NO A A PROFESSIONAL PAPER 11 



cumstances surrounding our 1976 data suggest that the 

 rate calculated by Segar and Berberian is an overestimate, 

 probably because much of the primary productivity con- 

 tributing dominantly to their calculation is in fact advected 

 out of the area rather than simply sinking through the 

 thermocline. 



Another time scale of interest is an analog for oxygen 

 of the advective flushing time for water. That is the ratio 

 of the oxygen storage in a segment to the advective input 

 of oxygen, CV/1,(CQ),„, which we will call the ventilation 

 time. This ratio is a measure of the flow-thru rate of oxy- 

 gen. Short ventilation times indicate that the availability 

 of oxygen by advective flow-thru is large relative to am- 

 bient oxygen storage. Ventilation times defined in this 

 way exhibit the same general features as the water flushing 

 time, but typically are 15 to 60 percent shorter, except in 

 the outer shelf segment and the shelf valley where they 

 are nearly equal. The implication of these results is that 

 advection is more effective in supplying oxygen to the 

 Bight than it is for simple renewal of water. These results 

 are due, in part, to the reduced oxygen concentration 

 values that occur regularly in the inner Bight during sum- 

 mer relative to higher oxygen concentrations existing else- 

 where available for transport to that region. 



An analysis of the errors in the calculation must be 

 made to determine the significance of the results. Standard 

 error of the means can be calculated for the D.O. con- 

 centrations, but the errors in volume, V, and transport, 

 Q, cannot be calculated from the available results. If the 

 error in volume and transport is assumed to be 10 percent, 

 it is the dominant error in calculation. The values of the 

 net utilization rates are greater than one standard error 

 of the rates for all the segments except L.3. The values 

 of IS for segments A and Jl are 4.0 and 4.8 times the 

 standard errors in S5, respectively. If the assumed errors 

 in Q are increased to 20 percent of their magnitude, then 

 the standard errors in I.S are still less than SS for the 

 segments of critical interest — A, Jl, J2, and L2. It would 

 be necessary to assume errors of 50 percent in Q to make 

 the errors in 25 greater than the magnitude of 25 in seg- 

 ment Jl. Thus, the important results of the calculations 

 can be taken as meaningful even if the errors in approx- 

 imating the transports are large. 



A drawback of diagnostic modeling is that it cannot 

 forecast how details of the circulation would change under 

 different wind or other forcing conditions. The model can 

 be used to describe and analyze only conditions for which 

 at least some current velocity data are available. 



The density field is held stationary over the entire 42- 

 day interval because only one complete set of density ob- 

 servations was available. This introduces errors in the 

 transport calculation in equations (8) and (9), but the near- 

 bottom velocity field is unchanged. Comparisons of ob- 

 served and modeled velocities and the effect of the density 



field on model accuracy will be the subject of a further 

 study. The results of our model are probably more ac- 

 curate than a prognostic model, since the calculation is 

 based upon a large amount of observed data. 



Another shortcoming of the method as presently used 

 is that it is applicable to only the steady or most slowly 

 changing components of the flow. Therefore, diffusion of 

 oxygen by higher-frequency movements such as tidal cur- 

 rents and storm-driven transient flows is not included in 

 the diagnosis. This effect arises from simultaneous vari- 

 ation of flow speed and oxygen concentration, the basic 

 mechanism of turbulent transport, within the four time 

 periods that were diagnosed and averaged. In defense of 

 the present application, incorporation of a gradient dif- 

 fusion mechanism of any sort into the model would in- 

 dicate additional oxygen flux into the oxygen-deficient 

 regions, thus strengthening the major conclusion of the 

 investigation. In fact, the respiration rate calculated by 

 Malone et al., using an independent approach (ch. 9, pt. 

 1), indicates that the sum of observed D.O. change and 

 advective import are of approximately the correct mag- 

 nitude to supply the oxygen utilization; hence, diffusive 

 flux of oxygen seems to be relatively unimportant. 



The circulation pattern, both in the Bight Apex over 31 

 days of the study and all along the inner New Jersey shelf 

 over 21 days of the study, shows convergent flow in the 

 lower layer and upwelling through the pycnocline. Though 

 the upper layer model results are not shown, the upper 

 layer flow was offshore as is seen in the current meter 

 data (fig. 8-5a, b, and d). The cause of the convergence 

 in the Apex is the unusual flow pattern shown in figures 

 8-5a and d, where the nearshore flows in the lower layer 

 off New Jersey and Long Island are both directed toward 

 the Apex. Away from the Apex, the bottom Ekman layer 

 transport, which is directed to the left of the alongshore 

 flow, creates a strong convergence off New Jersey. This 

 pattern of shoreward flow and convergence in the bottom 

 waters, compensated by divergent seaward flow in the 

 upper waters is kinematically similar to the circulation 

 commonly observed in coastal plain estuaries. Festa and 

 Hansen (1978) showed that particulate materials charac- 

 terized by a suitable particle sinking velocity will tend to 

 be concentrated by a flow field of this type, irrespective 

 of whether they are introduced from the river or from the 

 ocean. It is one of the causative mechanisms for the tur- 

 bidity maximum that has long been known to occur in 

 coastal plain estuaries. We expect that under the influence 

 of the circulation observed in the New York Bight in the 

 late spring of 1976, oxidizable particulate materials orig- 

 inating in the Hudson Estuary, or elsewhere throughout 

 the Bight, will have been concentrated in the Apex and 

 along the inner shelf off New Jersey. 



Ceratium tripos are ideally suited to couple to this con- 

 vergent pattern, as suggested in chapter 9, part 1, since 



190 



