Temperature proxies such as clumped isotope (Δ47) thermometry on biogenic carbonates are applied to the past with greatest confidence when the proxy-temperature relationship is shown to be robust within natural temperature conditions of the ocean. Especially well-suited for this purpose are biogenic carbonates sampled from well-constrained production period and oceanographic conditions of sediment traps. Since coccolithophorids have a cosmopolitan distribution and are major biogenic carbonate producers in the surface ocean, their coccoliths usually dominate the inorganic carbon flux in sediment traps and are sufficiently abundant in most traps for clumped isotope analysis. Here, we measured Δ47 in the coccolith size fraction of 18 sediment trap samples across a 75° latitudinal gradient and three ocean basins. To identify the upper ocean provenance region of the coccoliths in each trap, a simple model of coccolith transport by ocean currents was constructed. The coccolith Δ47 strongly follows the upper ocean temperatures, in particular the average temperatures from the maximum production depths of living coccolithophores from their provenance areas. There is no evidence for a coccolithophore species-specific effect on the Δ47-temperature relationship. Applying the recent coccolith-specific Δ47-temperature calibration (Clark et al., 2024a) to estimate calcification temperatures shows that inferred calcification depths match the depth of maximum coccolithophore production in the provenance area. Compared to other calibrations for biogenic carbonates, the coccolith-specific Δ47-temperature calibration yields the best agreement with the depth of maximum coccolithophore abundance and the expected depth of coccolith production.