program and sponsoring agencies should establish (with input from the community) priorities for moving long-time series and other observations initiated by the program into operational mode. Factors to be considered include data quality, length, number of variables, space and time resolution, accessibility for the wider community, and relevance to established goals.
The vast majority of the scientific capabilities represented by the major oceanographic programs are centered around observational platforms, laboratory facilities, and computers. Gaps in this infrastructure can translate into limitations in each program's ability to achieve the greatest possible scientific return on the substantial investment made in these efforts.
The capabilities of the fleet need to be matched to the scientific needs of the oceanographic community. UNOLS, in concert with the ocean science community and funding agencies, has developed and continues to update plans to maintain this fleet (UNOLS, 1990; UNOLS, 1995; Betzer et al., 1995). Over the past 20 years, the number of ships in the UNOLS fleet has declined slightly, with 31 ships in 1978 and 28 ships in 1997. The useful life of a research vessel is about 30 years, and older vessels have been replaced whenever possible. This is reflected in the present UNOLS fleet, which contains only 12 vessels that were in the fleet in 1978 and only one vessel that was constructed prior to 1970.
An example of this planning process is the development in the 1980s of plans to construct three large research vessels to replace aging large ships to be retired in the 1990s (R/V Thompson, Washington, and Atlantis II). The planned new vessels were to be available for the present generation of U.S. major oceanographic programs and were designed to meet those needs. The gestation period for an oceanographic vessel is 6 to 10 years; thus, the vessels that were initiated in the mid-1980s have come on-line in the past few years. The construction of these vessels was underwritten by the Office of Naval Research, consistent with plans to pursue "blue water" oceanography in the 1990s. The first of this AGOR-23 class of vessel, R/V Thompson, became available for scientific research in 1991. The other two, R/V Revelle and R/V Atlantis, became available in 1996 and 1997, respectively. The R/V Ron Brown also came on-line and, although operated by NOAA, is considered part of the UNOLS fleet.
At present, the major oceanographic programs are in the midst of a period of declining need for ships (see Table 5-1). The trend illustrates the problem of the planning for resources that require very long lead times and that have relatively long lives in comparison with the activities of the science community and its funding. As recently as the early 1990s, it was projected that there would be a