USF Libraries

The response of oceanic phytoplankton to nitrate flux in the eastern Gulf of Mexico : a simulation analysis


Material Information

The response of oceanic phytoplankton to nitrate flux in the eastern Gulf of Mexico : a simulation analysis
Physical Description:
xii, 144 leaves : ill. ; 29 cm.
Meyers, Mark B
University of South Florida
Place of Publication:
Tampa, Florida
Publication Date:


Subjects / Keywords:
Marine Phytoplankton -- Mathematical models -- Mexico, Gulf of   ( lcsh )
Primary productivity (Biology) -- Mathematical models -- Mexico, Gulf of   ( lcsh )
Nitrogen cycle -- Mathematical models -- Mexico, Gulf of   ( lcsh )
Dissertations, Academic -- Marine science -- Doctoral -- USF   ( fts )


General Note:
Thesis (Ph. D.)--University of South Florida, 1993. Includes bibliographical references (leaves 129-144)

Record Information

Source Institution:
University of South Florida
Holding Location:
Universtity of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 029495448
oclc - 29241269
usfldc doi - F51-00030
usfldc handle - f51.30
System ID:

This item is only available as the following downloads:

Full Text


Graduate School University of South Florida Tampa, Florida CERTIFICATE OF APPROVAL Ph.D. Dissertation This is to certify that the Ph.D. Dissertation of MARK B. MEYERS with a major in Marine Science has been approved by the Examining Committee on February 17, 1993as satisfactory for the dissertation requirement for the Ph.D. degree Examining Committee: Major Professor: John J. Walsh, Ph.D. Member: Eileen E. Hofmann, Ph.D. Member: Thomas L. Hopkins, Ph.D. Member: Joseph J. Torres, Ph.D. Member: Gabriel A. Vargo, Ph.D.


THE RESPONSE OF OCEANIC PHYTOPLANKTON TO NITRATE FLUX IN THE EASTERN GULF OF MEXICO: A SIMULATION ANALYSIS by MARK B. MEYERS A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Marine Science University of South Florida April 1993 Major Professor: John J. Walsh, Ph.D.


ACKNOWLEDGMENTS Major funding for this research came from the Office of Naval Research, with additional funding from the National Science Foundation and the National Aeronautics and Space Administration, in grants to Dr. John J. Walsh. Their support is appreciated. In 1991, I received the Department of Marine Science's (University of South Florida) John B. Lake Fellowship, created through the generous donations of Mr. John B. Lake and the Times Publishing Company of St. Petersburg. Their support, and that of the Department and its Faculty has been gratefully appreciated. Any scientific endeavor benefits from the generous, free exchange of ideas and comments. The members of our Ecosystems Analysis Lab, particularly Dwight A. Dieterle, J. Raymond Pribble, W. Paul Bissett, and Frank E. MullerKarger, have improved my research. The members of my advisory committee have been generous and helpful in their comments and discussions, and have sharpened my thinking. Lastly, and most appreciatively, I wish to acknowledge the abundant, generous advice and patience of Dr. John J Walsh. His knowledge and insight never failed to provoke and inspire me, on matters scientific and otherwise.


TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES ABSTRACT INTRODUCTION The New Production Paradigm for Open Ocean Ecosystems . . . . Producers, Grazers, the Microbial Loop, and the Bifurcation Model . . . . Grazing or Accumulation: Uncoupling Netplankton From Herbivory . Phytoplankton Sinking Rates Objectives . . CIRCULAT ION, HYDROGRAPHY, AND PRIMARY PRODUCTION IN THE GULF OF MEXICO . The Gulf of Mexico Loop Current Phytoplankton and Primary Production in the Gulf of Mexico METHODS One-Dimensional Model Framework . State Equations . . . Vertical Mixing . . . Solar Irradianc e and Ligh t Extinction Phytoplankton Growth and Irradiance Nitrogen Uptake . . . . Grazing L o sses and Recycling Efficiency Numerical Techniques RESULTS ... No Loop Current Loop Current Present During Spring Loop C urrent Present During Summer DISCUSSION The Model's Carbon Flow: Fate Determined at The iii iv X 1 3 8 13 13 1 5 1 8 19 2 3 37 37 40 45 47 5 0 5 3 5 6 63 66 6 7 8 3 9 8 112 First Bifurcation 114 Particulate O rganic Matter Flu x . 117 Energy A vail a ble t o Higher Trophic Levels 1 2 1 Reconciliation of Disparate Data S ets and Mod e l R esults . . . . . . 124 i


"Models and Muddles" REFERENCES ii 127 129


LIST OF TABLES Table 1. Recent primary production estimates for the Gulf of Mexico. Where the authors have done so, mean rates and annual estimates are reported, in addition to the observed ranges of daily integrated rates. Table 2. Phytoplankton b iomass (mg chl m-2 ) and primary production (g C m -2 d-1 ) across the southwestern Florida outer shelf and slope during an intrusion of the Loop Current (Yoder and Mahood, 1983). Table 3. Mid-monthly (16th day) values for wind stress (t), surface vertical eddy diffusivity (K0), mixed layer depth and noontime photosynthetically active radiation ( I0 max) used in the Gulf of Mexico ecosystem simulations (cf. Walsh, et al., 1989). Table 4. Biological parameters used in the Gulf -of Mexico ecosystem model. Table 5. Primary production from model year five for three cases. Production values (g C m-2 yr-1 ) are integrated over the entire 200 m domain. Nitrate flu xes (mmol N m -2 yr-1)are evaluated at 1 00 and 200m. Negative fluxes are upward. Table 6. Grazing loss, particulate matter export, and inorganic nitrogen fluxes from model year 5 for three cases. Grazing losses (g C m-2 yr-1 ) are integrated over the entire 200 m domain. Secondary export (g c m -2 yr-1 ) is the maximum potential e xport of feces and remains of herbivores based on recycling efficiencies used in the model. Fluxes are evaluated at 100 and 200 m for carbon (mg m -2 yr-1 ) and nitrogen (mmol m -2 yr-1). Negative fluxe s are upward. iii 27 34 39 42 68 80


LIST OF FIGURES Figure 1. The relationship of primary production by phytoplankton (P) and nitrogen cycling at the lowest trophic levels in a marine food web. New production is based upon the uptake of allochthonous sources of nitrogen, represented here by the upward flux of nitrate. Recycled production is based upon recycled, or autochthonous, sources of nitrogen, represented here as ammonium excreted by zooplankton (Z) The downward flux of particulate organic matter (POM) can include fresh phytodetritus, zooplankton fecal pellets, and body remains. This simplified model does not consider dissolved organic matter fluxes, nor lateral processes. Figure 2. The bifurcation model of marine trophic dynamics (modified from Legendre and Le Fevre, 1989). Figure 3. The Gulf of Mexico. Regions A and B are the central Gulf and west Florida margin locations of model output by Walsh, et al. (1989), for comparison with the present study, which focused on Region B. Region C i s a 200x200 km2 box which was subsampled from a CZCS time series (Muller-Karger, et al., 1990). Figure 4. The depths of the 22C isotherm and the N03 isopleth in the eastern Gulf of Mexico during November 1976 (from Walsh, et al., 1989). Figure 5. Sea-surface temperature (C) signature of the Loop Current during spring 1983 showing a sequence o f cold-core frontal eddies along the west Florida ma rgin, culminating i n a large c o l d perturbation o f f the Dry Tortugas, i v 4 10 17 20


Florida in mid-April (modified from Vukovich, 1986). Figure 6. A schematic diagram of the onedimensional model domain, illustrating the mechanism used to simulate eddyinduced nitrate injection across the pycnocline. The concentration within the injection is specified as the model's nitrate bottom boundary value; see the Methods section. Figure 7. The annual cycles of surface PAR irradiance and photoperiod used in the model. Figure 8. Growth versus irradiance described by a hyperbolic tangent model. Solid line = netplankton; dashed line = picoplankton. The initial slope, at low irradiance, is described by a. The value at saturating irradiance is See Table 4 for specific parameter values. Figure 9. Figure 10. The Michaelis-Menten nutrientlimitation function with ammoniuminhibition of nitrate uptake over ranges of ammonium and nitrate. See Table 4 for specific parameter values. Grazing loss functions. Netplankton loss is described by a rectangular hyperbolic expression (solid line) Picoplankton loss is described by a linear expression (dashed line). See Table 4 for specific parameter values. Figure 11. Integrated (0-200 m) primary production in the no-pulse case. Solid line = total production; dashed = netplankton; dash-dotted picoplankton. Lower panel: integrated (0-200 m) nitrate. Figure 12. F-ratio in the no-pulse case. Solid line = total phytoplankton; dotted line = netplankton; dash-dotted picoplankton. Lower panel: integrated (0-200 m) nitrate. v 22 38 49 54 57 61 69 71


Figure 13. Figure 14. Figure 15. Figure 16. Figure 17. Figure 18. Integrated (0-200 m)' chlorophyll in the no-pulse case. Solid line total chlorophyll; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0 -200 m) nitrate. The annual variation of phytoplankton biomass (mg chl m-3 ) and nitrogen in the no-pulse case. Top panel: total phytoplankton and euphotic zone depth (dashed line) Middle panel: netplankton and nitrate (dashed contour lines; deepest= 18 Bottom panel: picoplankton and ammonium (dashed contour lines) The annual variation of the subsurface chlorophyll biomass maximum (SCM) in the no-pulse case. Toy panel: total biomass (mg chl m) Middle panel: ratio of SCM biomass to surface (5 m) biomass. Bottom panel: depth (m) of the SCM. The annual variation of phytoplankton biomass (mg chl m-3 ) under seasonal w ind forcing and n itraterich Mississippi River effluent i n (A) the central Gulf of Mex ico, and (B) at the west Florida margin in Walsh, et al.'s (1989) model. The annual variation of surface (5 m) chlorophyll biomass in the nopulse case. Solid line = total phytoplankton; das h e d = netplankton fraction; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate. The annual variation of grazing loss in t h e no-pulse case. Soli d line total phytoplankton; dashed = netplankton; dash-dotted = p icoplankton. Lower panel: integrated (0-200 m) nitrate. Figure 19. The annual variation o f e xport loss in t h e n o -pulse case. Only particulate losses are considered in vi 73 7 4 75 77 78 8 1


this plot-see text. Solid line = total phytoplankton; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate. Figure 20. Integrated (0-200 m) primary production in the spring case. Solid line = total production; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate. Figure 21. Integrated (0-200 m) chlorophyll in the spring case. Solid line = total chlorophyll; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate-Figure 22. Figure 23. Figure 24. Figure 25. F-ratio in the spring case. Solid line = total phytoplankton; dotted line = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate. The annual variation of phytoplankton biomass (mg chl m-3 ) and nitrogen in the spring case. Top panel: total phytoplankton and -euphotic zone depth (dashed line) Middle panel: netplankton and nitrate (dashed contour lines; deepest= 18 Bottom panel: picoplankton and ammonium (dashed contour lines) The annual variation of the subsurface chlorophyll biomass maximum (SCM) in the spring case. Tof panel: total biomass (mg chl m) Middle panel: ratio of SCM biomass to surface (5 m) biomass. Bottom panel: depth (m) of the SCM. The annual variation of surface (5 m) chlorophyll biomass in the spring case. Solid line = total phytoplankton; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrate d (0-200 m) nitrate. vii 82 84 86 88 89 90 92


Figure 26. Annual variation of 'simulated surface chlorophyll in the eastern Gulf of Mexico from Walsh, et al.'s (1989) model, and of 1979-82 CZCS estimates (Muller-Karger, et al., 1991) of pigment concentration for a 200x200 km2 region in the eastern Gulf of Mexico (A and C in Figure 3). Modified from Walsh, et al., 1989. Figure 27. NOAA sea surface temperature (C) analysis. Top: 21 February 1980; dot is station 1 of El-Sayed and Trees, 1980. Bottom: 22 February 1981; dot is Ortner, et al.'s (1984) station. The dashed box is the 200x200 km2 region used to construct a CZCS surface pigment time series. Figure 28. The annual variation of grazing loss in the spring case. Solid line = total phytoplankton; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate. Figure 29. The annual variation of export loss in the spring case. Only particulate losses are considered in this plot-see text. Solid line total phytoplankton, dashed = netplankton, dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate. Figure 30. Integrated (0-200 m) primary production in the summer case. Solid line = total production; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate. Figure 31. The annual variation of phytoplankton biomass (mg chl m-3 ) and nitrogen in the summer case. Top panel: total phytoplankton and euphotic zone depth (dashed line) Middle panel: netplankton and nitrate (dashed contour lines; deepest = 18 Bottom panel: picoplankton and ammonium (dashed viii 93 94 96 97 100


contour lines). Figure 32. Integrated (0-200 m) chlorophyll in the summer case. Solid line = total chlorophyll; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate-Figure 33. F-ratio in the summer case. Solid line = that for total phytoplankton; dotted line = netplankton; dashdotted = picoplankton. Lower panel: integrated (0-200 m) nitrate. Figure 34. The annual variation of the subsurface chlorophyll biomass maximum (SCM) in the summer case. Top panel: total biomass (mg chl m-3). Middle panel: ratio of SCM biomass to surface (5 m) biomass. Bottom panel: depth (m) of the SCM. Figure 35. The annual variation of surface (5 m) chlorophyll biomass in the summer case. Solid line = total phytoplankton; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate. Figure 36. The annual variation of grazing loss in the summer case. Solid line = total phytoplankton; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate. Figure 37. The annual variation of export loss in the summer case. Only particulate losses are considered in this plot-see text. Solid line = total phytoplankton; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate. ix 102 103 105 106 108 110 111


THE RESPONSE OF OCEANIC PHYTOPLANKTON TO NITRATE FLUX IN THE EASTERN GULF OF MEXICO: A SIMULATION ANALYSIS By MARK B. MEYERS An Abstract Of a dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Marine Science University of South Florida April 1993 Major Professor: John J. Walsh, Ph.D. X


A one-dimensional nitrogen-phytoplankton model was used to explore the impact of nitrate injections into the euphotic zone, associated with cyclonic eddies present along the southwest Florida slope during the Loop Current's quasiannual cycle of extension into the eastern Gulf of Mexico, on size-fractioned primary production and vertical carbon flux. The annual cycles of nitrate, ammonium and two size fractions of phytoplankton standing stocks were simulated, with climatological forcings of solar irradiance, mixed layer depth, and surface mixing coefficients determined from wind stress. When nitrate pulses where provided, simulating a series of three cyclonic eddies during spring (the spring-pulse case), primary production increased by three-fold to 235 g C m-2 yr-1 relative to 84 g C m-2 yr-1 in a no-pulse case. The former value is similar to the upper range of annual primary production extrapolated from in situ measurements off of southwest Florida, impacted by Loop Current cyclonic eddies, while the latter is in the range of annual estimates from the eastern Gulf of Mexico, in the absence of Loop Current cyclonic eddies. The netplankton fraction of annual production increased from 54% to 72% when nitrate pulses were supplied. The f-ratio was 0.36 in the no-pulse case, xi


and 0.42 when nitrate pulses were supplied. Netplankton cells were also the dominant component of the phytodetrital flux, representing 90% of the 22 g C m-2 yr-1 of uneaten material settling out of the upper 100 m in the spring-pulse case. However, almost all of the phyto-plankton biomass was grazed prior to settling out of the model domain; >99% of the loss of phytoplankton carbon over the entire domain was accounted for by grazing. This underscores the importance of understanding trophic relationships and nutrient recycling in order to follow carbon flow within the upper ocean ecosystem. Added biological complexity, together with nutrient distributions which mimic the natural environment, has improved our efforts in modelling the nitrogen-phytoplankton dynamics in the Gulf of Mexico. Further improvements, coupled with realistic and relevant physical circulation models, will improve our abilities to interpret diverse data sets and test hypotheses of ecosystem dynamics. Abstract Approved: Major Professor: John J. Walsh, Ph.D Distinguished Research Professor Departm ent of Marine Science Date Approved: I z [ Z L,_J xii


1 INTRODUCTION The fixation of inorganic carbon dioxide during photosynthesis by different functional groups of phytoplankton and the subsequent downward flux, through a variety of mechanisms, of particulate organic matter facilitate the exchange of C02 between the atmosphere and the ocean. Indeed, the rapid increase of algal biomass and inorganic carbon uptake during springtime, as a consequence of decreased vertical mixing, in temperate and boreal regions creates a dramatic diminution in the surface total C02 content (HC03 co3=, C02). The resultant C02 partial pressure (pC02 ) of surface sea water is lower than that of the overlying atmosphere, driving an increased flux of C02 into the ocean (Watson, et al., 1991; Taylor, et al., 1991). This has been referred to as the biological pump (Longhurst, 1991), and, like many ecological processes in the ocean, it exhibits large variability in time and space, and even changes sign seasonally. Another major process of carbon exchange results from strictly physical and chemical factors, driving the formation and downwelling of cold, saline surface waters in restricte d regions o f the Arctic and Antarctic oceans. Because of their cold temperatures, these downwelling waters


carry a large load of dissolved inorganic carbon. 2 Elsewhere however, especially in the equatorial ocean, deep waters rise up toward the surface with consequent outgassing of C02 (Takahashi, 1989). This process has been referred to as the saline conveyor belt (Broecker, 1991; Broecker and Peng, 1992), the importance of which may be mitigated by global climate changes since the Industrial Revolution. The spatial and temporal dependencies of biological processes make it difficult to quantify the role of the marine biota in global carbon flux estimates. The expense of maintaining year-round sampling programs within even a few of the major ocean regimes has made the use of timeseries observations to place limits on annual estimates of carbon flux through the marine biota tenuous. For example, the U.S. Global Ocean Flux Study program only maintains time-series stations in two subtropical regions, off Bermuda and Hawaii. Ecosystem simulation analysis, with coupled physicalbiological models, instead provides a framework for the interpretation of disparate biological and chemical data sets, aliased over time and space. A sufficiently large computer would allow us both to interpolate missing information over annual cycles and to extrapolate local observations to basin l ength scales. Present computers and data assimilation techniques, however, require us to smooth the spatia-temporal signals of the physical environment in


3 order to drive simple biological models of bulk variables, e.g., chlorophyll, rather than of populations of individual algal species. Such modelling activities do allow us to consider various hypotheses of ecosystem functioning: for example, top-down versus bottom-up population control, the importance of the microbial loop, the role of various grazer populations, and the role of nutrient supply in mediating competition among functional groups of phytoplankton. The New Production Paradigm for Open Ocean Ecosystems An appealing paradigm of the open ocean ecosystem has been that the external supply of nitrogen can be coupled to the loss pathways (Eppley and Peterson, 1979) of photosynthetically-fixed algal carbon (Figure 1). Nitrogen is assumed to be the element which limits the standing stock of algal biomass (Ryther and Dunstan, 1971), and perhaps its growth rate (e.g., Falkowski, e t al., 1992), in the ocean. Dugdale and Goering (1967) compartmentalized phytoplankton nitrogen uptake into "new'' (also, "preformed") and "recy-ele ct" forms of n itrogen, principally nitrate (N0 3-) and ammonium (NH4+)1 respectively. Nitrate is produced, for the purposes of this idealized conceptual model, in the oxyge n minimum layer beneath the euphotic zone of the world 1Henceforth, the charges on these ions will be dropped for convenience.


Euphotic Zone New Production I N03 I ........ ...,._ ----z Aphotic Zone Figure 1. The relationship of primary production by phytoplankton (P) and nitrogen cycling at the lowest trophic levels i n a marine food web. New production is based upon the uptake of allochthonous sources of nitrogen, represented here by the upward flux of nitrate. Recycled production is based upon recycled, or autochthononus, sources of nitrogen, represented here as ammonium excreted by zooplankton (Z) The downward flux of particulate organic matter (POM) can i nclude fresh phytodetritus, 4 zooplankton fecal pellets, and body remains. This simplified model does not consider dissolved organic matter fluxes, nor lateral processes.


5 ocean by the bacterially-mediated process of nitrification. The external pool of ammonium is formed via the respiratory degradation of proteins and nucleic acids by animals, heterotrophic bacteria and protozoans in proximity to the grazing of algae, and is readily available for uptake and assimilation by algae. To some degree, bacteria compete with phytoplankton for utilization of ammonium (Harrison, et al., 1992). In its simplest application, the conceptual model ignores this processes, as well as those of nitrogen fixation, denitrification, and nitrification within the euphotic zone. Similarly, carbon fixation, the formation of organic carbon skeletons from dissolved inorganic carbon via photosynthesis, can also be divided into "recycled" and "export" components, based on the fate of the phytoplankton. That particulate organic matter (POM) which sinks below the euphotic zone is considered lost from the local system, and thus "exported," whereas that POM which is grazed within the euphotic zone is considered "recycled" (Figure 1) On an appropriately long time scale, the exported POM should be balanced by the stoichiometric equivalent o f the nitrate fluxing into the euphotic zone. The differentiation of primary production which occurs i n conjunction with N03 assimilation ("new production") versus that which occurs in conjunction with NH4 assimilation ("recycled production") i s described by the "f-


6 ratio" where f = new production/total production, as defined by Eppley and Peterson (1979) This, in turn, yields an estimate of the export of particulate organic carbon (POC) from the upper ocean by assuming a mass equivalence of assimilated new nitrogen with, firstly, the POC of new production, and, secondly, the POC of export production (eg., Suess, 1980; Betzer, et al., 1984). Observations of high growth rates, relative to maximum intrinsic growth rates, by open ocean phytoplankton, and of particulate C:N ratios near the Redfield stoichiometric ratio of 6 (mol/mol; Redfield, 1958) led Goldman, et al. (1979) to conclude that the in situ growth of individual algal cells was not nutrient limited, despite the low stock of nutrients present in surface waters of oligotrophic ecosystems.-They proposed that the primary producers are tightly coupled to their grazers, with respect to both NH4 supply and carbon fixation. That is, there is a rapid exchange of carbon and nitrogen between the low stocks of small autotrophs and their grazers, and the subsequent predators in thi s food web. Conversely, a potential disconnection, or mismatch, may occur between p rimary producers and grazers exists when the nitrogen source is nitrate supplied to the euphotic zone via physical mixing from deeper waters. A transient increase of algal b iomass in larger size fractions, i.e., o f largersized or chain-forming diatoms, may occur. This would


represent a functional group which escapes continuous grazing pressure of the established microheterotroph assemblage through larger size and faster growth rates. Aggregation of such phytoplankton, poorly utilized by a slow-developing copepod assemblage, may thus lead to faster and larger sinking fluxes of POC from the euphotic zone (Alldredge and Gotschalk, 1989). Subsequent research on the relationship of annual 7 primary production to nitrate flux into the euphotic zone (Lewis, et al., 1986; Jenkins and Goldman, 1985), oxygen signatures (extent of supersaturation in surface waters and utilization below the euphotic zone) (Shulenberger and Reid, 1981; Craig and Hayward, 1987; Emerson, 1987), and POM flux measured by sediment traps (e.g., Pace, et al., 1987) led to contradictory conclusions, depending upon the methods and measurements of specific investigators. A rationalization of these enigmatic relationships was proposed by Platt, et al. (1989), who showed that the disparate time scales (minutes to 1 year) of different measurements and processes confounded the simple mass balance paradigms. Averaging, extrapolating, and perhaps aliasing (undersampling the timedependent variability) shorter time scale measurements made in bottles (eg., nitrate uptake using 15N radiotracers, or photosynthesis by uptake of 14C-labelled carbon over a period of 4-24 hr) all yield inaccurate annual estimates. These are then compared with bulk geochemical estimates of


8 processes involved, for example, in effecting mid-water oxygen distributions or trace gas fluxes, subject to poorly resolved physical processes at longer time and larger space scales. My approach was to simulate some of these processes at daily time scales over an annual cycle of climatological forcing. Producers, Grazers, the Microbial Loop, and the Bifurcation Model The major hypothesis of my study is that the oceanic phytoplankton assemblage exhibits a size-dependent response (or set of responses) to a physical event of transient nitrate supply to the euphotic zone. Included in this ecologic response is an alteration of the fate of phytoplankton carbon in the ecosystem, depending on the size fraction from which it originates. The event time scale over which the perturbation operates is defined as an interval lasting days to two weeks, encompassing the passage of meteorological fronts and boundary current eddies (Haury, et al., 1977). On a somewhat longer, but also aperiodic, time scale, the development and passage of larger eddies and rings occur (Haury, et al., 1977). Jenkins and Goldman (1985) viewed the euphotic zone as a two-layered system, in which an upper layer exhibits greater exchange and recycling of nitrogen between small autotrophs and heterotrophs in a tightly-coupled "spinning


9 wheel" paradigm. The lower euphotic zone is postulated to be more susceptible to episodic impulses of N03 from the aphotic zone and is presumably dominated by phytoplankton with a higher potential growth rate. Because there is no a priori reason for an explicit depth-dependence in this twolayered system, it is more general to replace "two-layered" with "two food web'' and assume that both webs operate simultaneously (Goldman, 1988). The bifurcation model (Figure 2) of Legendre and Le Fevre (1989) illustrates the differing fates in such a system for a nitrogen, or carbon, atom entering these webs. Their conceptual model serves as the framework for the numerical simulation formulated for my study. The bifurcation model combines aspects of two concepts of the open ocean food web. The classic food web paradigm, featuring secondary production funnelled through copepods, and subsequently onto fish is represented by Steele's (1974) The Structure of Marine Ecosystems. Remineralization of dissolved and particulate organic matter (DOM, POM) was taken to occur in deep water or in sediments. Pomeroy (1974) called attention to an alternative paradigm, stressing the processing of DOM and POM within the upper water column through a microbial pathway of bacteria and a variety of protozoans. Azam, et al. (1983) then completed the framework for the "microbial loop" hypothesis: that DOM released by phytoplankton is taken up, along with


A Netphytoplankton Ultraphytoplankton No Sinking Aggregation A ;u: .. Macrozoop. Accumulation/ S1nk1ng Grazing Microphagy ""'-)? 'l4..Ji( Microbial Food Loop -------Recycled : Total Production .._ __________ F-ratio----------10 Figure 2. The bifurcation model of marine trophic dynamics (modified from Legendre and Le Fevre, 1989).


inorganic nitrogen, by bacteria; bacteria are consumed by flagellates, who in turn are consumed by other flagellates and ciliates; the microbial consumers have low utilization efficiencies for nitrogen and, thus, return inorganic nitrogen (NH4 ) to the water column. 11 An important concept in this paradigm is that evidence of efficient retention of nitrogen by bacteria precludes their direct role in the remineralization of organic matter and that, indeed, these bacteria compete with phytoplankton for inorganic nitrogen (Harrison, et al., 1992). The remineralization pathway requires the participation of microbial loop grazers and consumers both to crop bacterioplankton and to return remineralized nitrogen back into the system (Fuhrman and McManus, 1984) These protozoan grazers are capable of 24-hr generation times (Fenchel, 1982a) and have been shown to lag bacterial blooms by d (Fenchel, 1982b; Azam, et al., 1983). Furthermore, they can graze upon phytoplankton with cell diameters thus keeping picoplankton populations in check as well. The distinction between the large-cell based food web (netplankton to copepods to consumers) and the small-cell based microbial loop (picoplankton and bacteria to grazer and consumer microflagellates) serves as the initial branch of the bifurcation model and stresses the different fates of a carbon, or nitrogen, atom upon algal assimilation. The first branch in the model determines whether that atom will


be incorporated into a large or small phytoplankton cell (Figure 2) 12 The empirical support for this model can be seen in the size distribution of phytoplankton cells and their response to environmental perturbations or trends. Bimodality in the distribution of autofluorescent particles, assumed to be autotrophic cells, is clear in data from flow cytometry, capabl e of examining particles with diameters in the range 0.3-150 (Figure 7 in Chisholm, 1992). Chisholm's data suggest at least two dominant size fractions, centered around ca. 10-1 and 1 02 pg C cell-1 with the smaller being dominated by Synechococcus and Prochlorococcus, providing validation of the distinct foundations envisioned by the major b reak between large and small autotrophs in the bifurcation model (Figure 2). Elevated nitrate stocks can be correlated with increases in b iomass of larger size classes of phytoplankton (eg., Malone, 1 9 71; Bishop, et al. 1980; Herbland, et al., 1 985; Chavez, 1989; Hulburt, 1990). Although there is some evidence (eg. Eppley, e t al., 1969) w hich suggests that larger-celled species have a greater ability t o take up N03 much o f this m a y result from experimental artifacts; t h e more plausible explanation may be that enhanced vertical mixing b o t h elevates s urface N03 concentrations and enable s larger cells to remain i n the euphotic zone ( M argalef, 197 8 ; Chisholm, 1992 and ref erences ther e in) This hypothesis


suggests that physical perturbations that increase the vertical flux of inorganic nitrogen into the euphotic zone will tend to favor the growth of larger-celled phytoplankton. Grazing or Accumulation: Uncoupling Netplankton From Herbivory 13 The duration of a physical event, in part, will deter-mine whether the growth increment of the large size fraction is grazed, or sinks out of the system (bifurcation 2 in Figure 2) This is a function of the development time of copepod grazers, from egg through several larval stages to maturation. At temperatures of 15-20C, it may take 2-4 weeks to develop from egg to adult (Vidal, 1980; Landry, 1983; Smitfr and Lane, 1987; Hofmann and Ambler, 1988). Netplankton blooms, in response to events shorter than that time scale, will be minimally affected by copepod grazers. Instead, they will settle out of, or be advected from, the system. This result has been shown through simulation analysis by Hofmann and Ambler (1988). Phytoplankton Sinking Rates The alternative loss pathway to the consumption of large phytoplankton is their sedimentation. Laboratory (in vitro) measurements of cell settling rates, whether on


14 natural assemblages or on cultures, show rates which are slow in comparison with upwelling velocities or turbulent mixing (eg., Smayda, 1970). Also, in laboratory experiments, light and nutrient conditions have little or no affect on cell settling rate, except perhaps during silicalimited diatom growth (Bienfang, et al., 1982). In the ocean, however, settling rates have been observed to decrease for samples taken near the base of the euphotic zone (Bienfang, 1980; Johnson and Smith, 1986) Comparisons of settling rates and very short-term, shallow sediment trap deployments show reasonable agreement between calculated and measured biomass fluxes under certain conditions of algal assemblage and environmental conditions (Johnson and Smith, 1986; Pitcher, et al., 1989). Nonetheless, there are reports of-rapid sedimentation events to the sea floor in slope and deep sea regions, suggesting settling rates of ca. 100 m d-1 of fresh phytoplankton cells (Lampitt, 1985; Billet, et al., 1986; Gooday, 1988; Lochte and Turley, 1988; Rice, et al., 1986). Of potential importance to rapid and episodic losses from the euphotic zone, is the process of cell aggregation and mass settling (Alldredge and Gotschalk, 1989). Observations and application of coagulation theory, however, suggest that biomasses (or cell abundances) much higher than is observed in open ocean systems are required before this process can be invoked (Riebesell, 1991). On the other hand, very large, oceanic diatoms of the


genus Rhizosolenia have been shown to be able to become positively buoyant, with ascension rates of >100 m d-1 (Villareal, 1988). Similar behavior is exhibited by the intrinsically less-dense cyanobacterium, Oscillatoria (Villareal, 1988) Indeed, close physiological control of 15 buoyancy, through regulation of cell sap density (Smayda, 1970), can be a mechanism of enhancement for nutrient uptake at depth and photosynthesis near the surface on diel scales. The regulation of cell sap density, viz. the density difference imparted by the siliceous frustules, is only effective for larger diatoms, which have surface to volume ratios above 0.45. The largest of diatoms, therefore, may have an advantage, over smaller ones, with respect to suspension in the low mixing environment of the subtropical open ocean (Smayda, 1970; Villareal, 1988). Clearly, the enigmatic nature of phytoplankton sinking rates will be grossly simplified in my modelling experiments, but necessarily so, in the absence of consistent relationships with physical and physiological conditions. Objectives I have used mathematical models of nutrient cycling, photosynthesis, algal growth and loss (primarily f rom sinking and grazing), and particle flux in surface waters of the open ocean in order to study the potential ecosystem impacts


16 of the following: 1. The impact of N03 flux on annual primary production in slope waters, influenced by Loop Current eddies, in the eastern Gulf of Mexico. 2. The impact of N03 flux on the algal assemblage and the potential fate of algal carbon produced in a smaller size class versus a larger size class (with differing physiological characteristics and loss pathways) 3. The relationship of short-term eddy and mixing events to longer term mass balance and annual cycles of nitrogen and carbon in the surface open ecosystem. The primary modelling effort used a one-dimensional model of vertical mixing and eddy-enhanced nitrate influx coupled with two algal size fractions, grazing losses and nitrogen recycling. The results from this are compared with an earlier -three-dimensional basin-scale model of Loop Current circulation and nitrate-based primary production for a single phytoplankton component with no grazing losses and with a minimal description of nitrogen recycling. Because these models addressed the Gulf of Mexico ecosystem (Figure 3), specific features of the physical and biological regimes are discussed. Details of the model formulation and parameter choices then follow.


..JU ou 30 TEXAS 26 2 4 22 2 0 O L uu uu U ""l U L UIU . . .. .. .. .. . . . .. .:B DRY . .. . .. G;l < .:.. ... :" w .. FLORIDA ... ...... ._:_ : ... : . .. : .. ..... . . . ... ..... GVY.. z .... .... q:t. .-: .. : t-:;:( . ;q:oc: ,. Ot-, .. >-. . 30 28 2 6 2 4 22 2 0 Figur e 3. Th e Gul f of M exic o Reg ions A and B are the cent r a l Gulf and w e s t Florida margin locations of model output by Walsh, et al. (1989), for comparison with the present s tudy, which focused on Region B. Region C i s a 200x200 km2 box which was subsampled from a CZCS time serie s (Mul ler-Karger, e t al., 1990).


CIRCULATION, HYDROGRAPHY, AND PRIMARY PRODUCTION IN THE GULF OF MEXICO 18 As a means of assessing the impact of mesoscale (days-weeks, 100's of km) variability, eg., as a result of cyclonic eddies, on annual estimates of nitrate flux and primary production, the annual cycles of dissolved inorganic and of phytoplankton in the surface waters of the Gulf of Mexico were modelled (Walsh, et al., 1989). The Gulf of Mexico served as a useful test basin (see Lewis, 1992) for a simple physical model of eddy generation from the larger scale circulation, as observed along western oceanic boundaries. With few exceptions, in the absence of a coordinated program of in situ observations of Loop Current eddy impacts on nitrogen and phytoplankton stocks, and primary productivity, validation of the modelling efforts must rely upon approximating snap shot observations of the Loop Current and Florida slope ecosystems made by various investigators under various physical (Loop Current configuration, time-of-year) conditions. In particular, model performance will be compared with descriptions of the vertical structure of nitrogen and phytoplankton standing stocks, daily and annual primary productivity estimates, Coastal Zone Color Scanner (CZCS) estimates of surface biomass, and estimates of the vertical flux of dissolved


19 inorganic nitrogen and of particulate organic matter. In this section of background information on the Gulf of Mexico, a description of the Loop Current circulation and related hydrography is given, followed by descriptions of the basin's vertical structure of phytoplankton biomass, its productivity and similar descriptions for phytoplankton dynamics along the Florida slope region. The Gulf of Mexico Loop Current The Loop Current affects nitrate flux by influencing the depth of nitrate isopleths in the Gulf of Mexico, thereby altering the vertical nitrate gradient. This latter quantity directly affects the diffusive flux of nitrate into the relatively deep ( 75-100 m) euphotic zone (depth to which light is reduced to 1% of the amount just below the surface of the water) of the open Gulf of Mexico. Surface waters of the Loop Current, originating in the Caribbean (Nowlin, 1972), are warm and nutrient-depleted. Temperature and nitrate isopleths within this current are 200+ m deeper than their depths in surrounding Gulf Common Water (GCW) For example, the nitrate isopleth deepens to >400 m in LCW versus ca. 150 m in GCW (Figure 4). While the Loop Current was once thought to have an annual cycle of penetration and eddy shedding into the Gulf of Mexico (eg., Leipper, 1970; Maul, 1977), rings have been


20 I 129 j ,A 27 I 29' Figure 4. The depths of the 22C isotherm and the N03 isopleth in the eastern Gulf of Mexico during November 197 6 (fro m W alsh, et al., 19 8 9 )


21 observed to be shed over sub-and superannual ranges, in both winter and summer (Behringer, et al., 1977; Sturges and Evans, 1983; Vukovich, 1988; Auer, 1987). The maximum northward penetration of the Loop Current into the Gulf varies from 24-28N latitude (Vukovich, et al., 1979), although excursions up to the Mississippi-Alabama-Florida shelf break (north of 29N) have been documented (Schroeder, et al., 1988; Huh, et al., 1981). During periods when the Loop Current is extended into the Gulf, instabilities (warm filaments with cold-core frontal eddies-see Figure 5) form along its cyclonic (outer) margin enhancing vertical nitrate flux and primary productivity along both the Yucatan (Furnas and Smayda, 1987; French, et al., 1983) and West Florida (Pal us zkiewicz, et al. 19 8 3; Yoder and Mahood, 19 8 3) margins (see 2igure 3) Subsequent to the Loop Current's maximum penetration into the Gulf, a large cold perturbation (e.g., during April 1983; Figure 5) forms off the Dry Tortugas, Florida (Figure 3), initiating ring separation. The process of ring separation, occurring at intervals of 6-18 months (eg., Auer, 1987; Vukovich, 1988), appears, from numerical modelling experiments, to be related to the latitudinal variation of the Coriolis parameter (Reid, 1972; Hurlburt and Thompson, 1980). The resulting anticyclones translate westward at speeds of 2.5-8.0 em s-1 (Elliot, 1982; Kirwan, et al., 1984), impact the continental margin at the western


22 9 <983 12 APRI L 1 98 3 29 26 2 7 2 6 25 2 4 90 89 8 8 87 8 6 >O 89 88 87 86 1 9 M A R C H 1 983 26 APRIL 1983 : I &9 e-3 ':L .... :, .... e, so 2 ;"'PRIL i 983 9 0 89 88 8 7 86 85 84 83 82 8 1 Figure 5 Sea-surface temperature ( C ) signature o f the Loop Curr e n t during spring 1983 showing a s eque nce o f cold-core frontal eddies alon g t h e west Florida margin, culminating in a large cold perturbation off the Dry Tortugas, Florida in mid-Apr i 1 (modified from Vukovich, 19 8 6 )


edge of the Gulf of Mexico, and undergo frictional dissipation (eg., Vidal, et al., 1992). They appear to persist along the western margin for 3-5 months (Lewis and Kirwan, 1985). 23 Along the western margin, smaller cyclones (cold core rings) are observed in association with, and possibly engendered by (Smith and O'Brien, 1983), the warm core r ings. Whereas the anticyclones are downwelling i n thei r centers, the cyclones are upwelling i n their centers, and, as along the western Gulf margin, over a short spatial scale there can be large differences in nitrate profiles, phytoplankton and zooplankton biomasses, and primary p roductivitie s between these two structures (eg., Biggs, 1992). These are similar t o cyclonic eddy features foun d in the eastern Gulf of Mexico (Vukovich, 1 986) and in the South Atlantic Bigh t (Lee and Atk inson, 1 983). Phytoplankton and Primary Production in the Gulf o f M e xico As a consequence of the time-dependent, spatially v arying nutrient i njection effected by t h e Loop C urren t cycle and local wind mixing, the biomass, species composition, and productivity of phytoplankton exhibit signif ican t variability i n the Gulf o f Mex ico. I will briefly describe extant information o n the general features of the depth-integral s and of verti c a l structure of this


24 oligotrophic ecosystem. Then, the specific features of plankton dynamics of my study region, the eastern Gulf along the west Florida margin, will be discussed in relation to the Loop Current cycle. These prior observations will provide a validation of the results of the simulation analysis. Steele (1964) first investigated the nature of the subsurface chlorophyll maximum (SCM), using the Gulf of Mexico as a model of a subtropical oligotrophic system. His stations, sampled during April, 1962, were off the southwest Florida slope, near 25N, 84.5W. Natural phytoplankton populations were taken from depths of 5, 50 and 100 m and incubated on deck using artificial light (Sylvania Grolux fluorescent bulbs) to investigate the pattern of maximum productivity with depth. Rates ranged from 0.3 to 1.08 mg C m-3 hr-1 with the maximum recorded from the 50-m population. The use of artificial illumination and the paucity of details on the photosynthesis-irradiance performance for each sample make it difficult to compare these values with other data sets. With respect to biomass, for a transect off the west Florida shelf, there was little relationship between particulate carbon ( PC ) and chlorophyll. PC decreased both with increasing distance offshore and with increasing depth. In particular, despite a SCM of 0.2-0.3 mg chl m -3 at d epths o f 50-70 m, there wa s no similar maximum of PC, supporting the hypothesis that the


SCM is a phenomenon of photo-adaptation (decreasing algal C:chl with depth), not an increase in phytoplankton carbon biomass. Despite the uncertainties of the non-algal component in the determination of particulate carbon, recognized by Steele (1964), at depth, the bulk of the PC:chl ratios ranged between ca. 30-100. 25 Hobson and Lorenzen (1972) also observed a SCM in the eastern Gulf of Mexico, which generally tracked the pycnocline at depths of 50-100 m. At least for one station sampled during November 1969, phytoplankton carbon and chlorophyll both had maxima at depth (55-65 m) SCM chlorophyll values were 0.3-0.6 mg chl m-3 while surface concentrations were mg chl m-3 Phytoplankton carbon concentrations were 16-20 mg C m-3 in the SCM and 3 7 mg m-3 at 1 m and 6 mg m-3 at 46 m. Again, these may reflect decreasing algal C:chl with depth. Zooplankton, comprised mostly of ciliates, copepod nauplii and copepodids, had a maximum coincident with that of the phytoplankton. During this time they report that the algal assemblage was dominated by microflagellates and dinoflagellates, whereas a year earlier, similarly located samples were dominated instead by coccolithophorids and diatoms. El-Sayed and Turner (1977) later reported chlorophyll and primary production data for 34 Gulf of Mexico stations during three consecutive summers (1971-1973), and an additional 18 stations sampled during the same fall 1969


period as Hobson and Lorenzen. Mean primary productivity for all of these station i s 289 mg c m-2 d-1 yielding an annual estimate of 105 g C m2 yr-1 (Table 1). Summertime surface chlorophyll concentrations in the open Gulf in El-Sayed and Turner's (1977) data set were 0.08-0.20 mg m-3 For comparison, Louisiana shelf concentrations were 0.14-3.03 mg chl m-3 and those in the 26 Caribbean source waters of the Loop Current (an additional 6 stations south of the Yucatan Straits) were 0.05-0.15 mg chl m-3 The mean 0-200 m integrated chlorophyll during the summertime stations in the open Gulf was 31 . 0 mg chl m2 ; the mean value integrated down to the base of the euphotic zone (the one percent light depth in the Gulf is typically 75-100 m) was 12.7 mg chl m-2 Stations in the Caribbean had a mean euphotic zone chl of 7 .22 mg chl m -2 ; only one station was integrated down to 200 m and it had a value of 21.9 mg chl -2 m Surface and integrated (euphotic zone) productivities for the summertime, open Gulf stations averaged 5.84 mg C m-3 d-1 and 193 mg c m -2 d-1 ; this does not include one station located north of the Yucatan Straits with a value of 1568 mg c m 2 d -1 where surface nitrate was reported as 0.5 versus 0.1 most other stations. One station on the Louisiana shelf had a productivity of 605 mg C m -2 d -1 and two west Florida s h e l f stations averaged 335 m g C m-2 d-1 Six Caribbean stations averaged 390 mg C m-2 d-1


Table 1. Recent primary production estimates for the Gulf of Mexico. Where the authors have done so, mean rates and annual estimates are reported, in addition to the observed ranges of daily integrated rates. Study El-Sayed and Turner, 1977 Yoder and Mahood, 1983 Vargo, et al., 1987 Biggs, 1992 Ortner, et al., 1984 Region, time of year open Gulf summer, fall SW Florida margin April, 1982 September, 1982 west Florida shelf No red-tide Red-tide western Gulf warm-core (WC) ring outside we ring cold-core ring Loop Current February, 1984 Measured Rate (mg C m 2 d -1 ) 289 (18-1570) 500 (100-700) 525 (150-1200) 300-500 800 160 260 250 14-62 mg C m -2 hr-1 Annual Estimate (g C m-2 yr-1 ) 105 121-186 N ....J


28 The November 1969 survey, all in the western gulf, reported by El-Sayed and Turner (1977) had a mean surface biomass of 0.065 mg chl m-3 and mean euphotic zone and 0-200 m integrated biomasses of 6.54 mg chl m-2 and 15.5 mg chl m-2 Surface and integrated (euphotic zone) productivities averaged 4.41 mg C m-3 d-1 and 171 mg c m-2 d-1 Two stations in the central Gulf had productivities of 18.4 and 64.6 mg C m-2 d -1 ; all other stations had values over 100 mg C m-2 d-1 with a mean of 208 mg C m-2 d-1 leaving one to speculate whether an anticyclonic warm-core ring was in the western Gul f at that time. During the summer 1973 cruise, El-Sayed and Turner (1977) reported that nanoplankton (<20 represented 65-93% of the total biomass and 52-99% of the primary production. Taxonomic composition was studied from net samples. Small diatoms were most abundant, represented by Ceratulina pelagica, Thalassionema nitzschoides and Rhizosolenia alata and R. stolterfothii. Contemporaneous with El-Sayed and Turner's surveys of the Gulf basin, Ednoff (1974) examined the relationship of phytoplankton abundance and composition with the Loop Current in the eastern Gulf and the Florida Straits. He collected surface samples only, using two sampling methods: pump-collected water filtered i) through a 65 phytoplankton net into sample bottles, to which buffered, three-percent formalin was added, or ii) through 0.45


29 Millipore filters, which were stored frozen. Oscillatoria erythraea, forming colonies, or chains, of 5-300 cells, dominated most net samples, often representing 50-80% of the total cell count. Comparison of net samples with filtered ones showed that the latter collected on average 17 times more cells per volume. The dominants of the filtered samples often were pennate diatoms of the family Naviculacaea, in a size range of 5-20 These diatoms were either rare or absent in the net samples and were more abundant in open Gulf filtered samples taken during the fall. Furthermore, filtered samples contained more species than net samples and roughly half of the species found in the net samples were also found in the filtered samples. During a February cruise in the Florida Straits, Ednoff (1974) sampled a cold-core frontal eddy along the cyclonic edge of the Loop Current. At that station, Oscillatoria erythraea dominated, followed by the diatoms Rhizosolenia stolterfothii, Skeletonema costatum, Thalassiothrix delicatula, Hemiaulus membranaceus, and Nitzschia closterium. Two diatom species, Chaetocerus wighami and Bacteriastrum delicatulum, were more abundant within the Loop Current. Dinoflagellates, mostly of the genus Ceratium, were scarce, in comparison. These same patterns were observed latter in spring, apparently being more dependent on Loop Current position than season.


30 During summer, total cell abundance decreased by more than ten-fold. The diatom assemblage was less prominent, while there was an increase in the dinoflagellate component. During August 1973, however, a low-salinity (<35.0 per mil) surface lens in the middle of the Florida Straits was sampled, in which Rhizosolenia alata (ca. 10,000 cells 1-1), a coastal species, dominated. During this period, the Loop Current was extended fully into the Gulf, and was in the process of shedding a large anticyclone (Maul, 1977). Low salinity coastal water may have been entrained along the Loop Current's cyclonic edge (Atkinson and Wallace, 1975; Paluszkiewicz, et al., 1983). At other times of the year, greater cell numbers were also found at the cyclonic edge of the Loop Current and the assemblage there was dominated by both diatoms, including Thalassiothrix spp., Rhizosolenia alata and Navicula spp., and Oscillatoria. The presence of typically coastal/shelf diatom species such as Rhizosolenia alata and Skeletonema costatum suggests either cross-shelf advection into the open Gulf of coastal water and/or an enhanced nutrient and mixing regime at the cyclonic edge of the Loop Current. Similar diatom species also contribute to enhanced productivity within cyclonic frontal eddies of the Gulf Stream in the South Atlantic Bight (Yoder, et al., 1981, 1983), suggesting that while seeding from shelf populations may take place, enhanced local production must be responsible for the high cell


31 numbers observed in these offshore features. However, a coastal source for the occurrence of large diatoms, including variants of Rhizosolenia alata, is not necessarily required to explain their presence, as these have been observed in other open ocean regions (Clemons and Miller, 1984; Villareal, 1988). To my knowledge, only one study specifically has addressed the variability of primary production associated with Loop Current cyclonic frontal eddies (Yoder and Mahood, 1983). Using satellite and ship-based information, the response of the phytoplankton assemblage to nutrient injection by eddies was studied on the outer shelf and slope off of southwest Florida during April and September, 1982. Their method for determining primary production was the acid bubbling method of Wessels and Birnbaum (1979), which is significant in that it involves no filtration prior to scintillation analysis for 14c-labelled carbon uptake. This at least removes any potential filtration bias against phototrophic picoplankton. Filtration for chlorophyll was performed using Gelman type "C" glass fiber filters. A frontal eddy may be defined as a warm filament of Loop Current water shoreward of the main flow, with cooler shelf water between the filament and the main flow, and a thermal front between the filament and the main body of shelf water (Paluszkiewicz, et al., 1983). For the April eddy, the filament length was 220 km; its width varied


32 during the week-long survey between 35-62 km (Paluszkiewicz, et al., 1983). The horizontal length scale of the cold dome was estimated to be >100 km (Paluszkiewicz, et al., 1983) A translation speed of 30 em s-1 was estimated from the location of thermal fronts in successive sea surface temperature (SST) images recorded by the Advanced Very High Resolution Radiometer (AVHRR) (Paluszkiewicz, et al., 1983). During the April 1982 cruise, primary productivity was estimated at fifteen stations, across a range of isobaths and the cold-core eddy system, encompassing the mid-shelf, the eddy itself in both outer shelf and slope regions, and the main flow of the Loop Current (Table 2). These regions essentially corresponded with the isobath zones <100 m, 100-200 m, and >200 m; in the last region, one station was in the Loop Current, while three others were within the cold core of the eddy. A mean integrated primary productivity of 500 mg c m-2 d-1 was identical among the three depth zones. The one station within the Loop Current exhibited the lowest rate of the study, 100 mg C m-2 d-1 a value ca. one-sixth that of the other slope stations, indicative of the eddy's impact in slope waters. The September cruise began with a series of survey transects using XBT and CTD casts to map the Loop Current and shelf water mass structure, in the absence of AVHRR data. Summertime surface temperatures are uniformly warm in Gulf open ocean and shelf waters, making identification of


33 the Loop Current position by thermal imagery untenable. Muller-Karger, et al. (1991), however, do present an example of the use of color imagery during summer months for Loop Current delineation. Strikingly evident from the second survey was the presence of a warm, relatively fresh (<35.5 PSU) lens (ca. 20-m thick) of water covering the outer shelf region. This water mass was almost entirely dominated (>80% of total abundance) by the diatom Rhizosolenia alata, with cell counts of >10,000 1-1 Unusual, however, was the fact that despite the high cell number, chlorophyll concentrations were quite low (0.05-0.2 mg chl m -3), which may have been the result of i) high C:chl a ratios from photoadaptation in this very stable layer, or ii) from senescence or severe nutrient limitation (Yoder and Mahood, 1983). A subsurface chlorophyll maximum (SCM) was found at either 40 or 70 m, in which the coccolithophorid Coccolithus huxle y i usually dominated; occasionally, various diatoms prevailed. The actual depth of the SCM may have been aliased by their bottle sampling protocol (samples taken at the 100, 50, 25, 10, 5 and 1% surface irradiance depths). An in situ fluorometer was used only during the April cruise. At eight of fourteen stations, primary productivity (not normal ized to biomass) was maximal in the SCM. Average integrate d b iomass and productivity between the 100-m and 200-m isobaths, during September, were 16 mg chl


Table 2. Phytoplankton biomass (mg chl m-2 ) and primary production (mg C m -2 d -1 ) across the southwestern Florida outer shelf and slope during an intrusion of t h e Loop Current (Yoder and Mahood, 1983). Mean Mea n Isobath range (m) biomass Range (n) production Range (n) April 1982 <100 3.8 2 .5-5.0 ( 2 ) 500 400-500 ( 2) 100-200 9.8 6.0-15 ( 9 ) 500 200-700 ( 9 ) >200 8.5 5.0-10 ( 4) 500 100-700 ( 4) September 1982 <100 11 10-12 ( 3 ) 400 250-600 (3) 1002 0 0 20 13-34 ( 4) 900 600-1200 ( 4 ) >200 8.3 5.0-11 ( 7) 500 150-900 ( 7)


35 m-2 and 800 mg C m-2 d-1 (Table 2, Yoder and Mahood, 1983) The highest productivity recorded was 1200 mg c m-2 d-1. Recall that in the open Gulf, the mean values of these properties were 13 mg chl m-2 and 200 mg c m-2 d-1 within the summer euphotic zone (El-Sayed and Turner, 1977). Thus, at similar biomass, the enhanced carbon fixation along the slope implies differences in nutrient supply, as well as herbivory and POM export (Figure 2). As an example of the short-term variability of primary production at a specific site in the Gulf of Mexico, the hourly rate of photosynthesis increased four-fold with the passage of a meteorological front over the Loop Current region (Table 1, Ortner, et al. 1984). This observation was coincident with a >50-m deepening of the mixed layer (from 20-40 m to >110 m) and a shoaling of the 3-j.l.M N03 isopleth from 240 m to 140 m within three days. Surface water N03 concentrations remained below their detection limits (0.4 jl.M). Salt and temperature budgets suggested that some horizontal advection of a warmer, fresher water lens, similar to those observed by Ednoff (1974) and Paluszkiewicz, et al. (1983; see also, Yoder and Mahood, 1983), had occurred coincident with the passage of the meteorological front, perhaps contributing to the shoaling of deep N03 isopleths and a coastal seed population of phytoplankton. All of the above investigations, except for that of Ortner, et al. (1984), however, were performed prior to the


36 appreciation of the ubiquitous importance of the picoplankton size fraction's (<1-2 contribution to open ocean phytoplankton biomass and primary production (Waterbury, et al., 1979; Li, et al., 1983; Platt, et al., 1983; Murphy and Haugen, 1985). The importance of tracemetal clean techniques for measuring oceanic primary production (Carpenter and Lively, 1980; Fitzwater, et al., 1982; Ferguson and Sunda, 1984; Ortner, et al., 1984) was also unappreciated. Thus, the carbon fixation may have been underestimated, while the value of diatomaceous netplankton may have been overemphasized, neglecting the picoplankton contribution to carbon fixation and flux (Figure 2)


37 METHODS One-Dimensional Model Framework A one-dimensional, nitrogen-phytoplankton model was constructed for the upper 200 m of a water column at ca. 25N latitude in the eastern Gulf of Mexico. Its physical aspects were restricted to climatological descriptions of mixing (derived from wind stress and mixed layer depth) and solar irradiance (Table 3), based upon the same seasonal forcing functions of the previous three-dimensional model of basin-wide nitrate-phytoplankton dynamics in the Gulf of Mexico (Walsh; et al., 1989). This one-dimensional model focused instead on the impact of nitrate injection by cyclonic eddy events, with a more complex description of the potential biological response, in terms of nitrogen recycling and uptake, phytoplankton taxonomic composition, and carbon flow through different herbivore pathways. A schematic diagram of the one-dimensional model grid and simulated eddy-pulsing of nitrate is shown in Figure 6. The 200-m upper ocean domain was divided into 20 10-m thick fixed levels, or boxes. The system is defined by a set of coupled, nonlinear partial differential equations, which define the local rate of change of the state variables,


..c o_ Q) 0 200m 38 Time Dz=10 Figure 6. A schematic diagram of the one-dimensional model domain, illustrating the mechanism used to simulate eddy-induced nitrate injection across the pycnocline. The concentration within the injection is specified as the model's nitrate bottom boundary value; see the Methods section.


Table 3. Mid-monthly (16th day) values for wind stress Month Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec (t), surface vertical eddy diffusivity (K0), mixed layer depth (hm) and noontime photosynthetically active radiation (10 max) used in the Gulf of Mexico ecosystem simulations (cf., Walsh, et al., 1989). (Pa) 0.061 0.062 0.072 0.077 0.069 0.059 0.042 0.032 0.056 0.081 0.077 0.077 49 5 0 58 62 56 48 34 26 45 66 62 57 100 93 86 62 41 24 24 24 45 48 59 78 Io, max (l.i.Ein m-2 s-1 ) 1238 1419 1587 1699 1729 1727 1728 1713 1629 1466 1275 1177


described below, at the mid-point between each grid level. The fluxes of each of these were evaluated at level inter-faces. State Equations 40 The model state equations for netplankton (Pnl picoplankton (PP) using chlorophyll as an index of biomass, nitrate (N0 3 ) and ammonium (NH4 ) are defined in equations ( 1) ( 2) ( 3) and ( 4) Algal growth rate limitation is based on a Blackman model (Blackman, 1905; Platt, et al., 1977) assuming that either light (I) or nutrient, in the form of total inorganic nitrogen (N0 3 + NH4), is the single rate limiting factor at any given depth and time. The functl'ons a and a 1'n the n1'trogen equations refer to ox red those portions of the total algal growth demand supplied by N03 and NH4 respectively; they are defined by equations (5) and (6). Table 4 lists the values of various parameters used in the model. [ ( ani(z) l N03 e -'f "NH' NH4 l 1-Lm,n min tanh + pn apn 1-Lm,n kN,n+N03 kA,n+NH4 (1) ac gm,n (P -P ) a aPn aPn + -Kw --k +P n no az z az s,n az g,n n


41 = (2) + (3) (4) (5) (6)


Table 4. Biological parameters used in the model. Parameter Maximum growth rate Photosynthetic efficiency Chl-specific absorption coefficient Carbon chl a N03 half-saturation NH4 half-saturation N03-uptake inhibition Maximum grazing loss Grazing half-saturation Recycling efficiency Symbol a E Particle settling rate w5 Units day-1 rng c (rng chl a) -l hr-1 rn-2 s-1 1-1 1 11 1 rng chl a m -3 d -1 rng chl a m-3 (unit less) rn day-1 Netplankton 2.77 0.072 0.020 50 0.1 .025 1.5 1. 03 1.0 0.50 2.5 *This value is in units of day-1 ; i.e., it is a chl-specific loss rate. Pica plankton 1. 39 0.12 0.033 30 0.1 .025 1.5 0.696* N/A 0.80 0.0


43 Initial conditions for nitrate, ammonium, and total phytoplankton biomass were taken as composites of conditions in the open Gulf (Waddell, 1986; Biggs, 1988). Total phytoplankton biomass initially was apportioned evenly between netplankton and picoplankton. The model was then run for four years in order to ensure convergence and to attain an equilibrium among the state variables. The fifth year of the model runs are then presented. Model cases with nitrate pulses were initialized with state variable values at the end of the fifth year of the no-pulse case. Boundary conditions are required to define the model's behavior at the surface and at the base of the domain. For all state variables, no mass flux was allowed across the surface of the model ocean. For example, the model did not have an atmospheric source of nitrogen (e.g., Paerl, 1985; Fanning, 1989). At the base of the model domain, dissolved inorganic nitrogen was allowed to diffuse according to its concentration gradient. Bottom boundary concentrations for the state variables were specified at one grid box below the model domain (at a depth of 205 m) For nitrate, this value was 18 in accord with hydrographic observations (e.g., Waddell, 1986). Ammonium, on the other hand, was assigned a bottom boundary value of zero, as were both size fractions of chlorophyll. In the absence of an adequate description of the physi-cal generation of the observed small cyclonic eddies along


44 the cyclonic edge of a western boundary current, or of the upwelling velocities within their centers, some alternative means (deus ex machina) of creating a nitrate injection was required. Transient nitrate pulses were simulated by pro-viding an exogenous time-and depth-dependent source term, Q(z,t). The shoaling of nitrate isopleths seen in transects across cyclonic eddies in the eastern Gulf (Paluszkiewicz, et al., 1983; Waddell, 1986) was mimicked by providing grid boxes within a specified height above the model's base at 200 m (hereinafter referred to as.the ''pulse height") an additional source of nitrate, other than from diffusion, sufficient to maintain the nitrate concentration within those boxes at the nominal boundary value for nitrate, at each time step for a duration characteristic of frontal eddies (eg. 5 d) Observations across the Loop Current and its cyclonic eddies indicated that nitrate and temperature isopleths, normally at ca. 200 m i n Gulf Common Water, rise by as much as 50 m, to 150 m, defining the p u l s e height for the model simulation (Vukovich, 1986; Waddell, 1986). For comparison with the no-pulse case, the additive effect of this nitrate source i s e xpresse d below as if i t were an additional contribution to the influx of nitrate a t 200 m. This mechanism s e rved as a p r o x y for vertical advecti o n as a means o f enhancin g the nitrate i n flux I n this simplif i cation, how e ver, the lateral and vertic a l advective consequences for the other state variables, ammo-


45 nium and the two phytoplankton components, were ignored. Merely altering the bottom boundary value of nitrate and allowing a strictly diffusive exchange of nitrate across the model boundary at 200 m was insufficient to produce a significant impact on the depth profiles of nitrate or phytoplankton. Diffusive flux near the thermocline, without upwelling, is too slow to allow for a sufficient change of nitrate within the upper 1 00 m of the water column. Vertical Mixing Following Walsh, et. al. (1989), I empirically defined, using equation (7), the vertical structure of turbulent mixing with a form which decays exponentially from the base of the surface mixed layer to a specified value at the base of the model domain. The mixed layer eddy diffusivity and mixed layer depth were varied daily by linearly interpolating from monthly mean climatological values (Table 3). The mixed layer value for eddy diffusivity, K0 was assumed to be a function of the surface wind stress, t, and the Coriolis parameter, f, with an empirical relationship from Csanady (1976): K0 = 4.9 x 10-3 t /f. Blumberg and Mellor (their Figure 3a; 1985) reported monthly climatological e stimates for the wind stress vector i n the


46 Gulf of Mexico, the magnitude of which was used in the above formula (Table 3). The eddy diffusivity at the bottom boundary was estimated in two ways. First, investigations of mixing processes, microturbulent shear and nutrient flux across the pycnoclines of tropical and subtropical regions (Eppley, et al., 1979; King and Devol, 1979; Gargett, 1984) yielded eddy diffusivities, within the thermocline, of 0.1-1.0x10-4 m 2 s-1 Second, total annual primary production in most open ocean regions is limited by the upward flux of nitrate. Thus, a correspondence between strictly eddy diffusive nitrate flux and a minimum estimate of annual primary production (e.g., Jenkins and Goldman, 1985; Lewis, et al., 1986) The diffusive flux of nitrate scales directly with eddy and with the nitrate gradient across the thermocline, the bottom boundary in the model. With a bottom boundary value for eddy diffusivity, Kb of 0.2x10-4 m 2 s-1 the bottom boundary value for nitrate discussed above, and the set of biological parameters discussed below, the modelled primary production fell within the lower range (<100 g c m-2 yr-1) of recent primary production estimates for the eastern Gulf of Mexico. The vertical profile of eddy diffusivity, Kz was obtained by assuming an exponential decay from a uniform value within the surface mixed layer to the bottom value. This can be described by equation (7), where zb is the bottom of


47 the model domain, and hm is the depth of the mixed layer. Monthly climatological averages of (Table 3) were estimated from Levi tus' (1982) climatology at 25. 5 N, 4 o. 5 ow in the North Atlantic, which compare favorably with Gulf of Mexico temperature profiles from the National Oceanographic Data Center archives. (7) Solar Irradiance and Light Extinction In order to specify the light regime, daily solar declination angles for 25N latitude were first calculated; from these, daily values of the photoperiod, il (Figure 7), starting when model clock-hour, t equals sunrise, tr, were derived (Kirk, 1983). Maximum daily irradiance (at solar noon), Imax (Table 3) was calculated from Reed (1977) and converted to photosynthetically active radiation (400-700 nm; PAR) assuming that half of the surface irradiance is PAR; energy units were converted from joules to quanta (1 = 10-6 mol quanta) with a factor of 4.15 (Morel and Smith, 1974). The annual cycle of irradiance thus calculated is shown in Figure 7. A sine function was used to calculate the time-dependence of light over the photo-


period, in contrast to a sine3 function in the new production model of Walsh, et al., 1989). Light was extinguished exponentially with depth, as a function of attenuation from sea water itself and from the absorption and scattering properties of the algae. The attenuation coefficient for sea water, kw, was 0.033 m-1 while those for each of the algal components (kc,i), given in Table 4, reflect differing "packaging effects" (Kirk, 48 1983). The smaller picoplankton were deemed to be efficient at light absorption, per uni t chlorophyll, than the netplankton, w ith a chlorophyll-specifi c attenuation coefficient of 0.033 versus 0.020 m2 (mg chl)-1 for netplankton. The increased path length of diffusely scattered light was taken into account by assuming an inverse average cosine path length of 1.25 (Go rdon and Morel, 1983; Kirk, 1983). The attenuation of light by detritus and dissolved organic matter has been ignored. The time-and depth-dependent equation for light thus becomes I (z, t) = sin (lt ( t t,) n-l) e xp[ -1. 25( k.z+ L kc,i I p i (z) dz) l (8) T h e current state-of-the-art i n bio-optical modelling e xpands the PAR-based description o f the submarine light


,---..._ -I V1 N I E c w :::i.. '-.../ Q) u c 0 -o 0 c.. c.. 2000 1800 ------/ ..------1600 ./ / / 1400 / 1200 c ___ _.__ __ ___,_ ____ L_ ___ j _______ L___ 20 18 ,---..._ c.. _c -o 16 0 c.. Q) o_ 0 14 ....... --.. --------------------------L_ __ _c__ _ _L_ _________ L _____ L ______ l_ _______ _l____ _l_ __ _j__ __ _L __ __. ___ 0 _c ll. 12 Jan Mar lv1oy Jul lv1onlh Sep Nov Jan Figure 7. The annual cycles of surface PAR irradiance and photoperiod used in the model.


50 field to a wavelength-dependent one (eg., Sathyendranath and Platt, 1988, 1989). The advantages of such a description would be increased precision in the calculation of light attenuation and the ability to incorporate taxonomicallydependent light absorption for photosynthesis, which may affect the vertical structure of phytoplankton assemblages. However, in the absence of a specific bio-optical data set, including depth-varying estimates of spectral light, attenuation coefficients, and phytoplankton pigments, this level of detail was deemed to be beyond the scope of the present study. Phytoplankton Growth and Irradiance The netplankton, predominantly diatoms, were assigned a higher intrinsic maximum growth rate, in keeping with laboratory studies of growth rates among various phytoplankton taxa ( eg., Chan, 1978; Brand and Guillard, 1981; Falkowski, et al., 1985; Langdon, 1987). The model value for netplankton of 2.77 d -1 (4 div d-1 ) was taken from Eppley (1972), assuming a mixed layer temperature of 2 5 o c Brand and Guillard (1981) also reported growth rates of this order for diatoms cultured at 2 4 C. Maximum observed (in situ) growth rates for picoplankton, instead, range from ca. 0.5 to 2.0 divisions per day for the sargasso and Caribbean Seas and t h e subtrop-


51 ical North Pacific (Bienfang and Takahashi, 1983; Iturriaga and Mitchell, 1986; Glover, et al., 1987; Taguchi, et al., 1988; Marra, et al., 1992). The carbon-specific value used in the model, 1.39 d-1 reflects the upper end of this range, 2 div d-1 The smaller picoplankton have an intrinsic light harvesting advantage, per unit chlorophyll, (higher kc) than the netplankton, as a result of their pigment arrangement (Kirk, 1983), and, hence, were deemed more efficient at low light (higher a-(chl/C), Table 4). For example, Joint and Pomeroy (1986) found two-to three-fold greater photosyn-thetic efficiencies in the picoplankton size fraction than in the >5 size fraction. The carbon-specific growth rates and efficiencies used in this model result in saturating light efficiencies of ca. 100 and 250 m -2 s-1 for the picoplankton and netplankton, respectively. For comparison, laboratory and shipboard (Iturriaga and Mitchell, 1986; Joint and Pomeroy, 1986; Kana and Glibert, 1987) studies have shown picoplankton populations which reach light saturation at ca. 100-175 m -2 s -1 Furthermore, because both groups can exhibit steady photosynthetic rates up to several hundred m -2 s -1 (Kana and Glibert, 1987; Howard and Joint, 1989), potential photoinhibitory effects were ignored. Photosynthetic efficiency can also be analyzed theoret ically a = k where is the maximum quantum yield of m c m


52 photosynthesis (mole C per mole quanta) and with appropriate dimensional conversion factors. The minimum amount of light energy needed to fix one mole of C is equivalent to 8-10 Ein (1 mole quanta= 1 Einstein), though the practical limit may be closer to 12 Ein as a result of slippage/leakage processes in energy transduction (Kirk, 1983; Geider, et al., 1986). Laboratory experiments on algal cultures have measured in the range of ca. 12-20 mol quanta (mol C fixed)-1 with some shipboard determinations larger quantum requirements (i.e., lower m). For the model photosynthetic efficiencies, a (Table 4), it has been assumed that = 0.083 (Iturriaga and Mitchell, 1986) and that a= a(m kc), where a= 43.2 (12000 mg c mol-1-3600 s hr-1 -10-6 Ein per The c:chl ratios shown are minimum values, assumed for shade-adapted cells, reflecting prior observations of lower ratios in the SCM (Steele, 1964; Hobson and Lorenzen, 1972). Glover, et al. (1987) and Howard and Joint (1989) found that picoplankton C:chl ratios ranged from 20 to 40 even within the summertime, well-lit mixed layer. The netplankton ratio of 50 and the picoplankton ratio of 30 were used together with a constant C:N ratio of 6.625 (mol/mol) to convert the simulated pools of nitrogen, carbon and chlorophyll. The resultant competition based on growth rates and light harvesting abilities between the two size classes is illustrated in Figure 8. The modelled picoplankton grew


53 faster at, and saturated at, lower irradiance, whereas the modelled netplankton were capable of higher growth rates at greater irradiance. The highest surface irradiance was simulated during May-July, when the diffusive supply of nitrate is curtailed by low mixing rates and shallow mixed layer depths (Table 3); nutrient limitation may then prevail in surface waters. Nitrogen Uptake The two functional groups of phytoplankton were considerect roughly equal in their affinities for NH4 and N03 Although there may be an inverse relationship between cell size and nutrient uptake affinity, corresponding to the ratio of area to volume (eg., Eppley, et al., 1969; Stockner, 1988), species-specific short-term uptake dynamics may offset any strictly allometric relationship (eg., Dortch, et al., 1984; Callos, 1986; Chisholm, 1992). The two modelled groups thus were assigned the same nutrient half-saturation constants (Table 4). Half-saturation constants for oceanic species have probably been overestimated by the 15N technique, as a result of the relatively high concentrations of labelled "tracer" added to sea water samples from oligotrophic re-gions (Harrison, 1983). Also, precise measure m e nts of nitrogen stocks at low ambient concentrations have been


f-L = u t a n h ( o: I 1-Lm1 ) m 3.0 2 5 2 0 1 .5 / I --'--------' --'------'------'----L-----'-----100 300 400 500 1 -,-odionce (f-LEin m -2 sec-1 ) Figure 8. Growth versus irradi a nce described by a hyperbolic tangent model Solid line = netplankton; dashed line = picoplankton. Th e i nitial slope, at low irradiance, is described by a. The value at saturating irradiance is See Table 4 for specific parameter values.


55 difficult, until the recent development of more sensitive techniques (Garside, 1982). The consequence of poor resolution of uptake rates, at ambient nitrogen concentrations near their detection limits, is a highly uncertain value of the half-saturation constant for either nitrate, or ammonium uptake (Goldman and Glibert, 1983) Nonetheless, kN values of 0.1-0.2 oceanic species constitute the classic literature (Eppley, et al., 1969; Carpenter and Guillard, 1971). If phytoplankton are able to maintain high relative growth rates in the open ocean, in concert with rapid recycling of NH4 (eg., Goldman, et al., 1979), it seems reasonable to assume that, at the least, their kA must be very low. A value of 0.025 for ammonium was used here compared to 0.10 for nitrate (Table 4). In the nitrogen uptake expression, ammonium is prefer-entially taken up, 1.5 (Wroblewski, 1977). There is a major caveat in such a nitrogen selecti vity formalism. The specific choices of kA, kN can result in a nutri-ent limitation function that exceeds 1 .0. The nutrient limitation function, over the range of ammon ium and nitrate values in the model is illustrated in Figure 9 I n a multiplicative model, where light and nutrient limitation relations simultaneously limit this could result in a realized growt h rate tha t is actu ally larger t h a n unless checks are built into the mod e l code. Nevertheless, In the


single-limiting factor coding used here, realized growth cannot exceed the specified maximum growth rate. Grazing Losses and Recycling Efficiency 56 The grazing loss terms were the least constrained set of biological relationships in the model. Feeding and standing stock data, particularly for protozoan grazers, are not abundant for subtropical regimes, especially for the Gulf of Mexico. Such growth, respiration and grazing data are often from temperature-boreal coastal regimes (eg., Heinbokel, 1978; Fenchel, 1982a,b). Development of this portion of the model, therefore, was guided by a few key assumptions on the different nature of macrozooplankton grazing versus herbivory within the microbial loop, as developed conceptually in the bifurcation model (Figure 2). My constraints on the grazing parameters were both the ultimate survival of the two phytoplankton groups, and the amount of ammonium, from grazer activity, that could be allowed to accumulate below the euphotic zone. Based on few observations, it appeared that ammonium concentrations should remain <1.0 often <0.5 within the model domain (Sharp, 1983; Biggs, 1988). The assumptions on the implied microherbivore population dynamics, including mortality, are as follows:


Figure 9 :::J E :::J c 0 E E 'c> 0. \ \ \ I J ./ fKn + f'.l) e x p(-psi *A)+ A / (f < a +A) K n = 0 10 l

1. The herbivore growth rates must be similar to that of picoplankton; i.e., population doubling on the order of days, allowing a high ingestion rate of picoplankton. 58 2. The herbivores must have high mortality rates, with small fluctuations of their biomass. 3. The cumulative effects of grazing, predation and respiration within the microbial loop imply a high degree of retention of carbon and nitrogen, in dissolved and essentially non-settling small particulate (both living and detrital) forms, within the model domain. Primary production is thus respired back to C02 and NH4 locally. This is effected by a larger recycling efficiency, E, than that for netplankton (Table 4). I make no attempt, for lack of validation data, to model pools of either bacteria, or DOM. 4. The fractional share of POM exported from the microbial loop is 1-E. This is assumed to immediately (that is, on the daily scale of model output) exit the model domain, implying some mechanism of entrainment and flocculation of POM into rapidly settling marine snow. Similarly, the assumptions for the implied macroherbivore population dynamics are as follows: 5. Copepod population growth rates are much less than that of the netplankton, with doubling times of weeks-month. 6. Their ingestion rate is a saturable, prey-concentration-dependent function (Mullin, et al., 1975). 7. Unassimilated netplankton is form e d into fecal pellets, with a high settling velocity of order 100 m d-1 contributing to the export of carbon and nitrogen from the system (Walsh, et al., This is effected by a smaller E (Table 4), whlch contributes to the domain's NH4 stock. 8. Macroherbivor e mortality l eads predominantly to export, via migration of their predators.and rapidlysettling carcasses (Longhurst and This further suggests a lower recycllng efflClency


59 of netplankton versus picoplankton within the upper ocean model domain. Guided by these two sets of assumptions, the grazing loss for the netplankton, i.e., the last term in equation (1), was modelled using a rectangular hyperbolic function (Figure 10), similar to that for nutrient uptake. The grazing half-saturation constant, kg, was set to 1.0 mg chl m-3 chosen so that all but the largest of algal blooms, fell in the near-linear portion of the rectangular hyperbola (Figure 10) The slope, m of the grazing stress in this region was assumed to be similar to a nutrient-limited real-ized specific growth rate, ca. 10% of The maximum grazing rate, gm was then determined algebraically by recalling that kg is the chlorophyll concentration at which g = and assuming that m = i.e., gm = 2mkg. Tight coupling between phytoplankton and grazers is surmised to weaken at times when the larger phytoplankton size fraction comprises an increasing proportion of biomass and production under transient nutrient forcing. In the model, netplankton potentially could outgrow their macrozooplankton grazers, as a result of the saturating nature of the hyperbolic grazing loss term. As a result, netplankton biomass, under conditions which permit high growth rates, may accumulate until light or nutrient limitation occurs.


60 In addition, faster sinking rates, w5 (Table 4) and the implied larger, heavier fecal pellets produced by macrozooplankton, which would leave the domain on the order of one day, shorten the euphotic zone residence time of organic carbon and nitrogen in the larger size fractions. Thus, the nitrogen recycling efficiency, E of netplankton grazing loss was lower, 0.5, than that of picoplankton grazer loss, 0.8, given that packaged feces would escape from the model domain faster than the organic nitrogen contained within them could be recycled, and that a fraction of excreted nitrogen is also lost in the migratory nature of the grazers and their predators, contrasting with the physiology and motility of protozoan grazers on the picoplankton. In the case of the picoplankton population, grazed upon by rapidly growing protozoan herbivores, a linear, non-saturating function of the form -gPP was used (equation (2), Figure 10). In this function, gp is a mass-specific rate constant. This loss function should allow the grazing rate to track increases in biomass, as would be the case in a tightly coupled system. Reported carbon-specific grazing rates on chroococcoid cyanobacteria (e.g., Synechococcus spp.) are in the range of 0.2-0.83 d -1 (Landry, et al., 1984; Campbell and carpenter, 1986; Iturriaga and Mitchell, 1986). These studies indicated that as much as 90% of the phytoplankton growth rate and 40% of their standing crop were removed each day. The model's value of 0.696 d-1


I >, 0 lJ I') I E ..c u 0.4 r o.2 (/) (/) 0 CJ) c N 0 I.... <..? G = G "' N m,N G = p / / / / / chi I (G + chi) k / / G chi / "' / m,P / / / / / / / / / / / / / ----------0. 0 ""'------'----..._ __ __._ ___ .__ __ __._ ___ __j__ ____ _. ___ _._ __ ____J'-----_J 0 0 0 .2 0 4 0.6 0.8 1 .0 chi (mg m -3 ) Figure 10. Grazing loss functions. Netplankton loss is described by a rectangular hyperbolic expression (solid line) Picoplankton loss is described by a linear expression (dashed line) See Table 4 for specific parameter values.


62 applied t o IJicc;r:-.:.ZJ.nktcn size fraction, falls within the range of these ccservations. Lastly, with r espect to grazing, thresholds, or refuge populations, Pno and ? po were established for netplankton and picoplankton respectively. This prevented grazing to extinction, which have occurred in the absence of any explici t dynamics. were low, 0.004 mg c h l and mg chl m -3 respectively, for picoplankton a n d Phytoplankton Rates The and chain-forming, viable phytoplankto n cel:s an enigmatic phenomenon. Ignoring the possibilities c f a lgal diurnal migrations, two parsimonious assumpticns can te made about the size-fractioned populations in the mcdel First, ?iccplankton cells, those in diameters, do not sink at any appreciable speed; this fraction i s thu s assigned a settling rate of 0.0 m d-1 (eg., Bienfang and Takahashi, 1983). Second, larger cells settle faster, according to Stokes' Law for sinking spherical bodies. The assigned sinking rate for the model's netplankton is 2.5 m d-1 This is in the upper range o f estimates diatom settling rates of single cells, under a var1ety of experimental conditions (Smayda, 1970; Bienfang, et al., 1982). Field studies


63 suggest that i t mignt ce SOt higher than reported rates from some subtropical, sligotrophic areas (eg., Bienfang, 1980; Bienfang and Harriscn, 1984). Numerical A vertical ::.-.:.c :::;:-acin g (tiz) cf :o m (20 fixed layers over a 200 m chcsen sc there would be at least two model o:.-ic points with the seasonal mixed layer, 24 m at its s umme:.Time-dependent changes for the state variables (masses) were evaluated at the grid box centers, while mass were evaluated at level interfaces. The von stability condition for the diffusion terms solved explicitly with a forward-in-time, centered-inspace (FTCS) numerical integration method, with a given is one constant a time step (tit) must satisfy: 1 :or a maximum Kz of 70 cm2 sec-1 (Table 3) and = 10 m, a 2 hours would be required. The particulate (i.e., chlorophyll) settling term instead is solved with a modified upwind difference technique, particularly suitable for positive-definite properties, such as mass distributions (Smolarkiewicz, 1983). This method applies a correction for the artificial viscosity (Roach, 1976; O'Brien, 1986) that results from the upwind difference scheme by forming an anti-diffusion correction


64 from the estimated artificial viscosity. This numerical scheme for dealing with the velocity term is illustrated in equations (9), (!0 ) and (11); the starred mass variable, M* represents the solution of the upwind difference scheme, before the correction velocity, u* is applied. (9) = r ( iu:.y,i U j2.y,) (Mj.l Mj") l l (Mj + Mj". 1 + ) (10) "' Mj. + { [ (uj -y,+juj -Ytll M j 1 + (uj-y,-juj-Ytp Mj. ] (11) -[ ( u j ".y,+ lujYtl) Mj. + (uj.y,-jujYtl) Mj.l ] } The biological terms were explicitly with a :orward -in-time method. Th e nitrogen uptake term, partic u larly giv e n the low half-saturation values for NH4 uptake, placed another restrictio n on the time step, however. For large biological demand can outstrip supply, resulting in spurious, negative masses of dissolved n itrogen. While a of 1 hour was possible for a single a lgal component in a m odel o f nitrate-base d production (Wals h e t al., 1 989), experimentatio n in the


65 two-component model here showed that stable results required a shorter time step; a time step of 900 seconds was used. Mathematically speaking, equations of this type, where different times have dramatically different time scales (eg., here, mixin g versus remineralization) are known as "stiff" equations. Because the biologically-determined time step was not smaller than that which satisfied numerical stability :cr physical mixing, all of the terms in :hese equations were solved at the smaller time step. To ensure convergence c : :he model after startup from initial equations (1)-(8) were integrated i n t ime for four years; the fifth year results are presented below.


66 RESULTS The results of the one-dimensional, two-phytoplankton model, hereinafter referred to as the 10 model, simulations are presented below. Three scenarios of the biological consequences o f flux at the southwest Florida slope are examined: i ) a s imulated year within Gulf Common Water (GCW), without the of the cyclonic edge of the Loop Current, ii) a in which Loop Current extension into the Gulf of Mexico occurred during spring, and iii) one in which Loop Current extension occurred during summer. In these last two cases, the interaction of the seasonal mixing cycle with the timing of eddy and geostrophic uplift processes along the cyclonic edge of the Loop Current is illustrated. In addition, the results of the 10 model, with added biological complexity, will be compared with results from the previous three-dimensional, single-phytoplankton, new production model (Walsh et al., 1989), hereinafter referred to as the 3D model. The various model results also will be placed in context with in situ and remotely-sensed observations from the eastern Gulf of Mexico and other subtropical regions.


67 No Loop Current Under basin conditions within GCW, and with no intrusion of the Loop i.e., within a physical habitat of only seasonally-varying diffusive nitrate flux, the modelled annual primary production was 83.9 g C m-2 yr-1. This was evenly divided between the netplankton (54%) and pice plankton (46%) popu:ations (Table 5). With a minimal nitrate influx cf 384 :-:1mol n m-2 yr-1 (Table 5), the annual primary production was smaller than 14C-based annual estimates within either open Gulf, or along the west Florida slope and outer shel:. Total daily primary production varied between 118 and 386 mg C m-2 d-1 in May and January, respectively (Figure 1:). This January maximum was two-to four-fold less than in situ maximum rates (Table 1), measured during spring and summer (El-Sayed and Turner, 1977; Yoder and Mahood, 1983). In the 1D model, without Loop Current influence, new production was 30.5 g c m-2 yr-1 (Table 5), with an annual f-ratio of 0.36, similar to that required to reconcile primary production and nitrate flux estimates in the Sargasso Sea (Platt, et al., 1989). The 3D model, in which the f-ratio was empirically defined by a relationship with ambient nitrate concentration (Platt and Harrison, 1985), yielded a mean f-ratio of only 0.12 over the upper 200m domain (Walsh, et al., 1989).


Table 5. Primary production from model year five for three c .::Jses Produc tion values (g C m 2 yr-1 ) are integrated over the entire 2 uu m domain. Nitrate fluxes (mmol m -2 yr-1 ) are evaluated at 100 and 200 m. Negative fluxes are 11pward. Model Algal Primary N e w FN03 flux case group production production ratio ( 1 0 0 m) ( u 0 m) No Net 45.3 17.7 0.39 Loop Pico 38.6 12.8 0.33 Current Total 83.9 30.5 0.36 -337 -384 Net 169 75.0 0.44 Spring Pico 66.6 23.0 0.35 Total 236 98.0 0.42 -1210 -1240 Net 134 56.9 0.42 Summer Pico 76.3 25.8 0.34 Total 210 82.7 0.39 -1010 -1040 0"1 CX>


500 'i "0 JOO N I E u .... 0\ 200 '\ .. / E '\ / ...... / \ ., / -. )' .. ,. ....... .... / ,. ....... .... ,. ..... / ,. ..... / .. ...... ,. .. .... -0 --400 I I I I I N :-------I E 300 r-z r200 -0 -E 100 r-E or I I I ___ j_ L Jan Mar May Jul Sep Nov Jon Month Figure 11. Integrated (0-200 m) primary production in the no-pulse case. Solid line = total production; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate.


70 Examined on a daily basis, f-ratios in the more biologically complex 1D model were highest during autumn and winter, reflecting higher mixing (Table 3) and greater diffusive nitrate flux (Figure 12) Given a higher photo-synthetic efficiency at low light, and therefore growing better at the base of the euphotic zone where nitrate con-centrations were high, the picoplankton size fraction consistently had a higher f-ratio than did the netplankton; mean daily f-ratics f o r netplankton and picoplankton were 0.30 and 0.37, respec:.:..;ely (Figure 12). However, integrat-ing production annual basis, over which time mass balances are assumed to apply, netplankton had a higher f-ratio of 0.39, compared with 0.33 for picoplankton (Table 5). This reflects the much higher production of netplankton (Figure ll) during winter, when the diffusive influx of nitrate was greatest. Co-varying w ith cycle of daily carbon fixation, integrated (0-200 m) phytoplankton biomass had a minimum of 7 .45 mg chl m-2 in May and maxima in January and in October of ca. 24 mg chl m -2 (Figure 13), somewhat less than the 31.0 mg chl m -2 observed during summer in t h e open Gulf (El-Sayed and Turner, 1977). This phasing of biomass maxima and minima contrasts with the previous single-phytoplankton (3D) model (Walsh, e t al., 1989). The annual mean o f the 3D model, i n a case without a nitrate influx from the Mississippi River, was 52 mg chl m -2 along the west Florida


"' I 0

72 slope; the 1D model, i n the absence of a nitrate injection at the cyclonic edge of the Loop Current, yielded 17 mg chl m-2. As expected, the seasonal forcing functions simulated i n the 1D model had a profound effect on the relative importance of the two func:ional groups (Figure 13). By size fraction, netplankton comprised 67% of the total integrated biomass during the winter w hereas picoplankton dominated the assemblage duri::g sc:m.lT,er, comprising >95% of the total algal biomass. The component had a minimum integrated biomass of 0 4 9 !T.g c ::l :n-2 during August; the picoplankton component reached of 2.52 mg chl m-2 during April. Maximum integrated biomass occurred during January for netplankton versus October for picoplankton, with respective values of l6.7 and 20.9 mg chl m -2 The minimum in total biomass occurred durin g May, after the netplankton population had declined from its winter maximum and prior to the picoplankton population's summer increase. Although the vertical structure of chlorophyll also varied during the year (Figure 14), at all times there was a subsurface chlorophyll maximum (S CM; Figure 15). During the winter algal biomass maximum of ca. 0.23 mg chl m -3 the SCM was broadly centered around 45-55 m (Figure 14, Figure 15). Total biomass at this depth was less than 1.5-fold greater than that at 5 m, the c ente r of the uppermost model g rid box (Figure 15) During August-September the SCM was centered


-... I E 0' E 0 :c u 50 40 30 20 '"""-... / .:::/ / .. -QL_ ____ L_ ____ ____ _L ____ -E 3001_ z 1-2001-0 IE 100 IE 1-Jon Mar May Jul Sep Nov Jon Month Figure 13. Integrated (0-200 m) chlorophyll in the no-pulse case. Solid line = total chlorophyll; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate. ...J w


-E -s:. a. 41 0 -E -s:. a. 41 0 -E -s:. a. 41 0 100 150 200 0 150 200 10 0 0 10 150 200 Jon --10--------' Mer May 74 I f -01--1 o---I \' C1 \ 0 0 \ 0 / .I' o 'o --0.10 --Jul S ep Nov J o n Month Figure 14. The annual variation of phytoplankton biomass {mg chl m -3 ) and nitrogen in the no-pulse case. Top panel: total phytoplankton and euphotic zone depth {dashed line) Middle panel: netplankton and nitrate {dashed contour lines; d eepest= 18 Bottom panel: picoplankton and ammonium {dashed contour l ines).


0.50 _, E C7' O .JO E 0.1 0 0 .00 1 01 u 2 .. :;) VI :::l u VI s::. Q. 0 :::l u VI J I i 2 _,..; l Nov Jon Jon Mar May Jul Month Sep 75 Figure 15. The annual variation of the subsurface chlorophyll biomass maximum (SCM) in the no-pulse case. Top panel: total biomass (mg chl m-3). Middle panel: ratio of SCM biomass to surface (5 m) biomass. Bottom panel: depth (m) of the SCM.


76 around 95 m, at the 1% light depth (Figure 14), with total biomass greater than 0.30 mg chl m-3 six-to nine-fold higher than that at 5 m (Figure 15). In the 3D model, vertical structure was minimal; in particular, the presence of a sharp, summertime SCM was not simulated (Figure 16) Surface (5 m) chlorophyll concentrations varied seasonally in this model case, from ca. 0.18 mg chl m-3 during October-January to 0.03 mg chl m -3 in August-September (Figure 17). This compares well with a mean CZCS annual cycle of >0.18 mg m-3 during winter months (January-March) to ca. 0.06 mg chl m -3 during the summer (minimum in JuneJuly) These values were monthly means for the entire Gulf basin, exclusive of the continental shelves, derived from CZCS imagery, spanning the seven-year period of November 1978-November 1985 (Muller-Karger, et al., 1991). In contrast, surface chlorophyll values from the Loop Current region in the 3D model ranged from 0.40 mg C m-3 in February to 0.04 in August-September (Walsh, et al., 1989). The model's phytoplankton standing stock, as well as those from in situ and remotely-sensed observations, represents the summation of production and loss terms. The loss pathways in the model were both losses resulting from grazing activity and processes at higher trophic levels, and the settling and diffusive fluxes of uneaten phytoplankton cells (phytodetritus) (Table 6). Grazing losses (Figure 18), summed over the 200-m deep model domain, balanced total


, 1 .r ( lN\\ ..... -... l JUL . . . ....... j OCT DAYS I JAN I RPR 77 Figure 16. The annual variation of phytoplankton biomass (mg chl m-3 ) under seasonal wind forcing and nitrate-rich Mississippi River effluent in (A) the central Gulf of Mexico, and (B) at the west Florida margin i n Walsh, et al.'s (1989) model.


-..., I E 0' E ..._, 0 :c u 0 .25 0 .20 0 15 0 10 ' ' ' / r / / , / / / .... ---I I I 400-I I E 300 f_ z r-200-0 100 r-r--I I I Jon Mar May Jul Sep Nov Jon Month Figure 17. The annual variation of surface (5 rn) chlorophyll biomass in the no-pulse case. Solid line = total phytoplankton; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 rn) nitrate. ...J (X)


79 primary production to >99.9%; this is somewhat obscured by round-off in the values of Table 6. Phytodetrital flux out of the model domain at 200m represented <0.1% of the export via grazing activity (Figure 19). Sedimenting phytodetritus at 100 m, roughly the base of the euphotic zone, however, was 6.47 g C m-2 yr-1. The nitrate influx of 384 mmol N m-2 yr-1 at 200 m, yields a "new" production of 30.5 g C m-2 yr-1 assuming a C:N ratio of 6.625 (mol/mol). A zooplankton-derived detrital flux of 30.3 g c m-2 yr-1 the total secondary export of Table 6 roughly balances this new production. Such a rough mass balance of "new" production versus export does not include the dissolved component, i.e., DOC or DON, which was not explicitly simulated. However, the export flux, at 200 m, of NH4 can be calculated and must be included in the mass balance of nitrogen within the model domain. The extent of this dissolved flux (Table 6) is <1% of the secondary export, but ca. four-fold the phytodetrital flux of 0.0504 g C m-2 yr-1, when converted into a carbon equivalent of 0.194 g C m -2 yr1 via the Redfield ratio.


Table 6. Grazing loss, particulate matter export, and inorganic nitrogen fluxes from model year 5 for three cases. Grazing losses (g C m-2 lr-1 ) are integrated over the entire 200 m domain. Secondary export (g C m 2 yr) is the maximum potential export of feces and remains of herbivores based on recycling efficiencies used in the model. Fluxes are evaluated at 100 and 200 m for carbon (mg m-2 yr-1 ) and nitrogen (mmol m -2 yr-1). Negative fluxes are upward. Model Algal Grazing Secondary case group loss export No Net 45.2 22.6 Loop Pi co 38.6 7.72 Current Total 83.8 30.3 Net 168 84.0 Spring Pi co 66.6 13.3 Total 235 97.3 Net 134 67.0 Summer Pi co 76.3 15.3 Total 210 82.3 Phytodetrital flux (100 m) (200 m) 4680 44.7 1790 5.71 6,470 50.4 17,800 44.3 2920 4.69 20,720 49.0 14,800 44. 6 2760 4.94 17,560 49.5 Nitrate flux (100 m) (200 m) -337 -384 -1210 -1230 -1010 -1040 Ammonium flux (1 00 m) -66.3 -136 -125 (200 m) 2.45 5.95 5.30 00 0


500 I 300 ... I E u 0"1 E ' ...... / ........ .. .... / ........ --... / ... / 0 400 f-I I I I ... r---I -E 300 1--z I-200 I-0 --E 100 f.E I-0 I I _l J I Jon Mar M a y Jul Sep Nov Jon M onth Figure 18 The annual variation of grazing loss in the no-pulse case. Solid line = total phytoplankton; dashed = netplankton; dash-dotted = picoplankton Lower panel: integrated (0-200 m) nitrate. ()0 ....


200 150 -; '0 .. I E 100 u 0'1 E -z 200 0 \ \ .... .._ / ---------. / -.. ....... . -_ ... _... / < / / / / ........ ......... --100-____ _L _____ __ Jon Mar May Jul Sep Nov Jon Month Figure 19. The annual variation of export loss in the no-pulse case. Only particulate losses are considered in this plot-see text. Solid line = total phytoplankton; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate.


83 Loop Current Present During Spring Another model case was run for a subsequent five simulated years, to equilibrium, assuming the presence of the cyclonic edge of the Loop Current in proximity to the southwest Florida slope during spring. Within this time frame, the nitrate injection from a series of cold-core frontal eddies was simulated with three five-day pulses for a sixtyday period during March-May, similar in timing to the sea surface temperature s ignature of the Loop Current frontal structure along the southwest Florida slope in 1983 (Figure 5) This nutrient regime increased annual primary production by three-fold, to 235 g c m-2 yr-1 (Table 5), similar to the high end of the range of extrapolated 14cbased estimates for the outer shelf and slope (Table 1, Table 2) On an annual basis, the netplankton size fraction accounted for 72% of the model's production (Figure 20), compared with 54% in the previous case (Figure 11). On a daily basis, total primary production varied from 340 mg C m-2 d-1 in mid-March to 967 mg C m-2 d-1 in April, during the eddy-induced nitrate injections (Figure 20) This compares well with a maximum rate of ca. 800 mg C m-2 d-1 measured in April, 1982, under similar Loop Current forcing (Yoder and Mahood, 1983). The annual cycle of integrated algal biomass (Figure 21) again mirrored that for production, with a



PAGE 100

85 minimum stock during mid-March (18.7 mg chl m-2). soon thereafter, however, during the simulated eddy events, total biomass reached a maximum of 77.3 mg chl m-2 The summer biomass minimum, simulated in the no-pulse case (Figure 13), was not as small in this second case, because of the long residence time and extensive recycling of the Loop Currentinduced nitrate influx. The annual mean of 46.2 mg chl m-2 in this case compares well with that of 52 mg chl m-2 for the southwest Florida slope under Loop Current forcing in the 3D model (Walsh, et al., 1989). The netplankton component, in this case, represented 88-95% of the winter total biomass and >99% of the bloom following nitrate injection, while picoplankton comprised >85% of the summer and autumn stocks. An autumn bloom was also simulated, which began during September as the seasonal wind stress increased toward its maximum value during October (Table 3). Picoplankton, instead, comprised ca. 90% of this bloom's biomass. During late autumn, however, the netplankton fraction began to increase, dominating the phytoplankton assemblage by the end of November (Figure 21) New production was 98.0 g C m-2 yr-1 (Table 5), yielding an annually-averaged f-ratio of 0.42, somewhat higher than the no-pulse case. The nitrate influx, at 200 m, sup--2 -1 f. ft d porting this production was 1240 mmol m yr i een ays of a simulated eddy field resulted in an increased annual nitrate influx of three-fold (Table 5) In the spring,

PAGE 101

-... I "' I E 0 E u E z 0 E E 100 1000 500 0 Jon Mar ., "' ...:: ..... ..... I / / --'--May Jul Month / / / / \ / \ / \ I / :-,. / / / ,. ,. .... ---Sep Nov Jon Figure 21. Integrated (0-200 m) chlorophyll in the spring case. Solid line = total chlorophyll; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate. 00 "'

PAGE 102

under the influence of Loop Current-induced nitrate injection, the f-ratio increased to 0.63, but declined during summer, to 0.28 (Figure 22). 87 A combination of residual, unutilized nitrate in the aphotic zone (Figure 23, middle panel) and increased ammonium (Figure 23, bottom panel), recycled from grazing activity, continued to feed h igher production and algal stocks, relative to the no-pulse case, throughout the remainder of the year (Figure 20) Indeed, the fall bloom of this case {Figure 23), resulting from seasonally-increasing mixing {Table 3) of algal b i omass and nitrogen higher into the euphotic zone, was not simulated in the no-pulse case {Figure 14). As a result o f the residual nitrogen at the base of the euphotic zone {Figure 23), the biomass within the SCM increased by two-to three-fold throughout the year {Figure 24), relative to the first case (Figure 15). In addition, the depth of the SCM was shallower, 65-75 m (Figure 24), by 30-40 m than in the first case {Figure 15). In April, 1982, within a Loop Current cold-core eddy, a maximum of 0.7 mg chl m 3 within an SCM at 40 m was observed {Yoder and Mahood, 1983). Note that, relative to the 1% light depth, the SCM in this case still occurred somewhat shallower than the SCM in the no-pulse case (top panels in Figure 23 versus Figure 14). The maximum ratio of SCM-tosurface-chlorophyll concentration during the summer was 7.3

PAGE 103

0.8 0 0 6 .... t 0 '-1 lL. E .., 0 z 0 E E Figure 22. . ..... ' ..... .... 0 0 2000 1000 10 Jon Mar May Jul Sep Nov Jon Month F-ratio in the spring case. Solid line = total phytoplankton, dotted line = netplankton, dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate. CD CD

PAGE 104

e e /:. Q. 0 e /:. Q. 0 -tO 150 1 ......... I r '0 0 j -J \' J l I I IQ 0 _ I I 1 l I 200 r I' 10 0-0 -: I I i --j ' i ' ... i Nov Jon Jon Mar May Jul Month Sep Fiqure 23. The annual variation of phytoplankton biom ass (mg chl m-3 ) and nitrogen in the spring case. Top panel: total phytoplankton a n d euphotic zone depth (dashed line) Middle panel: netplankton and nitrate (dashed contour lines; deepest= 18 Bottom panel: picoplankton and ammonium (dashed contour lines) 8 9

PAGE 105

1.0 .., E 0' 0.6 E :c u ::::i u t/) ., u 0 ..... :l t/) u t/) ...... E .c Q. ., 0 u Vl 0.4 0.2 0.0 10 8 6 t 4t-I 2t --\_ j ., 0 i or t j I ..... I ] I .. I l ., I 150 Nov Jon Jon Mar May Jul Month Sep 90 Figure 24. The annual variation of the subsurface chlorophyll biomass maximum (SCM) in the sy,ring case. Top panel: total biomass (mg chl m-) Middle panel: ratio of SCM biomass to surface (5 m) biomass. Bottom panel: depth (m) of the SCM.

PAGE 106

(Figure 24), slightly lower than that in the first case, without the Loop Current influence (Figure 15). This is because summer surface chlorophyll concentrations (Figure 25) were five-fold higher than during the no-pulse case (Figure 17). Surface chlorophyll concentrations (Figure 25) were always higher than those in the no-pulse case (Figure 17). During the spring eddy events, surface biomass reached 0.5 mg chl -3 m declining to <0.1 mg chl m-3 during the summer, and rising again to 0.9 mg chl m-3 by October. These 91 simulated values no longer mimic the basin-wide means of the CZCS estimate, reflecting instead the results of the 3D model. If one subsamples the satellite data set, in space and time, however, the results of both the 1D and 3D models are corroborated by CZCS imagery in the eastern Gulf of Mexico during the winter of 1981 (Figure 26). At the beginning of 1981, almost 0.4 mg chl m-3 was recorded for this region by the CZCS, compared to ca. 0.2 mg chl m-3 in 1979, 1980, or 198 2 (Muller-Karger, et al., 1991). In this same region, satellite thermal imagery (Figure 27) recorded the cold water (<21C) surface expression of a large cyclonic perturbation along the edge of the Loop Current, on 22 February, 1981, compared to warm water (>26oC) of the Loop Current proper on 21 February 1980 (Vukovich and Maul, 1985). A series of cyclonic eddies

PAGE 107

-"') 1 0 0.8 E 0 6 0' E ...._, 0 :c u N E z 0 E E 0 4 I / / -------. -----1000 500 0 Jon Mar May Jul Mon th Sep Nov Jon Figure 25. The annual variation of surface (5 m) chlorophyll biomass in the spring case. Solid line = total phytoplankton; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate.

PAGE 108

-(") 35 I E .30 0'> .25 E -Model Chi -t-.20 z .15 6 czcs w .10 CJ .05 a.. 0 1981 1982 YEAR Figure 26. Annual variation of simulated surface chlorophyll in the eastern Gulf of Mexico from Walsh, et al.'s (1989) model, and of 1979-82 CZCS estimates (Muller-Karger, et al., 1991) of pigment concentration for a 200x200 km2 region in the eastern Gulf of Mexico (A and C in Figure 3) Modified from Walsh, et al., 1989. \0 w

PAGE 109

94 29 28 27 26 25 24 23 90 89 88 87 86 85 84 83 82 81 80 90 89 88 87 86 85 84 83 82 81 80 Figure 27. NOAA sea surface temperature (C) analysis. Top: 2 1 February 1980; dot is station 1 of ElSayed and Trees, 1980. Bottom: 22 February 1981; dot is Ortner, et al.'s (1984) station. The dashed box is the 200x200 km2 region used to construc t a CZC S surface pigment time series.

PAGE 110

95 leading into a similar, large cold perturbation was observed in thermal imagery during spring, 1983 (Figure 5). Photosynthesis within recently upwelled water and/or entrained coastal populations of phytoplankton, during subsequent offshore transport within the cyclonic perturbation, may have led to the high surface concentration of 0.4 mg chl m -3 sensed by the CZCS only in January and February, 1981 (Figure 26). The spring-pulse case of this model simulated such a transient signal, as sensed by satellite (Figure 25), unlike the no-pulse case (Figure 17), which typified the conditions of 1979, 1980, and 1982 (Figure 26). As a consequence, particulate fall-out increased in the second case of the model. The grazing loss of phytoplankton and the export of organic mat-ter in this second case are summarized in Table 6, Figure 28, and Figure 29. Grazing losses, summed over the 200-m deep model domain, again accounted for >99.9% of primary production. Phytodetrital flux at 2 00 m represented <0.1% of the total export. This same flux measured at 100 m, roughly the base of the euphotic zone, however, was 22.0 g c m -2 yr-1 a three-fold increase over that in the no-pulse case. Netplankton flux represented 90% of this increased tota l phytodetrital flux. There also was -2 -1 t a three-fold increase in POC export, to 97.5 g C m yr a 200 m, in response to nitrate injection, relative to the no-pulse case.

PAGE 111

I u N I E u C7' E N I 1000 E z 0 500 E E 0 Jon Figure 28. ,... \ .... ,. ...... >" / I \ \ I ,/ I \ /\ / / '\ \ / I ' / ..... "" / ........ Mar May Jul Sep Nov Jon Month The annual variation of grazing loss in the spring case. Solid line = total phytoplankton; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 rn) nitrate.

PAGE 112

-' "0 "" "' I E u C1' E E z 0 E E 500 \ '. / ..... / / / "" I' 0 -1 / 1000 -500 0 Jon Mar May Jul Month Sep Nov Jon Figure 29. The annual variation of export loss in the spring case. Only particulate losses are considered in this plot-see text. Solid line = total phytoplankton; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 rn) nitrate.

PAGE 113

98 For comparison, in the western Gulf of Mexico, mass flux at 50 m within a cyclonic ring was 24-fold that in an anticyclonic ring, originating from the Loop Current; at 200 m, this difference increased to 33-fold (Biggs, 1992). Some of the horizontal disparity i n mass fluxes in the western Gulf may, i n fact, be the result of offshore transport of shelf-derived material by anticyclone-cyclone eddy pairs in this region, as see in surface salinity (Nowlin, 1972), temperature and Legeckis, 1982), and color (Muller-Karger, et al., 1991) signals. The export, via diffusion, of ammonium was 5.95 mmol m-2 -1 y In carbon equivalents, this represents 0.473 g C m -2 y-1 of consumed phytoplankton, less than 0.5% of the POC efflux, but ten-fold greater than the phytodetrital flux at 200 m. Loop Current Present During Summer A third model case was run to assess the potential impacts when the Loop Current i s instead present along the west Florida slope during summer. The pattern of eddy simulation was the same as in the spring case, except that the eddies were introduced beginning on August 15. As before, the results are from the fifth year of the model simulation, after model initialization with conditions at

PAGE 114

the end of the fifth year of the first case, when the Loop Current was not present. 99 This seasonal shift of nutrient impact increased the annual primary production by ca. 2.5-fold over the no-pulse case, but the total production of 209 g c m-2 yr-1 was 11% less than that when eddies were present during the spring (Table 5). The netplankton share of total annual primary production (70%) was similar to that in the spring case (72%). Recall that with no nitrate injection, the netplankton share of a much smaller primary production was 54%. On a daily basis, primary production ranged from 149 mg C m-2 d-1 in May, prior to the eddy events, to 1220 mg C m-2 d-1 during the fall convective overturn in October (Figure similar to a maximum of 1200 mg C m-2 d-1 observed in September, 1982, within a Loop Current cyclonic eddy (Yoder and Mahood, 1983). Netplankton were responsible for most of the primary production during the winter and spring, while picoplankton produced a larger fraction during the summer and fall (Figure 30) Unlike the previous cases, the maximum daily production in this third case occurred during the fall, three months after the onset of the simulated summer eddy events. This result reflected the prevailing summer and fall vertical mixing conditions (Table 3), and in consequence, the store of nitrate (Figure 31) just below the euphotic zone. The

PAGE 115

-I u .... I E u 0' E N I E z 0 E E 1500 1000 """' "' \. "' 500 '-""'--------== ... . -/ -.: 0 ,____ __ __,_ ___ __L_ _ L.:=._ ...... 1000 5 0 0 Jon ------I Mar ---l May --...L Month r I I / I I / \1 I I \ I I I I / \ \ / -= !--=---.::::_ -_l -_...L___ _l_-=--.. J _l Sep Nov Jan Figure 30. Integrated (0-200 m) p r i mary production in the summer case. Solid line = total production; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated ( 0-200 m) n i t rate. ...... 0 0

PAGE 116

101 nitrate injected during the summer simulation, to a large extent, remained unused until the onset of stronger winds and more vigorous vertical mixing in the fall (Table 3) Given the lack of lateral advection, i.e., spatial resolution, in the model, this local storage of inorganic nitrogen may be an artifact, leading to prolonged higher algal production (Figure 30) and biomass (Figure 32) a few months after the simulated eddy events. New production in this case was 88.1 g c m-2 yr-1, three-fold higher than that in the no-pulse case, but 12% lower than that i n the spring-pulse case (Table 5) The annual f-ratio in this case was 0.41, similar to the 0.42 value of the spring-pulse case. The influx of N03 at 200 m, supporting this production, was 1040 mmol N m-2 yr-1 The daily f-ratio in this summer case (Figure 33) did not have a large spike during the eddy events, as was seen in the spring case (Figure 22), but was still somewhat higher than the no-pulse case (Figure 12). In fact, the fratio of the summer case was higher during winter than during the summer eddy events, a consequence of the storage of nitrate just below the euphotic zone, under summertime stratification. The annual cycle of integrated algal biomass reached a maximum of 98.1 mg chl m 2 in the fall, and a minimum of 8.66 mg chl m 2 during the late spring transition from a netplankton-dominated to a picoplankton-dominated community

PAGE 117

-E -100 Q. 0 -E -E -Q. 0 150 50 150 ' ' ' \ \ I May Jul Month ' Sep Nov -1 -J j I I -; Jon Figure 31. The annual variation of phytoplankton biomass (mg chl m-3 ) and nitrogen in the summer case. Top panel: total phytoplankton and euphotic zone depth (dashed line) Middle panel: netplankton and nitrate (dashed contour lines; deepest = 18 Bottom panel: picoplankton and ammonium (dashed contour l ines). 102

PAGE 118

100 80 "' .......... "' I E 60 (J"I E 0 :c u 2 0 0 N I 1000 E z 0 500 E E 0 L___ __ _.__ ___ ___l___ -'------'--.Jon Mar May 1 ,Jul Month lSep \ \ \ I I ( I I I I Nov / I I Jon Figure 32. Integrated (0-200 m) chlorophyll in the summer case. Solid line = total chlorophyll; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate. 1-' 0 w

PAGE 119

104 (Figure 32) The annual mean algal stock in this case was 44.0 mg chl m 2 in comparison with 46.2 mg chl m-2 from the spring case, and 52 mg chl m-2 from the prior 3D model (Walsh, et al., 1989). During the summer and early fall, picoplankton comprised 80-90% of phytoplankton biomass; during winter, netplankton constituted >90% of phytoplankton biomass (Figure 31). As in the spring-pulse case, the SCM was significantly enhanced by the additional source of nitrate, with concentrations between 0.8 and 1.3 mg chl m-3 following the summer eddies and into the fall (Figure 34). Again, the depth of the SCM was shallower, 25-95 m, than in the nopulse case, in concert with a shallower nitracline. The SCM:surface-chlorophyll ratio reached a peak of 7.5 within two weeks after the first eddy event. By mid-September, however, this ratio dropped to just 1.2-1.5. Total surface (5 m) algal biomass reached a maximum of 1.3 mg chl m -3 during late October (Figure 35), about twice that of the autumn bloom during the spring case (Figure 25) Picoplankton peaked, at 0.95 mg chl m-3 a couple of weeks prior to this combined bloom, while netplankton had a broad maximum of ca. 0.75 mg chl m-3 during early winter. As an extreme in the real world, autumn red-tides (Carder and Stewart, 1983; Vargo, et al., 1987), formed by Ptychodiscus brevis ( = Gymnodinium breve), which may be triggered, along the outer shelf, by such summer Loop Current intrusions

PAGE 120

1 0 0 8 06 0 0 ..... I LL. "' I E .., 0 z 0 E E 0 0 2000 1000 0 Jon r ........ Mar May \ \ \ \ ' Jul Month I --_, r I / I r Sep Nov Jon Figure 33. F-ratio in the summer case. Soli d line = total phytoplankton; dotted line = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate.

PAGE 121

2.0 I E at E -&. u u V1 0.5 u 6 2 ... :J V1 4 u V1 2 ol 0 ......... E s::. a. 0 u V1 150 200 Jon I ) l j \_ J J_]\ j I I --, / L__._._., Mor Moy Jul Sep Nov Jon Month 106 Figure 34. The annual variation of the subsurface chlorophyll biomass maximum (SCM) in the summer case. Top panel: total biomass (mg chl m -3 ) Middle panel: ratio of SCM biomass to surface (5 m) biomass. Bottom panel: depth (m) of the SCM.

PAGE 122

107 (Haddad and Carder, 1979), can lead to concentrations of 2.5-4.5 mg chl m-3 on the west Florida shelf, but inclusions of another, more complex functional group was beyond the scope of the present simulation analysis. Figure 31 instead summarizes the simple dynamics of this model's two-component phytoplankton biomass, in relation to the unutilized nitrogen and subsurface irradiance distributions of the third case. As in the previous cases, total biomass had a January SCM centered around 45 m predominantly composed of netplankton. During the summer, the vertical structure shifted to a picoplankton-dominated SCM at the base of the euphotic zone, coincident with the nitracline. The three eddy events can be seen in the middle panel of Figure 31 as three spikes in the nitrate contours (dashed lines), during July and August. While nitrate in the surface waters remained <0.1 there is a residual source of nitrate between depths of 150 and 200 m. Winter vertical mixing and subsequent algal growth, first by picoplankton, and then by netplankton, somewhat depletes this stock. Comparison of integrated nitrate stocks (Figure 30) with the no-pulse case (Figure 11), however, indicates that nitrate concentrations at depth in the eddy cases do not return to those in the no-pulse case. similarly, ammonium accumulated underneath the euphotic zone as a result of grazing (Figure 36) upon the sinking

PAGE 123

........... .., I E t:1' E .._, 0 :c u "' I E z 0 E E F igure 1 5 T--0 5 0 0 1000 500 \ \ / \ I I I \ I : = _l._ _j_.......J.._l __ /_1 1 1_L_ \ '_L_ ._ _: -=-'--' 0 ____ _L_ Jan 35. Ma r May Jul Sep Nov Jan Monl h The annual variatio n of s urface (5 m) chlorophyll biomass in t h e summe r case. Solid line = total phytoplankton; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 rn) nitrate. 1-' 0 CD

PAGE 124

109 phytodetrital flux, with >0.25 NH4-N simulated in both the spring (Figure 23) and summer (Figure 31) cases, compared with <0.15 NH4-N in the first case (Figure 14). Netplankton are the major source of recycled nitrogen in the aphotic zone of the model, allowing for some downward mixing of picoplankton. At a settling velocity of 2.5 m d-1 it would take netplankton cells 40 d to travel from 50 to 150 m, while subjected to grazing stress. On the other hand, picoplankton, whose production was greatest in this summer case (Table 5) contributed little to the ammonium pool at depth; recycling of nitrogen by the microbial loop takes place near the source of picoplankton production, within the euphotic zone. As a consequence, the ammonium efflux at 200 m of the third case,. 5.77 mmol NH4-N m 2 yr-1 i s somewhat larger than that in the no-pulse case, 4.68 mmol NH4-N m-2 yr-1 as a function of an increased source from faster settling netplankton, grazed upon below the euphotic zone (Table 6) Although there still was a significant phytodetrital flux of 17.6 g c m-2 yr-1 at 100m in this case, grazing (Figure 36) consumed most of these cells in their transit through the second 100 m of the model domain (Figure 37). The small phytoplankton flux out of the model domain of 0.496 g C m -2 yr-1 at 20 0 m was composed of 90% netplankton cells, as a result of their higher sinking speed.

PAGE 125

-' u N E u C1' E N E z 0 E E Figure 1000 / I I ....... I jl \I l 500 ', I \ I I I \ I -/ \ / / / 0 1000 500 0 .__ __ _,_ __ __ ._ ._._ ___ ____ __L ___ _L ___ __J_ __ __jt__ __ ....L.... __ ___. __ Jon 36. Mar May ,)ul Sep Nov Jon Month The annual variation of grazing loss in the summer case. Solid line = total phytoplankton; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 m) nitrate.

PAGE 126

-;-"0 N I E u 0" E .... I E z 0 E E 600 400 200 0 1000 500 0 "\ ""' r ----I / r-/ / / I I / .____ __. __ ___._ _ _.__ _ .. L r--._ _ __ _,_ __ l _j -------------' L._ _ ...._ _ --L. ___ J.. _ __. ____ IL---.---L-Jon Mar May Jul Month Sep Nov Jon Figure 37. The annual variation of export loss in the summer case. Only particulate losses are considered in this plot-see text. Solid line = total phytoplankton; dashed = netplankton; dash-dotted = picoplankton. Lower panel: integrated (0-200 rn) nitrate.

PAGE 127

112 DISCUSSION Discerning the cycles of b ioactive elements, notably but not exclusively carbon and nitrogen, may require an appreciation of processes on scales as small as those of the diffusion boundary layer for nutrient uptake by cells, and of turbulence, to those at least as large as the prominent oceanic current systems. The associated time scales range from seconds to years. Indeed, the residence time for dissolved inorganic carbon in deep waters is on the order of 1000 years (Broecker and Peng, 1982). Ecosystem simulation models typically focus on a subset of these space and time scales, necessarily simplifying processes which cannot be resolved or for which sufficiently developed theories or data are lacking. Longhurst (1991) has discussed the trophodynamically complex role that biota play in global biogeochemical cycles: The geochemical disequilibrium of our planet is due mainly to carbon sequestration by marine organisms over geological time. Changes in atmospheric C02 during interglacial-glacial require sequestration of carbon 1n the limited export flux from n7w 1n waters is the key process 1n th1s sequestrat1on. The most common model for export flux ignores important nutrient sources and export mechan1sms. Export flux occurs as a result of biological processes whose complexity appears not to be the principal classes of simulation models, th1s be1ng

PAGE 128

especially true for food webs dominated by singlecelled protists whose trophic function is more dispersed than among the multicelled metazoa. 113 Longhurst illustrated his point with examples of protistan species who function as both autotrophs and heterotrophs (mixotrophy), and similarly, of phytoplankton, notably among the dinoflagellates, who are heterotrophs. In addition to these confusions, from a modeler's viewpoint, of trophic function, there are ontogenetic and seasonal shifts in diet among the multicelled grazers. Summertime grazing on protists by copepods is both an example of linkages between the microbial loop and the macro food web, and another means of controlling protistan populations. These complexities have been left out of the current model, but only for lack of validation data. As data on the size structure and seasonal trophodynamics of the micro-, meso-, and macrozooplankton accumulate, this and other ecosystem models can be modified to incorporate increasing levels of complexity. While computer simulations will never capture the variability and complexity of the natural environment, they can still provide insight by examining presumed important aspects of physical forcings, elemental cycling, and ecological linkages or sinks, in conjunction with appropriate field measurements and process-oriented studies. Observations of phytoplankton biomass, e.g., as chlorophyll, and of nitrogen, e.g., as nitrate and ammonium, are the remnants of production, grazing, predation, and

PAGE 129

114 recycling processes. Furthermore, there are various pathways, food webs composed of subsets of the ecosystem's community, operating simultaneously, within which these processes take place. The bifurcation model, with a choice between a large, loosely-coupled food web and a smaller, tightly-coupled food web, represents but one level of complexity added to a classic nitrogen-phytoplanktoncopepod-fish paradigm of primary production and its fate in the sea. Certainly, many components, as suggested by Longhurst, can be added to follow the fate of an assimilated carbon or nitrogen atom (e.g., Pace, 1987; Michaels and Silver, 1988; Hofmann and Ambler, 1988; Fasham, et al., 1990). However, only four state variables were modelled here in order to focus on the initial "choice" in the bifurcation-model (Figure 2). Specifically posed, upon perturbation with a source of new nitrogen, what changes will there be in the balance of production between the large and small food webs? Furthermore, what are the consequences, in terms of export from the system, for primary production in this perturbed system? The Model's carbon Flow: Fate Determined at The First Bifurcation A specific hypothesis was that primary production would be shifted to phytoplankton within the larger size fraction. comparing the spring case with the no-pulse case, net-

PAGE 130

115 plankton primary production increased by 3.75-fold, versus a 1.72-fold increase in picoplankton production (Table 5). Similar increases favoring netplankton production were shown in the summer case (Table 5) In model cases with nitrate injections, netp1ankton accounted for >75% of new production, versus 58% in the no-pulse case (Table 5) Most of this netplankton production occurred between late fall and early spring, when higher mixing conditions prevailed (Table 3), in all cases (Figure 11, Figure 20, Figure 30). Surface phytoplankton biomass, with no enhancement of nitrate flux, i.e., simulating a water column that is not impacted by the cyclonic edge of the Loop Current, exhibited a persistent winter high of 0.18-0.20 mg chl m-3 between October and February, similar to values along the southwest Florida slope in monthly-composited CZCS scenes during 19791980 (Muller-Karger, et al., 1991). With nitrate injection in this one-dimensional, two-component phytoplankton model, surface phytoplankton biomass reached ca. 0.5 and 1.3 mg chl m -3 in the spring and summer cases, respectively. Surface chlorophyll in the spring case, however, had a bimodal annual cycle; a larger autumn bloom of >0.8 mg chl m-3 resulted from seasonal mixing of residual nitrate stored below the euphotic zone. In the previous singlephytoplankton, three-dimensional basin model, surface chlorophyll was unimodal, with a maximum of ca. 0.4 mg chl

PAGE 131

116 -3 h m at t e end of January (approximately, model day 250) in the eastern Gulf of Mexico (Figure 26; Walsh, et al., 1989). For the Loop Current cycling used in that model, this time period corresponded to the formation of a large cyclonic perturbation off of the Dry Tortugas, to be followed by eventual ring separation. Such a pattern can also be seen in thermal imagery. In particular, Vukovich and Maul (1985) present images for February, 1980 and 1981 which show such a perturbation (Figure 27). In the latter year, the perturbation was more strongly developed, with pools of surface waters of <23C, as compared with >25C Loop Current water, and extended westward into the region (Figure 3) subsampled for a CZCS time series (Figure 26; Muller-Karger, et al., 1991). The resulting monthly composited chlorophyll signature of 0.33 mg m -3 for February, 19 81 is two-fold greater than that for February, 1980 (Muller-Karger, et al., 1991). Contemporaneous with these remotely-sensed observations, and as an example of the extent of spatial and temporal variability in the region, Ortner, et al. (1984) made a week-long series of productivity and biomass observations, at a site (25.5N, 87.0W) within warm Loop current water, just outside of this same perturbation (Figure 27). During that time, conditions at the site were influenced by the passages both of a meteorological front, deepening the mixed layer to >110 m, and of a warm, low-

PAGE 132

117 salinity surface lens, similar in properties to those observed along the Florida slope (Ednoff, 1974; Paluszkiewicz, et al., 1983). The impact of these effects was a four-fold increase in productivity (Table 1) and an apparent increase in upper water column phytoplankton biomass (equipment failure prevented conversion of fluorescence to chlorophyll during a portion of the study). By the end of the study, surface chlorophyll values were 0.15 mg chl m -3 ca. half of that of the monthly CZCS composite's value in the region of the cold perturbation. Not only is this temporal and spatial variability indicative of the difficulties of reconciling data sets with one another and with model results, particularly those based on climatological forcings, but it also illustrates the interactionof various real-world physical mechanisms (eg., weather, vertical mixing, and lateral transport) with a water column's biological properties. Particulate Organic Matter Flux A corollary hypothesis was that a shift in production to the larger size fraction should yield a concomitant increase in the export of organic matter. Indeed, phytodetrital fluxes at 100 m, roughly the base of the euphotic zone, increased by over three-fold, ca. 85% of which was netplankton, with eddy-induced nitrate influx

PAGE 133

118 (Table 6) Relative to the no-pulse case, netplankton flux at 100 m increased by nearly four-fold, while picoplankton increased by ca. 1.5-fold, with increased nitrate flux along the cyclonic edge of the Loop Current (Table 6) Unlike the phytodetrital flux at 100 m, the flux at 200 m showed no such increase; in fact, this flux diminished by ca. 2% for model cases with nitrate injections over that in the no-pulse case (Table 6). Grazing activity between 100 m and 200 m consumed all of the resultant increased phytoplankton biomass. Thus, the model results imply that organic flux past 200 m is predominantly secondary export: organic matter processed through grazers, and their predators, as fecal matter, mortal remains, and transport via migration. These processes were not modelled, but are potentially significant pathways in the real ocean (Urrere and Knauer, 1981; Pilskaln and Honjo, 1987; Longhurst and Harrison, 1988). In the model, instead, such processes and pathways were subsumed into grazing loss and recycling efficiency parameters. While secondary export from the microbial loop increased by less than two-fold, that from the macro-grazer food web increased by ca. 3.5-fold and comprised ca. 85% of the total secondary export (Table 6, Figure 19, Figure 29, Figure 37). That the majority of the particulate flux in situ results from zooplankton grazing, and the subsequent formation of rapidly sinking fecal matter, can also be seen

PAGE 134

119 in the isotopic signature of nitrogen. Altabet (1988) examined the nitrogen isotopic composition of different size classes of both suspended (collected in Niskin bottles or by large-volume pumps) and settling (collected by sediment traps) particulate matter. Within the euphotic zone, suspended particulate nitrogen (PN) in the size fraction had an isotopic signature (S15N = -1 to 1 per mil2) similar to that for air. Additionally, the size fraction had values of ca. 1 to 0.0; most of the chlorophyll was in this size fraction. Suspended PN in size fractions greater than 75 however, had S15N values of 4-6 per mil (i.e., the material was relatively enriched with the heavier 15N isotope) The average S15N of total PN within the euphotic zone was -0.2 per mil. In contrast, PN in all size-fractions sampled from below the euphotic zone was enriched in 15N by ca. 4-6 per mil. Furthermore, settling PN had an average S15N of 3.7 per mil, similar to the 3.5 per mil value for nitrate. This isotopic shift in the presumptive fecal matter that comprises the bulk of the organic matter caught by sediment traps below the euphotic zone results from the biochemically preferential assimilation of the lighter 14N isotope during the digestion of phytoplankton by herbivorous zooplankton. It is also of interest that the minipellets described by Gowing and Silver 2 SlSN (per mil) = [ (R 1 /RN ) -1] x 1000, R = samp e 2, air 15N/14N.

PAGE 135

120 (1985), which are, i n part, comprised of fecal matter from radiolaria and other protistans, fall into the size range of 3-50 and may constitute some of the larger suspended PN, with high o15N, sampled by Altabet (1988). Further examination (Altabet, 1989) showed significant correlations among mixed layer depth, PN within the upper 100 m, and nitrate flux and PN flux at 100 m, based on a bimonthly sampling regime i n the Sargasso Sea. Seasonal correlations of hydrography (mixed layer depth and nitrate concentration, in particular) and primary productivity (Menzel and Ryther, 1961) were further supported by the positive relationships of mixed layer depth and euphotic zone PN with PN flux at 100 m (Altabet, 1988). Indeed, an 80% increase in integrated (0-100 m) PN was correlated with a seven-fold increase i n PN flux at 100 m (Altabet, 1988). If it is assumed that the increment of new nitrate-nitrogen from deep winter mixing resulted in a seasonal increase in the larger phytoplankton size fraction (Chisholm, 1992), then this increased particulate flux represents the correlated seasonal signal of macrozooplankton standing stocks (Deevey, 1971), supporting the basic premise of the bifurcation model (Figure 2; Legendre and Le Fevre, 1989) and its impact on the fate of carbon in the netplanktonmacrozooplankton food web versus the microbial loop, viz. the downwa r d POC flux.

PAGE 136

121 Whil e i t i s possible to envision the particulate organic matter (POM) flux resulting from the large food web, by the processe s o f egestion, migration, mortality, and predation, the microbial loop consists of organisms of a size unlikel y to produce particles with significant settling velocities. Although there are observations of mini-pellets from presumed microbial loop consumers (Gowing and Silver, 1985), the majority o f the organic matter flux from this food web may in f act, be of dissolved organic matter (DOM), transported downward b y d iffusion and downwelling (Toggweiler, 1989). I n the present model, however, grazing loss and secondary export were treated as implicit sinks for phytoplankton-derived POM, with no creation, consumption or transport of DOM. Energy Available to Higher Trophic Levels The model's grazing loss can be converted into energy units, to be c o mpared with trophic models of macrozooplankton and microzooplankton food webs. Total zooplankton b iomass in the eastern Gulf of Mexico is ca. 737 mg dry weight (dw) m-2 through the upper 200 m, and ca. 1240 mg dw m-2 i n the upper 1000 m (Hopkins, 1982). Of this total biomass, taxa identified as herbivorous, omnivorous, or detritivorous comprise 68% of the surface water (0-200 m) and 35% of the water column (0-1000 m) zooplankton standing

PAGE 137

stock (Hopkins, 1982). Copepods represent 49% and 60%, respectively, of the day and night zooplankton biomass, followed by euphausiids and chaetognaths; together, these three groups comprise ca. 80% of the total zooplankton standing stock (Hopkins, 1982). 122 In addition, this biomass i s highly structured with depth. Total zooplankton biomass declines, from the upper 3 0 m, by three-fold at 50-150 m, and by twenty-fold at 550 m (Morris and Hopkins, 1983). Furthermore, species within the various herbivorous and carnivorous functional groups delineate their niche spaces by remaining within specific depth zones, while others may migrate over 100 m per day (Morris and Hopkins, 1983). Still, it can be assumed that the energy supply over this extended water column originates from photosynthesis within the overlying euphotic zone. Conversion of grazing zooplankton (this included omnivores and detritivores; Morri s and Hopkins, 1983), all w ithin the larger food web, from biomass to caloric content yields 2.48 kcal m-2 within the upper 1000 m. This assumes that grazer biomass is roughly 50% of total zooplankton biomass, or 0.62 g dw m-2 within this same depth range. Conversion o f modelled netplankton primary productivity, ranging from 45.3 g c m-2 yr-1 (no-pulse case) to 169 g C m-2 yr-1 (spring case), yields an energy source to higher trophic levels of 516-1920 kcal m -2 yr-1 assuming a conversion for phytoplankton of 11.4 kcal (g C)-1 (Platt and

PAGE 138

123 Irwin, 1971). If, as occurred in the model, all of this primary production were consumed by zooplankton, with a gross growth efficiency (growth/ingestion) of 20% (Walsh, 1988), then one would expect the macro-zooplankton biomass annual increment available to predators, or for export by other means, to be 9.1-33.8 g c m-2 yr-1 or 18.1-68.0 kcal -2 -1 m yr over the range of cases in the 1D model, assuming conversions of 0.5 g C (g dw)-1 and 4 kcal (g C)-1 for copepods (Morris and Hopkins, 1983). Assuming that the consumed phytoplankton translates into zooplankton production, this would imply a zooplankton production-to-biomass (P:B) ratio of ca. 4-14. Although essentially all of the primary production in the present model was grazed within the upper 200 m of a simulatedslope water column, in a previous carbon budget for the Florida shelf, using a low primary production of 30 g C m-2 yr-1 and a P:B of 7 (see Steele, 1974), zooplankton grazed only 2% of the primary production (Walsh, 1983). As stated earlier, in the absence of aggregation mechanisms or other processes which would enhance the downward flux of uneaten phytoplankton cells, loss through grazing may be somewhat overemphasized in the present model, although not by much, if the nutritional demands of zooplankton are to be met.

PAGE 139

124 Reconciliation of Disparate Data Sets and Model Results In addition to hypothesis testing, another reason for modelling is to attempt to fill in the gaps, with respect to both time and space, in our observations of marine ecosystems. Lacking a more complete description of the three-dimensional temporal and space variability of the physical circulation in the Gulf of Mexico, the present "ecosystem" model merely sought to confirm the anticipated range in primary productivity and carbon flux possible given the disparate nitrate influxes in Gulf Common Water and at the cyclonic edge of the Loop Current, where the nitracline is sharper and shallower. Additionally, the model investigated potential pathways for the fate of the resultantcarbon by including two size fractions of primary producers, with size-dependent consequences for phytoplankton loss. In situ observations of relatively low salinity lens (<35.5 psu) associated with Loop Current-dominated circulation in the eastern Gulf of Mexico (e.g., Ednoff, 1974; Paluszkiewicz, et al., 1983; Ortner, et al., 1984), as well as time series of remotely-sensed sea surface temperature (SST; Vukovich and Maul, 1985; Vukovich, 1986) and pigments (Muller-Karger, et al., 1991), suggest strong laterally advective processes. A one-dimensional model cannot resolve processes resulting from such lateral

PAGE 140

125 influences. In the spring case of this study, the storage of injected nitrate beneath the euphotic zone prior to the onset of deep winter mixing led to autumnal blooms of algal biomass, ca. six months after the simulated injection events. It is possible, however, that the eddy field may move through the slope region and subsequently offshore before complete utilization of the upwelled nitrate, as happens along the southeastern U.S. margin (Yoder, et al., 1983; Hofmann, 1988; Lee, et al., 1991). If so, the predicted annual productivity, biomass and mass fluxes would be too high. Similar to the observations of lateral transport cited above, in a case of the previous three-dimensional model that included nitrate-rich effluent from the Mississippi River, but which did not resolve shelf circulation patterns normally responsible for advecting this effluent along the Louisiana-Texas coastline (eg., Cochrane and Kelly, 1986; Dinnel and Wiseman, 1986), persistent plumes of surface chlorophyll stretched towards the southeast (Figure 19 in Walsh, et al., 1989). While modelled currents at the Florida shelf break reversed directions from southeasterly to northwesterly at various stages of Loop Current penetration (Figure 20 in Walsh, et al., 1989), there were periods in the Loop Current cycle when southeastward currents were sufficiently persistent and strong enough to advect biomass along the southwest Florida margin, similar

PAGE 141

126 to observed low salinity water masses (Atkinson and Wallace, 1975), red-tide trajectories (Murphy, et al., 1975), CZCS imagery (Trees, 1985), and detached buoys (Schroeder, et al., 1987). This lateral transport of biomass into the region of interest for the present one-dimensional model, however, had little apparent impact on loca l annual primary production. In a comparison of model cases with and without Mississippi River nitrate effluent, primary production off of southwest Florida increased by 4%, whereas average integrated biomass and particle flux increased by 1 4 % and 19%, respectively (Walsh, et al., 1989); most of the phytoplankton advected into the region had already sunk below the euphotic zone (recall that this model was strictly nitrate-based, and had no grazing loss). In general, though, both models, with nitrate injection either from the Mississippi River, or from beneath the euphotic zone, have similar annual mean integrated chlorophyll and primary production, but dissimilar f-ratios and vertical chlorophyll structure. I n reality, lateral transport from the northern Gulf of Mexico of organic matter, in the form of pigments, and perhaps of colored dissolved organic matter, both of which contribute to plumes detected by the CZCS (Carder, et al., 1986; Muller-Karger, et al., 1988; Hochman, 1992), may allow for recycling and, thus, continued enhanced algal production along the plume's path. This would be in addition to the

PAGE 142

127 vertical nitrate influxes resulting from the Loop Current impinging on the southwest Florida slope. Indeed, Yoder and Mahood (1983) hypothesized such a recycled nitrogen component i n order to explain observations of outer shelf productivities as high as those observed along the Loop Current-impac:ed (Table 2), when no allochthonous nitrate source scppcrting those rates could be identified. "Models and effcr:s to discern the patterns of elemental cycling in :he ccean, and the important scales cf variability, both i n the forcings and the biological responses, require both observation and simulation. Clearly, the "answers" to be found in disparate in situ and remotely-sensed data sets, and the inclusion of processes within a computer sufficient to yield similar results, are often by the frustratingly high degree of variability at a wide range of space and time scales in oceanography. a better understanding of marine ecosystems still requires enhanced descriptions of both physical and biological processes. The title o f this section comes from Hedgpeth (1977), whose philosophical observations serve as both a caveat and a challenge, t o extract meaning and relevance from meager samplings and oversimplifications of the real world:

PAGE 143

128 There i s o f course no inherent evil in attempting to simplify what we know or suspect of nature so that we can handle t h e almost infinite variations of events in the natural world, and perhaps to arrive at some modest hope G f prediction. Unfortunately, however, many, and for most Fart those not directly concerned with modelling see in equations facts rather than ideas.

PAGE 144

129 REFERENCES Alldredge, A.L. and C.C. Gotschalk, 1989. Direct observations of the mass flocculation of diatom blooms: characteristics, settling velocities and formation of diatom aggregates. Deep-Sea Res., 36, 159-171. Altabet, !-1.A., Variations in nitrogen isotopic composition between sinking and suspended particles: implications for cycling and particle transformation in :he cpen ocean. Deep-Sea Res., 35, 535-554. Altabet, M.A., 1989. Sargasso Sea. ?articulate new nitrogen fluxes in the J. Geophys. Res., 94, 12771-12779. Atkinson, L.?. and 0. 1975. The source of unusuall y low surface sa11nities in the Gulf Stream off Georgia. Deep-Sea Res., 23, 913-916. Auer, S. J., 1987. Five-year climatological survey of the Gulf Stream system and :ts associated rings. Geophys. Res., 92, 11709-11726. Azam, F., T. Fenchel, J.G. Field, and F. Thingstad, 1983. The column microbes in the sea. 257-263. J.S. Gray, L.A. Meyer-Reil ecological role of waterMar. Ecol. Prog. Ser., 10, Behringer, D.W., R.L. Molinari and J.F. Festa, 1977. The variability of anticyclonic current patterns in the Gulf of Mexico. J. Geophys. Res., 82, 5469-5478. Bien fang, P. K. l 9 8 0. ?hyt oplankton sinking rates in oligotrophic waters off Hawaii, USA. Mar. Biol., 61, 69-77. Bienfang, P.K. and W.G. Harrison, 1984. Sinking-rate response of natural assemblages of temperate and subtropical phytoplankton to nutrient depletion. Mar. Biol., 83, 293-300. Bienfang, P.K., 1-J.G. Harrison, and L.M. Quarmby, 1982. Sinking rate response to depletion of nitrate, phosphate and silicate in four marine diatoms. Mar. Biol., 67, 295-302.

PAGE 145

130 Bienfang, P.K. and M. Takahashi, 1983. rates in a subtropical ecosystem. 218. Ultraplankton growth Mar. Biol., 76, 213-Biggs, D.C., 1392. Nutrients, plankton and productivity in a warm-core in the western Gulf of Mexico. J. Geophys. Res., 97, 2143-2154. Biggs, D.C., 1988. Hydrographic data from the Texas continental shel: and northwest continental slope of the Gulf of t-1e:-:ico: TA!1U Ecosystem Research Group "Rings" Cruise 88G-05. Tech. Rep. 88-05-T, Dept. of Oceanography, Texas A&M Univ., College Station, TX. Billet, D.S., A.L. Rice and R.F. Mantoura, 1983. seaimentation cf phytoplankton to the deep-sea Nature, 302, 520-522. Bishop, S.S., J.h. Yoder and G.-A. Paffenhofer, 1980. PhytoplanKton and nutrient variability along a crossshelf transect cff Savannah, Georgia, U.S.A. Estuarine Coastal :1ar. 2ci., 11, 359-368. Blackman, F.F., 1905. Optima and limiting factors. Ann. Botany London, 19, 281-295. Blumberg, A.F. and G.L. Mellor, 1985. circulation in the Gulf of Mexico. Sci., 34, :22-144. A simulation of the Israel J. Earth Brand, L.E. and R.R.L. Guillard, 1981. The effects of continuous light and light intensity on the reproduction rates cf twenty-two species of marine phytoplankton. :. Exp. Mar. Biol. Ecol., 50, 119-132. B roe c k e r S : 9 9 l. 4, 79-89. The great ocean conveyor. Oceanogr., Broecker, W.S. and T.-H. Peng, 1982. Tracers in the Sea. Eldigio Press, Columbia University, New York, 690 pp. Broecker, W.S. and T.-H. Peng, 1992. Interhemispheric transport of carbon dioxide by ocean circulation. Nature, 356, 587-589. Brooks, D.A. and R.V. Legeckis, 1982. A ship and satellite view of hydrographic features in the western Gulf of Mexico. J. Geophys. Res., 87, 4195-4206. carder, K.L. and R.G. Steward, 1985. A remote-sensing reflectance model of a red-tide dinoflagellate off west Florida. Limnol. Oceanogr., 30, 286-298.

PAGE 146

131 Carder, K.L., R.G. Steward, J.H. Paul and G.A. Vargo, 1986. Relcttionships between chlorophyll and ocean color constituents cts the affect remote-sensing reflectance models. :...imr.ol. Sceanogr., 31, 403-413. Campbell, and E.J. Carpenter, 1986. Estimating the grazing pressure o f heterotrophic nanoplankton on Synechococcus spp. using the sea water dilution and selective techniques. Mar. Ecol. Prog. Ser., 33, 121-129. Carpenter, S.J. and R.R.L. Guillard, 1971. Interspecific differences half-saturation constants for three species of marine phytoplankton. Ecology 52, 183-185. Carpenter, S.J. J.S. Lively, Review of estimates of algal Jsing l 4c tracer techniques. Primary Product i ; i t y i n 5 e a P G :a l k ow ski ( e d ) P 1 en urn, New YorK, :61-178. Chan, A.'l'., :9'72. physiological study of marine diatoms and in relation t o irradiance and cell size. I. Growth under continuous light. J. Phycol. 14, 396-402. Chavez, F., 1989. Size distribution of phytoplankton in the central and eastern tropical Pacific. Global Biogeochem. C y c les, 3 27-35. Chisholm, S.W., 1992. Phytoplankton size. Primary Productivity and Biogeochemical Cycles in the Sea, P.G. Falkowski and A.D. Woodhead (eds.), Plenum Press, New York, pp. 213237. Clemons, M .J. C.3. 1984. Blooms of large diatoms in the oceanic, subarctic Pacific. Deep-Sea Res., 31, 85-95. Cochrane, J.D. and f.J Kelly, 1986. Low-frequency circulation on the T exas-Louisiana continental shelf. J. Geophys. Res., 91, 1 0645-10659. Callos, Y., 1986 Time-log algal growth dynamics: biological constraints on primary production in aquatic environments. Mar. Scol. Prog. Ser. 33, 193-206. C d G T :976. Mean circulation in shallow seas. J. sana y, Geophys Res., 81, 5389-5399.

PAGE 147

132 Cullen, J .J., 1982. The deep chlorophyll maximum: comparing vertical profiles of chlorophyll a. Can. J. Fish. Aquatic Sci., 39, 791-803. Deevey, G B., :971. The annual cycle in quantity and composition of the zooplankton of the Sargasso Sea off Bermuda. :. upper 500 m. Limnol. Oceanogr., 16, 219-240. Dinnel, S.P. a n d W.J. Wiseman, 1986. Fresh water on the Louisiana and Texas shelf. Cont. Shelf Res., 6, 765-784. Dortch, Q., J.R. Clayton, S.S. Thoresen and S.I. Ahmed, 1984. Spec:es differences in accumulation of nitrogen pools:..:-. ;-:-,:_;,ktcn. 3iol., 81, 237-250. Ednoff, H.E., Surface phytoplankton communities and their rel a to the Loop Current in the Gulf of Mexico. unpubl ished Hasters Thesis, The Florida State University, Florida, 187 pp. Elliot, B.A., Mexico. 1982. Anticyclonic rings in the Gulf of J Phys. Oceanogr., 12, 1292-1309. El-Sayed, S.Z. and J.T. Turner, 1977. Productivity of the Antarctic and tropical/subtropical regions: a comparative study. P o lar Oceans, M.J. Dunbar (ed.), Arctic Institute o f America, Calgary, Canada, pp. 463503. Emerson, S., 1987. new produc:ion Pacific O cean. 6544. Seasonal oxygen cycles and biological in surface waters of the subarctic Global B i ogeochem. Cycles, 92, 6535-Eppley, R.W., 1972. Temperature and phytoplankton in the sea. Fish. Bull., 70, 1063-1085. Eppley, R.W and B.J. Peterson, 1979. Particulate organic matter flu x and planktoni c new production in the deep ocean. Nature, 282, 677-680. Eppley, R.W., E.H. Renger and W.G. Harrison, 1979. Nitrate and phytoplankto n production in southern California coastal waters. Limnol Oc e anogr., 24, 789-796. Eppley, R.W., J.N. R o gers, J.J. 1969. ration con s t a nts for uptake of nltrate a n d ammonlum by marine phy t oplankton L imnol. Oceanogr., 14, 9129 20.

PAGE 148

Falkowski, P.G., Z. Dubinsky and K. Wyman, 1985. irradiance relationships in phytoplankton. Oceanogr., 30, 311-321. GrowthLimnol. 133 Falkowski, P.G., Ziemann, z. Kolber, and P.K. Bienfang, 1991. Role vf eddy pumping in enhancing primary production ocean. tlature, 352, 55-58. Fanning, K.A., :989. of atmospheric pollution on nutrient limitation in the ocean. Nature, 339, 460463. Fenchel, T., 1982a. E:cology of heterotrophic microflagellates. II. Bioenergetics and growth. Mar. E co l P r c g r 8 2 2 5-2 3 1 Fenchel, T., :;_982b. of heterotrophic microflagel:ates. IV. Quantitative occurrence and importance as of bacteria. Mar. Ecol. Prog. Ser .. 9, Ferguson, R.L. and Sunda, 1984. Utilization of amino acids by planktonic marine bacteria: Importance of clean technique and low substrate addition. Lirnnol. Oceanogr., 29, 258-274. Fitzwater, S.E., G.A. Knauer and J.H. Martin, 1982. Metal contamination and its effect on primary production measurements. Limnol. Oceanogr., 27, 544-551. French, D.P., Furnas and T.J. Smayda, 1983. Diel changes in nitrite concentration in the chlorophyll maximum in Gulf of Mexico. Deep-Sea Res., 30, 707722. Fuhrman, J.A. and G.B. McManus, 1984. Do bacteria-sized marine eukaryotes consume significant bacterial production? Science, 224, 1257-1259. Furnas, M. and T. Smayda, 1987. Inputs of subthermocline waters and nitrate onto the Campeche Bank. Cont. Shelf Res 7 1 6 1 1 7 6 Gargett, A.E., interior. 1984. Vertical eddy diffusivity in the ocean u. Mar. Res., 42, 359-393. Geider, R.J., B.A. Osborne and J.A. Raven, 1986. Growth, photosynthesis and maintenance metabolic cost in the diatom, Phaeodactylum tricornutum at very low light levels. J. Phycol., 22, 39-48.

PAGE 149

134 Glover, H.E., B.B. Prezelin, L. Campbell, M. Wyman and C. Garside, 1988. A nitrate-dependent Synechococcus bloom in surface Sargasso Sea water. Nature, 331, 161-163. Goldman, J. C., 1988. Spatial and temporal discontinuities of biological processes in pelagic surface waters. Toward a Theory on Biological-Physical Interactions in the World Ocean, B.J. Rothschild (ed.), Kluwer Academic Pub., 3oston, pp. 273-296. Goldman, J.C. and P.M. Glibert, 1983. Kinetics of inorganic nitrogen uptake by phytoplankton. Nitrogen in the Marine Environment, E.J. Carpenter and D.G. Capone (eds.), Academic Press, New York, pp. 223-274. Gooday, A., l968. A response by benthic :oraminifera to the depositio n c : in the deep sea. Nature, 332, 70-73. Gordon, H.R. and J.. Y. Morel, 1983. Remote assessment of Ocean Color Interpretaticn of Satellite Visible :magery. ::.14 pp. Gowing, M.M. and M.W. Silver, 1985. Minipellets: a new and abundant size class of marine fecal pellets. J. Mar. Res., 43, 395-418. Haddad, K.D. and K.L. Carder, 1979. Oceanic intrusion: one possible initialization mechanism of red tide blooms on the west coast of Florida. Toxic Dinoflagellate Blooms, Taylor and Seliger (eds.), Elsevier, pp. 269-274. Harrison, W.G., 1983. Nitrogen in the marine environment: use of isotopes. Nitrogen in the Marine Environment, E.J. Carpenter and D.G. Capone (eds.), Academic Press, New York, pp. 223-274. Haury, L., J.A. McGowan and P.H. Wiebe, 1977. Pattern and processes in the time-space scales of plankton distributions. Spatial Pattern in Plankton Communities, J.H. Steele (ed.), Plenum, pp. 277-327. Hedgpeth, J.W., 1977. Models and Muddles. Helgo. wiss. Meeres., 30, 92-104. Heinbokel, J.F., 1978. Studies on the functional role of tintinnids in the Southern California Bight. I. Grazing and growth rates in laboratory culture. Mar. Biol., 47, 177-189.

PAGE 150

135 Herbland, A., A. Le Bouteiller, P.L. Raimbault, 1985. Size structure of phytoplankton in the equatorial Atlantic Ocean. Deep-Sea Res., 32, 819-836 Hitchcock, G.L., C. and T.J. Smayda, 1987. Short term changes in the biology of a Gulf Stream warm-core ring: ?hytoplankton biomass and productivity. Limnol. Oceanogr., :J2, 919-928. Hobson, L.A. Lorenzen, 1972. Relationships of chloropnyll maxima to density structure in the Atlantic Ocean ana Gul: of Mexico. Deep-Sea Res., 19, 297-306. Hochman, H.T., 1992. Interpretation of the Coastal Zone Color Scanner (CZCS) Signature of the Orinoco River. M.S. thes:s, of South Florida, Tampa, 82 pp. Hofmann, l96e. ?hytoplankton dynamics on the outer southeastern :ontinental shelf. Part III: A coupled model. J. Mar. Res., 46, 919-946. Hofmann, E.E. Ambler, 1988. Plankton dynamics on the outer southeastern U.S. continental shelf. Part II: A time-dependent biological model. J. Mar. Res., 46, 883-9l7. Hopkins, T. L. l 9 8 2. The vertical distribution of zooplankton in the eastern Gulf of Mexico. Deep-Sea Res., 29, 1069-1083. Howard, K.M. and 1989. picoplankt:n :n North Sea. 281. Physiological ecology of Mar. Biol., 102, 275-Huh, O.K., and L.J. Rouse, Jr., 1981. Intrusion cf Loop Current waters onto the West Florida continental shelf. J. Geophys. Res., 86, 4186-4192. Hulburt, E.M., 1990. Description of phytoplankton and nutrient in spring in the western North Atlantic Ocean. J. Plank. Res., 12, 1-28. Hurlburt, H.E. and J.D. Thompson, 1980. A numerical study of the Loop Current intrusions and eddy shedding. J. Phys. Oceanogr., 10, 1611-1651. Iturriaga, R. and J. Marra, 1988. Temporal and spatial variability of chroococcoid cyanobacteria Synechococcus spp. specific growth rates and their contribution to primary production in the Sargasso Sea. Mar. Ecol. Proq. Ser., 44, 175-181.

PAGE 151

136 Iturriaga, R. and B.G. Mitchell, 1986. Chroococcoid cyanobacteria: a significant component in the food web dynamics of the open ocean. Mar. Ecol. Prog. Ser., 28, 291-297. Johnson, T.O and W.O Smith, Jr., 1986. Sinking rates of phytoplankton assemblages in the Weddell Sea marginal i c e z one r-1 a r E co 1 P r o g S e r 3 3 1 3 1 -1 3 7 Jenkins, W.J. and J C. Goldman, 1985. cycling and primary production in Mar. Res., 43, Seasonal oxygen the Sargasso Sea. J. Joint, I. R. and Jl. J. Pomeroy, 198 6. Photosynthetic characteristics c: and picoplankton from the sur f ace m i :-: e d .:.. a 1 e r l-1 a r B i o 1 9 2 4 6 5-4 7 4 Kana, T.M. and P.:-1 Glibert, 1987. Effect of irradiances up to 2000 m -2 s-: o n marine Synechococcus WH7803. I. Growth, and cell composition. Deep-Sea Res 3 4 4 7 9-4 9 5 Kirk, J.T.O., 1983. =..i.ght and Photosynthesis in aquatic environments. Cambridge Univ. Press, 399 pp. Kirwan, A.D., Jr., W.J. Merrell, Jr., K.K. Lewis, R .E. Whitaker and R 1984. A model for the analysis of drifter data with an application to a warm core ring in the Gulf c f Mexico. J. Geophys. Res., 89, 4181-4185. Kullenberg, G.E., 1976. On vertical mixing and the energy transfer :he wind to the water. Tellus, 28, 159-165. Lampitt, R.S., 1985. Evidence for seasonal deposition of detritus to deep-sea floor and its subsequent resuspens ion. Jeep-Sea Res. 32, 8 8 5-8 9 8. Landry, M.R., 1983. The development of marine calanoid copepods with comment on the isochronal rule. Limnol. Oceanogr., 28, 614-624. Landry, M.R., L.W. Haas and V.L. Fagerness, 1984. Dynamics of microbial plankton communities: experiments in Kaneohe Bay Hawaii. Mar. Ecol. Prog. Ser., 16, 127-133. Langdon, c., 1987. On the causes of interspecific differences in the growth-irradiance relationship for phytoplankton. I. A comparative study of the growthirradiance relationship of three marine phytoplankton

PAGE 152

137 species: Skeletonema costatum, Olisthodiscus luteus and Gonyaulax tamarensis. J. Plankt. Res., 9, 459-482. Lee, T.N. and L.P. htkinson, 1983. Low frequency current and temperature variability from Gulf Stream frontal eddies and atmospheric forcing along the southeastern U.S. outer continental shelf. J. Geophys. Res., 88, 4541-4567. Lee, T.N., J.A. Yoder and L.P. Atkinson, 1991. Gulf Stream frontal eddy influence on productivity of the southeast U.S. continental shelf. J. Geophys. Res., 96, 2219122205. Legendre, L. and J. Fevre, 1989. Hydrodynamical singularities as controls cf recycled versus export production in oceans. Productivity of the Ocean: Present and Past, Berger, V.S. Smetacek and G. (eds.), New York, pp. 49-63. Leipper, D.F., 1970. Gulf of A sequence of current patterns in the J. Geophys. Res., 75, 637-657. Levitus, S., 1982. Climatological atlas of the world ocean. NOAA Professional Paper No. 13, 1-173. Lewis, J.K., 1992. Geophys. Res., physics of the Gulf of Mexico. 971 2141-2142 J. Lewis, J.K. and A.D. Kirwan, Jr., 1985. Some observations of ring topography and ring-ring interactions in the Gulf of Mexico. J. Geophys. Res., 90, 9017-9028. Lewis, M.R., W.G. Harrison, N.S. Oakey, D. Hebert and T. Platt, 1986. Vertical nitrate fluxes in the oligotrophic ocean. Science, 234, 870-873. Li, W.K.W., D.V. Subba Rao, W.G. Harrison, J.C. Smith, 0.J. Cullen, B. Irwin and T. Platt, 1983. Autotrophic picoplankton in the tropical ocean. Science, 219, 292-295. Lochte, K. and C.M. Turley, 1988. Bacteria and cyanobacteria associated with phytodetritus in the deep sea. Nature, 333, 67-69. Longhurst, A.R., 1991. Role of the marine biosphere in the global carbon cycle. Limnol. Oceanogr., 36, 1507-1526 Longhurst, A.R. and W.G. Harrison, 1988. Vertical nitrogen flux from the oceanic photic zone by diel migrant zooplankton and nekton. Deep-Sea Res., 35, 881-889.

PAGE 153

138 Longhurst, A.R. and W.G. Harrison, 1989. The biological pump: of plankton production and consumption in the upper ocean. Prog. Oceanogr., 22, 47-123. Malone, T.C., 1971. The relative importance of nannoplankton and netplankton as primary producer in tropical ccee:nic and neritic phytoplankton communities. Limnol. Cceanogr., 16, 633-639. Margalef, R., l978. Life-forms of phytoplankton as survival alternatives in an unstable environment. Oceanol. Acta, ... Marra, J., T. Dickey, W.S. Chamberlin, C. Ho, T. Granata, D.A. Kiefer, C. Langdon, R. Smith, K. Baker, R. Bidigare and Hamilton, 1992. Estimation of primary producti0n fr8m moored optical sensors in the Sargasso Sea. J. Geopf:ys. Res., 97, 7399-7412. Maul, G.A., 1977. The annual cycle of the Gulf Loop. 1. during a one-year time series. r-1a r. Ee s. 5, 21 9-2 4 7. Part J. Michaels, A.F. and M.W. Silver, 1988. Primary production, sinking fluxes and the microbial food web. Deep-Sea Res 3 5, 4 7 3-4 9 0 Morel, A. and R.C. Smith, 1974. Relation between total quanta and tctal energy for aquatic photosynthesis. Limnol. Oceancgr., 19, 591-600. Morris, M.J. and Hopkins, 1983. Biochemical composi tion of crGstacean zooplankton from the eastern Gulf of Mexico .;. Exp. Mar. Biol. Ecol., 69, 1-19. Muller-Karser, F.E., C.R. McClain and P.L. Richardson, 1988. The dispersal of the Amazon's water. Nature, 335, 56-59. Muller-Karger, F.E., J.J. Walsh, R.H. Evans and M.B. Meyers, 1991. Jn the seasonal phytoplankton concentration and sea surface temperature cycles of the Gulf of Mexico as determined by satellites. J. Geophys. Res., 96, 12,645-12,665. Mullin, M.M., E.F. Stewart and F.J. Fuglister, 1975. Ingestion by planktonic grazers as a function of concentration of food. Limnol. Oceanogr., 20, 259-262. Murphy, E.B., K.A. Steidinger, B.S. Roberts, J. Williams and J.W. Jolley, 1975. An explanation for the Florida east

PAGE 154

coast Gymnodinium breve red tide of November 1972. Limnol. Jceanogr., 20, 481-486. Murphy, L.S. and E.M. Haugen, 1985. The distribution and abundance of phototrophic ultraplankton in the North Atlantic. Limnol. Oceanogr., 30, 47-58. 139 Nowlin, W.O., Jr., :972. Winter circulation patterns and property distributions. Contributions on the Physical Oceanography of the Gulf of Mexico, L.R.A. Capuro and J.L. Reid (eds.J, Gulf Publ. Co., Houston, pp. 3-51. O'Brien, J.J. (eo.), ::.986. Advanced Physical Oceanographic Numerical Modelling, D. Reidel Publ. Co., Boston, 608 pp. Ortner, P.B., ......... :erguson, S.R. Piotrowicz, L. Chesal, G. Berberian e:nc ;. .. ;. Palumbo, 1984. Biological consequences o f and atmospheric advection with the Gulf Loop Ir.trusion. Deep-Sea Res., 31, 1101-1120. Pace, M.L., G.A. D.M. Karl and J.H. Martin, 1987. Primary production, new production and vertical flux in the easterr. Pacifi.c Ocean. Nature, 325, 803-804. Paerl, H.W., 1985. Enhancement of marine primary production by nitrogen-enriched acid rain. Nature, 315, 747-749. Paluszkiewicz, T., L.P. Atkinson, E.S. Posmentier and C.R. McClain, 1983. Observations of a Loop Current frontal eddy intrusicn cr.:o the West Florida Shelf. J. Geophys. Res., 88, 9639-9651. Pilskaln, C.H. and S. Honjo, 1987. The fecal pellet fraction of biogeochemical particle flux to the deep sea. G.i.::bal !3icgeochem. Cycles, l, 31-48. Pitcher, G.C., D.R. and B.A. Mitchell-Innes, 1989. Phytoplankton sinking rate dynamics in the southern Benguela upwelling system. Mar. Ecol. Prog. Ser., 55, 261-269. Platt, T., K. Denman, and A. Jassby, 1977. Modeling the productivity of phytoplankton. in The Sea, Vol 6., E.D. Goldberg, J.J. O'Brien and J.H. Steele (eds.), Wiley Interscience, 807-856. Platt, T. and W.G. Harrison, 1985. Biogenic fluxes of carbon and nitrogen in the ocean. Nature 318, 55-58.

PAGE 155

Platt, T. and ethers, 1990. Biological production of the oceans: the case for a consensus. Mar. Ecol. Prog. Ser., 52, 77-88. 140 Platt, T., D.V. Subba Rao, and B. Irwin, 1983. Photosynthesis of piccplankton in the oligotrophic ocean. Nature, 301, 702-704. Pomeroy, L., 1974. ':'he digm. ocean's food web, a changing para-24, 499-504. Redfield, A.C., The biological control of the chemical factors in environment. Amer. Sci., 46, 205221. Reed, R.K., 1977. esti:nating insolation over the ocean. J. Phys. Ocea!logr., 7, 482-485. Reid, 1972. s:mple dynamic model of the Loop Cur rent. Contritutions on the Physical Oceanography of the Gulf of Mexico, Capuro and J.L. Reid (eds.), Gulf Publ. Cc., Houston, pp. 157-159. Rice, A.L., D.S.M. Billet, J. Fry, A.W.G. John, R.S. Lampitt, R.F.C. Mantoura and R.J. Morris, 1986. Sea sonal depcs:t:cn cf phytodetritus to the deep-sea floor. ?roc. Royal Soc. Edin., 88B, 265-279. Riebesell, U., 1991. Particle aggregation during a diatom bloom. I. Physical aspects. Mar. Ecol. Prog. Ser., 69, 272-280. Roache, P.J., 1976. Fluid Dynamics, Hermosa, Albuquerque, 446 pp. Ryther, J.H., the sea. 1969. Photosynthesis and fish production in 166, 72-76. Ryther, J.H. and W.M. Dunstan, 1971. Nitrogen, phosphorus and eutrophication in the marine environment. Science, 171, 1008-1013. Sancetta, C., T. Villareal and P. Falkowski, 1991. Massive fluxes of rhizosolenid diatoms: A common occurrence? Limnol. Oceanogr., 36, 1452-1457. Sathyendranath, S. and T. Platt, 1988. The irradiance field at the surface and in the ocean: A model for applications in remote sensing. J. Geophys. Res., 93, spectral the interior oceanography 9270-9280. of and

PAGE 156

141 Sathyendranath, ana T. Platt, 1989. Computation of aquatic primary Extended formalism to include effect o f angular and spectral distribution of light. Limnol. 0ceanogr., 34, 188-198. Schroeder, W.M., S.?. Dinnel, W.J. Wiseman and W.J. Merrell, 1987. Circulation patterns inferred from the movement of detached in the eastern Gulf of Mexico. Cont. Shelf Res., 7 883-894. Schroeder, W.M., W.J. Wiseman, S.P. Dinnel and R.C. Heron, 1988. of shelf-break exchange from CZCS imagery. Trans. Amer. Geophys. Union, 69, 1283. Sharp, J. H., 96 3 ::;:..stribution of inorganic and organic : n sea. Nitrogen in the Marine Environment, E.J. Carpenter and D.G. Capone, (eds.), Academic Press, Y o r k pp. 1-35. Shulenberger, E. ctn d J Reid, 1981. The Pacific shallow oxygen maximum, deep chlorophyll maximum, and primary productio n reccr .s.idered. Deep-Sea Res., 28, 901-919. Smayda, T., 1970. The suspension and sinking of phytoplankton :n the sea. Oceanogr. Mar. Biol. Ann. Rev., 8, Smith, D C., : v and .:.J. O'Brien, ::.983. The interaction of a two-layer isolated mesoscale eddy with topography in the Western Gulf o f Mexico. J Phys. Oceanogr., 13, 1681-1697. Smith, S.L. and P.V.Z. Lane, 1987. On the life history of Centropages typicus: responses to a fall diatom bloom in the New York Bight. Mar. Biol., 95, 305-313. Smolarkiewicz, P,K., 1983. A simple, positive definite advection scheme with small implicit diffusion. Monthl y Weather Review 111, 479-486. Steele, J.H., 1974. The Structure of Marine Ecosystems. Harvard Univ. Press, Cambridge, MA, 128 pp. Steele, J.H. and E.W. Henderson, 1992. The role of predation in plankton models. J. Plankt. Res., 14, 157-172. stockner, J.G., 1988. Phototrophic picoplankton: An overview from marine and freshwater ecosystems. Limnol. Oceanogr., 44, 765-775.

PAGE 157

142 Sturges, W.A. and J.C. Evans, 1983. On the variability of the Lo o p in the Gulf of Mexico. J. Mar. Res., 41, 639-653. Taguchi, S., G.R. and E.A. Laws, Physiological characteristics and production of mixed layer and chlorophyll maximum phytoplankton populations in the Caribbean Sea and western Atlantic Ocean. Deep-Sea Res., Takahashi, T., 1989. Oceanus, 32, 22-29. Taylor, A.H., J.R.W. Harris and J. Aiken, 1986. The interaction of physical and biological processes in a model of the of phytoplankton under .'arine Interfaces Ecohydrodynamics, J.C.J. !Jihsul (ed.), Elsevier Oceanography Series, 42, 313-330. Taylor, A.H. and Joint, 1990, A steady state analysis of the 'microbial loop' in stratified systems. Mar. Ecol. Prog. Ser., ::.-17. Taylor, A.H., A.J. Watson, M. Ainsworth, J.E. Robertson and D. R. Turner, 1991. A modelling investigation of the role of phytsplankton in the balance of carbon at the surface o f North Atlantic. Global Biogeochem. Cycles, 5, Toggweiler, J.R., 1989. Is the downward dissolved organic matter (CCM) flux carbon transport? Productivity of the Ocean: Present and Past, W.H. Berger, V.S. Smetacek and G. Wefer (eds.), Wiley, New York, pp. 65-83. Trees, C.C., :985. Femote Sensing of Ocean Color in the Northern Gulf of Mexico. Doctoral dissertation, Texas A&M University, College Station, 258 pp. Urrere, M.A. and G.A. Knauer, 1981. Zooplankton fecal pellet fluxes and vertical transport of particulate organic material in the pelagic environment. J. Plankt. Res., 3, 369-387. Vargo, G.A., K.L. Carder, W. Gregg, E. Shanley and C. Heil, 1987. The potential contribution of primary production by red tides to the west Florida shelf ecosystem. Limnol. Oceanogr., 32, 762-767. Vidal, J., 1980. Physioecology of zooplankton. II. Effects of phytoplankton concentration, temperature, and body size on the development and molting rates of Calanus

PAGE 158

143 pacificus and Fseudocalanus sp. Mar. Biol., 56, 135-146. Vidal, V.M.V., F.V. Vidal and J.M. Perez-Molero, 1992. Collisio n c f a L o c p Current anticyclonic ring against the c ontinental s helf slope of the western Gulf of Mexico. Ge cphys. Res., 97, 2155-2172. Villareal, T.A., :988. r-ositive buoyancy in the oceanic diatom Rhizosolenia debyana H. Peragallo. Deep-Sea Res 3 5, l CJ 3 7 l 0 4 5 Vukovich, l986. Aspects of the behavior of cold perturbatio n i n : h e eastern Gulf of Mexico: A case study. C ceanogr., 16, 175-188. "Jukovich, =-.cop Current toundary variations. Geophys. !<. es., :::, 15,585-15,591. J. Vukovich, M. Eushnell and W.J. King, 1979. Some aspects o f :he oceanography of the Gulf of Mexico using satellite and in situ data. J. Geophys. Res a 4, 7 7 4 9 7 7 6 8 Vukovich, F.M. and G.A. Maul, eastern Gulf o f 117. 1985. Cyclonic eddies in the J. Phys. Oceanogr., 15, 105Waddell, E. (ed.), 1986. Gulf of Mexico Physical Oceanogra phy Program, Fin a l Years 1 and 2, vol. II: Technical technical report prepared for Minerals Management Service, U.S. Department of Interior, contract n o :4-12-0001-29158, Science Applications International Corp., Raleigh, N.C., 378 pp. Walsh, J.J., ::.983. phytoplankton 2eath i n the sea: enigmatic losses. Prog. Oceanogr., 12, 1-86. Walsh, J.J., D.A. Dieterle, M.B. Meyers and F.E. MullerKarger, 1989. Nitrogen exchange at the continental margin: a numerical simulation analysis. Prog. Ocean ogr., 23, ::::45-301. Walsh, J.J., D.A. Dieterle, J.R. Pribble, 1991. Organic debris on the continental margins: a simulation analysis of source and fate. Deep-Sea Res., 38, 805-828. Waterbury, J.B., S.W. Watson, R.R.L. Guillard and L.E. Brand, 1979. Widespread occurrence of a unicellular, marine, planktonic, cyanobacterium. Nature, 277, 293-294.

PAGE 159

144 Watson, A.J., C. Robinson, J.E. Robinson, P.J.leB. Williams and M.J.R. Fasham, 1991. Spatial variability in the sink for dtmospheric carbon dioxide in the North Atlantic. Nature, 350, 50-53. Wessels, C and E. 1979. An improved apparatus for use w1th the acid bubbling method of measuring primary prcducticn. Limnol. Oceanogr., 24, 187-188. Wroblewski, J.S., 1977. A model of phytoplankton plume formation during variable Oregon upwelling. J. Mar. Res., 35, 357-394. Yoder, J.A., L.P. S.S. Bishop, E.E. Hofmann, and T.N. Lee, 1983 .:::feet of upwelling on phytoplankton productivity s f outer southeastern United States continental shelf. Cont. Shelf Res., 1, 385-404. Yoder, J.A., L .?. T.!J. Lee, H.H. Kim, and C.R. McClain, 1981. ? .ole of Gulf Stream frontal eddies in form1ng patches on the outer southeastern shelf. Limnol. 'Jcear.ogr., 26, 1103-1110. Yoder, J.A. and A. Mahood, 1983. Primary production in Loop Current upwelling. Southwest Florida Shelf Ecosystems Study Year 2 Modification, Woodward-Clyde Cor.sultants and Skidaway of Oceanography, A final report submitted to the Minerals Management Service, 0 r 1 e an s LA con t r act no 1 4 -1 20 0 0 1 -2 9 1 4 4