SCOPE 35 - Scales and Global Change

ABSTRACT

Primary production in the sea accounts each year for at least 30 percent of the global plant fixation of carbon and is thus an integral part of major biogeochemical fluxes, such as the N₂ or CO₂ cycles. However, uncertainties exist in the magnitude of annual uptake of both carbon and nitrogen in the ocean, as well as in the regional or seasonal variability of these processes. The highly productive continental shelves of the world, where man's nutrient loadings are discharged, where 95 percent of the world's fishery is yielded, and where most of the organic carbon sink of atmospheric CO₂ occurs, may actually be two or three times more productive than presently estimated. Such uncertainties result from our past inabilities to accurately specify temporal and spatial inhomogeneities of phytoplankton biomass and associated productivities within shelf waters. As a result of the launch of the Nimbus- 7 Coastal Zone Color Scanner (CZCS) in October 1978 and the subsequent progress with data analysis, it is now possible to determine ocean chlorophyll concentrations from space to better than ±30 percent for values of 0-10 µg chl l^-1 in Case-1 waters (little sediment or humic matter). On a cloudless planet, calculations of net annual shelf primary production could be made from the local change with time of phytoplankton biomass as measured by the CZCS. Time aliasing by clouds of such a satellite-derived data set can be overcome with simulation models that incorporate in situ data, rate parameters of biological processes, and the spatially-rich, but sparse CZCS data sets in time. Significant improvement in the estimates of shelf productivity, and in understanding the controlling mechanisms for their temporal and spatial variations, can be achieved when satellite measurements of ocean colour are combined appropriately with in situ observations. Such a capability is clearly needed if we are to understand biological variability of the sea in response to discrete or climatic changes, such as the impact of anthropogenic nutrient inputs on coastal fisheries, and the effects of fossil-fuel CO₂ loading to the atmosphere in terms of global habitability.

INTRODUCTION

The potential of the carbon cycle of the sea to either yield fish or store atmospheric CO₂ is a subject of continuing controversy as man's ability to modify the marine environment increases. Because the actual amount of CO₂ fixed annually during marine photosynthesis is unknown, the fate of phytoplankton, serving as a precursor either to fish carbon or to sediment carbon, is also unknown. Debates over the amount of potential fish harvest (Ryther , 1969; Alversen et al., 1970) and CO₂ storage capacity (Broecker et al., 1979; Walsh et al., 1981) of the ocean thus hinge on the amount and fate of marine primary production. Current estimates of annual marine primary production range from 20 to 55 X 10⁹ tons of carbon per year (Ryther, 1969; Koblentz- Mishke et al., 1970; Platt and Subba Rao, 1975; DeVooys, 1979; Walsh, 1980). This range accounts for = 25 to 50 percent of the total net global carbon fixation, assuming a terrestrial primary production of -55 X 10⁹ tons C yr^-l (Woodwell et al., 1978). This possible range in marine primary productivity is thus 400 to 1000 percent of present fossil fuel emissions of 5.2 x 10⁹ tons Cyr^-l

Over the last century, man's ability to extract nitrogen from the atmosphere has also begun to rival that of N₂ fixation by plants. Between 1950 and 1975, world production of agricultural fertilizers increased tenfold. Anthropogenic nutrient input from agrarian runoff , deforestation, and urban sewage has already impacted local streams and ponds, some large lakes, major rivers, and perhaps even the continental shelves, with a tenfold increase of nutrient loading since 1850 (Walsh, 1984). Nitrogen is now only routinely measured in 25 percent of the world's 240 largest rivers, however, and few biological time series are available to document the coastal zone's past response to fluvial nutrient transients on even a decadal time scale.

The annual primary production of the Dutch Waddensea, for example, has apparently increased threefold between 1950 and 1970 (Postma, 1978) and presumably other shelves have responded to anthropogenic nutrient loadings as well. Rigorous analyses of past measurements of ocean colour by satellite over the last 5 years are now required to provide an adequate departure point for long-term chlorophyll time series to assess the fate of phytoplankton carbon and nitrogen biomass, as well as their productivity in the sea. With the present Nimbus- 7 satellite information and that from a follow-on sensor, one can begin to address serious problems of overfishing today, as well as the future and perhaps more ominous consequences of the above linked activities -man's accelerated extraction of nitrogen from the atmosphere and addition of carbon to the atmosphere.

Large areas of the ocean, such as the central gyres, have relatively low rates of production per unit surface area (Figure 14.1), but account for a major fraction of total carbon fixation because of their large areal extent (Table 14.1). In contrast, highly productive coastal and upwelling regions account for only 10 percent of the ocean by area and probably 25 percent of the ocean primary productivity. The coastal zooplankton populations (Figure 14.2) provide the basis for more than 95 percent of the world's estimated fishery yield, however, and a large part of the proposed organic carbon sink of the atmospheric CO₂ is located in adjacent slope sediments (Figure 14.3). These various ocean provinces exhibit pronounced differences in their phytoplankton species assemblages as the evolutionary consequence of their physical habitat. They also have significant differences in spatial and temporal variability of algal biomass as a function of nutrient input and grazing losses (Walsh, 1976), with perhaps very different fates of the fixed carbon (Walsh et al., 1981; Walsh, 1983): oxidation in the deep sea and burial along the continental margins.

There are two basic reasons for the uncertainty in the estimates of marine carbon fixation; both are of equal importance. First, the methodology used to estimate the rate of primary productivity (the ¹⁴C method) may be in serious error (Geiskes et al., 1979; Eppley, 1980). Secondly, the highly productive shelf regions exhibit a much wider range of spatial and temporal variability of biomass than the open ocean, on scales which have been very poorly sampled by classical shipboard programmes. It is in the oligotrophic (gyre) regions, where the biomass variability is not pronounced, that the methodology errors are greatest. This is because the oceanic phytoplankton are thought to be more sensitive to stresses related to capture and prolonged enclosure. However, the long food chains and the 90 percent recycling processes of the offshore regime provide insignificant fish harvest (Ryther, 1969) and little net biotic storage of CO₂ (Eppley and Peterson, 1980). 

In the coastal regions where phytoplankton productivity and zooplankton biomass are much higher (Figs 14.1 and 14.2), the results of the ¹⁴C methodology are probably more representative of the actual rate, but the spatial extent and the temporal character of the algal biomass fields are poorly known. For example, within 30 km off the Peru coast, the surface chlorophyll estimate of phytoplankton biomass ranges from 0.4 to 40.0 µg chl l^-1 and the integrated primary production from < 1 to > 10 g C m^-2 day^-1 (Walsh et al., 1980). Classical shipboard programmes consist of productivity measurements made once a day over a small spatial area, limited by the usual 10-12 knot speed of the research vessel and aliased within the above gradients (Walsh et al., 1987a). At present, other shipboard rate measurements, such as nutrient uptake, respiration, grazing, excretion, and sinking, can only be made at a few points of the sea as well, to be later multiplied by some inadequate estimate of mean biomass in order to calculate fluxes of elements within the marine food web.

Figure 14.1 The global distribution of phytoplankton primary production (mg C m^-2 day^-1) in five categories of > 500, 250-500, 150-250, 100-150, and <100 (after Koblentz-Mishke et al., 1970). (Reproduced with permission.)

Table 14.1 Aquatic photosynthesis and nitrogen fixation in relation to losses of sediment storage of organic carbon, of denitrification, and of methane production (after Walsh, 1984)

Figure 14.2 The global distribution of zooplankton abundance (mg m^-3) over the upper 100m of the water column in four categories of >500, 201-500, 51-200, and <50 (after Bogorov et al., 1968). (Reproduced with permission.)

Figure 14.3 The global distribution of organic carbon (% dw) within surface sediments in five categories of > 2.00, 1.01-2.00, 0.51-1.00, 0.25-0.50, and < 0.25 (after Premuzic et al., 1982). (Reprinted with permission from Organ. Geochem., 4, 1982, Pergamon Journals Ltd.)

However, phytoplankton species on the continental shelves can divide every 0.5 to 2 days; without significant losses, an algal population during the spring bloom could increase at the same rate. To resolve the temporal and spatial consequences of this resultant growth process, a Nyquist sampling frequency of at least 0.25 day^-l is required by sampling theory (Blackman and Tukey, 1957). If one sampled every 4 hours in a typical longshore upwelling flow regime of 30 cm sec^-l to resolve this process, at least 5 ships would be required every 20 km² for the necessary biomass measurements (Kelley, 1976). We thus arrive at a major reason for our analysis of CZCS colour data-the need to analyze distributions of biological properties at frequencies which can resolve causally the sources of their variance.

Approximately 10 percent of the annual shelf production (0.5 X 10⁹ tons C yr^-l ) is thought to be sequestered as organic carbon deposits (Table 14.1) on adjacent continental slopes (Walsh et al., 1981). Although the anthropogenic input of nitrogen to the shelves may have increased tenfold over the last 50 to 100 years, a sufficient time series of phytoplankton data is not available to accurately specify changes in primary productivity or shelf export to continental slopes. This lack of a proper spatial and temporal perspective has hindered our understanding and, therefore, our ability to make accurate estimates of coastal productivity and subsequent carbon and nitrogen fluxes to the rest of the food web. Understanding the coastal ecosystem processes has far greater significance than their areal extent or contribution to total marine carbon fixation (Table 14.1) would suggest because:

(a) The fate of carbon and nitrogen fixed in these highly productive shelf regions is quite different from the oceanic areas of the sea, sinking to slope depocentres instead of being grazed within the water column (Walsh et al., 1981; Eppley and Peterson, 1980), and
(b) Impacts of human activity are greater in the coastal region. Thus, there' is a strong motivation to obtain, for the first time, an analysis of synoptic biomass information, coupled with rate process data, required to study these highly dynamic oceanographic regions over longer periods at annual and decadal time scales in addition to the much higher Nyquist frequency for resolution of the basic biological processes.

SATELLITE IMAGERY

Preliminary investigations undertaken in the 1960s by Clarke, Ewing, Lorenzen, Yentsch, and others provided evidence that the quality of light reflected from the sea surface and remotely sensed by aircraft instrumentation might be interpreted as phytoplankton biomass, i.e. chlorophyll, in the upper portion of the water column. These workers (e.g. Clarke et al., 1970) were limited by their equipment to an altitude of 3 km. However, even at that altitude, the influence of the atmospheric backscatter was quite obvious as it began to dominate the colour signal reflected from the ocean surface. This raised the question of whether the rather poorly reflected ocean could be sensed through the entire atmosphere from a spacecraft, and if the contributions of the Rayleigh backscatter and aerosol backscatter could be effectively removed from the signal seen by a spacecraft. Additional NASA supported studies in 1971 and 1972 with Lear Jet and U-2 aircraft and a rapid scan spectrometer at altitudes of 14.9 and 19.8 km, demonstrated that this concept could be used to develop spacecraft equipment for the purpose of estimating ocean water column chlorophyll from earth satellites. This became possible through the realization that problems associated with the scattering properties of the atmosphere, as well as direct reflectance of the sun from the sea surface (glint), could be either avoided or corrected (Hovis and Leung, 1977).

The first satellite-borne ocean colour sensor, the Coastal Zone Color Scanner (CZCS), was launched aboard Nimbus-7 in October 1978 with four visible and two infrared (one of which is thermal) bands, allowing a sensitivity) about 60 times that of the Landsat-1 multispectral scanner. With failure of the sensor, transmission of CZCS data finally ceased in December 1986. Unlike many satellite sensors of ocean properties, the CZCS responded to more than the features of the mere surface of the sea and was sensitive to algal pigment concentrations in the upper 20 to 30 percent of the euphotic zone (Hovis et al., 1980; Gordon et al., 1980). The CZCS was specifically designed to detect upwelling radiance in spectral bands selected for the purpose of detecting variations in the concentrations of phytoplankton pigments. The theoretical and experimental techniques for describing the bio-optical state of ocean waters and its relationship to optical parameters that can be remotely sensed have been discussed by a number of workers (Morel and Prieur, 1977; Smith and Baker, 1978a, b; Gordon and Morel, 1983).

Simply stated, the CZCS radiance data can be utilized to estimate ocean chlorophyll concentrations by detecting shifts in sea colour, particularly in oceanic waters. Clear open ocean waters have low chlorophyll concentrations (0.01-1.0 µg chll^-l) and the solar radiation reflected from the upper layers of these waters is blue; conversely, waters with high concentrations of chlorophyll (> 1.0  µg chll^-l) are green (Morel and Smith, 1974). It has been demonstrated that this change in ocean colour can now provide a quantitative estimate of chlorophyll concentration (Gordon and Clark, 1980; Smith and Baker, 1982) for oceanic regions with an accuracy of 0.3 to 0.5 log C (where C is the chlorophyll concentration). The initial comparisons between CZCS imagery and surface pigments measured continuously along ship tracks carried out by Gordon et al. (1983) and Smith and Wilson (1981) had suggested that C could be retrieved from the imagery to within about a factor of two. Smith and Baker (1982) and Gordon et al. (1982) have shown, however, that accuracies on the order of ± 30 percent in C are possible for Morel's Case-waters (Morel and Prieur, 1977), i.e. areas of little sediment or humic matter within the water column.

Figure 14.4 The availability of CZCS data from relatively cloud-free regions (shaded areas) during the 14 orbits of the Nimbus-7 satellite on 28 February 1979 (after OAO, 1979).

We are now in a position to systematically exploit the rich CZCS data base, obtained somewhere on the world shelves (Figure 14.4) each day over the last 5 years. Figure 14.5 of the CZCS-derived (SASC, 1984) chlorophyll distribution during 28 February 1979 within the Mid-Atlantic Bight is an example of the information available from one of the many satellite images taken daily during the Nimbus-7 orbits around the planet (Figure 14.4). The chlorophyll isopleths of Figure 14.5 were contoured from a grid of-4500 data points between the 10 to 3000 m isobaths. Each grid point was 5 nautical miles apart and each chlorophyll value was thus the mean of-100 observations since the pixel resolution of the CZCS is-800 m. The cross-isobath and parabathic gradients of this synoptic CZCS-derived chlorophyll field (Figure 14.5) are 'typical' of composites of chlorophyll data taken at sea during the same season between 1975 and 1984 (Walsh et al., 1978; O'Reilly and Busch, 1984; Walsh et al., 1987a). On a cloudless planet, the relevance of these aliased shipboard observations could simply be addressed by assembling a daily time series of CZCS data, computing the local rate of change of phytoplankton at 4500 to 450000 grid points, and unravelling the rate processes responsible for the amount of algal biomass left behind in the upper water column.

Figure 14.5 The distribution of satellite-derived chlorophyll ( µg l^-1) over a grid of 4500 pixels, 5 nautical miles apart, within the Mid-Atlantic Bight on 28 February 1979  

Figure14.6 The CZCS estimate of chlorophyll distibution during a) 10 April 1979, b) 17 April 1979, c) 19 April 1979, and d) 21 April 1979 (after Walsh et al., 1987a). (Reprinted with permission from Deep-Sea Res., 34, 1987, Pergamon Journals Ltd.)

Unfortunately, it has not been possible to obtain CZCS measurements of the global oceans on anything close to a daily basis. On any given day (Figure 14.4), a major fraction of our watery planet is obscured by clouds. A qualitative estimate of realizable CZCS sampling characteristics was gleaned by the NASA Ocean Color Science Working Group (Walsh et al., 1982) from screening a few time sequences of CZCS data for which regular sampling was attempted. This experience suggests that in a month of data collection, useful satellite data can be obtained on several days within randomly distributed clear-sky domains which are a few hundred km in extent, less frequently > 1000 km in extent.

Of the nominal 2 hours of Nimbus- 7 CZCS coverage taken and recorded each day, an average of approximately 30-40 percent was rejected and not processed due to total cloud cover (no significant open water areas). Further- more, other data gaps for a particular site are derived from the inability of one satellite to provide sufficient daily overlap in swath width during the 14 orbits (Figure 14.4). As a result, only 45 useful CZCS images were available for the Mid-Atlantic Bight between 1 January and 30 June 1979, for example; the 25 percent data recovery was also not grouped in equal time increments. We present the results of 4 CZCS images taken on 10 April, 17 April, 19 April, .and 21 April 1979 as a representative time series (Figure 14.6) of the types of data sets that are likely to be available from some continental shelves.

Our experience to date suggests that global CZCS coverage would yield, on average, between 10 (at the equator) and 20 (at 40 degrees N) usable images per month. The upper estimate represents a mean sampling interval of about every 1.5 days, for a given 1000 km x 1000 km boreal ocean domain, with the majority of usable data in patchy subscenes of typically a few hundred kilometres in extent, excepting an occasional clear view of most of the domain in one image. Coverage frequencies, however, fluctuate seasonally (and regionally) around these nominal estimates; coverage gaps of 2 to 3 weeks are likely to occur several times per year, with less frequent gaps of longer duration. In winter, low sun elevations will cause sampling voids of several weeks to a few months (increasing with latitude) at latitudes above 40 degrees.

Such sampling constraints mean that the CZCS data sets violate stationary assumptions, i.e. time invariant probability density functions over the time domain of interest, inherent in most statistical approaches to time series analyses of phytoplankton processes. Clearly, the present data base collected with the Nimbus- 7 CZCS is thus inadequate to be directly applied to the global mapping of primary productivity except in a qualitative sense. It is limited both in terms of sampling frequency and in terms of concurrent oceanographic experimental data necessary to bridge the interpretive gap from phytoplankton pigment distributions to net primary production by statistical methods.

Adequate data do exist in certain shelf regions, however, to develop a modelling methodology for a future global productivity assessment programme, utilizing the present and follow-on CZCS-type sensors. We present the preliminary results of simulation models of the time-dependent flow field of continental shelf waters in the Mid-Atlantic Bight to accompany the satellite data set of April 1979. Such numerical analyses will eventually allow us to interpolate between the available CZCS images with sufficient accuracy that future changes of shelf ecosystems, induced by continued overfishing or discharge of anthropogenic nutrients, can be detected on other shelves as well, allowing realistic estimates of the ocean's role in the future habitability of the planet.

1979 SATELLITE TIME SERIES

Phytoplankton standing stock as indicated by CZCS estimates of pigment concentrations at a fixed time, within a particular spatial pattern, are the result of a complicated set of biological, chemical, and physical processes with time scales ranging from seconds to seasonal, and space scales ranging from global to microscopic. The shapes and locations of chlorophyll patterns delineating the synoptic-scale shelf features are dominated, however, by physical- dynamical processes, transporting the plankton populations left in the water column as the net result of birth and death processes. Therefore, mesoscale spatial patterns (10 < x < 100 km) tend to evolve over time scales ranging from several hours of phytoplankton cell division to a few days, i.e. at the time scale of a wind event (Walsh, 1976), while the synoptic-scale patterns (100 < x < 1000 km) tend to evolve over time scales of the order of a few weeks to a month, i. e. above the Nyquist frequency of the probable CZCS sampling time scale. Considering these separate biological and physical scales and the CZCS sampling characteristics together, we chose (Walsh et al. , 1987c) to simulate water motion with respect to concurrent CZCS images over 3 weeks during 5-25 April 1979 (Figure 14.7).

Figure 14.7 The 1979 depth-averaged velocity field on the Mid-Atlantic shelf during a) 5-12 April, b) 12-16 April, c) 16-20 April, and d) 20-25 April (after Walsh et al., 1987a). (Reprinted with permission from Deep-Sea Res., 34, 1987 Pergamon Journas Ltd.)

Over the last 50 years, estimation of marine primary productivity has involved some type of mathematical model, ranging from analytical process models of photosynthesis through ecosystem simulation with complex, coupled biological-physical numerical models. These models all shared a common parameterization of primary production as some function of an initial concentration of phytoplankton biomass and the regulation of photo- synthesis by light and nutrients. For example, the relation between vertical mixing, light intensity, and phytoplankton growth was quantified as a critical depth concept (Sverdrup, 1953), below which the 24 hr algal respiration of the water column exceeds the integrated daily photosynthesis. This critical depth was h_c=0.2 I₀(kI_c)^-1, where I₀ is the incident radiation, k is the extinction coefficient, and Ic is the compensation light intensity at which algal photosynthesis equals respiration ( ~0.3 ly h^-1). Sverdrup's concept was that if hc < h, the depth to which the phytoplankton are mixed as a result of wind and/or tidal stirring, no bloom would occur even in the presence of high nutrient content. At the beginning of March off New York at the 60 m isobath, the incident radiation is <250 g cal cm^-2day^-1 and h_c is 53 m within the well-mixed 60 m water column. Accordingly, high chlorophyll was not found at the 60 m isobath on 28 February 1979, but an order of magnitude more algal biomass was observed by the CZCS at depths £ 20 m (Figure 14.5).

The utility of the more sophisticated recent models (Platt et al., 1977) in predicting regional primary productivity is still largely hindered by the current meagre knowledge on the coupling of physical dynamics with biological processes on the appropriate time and space scales. Models of marine processes, whether physical, chemical, geological, or biological, are, in fact, inadequate theoretical constructs attempting to describe an incompletely known dynamic balance. Models thus lead to specific data acquisition and analysis which eventually suggest rejection of the original model, or hypothesis, in a cyclic process of oceanographic research (Walsh, 1972). Early ecological models of coastal upwelling (Walsh and Dugdale, 1971; Walsh, 1975; Walsh and Howe, 1976; Wroblewski, 1977), for example, described reasonably well the nutrient uptake and growth processes of phytoplankton, but their loss processes were poorly parameterized by unknown lateral export, grazing, and sinking terms (Walsh, 1983). With the synoptic CZCS data sets, we can begin to systematically place bounds on these loss terms by modelling the spatial changes of phytoplankton over discrete time intervals for comparison with successive CZCS images.

Rapid resuspension, offshore transport, and sinking events of phytoplankton can be inferred, for example, from the CZCS time series in April 1979 (Figure 14.6). During 17-19 Apri11979, 10 ships provided ground-truth chlorophyll measurements in the Mid-Atlantic Bight (Figure 14.8) as part of the LAMPEX experiment for calibration of aircraft and satellite overflights (Thomas, 1981). The R/V Kelez was occupying stations in the apex of the New York Bight during 17-19 April 1979, such that most of the chlorophyll observations on these days were taken within 25 km of the coast, south of Hudson Canyon. The rest of the Kelez cruise track (Figure 14.8) was performed after 19 Apri11979, such that the shipboard chlorophyll composite (Figure 14.9) is that of nearshore waters during the 17 April (Figure 14.6b) and 19 April (Figure 14.6c) CZCS overflights, but that of offshore waters during the 21 April overflight (Figure 14.6d). The high chlorophyll concentrations of the coastal zone ( < 20 m depth) and the low chlorophyll at the 60-100 m isobaths, southeast of Nantucket Island, measured aboard ships on 17-19 April (Figure 14.9) match quite well these regions of the two CZCS images on the same days (Figures 14.6b and 14.6c). However, the major mid-shelf resuspension of near-bottom chlorophyll, south of Delaware Bay, on 19 April (Figure 14.6c) and subsequent sinking by 21 April (Figure 14.6d) remained undetected (Figure 14.9) by the conventional shipboard surveys, i.e. the research vessels were then not in those regions.

Figure 14.8 Cruise tracks of the R/V Kelez, Advance II, Julius Nelson, Onrust, Kyma, Pathfinder, Shang Wheeler, Sub Sig II, Short Snort, and Lady Donna during 17-26 April 1979 (after Walsh et al., 1987a). (Reproduced with permission from Deep-Sea Res., 34, 1987, Pergamon Journals Ltd.)

The 10 April 1979 CZCS image of chlorophyll (Figure 14.6a) exhibits a decline of algal biomass with distance offshore during mean northwest wind forcing, i.e. from a direction of 296° True, of 1.07 dynes cm^-2 over 5-12 April 1979 (Figure 14.7a). The mean currents of the circulation model then (Figure 14.7a) are £ 5 cm sec^-1, with offshore flow between the 20-40 m isobaths and southwesterly flow between the 40-60 m isobaths at mid-shelf, south of the Hudson Canyon and north of Norfolk. In response to such northwest wind events, surface waters are pushed offshore, and the predominantly westward alongshore flow is slowed down (Beardsley and Butman, 1974). An upwelling circulation pattern is created, in which surface phytoplankton can be advected offshore and dissolved nutrients can be returned within subsurface waters to the shelf, providing the source of the next algal growth cycle (Walsh et al., 1978).

During a northeast wind event, the westward flow is instead intensified, however, and weak onshore flow usually occurs at the surface, with offshore flow of subsurface water (Beardsley et al. , 1983). Under a wind forcing of only 0.31 dynes cm^-2 from the northeast (068° T), the mean flow of the model's water column during 12-16 April 1979 was > 10 cm sec^-1 to the southwest over most of the shelf, except for offshore flows south of Long Island, near the Hudson Canyon, off Delaware Bay, and south of Norfolk (Figure 14.7b). During 17 Apri11979, tongues or streamers of 1-2 ug chl l^-1, in fact, extended within the CZCS image (Figure 14.6b) from the shelf to slope waters in these areas of the shelf, south of Long Island, south of New Jersey, off Delaware Bay, and off Norfolk, as previously observed within a March CZCS time'series (Walsh et al., 1987a). Similar to the northwest wind forcing, it appears that northeast wind forcing could also move algal cells in some shelf areas from the coastal zone to slope waters (Figure 14.7b). 

Following this northeast wind event, another mean wind forcing from the northwest (341° T) occurred during 16-20 April 1979 (Figure 14.7c), but with half the intensity (0.58 dynes cm^-2) of the first period (Figure 14.7a). In response to this shift in wind forcing, the mean alongshore flow is now weaker north of the Hudson Canyon and stronger south of Delaware Bay during 16-20 April, compared to the 12-16 April time period (Figure 14.7b). In the model, there was little or no offshore flow south of Long Island and New Jersey, but continued offshore movement of water occurred off Delaware Bay and Norfolk (Figure 14.7c). In contrast to observations of ~1 µg chl l^-1 on April 17 at mid-shelf within the latter region (Figure 14.6b), two days later , more than 16 µg chl l^-1 was detected by the CZCS on 19 Apri11979, from off Norfolk to off Cape Hatteras, 150 km south along the 40-60 m iosbaths (Figure 14.6c). At a population growth rate of only one doubling every two days (Walsh et al., 1987a), the order of magnitude change in algal biomass measured by the CZCS cannot all be attributed to in situ growth of phytoplankton; resuspension of phytoplankton, uneaten and sunk out of the water column, is a likely source term.

Figure 14.9 A chlorophyll (µg ug l^-1) composite of the distribution of phytoplankton biomass measured aboard ships during 17-26 April 1979 (after Walsh et al., 1987a). (Reproduced with permission from Deep-Sea Res., 34, 1987, Pergamon Journals Ltd.)

Few near-bottom chlorophyll data are available for April 1979 over the Mid-Atlantic Bight. During April 1984, however, near-bottom chlorophyll concentrations of > 25 µg chl l^-1 were found at the 60 m isobath (Figure 14.10). These data are similar to our observations of 9-17 µg chl l^-1 above the 21-63 m isobaths in April 1978, 10-12 µg chl l^-1 above the 15-35 m isobaths in April 1980, 8-14 µg chl l^-1 above the 21-49 m isobaths in April 1981, 10-29 µg chl l^-1 above the 17-39 m isobaths in April 1982, and 9- 15 µg chl l^-1 above the 13-43 m isobaths in April 1983. An accumulated, near-bottom chlorophyll concentration of 30 µg chl l^-1 within the lower 10 m of the water column at the 40 m isobath, and 0.5 µg chl l^-1 within the upper 30 m before a resuspension event, would yield 7.88 µg chl l^-1 after vertical homogenization in response to such a sequence of wind events (Figures 14.7b and 14.7c). A doubling of such a phytoplankton population after two days would then yield the estimated CZCS chlorophyll concentration of -16 µg chl l^-1 seen on 19 April 1979 (Figure 14.6c).

With southerly wind forcing, surface flow is offshore and to the east, reversing the predominantly westward currents if a storm is of sufficient intensity. During 20-25 April 1979, a mean wind forcing of 0.32 dynes cm^-2 from the south (174° T) was sufficient to drive weak mean currents ( < 5 cm sec^-1 ) to the northeast within the 20-50 m isobaths (Figure 14.7d), in contrast to the three previous flow fields of the model (Figures 14.7a-c).Within this simulated circulation pattern, which should lead to upwelling again, at least on the inner shelf like the case of 5-12 April 1979 (Figure 14.7a), the observed chlorophyll concentrations decline by an order of magnitude within 2 days, i.e. by 21 April 1979 (Figure 14.6d). Such a tenfold decline in near-surface chlorophyll concentrations between 19 April (Figure 14.6c) and 21 April (Figure 14.6d) implies rapid sinking and/or down- welling of the phytoplankton, since grazing stress removes less than 10 percent of the early spring bloom (Walsh et al., 1978).

Figure 14.10 The near-bottom chlorophyll distribution during 1-5 April 1984 as measured aboard the R/V Oceanus (after Walsh et al., 1987b). (Reproduced with permission from Cont. Shelf Res., in press, Pergamon Journals Ltd.)

Figure 14.11 The continuous, vertical distribution of chlorophyll fluorescence off the coast of Long Island during 4-8 April 1984 (after Walsh et al., 1987b). (Reproduced with permission from Cont. Shelf Res., in press, Pergamon Journals Ltd.)

No continuous fluorescence data with depth are available for Apri11979, but such a chlorophyll time series was obtained during 4-8 April 1984 (Figure 14.11) at the 60 m isobath, south of Long Island. It suggests a downward displacement of about 30 m day^-1 (3 x 10^-2 cm sec^-1) on 7-8 April 1984 (Walsh et al., 1987b). Furthermore, the vertically integrated biomass of phytoplankton, in suspension during this period at the 60 m isobath, was only 54-72 percent of that previously sunk out on 4 April 1984 (Figure 14.11), i.e. loss of algal biomass has occurred during this 8 Apri11984 resuspension event. After March-April wind transport events off Peru, the integrated chlorophyll biomass over the upper 40 m similarly decreased by 10-25 percent in 1976 and 50- 75 percent in 1977 (Walsh et al., 1980). We now attempt to quantify such a net offshore loss, or export, of phytoplankton biomass from the 60 m isobath, in terms of 1984 carbon fluxes to adjacent slope depocentres (Walsh et al., 1987b).

1984 MOORED FLUOROMETER TIME SERIES

Twice as much near-surface chlorophyll was found within the satellite images during the 19 April 1979 resuspension event (Figure 14.6c) above the 40-60 m isobaths, compared to a previous 21 March 1979 event detected by the CZCS above the 20-40 m isobaths (Walsh et al., 1987a). Such a temporal sequence over ~1 month implies both seasonal build up of chlorophyll within the aphotic zone and a gradual transfer of uneaten phytodetritus seaward. A 30-40 km offshore migration of the algal resuspension area, from the 20-40 m isobaths to the 40-60 m isobaths, within-30 days from 21 March to 19 April 1979, suggests a mean net seaward movement of algal particles of 1-1.3 km day^-1 ( ~1-1.3 cm sec^-1 ). Since much of the wind forcing is from the north in the Mid-Atlantic Bight during February-April, with more frequent northwest storms from off the North American continent, the average flow of the upper 30 m of the water column past 4 current meter arrays, between the 45-105 m isobaths off Martha's Vineyard (Beardsley et al., 1983), was a net 7.73 cm sec^-1 to the west and a net 1.43 cm sec^-1 offshore during February-April1980 and Apri1 1979. At a rapid sinking rate of 30 m day^-1 , the phytoplankton of the spring bloom would remain in the euphotic zone ( ~30 m depth), however, for one day until the next wind resuspension event.

During northeast transport events, offshore transport also occurs, but near the bottom; during April 1979 and February-April 1980, the mean subsurface flow (>30 m) was offshore at 1.98 cm sec^-1 past moorings at the 66-88 m isobaths, south of Martha's Vineyard (Beardsley et al., 1983). Vertical decomposition of the model's flow field during the northeast wind forcing event of 12-16 April 1979 also leads to offshore flow in the bottom layer, i.e. > 45 m depth (Figure 14.12). In fact, time series of the u and v components (UCMP, VCMP) of horizontal flow and chlorophyll from current meter and moored fluorometer observations tethered 3-5 m off the bottom at the 80 m , isobath off Long Island 5 years later during 17 February- 7 April 1984, i.e.seaward of the time series of vertical fluorescence (Figure 14.11), yield a mean offshore flow of 1.0 cm sec^-1 and a net seaward chlorophyll transport, up, of about 3.0 ng chl cm^-2sec^-1 (Figure 14.13).

Using another mooring of current meters and fluorometers, 12 km farther offshore at the 120 m isobath, the average chlorophyll flux at 80 m depth near the shelf-break, during February-April 1984, can then be calculated (Walsh et al., 1987b). We obtain 0.35-0.47 g C m^-2 day^-l, or 35-47 g C m^-2 over at least the 100-day period of the spring bloom, assuming a C/Chl ratio of 45/I and a bottom boundary layer of 30-40 m thickness. How long might such a shelf export continue to exit the Mid-Atlantic Bight in a combination of both offshore flow events at the surface, detected by satellites, and offshore flow events at the bottom, detected by moored instruments?

Figure 14.12 Vertical decomposition of the flow of the water column at the surface, at 20 m, and at 45 m under northeasterly wind forcing during 12-16 April 1979 (after Walsh el al., 1987a). (Reproduced with permission from Deep-Sea Res., 34, 1987, Pergamon Journals Ltd.)

Figure 14.13 A time series of currents, chlorophyll, and phytoplankton transport past a mooring 3-5 m above the 80 m isobath, south of Long Island, during 17 February- 7 April 1984 (after Walsh et al., 1987a). (Reproduced with permission from Deep-Sea Res., 34, 1987, Pergamon Journals Ltd.)

During October-May, the mean diabathic component of the surface flow (0-30 m) between the 45-105 m isobaths, south of Martha's Vineyard, was 1.77 cm sec^-1 offshore in 1979-80 (Beardsley et al., 1983); weak onshore flow of 0.05-0.20 cm sec^-1 occurred in the surface during June-August. An offshore flow of ~1.5 cm sec^-1 over the upper 30 m of the water column across the -1000 km length of the shelf -break between Cape Hatteras and Martha's Vineyard implies a net, seaward water transport of 0.45 Sverdrups (10⁶ m³sec^-1) from the Mid-Atlantic Bight. Assuming a CZCS-derived horizontal chlorophyll gradient of 2.94  µg chl l^-1on the shelf and 0.44  µg chl l^-1 within slope waters (Figures 14.6a-d), a C/chl ratio of 45/1 ;and a surface water flux of 0.45 Sv across the shelf break, a daily export of 0.44 X 10¹⁰ g C day^-1 might occur.

Over a year and the area of the adjacent Mid-Atlantic continental slope (-4 x 10¹⁰ m²), at least 41.0 gCm^-2yr^-1 might be imported to this continental slope, based on this CZCS estimate. Over the area of world slopes (Table 14.1), such an annual flux would amount to 1.3 X 10⁹ tons C yr^-1. Recent estimates of ²¹⁰pb and ¹⁴C rates of sediment mixing on the Mid- Atlantic slope, of vertical carbon gradients within the upper 10 cm of this sediment, and of anthropogenic nitrogen loading to the coastal zone suggest an accumulation rate of 9.9-16.7 gCm^-2yr^-1 on the Mid-Atlantic slope f(Walsh et al., 1985). This is only 24-41 percent of the estimated import, suggesting a possible world slope accumulation of at least 0.3-0.5 x 10⁹ tons C yr^-1 (Table 14.1), with perhaps the difference lost to oxidation in the water column and sediments.

To examine the long-term implications of a combined two-layered offshore transport of the spring bloom during the resuspension events detected by the CZCS time series (Figure 14.6) during April 1979 and by the FTD time series (Figure 14.11) and the moored fluorometer time series (Figure 14.13) in April 1984, we analysed our winter-spring shipboard data base, taken over ten years. The chlorophyll observations at each ship station (Figure 14.14) were subdivided into those from surface (0-20 m) and subsurface (20-50 m or 20- 75 m) waters on the middle shelf (20-50 m isobaths), the outer shelf (50-100 m isobaths), and the upper slope (100-1000 m isobaths). Although these station data were from different years (1973-82), each cruise was of sufficient duration (1-2 weeks) in February-May to capture the biological response to wind events at a frequency of 0.2 day^-1. We have plotted the envelope of maximum chlorophyll concentration encountered by the ship each Julian day (Figure 14.14), for all the cruise data. This analysis represents the statistical ensemble of resuspension events the CZCS would be likely to detect, as maximum surface chlorophyll, over a decade of infrequent, non-stationary time series (Figure 14.6).

Maximal shipboard chlorophyll concentrations as high as 16 µg chl l^-1 and as low as 4 µg chl l^-1 were found, similar to the CZCS images during April 1979, with the peaks separated in time by perhaps 5 days (Figure 14.14). The slanted lines of Figure 14.14 are thus drawn 5 days apart and suggest a seaward translation of maximal chlorophyll concentrations, from mid-shelf to the slope, in response to wind forcing of the same frequency. If the algal populations of the surface maxima (0-20 m) on the mid-shelf, outer shelf, and upper slope become the subsurface chlorophyll maxima (20-46 m) 5 Julian days later in time, a sinking velocity of at least 4-5 m day^-1 is implied. If the diatom populations of the mid-shelf spring bloom are translated 60-80 km across the the shelf to become part of the phytoplankton populations (i.e. including daughter cells produced en route) of the slope water 40 Julian days later in time, an offshore velocity of at least 1.5-2.0 cm sec^-1 is implied.

Such inferred vertical and horizontal diplacement rates of the chlorophyll ensemble over 10 years (Figure 14.14) are not inconsistent with estimates of sinking and offshore advection rates obtained from CZCS images, shipboard measurements and moored in situ data, during the 1979 and 1984 spring blooms. Such horizontal exchange rates also provide reasonable carbon budgets for export of particulate matter from the shelf. As additional analyses of spatially synoptic satellite data provide more time series on algal biomass changes in shelf and slope waters, we will be able to specify the interannual and possible long-term changes of shelf export from both the Mid-Atlantic Bight and other shelves. These data will provide a basis to refine our present understanding of global biogeochemical cycles between their oceanic, terrestrial, and atmospheric reservoirs (Walsh, 1984).

Figure 14.14 A chlorophyll (µg l^-1) composite of the vertical distribution of phytoplankton biomass across the shelf and slope of Mid-Atlantic Bight from over 50 cruises taken during February-May 1973-1982.

ACKNOWLEDGEMENTS

This research was sponsored by NASA for reduction of the czcs data, by DOE for acquisition of the shipboard rate data, and by NOAA for providing biomass data and the use of their research vessels.

REFERENCES

Alversen, D. L., Longhurst, A. R., and Gulland, J. A. (1970). How much food from the sea? Science, 168, 503-505.

Beardsley, R. C., and Butman, B. (1974). Circulation on the New England continental shelf: Response to strong winter storms. Geophysical Research Letters, 1, 181-184.

Beardsley, R. C., Mills, C. A., Vermersch, J. A., Brown, W. S., Pettigrew, N., Irish, J ., Ramp, S., Schlitz, R., and Butman, B. (1983). Nantucket Shoals Flux Experiment (NSFE 79) Part 2: Moored Array Data Report, pp. 1-140. WHOI-83-13.

Blackman, R. B., and Tukey, J. W. (1957). The Measurement of Power Spectra, pp. 1-190. Dover Press.

Bogorov, V. G., Vinogradov, M. D., Varonina, N. M., Kanaeva, I. P., and Suetova, I. A. (1968). Distribution of zooplankton biomass within the surficial layer of the world ocean. Dokl. Akad. Nauk. USSR, 182, 1205-1207.

Broecker, W. S., Takahashi, T ., Simpson, H. J ., and Peng, T. H. (1979). Fate of fossil fuel carbon dioxide and the global carbon budget. Science, 206, 409-418.

Clarke, G. K., Ewing, G. C., and Lorenzen, C. J. (1970). Spectra of backscattered light from the sea obtained from aircraft as a measure of chlorophyll concentration. Science, 167, 1119-1121.

DeVooys, C. G. N. (1979). Primary production in aquatic environments. In Bolin, B., Degens, E. T., Kempe, S., and Ketner, P. (Eds.) The Global Carbon Cycle, pp. 259-292. Wiley, New York.

Eppley, R. W. (1980). Estimating phytoplankton growth rates in the central oligotrophic oceans. In Falkowski, P. G. (Ed.) Primary Productivity in the Sea, pp. 231-242. Plenum, New York.

Eppley, R. W. and Peterson, B. J. (1980). Particulate organic matter flux and planktonic new production in the deep ocean. Nature, 282, 677-680.

Geiskes, W. W. C., Kraay, G. W., and Baars, M. A. (1979). Current ¹⁴C methods for measuring primary production: gross underestimates in oceanic waters. Neth. J. Sea Res., 13, 58-78.

Gordon, H. R. and Clark, D. K. (1980). Remote sensing optical properties of a stratified ocean: an improved interpretation. Appl. Optics, 19, 3428-3430.

Gordon, H. R., Clark, D. K., Mueller, J. L., and Hovis, W. A. (1980). Phytoplankton pigments derived from the Nimbus- 7 CZCS: initial comparisons with surface measurements. Science, 210, 63-66.

Gordon, H. R., Clark, D. K., Brown, J. W., Brown, 0. B., and Evans, R. H. (1982). Satellite measurement of the phytoplankton pigment concentration in the surface waters of a warm core Gulf Stream ring. J. Mar. Res., 40, 491-502.

Gordon, H. R., and Morel, A. G. (1983). Remote Assessment of Ocean Color for Interpretation of Satellite Visible Imagery, pp. 1-114. Springer-Verlag.

Gordon, H. R., Clark, D. K., Brown, J. W., Brown, 0. B., Evans, R. H., and Broenkow, W. W. (1983). Phytoplankton pigment concentration in the Middle Atlantic Bight: comparison of ship determinations and CZCS estimates. Applied Optics, 22, 20-36.

Hovis, W. A. and Leung, K. C. (1977). Remote sensing of ocean color. Opt. Engineer., 16, 158-164.

Hovis, W. A., Clark, D. K., Anderson, F., Austin, R. W., Wilson, W. H., Baker, E. T., Ball, B., Gordon, H. R., Meuller, J. L., EI-Sayed, S. Z., Sturm, B., Wrigley, R. C., and Yentsch, C. S. (1980). Nimbus-7 Coastal Zone Color Scanner systems: description and initial imagery. Science, 210,60-63.

Kelley, J. C. (1976). Sampling the sea. In Cushing, D. H., Walsh, J. J. (Eds.) Ecology of the Seas, pp. 361-387. Blackwell, Oxford.

Koblentz-Mishke, 0. J., Valkovsinky, V. V., and Kabanova, J. C. (1970). Plankton primary production of the world ocean. In Wooster, W. S. (Ed.) Scientific Exploration of the South Pacific, pp. 189-193. Nat. Acad. Sci., Washington, DC.

Morel, A., and Smith, R. C. (1974). Relation between total quanta and total energy for aquatic photosynthesis. Limnol. Oceanogr., 19,591-600.

Morel, A., and Prieur, L. (1977). Analysis of variations in ocean color. Limnol. Oceanogr., 22, 708- 722.

OAO Corporation. (1979). The Nimbus-7 CZCS data catalog: February 1-28, 1979, pp. 1-59. Goddard Space Flight Center Contract NASS-25937.

O'Reilly, J. E. and Busch, D. A. (1984). Phytoplankton primary production on the northwestern shelf. Rapp. P.-v. Reun. Cons. Int. Explor. Mer, 183, 255-268.

Platt, T. and Subba Rao, D. V. (1975). Primary production of marine microphytes. In Cooper, J. P. (Ed.) Photosynthesis and Productivity in Different Environments, pp. 249-80. Cambridge Univ. Press.

Platt, T., Denman, K. L., and Jassby, A. D. (1977). Modelling the productivity of phytoplankton. In Goldberg, E. D., McCave, I. N., O'Brien, J. J., and Steele, J. H. (Eds.) The Sea, Vol. 6, pp. 807-856. Wiley Interscience, New York.

Postma, H. (1978). The nutrient contents of North Sea water: changes in recent years, particularly in the Southern Bight. Rapp. P.-v. Reun. Cons. Int. Explor. Mer, 172, 350-352.

Premuzic, E. T., Benkovitz, C. M., Gaffney, J. S., and Walsh, J. J. (1982). The nature and distribution of organic matter in the surface sediments of world oceans and seas. Organ. Geochem., 4,63-77.

Ryther, J. H. (1969). Photosynthesis and fish production in the sea. Science, 166, 72-76.

Smith, R. C. and Baker, K. S. (1978a). The bio-optical state of ocean waters and remote sensing. Limnol. Oceanogr., 23, 247-259.

Smith, R. C., and Baker, K. S. (1978b). Optical classification of natural waters. Limnol. Oceanogr., 23, 260-267.

Smith, R. C., and Wilson, W. H. (1981). Ship and satellite bio-optical research in the California Bight. In Gower, J, R. F. (Ed.) Oceanographyfrom Space, pp. 281-294. Plenum Press, New York.

Smith, R. C., and Baker, K. S. (1982). Oceanic chlorophyll concentrations as determined using Coastal Zone Color Scanner imagery. Mar. BioI., 66, 269-279.

Sverdrup, H. U. (1953). On conditions for the vernal blooming of phytoplankton. J. Cons. Int. Explor. Mer, 18, 287-295.

Systems and Applied Sciences Corporation. (1984). Users Guide for the Coastal Zone Color Scanner Compressed Earth Gridded Data Sets of the Northeast Coast of the United States for February 28 through May 27, 1979, pp. 1-15. SASC-T-5-5100- 0002-008-84.

Thomas, J. P. (1981). Large Area Marine Productivity-Pollution Experiments (LAMPEX)-A series of studies being developed to hasten the operational use of remote sensing for living marine resources and environmental quality. In Gower, J . F. R. (Ed.) Oceanography from Space, pp. 403-409. Plenum Press, New York.

Walsh, J. J. (1972). Implications of a systems approach to oceanography. Science, 196, 969-975.

Walsh, J. J. (1975). A spatial simulation model of the Peru upwelling ecosystem. Deep-Sea Res., 22,201-236.

Walsh, J. J. (1976). Herbivory as a factor in patterns of nutrient utilization in the sea. Limnol. Oceanogr., 21, 1-13.

Walsh, J. J. (1980). Concluding remarks: marine photosynthesis and the global cycle. In Falkowski, P. G. (Ed.) Primary Productivity in the Sea, pp. 497-506. Plenum Press, New York.

Walsh, J. J. (1983). Death in the sea: enigmatic phytoplankton losses. Prog. Oceanogr., 12, 1-86.

Walsh, J. J. (1984). The role of the ocean biota in accelerated global ecological cycles: a temporal view. Bioscience, 34, 499-507.

Walsh, J. J., and Dugdale, R. C. (1971). A simulation model for the nitrogen flow in the Peruvian upwelling system. lnv. Pesq., 35, 309-330.

Walsh, J. J ., and Howe, S. 0. (1976). Protein from the sea: a comparison of the simulated nitrogen and carbon productivity in the Peru upwelling system. In Pat ten, B. C. (Ed.) Systems Analysis and Simu/ation in Ecology, Vol. IV, pp. 47-61. Academic Press, New York.

Walsh, J. J., Whitledge, T. E., Barvenik, F. W., Wirick, C. D., Howe, S. 0., Esaias, W. E., and Scott, J. T. (1978). Wind events and food chain dynamics within the New York Bight. Limno/. Oceanogr., 23, 659-683.

Walsh, J. J., Whitledge, T. E., Esaias, W. E., Smith, R. L., Huntsman, S. A., Santander, H., and DeMendiola, B. R. (1980). The spawning habitat of the Peruvian anchovy, Engrau/is ringens. Deep-Sea Res., 27, 1-27.

Walsh, J. J., Rowe, G. T., Iverson, R. L., and McRoy, C. P. (1981). Biological export of shelf carbon is a neglected sink of the global CO₂ cycle. Nature, 291, 196-201.

Walsh, J. J., Barnes, W. B., Brown, 0. B., Carder, K. L., Clark, D. K., Esaias, W. E., Gordon, H. R., Holyer, R. C., Hovis, W. A., Kirk, R. J., Lasker, R., McCarthy, J. J., McElroy, M. A., Mueller, J. L., Perry, M. J., and Smith, R. C. (1982). The MAREX (Marine Resources Experiment) Program. A report of the Ocean Color Science Working Group, pp. 1-102. Goddard Space Flight Center, Greenbelt, Maryland. 

Walsh, J. J., Premuzic, E. T., Gaffney, J. S., Rowe, G. T., Harbottle, G., Stoenner, R. W., Balsam, W. L., Betzer, P. R., and Macko, S. A. (1985). Organic storage of CO₂ on the continental slope off the Mid-Atlantic Bight, the Southeastern Bering Sea, and the Peru coast. Deep-Sea Res., 32(7), 853-883.

Walsh, J. J., Dieterle, D. A., and Esaias, W. E. (1987a). Satellite detection of phytoplankton export from the Mid-Atlantic Bight during the 1979 spring bloom. Deep-Sea Res. , 34, 675- 703.

Walsh, J. J., Wirick, C. D., Pietrafesa, L. J., Whitledge, T. E., Hoge, F. E., and Swift, R. N. (1987b). High frequency sampling of the 1984 spring bloom within the Mid-Atlantic Bight: synoptic shipboard aircraft, and in situ perspectives of the SEEP-I experiment. Cont. She/f Res. (in press).

Walsh, J. J ., Dieterle, D. A., and Meyers, M. B. (1987c). A simulation analysis of the fate of phytoplankton within the Mid-Atlantic Bight. Cont. Shelf Res. (in press).

Woodwell, G. M., Whittaker, R. H., Reiners, W. A., Likens, G. E., Delwiche, C. C., and Botkin, D. B. (1978). The biota and the world carbon budget. Science, 199, 144-146.

Wroblewski, J. S. (1977). A model of phytoplankton plume formation during variable Oregon upwelling. J. Mar. Res., 35, 357-394.