Ocean-Atmosphere Observations Supporting Short-Term Climate Predictions (1994)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

2 Panel's Purpose and Background TERMS OF REFERENCE The Terms of Reference are listed in Chapter I. To explain how the panel approached its task and to clarify what it saw as its restricted piece of the larger climate observation problem, it is useful to elaborate certain of these terms. Ocean-Atmosphere Observational System Despite being a panel for "operational ocean observations," we were clearly meant to consider a "global coupled ocean-atmosphere observational system." Taking up the atmospheric part of this assignment has confronted us with the large operational meteorological network devoted primarily to weather forecasting. Recommendation 2 above and further discussion be- low stem from this facet of the charge. Appropriate Scales We were asked to consider "appropriate space and time scales" and to build "upon the observational programs utilized in support of the Tropical Ocean and Global Atmosphere (TOGA) program." The first TOGA scien- tific objective (World Climate Research ProgrammelfOGA Scientific Steering Group, 1985) contains a crucial scale goal: 15 

16 OCEAN-ATMOSPHERE OBSERVATIONS EXTRATROPICAL ATMOSPHERE- OCEAN ANOMALIES sP\-\ERIC CIRC(JLA ~ "''loi~-,.c""~-,oo~~ ~,,TROPICAL ,)'J ... "'o~ \RAINFALl'} / JC.~ -~l!l!,·=l!! - •· Ill··1 ~RF~E ~N~S . . . . . . . . . . . . . :ft • • • , • • • • ,. llilii 1 -- • • • • • • • • • • ,. . . . . . . . . . . . . ._, •• , ==t> -=eo ~ SEA-SURFACE ~ TEMPERATURE FIGURE 1 Schematic of ocean-atmosphere interactions that lead to the time-depen- dent character of the climate system (NRC, 1990). "To gain a description of the tropical oceans and the global atmosphere as a time-dependent system, in order to determine the extent to which this system is predictable on time scales of months to years, and to understand the mechanisms and processes underlying this predictability." From this selection of time scales there follows a focus on the upper ocean (above the thermocline) and lower atmosphere (troposphere), as illustrated in Figure 1. The panel has retained this focus as its working definition of a vertically "appropriate scale." Nothing in TOGA research to date suggests that the deep ocean plays a first-order role in climate fluctuations of either the atmosphere or the upper ocean on the seasonal to interannual scale, although it is undoubtedly important in climate variations of longer periods. Discerning and then predicting changes in the deep ocean on a variety of long time scales are certainly important scientific and longer-term predic- tion problems, but they are ones that generally lie outside "appropriate scales" in the present context. Observational Systems Required for Climate Prediction The observational system is to be one "required for climate prediction." This suggests a way to guide choices about observations to add or observa- 

PANEL 'S PURPOSE AND BACKGROUND 17 tions to curtail, as noted in Recommendation 4. At the present time, the discipline is mostly potential. Indeed, predictions that now exist are experi- mental, or possibly quasi-operational, but largely under research auspices. The definition of observational requirements for truly operational predic- tions can be expected to remain in a state of some flux until such time as such predictions are regularly performed by a center such as the Interna- tional Research Institute for Climate Prediction. In asking how extensive an observing system is "required," one faces the fact that the ocean observing system as a whole is extraordinarily mod- est, with only rare examples of time series that could be considered long relative to the climatic variations of interest. Almost none of the observing systems can yet be shown to be unnecessary or redundant for prediction, and the use of prediction models to make such demonstrations of redun- dancy or nonredundancy has barely begun. But in the future the prediction standard does offer a consistent way to judge the value of reducing, con- tinuing, or extending a particular measurement system. For an existing system, the skill of predictions made with different proportions of the total data set withheld can be assessed. For a proposed system, similar assess- ments can be made using simulated data-an Observing System Simulation Experiment (OSSE). Caution is needed in applying the "prediction" standard too rigorously. Our present rudimentary models may lend themselves to misleading OSSEs, as discussed by the Provisional Working Group ( 1991). Important physical processes may be poorly parameterized or left out of the prediction scheme altogether. Fixing these faults might do more to improve predictions than any adjustments of the measurement system already incorporated into the prediction schemes and will generally require focused programs (process studies) having observation requirements and strategies different from (and often more intensive than) schemes founded in the needs of operational predictions alone. Development of Data Sets Required for the Use of Current and Anticipated Prediction Models This phrase further motivates our lenient approach to enforcement of the prediction standard. It is true that several TOGA data streams are not currently incorporated into experimental or quasi-operational predictions. It is far less clear that they need not be or never will be, and indeed some of the strongest advocacy for extending one such data stream (the TOGA Tropical Atmosphere Ocean array) has come from members of the modeling predic- tion community who are looking ahead to the use of the data (see below). Thus, for purposes of "anticipated prediction models," the continuation of these observations is important.