Salinity estimation from Gulf of Guinea, sediment core MD03-2707

The isotopic composition of surface seawater is widely used to infer past changes in sea surface salinity using paired foraminiferal Mg/Ca and d18O from marine sediments. At low latitudes, paleosalinity reconstructions using this method have largely been used to document changes in the hydrological cycle. This method usually assumes that the modern seawater d18O (d18Osw)/salinity relationship remained constant through time. Modelling studies have shown that such assumptions may not be valid because large-scale atmospheric circulation patterns linked to global climate changes can alter the seawater d18Osw/salinity relationship locally. Such processes have not been evidenced by paleo-data so far because there is presently no way to reconstruct past changes in the seawater d18Osw/salinity relationship. We have addressed this issue by applying a multi-proxy salinity reconstruction from a marine sediment core collected in the Gulf of Guinea. We measured hydrogen isotopes in C37:2 alkenones (dDa) to estimate changes in seawater dD. We find a smooth, long-term increase of ~10 per mil in dDa between 10 and 3 kyr BP, followed by a rapid decrease of ~10 per mil in dDa between 3 kyr BP and core top to values slightly lighter than during the early Holocene. Those features are inconsistent with published salinity estimations based on d18Osw and foraminiferal Ba/Ca, as well as nearby continental rainfall history derived from pollen analysis. We combined dDa and d18Osw values to reconstruct a Holocene record of salinity and compared it to a Ba/Ca-derived salinity record from the same sedimentary sequence. This combined method provides salinity trends that are in better agreement with both the Ba/Ca-derived salinity and the regional precipitation changes as inferred from pollen records. Our results illustrate that changes in atmospheric circulation can trigger changes in precipitation isotopes in a counter-intuitive manner that ultimately impacts surface salinity estimates based on seawater isotopic values. Our data suggest that the trends in Holocene rainfall isotopic values at low latitudes may not uniquely result from changes in local precipitation associated with the amount effect.表层海水同位素组成被广泛用于借助海洋沉积物中获取的有孔虫Mg/Ca与δ¹⁸O（d18O）配对数据，反演过去的海表盐度变化。在低纬度区域，采用该方法的古盐度重建工作大多用于记录水文循环的演变。该方法通常假设现代海水δ¹⁸O（d18Osw）与盐度的对应关系在地质历史时期保持恒定。模拟研究表明，这类假设可能并不成立，因为与全球气候变化相关的大型大气环流格局可在局地改变海水δ¹⁸O（d18Osw）与盐度的对应关系。迄今为止，这类过程尚未通过古气候数据得到证实，原因在于目前尚无手段重建海水δ¹⁸O（d18Osw）与盐度对应关系的过去变化。我们通过分析采自几内亚湾的海洋沉积物岩芯，应用多代用指标盐度重建方法解决了这一问题。我们测定了C37:2烯酮类化合物（C37:2 alkenones）中的氢同位素比值（δD，dDa）以估算海水δD的变化。研究发现，在距今10至3千年（kyr BP）期间，δD（dDa）值呈现约10‰的平稳长期上升趋势，随后在距今3千年至岩芯表层的时段内快速下降约10‰，最终数值略低于全新世早期的水平。这些特征与已发表的基于海水δ¹⁸O（d18Osw）与有孔虫Ba/Ca比值的盐度估算结果，以及通过孢粉分析得到的邻近区域大陆降水历史均不一致。我们结合δD（dDa）与海水δ¹⁸O（d18Osw）数据，重建了全新世的盐度记录，并将其与同一沉积序列中基于Ba/Ca比值得到的盐度记录进行对比。该联合方法得到的盐度变化趋势与基于Ba/Ca比值的盐度记录、以及通过孢粉记录推断的区域降水变化均更为吻合。我们的研究结果表明，大气环流的变化可通过与直觉相悖的方式改变降水同位素组成，最终影响基于海水同位素数值的表层盐度估算结果。我们的数据显示，低纬度区域全新世降水同位素数值的变化趋势，未必仅由与降水量效应（amount effect）相关的局地降水变化所导致。