to the structure and dynamics of phytoplankton ecology. Also, the variability of phytoplankton communities is thought to hold a key to understanding the relative importance of physical and biological factors in structuring the marine food web. In addition, there is evidence that the successful modeling of phytoplankton dynamics, and the predictive linkage of phytoplankton production to higher trophic levels, has so far been limited by a lack of synoptic data and by limited sampling strategies. A fundamental problem in marine ecology is to establish both the spatial and the temporal scales in which fundamental physical and biological processes occur and to sample the environment accordingly. Ships, aircraft, and satellites provide alternative, and complementary, strategies for sampling the environment. For example, if chlorophyll concentration, as an index of phytoplankton biomass, is the variable under investigation then ship, aircraft, and satellite "platforms" offer the opportunity to obtain diverse, and often mutually exclusive, experimental information. Shipboard data provide continuity with conventional oceanographic research techniques, can be relatively accurate, can include both vertical and horizontal measurements, but are comparatively limited in both space and time. Chlorophyll data from aircraft systems provide rapid spatial coverage of regional areas, can include both vertical and long-track measurements, can be relatively precise (however, accuracies are the subject of ongoing research), but are limited by the logistics of aircraft, and provide linear (as distinct from areal) coverage. Satellite chlorophyll imagery can provide worldwide coverage of cloud-free areas, can provide repeated routine coverage of regional areas (including those areas that are far from our oceanographic research institutions), but are relatively less accurate without concurrent ship or aircraft data, are limited by cloud coverage, and require more complex image and data processing. The key point is that the living. marine resources are unlikely to be assessed adequately without the synoptic perspective, the quantitative areal data, and the quasi-continuous temporal coverage provided by remote sensors. Some early use of the Nimbus 7 color images has shown very promising application to the studies of the food web and to illuminating the relationships between the planktonic distribution and the development of young fish. For example, off the California coast, such information has been used effectively to study plankton distribution and the distribution of anchovy spawning. More detailed studies of these kinds would clearly be important contributions to biological oceanography. Therefore, NASA's objectives with respect to ocean color scanners should include: • A program concerned with global marine ecology. Many nations are competing for various marine resources and as yet there is no sensible perspective, or even a coherent body of information, with which to judge conflicting interests and the magnitude and resilience of the resources. The "ecological boundaries" of the oceans are global in extent and the major regions of productivity widespread yet relatively small in area. These limited, yet critically important, areas are at least partially accessible to satellite observations and study. NOAA has some statutory responsibility here for U.S. waters, and a coordinated effort by NASA and NOAA would have an important long-term pay-off. Effective data and information transfer from the existing CZCS system to interested researchers and those concerned with living marine resource management. Selective utilization of the already existing data base can be an invaluable and cost-effective method for optimizing the accuracy of the information, resource assessment strategies, and the characteristics of future remote sensing systems. Development of an improved ocean color scanner (and associated system) that will allow for: additional bands providing further spectral information on the ocean and atmosphere for use in improved data processing algorithms; increased sensor signal-to-noise ratio for improved color gradient detection; accurate sea-surface-temperature images that are coregistered with those from the color scanner for obtaining both temperature and chlorophyll information with a minimum of data processing; operational capability, both in performance and with respect to availability of data. 2.2.4.2 Ocean physical applications The color scanner images available now also reveal some very exciting potential for broader oceanographic applications. For atmospheric scientists and operational meteorologists, cloud patterns and their time and space evolution represent unique signals of important atmospheric phenomena. The very first application of satellite observations to meteorology was the interpretation of cloud images, qualitative to be sure, but extremely important in delineating areas of important activity and in identifying what kinds of motions are involved. Cloud imagery still plays an important role in depicting in an easily interpreted way some of the important dynamics of the atmosphere, and we use carefully selected clouds as the only way, at the moment, of tracing atmospheric motions (wind) in some regions. Multispectral infrared and color remote sensing techniques are not well utilized at present, partly because of the lack of data and the required display/ analysis facilities. The few data that are available suggest that combined use of such techniques can greatly facilitate investigation of frontal structure variability on both the fine scale and the mesoscale. Figure 4 illustrates one such technique with visible (color) data from an area south of Georges Bank. The frontal features on the lineal front across the center of the scenes are observed to be displaced between successive scenes. Such displacements in ocean color can be interpreted as tracers of the horizontal flow field near the front. An unanswered question is whether the "motions" observed arise from advection, or from propagation of wavelike phenomena along the front. In either case we observe characteristics on time and space scales which are related to the frontal variability. These are very difficult to measure in any other way. Figure 5 is a composite of frontal outlines digitized from the scenes in Figure 4. Displacements in both the front delineating the eddy (a warm-core ring) and the front near the 200 m isobath can be noted. The eastward displacement of disturbances in the shelf break front is consistent with anticyclonic circulation of water about the warm-core ring. Ring streamers show similar behavior. The temporal variability seen in the ring frontal locus is interpretable as arising from rotation of the two-dimensional horizontal modal structure or as an alternating growth and decay. Other ring observations indicate that rotary motion is the correct interpretation. Consequently, using the color imagery, we can diagnose ring horizontal modal structure and rotation rate. A second example, from the Somali Current (shown in Figure 6) maps the evolution of a multiple separating coastal circulation into a single gyre system. The Somali Current is classically described as a low-latitude equator-crossing western boundary current which joins with a large eddy ("Great Whirl"), separates from the East African Coast (10° 12°N), and flows into the western Arabian Sea. Figure 6 illustrates the evolution of associated frontal structures for 1976 through 1979 as observed by satellite. The classical description is found to be applicable only in the later stages of the Indian Monsoon, while a multiple frontal separation is more often found early in the Indian Monsoon. Frontal translation rates range from 5 to 150 cm-1 over extents greater than 300 km. Such large-scale frontal variability cannot be observed from ships, and in fact was unknown until observed by satellite. These two examples illustrate two very different cases where satellite observation can make a significant input into our understanding of oceanic processes. Infrared and other emission-based thermal monitoring systems view a very thin layer on the ocean surface (~1 mm). Color scanners, on the other hand, sense to tens of meters into the water column. Both methods can be used independently to study frontal variability; however, little has been done to study differences in simultaneous observations made with the two methods; work (Mueller and LaViolette, 1981) on Grand Banks has shown strikingly different frontal patterns in the thermal and color scenes. It is suspected that it will be the subtle differences in structure functions observed in such multispectral measurements that |