The possibilities of adapting the Modern Analogs Technique (MAT) based on planktonic foraminifers for estimating sea-surface temperatures (SST) in the Mediterranean during the Pliocene are discussed in this article. The calibration database used comprises 684 core-top samples distributed in the North Atlantic and the Mediterranean. MAT estimates show an imperceptible bias (-0.03 °C; sigma=0.59) and a low mean error of estimates (0.42 °C; sigma=0.42) when applied over the samples of the calibration dataset. The procedure used for assimilating the Pliocene taxonomic categories to those of the modern assemblages results in an increase from 17 to 40 in the number of samples showing an error N2 °C when applied over the calibration database. However, the precision of MAT does not diminish when these samples are removed from the dataset. This methodology was used for obtaining SST estimates of late-middle Pliocene–earliest Pleistocene samples from ODP-site 975 (Menorca area), which have close modern analogs within the calibration database. In order to compare this technique with an additional proxy, we measured also delta18O values of G. bulloides tests from these samples. The results obtained show a good agreement on the whole, which corroborates the validity of the technical approach proposed.

Supplement to: Serrano, Francisco; Gonzáles-Donoso, José M; Palmqvist, Paul; Guerra-Merchán, Antonio; Linares, Dolores; Pérez-Claros, Juan A (2007): Estimating Pliocene sea-surface temperatures in the Mediterranean: An approach based on the modern analogs technique. Palaeogeography, Palaeoclimatology, Palaeoecology, 243(1-2), 174-188