A detailed δ18O and δ13C stratigraphy has been generated from analysis of well-preserved Albian–early Maastrichtian foraminifera from Deep Sea Drilling Project (DSDP) Sites 511 and 327 (Falkland Plateau; ≈58° S–62° S paleolatitude) in the southern South Atlantic, and Cenomanian and Coniacian–Santonian foraminifera from DSDP Site 258 (Naturaliste Plateau; ≈58° S paleolatitude) in the southern Indian Ocean. These results, when combined with previously published Maastrichtian stable isotope data from Ocean Drilling Program (ODP) Site 690 (Weddell Sea, ≈65° S paleolatitude), provide new insight into the climatic and oceanographic history of the southern high latitudes during middle–Late Cretaceous time. The planktonic foraminifer δ18O curves reveal a gradual warming of surface waters from the Albian through the Cenomanian followed by extremely warm surface waters from the Turonian through the early Campanian. Long-term cooling of surface waters began in the late early Campanian and continued through the end of the Maastrichtian. The benthic foraminifer δ18O record generally parallels changes in the oxygen isotopic curves defined by shallow-dwelling planktonic foraminifera. The vertical oxygen and carbon isotopic gradients were relatively low during the Albian–Cenomanian, high from the Turonian–early Campanian, and then low during the late Campanian and Maastrichtian.
Foraminiferal oxygen isotopic data from published sources and this study are averaged for each site, corrected for latitudinal changes in salinity based on modern-day surface-water values, and plotted versus paleolatitude for the late Albian, Coniacian–Santonian, and late Maastrichtian. Differences between lowand high-latitude surface-water paleotemperatures are estimated at ≈14 ° C during the late Albian and late Maastrichtian, but the Coniacian–Santonian reconstruction reveals only a 0–4 ° C latitudinal temperature gradient. Uncertainty regarding Cretaceous salinity gradients and possible diagenetic alteration of δ18O values introduce error into our estimates of paleolatitudinal thermal gradients; however, apparent low equator-to-pole temperature differences could indicate much higher poleward heat transport than at present.