ANALYSIS AND INTERPRETATION OF TIDAL CURRENTS IN THE COASTAL BOUNDARY LAYER by PAUL WESLEY MAY B. S., Southern Missionary College (1972) SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF SCIENCE at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY and the WOODS HOLE OCEANOGRAPHIC INSTITUTION May, 1979 Signature of Author Joint Program in Oceanography, Massachus6ts Institute of Technology - Woods Hole Oceanographic Institution, and Department of Earth and Planetary Sciences, and Department of Meteorology, Massachusetts Institute of Technology, May, 1979. Certified by ......... . . .. Thesis Supervisor Accepted by .......................................................... Chairman, Joint Oceano r mittee in Earth Sciences, Massachusetts In 5 hd4 1 - Woods Hole Oceanographic Institution. LIBRARIES MITLibraries Document Services Room 14-0551 77 Massachusetts Avenue Cambridge, MA 02139 Ph: 617.253.5668 Fax: 617.253.1690 Email: docs@mit.edu http://libraries.mit.edu/docs DISCLAIMER OF QUALITY Due to the condition of the original material, there are unavoidable flaws in this reproduction. We have made every effort possible to provide you with the best copy available. If you are dissatisfied with this product and find it unusable, please contact Document Services as soon as possible. Thank you. Pages are missing from the original document. Page 189 is missing. ANALYSIS AND INTERPRETATION OF TIDAL CURRENTS IN THE COASTAL BOUNDARY LAYER by PAUL WESLEY MAY Submitted to the Massachusetts Institute of Technology - Woods Hole Oceanographic Institution Joint Program in Oceanography on May 1, 1979, in partial fulfillment of the requirements for the degree of Doctor of Science ABSTRACT Concern with the impact of human activies on the coastal region of the world's oceans has elicited interest in the so-called "coastal boundary layer"-that band of water adjacent to the coast where ocean currents adjust to the presence of a boundary. Within this zone, roughly 10 km wide, several physical processes appear to be important. One of these, the tides, is of particular interest because their deterministic nature allows unusually thorough analysis from short time series, and because they tend to obscure the other processes. The Coastal Boundary Layer Transect (COBOLT) experiment was conducted within 12 km of the south shore of Long Island, New York to elucidate the characteristics of the coastal boundary layer in the Middle Atlantic Bight. Analysis of data from this experiment shows that 35% of the kinetic energy of currents averaged over the 30 m depth are due to the semidiurnal and diurnal tides. The tidal ellipses, show considerable vertical structure. Near-surface tidal ellipses rotate in the clockwise direction for semidiurnal and diurnal tides, while near-bottom ellipses rotate in the counterclockwise direction for the semidiurnal tide. The angle between the major axis of the ellipse and the local coastline decreases downward for semidiurnal and increases downward for diurnal tides. The major axis of the tidal ellipse formed from the depth averaged semidiurnal currents is not parallel to the local shoreline 2 but is oriented at an angle of -15 degrees. This orientation "tilt" is a consequence of the onshore flux of energy which is computed to be about 800 watts/m. A constant eddy viscosity model with a slippery bottom boundary condition reproduces the main features observed in the vertical structure of both semidiurnal and diurnal tidal ellipses. Another model employing long, rotational, gravity waves (Sverdrup waves) and an absorbing coastline explains the ellipse orientations and onshore energy flux as a consequence of energy dissipation in shallow water. Finally, an analytical model with realistic topography suggests that tidal dissipation may occur very close (2-3 km) to the shore. Internal tidal oscillations primarily occur at diurnal frequencies in the COBOLT data. Analysis suggests that this energy may be Doppler-shifted to higher frequencies by the mean currents of the coastal region. These motions are trapped to the shore and are almost exclusively first baroclinic mode internal waves. Thesis Supervisor: Gabriel T. Csanady Title: Senior Scientist Woods Hole Oceanographic Institution ACKNOWLEDGEMENTS This thesis contains contributions few of which can be mentioned in would like to thank. my thesis support, encouragement, and member of my this short space. advisor, patience. committee, thesis kindly only a from many individuals, Dr. Dr. G. R. In particular, I T. C. provided Csanady, for his Beardsley, another the from data the CMICE experiment as well as a number of helpful discussions. Tom Hopkins and Jim Lofstrand, Dr. Laboratory, COBOLT spar braved were very helpful buoy data. inclement In weather in both from Brookhaven National obtaining a useable version addition, Bert Pade and seasickness to and Neal perform of the Pettigrew the daily hydrographic surveys. The the data processing was Woods Pennington, Hole in aided Oceanographic particular, was greatly by Institution consulted on the "Data Buoy Dollies" Group. numerous of Nancy occasions and cheerfully corrected many of my programming errors. I would also like to thank Doris Haight and May Reese for their help in typing the various versions of this thesis. This Brookhaven 359133-S. work was National supported by Laboratory the Department subcontract of numbers Energy 325373-S under and TABLE OF CONTENTS ABSTRACT . . . . . . . ACKNOWLEDGEMENTS . . . TABLE OF CONTENTS. . . LIST OF FIGURES. . . LIST OF TABLES . I. II. . . . . The Coastal Boundary Layer and COBOLT Experiment A. Introduction . . . . . B. The COBOLT experiment. C. The experiment site. . D. Coastal measurements . E. The COBOLT instrumenta tion . F. COBOLT experiments and data. . . G. Data processing. . . . . . . . . .a Nearshore Tidal Current Observations A. Introduction . B. Tidal analysis . . C. The tidal ellipse. . . . . . . . . . 4. . . . . . . Tidal observations in the Middle Atlantic Bight. Analysis of the COBOLT tidal signal. . t. .a.. 5 F. Results of the semidiurnal analysis. . . . . . . . G. Band structure of the semidiurnal admittances. . . H. Results of the diurnal analysis. . . . . . . . . . I. Consequences and conclusions . . . . . . . . ... III. Tidal Dynamics and Theory A. Tidal theory . . . . . . . . a. .. . . ....... B. Vertical structure of tidal currents . . 89 C. Effects of friction on tidal propagation . . . . . 99 D. The Sverdrup-Poincare E. The Sverdrup wave--no reflection . . . . 101 . . . . 107 F. The Poincare wave--perfect reflection. . 107 G. The combination Sverdrup-Poincare wave . 113 H. Comparison of the model to observations. 123 wave model I. The effects of local topography. . . . . J. IV. Summary. . . . . . . . . . . . . . . . . . . . . . 128 144 Observations of Coastal Internal Tides Introduction . . . . . . . . . . . . . . . . . . . 147 B. Dynamical theory of the internal tides . . . . . . 148 C. Solutions for constant Brunt-Vaisala frequency . . 150 D. Solutions for an arbitrarily stratified fluid. . . 152 E. The mean fields of the COBOLT experiment ..... 156 A. 6 G. Internal tidal oscillations. . . . . . . . . . . 166 H. Modal structure of the internal tides. . . . . . . 178 I. Comparison to theory . . . . . . . . 190 J. Energy and flux of the internal tide . . . . . . . 190 K. Conclusions . . . . . . . . a.. .... REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BIOGRAPHICAL NOTE. . . . . . . . . . . . . . . . . . 192 . . . . . . . . . . . . 194 LIST OF FIGURES FIGURE 1-1 PAGE Cross section and classification of the Middle Atlantic Bight continental shelf 14 1-2 The South shore of Long Island 17 1-3 Configuration of a typical COBOLT spar buoy 22 1-4 Locations of the four spar buoys of the May, 1977 experiment 1-5 Depth profile off Tiana beach, location of spar buoys, instrument packages, and hydrographic survey stations 1-6 25 26 Calendar for the May, 1977 experiment showing buoy duration and hydrographic surveys 27 1-7 Configuration of MESA-CMICE mooring #5 30 2-1 Definition sketch of the tidal ellipse 42 2-2 Cophase contours of the Middle Atlantic Bight from to Swanson, 1976 2-3 Tidal ellipses in the New York Bight from Patchen, Long, and Parker, 1976 (MESA) 2-4 2-5 45 49 Variance preserving plot of the kinetic energy of depth-averaged currents at buoy 2 55 Semidiurnal tidal ellipses in the COBOLT experiment 59 2-6 Semidiurnal tidal ellipses in the CMICE experiment 60 2-7 Vertical profile of semidiurnal ellipticity 61 2-8 Vertical profile of semidiurnal orientation 63 2-9 Vertical profile of semidiurnal kinetic energy 64 2-10 Semidiurnal ellipses for depth-averaged currents 66 2-11 Band structure of semidiurnal admittances 70 2-12 Vertical profile of diurnal ellipticity 73 2-13 Vertical profile of diurnal orientation 74 2-14 Vertical profile of diurnal kinetic energy 76 2-15 Diurnal ellipses for depth-averaged currents 77 3-1 Theoretical vertical profiles of semidiurnal ellipticity 93 3-2 Theoretical vertical profiles of semidiurnal orientation 94 3-3 Theoretical vertical profiles of diurnal ellipticity 96 3-4 Theoretical vertical profiles of diurnal orientation 97 3-5 Geometry of the step shelf tidal model 3-6 Ellipticity versus offshore distance for perfect reflection (r'=1) and several positive incidence angles 3-7 110 Orientation angle versus offshore distance for perfect reflection (r'=1) and several incidence angles 3-9 109 Ellipticity versus offshore distance for perfect reflection (r'=1) and several negative incidence angles 3-8 102 112 Ellipticity angle versus offshore distance for imperfect reflection (positive incidence angle) 115 3-10 Ellipticity versus offshore distance for imperfect reflection (negative incidence angle) 3-11 Orientation angle versus offshore distance for imperfect reflection 3-12 139 Kinetic energy versus offshore distance for depth dependent model 3-21 138 Ellipticity versus offshore distance for depth dependent model 3-20 131 Orientation angle versus offshore distance for depth dependent model 3-19 130 Comparison of modelled depth dependence with actual depth profile off Tiana Beach 3-18 124 Plot of ln(H -H) versus offshore distance at COBOLT site, plus linear curve fit 3-17 122 Ellipticity versus Orientation angle at y=O with COBOLT observations plotted for three semidiurnal frequencies) 3-16 121 Ellipticity versus Orientation angle at y=O; varying incident wave frequency 3-15 120 Ellipticity versus Orientation angle at y=O; reflection coefficient varied while holding incidence angle constant 3-14 117 Ellipticity versus orientation angle at y=O; incidence angle varied while holding reflection coefficient constant 3-13 116 141 Dissipation versus offshore distance for depth dependent model 143 3-22 Free surface elevation versus offshore distance for depth 144 dependent model 4-1 Twenty-two day average sigma-t cross section at the COBOLT site 4-2 161 Twenty-two day average Brunt-Vaisala frequency profiles at buoys 2-4 164 4-3 Salinity time series from instruments 31 and 34 167 4-4 Sigma-t time series at all levels, buoy 3 168 4-5 Sigma-t energy density spectrum from insts. 170 21-23 & 31-33 4-6 Temperature energy density spectrum from insts. 21-23 & 31-33 171 4-7 Onshore velocity energy density spectrum, insts 32-32 172 4-8 Alongshore velocity energy density spectrum, 4-9 Kinetic energy density at 15 m for Sept., 4-10 First three vertical velocity modes at buoy 3 179 4-11 First three horizontal velocity modes at buoy 3 181 4-12 Fitted first baroclinic mode at buoy 2: insts 32-33 1975 experiment Fitted first baroclinic mode at buoy 2: velocities 177 onshore 184 velocities 4-13 173 alongshore 185 LIST OF TABLES TABLE PAGE 1-1 Summary of data returns from Tiana Beach site 31 2-1 The wavelength of the semidiurnal tide 47 2-2 The tidal components used to construct the reference time series 52 2-3 The semidiurnal admittances and coherences 56 2-4 The band structure of the semidiurnal admittances 69 2-5 The diurnal admittances and coherences 71 4-1 Numerical values of T, S, N, and sigma-t: buoy 3 161 4-2 Longshore mean currents at all instruments 165 4-3 Isopycnal displacements for buoys 2 & 3 for all levels: 12 hr., 4-4 18 hr., and 24 hr. periods 175 Energy distribution among modes 0-2; buoys 2-4; 12 hr., 18 hr., and 24 hr. periods 183 CHAPTER I THE COASTAL BOUNDARY LAYER AND THE COBOLT EXPERIMENT A. Introduction The coastal regions of the world's oceans have been the subject of increased interest among physical oceanographers This narrow band of shallow water surrounding been regarded as too insignificant deep ocean, theories. and as too The economic fisheries and energy scientific interest important study in its have also spurred (see by reviews conform the activities, dynamics own right. interest shallow water dynamics to and are not Niiler of to simple however, have the led has long the offshore promoted continental to dynamical of seas Improved measurement have decade. the great volume of the and environmental considerations related in the continents to affect complicated in the last new as an capabilities realization that as complicated as originally supposed (1975) and Winant A (1978)). complete understanding of the interaction of these regions with the rest of the ocean may yet prove the shelf's importance to the deep ocean if only as a boundary condition. The breadth of the continental areas within Fleming, shelf the 1942), break or one hundred meter shelf is by definition isobath (Sverdrup, limited Johnson, to and though shelf studies often pass beyond the continental continental slope in order to include important conditions in the transition of shallow to deep ocean flow. east coast of the United States, Middle Atlantic Bight, distance of 100 km. the a region known as the specifically in shelf extends Off the typically to an offshore A representative cross section of this particular region is shown in figure 1. The eastern continental areas depicted in figure boundary dissimilarities layer of the a 1: known as the shelf break; coastal shelf is often subdivided region of sharp further into the topographic inner and outer self regions; (CBL) inner close and to outer the shore. shelf, and and, a narrow The the change, dynamical shelf break, often noted as the basis of this classification scheme, are summarized in Beardsley, Boicourt, and Hansen (1976). The This layer. wide, region that interest here is the boundary coastal applied to a band of water on the order of 10 km term is which is of is small compared to the width of the continental shelf, but large compared to the several hundred meter width of the surf zone or littoral layer is zone. From a physical the region where offshore standpoint, the coastal boundary to the presence of currents adjust the coast. Early features work which on are Great the peculiar observational particular, (Csanady, Lakes to evidence the and coastal primary transient mechanism meteorological by which forcing. the layer. modelling led In to 1977 for more details)--- nearshore With revealed has boundary theoretical the concept of a coastal "jet" (see Csanady, the 1972) regard waters to the respond to relatively 20 40 60 80 100 \ A!00 150 200 250 Figure 1-1 Cross section and classification of the Middle Atlantic Bight continental shelf uncomplicated dynamics of large lakes, this model has substantially increased the understanding of coastal boundary layer processes. While application boundary layers difficult are presence the And, than of what the strong current equivalent tidal are important in theory observations do Great currents So, jet observational observations with the rest of the shelf. scales coastal straightforward, since suitable rare. interpret is of exist oceanic confirmation in the are Lakes more appears scale that coastal is more coastal region difficult observations and large it to due flows to to the associated two additional the oceanic coastal boundary layer: time the mean circulation, and tidal frequency motions. As part of the Coastal thesis is directed toward Boundary Layer developing Experiment an understanding frequency motions of the coastal boundary layer. by presenting a description of the followed by a conceptual model features or of the barotropic baroclinic tides is that reproduces addressed tide. of the This goal is tidal currents or surface (COBOLT), tidal pursued of the coastal many of this zone the observed The question of internal a detailed description and Transect (COBOLT) experiment was with comparison to existing models. B. The COBOLT experiment The designed COastal BOundary specifically to Drawing from experience of newly developed Layer study the complexity of the coastal zone. gained on the Great Lakes and taking advantage instrumentation, it was planned to provide a 16 detailed spatial boundary layer, with the the large motivation for stations little was currents picture induced circulation experiment offshore, known project temporal scale the power The and represents the the with joint by the efforts and the shelf. proposals to realization coastal the of coastal interaction continental the about wind-driven tides, provided together the by of was observationally of locate that boundary Woods The very layer. Hole Ocean- ographic Institution and Brookhaven National Laboratory (BNL). C. The experiment site The southern coast of Long Island was chosen for the site of the COBOLT experiment because of coastline. This region is The approximate 30'W. The figure 2. shown in km west of Montauk Point, and 60 Apex, coordinates site enjoys the of easy idealized Tiana Beach, and is the terminus from straight the shore of Long Island. 400 is 720 and of waters about 6 km east of distance reasonable 45'N protected the Inlet which also within are experiment access Shinnecock Bay through Shinnecock Tiana Beach, an 135 km east of New York City and the New York Bight is location point, its similarity to of Brookhaven National Laboratory. Geographically, the coast of Long northern boundary of the Middle Atlantic Bight. virtually continuous barrier for entrances water to topography protected is formed sand bar, bays from in forms Island The coast itself with only its loose, part the is a four or five breaks km extent. The large-grained sands 195 of shallow and is AN 0 50 uN 0 410 0 NG TIANA BEACH OCEAN ATLANTIC 40*30' Figure 1-2 The south shore of Long Island ffimffikwwwwmiav 18 as 1973) et al., (Swift features smooth with minor "swale" remarkably the only irregularities. topographic While uncomplicated are smooth there are other dynamics, and Swanson, show 1976) close-to-resonant response vicinity of Montauk Point to point. aberrations strong gives and rise the Sound Also, (Ketchum and Corwin, 1964). phases tidal of the data. is (Redfield, 1958 a major propagation tidal in to the source of the in rapidly change A Sound. currents large that COBOLT the likely to have some is entrance to very relatively of observations to 50 km away from the up characteristics point Tidal COBOLT measurements. on to lead features for example, The presence of Long Island Sound, effects and the interpretation complicate which may site experiment features from fresh water Since the runoff from Long Island itself of any fresh- is relatively minor, the Sound is probably the origin ening that occurs at the COBOLT site. In addition, measurements. Inlet the proximity of Shinnecock is Though it narrow (about may influence 200 m wide) and less the than 5 m deep at most points, visual surveys indicate that the plume of tidal discharge as 2-3 km out to sea and is reaches Thus, 6 km. it conceivable is visible as that moorings which far down-shore are close to shore may show the effects of being near to the inlet. D. Coastal measurements One of measurement in mounting major hurdles encountered program is that of choosing adequate the a near-shore instrumentation. It 19 is well known profoundly carefully that affected by conceived Instruments which vane), current such as meters high sampling mounted frequency and near surface are waves even when gravity the averaging schemes are employed. sample speed and direction (via Savonius the VACM or Aanderaa current meters, rotor and are particularly susceptible to rectification of wave-induced orbital velocities, even when mean velocities Taut rope moorings are Strong currents, cause sizable layer comparable magnitude (McCullough, also contribute to measurement ways. surface of errors such as those encountered in vertical excursions fluctuations to contaminate measurements of the in several the coastal zone, instrumentation. can be transmitted at deeper 1977). down Also, the flexible instruments. Finally, rope the lack of torsional rigidity may introduce directional errors. The presence of a nearby coast In own. the addition to the increased nearness polarized in of the coast adds measurement problems error causes low current of orientation velocity--an frequency currents to be the alongshore direction and increases the probability of alongshore in its possibility of human interference, measuring important- onshore velocities strong of can amount which is 50 incorrectly. cm/sec, cause a 1 as little cm/sec comparable For example, as error in one degree in the a of onshore to the true mean value of the onshore currents. To mooring the list of design must salinity be measured. difficulties be added the to be demand overcome that Unlike the deep ocean, in both instrument and temperature and where tight temperature- 20 salinity properties make possible and eliminate shallow coastal a functional (somewhat) waters have relationship between the need for salinity -no such links. Density the two time series, variations are controlled by salinity at certain times of the year and by temperature at other times, and both signals are usually large. In order to separate dynamic effects, time series of both parameters are essential. Despite the difficulties, carried out in the coastal several zone of conventional measurement techniques. useful experiments the Middle Atlantic have been Bight using Two of the most notable of these are the EG&G Little Egg Inlet experiment (EG&G, 1975) and the New York Bight MESA project (Charnell and Hansen, successes, eliminate a concerted the potential ation and moorings, in effort earlier was sources of 1974). made in These view of these the COBOLT experiment error in conventional and to add measurement studies. Even in capabilities requirements to instrument- not available necessitated a radical departure from common deep water mooring design and instrumentation. E. The COBOLT instrumentation The Spar", mooring platform the COBOLT instruments, was developed for coastal work off La Jolla, constructed of sections Brush, for 1972). The the California. of 2 1/2" diameter PVC pipe (Lowe, moorings utilize specially without sacrificing designed pipe. Since it is too much of -the inherent torsionally rigid (torsional It Inman, joints to allow the spar to articulate freely at the several points, "Shelton is and universal junction rigidity of the variations are estimated by the manufacturer requires only one compass current meters buoyancy mounted element typically 100 in to be less than 10), the mooring to determine the orientation of the on it employed, in the rigid steel cages. mooring a 50 cm/sec current. also four With the large tilts very little; Thus much of the vertical and rotational movement of conventional moorings is eliminated. Instrument packages consist of two temperature probes--one "local" and one "remote"--and Marsh-McBirney, Inc. induction-type Model 711 conductivity electromagnetic sensor, and current meter. a The current meters have two orthogonal sets of electrodes mounted on a 2 cm diameter vertically oriented cylinder. The principles of operation of the electromagnetic current meter are discussed in Cushing (1976). A typical mooring configuration, pictured four of the instrument packages described above, orthogonal tilt transmitter. sensors, an in situ data in figure 3, employs plus one compass, processor, and two a radio Sensor outputs are low-pass filtered in real time with a five second time constant (the stated response time for the sensors is typically processor. one second) Averaged and continuously values of the integrated measured in the data parameters are then transmitted, on command, to a shore station at Tiana Beach. can therefore adjust the sampling rate or detect faulty instruments while the experiment is in progress. typically to one month periods, transmitter. al. (1976). Operators Experiment duration is limited, by the large power consumption of the Further technical details are available in Dimmler, et. Radio Telemetry link to BLN Floatation collar 3.8m- ECM 7.8 m- ECM Data processor 16.0 m ECM Compass and inclinometer Shelton spar 28.4m~ -* Fu *- Configrat 6 io . ECM .- Battery container ---.- * . o a t C - 30.8m spar buoy Figure I1--3* Configuration of a typical COBOLT spar buoy In spite of the care taken in its design, the COBOLT moorings have not been perfected yet. An experiment somewhat related to COBOLT, the Current Meter Inter-Comparison- Experiment (CMICE), was conceived as an opportunity to test the merits of the spar system against coventional moorings and instruments. In this experiment, Beardsley, six moorings were deployed off Tiana Beach et. al. (1978), described in in a line parallel to the shoreline and 6 km from the beach. detail by Four of the moorings were conventional taut rope moorings instrumented with a variety of current type), while meters (mostly of the Savonius the remaining two moorings were rotor and vane the Shelton spars. A comparison of the measurements of these instruments suggest that there are some deficiencies in the COBOLT moorings and intrumentation. The sources of possible error in the COBOLT velocity measurements are: 1. Errors due to mis-orientation of the single compass or misalignment of current meters with respect to the compass. 2. Errors due to a shift in the zero point of either or both of the current meter axes. 3. Errors due to asymmetric gain adjustment of the two current axes or non-cosine response of the sensors. F. COBOLT experiments and data After some pilot studies, the full COBOLT array of four spar buoys was first deployed in May, 1977. buoys and their relationship to The location of each of the four surrounding features is shown in 24 figure 4. The buoys were placed approximately 3 km, km away from the shore, and stand in 20 m, 28 m, 6 km, 30 m, 9 km and 12 and 32 m of water respectively. The instrument configuration, daily hydrographic casts ically in figure 5. numbers: the first the instrument Instruments mounted, starting at the top, place instruments below the (described corresponding is in at surface; bottom profile, subsequently) are and is standard identified by a sequence of two to the number of the buoy on which mounted. is depths: intermediate the to the order, An attempt was made shallowest instruments at 7.4 at to this rule with one instrument to 3.8 meters meters meters; and the deepest at 2.4 meters above the bottom. exception of schemat- shown and the second corresponding which it location and 16.0 Buoy 1 is the at 12.3 meters instead of 16.0 meters. The spars were launched on recovery from all four buoys was non-uniform power significantly. until May 24; April initiated drain, endurance Buoys 29, 1 and 3 were of 1977, and on April the different operational until and buoy 4 until May 17. regular 30. data Because of moorings varied May 29; buoy 2 The operation period of the experiment is summarized in figure 6. The quality temperature(2), of instrument records (containing temperature(l), salinity, X velocity and Y velocity).is good, with the exclusion of buoy 1 which suffered numerous irrecoverable data gaps. These gaps were uniformly spread throughout the data and amounted to a total of 140 hours out of a total duration of about 700 hours or I I I 1 0 1 2 3km 3. 4, 30' Figure 1-4 Locations of the four spar buoys of the May, 1977 experiment MAY 1977 hydro. 20 .c S 4- 30 40 5 10 km Figure 1-5 Depth profile off Tiana beach, location of spar buoys, instrument packages, and hydrographic survey stations MRY 12 3 APR. 30 START BUOYS H 77 7 HH H 2 3 4 a H H H 6 hr gap buoys 3, 4 H 16 H H H H H H H H H END BUOY 4 22H 23 H 2 4H 2 S END BUOY 25 H H 2 H E ND BUOY 3 H Hydrographic survey conducted Figure 1-6 Calendar for the May, 1977 experiment showing buoy duration and hydrographic surveys one-fifth of the total time. Eree of isons, long but gaps and consequently the rest of tidal analysis. One stretch of the record ten days was can be used was abandoned for as Data from the other three moorings, only occasional, short data gaps during periods relatively limited compar- unacceptable buoys 2-4, for showed of high speed flow. measurements, daily These gaps never exceeded 6 hours in length. In conjunction hydrographic surveys ments were made along a line stations, with of continuous the area were from a small vessel coincident with the about 1 km, the coastal the was chosen boundary layer the spar buoys. Although buoy conducted. at spar These STD measure- ten semipermanent transect. to give more The spacing of the detailed resolution of than was provided by the 3 km spacing of they were performed only in and although they are aliased by tidal fluctuations, surveys are a valuable locations source of information in fair weather, the hydrographic interpreting the spar data. In view of the questions quality of the spar system, of the tidal analysis that and in to follow, be included in the discussion: have arisen an effort results concerning the data to assure the generality from two other moorings will a "reference" mooring from the CMICE experiment, and the COBOLT pilot mooring. The mooring chosen from the CMICE experiment was deployed by the MESA New York Bight project and has been used extensively in their field program. The current meters; three mounted on a subsurface taut wire mooring, and a instrumentation consisted of four Aanderaa RCM-4 fourth mounted biases. The instrument at experiment was beneath mooring 11 a surface is shown meters below conducted at spar buoy to schematically the surface the COBOLT reduce in did site figure not convention used here, it is One function. The 1976 with at approximately the same location as buoy 2 of the May COBOLT experiment. the 7. in February, this particular mooring positioned 6 km offshore designated as #5 in the CMICE experiment wave-induced The mooring was and since this conforms retained in Table I and to in further references. The COBOLT pilot mooring, launched in September, single mooring placed 11 km offshore at roughly buoy 4 of the May, packages at 7.8 m, details 1977 16.0 m, concerning this experiment. It the same had and the was instrument 32 m of water. others a location as working and 27.0 m and was in mooring 1975, employed in The this analysis are summarized in Table 1. Although it seems a bit capricious taken during different year, the there same seasons and are elements throughout stratification are istically related the the COBOLT separated Aanderaa current meters time by more than year. Even the if tidal forcing meteorological signal at all should times. be Including these allow comparison between certain spar buoys and the relatively of the CMICE experiment, and determin- aspects of well-understood and will also assure that measurements are somewhat representative of different seasons conditions. a signal which are expected to remain to well-known forces experiment in of the different, additional moorings will to compare current observations and Spar buoy W~AJ Auxiliary float 9.0 M- 51 3.0m I1# buoyancy 10.0 m-( 955t#: buoyancy 1 1.0 m- Aanderoo 52 3/16" Wire rope 15.7 m~n Aanderoa 53 25.0 in- Aande raa 54 A coustic release Dan forth anchor Figure 1-7 Clump anchor Configuration of MESA-CMICE mooring #5 TABLE 1-1 SUMMARY OF DATA RETURNS FROM THE TIANA BEACH SITE Water Dist. Depth of working Depth offshore Current meters No. Duration Sept., 1975 0 640 hr 32.6 m Feb., 1976 5 697 26.5 6 3.0, 15.7, 25.0 May, 1977 1 240 20.3 3 3.8, 7.8, 12.3, 17.9 2 577 27.7 6 3.8, 7.8, 16.0, 25.3 3 700 30.8 9 3.8, 7.8, 16.0, 28.4 4 385 32.3 12 Date of Exp. 11 km 4.2 m, 16.5 m, 29.7 m 3.8, 7.8, 16.0, 29.9 G. Data Processing The data sampling scheme is unique to the spar system and presents minor problems and at of its own. intervals that Buoys were ranged from interrogated at separate five minutes to several times hours. Since ordinary time series analysis demands that sampling intervals be uniform time, and the one-hour that measurements COBOLT data sampling were over a one-hour closest whole comparison adjusted interval interval in the data). for (one to hour a time period hour. The and then taken common was This was achieved be by base the most scheme common with a common averaging all data interpolating interpolation a time far by first at was values a to third the order polynomial that used four data points (two on either side of a gap) to determine the value of the function on the hour. advantage of interpolation, consistent eliminating and curves. of For the filling sharp bends gaps periodic in one-half of a period. Using this as introduced strong functions, nomial interpolation gives a good visual This method has the fit for by tidal linear flows with the poly- example, for record gaps of up to a guideline, COBOLT data gaps were filled only if they were less than or equal to 6 hrs in duration; that is, half a semi-diurnal tidal period. The X and Y component were converted velocities output from to east and north components the current meters using the headings the single on-board compass. Then the coordinate system was by local 220 to uncertainties conform usually to the associated coastline with this at Tiana maneuver are from rotated Beach. quite The small here The to due to result the uniformity of the coastline is and features. system with the X axis aligned alongshore the Y axis the north- a coordinate the east-northeast and topographic pointing onshore to northwest. The special salinity time attention. conductivities) ments series and by as much measurements. salinities Mean differed from the May, 1977 experiment by as much (computed as 3 o/oo These aberrant suggest that these salinity measurements aberrations constant calibration offset, two different procedures. to eliminate all density were the the first, inversions, instru- from nearby resulted in STD large, Since there was nothing other than the result an effort was made to correct In measured from adjacent as 2 o/oo from values obtained persisent inversions in the computed density. to from required of them using salinities were offset enough while in the second, salinities were made to conform to nearby daily hydrographic survey salinities a least-squares sense. credence to the notion a in These adjustments agree quite closely and give that errors were due only offsets and not to instrument drift or malfunction. to calibration CHAPTER II NEARSHORE TIDAL CURRENT OBSERVATIONS A. Introduction An examination of the current records from any coastal experiment in at the Middle Atlantic Bight shows that they are dominated (vis.ually least) by oscillations properties large tidal do of not the amplitude oscillations. transport water of the tidal the records--particularly consequence, column if an understanding Even mass, such momentum, (except signal though or short other in non-linear often obscures period passive cases), the other aspects the observation period is short. of As a of some of the slower and less obvious processes of the coastal region may be improved by an understanding of the tides. Certain controlled aspects of coastal or influenced by dynamics may the surface tides. also be Internal directly waves, for example, are known to be generated by tidal currents interacting with the topographic There is currents coastal found in coastal control the high background level regions--acting, in effect, like to the question of tidal presumed to occur on the continental (Munk, areas (Rattray, also evidence (Bowden and Fairbairn, 1956) closely related is features 1968). Little is that 1960). the tidal of turbulence observed in a stirring is dissipation, rod. This much of which shelves of the world's oceans known about the mechanisms by which this is accomplished or the regions in which it occurs. A study of coastal tides may serve to illuminate the subject. Because of the deterministic nature and relatively high frequency of tidal currents, information can be extracted from relatively short duration experiments. 1977 is The thirty days of data gathered during May suitable for some forms of tidal analysis and will be used in the hope of elucidating region, comparing other systems, the some of the local performance and as a first of dynamics the step in of the nearshore COBOLT mooring system obtaining detided records to for analysis of low frequency phenomena. B. Tidal Analysis Tidal method which analysis introduced forcing is traditionally carried by Lord Kelvin in 1867. occurs, are obtained from out using The the harmonic frequencies, expansions of ., at the tidal potential (Doodson and Warburg, 1941) and used in the expression F(t) which is adjusting = a. cos (W.t + #.) 1 then fitted the constants to the a. 1 1 data in and *. , (1) a least-squares This method sense requires by long records, typically greater than a year, in order to resolve some of the closely spaced constituents, and to provide statistical stability since weak Also, the similarities tidal "lines" in are often obscured by background noise. responses to given forcing are concealed in the multitude of different amplitudes and phases. So it is not well suited to the analysis of short records. In harmonic analysis, statistical stability is at the expense of resolution. different in frequencies (Bendat spectra but and improves Piersol, the In differences adjacent example, fifteen days the since components these Averaging spectral Thus, minimum lunar differ estimates resolving capabilities analyzing to resolve different the spectral estimates short frequencies. the principal spectral ability time series the averaging procedure obscures is resolution of the reliability of 1971). since between or averaging across frequency bands reduces problem is critical days. That is, averaging the spectra of many pieces or realizations, individual usually maintained In tidal record by over one n cycle frequency that solar in spectral analysis, length and principal estimation or harmonic- techniques long term observations. of the allows fifteen bands tidal for constituents days. limits to signals which differ by n cycles in reliable this the fifteen constituents by depends on the availability of fairly If this criterion is not met the so-called "admittance approach" offers a viable alternative. The approach, one method is used x(t), the "output", analyze the COBOLT data, described by Munk and Cartwright (1966). hypothesizes series, to a linear, which is the most causal termed general relationship the "input", linear can be defined by the convolution integral, admittance Basically, if between and y(t), relationship the which between two is time termed the two y(t) where h(t) is f x(t') h(t-t') = known as the dt' (2) impulse response function. Defining the Fourier transform by capital letters, i.e., = F() f f(t) e dt , (3) and taking the transform of equation (2) gives = Y() H() X(W) , (4) where H(o) is the transfer function or admittance. Since one rarely works with direct transforms, but rather with spectra, the following definitions are useful: AUTO-SPECTRUM S (W) = X(o) X*(o) (5) CROSS-SPECTRUM (where (4), it * indicates S a xy (o) complex = X*(W) Y(W) conjugate) ; from which, using equation follows that S (W) = H(w) S (w). (6) 38 If x(t) is a periodic function, say x(t) a exp iwt = (7) , equation (2) assumes a particularly simple form y(t) This form since it the is especially = H(W) x(t) . useful in (8) generating the output is more easily computed than equation (2). conceptual basis of the admittance; it function, It also reveals is a measure of the spectral linkage between the input and the output functions. The primary advantage of the admittance analysis to reduce noise to well-defined levels without tion. This Smoothness" amplitudes bands. is accomplished by (Munk and Cartwright, and phases This is invoking the are relatively that are being forced at sacrificing so-called smooth over resolu- "Credo of broad frequency that the response of most physical systems does not change too abruptly if altered. the ability 1966) which states that admittance based on the observation forcing or input is is the frequency of the Exceptions to this argument are systems close-to-resonant successful use of the admittance approach frequencies. The does not depend on a high degree of resolution because the admittance function varies so slowly with frequency that any structure in records or low resolution analysis. it can be discerned with short Because the input is generally a 39 well-known function for resolution analysis of equation (8) which long output time time series series are can available, be high obtained once the form of the admittance function is from known. Instead of resolution and stability, the questions to be answered in the admittance function. output The that Smoothness"; time ideal the it periods approach center input function admittances is to available allow the on the is proper related necessarily (or choice so conform of an input closely to can be constructed) the to the "Credo of for long enough desired resolution; and, another important advantage. it is free of noise. The analysis admittances offers are formed from ratios, they tend to divide out some of the numerical effects of the finite Fourier transform. of interest in Because This is again the processing of short time series where information from narrow frequency bands is spread out into relatively broad bands by the effective transform manner, alters filtering both the of the input transform and output process. functions Because in the a similar these effects are minimized with the use of the admittance. Finally, the analysis provides a measure output is coherent or phase-locked to the input. of how much of the This measure is the squared coherence, defined as (Bendat and Piersol, 1971) Y2 S (m) 2 W(SS x 2 xy ) y (9) )S That part of the signal which has random variations phase, such as weather fluctuations or in amplitude and intermittent baroclinic 40 effects, is summarily classified admittances have, as a result, of the coherence. real and as noise. Ensemble well-defined averages of the errors expressed in terms A particularly simple form for the variance of the imaginary parts of the (which admittance are distributed normally) is given by Munk and Cartwright as JH(W) ~ 1 - y( 2 V where N is 2N 2 the number of statistical degrees of freedom and Y is the true coherence. Traditionally function when the equilibrium analyzing short tide duration is tide chosen gauge as or the input current meter data (see Filloux, 1971 and, Regal and Wunsch, 1973). The equilibrium tide, however, assumption that ocean. the of the from the entirely presence it astronomical forcing comes the the earth is coastal waters forcing by at computed tidal covered by of is land masses well-known under bodies plays and 1961) only a minor instead through interaction with the continental shelf outer continuity topography. boundaries. Here to drive more energetic in flows that role. the on the discussion of this subject is the following chapter.) For this tides water depth and act ical (Further The main pceanic tidal through forcing alone. In direct deep ocean shelf than could be achieved in the an infinitely deep (Defant, currents are constricted by the rapid decrease through potential It consequently does not account for variations that occur as result shallow, is continental the action of direct astronomcontained reason a series of coastal tide height function analysis for procedure suggested a presumably is observations of by coastal tidal Cartwright, series station is used as the input function in So, and input a following Zetler computed from the tidal harmonics (1969), a of a nearby tide analyzing the COBOLT data. The tidal ellipse The presentation of the data is the use of the tidal ellipse. are periodic v, which iv appropriate fields. Munk, reference C. more much can be formed. complex (-) Given the orthogonal velocities A+ and and A~, decomposed rotating in (+). counterclockwise into two opposite u and vector u + with some frequency w, the complex This vector may be vectors clockwise conveniently accomplished through constant, directions: Algebraically this is expressed as u + i v + A = iWt e - + A -iWt add to, alternately These rotating vectors e (11) or subtract from one another producing the characteristic shape of the tidal ellipse. phases of the vectors which determine direction the The is ellipse oriented. The parameters ellipticity and the succinctly which orientation. describe They are the ellipse illustrated in are the figure and defined (respectively) by: - LA= i- A A+ + A- 12) I ELLIPTICITY E= m/M ORIENTATION TIDAL Figure 2-1 ELLIPSE Definition sketch of the tidal ellipse argA + arg A 2 (13) In geometric terms, the ellipticity is the ratio of the minor axis of the ellipse to its axis. major It is: positive if the complex current vector u + iv rotates in a positive sense (counterclockwise); negative if the vector rotates in a negative sense (clockwise); equal to one if the vector traces a perfect circle; and equal to zero if the ellipse degenerates into a line. The orientation measures the angle between the major axis of the ellipse and the positive x axis. (The x axis will point alongshore and the y axis onshore throughout this work.) definition, to fall between These easier quantities are to visualize, but dynamics of the tides. constant phase) and It is by constrained, ±900. introduced, not only to make the results as tools diagnostic for determining the While the free surface co-phase (lines (lines co-amplitude of constant of amplitude) contours are valuable in this respect, the velocity field is quite sensitive sensitivity introduces dynamics. to other dynamic is a consequence ellipse (e.g., of the characteristics frictional) rotation that are of effects. the peculiar This earth which to certain Thus, it is advantageous to employ information from both surface and velocity fields in attempting any interpretations. D. Tidal observations in the Middle Atlantic Bight Under a common classification from the amplitudes constituents, characterized the as of four tides of predominantly scheme which uses a ratio prominent the semidiurnal Middle semidiurnal. and Atlantic This formed diurnal Bight ratio are (Defant, 1961), Ky1 +0O1 M + S ' 2 2 (14) ranges from 0.19 at Sandy Hook, New Jersey to 0.33 at Montauk Point, New York, and averages about 0.25 for the Middle Atlantic Bight in general. The M2 constituent is the largest; the ratio M2 :S2 typically being about 5:1 (Shureman, 1958). The Atlantic Ocean semidiurnal tide arrives everywhere at the edge of the continental shelf at roughly the same instant (Dietrich, 1944) and progresses with Jersey-Delaware shore. Sound affects cophase To the north, propagation contours the characteristics to-resonant response (Swanson, 1976). paralleling the New presence of Long Island markedly with its near- Cophase lines (see figure 2, taken from Swanson's work) bunch up around Montauk Point and distort normal tidal patterns many kilometers away from the Sound itself. a consequence, the tide propagates to the east (towards As the entrance to the Sound) along eastern Long Island and the west along central and western Long Island (in conformity to the rest of the shelf). The contours also show that the propagation pattern divides somewhere near the COBOLT region (station 20 on Swanson's map). Thus this area marks the transition between the tidal regime of the Bight and that Figure 2-2 Cophase contours in the Middle Atlantic Bight (from Swanson) TraWW-- M - tloction of the Sound, and complicated interactions between the regions may be expected. A crude estimate of the semidiurnal tidal wavelength, which will be valuable in the ensuing discussion, may be made by using the phase lags from the NOAA Tide Tables with the kinematic relationship between wave speed and wavelength, = Wavelength Phase Speed x Period. These figures suggest that this wavelength is The but diurnal seem to perpendicular tides to the from is difficult to north to south as with the semidiurnal cophase contours isobaths and coastline rather than parallel them (Dietrich, 1944). it about 1500 km (Table 1). are not so well documented progress (15) to In view of the lack of published information, characterize them propagation patterns differ noticeably except in noting from those of the that their semidiurnal constituents. Tidal current measurements appropriate analysis, current measurements in one of lightships the on from about fifty located dangerous shoals nearshore tidal currents. however. First of and shelf, are generally sparse. earliest studies were the at of the consequently all, Some tidal accompanied Haight light ships tidal large Most observations are compiled usually of harbors very complicated general ellipses to the on the East Coast currents. entrance are (1942) by these or on examples of may very be made, elongated TABLE 2-1 SEMIDIURNAL WAVELENGTH COMPUTATION Guage Location Distance from High & Low Water Interval Phase Speed Wavelength Sandy Hook Shinnecock Inlet 138 km Fire Island Jones Inlet 39 0.83 hr 1.13 hr 144 km/hr 1791 km 0.63 0.48 114 1413 0.32 0.45 104 1295 (the ellipticity more circular almost is at much less offshore exclusively in the than one) at nearshore And, points. clockwise velocity direction; locations vectors at and rotate 94% of Haight's observation points, according to Emery and Uchupi, 1972. Form measurements on the to 50% of the total variance the combined (1976), using ellipses are direction, effects the of while data and set, circular diurnal Flagg (1977)found at individual diurnal same virtually outer shelf, up current meters was due to semidiurnal notes and ellipses that that oriented are tides. semidiurnal in very Traschen the tidal cross-isobath elliptical and are oriented along isobaths. Nearshore current measurements, and Parker (1976) effects a nearby shoreline, of not influenced If there is diminished condition (low a to the solid the ellipticity) Apex, of harbors show if or bays onshore tidal boundary condition tidal ellipses to elongate Figure semidiurnal 3, tidal taken along at the axes show a noticeable orientation. Typically anywhere 5 0-10 -- and, from this it deviation orientation does not Patchen, local topographic or shoreline features. from to shore. must shore. from their "tilt" seem that are be This into very eccentric addition to the elongation of the tidal ellipses, major pronounced velocities from ellipses the Long, the measurements the forms. the those of Patchen, particularly boundary, satisfy causes as the New York Bight by the presence Parker, shows In in such it a is Long, and experiment. noted that shore-parallel angle is correspond small-to any 8 meters above bottom 73*30' 74%0 S New%York Rockaway Pont Sandy Hook 50 0 cm/sec New Jersey 1 520.000 N IStatute Miles 0 5 0 f Nautecalm 74%0' 0 73*MY 3 meters or less above bottom 74*0D' I uperO Say New York Rockaway Pont Sandy Hook S6 50 0 cm/sec New Jersey I 520.000 0 5 10 K 5 N0 74*00' Rotation is predominantly clockwise at 8 m (26 ft) above the bottom, counterclockwise at 3 m (10 Ift)or less above the bottom. The centers of ellipses are the station locations. m0 73*M' Arrows indicate direction of progress of the maximum M2 flood current velocity. Current velocity vectors rotate 3600 in 12 to 42 hours. Figure 2-3 Tidal ellipses in the New York Bight Source: From Patchen et &11976 Metcatr Poject ion Other tidal generalizations ellipses ellipses near ellipticities circular in direction. (8 can be m) are regarding made from the bottom (3 than those near shape) and By contrast, almost this m rotate, the away) the structure experiment. It usually surface generally, tidal ellipses always more vertical in the further eccentric appears exhibit (here of the that different they are more counterclockwise away from the bottom and rotate in the clockwise direction. Measurements coastal near Little Egg Inlet, series available for comparison, presence of the inlet. exception) J. (EG&G, 1975), another are highly influenced by the This, as was the case with Haight's analysis, makes generalizations difficult. predominantly N. clockwise The experiment does show, however, rotation and emphasizes the point of tidal that ellipses large (with amounts of one variance are due to the tides; 33% for year-long records in this case. There coastal appear tidal increased to currents interest usually focus mention of be on tidal in the few in this lower phenomena. other the relevant Middle Atlantic frequency on signal other of Bight Measurements region. Work studies despite which and neglect shelves nearshore do the exist altogether (e.g. Petrie, 1975), while serving as a useful comparison, will not be pertinent to the Middle Atlantic Bight because forcing and topographic features. of different deep ocean tidal E. Analysis of the COBOLT tidal signal the step in first The gauge Sandy at the assumed to be in moorings are operated over one to have closest the also necessary to obtain stable values the has, It prediction. in tidal regime of the open shelf and to that of Long Island Sound. to be too closely related not the COBOLT if a reasonable choice This is west of the COBOLT site. to the 140 km approximately New Jersey, Hook, the input The station chosen was the time series for the admittance procedure. tide the involved COBOLT data the from which to generate tide station of a reference choice of analysis It is of time the long period of tidal amplitudes and phases been operational for more than fact, for a hundred years. used to construct the reference time series constants The tidal were taken from Shureman (1958) and represent the results of harmonic The components employed are listed in analysis of ten years of data. table 2 along with appropriate phase relative to the of transit Non-astronomical meridian). amplitudes, periods, tides, over the the mean moon such as due those and components with amplitudes and radiational effects, (the epochs and to Greenwich non-linear that are less than 2% of the M2 amplitude were ignored. The input procedures as overlapping Fast by Fourier 1/15 generated function were data followed 360 hours pieces Transform cycles per with routine day. These was then to give meter current the (15 subjected days) long spectral estimates fall were to same the data; i.e., used with estimates a separated approximately on the TABLE 2-2 TIDAL COMPONENTS OF REFERENCE TIME SERIES COMPONENT PERIOD 11.96723 AMPLITUDE 2.9 cm PHASE 243 deg 12.00000 13.0 246 12.19162 3.4 203 12.42060 70.0 218 12.65835 15.9 201 23.93447 9.0 102 24.06589 3.2 105 25.81934 4.3 98 S2 and N2 ' M. the spectrum, 12.00 hr; diurnal, with and adjacent estimates approximately with estimates on at -at the 25.71 12.86 0 hr portion semidiurnal the for bands frequency and hr, K -P 24.00 and of 12.42 hr, and bands for the hr. The Fourier coefficients were hanned to reduce leakage of energy from the strong tidal into lines spectral the estimates weaker between ones, the and then reference used series to and form the cross- individual The admittances for each 360 hr piece were then velocity components. calculated according to equation (6). Three Besides types the different of averaging standard pieces and practice across among the COBOLT moorings effects the are dynamic effects frictional error, and these show baroclinic averaging used in this and real enterprise vertical systematic topography variations instruments presence of on different lie in moorings the same background horizontal "noise" are and expected due on admitto the due to sources of presumably subject be (see small; acceptable plane. cross-isobath vari ations at these locations. since other of the region seems taken lengths variations the should over were variations Unlike data. done to reduce the for buoys 2-4 where the depth changes only 4 m in 6 km. instruments COBOLT averages This was horizontal should be that horizontal the cross-spectra bands, influences. Examining the on errors and short record real topography variations to prediction. 1) suggests of to of themselves. Caution must be expected utilized frequency of individual instrument results. tances were especially So averaging provided Other errors to chapter swamp the and the any real As mentioned mooring 5 before, from represent two the parallel CMICE different analysis experiment. seasons--winter and provide a check for baroclinic effects. will furnish comparison between the CMICE performed these Since will a also was experiments spring--comparing Also, on the them experiments instrumentation and the data is newer COBOLT instrumentation. F. The results of the semidiurnal analysis The presence exhibited by of tidal spectral frequency analysis averaging was independent or variance preserving" combined, plot. the In time this account for about "barotropic" velocities--a the to series. some extent, the The area in this 35% of the total fairly typical beneath it makes and energy each to the "variance diurnal tides observed in proportion 4. true depth so-called semidiurnal The figure shown. in the contribution the COBOLT time at buoy 2 is series case, in velocity velocities. proportional to of the to isolate, barotropic frequency band is total done of currents spectrum of depth-averaged This motions in the coastal waters. The table admittances 3 for averaged over 12.86 hr, hours at and each for of the the semidiurnal instruments components deployed. are These entered have in been three frequency bands covering periods from 12.00 hr to over T/360 the buoy in pieces, question. where is Also included intervals computed from equation (10) oF the true coherqnce, T the record length in are the 95% confidence and using the unbiased estimate .8 Ld U) .6 .4 (9 a: LU LU C) FLU .2 0 .05 .10 FREQUENCY (C/HR) ,C1 .50 Figure 2-4 Variance preserving plot of kinetic energy of depthaveraged currents at buoy 2 TABLE 2-3 ADMITTANCE AMPLITUDES AND PHASES FOR THE SEMIDIURNAL TIDE INST. 21N 21E 22N 22E 23N 23E 24N 24E AMPL. 0.058 .162 .037 .153 .053 .148 .053 .095 ERROR 0.019 .017 .015 .011 .009 .010 .013 .007 PHASE -65 58 -79 58 -110 47 -170 34 32N 32E 33N 33E 34N 34E 0.052 .154 .052 .138 .048 .106 0.010 .010 .010 .012 .005 .007 -85 57 -92 50 41N 41E 42N 42E 43N 43E 44N 44E 0.028 .153 .039 .150 .070 .122 .052 .104 51N 51E 53N 53E 54N 54E 0.047 .147 .054 .184 .036 .118 ERROR 18 6 21 4 10 4 13 4 COH1. 0.83 .98 .78 .99 .94 .99 .90 .99 36 11 4 11 5 6 4 0.92 .99 .92 .98 .97 .99 0.029 .027 .029 .027 .014 .033 .009 .018 -44 61 -15 65 -99 54 -152 30 46 10 37 10 11 15 10 10 0.55 .97 .68 .97 .96 .93 .97 .97 0.009 .016 .016 .020 .009 .015 -88 53 -85 50 -150 40 11 6 16 6 13 5 0.91 .97 .85 .97 .88 .98 -153 Quantities listed under ERROR are the 95% confidence limits of the admitance amplitude and phase. COH is the true coherence of tidal currents with the reference series ^2 Y where N is, = NY Y2__ N - 1 (16) as before the number of degrees of freedom, and Y is the coherence estimate formed from equation (9). While the number of degrees of freedom are fairly low (generally less than 12) due to the short record lengths, the coherences are high, especially for the alongshore components. statistical nents, errors are kept to manageable As a consequence, Onshore levels. compo- though much noisier, exhibit a certain stability (with the exception of instruments 41 and 42) that suggests that these numbers are also trustworthy. Though longer records would do much to clear up doubts about the apparent discrepancies, certainly reliable. the admittances exhibit some trends which are Most noticeable is the disparity between onshore and alongshore admittance amplitudes. of the adjustment of velocities This is the a consequence to the presence of the shore and may partly cause the lower coherences evident in the onshore admittances since signal to noise ratios are presumably decreased also. Vertical trends are also evident. decrease towards drastically for the onshore bottom, while components. Admittance amplitudes usually phases This decrease also--quite tendency is always most evident in the current meter that is nearest to the bottom and may indicate the presence of a frictional boundary layer. Systematic differences between the May 1977 data and the February 1976 data experiments. are small Slightly despite larger the differences admittances for between the CMICE the two data and slight discrepencies in phase might be noted, but all variations are well outside the resolving capabilities of the analysis and conse- quently cannot be argued with much certainty. The semidiurnal tidal ellipse for each of the available instruments is moorings shown in its appropriate location in figure 5 for the COBOLT and in figure 6 for the CMICE mooring. The striking features of nearshore tidal flow are immediately apparent from these The ellipses are all very eccentric but are not oriented diagrams. negative inclination near the bottom. for they have a small but persistent Instead, parallel to the shoreline. (-100 to -150) Moreover, which becomes more the sense of rotation, which is noticeable clockwise all shallow and intermediate instruments, reverses to counter- clockwise for all bottom instruments. Some of the ellipses have noticeably different characteristics from ellipses at the same level on other moorings, or from adjacent Ellipses at two instruments already ellipses on the same mooring. alluded to, 41 and 42, have slightly different orientations than other instruments, while the ellipse at instrument 21 appears to have a slightly different ellipticity. consequence of problems outlined These discrepencies are probably a in chapter 1 (i.e., instrument related errors). Horizontal averaging provides a mean vertical profile of ellipse and tidal characteristics. The vertical profile of averaged ellip- ticity in figure 7, for example, shows plainly the characteristics described above. Ellipticities through most of the water column are COBOLT 21 10 cm/sec 41 22 32 42 23 33 43 24 34 44 Figure 2-5 Semidiurnal tidal ellipses inthe COBOLT experiment -4-- 51 CMICE 10 cm/sec 53 54 Figure 2-6 Semidiurnal tidal ellipses in the CMICE experiment 61 ELLIPTI CITY 0 -0.2 -0.4 0 0.2 95% 10 E 20 - 30 Figure 2-7 Vertical profile of semidiurnal ellipticity constant change and negative abruptly, indicating however, clockwise somewhere surface. Ellipses near the bottom, positive ellipticities indicating rotation. Unfortunately, this rotation. between in to those above, change feature values 16 m and 25 m below the contrast a These is to not have counterclockwise well resolved and little can be said about the structure in the region where it changes most rapidly. Also agreement between though the from this diagram the COBOLT and CMICE ellipticity is subject it is data to apparent sets large is that very errors good because the even it is formed from a small difference of two large numbers (equation (12)). The averaged orientation angle with respect to alongshore the direction (figure 8) has an almost linear trend with depth instead of changing It is more homogeneous suddenly. than for the COBOLT experiment, uncertainties involved distribution. More these though profiles importantly, for the CMICE experiment in view of could be orientation the the statistical part of angle is the same negative and significantly different from zero at all levels. Figure energy. 9 This shows the energy vertical was structure formed from of the averaged the admittances kinetic by multiplying them by the 70 cm amplitude of the principal semidiurnal constituent, M2, at Also Sandy Hook. the is included energy from the harmonic analysis of a four day period when the stratification, the internal tidal oscillations, was admittance approach and averaging in baroclinic signal is evident by strongest. and presumably The success of the eliminating much of the unwanted the vertical uniformity of the 63 ORIENTATION -30 -20 (0) -10 0 95% 10- E - CL 20 & COBOLT o CMICE May 77 Feb 76 30 Figure 2-8 Vertical profile of semidiurnal orientation ENERGY (Cm2/sec 2) 60 40 20 80 x 95% 10 E 1- 20 o COBOLT o CMICE May 77 Feb 76 X 4 day harmonic mean 30 Figure 2-9 Vertical profile of semidiurnal kinetic energy 65 energy. The CMICE mooring does show some variability in the vertical structure, but this is most likely due to the fact that the surface instrument was not attached directly to the mooring but to a tethered spar. As a result, the top meter shows much less of the surface wave contamination common in conventional moorings and instrumentation. In view of the homogeneity of well-covered water column, it tidal energy and the reasonably seems natural to form the "barotropic" ellipse by integrating the data vertically. Accordingly, the real and imaginary parts of the admittances were summed using the ordinary trapezoid rule to approximate integration. The velocity equations (using again the 70 cm tidal amplitude to convert admittances to velocities) are: COBOLT (17) U = 9.0 cos (Wt- v = 2.8 cos (Wt+ 1020) 510) and CMICE (18) u = 10.2 cos (Wt - 500) v = 2.9 cos (Wt+ 990). The depth integrated ellipses and their respective parameters are compared in figure 10. Although it is difficult uncertainties of the integration procedure, agree very well, to define the the two ellipses seem to further supporting the assertion that the barotropic 66 SEMIDIURNAL T COBOLT 10 cm/sec E -0.13 *.02 db= - 16 ± 40 CMICE 10 cmn/sec E= >= Figure 2-10 -0.15 E.03 -13**5* Semidiurnal ellipses for depth-averaged currents primary features of the mid-depth i.e., depth averaged ellipses; local shoreline sense. These regime and agree well and seem features that the tidal ellipses are preserved in the with also shows figure they are significantly inclined to the contours, depth mooring spar COBOLT the that The results. given accurate has system and resolved have been currents tidal be to the they rotate and a of part in nearshore tidal of Patchen, Long, the MESA measurements a clockwise and Parker, 1976. The band structure of semidiurnal admittances G. The to of the structure be explored. This admittances is structure Certain conditions may alter this place, the height tidal at to such factors as distance vary Secondly, considerably; 17:70:13 about energetic cm. the the to while be fairly amplitudes implies more 2, to the constituents example for uncertainties which do not enjoy due are passed along in N2 :2 : ratio closely or proximity the different of first fine differences Island Sound from Long yet smooth. In the undoubtedly probably contains amplitude This constituents, Hook, These differences the Hudson River estuary. admittances. expected has bands statement slightly. Sandy related to that at Shinnecock, frequency across in the high signal the is less to noise ratio of the M 2 tide. The admittances were obtained the admittance by and depth and then phases for the three integrating the real averaging the values most and energetic imag.nary for buoys bands parts 2-4. of The results, including ellipse parameters, are entered in table 4 and displayed in figure 11. Li Both increased and v admittance frequency. The amplitudes phases of both decrease as the frequency increases. viewed with errors. As of slight far as frequency decreases skepticism the ellipse changes are tend These results, parameters the are visible dramatically with increased decrease components considering most to in also the frequency. tend however, to must be magnitude concerned with of the the effects ellipticity which The orientation, by contrast, is totally unaffected. H. The results of the diurnal analysis Analysis of the semdiurnal band is much simplified by reason of its great about energy 200:1. content. Furthermore, variance observed. of the The total By M2 it contains contrast, variance and signal:noise almost ratio 30% is of in the fact total the diurnal band has only about its principal resolved component, 5% K - P1, has a signal:noise ratio of only 4:1 in the COBOLT experiment. As a consequence, more uncertainty the diurnal associated with over the four frequency bands table 5. It is the semidiurnal fact, coherences confidence) for admittances are liable apparent them. the coherences analysis, especially in are not significantly most of These admittances, from 22.50 hr to 27.69 hr, that the to have much are shown in are much lower than for the onshore components. different from zero admittances averaged of buoy 4 In (with 95% and for 3 69 TABLE 2-4 BAND STRUCTURE OF SEMIDIURNAL ADMITTANCES ADMITTANCE PERIOD 12.86 hr 12.42 12.00 COMP. AMPLITUDE PHASE ELLIPT. U 0.137 .007 51 3 V 0.051 .007 -96 7 U 0.132 .005 51 2 V 0.041 .003 -102 5 U 0.131 .009 V 0.045 .007 48 -113 4 9 -0.19 ORIENT. -18 -. 13 -16 -. 10 -18 0 ADMITTANCE AMPLITUDE I I PHASE I I I I .161- -A60* .14 ~-440* U .121- 200 10- .08 .06 -80* .04i -100* V 021- -120* I I I N2 M2 S 2 Figure 2-11 I II I N2 M2 S 2 I Band structure of semidiurnal admittances TABLE 2-5 ADMITTANCE AMPLITUDES AND PHASES FOR THE DIURNAL TIDE INST. 21N 21E 22N 22E 23N 23E 24N 24E AMPL. 0.16 .31 .09 .35 .11 .37 .06 .23 32N 32E 33N 33E 34N 34E 0.19 .39 .07 .34 .10 .24 41N 41E 42N 42E 43N 43E 44N 44E 0.30 .48 .20 .38 .20 .43 .17 .33 51N 51E 53N 53E 54N 54E 0.17 .48 .29 .44 .15 .17 ERROR 0.09 .10 .06 .13 , .35 .05 PHASE -75 42 -74 51 -122 26 59 .10 55 0.10 .07 -68 44 -53 34 24 60 .07 .03 .09 ERROR 28 18 35 20 43 37 23 27 11 12 16 21 -144 .08 .10 .09 .19 .26 0.08 .19 .12 .26 .07 .10 1 122 25 -2 43 26 76 -25 47 -106 43 134 19 10 15 24 24 38 25 21 23 31 25 31 Quantities listed under ERROR are the 95% confidence limits of the admitance amplitude and phase. COP is the true coherence of tidal currents with the reference series Starred coherences indicate that these quantities are not significantly different from zero at the 95% confidence level COH. 0.67 .83 .58* .80 .10* .46* .54* .50* 0.65 .92 .10* .90 .84 .75 0.26* .97 .00* .93 .84 .84 .32* .66* 0.69 .75 .72 .59 .69 .59 components number on of buoys 2 degrees and of The 3. freedom, CMICE shows mooring, with significant, a greater albeit low, coherence at all levels. Because of the large considered separately structure of but band-averaged diurnal only as the admittances resolution) semidiurnal but uncertainties, useful and case. is components lumped diurnal consequently information depth-averaged is This averaging, it appears, admittances since it is the is be The (and it the from the in the for the with available admittances, not admittances. lost still will just as essential only way to achieve significant coherences. The figure surface vertical 12. of the Like the semidiurnal and furthermore, at all structure increases with diurnal ellipticity depth. This makes the diurnal the semidiurnal ellipses at all levels. mooring but The never become positive and in depths. follow abruptly the COBOLT ellipticities go offscale at ellipticity the it is is shown negative COBOLT at in the ellipticities, fact remain less than -0.2 ellipses more circular Ellipticities at the CHICE at surface and middle bottom. This is than depths undoubtedly a spurious result. COBOLT orientation angles (figure 13) progress almost from large negative values at the surface to a positive bottom. roughly By contrast, the the opposite slope. in noise or CICE mooring is linearly angle at the non-monotonic and has These variations appear to be submerged influenced by non-barotropic effects. indicate the consequences of the low coherence. Large error bars ELLIPTICITY 0 -0.2 -0.4 0 1 1 1 0.2 1 0.4 1 % -95 1-9 * COBOLT May 77 o CMICE Feb 76 I0E a20 30 Figure 2-12 Vertical profile of diurnal ellipticity 74 ORIENTATION -30 -20 -10 0 (0) 10 20 30 95% 10 0 COBOLT May77 0 CMICE Feb76. E LL 20 30 Figure 2-13 Vertical profile of diurnal orientation The vertical distribution constant values except decreases very sharply near at of the energy bottom. the bottom (figure The 14) CMICE instrument, shows mooring again almost energy calling the result into question. In the spite of major differences depth-integrated similar. Compared to ellipses the in ellipse parameter distributions, (figure 15) semidiurnal appear ellipse, to be the diurnal quite ellipse rotates in the same direction but is oriented at less of an angle to the shoreline and is slightly-more circular. The velocity equations, with phases relative to Sandy Hook high tide, are: COBOLT (19) u = 3.5 cos (Wt- y = 1.3 cos (wt + 800) 400) and (20) CMICE T. = 3.7 cos (wt - v = 1.1 cos (Wt + 900) 420) Consequences and conclusions It should be apparent coastal by now that in the COBOLT data. the ellipticity (including non-zero the main features of previous tidal current observations in the Middle Atlantic also evident are U orientation angle. Bight are The most prominent characteristics sense of rotation) The vertical and the structure small, of but these 0 ENERGY 4 2 0 I-*--*---1----**-*---------1*--------------1----* (cm2 /seC 6 I I 2 ) 10 8 I I I 95% 10 - E LU 20 30 I I Figure 2-14 I I I 0 COBOLT May 77 o CMICE Feb 76 I I I I Vertical profile of diurnal kinetic energy - DIURNAL COBOLT - 5 cm/sec (Z:E E= < = -0.27 *.08 - 5* + 6* CMICE 5 cm/sec -0.20* d>= -Il* * E= Figure 2-15 .l1 80 Diurnal ellipses for depth-averaged currents parameters is another feature that agrees qualitatively with other experiments. A curious, presence of and small possibly scale, loose-grain sediments. that, on average, with the is features it difficult further Swift, topographic Duane, 220 ±160 of conceivable these is wave-like observation, and McKinney that and may account to argue evidence for in the Middle tidal currents the inclination the formed (1973) Atlantic are have noted Bight. It responsible for And, although that this swale topography angle tidal ellipses, of by form an acute angle for their persistence. persuasively concerns features the "crests" of these features shoreline certainly related is the possibility is rather intriguing. These features seem, or surface Bight. The tides of depth structure the energy profile) nature of then, the the to be characteristic nearshore region of the ellipse tide itself, to relatively unstratified insignificant have levels. conditions reduced Also, of Middle parameters Finally, depth integration the the February support the conclusion that the barotropic baroclinic "noise" the Atlantic (particularly suggests that the time and buoy averaging, velocities resolved. in of the barotropic baroclinic comparison is good or the tidal with the enough to tidal components have been has certainly reduced remained and has exposed the true barotropic what tidal currents. This tant is not or do not to say that influence the internal these tidal currents are unimpor- calculations. This energy primarily affects the barotropic admittances in manageable introducing low coherences into the measurements. particularly true diurnal for tide--both baroclinic the of tides. ways; Further show evidence discussion of of this by This appears to be on-shore velocity components which namely and for interference the from matter, however, is deferred until chapter 4. The vertical structure of the ellipse parameters point to the importance of friction. occur the in parameters are smooth seems also to The depth variations that do (at least for tides) and are accentuated near the bottom where should be strongest. This matter will also be the semidiurnal frictional effects explored in more detail (chapter 3). In terms of importance to continental layer dynamics, the most interesting shelf and coastal boundary and significant observation is that the depth-averaged tidal ellipse has a marked inclination to the shoreline. of both within Such an inclination is energy and momentum. 12 km of shore, vanishingly small these in order indicative Since fluxes to conform all of shoreward the COBOLT are normally transport moorings considered are to be to zero flux boundary condi- This appears not to be the case, however. tions at the shore. The energy equation for long waves is obtained from the Laplace tidal equations (see chapter 3) by forming the vector dot product bv with the momentum equation, and adding it the continuity equation. 2t (V to the product of g and This gives the energy conservation equation, V + gC ) + g V hv 0, (21) or aE+ where E is = - + 0 , (22) the kinetic plus potential energy of the water column, and F is the energy flux. Averaging over a wave period, indicated by brackets, < >, gives the average tidal energy flux, F = gh <cv> . (23) Before forming this product for the COBOLT data, the reference tide amplitude and phase must be shifted in some manner from Sandy Hook to Shinnecock Inlet. This shift to local tide is the tenuous phase and measurements in hour in is rather the Shinnecock area. estimating directions due the phase for the energy flux. can to the lack primarily in of Errors of as little result in tide as half an radically The calculation, gauge however, different is very revealing even if some errors are present. Analysis by Swanson (1976) suggests that the tide offshore of Shinnecock Inlet precedes that at Sandy Hook by about 1.0 ± 0.1 (error inferred from Swanson), about the 50 cm. semidiurnal available This amounts tide. to this adjustment, and that the M2 tidal amplitude is a phase correction Unfortunately, for the diurnal tide. hr no such of 300± 30 for information is Despite the obvious shortcomings of these figures are used in obtaining a rough estimate of the tidal energy flux for the semidiurnal tide. Noting the 30 phase correction in the free surface equation, (17) can be used with equation (23) to give the energy flux, = Flux onshore 800 150 watts/m (24) = Flux alongshore The energy alongshore flux has component 1000 300 watts/m a significant to the east onshore (towards component Long Island and an Sound). Furthermore, the magnitude of the energy flux is quite large compared with shelf-wide estimates such as those of Miller (1966). Miller, using a frictional dissipation equation due to Taylor (1919), CdU = E (with Cd = 0.002 and typical 3 , (25) tidal current speeds), found that the energy flux on the eastern coast of the United States averaged less that 250 scale. watts/m Though and was comparisons argument and the direct COBOLT relatively calculations between local seem to unimportant this on a world-wide shelf-wide dissipation flux calculation are difficult, the indicate that Miller's values are an underestimate. A more unusual fact is distance from the shore. Taylor's is supposed, that the flux is so high at such a short If a bottom friction mechanism similar to this rate requires tidal current amplitudes of 30 cm/sec shoreward of the COBOLT moorings--about than those observed. three times higher Another candidate for this dissipation, Shinnecock Bay, has only the narrow (200 m) and shallow (5 m) inlet to admit energy. Even a gross overestimate of energy entering the bay, made by assuming that all the energy of the incoming tide is dissipated, results in an energy flux of only 10 watts/m at the COBOLT site. It is possible that the flux can be accounted for by considering the divergence of alongshore energy flux. This supposition requires that the alongshore flux increase towards the entrance to Long Island Sound Sound. by roughly While divergence of 100 there watts/m is no for direct this magnitude, the every kilometer evidence that dissipation closer might rates to the dispute a in the Sound would have to be 5-10 times greater than are expected in order to accomodate the divergence. the alongshore flux is As suggested, of Long Island Sound. The large tidal probably due to the presence currents of the imply relatively large dissipation rates. Sound also This flux does not mean that tidal currents at the COBOLT site are dominated by Long Island Sound tidal flow. Atlantic Bight, Co-oscillating tides, generate such as those of the Middle substantial velocities but transport little energy because of their standing wave characteristics. The uncertainties of evaluation the flux calculations are absent of Reynold's stress terms due to tidal velocities. in the These depend only on the time averaged product of the velocity equations, (17) and (18). For the results in tide, semidiurnal difference phase of 1500 about a momentum flux of <uv> = This is a onshore transport -13 2 cm /sec 2 (26) . of westward momentum comparison with other forces if its divergence and is is very small uniform across in the 10 km coastal region. The momentum flux does not show any divergence across within buoys 2-4 (to 6 km of shore), however, and may yet prove of significance very close to shore. The diurnal momentum flux is the semidiurnal negligible, and Mann flux it makes (1978) and an who has an order the interesting found large components on the Scotian shelf. ments, the Scotian shelf of magnitude same sign. In semidiurnal Although comparison with momentum fluxes contrast smaller totally Smith, in than Petrie, both tidal to the COBOLT measure- and diurnal fluxes had opposite signs. To analysis summarize: of the COBOLT performance conventional experiment boundary with of the a condition COBOLT continental shelf. From be a physical questions applied exceptions, considerable instrumentation important to well-defined data promotes techniques. raises few at in the confidence comparison standpoint, in with the concerning the the boundary inner tidal the more COBOLT proper of flux the CHAPTER III TIDAL DYNAMICS AND THEORY A. Tidal dynamics Theoretical ment of the Although interest in tidal the the tides dates back to Newton's develop- potential potential which did appeared explain the in Principia origin of in 1686. tide-producing forces, the real beginnings of the dynamic theory of the tides can be traced to Laplace; appeared in done that is particular So 1799. it in to the subject difficult to is his old, Mechanique Celeste and work has find a problem enough that has which been not been ap- proached in some manner before (see e.g., Ferrell's (1874) discussion of non-linear bottom stress tidal theory is (1960), (1970) while and summarized in more recent Hendershott and tidal friction). Lamb (1932), reviews, (1973), such emphasize Much of the early Proudman (1953), as Munk the and areas or Defant Hendershott that are of concern to modern investigators. A dynamic theory of the tides begins by considering equations of motion for a fluid in the Eulerian a rotating frame of reference. their most general form they are (Krauss, 1973): ( + (V - V)v + 2Q x v) -Vp- VC} + V (1) -; +V- v =0 , where the symbols are defined as follows: In 85 p is the density of the fluid, v = (u,v,w) is the Eulerian velocity, is the rotation vector of the observer's frame of reference, is the fluid pressure, p is the gravitational potential, is the stress tensor, and the other quantities have well known meanings. As traditionally are and earth. these stand, they The equations, usual equations for simplified lead fluid is not the to have been critically examined by Miles used here with one exception--the to application which approximations, to difficult too much are tides Laplace (1974) solve on the tidal and will be considered friction- less. Basically, important the approximations and idealizations employed, and the modifications required of (1) are: 1. a homogeneous, incompressible fluid: -t + v ' p= 0; (2) 2. small disturbances relative to uniform rotation: (V 3. V)v = 0; (3) a uniform gravitational field (which implies the neglect of tidal self-attraction); 4. a rigid ocean bottom; and 5. a shallow, or hydrostatic, ocean: p g . known as (4) 9z This tion" last simplification, (Eckart, 1960), involves not the "traditional only the neglect approximaof accelerations but also the neglect of the vertical Coriolis to the horizontal velocity. vertical force due The omission of the latter term (and the approximation itself) has come under some attack (Phillips, 1966) no instance use would be the nature of specific has come to light where its but misleading. In tidal addition to dynamics on simplifications. 6. the simplifications the continental allows certain other a plane earth coordinate system: f k; (5) retention of stresses on horizontal planes only: V - T where 8. shelf above, They are: 2 Q 7. listed T= =3T/z (6) X i + Yj; and, the omission of direct tidal forcing: V+ = 0 . (7) The plane modelling after earth, coordinate a of the f-plane approximation the dynamics of small scale oceanic noting as or that the Coriolis function of the Taylor earth, latitude, G. expansion of f commonly phenomena. parameter system on -a spherical is of f = a a 21Q[ sinG, given It local The ratio of the around used is for made Cartesian varies first slowly two terms latitude forms the criterion for applicability of equation (5), ta t an 6where AG is a latitude <<1, increment and RAG is problem (R = 6000 km being the radius dimension of 100 (8) km, for example, the length of the earth). gives scale of the A typical shelf (8) a value of about 2 x 10-2 for mid-latitudes. According equation (1), are "stresses" known stresses (except possibly in the frictional familiar Reynold's forces in the oceanic case, In of stress gradients. consequence turbulent to to dominate a fluid to viscous very thin layers near boundaries). decomposition (Bowden, these 1962) a due to forces due forces are With turbulent forces are (in the x direction) (V the bracket where velocity and . T) - i a + = indicates perturbations. vertical length right-hand side of (9), a time average Introducing scale, is H, (9) , a the and the horizontal ratio of indicate primes length the scale, terms on L, the 88 H < UVI> (10) < u'w'> L The validity of assumption For tidal waves 104 ;while on the . 7 requires 11 = shelf, measurements direct and <u'w'> = 1 cm 2 /sec 2 . this ratio be very small. that 100m and L = 1000 that <u'v'> show km, so cm2 /sec 2 1O = I/L = Thus, this criterion is met. Finally, tidal phenomena on the shelf are generally assumed to be independent 1960). of the direct of the tidal potential (Defant, They are instead generated by the inertia of the deep ocean, acting through continuity, tides forcing are then termed propagating waves. at the edge of the continental "co-oscillating" and are treated shelf. as The freely The condition for the validity of this approxima- tion is that - gVC where V << , (11) VC n is known as the equilibrium tide and is given by Lamb (1932) as 3 M where M/E is earth, and the ratio of R/D the is distance to the moon. 1 a k _ - R 3 (E) 2ED = I the mass ratio 1 3 R (cos 3 of - (12) 3), of the moon the radius of to the the mass earth of the to the The gradient is typically 3M R3 -8 sin 20 = 5 x 10 (13) 89 of this A comparison latitudes. for middle quantity with the order of magnitude of the surface gradient made from estimates of the tidal (VC wavelength 10- ) the that indicates ratio is (11) small-- about 0.05. With the using the free relation to replace the pressure with equation (1) and above, described changes surface to the x and y momentum equations, reduces hydrostatic the function, C, and continu- ity: Bu at - -g fv + fu = 3x - au av 3x ay p az 1y p 3y at B. + (14) 3z aw + z The vertical structure of tidal currents To examine the effects of tidal friction on the vertical currents, the stress must be related to manner to close the set of equations (14). = where K is Kv structure of the currents in some The form chosen, (15) , a constant eddy viscosity, models a well-mixed water column, away from boundaries. turbulent effects in It be is treated (14). apparent that the free surface, as a forcing This results prescribed in a "local" on the horizontal of z, can velocities in calculation of current structure for tidal height variations. by i and adding the two, (15)) function being independent Multiplying leads (including the second equation the stress parameterization to the single, complex, second-order equation, + i fq - K ~2 = P , (16) where q = u + i v and P = g( ) + i i ((w+f)A+eit- (w-f)A e is the arbitrary forcing function expressed in ) terms that will relate it easily to the tidal ellipse (see chapter 2). The solution to this equation can be expressed as the sum of a clockwise and a counterclockwise rotating solution, q(z) = qj(z) e Wt + q 2 (z) e-it (1) where q(z) = C exp((1+i)z/d ) + C2 exp(-(1+i)z/d) + A+ for all W, and (18) q2 (z) = C 3 exp((1±i)z/d 2 ) + C exp(-(1±i)z/d 2 ) + A (using the plus sign for W < f and the minus sign for o > f) integration constants Cn, and defining the parameters d = K/(w+f) d /-fl with the This problem has been treated before by Sverdrup recently by Butman lies in took to be non-slippery. does not model account the for The difference (1975). choice of the a between (1926) and more the two analyses Sverdrup bottom boundary condition, which From a physical standpoint, this condition turbulent boundary layer correctly. the existence of the so-called "wall In layer", order to Butman used instead a slippery boundary condition, K where r is = (19) at z = 0, r q an adjustable drag coefficient. Both investigations used the same no-stress surface boundary condition 3q/az = 0 at z = H. (20) With these boundary conditions, equation (18) becomes r d q (z) = A+ eiot (1 ~ K1 cosh(1+i)(z-H) v'1 (21) q2 (z) = A~ e-i i( - r d K 2 sinh(1±i)( -), vQ2 2 where Q for n = 1, = (1±i) sinh(1±i)H/d + (rd /K ) cosh(1±i)H/d 2 and where the minus sign is used for W > f and the plus sign for o < f. The analogy is to seen clearly in the ellipse equations equation (equation (17) and (21). (11) Indeed, of chapter 2) for the case of no bottom equation stress, and (17) shows reduces that the exactly forcing to the parameters, complex A+ ellipse and A~, can be identified with the frictionless, barotropic tidal current ellipse. Because the description, forcing the actual parameters numerical orientation are also arbitrary. are arbitrary values of independent parameters vertical structure frictional forcing of the ellipse constants, "local" and ellipse However, it has been shown by Butman tidal ellipticity the the ellipticity that the vertical structure of of in A+ parameters and A-. depends a fact which was r and Kv; and orientation is Thus solely the on the used by Butman to make estimates of these constants. The vertical profiles of ellipticity and orientation are shown in figures 1 and 2 for values of the dimensionless quantities, Y = r/fH and (22) 1 that gave good visual fits / 2K to the observed semidiurnal profiles of the same ellipse parameters (presented in chapter 2). These of chapter figures were made to correspond roughly to figures 2 by adjusting A to match the observed thickness tional influence to the total ( A depth of is actually the the water ratio column) 7 and 8 of fric- of the Ekman layer depth and then varying Y to match ELLIPTICITY -0.2 -0.4 0.2 0 0.2- 0.4- ~ z/H 0.6- 0.8- 0. 2 y =0.5 A =0.2 1.0 Figure 3-1 Theoretical vertical profiles of semidiurial ellipticity ORIENTATION -100 -23* 0* 0.2- 0.4- _ z/H =0.2 =O.5 0.6 0.8 A=0.2 1.0 Figure 3-2 Theoretical vertical profiles of semidiurnal orientation 95 the range of values assumed by the ellipticity or orientation. was found that large values of Y, which correspond (in approaches result infinity) to Sverdrup's The actual the limit as Y no-slip bottom boundary condition, in excessively large ranges ellipticity. It observed for the range ellipse of this orientation parameter and supports the use of the slippery boundary condition (19). The model reproduces the main features of the observations well considering constant values near its the (currents simplicity. surface The and then changes can even rotate in the middle of the water Thus, it rather for near is to higher the bottom. slope towards negative angles as column, but diminishes shows the effects example, sharply counterclockwise) The orientation has an almost linear bottom. ellipticity, fairly sharply near the of friction a little further away from the boundary than the ellipticity does. Since the values of Y and A were vertical structure of semidiurnal chosen to fit the observed ellipticities and orientations, and since the actual values of these numbers were determined by judicious choice of surprising the that forcing the parameters, agreement A+ between good. An independent test may be made, values of y and A to the diurnal tidal and A~, theory however, ellipse it may not and be observation too is by applying the satme vertical structure. Figures 3 and 4 are these predicted vertical profiles. The except layer is diurnal that it ellipticity has deeper for profile is much a slightly more gradual frequencies that like the slope since are closer semidiurnal, the boundary to inertial (see the ELLIPTICITY -0.4 -0.2 0 A =0. 2 0.2- 0.4- z/H 0.6- 0.8 -=0.2 1.0 Figure 3-3 Theoretical vertical profiles of diurnal ellipticity ORIENTATION - 20C 0 -10*0 10* 200 A 0.2 0.4- ~ z/H 0.6- 0.8y=0.2 1.0 Figure 3-4 Theoretical vertical profiles of diurnal orientation for expressions tation, on the d following n other hand, f--it increase.s slope is 12-13 in with chapter is is rather found 2). The (18)). The totally different depth exactly what equation in than the model diurnal orien- at frequencies decreases. below This opposite COBOLT observations (figures also shows a greater range of orientation angles--again a feature that is found in the data. Given the viscosity model observations, provide independent it and is acceptable the agreement COBOLT COBOLT the best site, to suppose values the and agreement semidiurnal reasonable of Using the definitions of equation the between the range and that friction (22), (A = constant diurnal parameters, K depth of for y and A that 0.20± .05 eddy frequency this model will the nominal of values with observations the and y = also and r. 30 m at produced 0.4± .1) gives K v = 10-25 cm 2/sec (23) r These (see values e.g., estimate agree Butman, made by = 0.09-0.014 cm/sec well with 1975). Scott low frequency currents. other The estimates value and Csanady for (1976) r at made in also shallow water agrees the COBOLT with an site using C. The effects of friction on tidal propagation Besides affecting friction currents, istics of the the also local vertical affects tides (this the "global" structure global surface subject has investigators important in In space have time, conditions. ignored friction It resistance more apparent. also field order in the of attention the character- of equation (16)). basic in While past, the conditions on as the shortens the expected nature the water by most that are of attenuating waves the wavelength column has less the It largely ignored fact, is this effects shallow water, to explore simplest terms, of bottom that is the global or initial 1955). inertia In and hence, friction are much the enhancement vertically of the tidal at fv = -g + fu = -g 3x ay +5 + integrated versions pH wh e re H is the depth of the fluid; B is the bottom stress; and (24) PH - + -(Hu) + -c(Hv) at 3x 9y of elevation equations (14) will be used: at in investigated here. structure the boundary (Proudman, to change, frictional effects in In some acts depending shallow water, where less considerable propagation tidal the coastal region; namely rotation and depth variations. general, or received the tide was impressed on the previous solution as the arbitrary forcing this of 0 of 100 u and v are the vertically averaged velocities defined by H v(x,y,t) = 1/H v(x,y,z,t) dz 0 In direct analogy to equation (19), bottom friction will be taken as proportional to the depth-averaged velocity, i.e., B Although this introduce a form has Physically, should used be some dissipative dynamics. in -r v . (25) shortcomings mechanism, velocities But (25). albeit above for (Rooth, a the tidal crude 1972), it one, into frictional oscillations, does the wall layer the depth- integrated velocities are a good approximation. The equations can be solved quite simply if be constant. The elimination of all the depth is variables except taken to the surface elevation leads to the equation, (( at + r)2 + f 2 )_-_ H at gH( 3t + r H =2C 0. (26) Rearranging this into frictional and non-frictional expressions a((a + f2) at t 2 )gv) 22 gH2) gives + (27) + r (2 II Thus the equation 2 at2 + r G H at separates gHV 2C) neatly For small bottom friction or large depth, (upper part) governs the dynamics of 0. into two well-known forms. the familiar wave equation tidal waves, whil- for large 101 friction or small .equation (lower solutions. depth, the dynamics part) which is fact for In are governed by the telegraph known to have wave-like the large friction-shallow and diffusive water case, (27) reduces approximately to the parabolic partial differential equation, 2 2 gi_ r Dt The is scaling factor This r/wil. 0.2-0.4) but that factor rapidly is r/wH This point, marks the reaches which is 0 indicates small becomes decreasing water depth. structure, = at which dynamics are the COBOLT mooring important Using r = 0.1, a (28) value of nearshore appropriate site because (about of the as inferred from the vertical unity in about about 1 km from the shore at outer boundary of frictional dominance 10 m of water. the COBOLT site, in depth dependent dynamics. D. The Sverdrup-Poincare wave model Analytical solutions of (24) are extremely complicated if both rotation and variable depth are retained in the model. To simplify, one must two or to model either make a choice between the try the frictional effects of shallow water in some other manner. The obvious choice for an alternate model is a step discontinuity to simulate shoaling water. coordinate system envisioned, dynamics the that deeper dominate, and The will of the be the geometry of used two shallower is sides a solution dominates\ (t'ough this region is this situation and the shown in is region region a figure where not examined in 5. As where wave the diffusive detail). 102 Figure 3-5 Geometry of the step shelf tidal model 103 In the deep region, the familiar this wave refraction problem is posed in problem with the incident, reflected, and transmitted waves. the same manner solutions formed determine of the (29) H'v' the relationships between the various components. first condition, Kelvin direction (alongshore), must they will decay shallow in the from The boundary conditions, at y = 0 , Hv as waves, be which excluded region travel from but not only Because in consideration in the deep, the x since and a mismatch at the boundary will inevitably result. The governing equation in the deep region is (from (24)) 2 ( + f2)C - gH V2C 0 , (30) at with the velocity fields defined by 2 (a2 2 + f2)u =_g( + f a) y tx at22 (31) 22_2 (D at 2 2 + f2)v =-g( at3y ax Assuming a solution consisting of an incident and a reflected wave, c where k = a (exp i(kx+ty-wt) + R' exp i(kx-ty-wt) , and k are equation (30), real wave numbers, gives the dispersion and substituting (32) it relation for the elementary into long 104 gravity wave in a reference--the rotating Sverdrup (Sverdrup, wave 1926): 2 _ The friction solution is in assumed across the step. It g = shallow to (k water absorb must, (33) + 22) is relatively all the however, have energy ' where 2' can be = T' complex exp i(kx+9'y-wt) to allow for that the same the x direction as the deep water solution. unimportant is since transmitted functional form in It is taken to be , (34) frictional attenuation in the offshore direction. Applying the boundary conditions, with the aid of equation (31) prescribes the relationships: R' + 1 =T' (35) gi 22 (Wl(R'-1)+ifk(R'+1)) = igH' T' ((r/H-i22 (fk-(r/H--io)) which can be solved for the complex amplitude of the reflected wave, R' H -1 IQ -- I'Q' j~ HQ* + H'Q'( where Q Q' Eqiati on (36) oz - = ifk wZ' - ifk can be reformulated into the more conveni rnt form, (36) 105 (Q/Q*) r' R'= where r' will be called (37) , the "reflection coefficient" and is defined by In the present (38) 1 - (H'Q'/HQ) 1 + (H'Q'/HQ*) r context H'/H is presumed to be small in keep the absorption of energy to small values. In this order to case r' can be approximated to order (H'/H)2 by r' r = 1 -1 H' Real (Q) Q, .(39) In r' general is complex, but for small H'/H it can, for all practical purposes, be considered real since the imaginary part of r' is very small in real since both For the case of no rotation, this instance. Q and Q' become real. Also, r' is less r' is than unity for H' < H, indicating an absorbing boundary. Replacing R' in equation (32) C (r'Q e '+ with equation (37) gives the wave solution, Q* e ) ei(kx-tt) (40) and the velocity distributions, V ag Q12 1 , e-ity ity) i(kx-ot) o -_f ~ Q* (41) u = ag 2 o- 1 (r'SQ e 2 Qe f Q* - S*Q* e ) ei(kx-ot) 106 where S These solutions whose elementary ift. - consist of incident and reflected are characteristics well-known invoked as an explanation of tidal phenomena. Sverdrup waves but are seldom The ellipse character- istics are easily examined from equations (40) and (41). This boundary condition bears a close relation to one proposed by Proudman (1941) continental in shelf an attempt to model the dissipative effects of the as a boundary to the deep ocean. The Proudman boundary condition, V c at y = 0 , takes the place of the boundary conditions some of the numerical models It 1977). was solutions California with included and has been used in deep ocean tides (Hendershott, also used by Tendershott and Speranza (1971) strongly localized These of the (29) (42) coastal dissipation at the end of a long channel. were applied some success. implicitly in to model to the Adriatic Finally, the (1978) model Redfield's Sea and absorbing the Gulf boundary of Long Island of is Sound and several other basins. To show the analogy, the solution (32) is used in with (42) as a boundary condition to give - Q + a(2_ f2) 2 2 gQ gQ* - a(o f 0*3' ) where 22 1 + a(wo ~f )/gQ 1 - _ ?) , l equation (31) 107 is a clear analogy to equations (37) r II + + 2 2 = and (38). - ) Real _(~-f g Thus a must be complex to agree If a/g is small, ( (45) . jQ[ completely with equation (38) and also must be negative to assure that r' < 1. E. The Sverdrup wave--no reflection The plane Sverdrup wave is 0, which is the case of investigated by setting r' no reflection or perfect = 0 and k absorption. = For this wave, the ellipticity is defined by C = -i (v/u) = -(f/o). For middle latitudes, of about -2/3--i.e., minor axis:major (46) the Sverdrup tidal ellipse has an ellipticity a clockwise axis ratio of rotation 2:3. of the current The orientation vector and angle of Sverdrup tidal ellipse coincides with the direction of propagation the so in this instance is identically zero. F. The Poincare wave solution--perfect reflection Examining the opposite extreme, r' = 1, shows the characteristic the case of perfect reflection or tidal ellipses of two superimposed Sverdrup waves whose onshore velocities exactly cancel one another at the shore. 1971). result wave This combination is Often in invoked as an eigenvalue numbers (Defant, a known as the Poincare wave (Platzman, solution in a channel, where problem and a set of discrete 1960), this solution is boundaries cross-channel also valid for long, 108 straight coasts and Wimbush, or 1970) continental topographies (Munk, Snodgrass, where only one boundary plays an important The resulting the dynamics. shelf free surface role in and velocity distributions are: (o, cos ky + fk sin ky) ei(kx-wtl = -i2ag 2 v = u = 2 2 22 2 ((W -f) f ( 2 2ag 22 2 i(kx-wt) (k +P, )) sin ky e )Wf sin 9y) ei(kx-ot) cos Ly + (k2 +, 2)k (47) o -f ) As in the non-rotating, standing wave problem, surface nodes is apparent from the first of the Otherwise, the wave complicated than the non-rotating rotating standing case. the possibility of equations characteristics Turning again to in (47). are more the tidal ellipse to elucidate the signature of the Poincare wave, v _ u where Besides highly dependent 2 ? + f /g) sin ky kk cos ky + wf/gH sin Zy the dispersion sion. ( relation has been used to simplify the expres- depending on the distance offshore, on the wave numbers, or 48) ' the ellipticity is equivalently the angle By the angle of of incidence of the incoming incidence as the angle between the wave vector and the shoreline (see Sverdrup wave. defining figure 5), I the wave properties = tan~(- ), (49) can be plotted as a function of offshore distnace 1.0 0.8 r'= 0.6 0.4 I= 30* I= 60* 0.2 00 CL -J 0.2 0.4 0.6 0.'8 ky3 -0.2 -0.4 -0.6 -0.8 -1.0 Figure 3-6 Ellipticity versus offshore distance for perfect reflection and several positive incidence angles 1.0 0.8 0.6 0.4 0.2 0 -0.2 -0.6 -1.0 Figure 3-7 Ellipticity versus offshore distance for perfect reflection and several negative incidence angles 111 The negative values (the is ellipse distance offshore velocity becomes becomes circular. but equal the to this From the on, point and velocities the onshore the and velocity alongshore short or negative if .k > 0 away from shore, some distance At a positive becomes to the right) zero identically the ellipticity is polarized) alongshore than greater = 0) k < 0 (shore (shore to the left). and positive both assuming greatly, (y linearly if Note on the angle of incidence and the distance depending At the shore offshore. for scaled by the wavelength. these plots is varies ellipticity 6 and 7 figures and positive incidence angles. the direction of propagation) that the offshore distance in in done on the right-hand when looking representative negative (shore several in is This angles. incidence various for ellipse velocities onshore again tend ellipticities are toward smaller values. this at Also changes ellipse tion between is tation the point, abruptly u and v is constrained to always be for the same values to 00 from from shore and all incidence angle orientation 900 or 900 incidence the rela- the orien- all for of phase the Figure 8 shows angles. axis major equation (47)) (see 00 the Since 900. either of the of distances the orientation angle (both positive and negative) as the previous figures. It clear that is the COBOLT data differs substantially Sverdrup and Poincare wave models. ellipse does ellipse is rotate in . the much too circular While clockwise the Sverdrup direction, the from both wave shape current of to agree with COBOLT observations. the The I=60* 80*0 I=30 60 40 r= 20 00 0.4 0:2 1 -20 0:6 0:8 ky -40 -60 - - 80* 0 I -60* Figure 3-8 Orientation angle versus offshore distance for perfect reflection and several incidence angles 113 Poincare current observations, ellipse, is orientation unlike any oriented exactly alongshore tendencies close to the wave progressing with the shoreline wave model ellipses degenerate rotate counterclockwise So as of the they stand, ellipses in the and exhibits shore. Also, for on its right-hand COBOLT no onshore an incident side, Poincare into lines very near to the shore and further away, again unlike the observations. the Sverdrup and Poincare waves are not capable of reproducing the results of the COBOLT experiment individually. G. The combination Sverdrup-Poincare wave A combination of these waves offers a third alternative, the case of 0 pure Poincare and Sverdrup waves as references, 1 is less difficult to plus incident Sverdrup wave wave; or a equivalently, - consists amplitude reflected plus smaller wave a r' of an Sverdrup amplitude The solutions for this case are: incident Sverdrup wave. = - a smaller Poincare solution This interpret. and with a((r'+1)(oiZ cos Zy + fk sin Zy Q- - i(r'-1)(fk cos iy + wk sin ky)) e (50) u = -a ag Q*(w"-f - v = 2 _ 2 ((r'-1)((o -f2)k 2 2 cos Ly + (k +z. )Wf sin ky) - ) i(kx-)t) 2 2 2 2 i(r'-1)((k +9 )of cos Ly + (W -f )kZ sin ky)) e ~ a g (U2 2 2 + f2 2 2 Q*(2'-f 2) -((r'-1) cos Ly - . . i(kx-wt) i(r'+1) s tn Zy) e. 114 Tn natural cient will situations generally be it is close likely that the reflection to unity because the to reflect much of the energy impinging on it. true very all possible near to shore. Nevertheless, reflection coefficients it shore This is is the two Sverdrup diagrams are waves of representative Sverdrup waves of equal be interpreted as waves. of amplitude all to ellipse the range the range of possibilities Sverdrup waves ellipses that can occur under and may provide between parameter interactions the Poincare distance the of relection that exist two of two can then interacting coefficients for some insight into a wide as consider interactions The offshore difference entire (just possible amplitude). the phase Considering establish unequal certain particularly diagrams will then be representative of all the possible of is valuable since coeffi- two will interacting the forms of tidal range of reflection conditions and distances from shore. Plots of the for different the the sharp transitions the parameters real values of combination between tidal two waves the Poincare wave exhibit becomes smoother less the ellipse parameters reduce Sverdrup wave orientation with an figures 9 through 11 the reflection coefficient. noted in waves are shown in to of ellipse pairameters, profiles important. Tnstead In as interference the r' = 0 limit those of the uniformly srooth pure angle equal to the complement of the angle of incidence and an ellipticity equal to -f/w. The ellipticity, characteristics, tion coefficient as a result increasi-ngly drops. of favors Thus, the transition negative clockwise to Sverdrup wave values rotation as of th. the reflectidal 1.0 0.8 0.6 0.4 0.2 0 -0.2 -0.4 -0.6 -0.8 -1.0 Figure 3-9 Ellipticity versus offshore distance for imperfect reflection 1.0 0.8 0.6 0.4 0.2 0 -0.2 -0.4 -0.6 -0.8 -1.0 Figure 3-10 Ellipticity versus offshore distance for imperfect reflection 800 60 40 z 0 20 O0 z -20 -40 -60 -80* Figure 3-11 Orientation angle versus offshore distance for imperfect reflection 118 ellipses graph becomes over observing more which probable the (as ellipticity perfectly the extent negative). The probability of is (c = circular measured by 1) is quite small seen sharp that orientation angles jumps With these from 00 general to 900 trends (figure as (11)) were no longer seen established, in the the Also, it exhibit Poincare extrapolations the since ellipticity tends to range somewhere between zero and -f/w. is of can the wave. easily be made for cases not shown. Considering that to the tidal wavelength is shelf dimensions, parameters 9-11 may that, allows be most interesting found for ky << 1. the Poincare wave unlike both the non-zero generally large orientation It aspects is model, angles and of apparent the compared the ellipse from figures combination non-zero wave ellipticities near the shore (near y = 0)--features which reproduce the results of the COBOLT experiment. In particular, interest to plotting the angles as the a angle constant. axis, conditions COBOLT possible function coefficient. the the of experiment, of incidence the are combinations The various Since at y = the angle curves while diagram in of of which ellipticity incidence angles is limited to negative the most summarized and by orientation reflection 12 are formed by varying the symmetric and of the reflection about only negative orientation angles are shown. incidence are conveniently figure holding is 0, the 00 orientation Also, values, coefficient the range of -80o < I < -100 119 with the indicated by dots), are increments of incidence angle (100 exception of the asymptotes which are indicated by dashed lines. As alternative an r', constant a constant figure r' is examining can plot similar incidence 13 for while in ellipse ellipse characteristics by the made the from 0.6-1.0 varied of be varying holding reflection angle coefficient. for of In dots indicating reflection with By considering both of these figures, coefficient increments of 0.1. the variations the characteristics that occur due to changes reflection coefficient or angle of incidence should be evident. From figure that while angles the from 12 -900 regardless ficient. For higher angles since negative the ellipticities angles to +900, of the angle in result is plot can about cover 00) the ellipticities of incidence small is always furthermore, angles small, and is quite evident angles are coefficients that small alongshore wave numbers) figure difficult in incidence range negative coef- the bulk (positive negative or ellip- and small any real situation where It is also evident from angles (waves with larger orien- associated with small negative orientations also show variations are of 13 and would make determin- close to unity. negative 13 in apparent or reflection orientation symmetric) it complete are coefficients, reflection incidence of reflection figure reflection This convergence of lines to small orientations ticities. ation its orientation nearshore of incidence (and tation angles. The ellipticities and depend on the frequency of the incoring wave. that Figure 14 shows curves 120 ORIENTATION -80 -60 I I I=-8* -40 I I -20 00 I 1~; -0.2 I P T I=- 10 -0.4 -0.6 0.8 Figure 3-12 E L L Ellipticity versus orientation angle at y = 0; incidence angle varied while holding reflection coefficient constant C I T Y 121 ORIENTATION -80 -60 -40 -20 O0 -0.2 E L L P T C - 0 .4 c T Y -- 0.6 - Figure 3-13 Ellipticity versus orientation angle at y = 0; reflection coefficient varied while holding incidence angle constant -0.8 122 ORIENTATION -10 -20 -30 0 E L L P T --0.1 1 /I / /1 T Y I=-80* r'=0.9 -- 0.2 Figure 3-14 Ellipticity versus orientation angle at y = 0; varying the incident wave frequency 123 angle -700 < I < -20'. range, previous figures is from clear different waves as ellipse that 14 o/f = 1.5 was used in Although representative figure = 0.9 and for the incidence for r' frequencies different for several the mid-latitude of characteristics. particular, In higher it slightly frequency the (i.e., ellipticities less negative nearshore tend to have tide, have frequencies tidal other M2 the ellipticity approaches zero) than lower frequency waves. it Again, angles incidence be should on (shore around the this operation is 00 left angle orientation that of wave a positive looking if in the can be constructed simply by reflecting direction of propagation) plots the for diagrams similar that noted all The axis. angles orientation affect main become the of positive when the coast is on the left of an incoming wave. H. Comparison of the model to observations comparison A ellipticities and on components observations COBOLT by plotting accomplished is Poincare wave model tations the of 12. figure of three the 15 Figure the observed semidiurnal resolved shows an Sverdrup- the with orientidal corner enlarged of figure 12 with these observed values in place. The positions of the observations that of incidence an incident angle) fact predicted hand side oE by the wave since eminating observed on the plot are consistent with from the orientation deep angles ocean negative; are the modelling of waves with a coast on wave (looking in the direction of (negative a the right- propagation). 124 ORIENTATION -20 -30 -10 O0 -- 0.2 E L L P T -- 0.4 -- 0.6 Figure 3-15 Ellipticity versus orientation angle at y = 0 with COBOLT observations plotted for three semidiurnal frequencies I T Y 125 Furthermore, incidence data are the angles (since also consistent with predictions observed large reflection coefficients to zero). Finally, suggested tidal by the component, the (since data is the closest angles are small) observed ellipticities exhibit theory--i.e., S2, orientation for large the to are close dependence highest zero and on frequency frequency semidiurnal while ellipticity, the lowest frequency component, N 2 , is furthest away. The notion dently, to of some small degree, It chart (figure 2-2). incidence by angle with exists, the south examining shows dicular to cophase contours, shore due to the curvature angle can be Swanson's confirmed semidiurnal that the wave vector (which or wave crests) of Long is cophase perpen- does indeed form a small Island. of cophase indepen- lines, While some this angle ambiguity appears to be approximately -20O. Physical cient is considerations also a reasonable suggest that result. By a large requiring onshore energy fluxes match observed energy fluxes, compute (chapter a reflection 2) and the coefficient. From Sverdrup-Poincare reflection the free that it energy surface coeffi- theoretical is po>sible to flux and equation onshore velocity distributions (equation (50)), the onshore flux is F = gH Real <Cv*> - (51) 2 2 _ag H _1 2 2 - r' 2 2 Substituting the following numerical values into (51): 126 and equating it to a = 50 cm H = 30 m o = 2 /24 k = 2 /1500 km, the energy hr flux estimates of chapter 2 (COBOLT measurements and Miller's shelf-wide estimate) gives = r' a value that supports 0.98 ± .01 , the large (52) reflection coefficient deduced from the tidal ellipse parameters. Since the flection ellipse coefficient determine the parameters and values angle of In could be used to find the values principle, the three because of the on an this obtained angle reflection is from Swanson's known it coefficient orientation. Doing is independent of r' this for re- is impossible at a single to fre- frequencies and T, but this procedure complexity of work offers a it unknowns, semidiurnal the equations is involved The estimate of incidence an a straight-forward from two observation and the uncertainties of the observations. angle the incidence, of both with quency. very difficult depend alternative since once task to determine single observation the M2 tidal of frequency ellipticity with the and I = -200 gives a reflection coefficient of = r' 0.95 , (53) which is close to the estimate derived in (52). Other observations Sverdrup-Poincare in wave model the too. Middle Atlantic The model Bight suggests support that the clocktwise 127 current should ellipses rotating tidal stances. Emery and Uchupi (1972) long gravity phases in the waves (without appropriate of the for most to implies modelling tidal ellipses left-hand and for side (New a wave Jersey) that the Sverdrup to a rotating system, tidal phenomena propagating looking in for a wave propagating with (Long Island). Both ellipse orientation in angle for the of with in cases a show to the nearshore measurements--a shore-to-the-left case and wave, is the the region. (figure 2-3) shoreline the direction a shoreline these fit elevations Finally, the results of Patchen, Long, and Parker (1976) show circum- this was true at 94% of rotation) long gravity wave solution in Redfield's (1958) empirical the Middle Atlantic Bight extention norm the found that the sites surveyed by Haight (1942). of be to the of propagation, the right-hand predicted side sense of positive orientation and a negative angle model is clearly an oversimpli- region, even for the shore-to-the-right case. The two-wave, fication of does Sverdrup-Poincare the complex tidal explain many of the observed does not account for geometries coasts. Corners, the coast from regime which scatter not features. other considered. two-wave model the wave propigation order of magriitude The model, than infinitely additional and the Small incoming wave energy And, the though it for example, long, straight such as that formed by the coast of Long Island and of New Jersey, them are of (towards predicts scale (Mysak, an energy the west at the reflections are also ignored. the direction of flux in COBOLT result irregularities, coastal 1978), that site) Larger than that which is observed. that is an 128 Kelvin waves have also been ignored, a serious flaw. The generally progress fluxes that are Kelvin wave, attenuated in semidiurnal tides this is not likely to of the Middle Atlantic Bight the onshore direction and have the small energy characteristic of standing waves. a possible solution (though offshore though and have a large not in By contrast, this model), alongshore energy the would be flux. These features are not seen in the observations. Although necessary to more complicated account for models these are additional possible features, and the may be comparison with observations is good enough to suggest that: 1. The classical Sverdrup wave tion for the semidiurnal it accounts for many is the fundamental mode of propaga- tides of the Middle Atlantic Bight since of the observed features of the tidal currents. 2. The coastal region absorbs a small of the Sverdrup wave, probably water 10 shallower than amount through This m. of the incident energy frictional absorbed dissipation energy, small fraction of the incident wave energy, is larger although in a than some previous studies have suggested. I. The effects of local topography It is of considerable Sverdrup-Poincare that is interest evident interest wave model in of to consider one particular the complicating the COBOLT region--topography. to consider depth variations effects to determine on the feature Not only is it of how far away from 129 the step model, shore can be correctly considered a small distance in but some to shed also In dissipation may occur. the on light other words, it the of where question is inferred hoped that this model how good is the vertical wall assumption, will answer the questions: and where does the large amount of energy observed to be propagating onshore dissipate? The long wave equations, with no bottom for a variety of bottom profiles. particularly well-suited tical expression. bottom The COBOLT region, however, has a profile to = H (1 - straight-line of fit 16. In the distance values of H linear fit of area. y is a simple analy- These the this outermost depth parameters (54) the best were constants ln(H -H) fit to the actual chosen from versus y shown in a figure only extended to 14 km--slightly greater than shown, H0 = for approximate the give on the graph of this plot, by b e~s ) , s, and b are chosen to profile approximate This function is H(y) where H0 , friction, can be solved in 35 m COBOLT mooring. appears to give b = and gives equation sulting computed profile with the 0.9 (54). actual Of the the best s = different fit. The 0.2 km~ as the of the re- A comparison depth profile is shown in figure 17. With omitted the in depth equations variations (24), retained elimination equation for the free surface, of and the bottom the velocities stresses gives the 130 100 50 E 10 0 X 0. 0 2 4 6 8 10 12 OFFSHORE (km) Figure 3-16 Plot of in(Ho- H) vers us offshore distance at the COBOLT site, plus linear curve fit 10- 20 ac -C 4.- bc 'I, 30- H(y)=H0 (I-e-sy IJ 10 0 5 10 km Figure 3-17 Comparison of modelled depth dependence with actual depth profile off Tiana Beach 15 132 Taking gHV 2 + f2 )% ( the depth to - be VH - a only - (V function of = 0. fk x V) the offshore (55) variable allows a solution, F(y) exp i(kx-wt) , C(x,y,t) which, when substituted into (55), results in (56) the second order, ordinary differential equation, 1 dH dF d2F +---+ ( W2f 2 g H dy dy dy2 fk dH 2 +-- - - -k)F w H dy . =0 (57) A change of the independent variable to z = b exp (-sy) (58) and substitution of the depth profile (54), transforms 2 d2F dF 2 z (l-z) 2 + z(1-2z) - + (a2 + Sz)F dz2 (57) 0 , to (59) dz where 2 22 _ 2 1 (W -f 2 - k f k k ( f S S Equation (59) is k2 g H o ) one of the many variations equation (Morse and Feshbach, 1953). z = 0, 1, dom.tin of +o (or at sy = +o, ln b, intere.-t provided b < 1. of the hypergeometric It has regular singularities -o); all of which lic, outside at the 133 by solutions, Stegun much more illuminating to equation solve the (59) Abramowitz and in tabulated to obscure as so general but are (1964) are functions, hypergeometric Its (1970). Wimbush and Snodgrass, Munk, by and (1967) Ball context before an oceanographic in This equation has been studied results. directly It using is the method of Frobenius. Substituting the infinite series, zp = F(z) c (60) zn, and equating lowest order terms gives into equation (57) the indicial equation, p2 This two independent higher (61) which are associated of the second order problem. solutions determines terms order 0. roots distinct two has equation = 2 the constants the Equating the for n > 0, cn, with by the equation c n = (p+n)(p+n-1) 2 (p+n) 2 to eliminate p, Or, with the use of (61) c n =c (62) 6)1 n-i (n(n-l) For large sy (i.e., z - n(n -+ - a22 ) 12a) . i (2n-1) n-1 0) the solution becomes (63) 134 F(y) lim (64) b e Iasy sy +0 a2 > 0. provided This solution which is flat the steep solution appropriate, portion of the be pursued by examining and a reflected apparent from topography) or that plane Sverdrup for large y (that large s this and (that is, over for a step model the just as before. sa = k is wave is, (using the plus sign in wave (minus sign) the for between an incident and (59) (63) simply apparently, The similarities slope). (64)) is equation It an appropriate can is also definition for the wave number in the offshore direction. Both of the previous studies have examined the case for b = 1 and emphasized consisting modes merely Wimbush studies previous topographic, of These are the and Kelvin, shore-trapped edge point examined that they shapes of out the exist. Snodgrass, Neither of the only their specific depth solutions; the Ball waves. ignores the trigonometric solutions while Munk, completely and where a2 < 0. solutions spectra. In form of the final profile, to interest of examining the consist of an incident the of this effects solutions the trigonmetric and reflected wave. is rearranged From equations (57) and (60): = where hofore, (A e c = ity 1. c + b The by demanding n -sny en B + B e absorbing that -ity X- c boundary the reflected b n - sny) en) condition e i(kx-ot) , is wave amplitude (65) applied be less as than 135 the incident wave amplitude for r' vanish = at the coast for r' of the reflected wave.). v 1 (This and that the onshore velocity condition determines the phase The onshore velocity is -sny-(A1 y = c bn(Q +iswn)e-sny 2_ 2 (Ae V < 1, (66) where Q is c bn(Q*-isn)e-sny) ei(kx-Wt) n (Be ity - defined in equation (36). This determines the coefficients, = A c* bn(Q*.-iswn) n r' r'B* (67) c = B bn(Q +iswn) Here it has been noted from equation (63) that c = c The final form of the solution closely parallels c. that of the step model (compare to equation (40) and (41)): v = - u ei(kx-wt) + r'F(y) eY) (F*(y) e g W -f (r'e'y (Q F(y) + isw G(y)) e (Q* (68) F*(y)- isw G*(y))) e i(kx-wt) (S* F(y) - 2(r'e'k = - sf G(y)) + W -f + e zy (S F*(y) + sf G*(y))) ei(kx-ot) where e-sny F(y) = B* c bn G(y) = B* c bn n e-sny 136 and S is defined calculated as in numerically, equation (41). but several These features functions are easily can be noted before the solution is presented. One of topography is the is to be answered an adequate accomplished terms in questions by is whether or not the step representation of the COBOLT recion. examining the ratio c /c since the This successive the series are a measure of the extent to which the solution deviates from a plane wave. This ratio, c 1C +2. - + a ) + ia (69) 1 + i2a c 0 is small if both a and 3 are small. These parameters are indeed small for the semidiurnal frequency waves since c k/s a a P/s are the ratios lengths. result to within the topographic length scale tidal the does not imply characteristics region characteristic of that very near topographic topographic the topography scale is to change. the Off tidal the the COBOLT moorings Far the from the coast, for tidal models. however, will be wave- shore, unimpor- particularly Tiana Beach, 1/s = 5 km so variations current meter records between wall to The ratio (69) is estimated to be no greater than 0.03. This tant of (70) shore are in the the a possibility. can be considered a vertical 137 For perfect reflection simplify consiAerably ence of a vertical since = I) the expressions all three are number and its term, exp i(kx-ot), dependendent quantities (r' and v is complex boundary model, the So, 900. Again, necessary to an energy provide just velocities, phase and ellipse orientation angles or means as C in and and u the v, the are x real flat bottom, 900 out-of- are are restricted to be either 00 absorbing boundary the differ- Disregarding that u equation (68) the suri or conjugate. this imaginary. either in orientation condition (r' needed to "tilt" < 1) fit is the observations. The absorbing shoaling water, boundary however, condition shows that angle this as a effect much more effective in since the incoming wave crests are refracted by the topography to parallel orientation is the shoreline. function is of Moreover, offshore a plot of the distance only evident very near (figure to the 18) shore--too close to be detected by the COBOLT moorings. A plot absorbing of the computed boundary topography ellipticity condition is also (figure needed 19) in the to bring the ellipticity to the magnitude observations. Again, the greatest variations shows occur that an presence of and sign of the closer to shore than could have been detected with the operational COBOLT moorings. These computations observed also in support crease suggest that a region very close the inter-mooring the statistical the strongest (within 2-3 km) averaging confidence in the used variations may be to the coast. in chapter tidal current 2 They to in- measurements 138 00 -10 -20 -30 -40* 5 10 y (km) Figure 3-18 Orientation angle versus offshore distance for the depth dependent model 139 0.4 0.2 0 -. 02 = 45* -0.4 SI I I I I i i 1 10 5 0 i y (km) Figure 3-19 Ellipticity versus offshore distance for the depth dependent model 140 the since depth-varying model suggests little variation in ellipse parameters beyond 5 km from shore. The question and whether it cients sy = the inferred great enough can be answered previously, at is of where in to account the reflection (59). this equation has a singularity at z = 0, In kept for part by examining equation b, which lies very close exactly at the shoreline for b = 1. are tidal dissipation might occur finite constants, at A and B, this energy be reflected case of perfect (the reflection the i.e., by by Sommerfeld radiation (r' = 1) leads to a the that on the other hand, become singularity, infinite at the surface integration all incoming condition). standing fairly uniform energy over the entire nearshore region. purely progressive wave, b < 1 and for and free choosing demanding As noted or equivalently shore The velocities singularity correctly; wave to coeffi- This wave and Allowing permits the velocities which in turn allows a a to large energy level near to the singularity. Examining for several the kinetic different energy as a function of offshore distance values of r' confirms this. Figure 20 is a plot of 1 2 2 2 H (u + v2) KINETIC ENERGY computed decreases energies for the depth-dependent nearshore for r' for < 1 show the the model. case The kinetic of presence (71) perfect of the energy actually reflection, singularity while by rising 141 1.5 KINETIC ENERG 1.0 1.0 I= -45* 0.5 - 0.9 0.8 I A I I I. -I I- 10 5 0 I y (km) Figure 3-20 Kinetic energy versus offshore distance for the depth dependent model 142 slightly very close to the shore. The depth in factor (71) diminishes this effect somewhat. It is reasonable to assume that dissipation rates are great where the energy content of is not included in the wave the model. field is great, even The dissipation rate though in friction this instance can be estimated using Taylor's (1919) dissipation equation, DISSIPATION Figure 21 shows varying model, uniform rate, the Cd U3 = Cd (u2 + v2)3/2 computed values again for offshore as = rate of at about cannot be detected 1. function values The r' < 1). km by the COBOLT moorings. times greater than the offshore from the depth- of r' and with a dissipation the shoreline only This increase starts to be 5 distance for nearshore a marked increase near for imperfectly reflected waves apparent this several different dissipation expected, shows of (72) the shore and, once Furthermore, rate and is it again, is many certainly able to account for the additional dissipation needed to explain the observed onshore energy flux. In surface contrast to (figure 22) the kinetic does not shore for any of the realistilc It energy and show a rise in dissipation, amplitude values of the reflection exhibits instead the linear trend characteristic tides (Petrie, 1975). with coastal the near free to the coefficient. o-f co-oscillating The relatively flat wave amplitude also agrees observations and implies that high dissipation could not be detected by observations taken from tidal stations. rates 143 0 5 10 y (km) Figure 3-21 Dissipation versus offshore distance for the depth dependent model 144 1.6 1.4 r 51. O,0.9, 0.8 1.2 I= -45* I I I 0 I t I I 5 I I~ I I 10 y (km) Figure 3-22 Free surface elevation versus offshore distance for the depth dependent model I - 145 Retaining appears the to a real istic confirm COBOLT tidal moorings are vertical-wall topography some notions records; minimal coastline, be possible in shallow water that can and that the Laplace that have namely, and in been advanced variations be tidal modelled the inferred equations to across explain the three by effectively tidal dissipation a may inside the coverage region of the COBOLT experiment. J. Summary Three idealized analytical models have been examined ii to illuminate coastal the effects region. structure friction They were: model; vertical of a a reflected coastline; and, on tidal measurements constant Sverdrup an effort eddy viscosity, wave model with an a reflected Sverdrup wave model in the vertical absorbing vith simple topography. From the first of these models, viscosity parameterization description of the of vertical ellipticity and orientation, are applied. (1926) must layer bottom appears that a constant frictional stresses structure of provided boundary the be modified to near the bottom. the law, bottom results the observations. which for the presence to vertical the boundary bottom velocity profiles which agree diurnal and in no-slip of eddy an adequate proper boundary condition, A more appropriate stress in \ account allows semidiurnal investigation was the physically inadequate relates drag The it conditions Sverdrup's condition, a turbulent wall condition which through a quite well linear with 146 a boundary which perfectly reflects In the second model, Sverdrup tidal was waves ellipse not capable characteristics. of Specifically, the correct sense of current vector orientation nearshore or the effects in condition. which shallow water, The resulting nearsbore the model rotation, presence dissipative mechanism, frictional boundary angle, explaining barotropic did not show the correct major axis of onshore is due was incident energy to the modelled flux. A enhancement by an of absorbing Sverdrup-Poincare waves success- fully reproduced the features mentioned above. Finally, was realistic topography found that the vertical wall geometry is reality, to by including the if it is stipulated that wavelengths topographic scale and the the in a third model, it a good approximation are large in to comparison region of interest is outside this scale. For the COBOLT region these conditions are met since the topographic length suggests scale is short; that dissipation rates about 5 km. are large within and could account for the apparent coastal region. The the analysis scale also distance absorption of tidal energy in the CHAPTER IV OBSERVATIONS OF COASTAL INTERNAL TIDES A. Introduction With the ability to measure deep ocean realization that tidal frequency motions the surface or barotropic tide. tidal currents are not has come the exclusively due to Internal tides, or internal waves at or near-tidal frequencies, are present in almost all oceanic current records and may even dominate velocity measurements at some locations and frequencies (Gould and McKee, 1973). By obvious extension, much of the observational literature The of extent internal waves of this internal waves specific to by the material Briscoe internal generation process, is applicable is evident (1975). tides, In much to in the a internal general addition, of which and theoretical the is is summarized in an excellent tides. review of literature devoted to the review by Wunsch (1975). Observations e.g., Petrie, on 1975, the continental Halpern, 1971, shelf and in shallow seas Lee, 1971, or Apel, et. al., (see 1975) show that internal tides are common--probably because nearby regions, particularly the are generated despite slope, (Prisenberg, Wilmot, Sandstrom, 1962). rare, continental are areas where and Rattray, internal 1974, tides and Cox and Current measurements near the coast, however, are theoretical interest in this so called "corner region" 148 Measurements (Wunsch, 1969). that have been made (Winant, primarily concerned with freely propagating waves. motions, such as those of the mid-latitude observed near topographic features are Trapped internal diurnal such as 1974) seamounts tide, have (Hendry, been 1975), however, and presumably may be important near the coast. The presence or absence of an internal tidal signal in the COBOLT data It is is of relevance well-known source that velocities of confusion in (Regal and Wunsch, addition mittent both to the barotropic efforts 1973 and Magaard and McKee, of internal amplitudes structure vertical of to internal and the internal structure of tidal current tides phases tides are a major to interpret deep-ocean tidal currents of energy to barotropic nature due tidal analysis of chapter 2. results in these tides in large affect tidal Besides estimates, the the such of vertical analysis parameters the the inter- uncertainties Also, estimates. may barotropic 1973). of the as those a brief associated with the tidal ellipse. B. Dynamic theory of the internal tides The analysis of the COBOLT data is strengthened The development exposition of the theory relevant to internal tides. of the theory here follows the subject such as closely Eckart approximations (1960), involved (1966). The equations are similar to those made (see chapter 3). those given in Phillips in standard (1966), formulating for the Laplace The primary differences are by texts on or the Krauss model tidal equations that the fluid is no 149 longer considered to be homogeneous and the effects of a mean current will be included in the dynamics. Consider then a stratified ocean with a uniform mean current U, in geostrophic equilibrium, and in hydrostatic equilibrium. case letters) waves, Small from this mean state, governed by are a mean density distribution p 0(z), deviations lower such as those caused by internal Boussinesq incompressible, linearized, the (indicated by equations: fk x v 5t+ Dt p = az Dt V -v a where p is density is density to equal VP- ! 0 p k 0 0 (1) 0, perturbation the sum such of p and that p << p0 p0 , k is and a the unit total vector pointing in the +z direction, and _ D Dt a at + U (2) ax The equations which relate the velocity and pressure fields are: D2 2 +* D( 2+ f ) V = Dt 1 (DV h - ( - Vh 0 (3) and 2 + N2) W 2- +Nw Dt2 where Vh is the horizontal 1 D Dt 3zp 0 gradient ' operator D /D x i + a / y the Brunt-Vaisala frequency is defined as j and 150 N 2 (z) Combining = (1) into - a g (p/az)/ single p0 (4) equation in the vertical velocity gives 2 2 ( f2) + ( Dt2 az Dt2 + N2) V2 w h 0, = (5) with associated boundary conditions w(z=0) = = w(z=-H) 0 This "rigid lid" approximation removes . (6) the surface wave solutions of equation (5). C. Solutions for constant Brunt-Vaisala frequency For constant 2 N (z) 2 N a 0 solution to equation (5) is the plane wave, w(x,y,z,t) where the wave vector = K W0 exp i(kx+ky+mz-cot) ki + 2 j + mk, and measured by a stationary observer. , w is (7) the wave frequency The boundary conditions demand a discrete set of vertical wavenumbers m mH = n while substuting the solution (7) sion relation, (8) i into equation (5) gives the disper- 151 ((G- Uk)2_ denote this case) sec N N2- Q- Uk)2)(k2 2 from a to -1 the dispersion relation can be further and N 2 10 = 0 -4 sec where regions, coastal and frequencies tidal 2 2 = the absence of a mean current (removing the tilda In for 2 -2 2 10-8 = that W assuming by , simplified 2 << Then 0 (W2 _ 2 (k2 + 2 2 = 2 The similiarity between this surface gravity waves (10) 0 dispersion relation and that for long chapter 3) (see is apparent if an "equivalent depth" (N /m ) 2 = (N0H g h mT) 2 (11) is defined, reducing (10) to W2 these equations Because +2 ). solutions (12) are similar to the Laplace tidal equa- any solutions implicit in tions, as = g h (k 2 the tidal equations can also appear In particular, there of the internal wave equations. are free and trapped modes that are equivalent to the Sverdrup Kelvin wave solutions of the Laplace tidal equations. is clear numbers case that k for w2 and P,--both both k2 and 2 2 two possibilities demanding are positive that k 2+ 2 From (12) and it exist for the wave- > 0. In the first and freely propagating internal 152 waves result. In the second case k2 < 0 1 and <1k 21 imply exponential decay or growth for the solution (7). 2 For w2 < Ik21 --implying freely in coast. only one waves of possibility this frequency the y-direction but are trapped In addition, the exists-- wavelengths range 1221 £2 < 0 and do propagate not to a boundary such as of both free and the trapped internal motions are considerably smaller than the equivalent surface waves since g h n Taking . N 2 0 = 10 that first mode about 2 = -4 (N H/nl) <<g o sec (n = -2 and H . = (13) 30 1) internal waves 10 km--a factor of 100 less m, for example, should have suggests wavelengths than the equivalent of surface wave- lengths. D. Solutions for an arbitrarily stratified fluid For an arbitrarily stratified fluid equation (5) has a solution w(x,y,z,t) = W(z) exp i(kx+ky-wt) , (14) if (with w2<< N2 and notation of equation (16) retained) d2W N2(z) K2 + 2 2 dz 2 - f2 (15) 0, W where K = ki + tj is now the horizontal wave vector (1K = K) and the boundary of a conditions equation (6) remain. This is classical > 153 eigenvalue problem whose solutions are a countable number of eigenfunctions, Fn(z), eigenvalues and each with an associated eigenfunctions, for eigenvalue arbitrary Kn. N2(z), These can be computed numerically by one of any number of integration techniques. One such method, Krol (1974). a matrix diagonalization technique, The rigid lid boundary conditions is also described by assure that eigenfunctions are orthogonal, K2 Kn H =(--nP) N2(Z) W2( z) dz 2 2 f 2 J )dW H n 0 dz dz 0 H N2(z) Wm (z) Wn (z) dz f = (16) 0 0 dW dW H m n dz- 0 = for m n. 0 Other perturbation fields are related to the eigenfunctions by n (Z) wk + 2i f dWn dz = n v (z) n n p 1 dW n - i fk wK 2 dz n . o- .w 2 nn f 2 n 2 (17) dW n dz N2 pn (z) = i p 0 g W W(z) n 154 Important relative relationships phases between developed from (17). defining the the various amplitude measured ratios quantities and can be For trapped waves, where Z2 < 0, u n _ v o k + fX + fk 2 p pKN on 2 W n ' dz where the decay parameter has been redefined as: able equations for free waves (18) ndWn k + fA1 un X = -i2.. are found by taking Compar- X= 0 since the direction of propagation is arbitrary. The phase propagation relation (u) and that types of waves, 0 0 or 180 o between while depending the velocity perpendicular to in the direction it (v) is 900 for both the phase relation between u and on the sign of dW /dz. n of These p is either phases are an important consideration in determining the propagation direction of the internal wave. The energetics of an internal wave field are examined by multiplying the momentum equation by v, the continuity equation by -p, the density conservation equation by pg/N , and adding and the results to obtain the energy conservation equation, Po D 2 2 2 2 Dt (u + v + w Making vertical the assumption kinetic 2 2 + ( pg/0 ) IN ) +V-pv that w2<< N2 energy, w2. allows Integrating one (19) 0. to (19) disregard over the the water 155 column and for free and trapped waves slightly different expressions results in are substituted into (19): when the relations of equation (17) 2 2 (<u > + <v >) PO = KINETIC ENERGY - 2 4 W2 K 2 f dz dz dz + ( f +fk) 2 4 4 FREE 2 2 _1 (ok +fX) dz 2 dW 2 1 2 dW dz (_n) dz = 4o2 ENERGY FLUX = 0 2 of that there (k i + 1 wK2 po internal is ) >/N waves.) N ratio (20) FREE TRAPPED W FREE dz n 2 wk + fX KN2 2K The = = kj) no vertical dz & <py>dz / = = (Note <(g P / N 2 W2 nd 1 TRAPPED 2 p2 POTENTIAL ENERGY brackets < >) by (indicated average time a performing 2 Wn energy flux in of TRAPPED dz potential the modal description to horizontal kinetic energy, another important diagnostic quantity, is P.E. K.E. _ w2 _ 2 + 2 FREE F( (21) 156 (W _ )(k2 _ ) 2 2 W (cok + fX) (coX + fk) TRAPPED. (21) cont. Finally, the total energy for both wave types is E = K.E. + P.E. = 1 2(w 2_f TN WR dz 2 ) FREE (22) 22 _I (wk+fX) + (wX+fk) 4w2 ( + N2 W2 dz TRAPPED. K2 (W2_ 2) This quantity, the total energy per unit surface area, is that energy carried along by a wave packet. p v = Consequently E C, (23) 4 where C is the group velocity, C C = 3k i +W/UZ A A (24) .j+Dw/@m k E. The mean fields of the COBOLT experiment The average sigma-t cross section formed from the twenty-two days of profiling figure 1. The most noticeable feature is a fairly distinct pycnoBecause this cline about 12 m deep which broadens toward the shore. feature exists in the individual daily cross sections, able to assume that this reduction of stratification is physical process (such as the enhanced mixing) and is artifact of the averaging procedure. shown in is that coincided with the buoy measurements it is reason- due to some not solely an MAY 1977 AVERAGE 0 24.8 8_25.2 16 24.6 25.0 25.4 2 5.6 25.8 5. 24 26.0 32 40 Ut 0 5 10 km Figure 4-1 Twenty-two day average sigma-t cross section at the COBOLT site 15 158 salinity, and sigma-t at the The numerical values of temperature, location listed of spar buoy 3 are a profile representative as in day twenty-two the possible To determine 1. table as average profiles of the two transect stations located on either side of a given mooring (averages were also performed for buoys 2 and 4) were Also included in the combined to make the estimates in the table. table are the standard deviations of the averaged temperatures and salinities. These profiles suggest that density during the month of May was primarily controlled by the salinity. This can be checked quantita- tively by noting that the partial derivatives of density with respect to temperature and salinity (obtained directly from the equation of state) are quite different, i.e., = 7.5 x 10- 4 ap as = DT the 10 -4 3 (25) 3 10 c.3 ./0 cm o/oo At buoy 3 a temperature contrast across of about 30 results in a density change of 4.2 x thermocline halocline -1.4 x and S = 32 o/oo. gm/cm3. 3 cm 0/00 ap at T = 8 gm The giving a salinity change was density change of about 7.5 twice as great as that due to temperature. x 1 o/oo 10~4 across the gm/cm 3--almost The late-spring measure- ment period and the large amounts of fresh water discharge from the Connecticut River (Ketchum and Corwin, 1964) are probably responsible for this result. 159 TABLE 4-1 NUMERICAL VALUES OF T, S, SIGMA T, AND N2 AT BUOY 3 DEPTH 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Standard deviation: TEMP. SALIN. ( 0C) (0/ 0) 10.71 10.68 10.60 10.50 10.39 10.26 10.08 9.86 9.63 9.40 9.17 8.94 8.70 8.46 8.25 8.07 7.92 7.80 7.67 7.56 7.47 7.40 7.32 7.23 7.15 7.08 7.02 6.97 6.91 6.86 6.84 31.96 31.96 31.97 31.98 31.99 32.01 32.05 32.09 32.15 32.21 32.28 32.36 32.44 32.52 32.59 32.65 32.70 32.75 32.78 32.81 32.85 32.88 32.90 32.92 32.94 32.95 32.96 32.97 32.99 33.01 33.02 1.5 0.1 SIGMA T 24.46 24.46 24.48 24.51 24.54 24.58 24.63 24.70 24.78 24.87 24.96 25.06 25.16 25.26 25.34 25.41 25.48 25.53 25.58 25.62 25.65 25.69 25.72 25.75 25.77 25.79 25.81 25.83 25.85 25.87 25.87 N2 PERIOD (SEC-2) (MIN) 0.73x10~4 1.78 2.27 2.85 3.99 5.49 6.86 7.86 8.57 9.14 9.60 9.75 9.32 8.37 7.23 6.17 5.27 4.54 3.97 3.58 3.34 3.11 2.76 2.30 1.87 1.67 1.76 2.01 1.89 0.81 12 8 7 6 5 4 4 4 4 3 3 3 3 4 4 4 5 5 5 6 6 6 6 7 8 8 8 7 8 12 160 Also included in table 1 (and displayed in figure 2 for all three mooring The locations) are equivalent the values of periods less .quite short--generally are frequency. the Brunt-Vaisala ten than minutes--implying that resolution of high frequency internal waves is not possible because of the one hour buoy averaging period. high frequency internal waves will alias not the However, low frequency signals either. The COBOLT instruments also recorded significant mean currents. Onshore mean currents are generally quite small (less than 2 cm/sec) The alongshore and can be ignored as an influence on internal tides. mean currents, intervals, in presented table two 2 for averaging different are substantially stronger than onshore currents and more comparable to the phase speed of internal flowed to the west during the experiment, These waves. and, currents given the distance to Montauk Point and the entrance to Long Island Sound (60 km), suggest an upper layer advective time scale of 7-14 days. Although the measurement period over which these averages were made was quite variable, been included. Figure 3, showing the salinity time series of buoy 3 at 3.8m and 25.Om (instruments 31 changes measured during the crossed the site causing the Higher salinities approximately ten days, from about 32.5 and 34), experiment. homogeneous. sharply that the major time scales have appears it water to the abrupt On May 10 a large column 20, when 31.7 o/oo. storm to become practically then persisted at the until May o/oo illustrates the surface salinity This for changed freshening is 161 N (sec-2 ) 10 15x10 10 E 15 0 C 20 25 30 Figure 4-2 Twenty-two day average Brunt-Vaisala frequency profiles at bouys 2, 3 and 4 162 TABLE 4-2 MEAN ALONGSHORE CURRENTS FOR BUOYS 2, 3, & 4; MAY 1977 CURRENTS AT BUOY NO. 2 3 4 APR. 30-MAY 15 LEVEL - -4.5 cm/sec 2 -6.2 3 -2.8 -6.6 -7.4 4 -0.6 -2.4 -2.6 -10.5 APR. 30-MAY LEVEL -11.9 cm/sec 1 - 1 -3.2 cm/sec 2 -5.6 -8.2 3 -2.9 -5.0 4 -1.2 -2.3 *Negative values are to the west -10.2 33.5 33.5 33.0 33.0 32. 5 32.5 9.. z 32.0 32.0 z cc an 31.5 31.5 31.0 31.0 Figure 4-3 Salinity time series from instruments 31 and 34 164 due probably from Long Island an influx of water to nearest and most logical source of fresh water). Sound (the Furthermore, the ten day time estimate obtained from the salinity series agrees with the advective time scale suggested by the mean currents and supports the notion that a thirty day averaging period is reasonably representative. F. Internal tidal oscillations COBOLT The internal tides. experiment is well-suited for observations of the The three kilometer spacing of moorings was intended to provide a coherent array for internal (albeit very low frequency) While the orientation of the transect line (i.e., motions. to dicular shore) limits sensitivity directional to perpen- the offshore direction, the propagation of free waves is probably biased in this direction by the local topography. The sigma-t time series of all four instruments on buoy 3 (figure 4) shows plainly the variations of interest here: the internal tides. The large regular oscillations at instruments 32 and 33 are typical of measurements tions at instruments near the pycnocline while smaller oscilla- 32 and 33 indicate reduced stratification and adjustment to top and bottom boundaries. To identify the major periodicities of pycnocline oscillations, energy spectra of sigma-t time series were computed. the averaged sigma-t Figure 5 shows energy density of instruments 21-23 and 31-33 for the twenty-five day period over which both buoys were operational 26.5 26.5 26.0 - 26.0 2.4 m 25.5 - 25.5 25.0 25.0 a- 24.5 24.5 a- (f) 0 24.0 24.0 23.5 - I 01 77 - I r- 06 11 r1 AA gureu - oVaL61t:bVt I i cI 16 - 6 I 21 Ld I levelsbuo 26 3UY% 23. 5 166 o-t 100 D I SD A 10-1 10-2 10-3 195% 0.01 0.10 1.00 FREQUENCY (cycles/hr) Figure 4-5 Sigma-t energy density spectrum from insts. 21-23 & 31-33 167 (April 30-May 24), with the values of diurnal inertial (18.35 hr), for reference. and semidiurnal (12.42 hr) (D = 24.00 hr), local included frequencies These instruments were selected as representative of upper layer and pycnocline fluctuations. The internal oscillations have two resolved maxima at which energy rises above the background level: a broad peak centered around the inertial frequency and a sharp peak centered on the principal semidiurnal frequency. the "inertial" frequencies, it peak is so much contains almost broader three than times accounts for most of the regular oscillations figure 4. By contrast, the temperature that as at much Because semidiurnal energy and that catch the eye in spectrum from the same instrument packages (figure 6) does not show any energy significantly above the continuum. Further insight into the nature of the internal oscillations can be obtained by examining the kinetic energy spectrum. A comparison of the energy density of onshore velocities (figure 7) with that of alongshore velocities (figure 8) shows a marked disparity. energy is much less than alongshore semidiurnal frequency ranges, In fact, alongshore currents energy for both Onshore diurnal and but not for frequencies near inertial. are almost ten times as energetic as onshore currents for the semidiurnal band and almost three times as energetic for the diurnal. Currents in the inertial range, on the other hand, have comparable energies in both directions. A convenient way to characterize the sigma-t variations is to convert the energy in a particular frequency band to an equivalent isopycnal displacement using the differential 168 T 101 D I SD 100 10- 10-2 195 0.01 0.10 1.00 FREQUENCY (cycles/hr) Figure 4-6 Temperature energy density spectrum from insts. 21-23 & 31-33 169. PERID 9 HRS. 10 +4100 10 ii (II ED H- CrN +3 10 CD +2 10 cn LLJ Z 11 10 cz CD +0 10 Figure 4-7 Onshore velocity energy density spectrum from insts. 32 and 33 1 1-1 -I II FUN0.01 0.1 FREQUENCT,.CYCLES/HRS. 170 PERIDD, +4100 10 11 1i HRS. 10 1 (NJ CL LL (iT N (NJ +3 10 (I 10 LJ c> Lu +1 10 LJ Li -- +[) 10%-1 Figure 4-8 Alongshore velocity energy density spectrum from insts. 32 and 33 1 0-1I '.0 1 O. 01 0.1 FREQUENCY, CCLES/HRS. 171 Aa These quantities = Da/Dz Az = N2 AC . are averaged .for buoys (26) 2 and 3 at each instrument level and presented for the three periods of interest in table 3 with the isopycnal displacement expected for a long surface wave of 100 cm amplitude. This last item is velocity decreases computed by assuming that the vertical linearly to zero at the bottom and that it is equal to the time derivative of the displacement, i.e., w(z) = aw (H-z)/H and (27) ;/t where a is = w(z), the amplitude of the surface wave, ? is the isopycnal displacement, and H is the depth of the water. Since the semidiurnal surface tide has an amplitude of about 1 m in the COBOLT region and the diurnal surface tide an amplitude of about 10 cm, these numbers give a fair indication of the isopycnal displacements due to surface tides alone. A comparison of the measured and computed displacements shows to what degree the surface tide can account for the sigma-t variations at each frequency. In the semidiurnal band displacements due to the surface tide and those computed from density variations are virtually identical (except into for instruments at level 4 where energy peaks fade the continuum) implying that the narrow energy peak at this 172 TABLE 4-3 ISOPYCNAL DISPLACEMENTS FOR BUOYS 2 & 3; ALL LEVELS INST N2 12 HR. 18 HR. 24 HR. (cM) (CM) (CM) 3.9x10- 2 70 2.3x10-2 73 2.4x10- 2 45 2.8 79 4.9 11 4.4 4.8 50 2.4 100 4.8 52 2.5 24 2.4 96 2.3 67 1.6 42 1.0 31 2.3 170 4.0 130 3.1 91 2.1 32 6.9 62 4.3 78 5.4 78 5.4 33 7.2 60 4.3 86 6.2 60 4.3 34 2.0 180 3.5 65 1.3 110 2.2 21 3.3x10 4 22 6.2 23 102 DISPLACEMENTS FROM SURFACE WAVE WITH 1 METER AMPLITUDE DEPTH AC~ 1 4 m 87 cm 2 8 73 3 16 53 LEVI L 173 is almost frequency tide. The greater exclusively band diurnal can be than a consequence however, displacements, expected from is there reasonable to no assume that horizontal barotropic 5-10 are times tide. And finally, it displacement around 18 hours, this energy is baroclinic all of and is in the large horizontal scale (barotropic) and In this sense nature. small surface the surface wave contributions therefore must be in part due to the baroclinic since of (baroclinic) scale fluctuations are sorted by frequencies in the May experiment. Unlike the May This sorting of dynamics does not always occur. 1977 experiment, where kinetic energy is found in a broad band at and above diurnal density for spread frequencies, the September a comparable 1975 spectrum (figure experiment fairly symmetrically about 24 hours, inertial frequency. of kinetic 9) shows energy energy and not as high as the The semidiurnal energy peak is again very sharp, as in May, 1977, and is centered at 12.42 hours. G. Modal structure of the internal tides It flows, is possible to discriminate between barotropic and baroclinic and between the examining the modal different modes structure of density distribution (see table 1). three vertical velocity density distribution at (1974). These modes the (figure 10) the computed currents Consider, eigenfunctions, buoy 3 by of baroclinic motions, for example, Wn(z), computed from by the the first from the procedure outlined by Krol generally have large amplitudes lwm' j 174 PERIOD, 1 U I 10 100 I HRS. I I I C\i C IL 162 Z cc LL) w CD HU L- 1 1 Figure 4-9 10 Kinetic energy density at 15 m for September, 1975 experiment F-I 0.01 0.1 FREQUENCY, CTCLES/HRS. 175 VERTICAL MODES Figure 4-10 First three vertical velocity modes at buoy 3 176 where N2 is large, toward top the i.e. and in the pycnocline, bottom, where velocity adjusts to the boundaries. N2 and smaller amplitudes weakens and the vertical These are the features that are evident in the sigma-t variations of the COBOLT experiment. The dW /dz, n horizontal velocity modes, are shown in figure 11. which are proportional to Unlike vertical velocities, which must be extrapolated from other fields, the horizontal velocities are directly measured quantities and can be used in a straightforward manner in interpreting the distribution of energy among the modes. The COBOLT data were analyzed by fitting the calculated eigenfunctions to the observed velocities in a least squares sense. a continuous ments u. at eigenfunction each U (z) = n M different of dW /dz n points Given and velocity measurein the vertical, it is possible to determine a coefficient an, for each mode, such that ( u. i is a minimum. The - aU (z.))2 h. n n i (8 i (28) factors, weighting h., trapezoid integration rule to favor instruments vertical range. are that chosen by the cover a large Minimizing (28) with respect to an determines the values of the coefficients, Nh.u.U (z.) ai n i(29) i n h.U2 (z i=1 i n 1 Because the eigenfunctions have arbitrary amplitudes, a more useful 177 HORIZONTAL Figure 4-11 MODES First three horizontal modes at buoy 3 178 quantity is the ratio of the fitted eigenfunction variance to the observed variance, M Pn 2 2 h.a U (z.) i n n2 i=1 (30) 11 which is expressed as the percentage of observed variance (or energy) that can be accounted for by a fit of the nth eigenfunction. The results of the eigenfunction analysis and mode fitting for three different frequency bands are summarized in table 4. With only four instruments measuring velocity (three than four modes is not possible. (n = 0) and barotropic first at buoy 3) a fit In practice it two baroclinic accounted for virtually all the variance, of more was found that the modes (n = 1 & 2) so fitting of higher order modes was not necessary. Table 4 confirms analysis by showing: largely barotropic; clinic; and, also shows the results of the isopycnal displacement that the vertical variance in the 12 hr band is the variance in the 18 hr band is largely baro- that the variance in the 24 hr band is mixed. that onshore velocities at The table 12 and 24 hr have a higher percentage of variance in the baroclinic modes than do alongshore velocities. The mode fitting of assertion that energy in 18 hour variance this band is is consistent with the internal tidal energy that has been Doppler-shifted away from the diurnal band, since 80-95% of the variance at all three COBOLT moorings can be attributed to the first 179TABLE 4-4 PERCENTAGE OF VARIANCE IN BAROTROPIC AND FIRST TWO BAROCLINIC MODES % AT 12 HR. B U M 0 0 D Y E NO. EA. 2 0 55 95 16 2 2 0 0 0 81 96 1 3 1 6 1 3 4 0 % AT 18 HR. WAVE % AT 24 HR. LENGTH (KM.) 12 HR. 24 HR. 33 12 17 0 1 5 7 4 90 64 92 82 4 28 15 22 0 1 6 2 3 6 9 2 4 3 11 71 51 NO. EA. T. NO. EA. 1 0 1 5 87 63 3 97 99 98 90 9 0 0 0 0 2 2 0 1 1 89 95 2 2 T. 90 95 86 T. 44 94 33 1 6 81 81 81 55 9 24 16 23 1 0 0 0 0 0 1 0 0 7 9 NO. = Onshore component EA. = Alongshore Component TOT. = Total variance 180 baroclinic mode. mode, The 18 hr velocities at buoy 2 and the fitted first shown together in figures 12 and 13, visually confirm the baroclinic nature of this freqency band. The modal eigenvalue, analysis also supplies Kn = (27r /wavelength). the magnitude These wavelengths (see are approximately 15 km for the first internal mode in of the table 4) the COBOLT region, and around 5 km for the second internal mode. For trapped waves, more information is needed to determine the wavelength--i.e., there must be some method of choosing the decay scale or e-folding distance. Traditionally the Kelvin wave problem is modelled boundary in an ocean condition. with Equation a vertical (17) wall shows and that no-normal-flow this condition (assuming y is the onshore direction), v &w- ifk -i (wX + fk) (31) , demands that A =-fk/w, (32) so that the eigenvalue is K2 n W2_f2 o -f As a result, Kelvin wave the wavelength of k2 n W2 (33) a first mode semidiurnal internal is almost half that of a free wave. In comparison, a first mode diurnal internal Kelvin wave has a wavelength of around 20 km. Furthermore, it is necessary, in order to have exponential offshore decay, for these wave to propagate alongshore to the west (the negative x direction)--the same direction as the mean current. 181 -4 4 cm/sec -2 Figure 4-12 Fitted first baroclinic mode at buoy 2: onshore velocities - -ffi~ 182 -4 -2 Figure 4-13 0 2 Fitted first baroclinic mode at buoy 2: alongshore velocities 4 cm/sec 183 H. Comparison to theory This subject last fact to Doppler Kelvin wave makes the shifting. internal The Kelvin frequency wave particularly of a diurnal internal superimposed on a mean current of 10 cm/sec would be measured by a stationary observer as w = + Uk = 1/24 hr + 10 cm/sec x 1/20km = 1/17 hr This estimate, using (34) . realistic values for all of the parameters, results in a Doppler-shifted frequency that is remarkably close to the sigma-t energy maximum observed in the COBOLT data. In fact, smaller amplitude mean currents would bring the frequency estimate more into line with the observed sigma-t energy peak at 19-20 hour periods. This evidence again favors the hypothesis that energy peaks at near inertial frequencies in the May 1977 data are a result of Doppler-shifted internal tides of diurnal period. Vertical coherence at 18 hour periods is high among the COBOLT instruments confidence (always level) significantly so computed accurately. equation (18). u-velocities phases These An average different between phases of (alongshore east) all from measured agree with COBOLT zero quantities the 50, indicating a clockwise rotation of ellipses. the 99% can be predictions of instruments lag v-velocities at shows that (onshore) by 910 + Also, lower layer 184 (level 3) alongshore variations by velocities lead 1700 + 10 --again in thermocline (level 3) sigma-t close agreement to predicted values. Horizontal coherence periods and phases between buoys is also generally small (less high at 18 hour than 200) indicating that the wave crests are parallel to the mooring transect (perpendicular Phase differences to the shore). that do exist can be explained by considering the different mean alongshore velocities at each of the moorings. The comparison of observations boundary Kelvin wave model condition used demands that Furthermore, to choose fails in onshore the the the idealized certain respects. vertical The boundary offshore decay scale (equation (33)) velocities magnitude with of be the identically decay scale, zero everywhere. using appropriate values of W, f, and k, suggests that the e-folding distance should be less than 3 km. velocities Observations, by contrast, are comparable to alongshore indicate that onshore velocities (u/v ~ 0.9) and that their magnitudes do not show any measureable decrease offshore, even out to 12 kilometers (buoy 4). It is apparent from the discussion that led to equation (34) that the absence of onshore currents and the choice of a decay scale are both a consequence of the no-normal-flow boundary condition. of the relatively gentle bottom slope (see chapter 3), In view this boundary condition and the traditional Kelvin wave model are probably inappropriate for the COBOLT region. The pycnocline intersects the bottom 185 several kilometers from the and shore suggests that the boundary condition may be modelled more correctly by demanding that v dz =0. (35) already met by the baroclinic modes and cannot be This condition is Unlike the Kelvin wave, used to determine a decay scale. the integral condition does allow onshore velocities however, (see equation (18)) for decay scales other than that obtained in equation (32). Although the traditional, vertical boundary, internal Kelvin wave model fails for to account possible that similar, observations. Also, some of the observed trapped-wave dynamics are responsible because of the strength and it is features, for the persistence of coastal mean currents, it is reasonable to assume that the broadening of the kinetic energy of around surface diurnal and frequencies Doppler-shifted is due to internal the tidal combined presence motions. Wunsch (1975) suggests that a broadening of energy peaks is one of the noticeable features of the internal tide, and that it can be used to distinguish the respective contributions of barotropic and baroclinic tides to current meter records. peak in both fall and spring measurements The presence of a broad suggests a persistent generation mechanism such as the barotropic tide since other possible generation processes (e.g., wind stress) are intermittent and quite different for the two seasons. 186 Despite these arguments, however, it is difficult to establish unequivocally the reality of the Doppler-shifting mechanism without examining other effects which may be direct forcing, mean shear, topography, important. etc., Factors such as may be responsible for the unusual results of the May, 1977 experiment. Only more inclusive models and further examination of the data will resolve this question. I. Energy and flux of the internal tide The energy content of the internal diurnal the aid of equation (24). tide is examined with From table 4 (assuming that the Doppler mechanism is operating) it is apparent that not all of the diurnal shifted to the 18 hour band, internal tidal energy is since 24 hour period velocities still show a substantial amount of variance in the first baroclinic mode (about 30%). It is estimated that this contribution is about one-fourth of the contribution from the 18 hour band and is ignored in the following calculations. Depth integrated values of potential and kinetic energies for the 18 hour band are averaged for buoys 2-4 to give This is the K.E. = 33 ±5 P.E. = 5 ± 1 Joules/m 2 Joules/m 2 (36) . roughly half of the energy content that can be computed for surface diurnal tide and less surface semidiurnal tide. The ratio of energies is than one-tenth of that of the 187 = P.E./K.E. 0.15 ± 0.05 , (37) again in contrast to the Kelvin wave model. decay scale in equation (21) Using the Kelvin wave suggests equipartition of energy (i.e., P.E./K.E. = 1) while a larger decay scale, more in line with the observations, gives an energy ratio of less than one, as observed. It is also possible to perform a crude energy flux calculation. Unlike ties the surface wave flux calculation, (free surface computation, the knowledge of equation (23) elevation and flux calculation the dynamics. can be For used if velocity of the wave is not velocities) for internal a it is assumed quanti- were used waves requires shore-trapped that for some internal wave, the group the i.e. C = 20 km/24 hr = 23 In this case Energy Flux = 9 watts/m alongshore to the west. While this figure is (38) small with respect to computed surface semidiurnal flux rates (chapter to the too different than the phase velocity (they are identical for the Kelvin wave); cm/sec. where measureable deep water internal tide 2) it is comparable fluxes measured by Wunsch and Hendry (1972). If the internal wave west, the Topographic source of features this is assumed to progress energy is probably at the entrance alongshore to the Long Island Sound. to the Sound itself are quite 188 pronounced and undoubtedly provide the correct length scale for The alongshore length scales of topographic features to generation. the south of Long Island, by contrast, are not well-matched of the internal tide, but are generally much longer. to those Thus a topo- graphic generation process such as that proposed by Baines (1973) is more likely to occur at the entrance to the Sound than locally along Furthermore, the South Shore. wide with respect to the entrance to Long Island Sound is the internal tide wavelength and undoubtedly will prevent any transmission across from the southern coast England (see Buchwald, 1971). of New It is also possible, though the matter is open to speculation, that the semi-permanent density front known to exist where the fresh waters of the Sound come into contact with the saline waters of the Mid-Atlantic Bight, play a role in the generation process. It is not likely that the internal tide evident periods results from generation in Long Island Sound. at 12 hour Internal waves at semidiurnal frequencies are free waves and are therefore able to radiate away from the .generation region in likely source would be onshore generation from offshore regions. Because barotropic currents, analysis of difficult. nearshore (e. all directions. g., A more the shelf break) or dominated by the records are the internal oscillations is very Even so, there are indications in the COBOLT data that density flucuations generation in the coastal zone. lead those further out; evidence of Until longer records are available, it is not feasible to resolve this question fully. PAGES (S) MISSING FROM ORIGINAL 190 J. Conclusions effects The the of internal tides chapter 2 should be clear at this point. little baroclinic analysis. energy present Onshore velocities, on the Diurnal velocities, of Semidiurnal velocities show to interfere with the barotropic where the baroclinic effects were the strongest, were indeed subject to the most variations 2). analysis tidal during the May 1977 experiment, (see chapter experienced a bit more interference from internal tides but not nearly so much as might have occurred in the absence of a mean current. In either case, the fact that baroclinic variance is primarily in the first mode promotes the success of the vertical integration as a way of reducing the effects of the baroclinic tides on the results of the barotropic analysis. REFERENCES Apel, J. R., H. M. Byrne, J. R. Proni, and R. L. Charnell. 1975. Observations of oceanic internal and surface waves from the Earth Resources Technology Satellite. Journal of Geophysical Research. 80:865-881. . Abramowitz, M. and I. Stegun. 1964. Handbook of Mathematical Functions. National Bureau of Standards, Applied Mathematics Series 55. 1046pp. Ball, F. K. 1967. Edge waves in Research 14:79-88. a ocean of finite depth. Deep-Sea Baines, P. 1973. The generation of tides by flat bump topography. Deep-Sea Research. 20:179-205. Beardsley, R. C., W. Boicourt, L. C. Huff, and J. Scott. 1977. CMICE 76: A current meter intercomparison experiment conducted off Long Island in Feb.-Mar. 1976. Woods Hole Oceanog. Inst. Technical Report WHOI 77-62: 123 p. Beardsley, R. C., W. Boicourt, and D. Hansen. 1976. Physical oceanography of the Middle Atlantic Bight. in Middle Atlantic Continental Shelf and the New York Bight. vol 2. M. Grant Gross, ed. American Society of Limnology and Oceanography. p. 20-34. Bendat, J. S. and A. G. Piersol. 1971. Random Data: Measurement. Wiley-Interscience, N. Y.: 407 p. Bowden, K. F. 1962. Turbulence. The Sea. Vol. 1. Hill. Interscience Pub., N. Y.: 802-825. Analysis and Ed. by M. N. Bowden, K. F. and L. A. Fairbairn. 1956. Measurements of turbulent fluctuations and Reynolds stresses in a tidal current. Proceedings of the Royal Society of London. A 237:422-438. Briscoe, M. 1975. Internal waves in the ocean. Geophysics and Space Physics. 13:591-645. Reviews of Buchwald, V. 1971. The diffraction of tides by a narrow channel. Journal of Fluid Mechanics. 46:501-511. Butman, B. 1975. On the Dynamics of Shallow Water Currents in Mass. Bay and on the N. England Continental Shelf. Doctoral Dissertation. Massachusetts Institute of Technology and Woods Hole Oceanographic. 192 Cartwright, D., W. Munk, and B. Zetler. measurements. EOS. 50:472-477. 1969. Pelagic tidal Charnell, R. L. and D. Hansen. 1974. Summary and Analysis of Physical Oceanography Dat'a Collected in the New York Bight During 1969-1970. NOAA/MESA Report No. 74-3. Boulder, CO. Cox, C. and H. Sandstrom. 1962. Coupling of internal and surface waves in water of variable depth. Journal of the Oceanographic Society of Japan. 30th Anniversary Volume. 449-513. Csanady, G. 1972. The coastal boundary layer in Lake Ontario. Journal of Physical Oceanography. 2:41-53, 168-176. Csanady, G. 1977. The coastal jet conceptual model in the dynamics of shallow seas. in The Sea. Vol. 6. E. D. Goldberg, Ed. John Wiley and Sons, N.Y. 117-144. Cushing, V. 1976. Electromagnetic water current meter. Oceans 76 Conf. Proc 25C. 1-7. Defant, A. 1961. N.Y. 560 p. Physical Oceanography. Vol. 2, MTS and IEEE Pergamon Press, Dietrich, G. 1944. Die Schwingungssystems der half - und eintagingen tiden in den ozeanen. Veroeffentl. Inst. Meerskd. Univ. Berlin. A41:7-68. Dimmler, D. G., D. Huszagh, S. Rankowitz, and J. Scott. 1976. controllable real-time collection system for coastal oceanography. Ocean 76 Conf. Proc. 14E-1:1-13. A Doodson, A. T. and H. D. Warburg. 1941. Admiralty Manual of the Tides. His Majesty's Stationery Office, London:270 pp. Dronkers, J. J. 1964. Tidal computation in waters. North-Holland Pub., Amsterdam. rivers and coastal 173 pp. 1960. Hydrodynamics of Oceans and Atmospheres. Eckart, C. Press, N.Y. 290 pp. Pergamon EG&G, Environmental Consultants. 1975. Summary of Oceanographic Observations in New Jersey Coastal Waters near 39 28'N Latitude and 74 15'W Longitude during the period May, 1973 through May, 1974. A Report to the Public Service Electric and Gas Co., Newark, N.J. Emery, K. 0. and E. Uchupi. 1972. Western North Atlantic. Assoc. of Petroleum Geologists, Tulsa, Okla. 532 pp. Amer. 193 Ferrel, Wm. 1874. Tidal Researches. U. S. Coast Survey Rept. Govt. Printing Office, Washington, D.C. 268 pp. U. S. Filloux, J. H. 1971. Deep-sea tide obs. from the northeastern Pacific. Deep-Sea Research. 18:275-284. Flagg, C. N. 1977. The Kinematics and the Dynamics of the New England Continental Shelf and Shelf/Slope Front. Doctoral Dissertation, Massachusetts Institute of Technology-Woods Hole Oceanographic Institution. 207 pp. Gould, W. and W. McKee. Vertical structure of semidiurnal tidal currents in the Bay of Biscay. Nature. 244:88-91. Haight, F. J. 1942. Coastal Currents along the Atlantic Coast of the United States. U. S. Coast and Geodetic Survey Spec. Publ. No. 230. 73 pp. Halpern, D. 1971. Semidiurnal internal tides in Massachusetts Bay. Journal of Geophysical Research. 76:6573-6584. Hendershott, M. C. 1973. Ocean tides. EOS 54:76-86. Hendershott, M. C. 1977. Numerical models of ocean tides. Sea. Vol. 6. E. D. Goldberg, Ed. John Wiley and Sons, 47-95. The N.Y. Hendershott, M. C. and W. Munk. 1970. Tides. Annual Review of Fluid Mechanics. Vol. 2. M. Van Dyke, Ed. Annual Reviews, Inc. Palo Alto, CA. 205-224. Hendershott, M. C. and A. Speranza. 1971. Co-oscillating tides in long, narrow bays; the Taylor problem revisited. Deep-Sea Research. 18:959-980. Hendry, R. M. 1975. The Generation, Energetics and Propagation of Internal Tides in the Western North Atlantic. Doctoral Dissertation, Massachusetts Institute of Technology-Woods Hole Oceanographic Institution. 345 pp. The persistence of "winter" water 1964. Ketchum, B. and N. Corwin. on the continental shelf south of Long Island, New York. Limnology and Oceanography. 9(4):467-475. Krauss, W. 1966. Methoden und Ergebnisse der Theoretischen Ozeanographie: Interne Wellen. Gebruder Bor'ntraeger, Berlin. 248 pp. 194 Krauss, W. 1973. Methods and Results of Theoretical Oceanography: Dynamics of the Homo. and Quasihomo. Ocean. Gebruder Bor'ntraeger, Berlin. 302 pp. Krol, H. 1974. Some numerical considerations: Eigenvalue problems in the theories of underwater sound and internal waves. NATO SACLANT ASW Memorandum SM-40. Lamb, H. Lee, 1932. Hydrodynamics. Dover Pub., N.Y. 0. 1961. Observations on internal waves in Limnology and Oceanography. 6:312-321. 738 pp. shallow water. Lowe, R. L., D. L. Inman, and B. M. Brush. 1972. Simultaneous data system for instrumenting the shelf. 13th Conf. on Coastal Eng., Vancouver, B.C. Magaard, L., and W. McKee. 1973. Semidiurnal tidal currents at Site D. Deep-Sea Research. 20:997-1009. McCullough, J. 1977. Problems in measuring currents near the ocean surface. MTS & IEEE Ocean 77 Conf. Proceedings 468:1-7. Miles, J. 1974. On Laplace's tidal equations. Mechanics. 66:241-260. Journal of Fluid Miller, G. 1966. The flux of tidal energy out of the deep ocean. Journal of Geophysical Research. 71(10):2485-2489. Mooers, C. N. K. 1975. Several effects of a baroclinic current on the cross-stream propagation of inertial-internal waves. Geophysical Fluid Dynamics. 6:245-275. Morse, P. and H. Feshback. McGraw-Hill, New York. 1953. Methods of Theoretical Physics. Munk, W. 1968. Once again--tidal friction. Astron. Soc. 9:352-375. Quart. Jour. Roy. Munk, W. and D. Cartwright. L966. Tidal spectroscopy and prediction. Proc. Roy. Soc. London. A259:533-581. Munk, W., F. Snodgrass, and M. Wimbush. 1970. Tides off-shore: Transition from California coastal to deep-sea waters. Geophysical Fluid Dynamics. 1:161-235. Mysak, L. 1978. Wave propagation in random media, with oceanic applications. Reviews of Geophysics and Space Physics. 16(2):233-261. 195 Niiler, P. 0. 1975. A report on the continental shelf circulation and coastal upwelling. Reviews of Geophysics and Space Physics. 13(3):609-659. Patchen, R., E. Long, and B. Parker. 1976. Analysis of Current Meter Observations in the New York Bight Apex, August, 1973-June, 1974. NOAA Technical Report 368-MESA 5. U. S. Department of Commerce. 24 pp. Peregrine, D. H. 1976. Interaction of water waves and currents. in Advances in Applied Mechanics, vol.16. C. S. Yih, Ed. Academic Press, N. Y. p9-117. Petrie, B. 1975. M2 surface and internal tides on the Scotian shelf and slope. Journal of Marine Research. 33:303-323. Phillips, N. A. 1966. The equations of motion for a shallow rotating atmosphere and the "traditional" approximation. Journal of Atmospheric Science. 23:626-628. Phillips, 0. 1966. Press. 261p. Dynamics of the Upper Ocean. Cambridge Univ. Platzmann, G. 1971. Ocean tides and related waves. Lectures on Applied Mathematics, v. 14. Ed. by W. Reid. American Mathematical Society, Providence, R.I. Prinsenberg, S. J., W. Wilmot, and M. Rattray, Jr. 1974. Generation and dissipation of coastal internal tides. Deep-Sea Research. 21:263-281. Proudman, J. 1941. The effect of coastal friction on the tides. Geophys. Suppl. Mon. Nat. Roy. Astrono. Soc. 5(l):23-26. Proudman, J. Rattray, M. Tellus. 1953. Dynamical Oceanography. John Wiley, N.Y. 409 pp. 1960. On the coastal generation of internal tides. 12:54-62. Redfield, A. 1950. The analysis of tidal phenomena in narrow embayments. Papers in Physical Oceanography and Meteorology. 11(4):1-36. Redfield, A. 1958. The influence of the continental shelf on the tides of the Atlantic coast of the United States. Journal of Marine Research. 17:432-448. Redfield, A. 1978. The tide in Research. 36:255-294. coastal waters. Journal of Marine 196 Regal, R. and C. Wunsch. 1973. M 2 tidal currents in the western North Atlantic. Deep-Sea Research. 20(5):493-502. Rooth, C. 1972. A linearized bottom friction for large scale oceanic motions. Journal of Physical Oceanography. 2(2):509-510. Scott, J. and G. Csanady. 1976. Nearshore currents off Long Island. Journal of Geophysical Research. 81(30):5401-5409. Scott, J., T. Hopkins, R. Pillsbury, and E. Davis. 1978. Velocity and Temperature off Shinnecock, Long Island in October-November, 1976. Oceanographic Services Division, Brookhaven National Laboratory (unpublished manuscript). Shureman, P. 1958. Manual of Harmonic Analysis and Prediction of Tides. U. S. Coast and Geodetsic Survey, Special Publication No. 98. U. S. Government Printing Office, Washington, D.C. 317 pp. Smith, P., B. Petrie, and C. Mann. 1978. Circulation, variability and dynamics of the Scotian Shelf and Slope. Jour. Fish. Res. Board Can. 35:1067-1083. Sverdrup, H. 1926. Dynamics of tides on the North Siberian shelf. Geophys. Publ. 4(5):1-75. Sverdrup,- H., M. Johnson, and R. Fleming. 1942. Prentice-Hall, Inc., Englewood Cliffs, N.J. The Oceans. 1087 pp. Swanson, R. 1976. Tides. MESA New York Bight atlas Monograph 4. New York Sea Grant Inst., Albany, N.Y. 34 pp. Swift, D., D. Duane, and T. McKinney. 1973. Ridge and swale topography of the Middle Atlantic Bight, North America: Secular response to the Holocene hydraulic regime. Marine Geology. 15:227-247. Taylor, G. I. 1919. Tidal friction in the Irish Sea. Roy. Soc. A220:1-93. Phil. Trans. Traschen, J. 1976. Tides of the New England continental shelf. Woods Hole Oceanographic Institution (unpublished manuscript). Whitham, G. 1960. A note on group velocity. Mechanics. 9:347-352. Journal of Fluid Winant, C. 1974. Internal surges in coastal waters. Geophysical Research. 79:4523-4526. Winant, C. 1978. Coastal current observations. of Geophysics and Space Physics. Journal of To appear in Reviews 197 Wunsch, C. 1969. Progressive waves on slopes. Mechanics. 35:131-144. Journal of Fluid Wunsch, C., and R. Hendry. 1972. Array measurements of the bottom boundary layer and the internal wave field on the continental slope. Geophysical Fluid Dynamics. 4:101-145. Wunsch, C. 1975. Internal tides in the ocean. and Space Physics. 13:167-182. Reviews of Geophysics BIOGRAPHICAL NOTE Paul May was born on June spent most of his younger years 14, in 1950 in Chicago, the suburb of Illinois. Hinsdale. Paul graduated from high school in 1968 with high scholastic honors. attending Southern Missionary four years, class of in in Collegedale, 1972. As a graduate Sciences coming to at student, he enrolled the University Woods Hole of in for of his the Department Chicago Oceanographic After Tennessee he graduated as a physics major and as president Geophysical before College He for one year Institution and Massachusetts Institute of Technology. Paul is interested in a variety of outdoor activities, including back-packing, scuba-diving, photographer. He Falmouth, months!). is Massachusetts married and to skiing. the and will have former a child He is Sandra in also an Watkins the near avid of E. future (5