Horizon 1 1 17 ACL : Articles dans des revues avec comité de lecture répertoriées par l'AERES Geochimica et Cosmochimica Acta 1282-1293 2010 fdi:010049238 ENG Geochimica et Cosmochimica Acta 0016-7037 ISI:000273824800007 4 10.1016/j.gca.2009.11.002 https://www.documentation.ird.fr/hor/fdi:010049238 https://www.documentation.ird.fr/intranet/publi/2010/02/010049238.pdf 74 Horizon (IRD) A geochemical proxy for surface ocean nutrient concentrations recorded in coral skeleton could provide new insight into the connections between sub-seasonal to centennial scale nutrient dynamics, ocean physics, and primary production in the past. Previous work showed that coralline P/Ca, a novel seawater phosphate proxy, varies synchronously with annual upwelling-driven cycles in surface water phosphate concentration. However, paired contemporaneous seawater phosphate time-series data, needed for rigorous calibration of the new proxy, were lacking. Here we present further development of the P/Ca proxy in Porites lutea and Montastrea sp. corals, showing that skeletal P/Ca in colonies from geographically distinct oceanic nutrient regimes is a linear function of seawater phosphate (PO4 (SW)) concentration. Further, high-resolution P/Ca records in multiple colonies of Pavona gigantea and Porites lobata corals grown at the same upwelling location in the Gulf of Panama were strongly correlated to a contemporaneous time-series record of surface water PO4 SW at this site (r(2) = 0.7-0.9). This study supports application of the following multi-colony calibration equations to down-core records from comparable upwelling sites, resulting in +/- 0.2 and +/- 0.1 mu mol/kg uncertainties in PO4 (SW) reconstructions from P. lobata and P. gigantea, respectively. P/Ca-Porites (lobata) (mu mol/mol) = (21.1 +/- 2.4)PO4 SW(mu mol/kg) + (14.3 +/- 3.8) P/Ca-Porites (lobata) (mu mol/mol) = (29.2 +/- 1.4)PO4 SW(mu mol/kg) + (33.4 +/- 2.7) Inter-colony agreement in P/Ca response to PO4 SW was good (+/- 5-12% about mean calibration slope), suggesting that species-specific calibration slopes can be applied to new coral P/Ca records to reconstruct past changes in surface ocean phosphate. However, offsets in the v-intercepts of calibration regressions among co-located individuals and taxa suggest that biologically-regulated "vital effects" and/or skeletal extension rate may also affect skeletal P incorporation. Quantification of the effect of skeletal extension rate on P/Ca could lead to corrected calibration equations and improved inter-colony P/Ca agreement. Nevertheless, the efficacy of the P/Ca proxy is thus supported by both broad scale correlation to mean surface water phosphate and regional calibration against documented local seawater phosphate variations. 032