In deriving the coefficients of dissipation, I 

 tried to balance the fit obtained by looking- at 

 strictly the inertial transports versus the predic- 

 tion and the fit obtained when recombining the 

 prediction and the net sea current. For this rea- 

 son, the plot of predictions is slightly to the right 

 of the inertial period data, while the prediction 

 recombined with the net sea current is slightly to 

 the left of the data. The runs illustrated here 

 provided the best fit overall. The coefficients 

 required for this fit are: 



K, = 1 X 10 '; second order dissipation term 

 K., = 5 X 10 °; first order dissipation term 



Runs from data Windows II and III did not pro- 

 vide as useable testing values as Window I. Due to 

 the variable nature of the winds, there was little 

 overall transport, as shown in Figure 4.9. Figure 

 4.9 begins after 4 days have passed. Data Window 

 II was used for this run. Both plots show the 

 inertial period component of the Lagrangian 

 particle trajectory. 



The initial bathythermal profiles for all the 

 runs illustrated above were taken at Ocean 

 Weather Station Hotel near the dates in question. 

 OWS Hotel was "upstream" of the mooring, so the 

 profiles experienced at Hotel would be advected 

 onto the mooring. Because of the lag involved 

 between the measurements at Hotel and the 

 mooring, I used BT traces preceding the time of 

 the start of the model by one week to initialize the 

 program. 



Figure 4.11 shows the total data set. All the 

 curves shown on the graphs are the data from the 

 mooring. I was unable to superimpose on this plot 

 the predictions of the model, unfortunately, due 

 to time constraints. It is easily seen that large 

 inertial oscillations occurred throughout the data 

 record, and that a sea current towards the nor- 

 theast existed throughout the record. 



One could try to postulate what the mean cur- 

 rent would be in the area by trying to set the 

 mean velocity to zero within a certain window, 

 but this would ignore currents that may be pro- 

 duced by steady winds, which are an important 

 part of the testing of this type of model. Complex 

 Fourier analysis would also be an answer, but it 

 still would not answer the problems I showed 

 earlier with Lagrangian vs. Eulerian conflicts 

 and the testing of the velocity shears that have 

 been shown to exist. 



EXPERIMENTAL DESIGN 



The problems I experienced with testing the 

 model have prompted me to formulate an exper- 

 imental design proposing a method to test an 

 integrated model of this sort. To eliminate prob- 

 lems with the isolation of inertial frequencies 

 from the constant current, the data should be 

 taken in an area with a very low overall sea cur- 

 rent, or no current at all. The area should have 

 steady winds at most times to get a consistent 

 transport for ease of testing. The mixed layer 

 that is predominant in the area should be less 

 than 100 meters deep, so that the velocities pro- 

 duced by the wind forcing are of a measurable 

 and distinct character. 



Simultaneous with the current measurements, 

 the bathythermal profiles should be measured in 

 both surface layers and below the thermocline, to 

 determine surface warming and the tempera- 

 ture gradient at the thermocline. A pyroheliome- 

 ter and some instrument measuring back radia- 

 tion should also be included. 



The most fundamental problem in the experi- 

 mental design is the problem of Lagrangian vs. 

 Eulerian measurement. Lagrangian is most 

 desireable, but problems occur with measure- 

 ment of currents at different depths. Accuracy in 

 measurement must be = 1 km, so that the inertial 

 oscillations can be accurately mapped. Drift 

 buoys with drogues lack this accuracy, if satellite 

 tracking were used. Satellite tracking also has 

 the disadvantage that data points are much too 

 far separated in time to be valid in inertial fre- 

 quencies. The buoys could be followed with a 

 ship, but the cost for such an operation would 

 probably be prohibitive. Sofar floats could not be 

 used due to the constantly changing density 

 structure of the mixed layer. Therefore, we come 

 back to an Eulerian measurement of a mooring. 

 Eulerian data could be used reasonably, as long 

 as one can assume horizontal homogeneity of the 

 boundary conditions . . . insolation, winds, and 

 bathythermal structure. The extent of the homo- 

 geneity must be greater than the total possible 

 inertial transport over the length of the data 

 record. If we assume an average inertial velocity 

 of 5 cm sec , the total transport over a 48 day data 

 record would be 207 km. 



I feel that the center of a gyre would provide all 

 the factors I have outlined as necessary. The 

 water conditions here are homogeneous over the 

 extent of the gyre center. The North Atlantic 



19 



