Scientific studies that describe interannual-to-decadal variability of the temperature-salinity structure of the world ocean are reviewed. Studies published during the past twenty years indicate that the temperature-salinity structure of the world ocean changes more quickly, and to greater depths, than had been previously thought. Temporal variability has been observed in the deep waters of all the major ocean basins of the world and in many of the adjacent seas such as the Mediterranean Sea and Arctic Sea. For example, there has been a major redistribution of heat and salt in the North Atlantic Ocean from the surface through 2500 m depth during the 1960-1990 period. The deep water of the western part of the subtropical gyre has increased in temperature while the deep water of the subarctic gyre has cooled.
Some new results based on historical data recently made available through digitization projects and data exchange projects are presented. For the first time estimates of the interannual variability of upper ocean heat storage for the world ocean has been computed (for the 1960-90 period). Based on multivariate analysis of these heat storage estimates we note that during the this period, the first principal component of the heat content of the upper 200 meters of the North Atlantic Ocean (20N-70N) varied closely with the North Atlantic Oscillation in sea level pressure. Such observed variability has important implications for the possibility of decadal forecasts of the earth's climate system.
Observations of sea surface temperature are much more numerous than subsurface observations and are available for a longer historical period. It is through the study of both sets of observations that we may be able to infer past states of the subsurface ocean during periods when direct subsurface observations are not available in sufficient numbers for analysis.
A data analysis technique introduced by Levitus (1989) is being used by various researchers to diagnose the observed changes in the temperature-salinity structure of the ocean. This helps determine the type of forcing at the sea surface that caused the observed subsurface temperature-salinity anomalies. For example, interannual-to-decadal temperature-salinity differences may be due to mechanical vertical displacements of the temperature-salinity structure of a water column and/or may be due to changes in potential temperature and salinity on potential density (or neutral) surfaces. Being able to discern the forcing mechanisms of variability of subsurface ocean temperature-salinity anomalies is of importance to understanding air-sea interaction and climate variability.
The outlook for continued progress in describing the variability of the temperature-salinity structure of the world ocean is good due to the increased awareness of the importance of the role of the ocean in climate change [(IPCC (1996); WCRP (1995)]. This awareness has led to increased support during the past eight years for projects to build the global ocean databases needed to describe ocean variability. However, there are still substantial amounts of ocean data not yet available in electronic form that will lead to a better description and increased understanding of ocean variability. Such data are invaluable for describing time and space scales of ocean variability and their availability will lead to the design of better observing systems and also provide insight into the physics of ocean variability on interannual-to-decadal time scales.