Abstract
The area of the ocean, as derived from a new analysis of two digital data sets, is near to 362.5 Mm2 (1Mm2 = 106 km2). The decimal digit is meaningful: uncertainty is about ±0.1 Mm2 or ±0.03%. Although it is impractical to quantify their dominant errors precisely, the ocean areas presented in both of the canonical sources are significantly more uncertain. The often-quoted figure of 361 Mm2 probably derives from the work of Kossinna in 1921 and represents the ocean without the ice shelves and floating glacier tongues, which have an area at present of 1.561 Mm2. The often-quoted figure of 362 Mm2 probably derives from the work of Menard and Smith in 1966 and includes the ice shelves, as it should for the purpose of converting masses of cryospheric or terrestrial water to sea-level equivalent units. However, when Kossinna's estimate is corrected by adding the ice shelves, the two canonical estimates are seen to be inconsistent. The discrepancy implies that measurement errors are much larger than estimated in one source or in both, but Kossinna's estimate, when corrected, agrees much more closely with the modern digital estimates than does the Menard-Smith estimate.
Acknowledgements
I thank the anonymous reviewers for careful comments.