Jiang et al. (2007), from the Mangshan loess–palaeosol sequence and former Sanmen Lake sediments, concluded the gorge was downcut only ca. 150 ka (late Pleistocene). Kong et al. (2014), using cosmogenic burial dating of the Weihe paleolake drainage transition and gorge terraces, constrained the cut-through to 1.3–1.5 Ma. Hu et al. (2019) read the gorge gravel provenance as reinforcing a minimum age of 1.2 Ma. Wang et al. (2022) analyzed a drill core near the gorge and found Ordos-sourced channel deposits beginning ~1.25 Ma, tying final integration to the onset of the Mid-Pleistocene Transition (MPT; the shift, roughly 1.2–0.8 Ma, of glacial cycles from 41-thousand-year to 100-thousand-year periodicity) and arguing that accelerated eustatic sea-level fall during the MPT mattered as much as tectonics. Liang et al. (2022) correlated dated terrace profiles along the gorge and the Fenwei Basin and instead proposed a bottom-up integration: headward incision initiated at least ca. 621 ka, breaching the Fenwei Basin only ca. 245 ka, driven by subsidence of the North China Plain. Most recently, Huang et al. (2026) used borehole YJ-1 in the Yuncheng Basin (Fenwei Graben) to argue for initiation of integration at ~1.8 Ma, Chen & Miao (2026) showed that rapid Mangshan loess deposition began earlier than the 150 ka hypothesis allows, and a 2026 Bohai Strait provenance study identified a 0.94 Myr shift it attributes to river integration. Xiong et al. (2024) add a methodological caution: detrital-zircon provenance shifts in the region can be produced by enhanced loess reworking and bedrock unroofing without any drainage integration, so provenance signals must be used carefully. Two dedicated syntheses — Lin et al. (2025) on the Cenozoic evolution of the whole river and Lai et al. (2026) specifically on Sanmen Gorge incision and the transition from endorheic to exorheic drainage — survey this evidence, and the persistence of mutually incompatible ages in 2024–2026 primary literature shows the question remains unsettled.

Problem Statement

Determine when the Yellow River incised the Sanmen Gorge and thereby established a continuous, exorheic drainage from the Fenwei Basins and Loess Plateau to the North China Plain and the Bohai Sea — that is, establish the best-constrained age (or bounded age interval, with the uncertainty supported by the evidence adduced) of this incision/final-integration event, the 'birth of the modern Yellow River' in the sense of Wang et al. (2022) — and identify the dominant forcing of that integration (Tibetan Plateau / North China Craton tectonics, Mid-Pleistocene Transition climatic and eustatic change, or bottom-up headward erosion and paleolake-overflow capture). The published constraints on this event conflict — the Mangshan loess chronostratigraphy and the ca. 150 ka downcutting claim of Jiang et al. (2007); the 1.3–1.5 Ma cosmogenic burial ages of the Weihe paleolake drainage and gorge terraces (Kong et al. 2014); the ≥1.2 Ma gravel provenance constraints (Hu et al. 2019); the ~1.25 Ma onset of Ordos-sourced channel deposition in the drill core near the gorge (Wang et al. 2022); the ca. 621–245 ka terrace-series headward-integration model (Liang et al. 2022); the ~1.8 Ma Yuncheng Basin borehole evidence (Huang et al. 2026); and the offshore Bohai provenance-shift chronology — and a proposed age and forcing must account for them, whether by reconciling them or by demonstrating that particular constraints do not date the integration event. Ages of related process stages (initiation of headward incision, paleolake drainage, first arrival of upstream-derived sediment downstream of the gorge, terrace abandonment) are evidence the answer must interpret and reconcile with its proposed age for the incision/final-integration event, not alternative definitions of the target; new well-dated local evidence remains admissible as part of an answer.

References

  1. Jiang, F., Fu, J., Wang, S., Sun, D., Zhao, Z. (2007). Formation of the Yellow River, inferred from loess–palaeosol sequence in Mangshan and lacustrine sediments in Sanmen Gorge, China. Quaternary International 175, 62–70. DOI: 10.1016/j.quaint.2007.03.022. https://doi.org/10.1016/j.quaint.2007.03.022
  2. Kong, P., Jia, J., Zheng, Y. (2014). Time constraints for the Yellow River traversing the Sanmen Gorge. Geochemistry, Geophysics, Geosystems 15, 395–407. DOI: 10.1002/2013gc004912. https://doi.org/10.1002/2013gc004912
  3. Hu, Z., Li, M., Dong, Z., Guo, L., Bridgland, D., Pan, B., Li, X., Liu, X. (2019). Fluvial entrenchment and integration of the Sanmen Gorge, the Lower Yellow River. Global and Planetary Change 178, 129–138. DOI: 10.1016/j.gloplacha.2019.04.010. https://doi.org/10.1016/j.gloplacha.2019.04.010
  4. Liang, H., Zhang, K., Li, Z., Fu, J., Yu, Z., Xiong, J., Li, X., Ma, Z., Huang, P., Li, Z., Zhang, Y., Hui, G., Tian, Q., Wang, W., Zheng, W., Zhang, P. (2022). How headward erosion breaches upstream paleolakes: Insights from dated longitudinal fluvial terrace correlations within the Sanmen Gorge, Yellow River. GSA Bulletin. DOI: 10.1130/b36537.1. https://doi.org/10.1130/b36537.1
  5. Wang, X., Hu, G., Saito, Y., Ni, G., Hu, H., Yu, Z., Chen, J., Wang, M., Yuan, X., Wang, L., Hu, Z., Nie, J., Pan, B. (2022). Did the modern Yellow River form at the Mid-Pleistocene transition? Science Bulletin 67, 1603–1610. DOI: 10.1016/j.scib.2022.06.003. https://doi.org/10.1016/j.scib.2022.06.003
  6. Xiong, J., Zhang, P., Deng, C., Picotti, V., Liang, H., Ren, Z., Wang, W., Kang, H., Liu, Q., Zhao, X., Zhang, X., Zhang, Y., Li, Y., Zhang, H., Zhao, X. (2024). Neogene–Quaternary Channel Evolution and Provenance Shift of the Middle Yellow River. Journal of Geophysical Research: Earth Surface 129. DOI: 10.1029/2023JF007532. https://doi.org/10.1029/2023JF007532
  7. Lin, X., Dröllner, M., Barham, M., Liu-Zeng, J., Jolivet, M., Liu, H., Guan, K., Hu, C., Chen, X. (2025). The Cenozoic evolution of the Yellow River. Earth-Science Reviews 261, 104997. DOI: 10.1016/j.earscirev.2024.104997. https://doi.org/10.1016/j.earscirev.2024.104997
  8. Hu, H., Yu, Z., Yang, J., Lin, J., Yang, W., Li, B., Hu, G., Wang, X. (2025). Grain size and geochemical evidence for the integration of the Yellow River through the Sanmen Gorge. Palaeogeography, Palaeoclimatology, Palaeoecology 680, 113348. DOI: 10.1016/j.palaeo.2025.113348. https://doi.org/10.1016/j.palaeo.2025.113348
  9. Chen, Y., Miao, X. (2026). Early-onset of rapid deposition rates of the Mangshan loess challenges the classical hypothesis of full Yellow River connection during the late Pleistocene. Geomorphology 495, 110149. DOI: 10.1016/j.geomorph.2025.110149. https://doi.org/10.1016/j.geomorph.2025.110149
  10. Huang, T., Li, Z., Dong, X., Zhang, S., Guan, Y., Lei, Y., Wu, Z., Liu, Z. (2026). Multi-proxy evidence for the initiation of Yellow River integration through Sanmen Gorge at ~1.8 Ma: Insights from borehole YJ-1 in the Yuncheng Basin, Fenwei Graben. Palaeogeography, Palaeoclimatology, Palaeoecology 698, 113965. DOI: 10.1016/j.palaeo.2026.113965. https://doi.org/10.1016/j.palaeo.2026.113965
  11. Xiong, W., Mei, X., Huang, L., Bi, N., Han, Z., Zhang, Y., Wang, Z. (2026). A Mid-Pleistocene Provenance Shift in the Bohai Strait at 0.94 Myr: Implications for Synergistic Controls on Yellow River Integration. Journal of Ocean University of China 25, 1270–1284. DOI: 10.1007/s11802-026-6310-z. https://doi.org/10.1007/s11802-026-6310-z
  12. Lai, Z., Song, Y., Li, Z., Liu, B., Xiao, G., Pan, B. (2026). Sanmen Gorge incision and the formation of a super Yellow River: A review of evidence, age, and geomorphic models from endorheic to exorheic. Earth-Science Reviews 277, 105458. DOI: 10.1016/j.earscirev.2026.105458. https://doi.org/10.1016/j.earscirev.2026.105458