Historical Introduction
5
layering of these features. From such evidence Merz and W¨ ust (1922) inferred a
layered scheme of meridional circulation in the Atlantic which was not challenged
until the 1960s. [I sketched earlier the development of ideas about the deep ocean
circulation (Warren, 1981), and so will not say much more about it here.] Notable
atlases displaying color plates of property sections were compiled by W¨ ust and
Defant (1936) and Wattenberg (1939) for the German Meteor Expedition in the South
Atlantic (1925–1927), and by Fuglister (1960) for the I.G.Y. (1957–1958) cruises in
the Atlantic; they have afforded several generations of oceanographers immediate,
vivid impressions of how the Atlantic Ocean is structured.
As observations accumulated, it became practical to prepare maps of the horizontal distribution of properties at depth as well as at the surface. In the Challenger
Reports Buchan (1895) included global maps of temperature and specific gravity
(salinity) at the sea surface, and temperature maps at depths to 1500 fathoms. The surface maps look pretty good, but there are bullets in the subsurface ones, which suggest
some bad data. However, the maps at 500 to 900 fathoms clearly depict the westwardpointing, high-temperature tongue of Mediterranean outflow water. A Mediterranean
undercurrent in the Strait of Gibraltar had been conjectured for two centuries (M. Deacon, 1971, Chap. 7), but it was not actually observed [with drogues and with temperature and salinity (specific gravity) measurements] until 1870 (Carpenter and Jeffreys,
1871); so Buchan’s maps may have been the first to show its subsequent spreading in
the open Atlantic. It was later depicted with much better observational coverage, in
salinity as well as temperature, in the excellent maps of W¨ ust and Defant (1936).
At this time Murray (1899) constructed the first global map of bottom temperature, apparently at the request of a geologist investigating the contrasting geographies
of living conditions between the near-surface and bottom waters for organisms found
as fossils in the sediments. Bottom-property maps can be misleading about flow fields,
though, because on account of stratification their tonguelike patterns register bottom
topography as much as currents; but physical oceanographers have continued plotting
them anyway—perhaps out of habit.
Other types of maps were tried too, in attempts to depict property variations
in surfaces or layers that better parallel the flow field. W¨ ust (1935) identified “core
layers”—layers in which a vertical extremum, like an oxygen maximum, occurred—
and mapped the extreme values of the properties observed in these layers over the
Atlantic. Montgomery (1938) advocated plotting on surfaces of constant potential
density instead, because to the extent that potential density is conserved in the subsurface circulation, flow and the mixing of water properties should take place along
such surfaces. These have at least the advantage over core layers that there are infinitely many of them; but when core-layer extrema do not track isopycnals (as they
frequently do not), interpretation of one or the other may be labored. Moreover, both
types of map are often ambiguous about the relative degrees to which mean advection
and lateral mixing shape the property variations. Nevertheless, they have been useful
in tracing property features back to sea-surface locations where characteristics are
imparted by air–sea exchange, or to sites of deep sinking.
5
layering of these features. From such evidence Merz and W¨ ust (1922) inferred a
layered scheme of meridional circulation in the Atlantic which was not challenged
until the 1960s. [I sketched earlier the development of ideas about the deep ocean
circulation (Warren, 1981), and so will not say much more about it here.] Notable
atlases displaying color plates of property sections were compiled by W¨ ust and
Defant (1936) and Wattenberg (1939) for the German Meteor Expedition in the South
Atlantic (1925–1927), and by Fuglister (1960) for the I.G.Y. (1957–1958) cruises in
the Atlantic; they have afforded several generations of oceanographers immediate,
vivid impressions of how the Atlantic Ocean is structured.
As observations accumulated, it became practical to prepare maps of the horizontal distribution of properties at depth as well as at the surface. In the Challenger
Reports Buchan (1895) included global maps of temperature and specific gravity
(salinity) at the sea surface, and temperature maps at depths to 1500 fathoms. The surface maps look pretty good, but there are bullets in the subsurface ones, which suggest
some bad data. However, the maps at 500 to 900 fathoms clearly depict the westwardpointing, high-temperature tongue of Mediterranean outflow water. A Mediterranean
undercurrent in the Strait of Gibraltar had been conjectured for two centuries (M. Deacon, 1971, Chap. 7), but it was not actually observed [with drogues and with temperature and salinity (specific gravity) measurements] until 1870 (Carpenter and Jeffreys,
1871); so Buchan’s maps may have been the first to show its subsequent spreading in
the open Atlantic. It was later depicted with much better observational coverage, in
salinity as well as temperature, in the excellent maps of W¨ ust and Defant (1936).
At this time Murray (1899) constructed the first global map of bottom temperature, apparently at the request of a geologist investigating the contrasting geographies
of living conditions between the near-surface and bottom waters for organisms found
as fossils in the sediments. Bottom-property maps can be misleading about flow fields,
though, because on account of stratification their tonguelike patterns register bottom
topography as much as currents; but physical oceanographers have continued plotting
them anyway—perhaps out of habit.
Other types of maps were tried too, in attempts to depict property variations
in surfaces or layers that better parallel the flow field. W¨ ust (1935) identified “core
layers”—layers in which a vertical extremum, like an oxygen maximum, occurred—
and mapped the extreme values of the properties observed in these layers over the
Atlantic. Montgomery (1938) advocated plotting on surfaces of constant potential
density instead, because to the extent that potential density is conserved in the subsurface circulation, flow and the mixing of water properties should take place along
such surfaces. These have at least the advantage over core layers that there are infinitely many of them; but when core-layer extrema do not track isopycnals (as they
frequently do not), interpretation of one or the other may be labored. Moreover, both
types of map are often ambiguous about the relative degrees to which mean advection
and lateral mixing shape the property variations. Nevertheless, they have been useful
in tracing property features back to sea-surface locations where characteristics are
imparted by air–sea exchange, or to sites of deep sinking.