| Summary: | We assess d18O‐ and Mg/Ca–based paleothermometry in relation to seasonal variations in the shell fluxes of the surface‐dwelling planktonic foraminifera G. ruber and G. trilobus. Using deep‐moored sediment traps in the Mozambique Channel, upstream of the Agulhas Current, we find a distinct antiphase response to the annual cycle in sea surface temperature (SST) in G. ruber and G. trilobus. Maximum fluxes of G. ruber occur in late austral summer (February–March) when SST ranges between 28.7°C and 30.3°C. By contrast, G. trilobus maxima appear in early winter (June–July) at a lower SST between 25.3°C and 27.0°C. Cross‐correlation of the 2.5 year time series data confirmed that the G. ruber/G. trilobus ratio (R/T ratio) closely followed the annual cycle in SST, as did their paired d18O and Mg/Ca. In all proxies, G. ruber showed a 3 week phase lag against SST and G. trilobus of 6 weeks consistent with (semi‐) lunar population turnover rates in combination with a 1 week settling time to the 2500 m deep ocean floor. After correcting for these lags, we derived five independent equations for improved paleothermometry in the southwestern Indian Ocean that specify fluxweighted annual mean SST. The offset between flux‐weighted SST of G. ruber and G. trilobus and annual mean instrumental SST is about 0.8°C despite their strong seasonality in shell export. This is far less than the mean difference in instrumental SST between summer and winter of 3.2°C. Both species therefore closely reflect annual mean SST in sediments within a margin of +0.5°C for G. ruber and –0.3°C for G. trilobus. We also find that multiple linear regression of the five proxies (R/T ratio and the d18O, Mg/Ca of G. ruber and G. trilobus) strongly reduces the effects of fast rotating eddies and improves the correlation coefficient from an r2 of 0.4 to an r2 of 0.8.
|
|---|