ography makes it difficult to fund projects of intermediate size and this difficulty is compounded if the project is interdisciplinary.

Despite these problems, there was consensus that the National Science Foundation's core program is an invaluable asset of the field. The peer-review system maintains a balance between scientific rigor and responsiveness and ensures continuing support for innovative and fundamental science.

Challenging as it may be to make progress on any scientific problem, it is even more difficult to predict the future course of scientific progress. One might say that every important discovery in science is, almost by definition, unpredictable and so it is futile to guess at future triumphs. Indeed, it is worse than futile if these guesses are used to ''manage" the direction and content of science. It is our belief that basic research, independent of any practical concerns, is critical to the advance of science and the development of technology. Science is the most serendipitous of human enterprises and the ability of physical oceanography to solve problems of social concern depends on a healthy commitment of resources to basic research on fundamental scientific issues.

The economic benefits of understanding the role of the ocean in the climate system are enormous. And accumulating evidence of man-made climate change has brought these issues to the attention of the public. These concerns coincide with recent successes in long-term weather forecasting associated with El Niño, and with advances that enable detailed measurement of climate variables. (For instance, in the last ten years, the errors in surface heat fluxes obtained from moorings have been reduced by a factor of forty so that the present uncertainty is 5 Watts per square meter.) These factors imply that climate studies will be a significant path for future research in oceanography.

The development of long-term forecasting skill raises challenging scientific problems. These include: understanding and quantifying turbulent mixing, convection, and water-mass formation and destruction; the thermohaline circulation and its coupling to the wind-driven circulation; the generation, maintenance, and destruction of climatic anomalies; climatic oscillations and the extratropical coupling of the ocean and atmosphere on seasonal, decadal, and inter-decadal timescales; and the physics of exchange processes between the ocean and the atmosphere. All these problems are of fundamental scientific and practical importance.

Will there be substantial progress on these issues during the next decade? Many physical oceanographers have already begun an enthusiastic frontal assault under the banner of CLIVAR. It is likely that the economic issues that surround global change and climate prediction will motivate continued financial support from society. If people and money are what counts, then we have every reason to be optimistic.

The problem of global climate prediction is the most difficult that our field has encountered. Unlike equatorial oceanography and El Niño, there is not going to be a theory based on linear waveguide dynamics that decisively identifies timescales and cohesively binds oceanography and meteorology. Further, the decadal timescale of extratropical dynamics means that scientists see only a few realizations of the system within their own lifetime. This is bad for morale, but even worse, we cannot wait to gather enough data to reliably verify the different predictions of climate models. Could meteorologists have developed daily weather prediction models if these scientists saw only three or four independent realizations of the system in a lifetime? The only way around this statistical problem is to expand our data base and frame hypotheses about past climate change and ocean circulation using paleoceanographic studies. An important challenge is to test the dynamical consistency of these hypotheses.

An emerging theme, which is strongly related to climate, is the ocean's role in the hydrologic cycle. New satellite technologies promise to measure sea surface salinity and precipitation. These, coupled with improvements in the computation of evaporation via indirect methods, will improve our picture of the freshwater flux in the oceans. The freshwater sphere is an encompassing topic that spans oceanography, the atmospheric sciences, polar ice dynamics, and hydrology. Our knowledge of the oceanic freshwater source-sink distribution is far poorer than our knowledge of the source-sink distribution of heat. Yet salinity and temperature contend in their joint effect on the density of seawater and in their influence on the ocean circulation, and the climate system. Knowledge of freshwater input from continents, precipitation, and sea-ice is poor. Observational techniques addressing these issues (for example, the use of oxygen isotopes, and tritium/helium to diagnose freshwater sources) herald progress.

Coupled with improved estimates of the freshwater sources at the surface, will be an increased understanding of water-mass dynamics and transformations. We can look for advancement on such fundamental issues as the causes of the temperature-salinity relationship, thermocline maintenance, and interhemispheric water-mass exchanges.

We will see explosive development of new observational tools, such as those used by the TOPEX/POSEIDON