•We reconstruct deglacial/Holocene sea–ice covers north of Iceland using biomarkers.•Sea–ice conditions reveal unstable and complex climate situation in the region.•We find a deglacial sea–ice seesaw pattern between the western&eastern Nordic Seas.•The sea–ice variabilities were likely linked to the oceanic circulation changes.Sea–ice conditions on the North Icelandic shelf constitute a key component for the study of the climatic gradients between the Arctic and the North Atlantic Oceans at the Polar Front between the cold East Icelandic Current delivering Polar surface water and the relatively warm Irminger Current derived from the North Atlantic Current. The variability of sea ice contributes to heat reduction (albedo) and gas exchange between the ocean and the atmosphere, and further affects the deep-water formation. However, lack of long-term and high-resolution sea–ice records in the region hinders the understanding of palaeoceanographic change mechanisms during the last glacial–interglacial cycle. Here, we present a sea–ice record back to 15 ka (cal. ka BP) based on the sea–ice biomarker IP25, phytoplankton biomarker brassicasterol and terrestrial biomarker long-chain n-alkanols in piston core MD99-2272 from the North Icelandic shelf. During the Bølling/Allerød (14.7–12.9 ka), the North Icelandic shelf was characterized by extensive spring sea–ice cover linked to reduced flow of warm Atlantic Water and dominant Polar water influence, as well as strong meltwater input in the area. This pattern showed an anti-phase relationship with the ice-free/less ice conditions in marginal areas of the eastern Nordic Seas, where the Atlantic Water inflow was strong, and contributed to an enhanced deep-water formation. Prolonged sea–ice cover with occasional occurrence of seasonal sea ice prevailed during the Younger Dryas (12.9–11.7 ka) interrupted by a brief interval of enhanced Irminger Current and deposition of the Vedde Ash, as opposed to abruptly increased sea–ice conditions in the eastern Nordic Seas. The seasonal sea ice decreased gradually from the Younger Dryas to the onset of the Holocene corresponding to increasing insolation. Ice-free conditions and sea surface warming were observed for the Early Holocene, followed by expansion of sea ice during the Mid-Holocene.

•Organic temperature proxies were applied in the Northern South China Sea.•Surface and subsurface temperature records were reconstructed over the past 2.8 Myr.•An East Asian winter monsoon record was generated by upper water thermal gradient.•400 kyr cycle was found in East Asian winter monsoon records during 2.8–1.2 Ma.•400 kyr cycle was generated from remote forcing by super tropical Pacific ENSO.Variability of the East Asian winter monsoon (EAWM), stronger during glacials and weaker during interglacials, has been tightly linked to the wax and wane of the Northern Hemisphere ice sheets (NHIS) via the Siberian High over the last 2.8 million years (Myr). However, the long eccentricity cycle (ca. 400 kyr) in the EAWM record from the late Pliocene to early-Pleistocene (2.8–1.2 Ma) could not be linked to NHIS changes, which lacked the long eccentricity cycle in the Pleistocene. Here, we present the first low latitude EAWM record of the last 2.8 Myr using surface and subsurface temperature difference from the northern South China Sea to evaluate interactions between tropical ocean and EAWM changes. The results show that the EAWM variability displayed significant 400 kyr cycle between 2.8 Ma and 1.2 Ma, with weak (strong) EAWM during high (low) earth orbital eccentricity state. A super El Niño–Southern Oscillation (ENSO) proxy record, calculated using west-east equatorial Pacific sea surface temperature differences, revealed 400 kyr cycles throughout the last 2.8 Myr with warm phase during high eccentricity state. Thus, we propose that super ENSO mean state strongly modulated the EAWM strength through remote forcing to generate the 400 kyr cycle between 2.8 Ma and 1.2 Ma, while low NHIS volume was not sufficient to dominate the EAWM variation as it did over the last 0.9 Myr with 100 kyr cycles in dominance.

•Highest concentration of GDGTs in particles from the YS bottom layer.•Sedimentary TEX86 correlated better with bottom water temperature than with SST.•Regional calibration for TEXL86 proposed for reconstructing mean annual BWT.The TEX86 paleothermometer was applied for reconstructing sea surface temperature (SST) from glycerol dialkyl glycerol tetraethers (GDGTs) in marine sediments. It has become clear that GDGT-producing archaea live throughout the water column, with maximum concentration usually well below the surface mixed layer of the ocean. Why the TEX86 parameter correlates well with SST remains poorly understood. Here we evaluate the fidelity of the TEX86 thermometer using surface sediments and suspended particles from the Yellow Sea (YS), a shallow marginal sea between China and the Korean peninsula. The highest concentration of GDGTs in the water column at a site (A02) from the middle of the YS occurred in the bottom layer, at 70 m. This contrasts with phytoplankton lipids, which were most abundant near the surface. Consistent with the maximum abundance of GDGTs in bottom water, TEX86 values in surface sediments correlated better with mean annual bottom water temperature (BWT, R2 0.81) than with mean annual SST (R2 0.74). Moreover, TEXH86 temperature derived from a global core top calibration gave values 0.1–9.4 °C colder than mean annual SST but much closer to mean annual BWT (2.7–4 °C). Lastly, TEX86 and U37K′ paleotemperature values displayed distinctly different trends over the last 12 kyr for sediments from the shelf between the YS and the East China Sea (ECS), consistent with the notion that TEX86 and U37K′ reflect different temperature signals. This preponderance of evidence supports the use of TEX86 as a proxy for BWT in the YS. Therefore, we propose a local calibration of TEXL86 for reconstructing mean annual BWT (TEXL86 = 0.03 BWT-0.94; R2 0.86, n = 22, P < 0.0001). The combination of alkenone (U37K′) derived SST with TEXL86 BWT yielded a quantitative reconstruction of the vertical thermal gradient in the YS, and an insight into understanding the impact of the Kuroshio Current and East Asian monsoon on the YS.

•Spatial patterns of BIT, TMBR and δ13C values were similar for the southern YS.•Terrestrial OM from a binary model was lower than from a three end member model.•Plant OM input was higher than soil OM input to the YS.•Multi-proxies provided means of quantitatively reconstructing soil OM and plant OM.Marginal seas play an important role in the global carbon cycle as organic matter (OM) buried in shelf seas accounts for ca. 80% of marine sedimentary OM. The Yellow Sea (YS) is a significant sink for both terrestrial OM (TOM) and marine OM (MOM), so the source and spatial variation of sedimentary OM in the southern YS were investigated by analyzing several parameters for 54 surface sediment samples. Spatial ranges were 5.4–12.8 for total organic carbon/total organic nitrogen (TOC/TON), −22.8‰ to −19.9‰ for TOC δ13C, 421–4515 ng/g TOC for ∑(C27 + C29 + C31) n-alkanes, 33–152 ng/g TOC for branched glycerol dialkyl glycerol tetraethers, 316–8550 ng/g TOC for total marine biomarkers. TOC/TON displayed no distinct spatial pattern, but the BIT (branched isoprenoid tetratether index) proxy, TMBR [terrestrial and marine biomarker ratio: (C27 + C29 + C31n-alkanes)/(C27 + C29 + C31) n-alkanes + (brassicasterol + dinosterol + alkenones)] and δ13C values revealed similar spatial distribution patterns, indicating higher TOM near coastal regions and especially near the old Huanghe Estuary. Quantitative estimates using a binary model revealed higher %TOM (avg. 34%) from TMBR than from δ13C (avg. 26%) and BIT (avg. 12%). TMBR is a proxy incorporating both soil OM (OMsoil) and plant OM (OMplant) while BIT is a proxy for OMsoil, so quantitative estimates using a three end member model indicated higher OMplant and OMsoil values near the old Huanghe Estuary, but %OMplant (avg. 23%) was higher than %OMsoil (avg. 13%). Our study suggests that these proxies can be used to study the spatial and temporal variation and delivery mechanisms for both OMsoil and OMplant in marginal seas.

This paper reports on the spatial distribution patterns and investigates the controlling mechanisms of phytoplankton biomarkers (brassicasterol for diatoms, alkenones for haptophytes, dinosterol for dinoflagellates) and terrestrial biomarkers (odd C number long-chain (C27 + C29 + C31) n-alkanes) in surface sediments from the southern Yellow Sea (SYS). The contents of the phytoplankton biomarkers in the SYS surface sediments reveals a clear spatial pattern, with low values near the coasts and increased values seaward, caused by higher phytoplankton primary productivity and low sedimentation rates in the basin. The contents of terrestrial biomarkers show high values in the northern part of the study areas off the Shandong Peninsula and Jiangsu coast, caused by inputs of materials from the modern Huanghe River and the old Huanghe delta, respectively. The results also indicate that biomarker ratios offer the best approach for reconstructing marginal sea C cycles, as these proxies can be used to estimate the contributions of both terrestrial and marine organic matter and to reconstruct paleoproductivity and paleoecological changes in the SYS.Highlights► Distributions of terrestrial and marine biomarker in Yellow Sea sediments are unlike. ► Productivity and sedimentation rate control the distribution of marine biomarkers. ► Biomarkers are reliable proxies for reconstruction of carbon cycle and ecosystem changes in the YS.

•TEX86 and U37K' temperature records of the 356 kyr for the SCS.•Both TEX86 and U37K' records display similar glacial–interglacial difference of 4–5 °C.•TEX86 temperature is consistently lower than U37K' temperature over the past 356 kyr.•Lower U37K' and TEX86 differences during glacials were driven by stronger winter monsoon.Proxy records from the South China Sea (SCS) have played an important role in understanding late Quaternary SCS environmental changes and the East Asian Monsoon (EAM) variations, but few long sea surface temperature (SST) records have been generated from the SCS. The recently established TEX86 index provides an opportunity to generate long SST records, especially for tropical oceans, but the applicability and climatic interpretation of the TEX86 index for the SCS need further study. We report and compare orbital scale resolution (ca. 5.3 kyr) TEX86 and U37K' temperature records of the past 356 kyr for ODP Site 1147 from the northern SCS. Both records display clear and similar glacial–interglacial patterns. The U37K' SST fluctuated between 22.3 °C and 28.3 °C, with a glacial–interglacial difference of 4–5 °C; while the TEX86 temperature fluctuated between 20.3 °C and 25.4 °C, with a slightly smaller glacial–interglacial difference of ca. 4 °C. The TEX86 temperature has been consistently lower than the U37K' temperature over the past 356 kyr, consistent with previous studies suggesting that the U37K' index is a proxy for SST but the TEX86 index is a proxy for subsurface temperature. Thus, it is proposed that the temperature difference (ΔT) between the U37K' and TEX86 records reflects mixed layer depth (MLD) changes in the northern SCS, which fluctuated between 0.3 and 4.0 °C with high values mostly during interglacials and lower values mostly during glacial stages. MLD changes in the northern SCS are interpreted to be driven by EAM changes, with stronger winter monsoon resulting in a deeper MLD and a smaller ΔT. Thus, the combination of U37K' and TEX86 records and their ΔT can be used to reconstruct East Asian winter monsoon variations, as well as their influences on the hydrographic structure of the SCS.