Planktonic foraminifera, oceanic amoeba-like single-celled organisms that produce a calcareous shell or test, are often well preserved in ocean sediments. Some foram species are characteristic of cold water, others live only in warm water. Consequently, foram assemblages that accumulate in the sediment reflect the environment near the surface, and fossil forams extracted from sediment cores can be used to estimate past environmental conditions, usually sea-surface temperature, based on the modern relationship between forams and the environment. These foram-assemblage derived estimates of sea surface temperatures are widely used: over half of the Last Glacial Maximum temperature estimates in the MARGO compilation are derived from planktonic foram assemblage data. The MARGO LGM temperature estimates have been used to estimate the climate’s sensitivity to CO2 increases. As any biases in the foram-assemblage derived sea-surface temperature estimates will bias both the estimates of LGM climate and climate sensitivity, it is important to test for potential biases.
Planktonic foram assemblages are usually calibrated against near-surface temperatures, although many species live far below the surface. Initially this was done from necessity: when Imbrie and Kipp (1971) developed their method for making quantitative estimates of environmental conditions from foram assemblages, only a map of sea-surface temperatures for February and August was available. Imbrie and Kipp had no choice but to calibrate against surface temperatures. In the 50 years since Defant (1961) published these maps, an enormous amount of data on ocean temperature has been collected and compiled into an atlas with depth and season. Despite it now being possible to calibrate foram assemblages against any depth, most authors have continued to use near-surface temperatures, typically 10m depth, and there has been little attempt to check the impact of this decision.
In a paper published in Climate of the Past yesterday with Camille Li and Michal Kucera, we explore the impact of the choice of calibration depth. In the mid to high latitudes of the North Atlantic, it really doesn’t matter which depth in the top few hundred metres of ocean is used to calibrate the forams, as all the reconstructions have a similar pattern. In the tropics, reconstructions calibrated to different depths can be very different. At site V30-36, halfway between Africa and South America, reconstructions calibrated against surface temperature suggest relatively little temperature change since the Last Glacial Maximum. In contrast, reconstructions calibrated against 100m temperature indicate >5°C of warming at the end of the last glaciation. This dramatic difference between reconstructions calibrated against surface or subsurface temperatures reopens the long-standing debate about how cool the tropical oceans were during the Last Glacial Maximum, but also exposes the limitations of current methods for reconstructing past environmental conditions from species assemblages, especially when we don’t know to which environmental variables the species are responding.