Sr/Ca, U/Ca, and stable isotope data and bimonthly records of Ogasawara coral core OGA-02-1

Instrumental climate observations provide robust records of global land and ocean temperatures during the twentieth century. Unlike for temperature, continuous salinity observations in the surface ocean are scarce prior to 1970, and the magnitude of salinity changes during the twentieth century is largely unknown. Surface ocean salinity is a major component in climate dynamics, as it influences ocean circulation and water mass formation. Here we present an annually resolved reconstruction of salinity variations in the surface waters of the western subtropical North Pacific Ocean since 1873, based on bimonthly records of d18O, Sr/Ca, and U/Ca in a coral from the Ogasawara Islands. The reconstruction indicates that an abrupt regime shift toward fresher surface ocean conditions occurred between 1905 and 1910. Observational atmospheric data suggest that the abrupt freshening was associated with a weakening of the winds that drive the Kuroshio Current system and the associated subtropical gyre circulation. We note that the abrupt early-twentieth-century freshening in the western subtropical North Pacific precedes abrupt climate change in the northern North Atlantic by a few years. The potential for abrupt regime shifts in surface ocean salinity should be considered in climate predictions for the coming decades.

Palynology and dinoflagellate counts of surface sediments, and sea surface conditions, of sites in the North Pacific

Palynological analyses were performed on 53 surface sediment samples from the North Pacific Ocean, including the Bering and Okhotsk Seas (37-64°N, 144°E-148°W), in order to document the relationships between the dinocyst distribution and sea-surface conditions (temperatures, salinities, primary productivity and sea-ice cover). Samples are characterized by concentrations ranging from 18 to 143816 cysts/cm**3 and the occurrence of 32 species. A canonical correspondence analysis (CCA) was carried out to determine the relationship between environmental variables and the distribution of dinocyst taxa. The first and second axes represent, respectively, 47% and 17.8% of the canonical variance. Axis 1 is positively correlated with all parameters except to the sea-ice and primary productivity in August, which are on the negative side. Results indicate that the composition of dinocyst assemblages is mostly controlled by temperature and that all environmental variables are correlated together. The CCA distinguishes 3 groups of dinocysts: the heterotrophic taxa, the genera Impagidinium and Spiniferites as well as the cyst of Pentapharsodinium dalei and Operculodinium centrocarpum. Five assemblage zones can be distinguished: 1) the Okhotsk Sea zone, which is associated to temperate and eutrophic conditions, seasonal upwellings and Amur River discharges. It is characterized by the dominance of O. centrocarpum, Brigantedinium spp. and Islandinium minutum; 2) the Western Subarctic Gyre zone with subpolar and mesotrophic conditions due to the Kamchatka Current and Alaska Stream inflows. Assemblages are dominated by Nematosphaeropsis labyrinthus, Pyxidinopsis reticulata and Brigantedinium spp.; 3) the Bering Sea zone, depicting a subpolar environment, influenced by seasonal upwellings and inputs from the Anadyr and Yukon Rivers. It is characterized by the dominance of I. minutum and Brigantedinium spp.; 4) the Alaska Gyre zone with temperate conditions and nutrient-enriched surface waters, which is dominated by N. labyrinthus and Brigantedinium spp. and 5) the Kuroshio Extension-North Pacific-Subarctic Current zone characterized by a subtropical and oligotrophic environment, which is dominated by O. centrocarpum, N. labyrinthus and warm taxa of the genus Impagidinium. Transfer functions were tested using the modern analog technique (MAT) on the North Pacific Ocean (= 359 sites) and the entire Northern Hemisphere databases ( = 1419 sites). Results confirm that the updated Northern Hemisphere database is suitable for further paleoenvironmental reconstructions, and the best results are obtained for temperatures with an accuracy of +/-1.7 °C.