MARINE BIOLUMINESCENCE AND UPPER OCEAN PHYSICS : SEASONAL CHANGES IN THE NORTHEAST ATLANTIC

A N OCCASIONAL form of nighttime entertainment at sea, perhaps during a transit, after the work of a particular leg is finished, is to spend time away from the ship's lights and look at the sea. Many of us have gone up to the bow (or descended into the bow dome) and, after allowing time for pupil adaptation, watched the light show. The "show," however, also illuminates one of the more intriguing problems in biological oceanography: bioluminescence. The wheres and whys of this phenomenon remain little known, despite its apparent ubiquity throughout the marine realm. Although it has seemed merely a curiosity, such a common behavioral phenomenon must have implications for life in the sea. Studies of bioluminescence within an oceanographic context began only in the last several years. Beginning in 1989, a group of oceanographers will continue the work of Biowatt (see Marra and Hartwig, 1984) in a program entitled Marine Bioluminescence and Upper Ocean Physics, or "Marine Light-Mixed Layers" (ML-ML). The point of view of ML-ML is that predictions of the spatial and temporal variability of bioluminescence are best achieved through better knowledge of oceanographic processes, from the forces driving the mixed layer to the waxing and waning of plankton populations. There are two major reasons for this approach. First, due to their large size, observed bioluminescent signals can be associated with zooplankton (Marra and Hal-twig, 1984; Swift et al., 1985a; Swift et al., 1985b). Although not strictly true, this is a useful simplification. Thus, the seasonal cycle of heterotrophic populations, and the predictions regarding secondary production, will be an important first step in predicting bioluminescence in the ocean. The second reason has to do with how biolumines-

AN OCCASIONAL form of nighttime entertainment at sea, perhaps during a transit, after the work of a particular leg is finished, is to spend time away from the ship's lights and look at the sea.Many of us have gone up to the bow (or descended into the bow dome) and, after allowing time for pupil adaptation, watched the light show.
The "show," however, also illuminates one of the more intriguing problems in biological oceanography: bioluminescence.The wheres and whys of this phenomenon remain little known, despite its apparent ubiquity throughout the marine realm.Although it has seemed merely a curiosity, such a common behavioral phenomenon must have implications for life in the sea.
Studies of bioluminescence within an oceanographic context began only in the last several years.Beginning in 1989, a group of oceanographers will continue the work of Biowatt (see Marra and Hartwig, 1984) in a program entitled Marine Bioluminescence and Upper Ocean Physics, or "Marine Light-Mixed Layers" (ML-ML).The point of view of ML-ML is that predictions of the spatial and temporal variability of bioluminescence are best achieved through better knowledge of oceanographic processes, from the forces driving the mixed layer to the waxing and waning of plankton populations.
There are two major reasons for this approach.First, due to their large size, observed bioluminescent signals can be associated with zooplankton (Marra and Hal-twig, 1984;Swift et al., 1985a;Swift et al., 1985b).Although not strictly true, this is a useful simplification.Thus, the seasonal cycle of heterotrophic populations, and the predictions regarding secondary production, will be an important first step in predicting bioluminescence in the ocean.The second reason has to do with how biolumines-John Marra, Lamont-Doherty Geological Observatory of Columbia University, Palisades, NY 10964.cence is measured.Bioluminescent light flashes commonly arise when an organism senses a disturbance in the water.Since measurements of these light flashes involve disturbing the local environment of an organism, the measurements are of bioluminescent potential.Obviously, this presents difficulty, since the strength of the disturbance affects the measured signal (natural levels of bioluminescence have not been quantified).Also, the elicited signal will depend upon the sizes of the organisms involved, and certain measuring devices may not detect the dim signal from otherwise abundant forms.For these reasons, establishing which organisms exhibit a bioluminescent response under what conditions is as important as measuring the emitted photons of a bioluminescent flash.
Both of the major oceanographic processes considered in ML-ML, mixed layer dynamics and (bioluminescent) plankton production, can have a strong seasonal component.That is, changes in mixed layer depth result from successive wintertime storms and low solar irradiance, followed by calmer winds and increased solar heating in the summer.Concordantly, the seasonal changes in mixed layer depth cause immense changes in plankton biomass, the envelope of which encompasses perhaps sharper changes in individual populations.ML-ML will focus on seasonal changes in upper layer physics and the successive populations that are responsible for bioluminescence.
Where in the open ocean are we likely to observe large signals in the seasonal cycle?Early in the planning for ML-ML, it was decided that the northeast Atlantic would be an optimal locale.From various oceanographic atlases (e.g., Robinson et al., 1979: Isemer andHasse, 1987), it is known that during spring the mixed layer shoals from several hundred to a few tens of meters.Signals in the plankton are correspondingly large.Some have viewed the vernal increase in phytoplankton in the North Atlantic, as seen from satellite (Esaias et al., 1986), as one of the largest transient signals on the planet (Ducklow, 1989).Bioluminescent potential also shows a large signal (E.Swift, personal communication), although its importance seasonally is not known.(59°N, 19°W).ATis the temperature difference between T and T125. Figure re-drawn from Lambert and Hebenstreit (1984).
As a result of planning meetings thoughout 1988, a committee of scientists settled on a site in the vicinity of 59°N, 21°W as suitable for our studies.This location was selected for several reasons; for example, logistical access, availability of historical data (e.g., Colebrook, 1984;Lambert and Hebenstreit, 1985), water depth, and distance from the subarctic front.The general oceanographic setting of the ML-ML site is cold, cloudy, windy and given to mesoscale variability.Water temperatures range from 9-13°C.On the average, three-fourths of the sky is cloud-covered.Seasonally-averaged wind speeds range from 8 m s-m in the summer to 18 m s-~ in winter.The North Atlantic Current, with its extensive field of meanders and eddies (Krauss and Meincke, 1985), flows through the area.

Synopsis of the Seasonal Cycle
The January-March period in the northeast Atlantic is characterized by a net loss of heat from the ocean to the atmosphere, resulting in convective cooling of the water column.Estimates of the heat loss average about 200 W m -2 in January (Isemer and Hasse, 1987).This magnitude can be appreciated, as pointed out by Stommel (1979), by comparison with the solar flux at 50°N latitude in June, under cloudless skies, of 375 W m-L Convective cooling during January-March produces deep mixed layers of several hundred meters (Robinson et al., 1979).These depths are difficult to specify more exactly because the mixed layer is difficult to define, and because the actual physical process producing the deepening is uncertain.Also apparent in the charts in Robinson et al. (1979) is an advective heat input from the Gulf Stream and North Atlantic Current.The advective heat input is necessary to balance the vertical heat flux.During the January-March period, the low water column irradiance prevents phytoplankton growth, and deep mixing establishes the total new nutrients available to the ecosystem as it develops through later warming and periods of low vertical exchange.Although measurements in winter have not been published, phytoplankton populations are expected to be very low; data from the Coastal Zone Color Scanner suggest values in the range of 0.1-0.2mg m 3 (W.Esaias, personal communication).Although we know very little of over-wintering populations of zooplankton, they might consist of species such as those of the genus Metridia, whose life cycle is less tuned to the spring phytoplankton increase, or large carnivorous forms such as Euchaeta and Pleuromamma.
By April, the water column begins to stratify (Fig. 1) and the winds begin to subside.The first increases in chlorophyll have been noted in the data at this time (Williams and Hopkins, 1974), probably dominated by early forms such as Skeletonema costatum, Thalassiosira spp.and Chaetoceros spp.May is typically the month of the phytoplankton bloom at this latitude, and chlorophyll values can be as high as 2 -5 mg m -3.This is also the time when Calanus finmarchicus and C. helgolandicus ascend to the surface for maturation, grazing and reproduction.Heterotrophic dinoflagellates, of the genus Protoperidinium, have been observed to graze on diatoms, sometimes exclusively (Jacobsen and Anderson, 1986).
By early June, the various plankton populations characteristic of spring have reached their peak in biomass.Other planktonic forms, such as euphausiids, make their appearance (Tett, 1972;Colebrook, 1984;Williams and Lindley, 1982), as do siphonophores and ctenophores (Williams and Conway, 1981); thus begins a transition to the summertime community, in the phytoplankton commonly characterized by dinoflagellates and a few species of Rhizosolenia (Colebrook, 1984).The spring increase in phytoplankton appears to give way to greater spatial heterogeneity in chlorophyll (Strass and Woods, 1988; also seen in the data of Williams and Conway, 1981).This may be a manifestation of phytoplankton growth associated with upwelling along the borders of anticyclonic mesoscale eddies (Woods, 1988).
August is the month of greatest water column stability (Fig. 1), and both August and September are the months when the summer plankton community species (as described above) reaches its maximum.In September, there is some evidence for a fall bloom as stratification declines.By late October, although surface water temperatures are above their seasonal minimum, mixed layer depths have increased to greater than 100 m.

A Program of Study
Bioluminescence is often correlated with the abundance of heterotrophs.Thus maximum bioluminescent potential, according to the current data, should occur in August when heterotrophic populations are at their maximum.In an operational sense, we might develop a crude model for the seasonal change in bioluminescent potential based on sea-surface temperature or water column stability.This is a useful A relationship between bioluminescence and chlorophyll may not be readily apparent.
first step, but it omits several important factors.First, spatial variability in the physical fields, which are driving forces in the variability of production (e.g., Woods, 1988), with attendant time lags, will probably obscure any simple relationships with physical properties.
Second, diurnal vertical migration is an important component of bioluminescent potential (Marra and Hartwig, 1984;Swift et al.', Swift et al., 1985), and the causative factors (e.g., avoidence of predators) and environmental signals (e.g.irradiance levels, food availiabilty) for this phenomenon continue to be an ongoing research concern.One interesting facet of vertical migration in the northeast Atlantic involves Metridia lucens and Pleuromamma sp., two species we believe to be particularly important to bioluminescence in the ML-ML area.According to Roe (1984), only the females of these species migrate.Diurnal vertical migration is one of many types of interactions within and among the various plankton groups that will influence bioluminescent behavior.Although we have made significant progress in identifying these (e.g., Buskey and Swift, 1987;Swift et al.. 1988), the problem requires considerable study.
Heterotrophic dinoflagellates, including Protoperidinium, may be a significant fraction of the bioluminescent signal in the ML-ML area (Swift et al., personal communication).If so, this would simplify the understanding of bioluminescence, since, as weakly-swimming grazers, they will be expected to have a more direct relationship, temporally and spatially, to chlorophyll-a, a representative property of their prey.Some puzzles exist, however.As stated above, many species could be grazers on diatoms.But the dinoflagellate and diatom species identified by Colebrook (1984) are seasonally out of phase with one another.Lessard and Swift (1985), from isotope incorporation data, suggest that heterotrophic dinoflagellates feed on bacteria as well as phytoplankton.Thus, a relationship between bioluminescence and chlorophyll may not be readily apparent, meaning that a broader range of bio-optical indicators associated with phytoplankton will have to be investigated (Carder et al., 1989).
An interesting and unexplored area of bioluminescence is that of signal transmission.This will depend not only on the strength of the bioluminescent flash (a biological property), but also on the optical characteristics (absorption, scattering) of the water itself.Is bioluminescent potential related to the magnitude of absorption or scattering?Do the inherent optical properties affect the empathic or behaviorally-propagated (e.g., Neshyba, 1967) bioluminescent displays?
The general program for ML-ML, as currently scheduled, is as follows.For 1989, we are embarking on a pilot mooring experiment (see Dickey et al., 1986), plus cooperative field work with the JGOFS North Atlantic Bloom Experiment (Ducklow, 1989).In 1990, we are planning a cruise for the period of the spring phytoplankton bloom to access the detailed spatial variability of the bloom and to identity the major bioluminescent organisms.There will also be a collaborative cruise with the Naval Ocean Research and Development Activity (NORDA) for September of that year for further identification of organisms responsible for bioluminescence.Both of these major experiments will be done in preparation for 1991, the major field year, which will include the installation of a mooring and cruises spanning the productive seasons.