Tropical Pacific Ocean thermocline depth reconstructions for the Last Glacial Maximum

Abstract

We evaluate the relationship between ten surface ocean (0–300 m) hydrographic parameters and the spatial distribution of factor‐analyzed core top planktonic foraminiferal abundances in the tropical Pacific Ocean (24°N–24°S) for core tops <3800 m. The spatial distribution of the first three faunal factor loadings (88% of the variance) are most highly correlated to subsurface variability (mixed layer depth, thermocline depth) and resistant species percent (RSP). However, RSP is not related to dissolution but is related to thermocline depth. Factor I (mixed layer species G. glutinata, G. ruber) and factor III (G. ruber) can be distinguished from each other by low abundances of G. glutinata in factor III. Both assemblages spatially comprise the deep mixed layer region of the western tropical and equatorial Pacific Ocean, but are associated with distinct water mass properties. A combination of Factor I and III loadings shows a higher correlation to thermocline depth (R² = 0.70). Factor II loadings (dominated by thermocline dwelling species N. dutertrei) are most significantly correlated with the thermocline depth (R² = 0.73). Most factors show only marginally significant correlation to sea surface temperatures (SSTs), indicating that SST is not the primary forcing factor on the planktonic foraminiferal species distributions in the tropical Pacific. A new transfer function was calculated to predict tropical Pacific thermocline depth from planktonic foraminifera abundances using the Imbrie‐Kipp Method (IKM) (standard deviation of residuals ±22 m (1σ)). An additional ±5‐m error is attributed to low species counts in the core top database. The modern analog technique (MAT) was also used to predict thermocline depth (standard deviation of residuals ±21 m). While last glacial maximum (LGM) thermocline depth changes by IKM and MAT were generally within error, estimated changes were geographically uniform, suggesting an oceanographic response to climate forcing. We estimate that the thermocline depth of the LGM was shallower than present by ∼20 m south of 8°S, possibly due to a shift in the South Pacific anticyclone to the northeast. Both the IKM and MAT estimate a steeper east‐west thermocline slope along the equator, suggesting that zonal wind stress (Walker circulation) was intensified during the LGM. Collectively, the thermocline estimates for the LGM suggest an equatorward compression of the climate zones in both hemispheres.