Collision between Indian and Eurasian plates has created the largest plateau on our planet, the Tibetan Plateau, with an average altitude of 4000 m. Its formation has exerted large impacts on the regional, especially the arid Asian interior, and global climate evolution by changing the distribution pattern of land and sea and atmospheric dynamics. Several geodynamical models have been proposed to interpret the plateau uplift processes, which would predict different uplift histories for the northern Tibetan Plateau. Hence, determining the detailed topographic history of the northern plateau would provide vital insights into the underlying mechanisms of plateau uplift. However, commonly used methods, mainly stable isotopes and plant fossils, yield contrasting paleoaltitude estimates for the Qaidam Basin on the northern plateau, ranging from near-present elevation by the early Oligocene to substantial uplift since the mid-Miocene. Therefore, with known limitations in the two main paleoaltimeters, alternative indicators are desired to further constrain the paleoaltitudes of the Qaidam Basin.

Discovery of marine alkenones by chance

We started this project in 2010. Low-resolution carbonate oxygen and carbon isotopic records from the Dahonggou section in the Qaidam Basin were generated first, but a couple of high carbon isotopic values within the mid-Miocene did not attract much of our attention at that time. Driven by curiosity, we then analyzed the lipid fraction (which contains long-chain alkenones if they existed) of the same sample set. To our surprise, we found marine-type alkenones in a few samples, aged around the mid-Miocene, out of the total ~200 samples analyzed for the whole section. Puzzled by the occurrence of marine alkenones in a typical terrestrial lacustrine setting, we densely resampled the mid-Miocene Youshashan Formation at the Dahonggou section, and the subsequent analysis confirmed our initial finding (Fig. 1).

Since then, we (a group formed by many researchers specialized in various subfields) recognized that more field work and analyses were needed in order to address the occurrence of marine alkenones. Through more than ten years of investigations, three field sections in total, including Dahonggou, Honggou, and Hongliugou sections, and KC-1 well (to partially address the current chronological controversies in the northern basin), have been studied using both organic (biomarkers, carbon and hydrogen isotopes of leaf wax) and inorganic (carbonate oxygen and carbon isotopes) techniques. Our investigations across the basin indicate that the typical marine alkenones started to occur around the mid-Miocene, accompanied by anomalously high carbonate carbon isotopic values (Fig. 1).

Fig. 1 Cross-section profile correlating the stratigraphic positions of marine signatures in the Qaidam and Tarim Basin. a, DHG (Dahonggou). b, HG (Honggou). c, HLG (Hongliugou). d, KC-1 in the western Qaidam Basin. e, MR (Miran River) in the southeastern Tarim Basin. Identified marine signatures include marine alkenones from the DHG and HG sections and KC-1 well, carbonate carbon isotopic anomaly from the DHG and HG sections, and planktonic foraminifera from the MR section. The blue line marks the boundary between the Shang Youshashan (SYSS) Formation and Xia Youshashan (XYSS) Formation.

Interpretation of geochemical indicators

Marine alkenones are ubiquitous in marine settings, but not in modern inland lakes. Hence, as the Qaidam Basin became a terrestrial sedimentary basin after the early Cenozoic, the basinwide occurrence of marine alkenones during the mid-Miocene needs to be critically assessed. Sediment reworking of nearby pre-Cenozoic marine strata can confidently be excluded based on alkenone concentrations, the shift from marine to lake alkenones, variable temperatures estimated from marine alkenones, and alkenone occurrence in contrasting lake facies at the three sections. A proposal of the remnant species associated with the last major marine transgression during the late Eocene could also be refuted due to no alkenones detected in a wider range of lake facies before the mid-Miocene. As marine alkenones are common in coastal oceans (i.e., the Black Sea in particular), we thus use this modern analogue to infer possible seawater incursion during the mid-Miocene to account for the unusual marine signature found in the Qaidam strata. The abnormally high carbonate carbon isotopic values (> 0‰) within the same time period also appear to be best explained by seawater incursion.

Fig. 2 Hypothesized path of the seawater incursion from the Paratethyan Sea to the Qaidam Basin during the mid-Miocene. The blue dash line indicates a possible incursion path, a paleoriver along the southern Tarim Basin identified previously.

Based on available evidence, we hypothesize mid-Miocene seawater incursion as an accidental event, coincident with sea level rise and reactivated offset along the Altyn Tagn Fault serving to make the connection possible or deepen connecting channels at the time (Fig. 2). Many details remain to be explored in future studies. For instance, it is unclear whether the Tarim Basin remained connected to the Paratethys before the mid-Miocene. It also remains uncertain whether the Qaidam Basin was connected to the Tarim Basin before the mid-Miocene, although we consider this scenario less likely as we have not found any unusual marine geochemical signature in the Qaidam Basin for the concerned time period. Fortunately, these details do not substantially affect our main conclusion- mid-Miocene sea level altitude of the Qaidam Basin.

Our inference agrees well with the suggestion that the mid-Miocene southern Tarim Basin was at sea level and connected to open marine waters, inferred from planktonic foraminifers discovered at the Miran River section. Indeed, our study was inspired by this pioneering work and further suggests even higher amplitude of the northern plateau uplift (~1500 m vs. ~2800 m) since the mid-Miocene. Our inference is also largely consistent with numerous reports of accelerated uplift of various parts of the northern plateau since the mid-Miocene, and the suggestion that a profound geodynamic event seems to have occurred on and around the plateau during the mid-Miocene. At present, we consider that our hypothesis represents the most plausible explanation for the unusual marine signatures found in the mid-Miocene Qaidam strata, and if correct, it would effectively constrain the mid-Miocene Qaidam Basin close to sea level. With most of current geodynamical models placing the mid-Miocene Qaidam Basin at ~1000‒2000 m altitude, we wish that our new finding could be incorporated into models to better understand plateau uplift history and mechanisms.