A technique is described for preparing detailed sea surface temperature analyses for large ocean areas. These analyses utilize injection tempera ture observations taken by commercial ships. The inadequacies of analyses based on averaged data and some difficulties inherent in contouring scalar fields are discussed. Sea surface temperatures are interpreted according to some concepts derived from cross-sectional profiles and surface current data. Isotach analyses of mean current drift are considered as flow pat terns to aid temperature analysis in areas where data are sparse. (Author).

A series of tests was conducted to determine data requirements for synoptic sea surface temperature analyses. Random plots of various concentrations of perfect temperature data were analyzed manually and by computer techniques to establish analysis accuracies as a function of the quantity of data input. A random error function was utilized to introduce errors in the temperature data + or - 1 to 7 F in accuracy and + or - 12 miles in positioning. Analyses of various quantities and qualities of input data show that the present data input would have to be increased by a factor of 5 or 6 to ensure reliable analyses of a complex thermal area. By using improved instrumentation, such as the airborne radiation thermometer (ART), buoys, and the near-surface reference temperature device (NSRT), the requirements for data can be halved. That is, 25 to 30 observations per day per 5-degree square of ocean would be required for areas like the Gulf Stream. For less complex areas, such as the Sargasso Sea, only 6 to 8 reports per day are required for the same size area, assuming the observations are taken by reliable instrumentation. Ship injection temperatures (not NSRT) in any quantity will not provide the desired reliability of analysis in these areas. ART surveys are highly recommended for complex areas where large quantities of fairly reliable data are required. These data would be wasted in less complex areas, where a few buoys of high accuracy can provide sufficient information. (Author).

The primary purpose of this research was to perform an analysis of the constancy of the sea-surface temperature in candidate locations for OTEC siting. Most in situ data available to determine the constancy of the sea surface temperature were obtained using point sensors that recorded surface temperature at specific points and at specific times or for some specific period. These data, while establishing certain gross or average patterns, are inadequate for describing short-term temperature variations for large areas and are insufficient for assessing the constancy of the temperature patterns relative to selecting potential sites for OTEC. The only data available for analysis of the synoptic ocean surface temperature pattern are those obtained in the last five years by infrared scanning radiometer systems on earth observational satellites. For this study, the satellite infrared (ir) data for the period 1973-1977 were used to identify and locate major thermal fronts at the sea surface. These data were then used to glean certain statistics on these fronts. In situ data were used, when available, to determine temperature difference across specified thermal fronts. This study was performed in three regional locations of interest to OTEC. The specific regions are: Region I, Eastern Gulf of Mexico (bounded by 22 N to 30 N and 91 W to the west coast of Florida including the Straits of Florida); Region II, the east coast of Florida (bounded by 23 N to 32 N and the east coast of Florida to 79 W); and Region III, Puerto Rico and vicinity, (bounded by 17 to 19 N, 64 to 68 W). Results are presented and discussed.