A finite element model for tides and resonance ... - Princeton University

5 downloads 1 Views 1MB Size Report
not felt that more accurate calculations [e.g., Francis and. Mazzega, 1990; Hendershott ...... We thank Bill Crawford, Mike Woodward,. Rick Thomson, and Bodo ...

JOURNAL

OF GEOPHYSICAL

A Finite

RESEARCH,

Element

VOL. 98, NO. C2, PAGES 2509-2531, FEBRUARY

Model

15, 1993

for Tides and Resonance

Along the North Coast of British Columbia M. G. G. FOREMAN

AND R. F. HENRY

Institute of Ocean Sciences, Sidney, British Columbia, Canada

R. A. WALTERS U.S. Geological Survey, Tacoma, Washington

V. A. BALLANTYNE Institute of Ocean Sciences, Sidney, British Columbia, Canada A finite element, barotropic, tidal model is developed for the north coast of British Columbia.

The modelis run with eight tidal constituentsand the resultsare comparedwith the Flather (1987) finite difference model, and with extensive tide gauge and current meter observations. Although the tidal potential, Earth tide, and loading tide are included in the forcing, their inclusion is shown

to changethe largestM2 amplitudesby only 2.5% and the largest K1 amplitudes by lessthan 1%. Root mean square differencesbetween observed and calculated sea level amplitudes and phases are within 1.9 cm and 2.9ø for all but one constituent, but the model currents do not in general, compare as favourably. The barotropic currents observedin Hecate Strait are reproduced well, but elsewhere

evidence

is shown

that

model

inaccuracies

are due to baroclinic

effects.

Tidal

residual

currents calculated by the model suggest the existence of eddies off the tip of Cape St. James, Cape Chacon, and around Goose Island and Learmonth Banks. The shallow water constituents in Hecate Strait are shown to have significant contributions from the constructive interference of signals propagating into Dixon Entrance and Queen Charlotte Sound. Using the model, the longest resonant period of the system is estimated to be 7.6 hours with an energy dissipation parameter, Q, of 9.5.

1.

INTRODUCTION

and 2). These islandsare separatedfrom the mainland

This paper describesthe developmentand validation of a barotropic tidal model for the northern coastal waters of British Columbia. This work is part of a project

by three bodies of water: Dixon Entrance, Hecate Strait, and Queen Charlotte Sound. Whereas Dixon Entrance is approximately 150 km in length, 55-65 km in width and

has depthsas large as 400 m, the other two inland waters (henceforthPERD) to study surfacecurrentsin Queen are not so regular. Hecate Strait is very shallow (less Charlotte Sound,Hecate Strait, and Dixon Entrance (see than 30 m) on its western half and has only a narrow Figures1 and 2). Exploratoryoffshoredrillinghasindicated north-south trench with depths greater than 100 m on the possibilityof oil reservesin the region and this project its eastern side. Queen Charlotte Sound is characterized

funded by the Panel of Energy Researchand Development

aims to establish a basic understandingof surface flows in the event of an oil spill. In addition to the tidal model to be describedshortly, a wind-driven model has also been

developedfor the sameregion [Hannahet al., 1991], and

by three relatively deep troughs that separate Cape St. James, Middle Bank, Goose Island Bank, and the Scott Islands. Other notable features of the region are the highly irregular mainland coastline comprising numerous narrow passesand fiords, the narrow continental shelf off the west coast of the Queen Charlotte Islands, and the

a buoyancy-drivenmodel will follow. The P ERD project also comprisesseveral years of observationalwork which includes current meter and tide gauge deployments, ship- presenceof seamounts(e.g., Union and Bowie) beyond mounted acoustic Doppler current profiler measurements, the continental slope. Detailed descriptionsof the physical of the regionhavebeengivenby Crean[1967], CTD surveys,and the deploymentand trackingof drifters. oceanography Dodimead [1980], Thomson[1981],Freelandet al. [1984], Only the current meter and tide gauge observationswill Crawfordet al. [1988],and Crawfordand Thomson[1991]. be discussedhere as they are used in the validation of the There have been severalprevioustidal models for this retidal model. gion. Bell and Boston[1962,1963]developeda non-rotating The Queen Charlotte Islands are a prominent feature hydraulic model for Hecate Strait, Dixon Entrance, and of the north coast of British Columbia (see Figures i Queen Charlotte Sound that gave a reasonably successful This paper is not subject to U.S. copyright. Published in 1993 representationof the diurnal and semi-diurnaltidal heights. However, their M2 tidal currents had large phase shifts by the American Geophysical Union. relative

Paper number 92JC02470.

to observations

in Dixon

Entrance

and inaccurate

residual currents near the forcing boundaries. The first nu2509

FOaEMAN ET AL.' TIDESANDRESONANCE, NORTHCOAST OFBRITISH COLUMBIA

2510

126

130

138

57

Depth contours inmetres

• Learmonth Bank

oo

•)

50

Union --•

49 49

138

130

126

Fig.1. Geography anddepth contours (m)forthenorthcoast ofBritish Columbia andthesouthern coast of Alaska.

modelof tidally-rectifiedcurrentsaround mericalmodel,developed by Kinneyet al. [1976Jto predict three-dimensional an idealized Cape St. James. Their resultscompared oil spill movementsfor the Kitimat PipelineProject, was a M2 barotropicmodel with 9.3 km resolution. Perry et well with both current meter observations and satellite

Finally,Bowmanet al. [1992]useda model al. [1983]usedcurrentspredicted by this modelto study photographs. that simplified the geometry of DixonEntranceandHecate Strait in order to study tidally-rectified currents(the Rose Flather[1987](henceforth F87) developed a finite dif-

tidal frontsand the plankton distributionin HecateStrait.

ference model with 18 km resolution for the region north Spit Eddy) in Dixon Entrance. All the foregoing modelsusedfinitedifference techniques of 45ø and east of 136ø. He included only constituents

of the coastline M:• and Ki and obtained generally excellentagreement and a regulargrid. Giventhe complexity with the elevationamplitudesand phasesmeasuredat 63 and bathymetryand the scalesof variationin the tidal

to followthe example of tide gaugelocations.However,no verification of the tidal flows(e.g.,seeF87), we decided FW) and employ currentswaspresented.Thomsonand Wilson[1987]used Foremanand Walters[1990](henceforth for this region. As explainedin the results of Flather's model to force a finer resolution finite elementtechniques

FOREMANET AL.' TIDES AND RESONANCE, NORTH COASTOF BRITISH COLUMBIA 130

136

,

56

2511 126

,

I

i

56

ß Current Meter

ß TideGauge

ß DIXON ,

ß

ß

"•

ßß

ß

ß

ß

-•

"d•/-

ß

Rose

I[ sp.



) ß

HECATE

BRITISH

ß,',

COLUMBIA

•• ßQUEEN ß

CHARLOTTE

CapeSt. Jamesß

ß ß SOUND i

i

Scottoo Islands ß ß

ß

VANCOUVE 1 ISLAND

50--

5O

ß

49

'

'

'

'

'

I

136

I



i

130

49 126

Fig. 2. Place names,and tide gauge (square),and current meter (circle) observationsites for evaluatingthe performanceof the tidal model. Tide gaugesitesmarkedI (QC14) and 2 (Cape Ball) are referencedin the text. FW, a triangular grid permits more flexibility in reducing numerical errors and providing a better fit of an irregular coastline and bathymetry with elements of an arbitrary

size, shape, and orientation (within certain regularity constraints). In particular, this means that smaller

observations. In section 4, two shallow water constituents are examined

and the model is modified

results are summarized

elements can be placed in regions where the elevation or velocity field is changing rapidly and higher accuracy is desired.

This paper is organized as follows. In section 2, we briefly describe the numerical technique and the triangular grid. In section 3, tidal height and current results are presentedfor eight astronomicalconstituentsand compared with both F87, and extensive tide gauge and current meter

to find the resonant

frequency of the Queen Charlotte Sound, Hecate Strait, and Dixon Entrance system. Finally, in section 5, the

2.

and future

work is outlined.

NUMERICAL TECHNIQUE

Tidal heights and currents were calculated with a finite element

model

that

solves

the

two-dimensional

shallow

water equations:

oq--• + •7-[(H+ r/)u]-- 0

(1)

2512

FOREMAN ET AL.' TIDESANDRESONANCE, NOR• COASTov BRITISHCOLUMBIA culatedby Wahr [1981]and summarizedin Table 1. These ulul (2) values are recommended by the TOPEX/Poseidon Satellite +. _ AhV2U Tidal Committee(P. Woodworth,personalcommunication,

where (x,y) are eastwardand northwardspatialcoordi- 1992),and showthat the P1 and K1 h valuesare affected nates,t is time, r/(x,y, t) is surfaceelevation,u(x, y, t) is by the narrow "freecorenutation"(or "nearly-diurnalfree horizontalvelocity,H(x,y) is depth, f(y) is the Coriolis wobble")resonance at approximately 0.0419cycles/hour. parameter,g is gravity,• is the bottomfrictioncoefficient, The equilibrium tide for constituent n, namely •n, is Ah is the horizontalviscosityparameter,{(x,y,t) is the expressed as [Cartwright, 1978;Schwiderski, 1978] equilibriumtide, and a, •3 are parametersto accoun•for ½.= co?cos(.t + 2x+ v.) (4) the tidal potential,Earth tide, and loadingtide. As in FW, we set •- 0.003 thoughout this study. (An experiment with • = 0.004 scarcelychangedthe results from those reported herein.) The valuesfor a, •3, and Ah will be

for semi-diurnals

- Hn sin2•bcos(wnt+ X -t-Vn)

(5)

discussedshortly. for diurnals, where the amplitudesHn are given in Table The numerical technique is virtually the same harmonic 1; (•b,X) are the latitude and east longitude;•vn is the

wave equation method [Walters, 1986, 1987] as was frequency;and Vn is the astronomicalargument. In used by FW for their model of the southwestcoast of particular, the Vn values are the same as those in the Vancouver Island. The only differencesare the inclusion boundary forcing and are calculatedfor the arbitrary time of tidal potential forcing and sea bed deformationdue 0000 UT, August 1, 1984. to the Earth and loading tides, and a true variation of f with latitude.

Although the governing equations are

not expressedin sphericalcoordinates,the grid units are measuredwith respectto a Universal TransverseMercator

TABLE 1. Love Numbersand Tidal Potential Amplitudes

(UTM) projection[McDonnell,1979]. This projectionis widely usedfor mappingand military purposes.Although the longitudinalrange of our grid (from approximately

127øEto 139øE)is largerthan the 6ø width of standard UTM zones,choosingthe centralmeridianto be 133ø producesonly the smallrangeof 0.9996to 1.002074in the scalingfactor. Thus our grid distancesand elementareas are very closeto their true values. The eight tidal constituentsM2, S2, N2, K2, K1, O1,

P1, and Q1 are includedin the boundaryand astronomical forcing, howeverthe presenceof nonlinearterms in the governingequationsmeansthat a residualtide and several compoundtidesand overtideswill alsobe generatedwithin the model domain. At Prince Rupert (seeFigure 2), the eight forcingconstituentsaccountfor 89% of the tidal heightrange in the daily and half-daily frequencybands. Boundaryconditionsfor the tidal model are zero flow normal to the coast and specified elevations along the

Constituent

k

h

Hn

Q1

0.298

0.603

0.01940

O1 P1 K1 N2 M2 S2 K2

0.298 0.287 0.256 0.302 0.302 0.302 0.302

0.603 0.581 0.520 0.609 0.609 0.609 0.609

0.10128 0.04713 0.14246 0.04674 0.24408 0.11355 0.03090

In accordance with Pekerisand Accad[1969],the ocean loading tide (the deformationof the elastic earth due to the redistribution of mass in the ocean tide) is assumed to be in phase with the ocean tide and a fixed percentage of that value. Following the recommendationof Ray and Sanchez[1989] and the TOPEX/Poseidon Satellite Tidal Committee(P. Woodworth,personalcommunication, 1992), we set a- 0.940,0.953 for the diurnal and semidiurnal constituents, respectively. As the load tide is

open sea. Theseelevationswere computedpredominantly expected to have a relatively minor influence, it was from observations with the tidal harmonics from Bowie not felt that more accuratecalculations[e.g., Francisand and Union Seamounts(Figure 1) playing a critical role Mazzega,1990; Hendershott,1972]werenecessary. along the westernboundary. Values for the northern and The triangular grid for the l•nite element model was northwestern boundaries, where there are no observations, created using the Henry and Walters [1992] packageof wereinitiallyestimatedfromF87 andthe Schwiderski [1978] interactive computer programs and is shown in Figure 3. world tidal model, and then refined(via trial and error) It has 4893 nodes, 8004 elements, and triangle sides that in accordance with the ensuing model results in Dixon vary from approximately 35 km in the deep ocean to about Entrance. This boundary condition tuning may be viewed 1.2 km off Cape St. James. The grid generally representsa as a crude form of the inverse methods approach used by compromiseamongthe followingdesignrequirements:(1)

Bennettand Mcintosh[1982]. closefitting of the coastline,(2) near-equilateralelement In addition to boundary forcing, the effectsof the tide shape;(3) an elementsizethat decreases with rapid changes generatingpotential, the Earth tide, and ocean loading in the velocity and elevation fields.

tide are also included in the model. Assuming the Earth

In addition to a generally finer resolution, the model

to be elastic,the first two forcingterms are usually(e.g., used in this study has other features that should lead to F87, Schwiderski[1978])expressedas f]•, where more accurate results than were described in F87. They

f] -- 1 -t-k- h

(3)

are as follows.

1.

More

constituents

are included

in the

simulation.

This will not only give a more completepicture of the total (andassociated perturbations) to the potential.Although tide but will also improve the nonlinear representation the values k- 0.3 and h- 0.61 are often assumed for all of bottom friction and advection, as more constituent

and k, h are Lovenumbersrelatingthe body Earth tide

constituents, we usedthe frequencydependent valuescal-

interactions

will be included.

FOREMANET AL.: TIDES ANDRESONANCE, NORTHCOASTOF BRITISHCOLUMBIA

2513

Fig. 3. Triangular grid for the finite element model. Numbered open boundaries are referenced in the text.

2. More accurateLove numbers(particularlyfor K1)

3.

MODEL

EVALUATIONS

FOR THE MAJOR

CONSTITUENTS

are used in the tidal potential and Earth tide calculations.

Also an oceanloading tide is now included as an additional forcing. 3. The specified boundary conditions have been im-

The model was evaluated by comparing its results with observations at the 63 tide gauge and 67 current meter sites shown in Figure 2. Actually, observations exist at proved. ObservationssinceF87 and a careful re-analysis more sites, but for reliability, only those of at least 50 of the Bowie Seamount record that corrected for a timdays duration were chosen. The set of tide gauge sites is ing problem have enabled more accurate values to be about double the number used for the same region in F87. specified along the offshore boundaries. Whereas Flather For tidal heights, the amplitudes and phases of the eight

obtainedhis boundaryconditionsfrom $chwiderski[1978], major constituentswere calculatedusing Foreman's[1977] the boundary conditionsused in this model were calculated mainly from observations.

harmonic analysis programs and compared at each site.

For tidal currents,the ellipseparameters(major semi-axis,

FOREMAN ET AL.: TIDESANDRESONANCE, NORTHCOASTOFBRITISHCOLUMBIA

2514

minor semi-axis, angle of inclination, and Greenwich phase

lag) were calculated[Foreman,1978] and comparedat

AlaskanPanhandlein Figure 4b arisebecausethe northern boundaryelevationsare not perfectly compatiblewith the

each site. Observedbarotropic currents were calculated by

northwardpropagatingdiurnal shelfwave. Flather (F87)

weightedaveragesat a minimumof two depths(unlessthe water depth was lessthan 100 m), with the weightsbeing

had similar problemsin the sameregion. Though co-amplitudeand co-phaseplots for the remaining six constituentsare not shown, constituents$2, N2, and K2 havecharacteristics similarto M2; whereasO1, P1, and Q1 are similar to K1. In particular, $2, N2, and K2 have amplitudesthat are approximately30, 20, and 8% of

proportional to the vertical separation between the current

meters. For both tidal elevations and currents, model values were calculated by interpolating from the nodes to the exact observation

location.

As discussedand illustrated in FW, it shouldbe pointed

the M2 amplitudes; and phase lags that are approximately

out that the observed harmonics are only estimates of the true tides. Thus discrepancies between the model

30ø, -25 ø, and 25ø greater. Analogously, the Ox, P•, and

results and the observations

are due to both uncertainties

in the observed harmonics, and inaccuracies in the model. In addition to measurement errors, uncertainities in the observed harmonics can arise from both real influences, such as seasonal changes due to the nonlinear interaction between tides and wind- or buoyancy-driven currents, and from statistical factors such as the length of the time series and the number

of constituents

included

in the harmonic

analysis. The harmonic analysis provides estimates of the latter

uncertainties

in the form of standard

deviations

for all

the cosineand sine (constituent)coefficients.For example,

Qx amplitude ratios with respect to K1 are approximately

60, 31, and 11%, and the phaselags are 16ø, 4ø, and 23ø smaller.

It is interesting to also compare our M2 and K• results with those in F87. Since Flather compared observations

with the nearestgrid points (rather than interpolatingto the precisetide gaugesites),we haveattemptedto provide a fair comparison by interpolating our model results to those same Flather locations. Employing our set of

observedharmonics(which in somecasesare more recent than Flather's), rms errors for the 34 sites common to both

models

are shown in Table

3b.

The

results

from

our

the one year analysis of elevations at Beauchemin Channel model are seen to be consistently more accurate for both and the four month analysis at Cape Ball respectively constituents. estimated these standard deviations to be about 0.16 cm It is also interesting to compare the effects of the tidal and 0.43 cm for all constituents. As these estimates are potential, Earth tide, and loading tide with those noted in virtually identical for all constituents, they are not the F87. This was done by repeating all model calculationswith best measures of variability. A better indication, given a---- 1 and/• = 0 in equation(1). WhereasFlather found a sufficiently long time series, is obtained by performing that the tidal potential, Earth and loading tides increased a seriesof short (e.g., monthly) analyses(with inference) the maximum M2 elevation amplitude in Hecate Strait by and calculating averages and standard deviations. This about 5% and increasedthe phase by approximately2ø, technique was used with three time series in FW, but it we found the inclusion of these terms only increased the

has not been appliedhere. See Foremanand Henry [1989] amplitudes by about 2.5% and increasedthe phasesby for further

discussion

of the

confidence

limits

associated

with the harmonic tidal analysis of time series. Table 2 shows observed and calculated M2 and K1 elevation amplitudes and phases for the 63 sites shown in Figure 2. Differences between the two sets of harmonic constants are calculated as distances in the complex plane; that is,

about 0.5 ø.

However, as in F87, we found the effects of these extra

forcingterms on Kx (and all diurnal constituents)to be very small. When the tidal potential, and Earth and loading tides were removed from the model forcing, the Kx amplitudes decreased by a maximum of 0.4 cm in Hecate Strait, and the phases decreased by a maximum of about

D ---{(Aocosgo-Am cosgm)2•-(Aosingo-Amsingm)2} 1/2, 1ø, both alongthe AlaskanPanhandlejust north of Dixon Entrance and in eastern Hecate Strait. As concluded by where Ao, Am, go, and gm are the observedand modelled amplitudes and phases. The largest M2 differencesoccur at Bella Coola, Kitimat, Kemano Bay and Queen Charlotte City; sites at the end of fiords or channels where, presumably, the model resolution is inadequate and the

Flather (F87), the tide-generatingpotential and the body and Earth tides should be included to obtain high accuracy for the semi-diurnal constituents, but their presenceis not essential

for the diurnals.

Figures 5a and 5b show the major semi-axis magnitudes

true dissipation(via bottom and side-wall friction, and for M2 and Kx. Theseplots havethe samefeatures(but in the generationof internal waves over sills) has not been greaterdetail) as Figure 4 in F87. In particular,noticethe accurately represented by the model. Table 3a shows root

meansquare(rms) elevationdifferences betweenmodeland observationfor each of the eight major constituents. Notice that with the exception of M2, all the rms amplitude differences are within 1.9 cm, and with the exception of

Qx, all the rms phasedifferencesare within 2.9ø. The M2 and K• model results are also displayed in Figures 4a and 4b. They are seen to have the same features as the analogous plots in F87. Most notable are the large M2 amplitude increases eastward in Dixon Entrance and northward in Hecate Strait. K1 is also seen to have an amplitude increase in the same region, but to a lesserdegree than M2. The irregular patterns along the

large M2 values off Rose Spit, Cape St. James, Cape Scott, the mid-western shoreline of Hecate Strait, and around Goose Island and Learmonth Banks. Observed vertically averaged major semi-axes are also shown in Figure 5 and are generally seen to be in good agreement with the model results. Notable exceptions for M2 are the large observed currents off Cape Chacon and on the continental slope to the west of Dixon Entrance; and the large model currents at the most

western

site in the central

Hecate

Strait

line

and the two sites nearest Rose Spit. Agreement for K1 is not as good; however, to some degree this may be a problem with the observations. The Kx confidence limits are proportionately much larger than those for M2 and

FOREMANET AL.: TIDES ANDRESONANCE, NORTHCOASTOF BRITISHCOLUMBIA

2515

TABLE 2. Observed and CalculatedM2 and K• ElevationAmplitudes(Amp.) and Phasesfor the 63 SitesShownin Figure2 M2

Site Dixon Entrance Brooks Pen. S. Brooks Pen. N. Kyuquot Hunt Islets Q05 QC10 QC12 QC14 QC15 Hakai Passage Beauchemin Ch. Hecate Sandspit Welcome Hat. Foggy Bay McLean Arm Cape Chacon Cape Muzon Security Cove Houston Stewart Dawson Harbour Nesto Inlet Heater Harbour Atli Inlet W05 Cape Ball HEC7 Dixon Rose Spit Egeria Bay Langara Island Union Seamout

Alert Bay Port Hardy Cape Scott Egg Island Wadhams Bella Coola Bella Bella Milne Island McKenney Is. Gillen Harbour Kitimat Kemano Bay Qlawdzeet Brundige Inlet Prince Rupert Port Simpson Tasu Sound Copper Island Q.C. City Wiah Point Langara Island Ketchikan Davis River Bowie Seamt. QC13 Nakat Harbour Kelp Island Nichols Bay View Cove G01 G05 Q06

Latitude

Longitude

54.562 49.837 50.267 50.217 50.473 51.367 51.847 51.005 51.053 51.120 51.690 52.782 53.275 54.023 54.935 54.793 54.667 54.640 54.747 52.077 53.163 53.557 52.120 52.713 53.173 53.717 54.122 54.260 54.225 54.242 49.583 50.583 50.717 50.783 51.250 51.517 52.383 52.167 52.617 52.650 52.967 53.983 53.517 54.217 54.617 54.317 54.567 52.750 52.350 53.250 54.117 54.250 55.333 55.767 53.317 51.680 54.820 54.877 54.717 55.083 51-.600 51.363 51.483

132.937 127.943 128.352 127.353 128.317 130.017 129.578 128.698 129.438 127.945 128.107 129.298 130.982 130.613 130.960 131.968 132.025 132.688 132.845 131.123 132.458 132.925 131.033 131.577 131.275 131.733 131.024 131.950 132.965 133.163 132.783 126.933 127.483 128.417 127.833 127.517 126.800 128.133 128.767 129.483 129.600 128.700 128.117 130.767 130.850 130.333 130.433 132.033 131.167 132.067 132.317 133.050 131.633 130.167 135.633 130.565 130.702 131.300 132.133 133.017 128.883 128.915 130.483

K1

Observed Amp. Phase

Model D, Amp. Phase cm

128.2 98.5 98.3 99.0 101.1 107.9 120.1 113.2 106.8 123.1 126.0 145.3 167.8 190.7 182.9 170.9 162.9 137.3 125.7 103.4 103.1 105.2 119.4 151.4 167.3 191.5 186.0 160.0 125.3 119.1 91.2

268.7 242.0 244.0 241.4 243.6 250.5 251.9 245.9 248.1 247.1 248.0 253.9 263.0 264.8 268.1 269.9 270.3 270.2 270.9 252.8 259.8 262.1 264.3 267.6 266.6 265.2 264.1 263.6 263.0 260.9 251.5

128.7 98.2 100.5 100.0 101.5 106.5 119.4 111.1 105.6 120.0 124.2 145.3 168.6 192.6 183.6 168.6 161.3 136.8 125.4 102.1 104.2 105.7 117.3 151.6 168.1 194.5 187.6 158.7 130.9 119.8 89.5

127.1 133.0 108.0 118.1 124.7 142.6 129.7 128.2 137.8 144.6

259.5 251.9 244.2 249.0 249.0 252.6 249.8 250.7 255.1 255.8

131.6 132.9 106.4 120.0 123.7 147.4 133.0 138.5 140.5 148.1

165.0 172.5 183.8 179.4 195.1 186.2 101.9 131.7 197.4 146.5 130.6 187.4 203.8 100.9 114.4 181.3 177.6 163.6 152.2 119.0 115.0 111.7

257.8 258.5 267.0 270.0 267.4 266.9 257.9 267.5 271.5 263.0 262.8 270.3 268.4 265.9 257.7 271.7 270.6 269.8 272.1 249.0 247.4 256.8

160.9 162.2 189.7 184.8 197.5 190.4 102.3 132.0 206.5 150.2 125.0 180.3 206.0 99.5 106.9 186.5 178.0 156.8 138.7 118.5 113.8 111.5

268.3 241.9 244.4 240.0 244.7 251.4 251.9 246.1 248.6 246.6 246.9 252.5 263.4 264.5 270.7 271.4 270.7 270.9 271.4 256.9 260.2 261.6 264.9 268.6 266.7 266.8 264.8 264.3 261.6 262.3 251.8 261.1 252.7 243.3 246.6 246.2 247.3 248.4 250.6 253.6 254.8 253.6 253.3 265.7 267.9 266.2 268.0 258.7 266.9 274.2 261.8 261.6 271.8 268.7 266.2 254.8 269.2 270.3 271.4 272.9 249.1 248.2 256.2

1.1 0.3 2.3 2.6 1.9 2.2 0.6 2.1 1.6 3.3 2.9 3.6 1.5 2.3 8.3 5.0 2.0 1.8 1.2 7.4 1.3 1.0 2.4 2.7 0.8 6.1 2.7 2.4 6.5 3.0 1.8 5.8 1.9 2.3 5.4 6.1 14.3 4.6 10.3 4.6 4.4 12.8 18.4 7.2 8.6 4.7 5.7 1.5 1.4 13.2 4.7 6.2 8.5 2.5 1.5 9.3 9.4 1.0 8.1 13.7 0.6 2.1 1.3

Observed Amp. Phase 48.1 44.8 45.2 47.9 43.5 44.5 46.9 45.5 45.2 46.5 45.9 47.4 49.7 51.7 50.5 50.2 49.6 47.8 46.4 43.6 44.9 48.0 44.4 47.9 48.3 49.7 50.8 48.9 45.3 46.5 43.3 51.6 49.9 44.3 44.6 46.7 48.2 45.8 43.6 44.5 45.6 48.5 48.8 49.4 48.7 50.7 50.5 43.7 44.1 51.0 45.9 49.1 50.7 51.2 44.6 49.6 48.6 48.9 47.2 47.3 44.9 44.5 45.8

260.6 243.7 244.8 245.8 247.0 249.2 250.0 246.8 246.9 248.2 250.1 252.8 258.8 258.4 260.2 260.8 261.4 261.5 261.1 247.6 252.7 253.5 257.7 259.6 261.0 260.3 258.9 258.6 257.1 257.2 248.9 256.6 249.8 246.4 251.8 251.1 253.0 251.7 253.7 253.9 256.1 254.9 255.4 259.4 260.6 259.4 259.3 256.0 259.8 261.9 258.1 258.0 261.6 260.2 256.9 248.6 262.3 261.1 257.0 262.5 251.5 250.5 253.4

Amplitudes arein centimeters, andphases arein degrees, UT. D isdefined byequation (6).

Model D, Amp. Phase cm 47.1 44.5 44.9 44.9 45.2 47.6 47.7 45.9 45.9 46.4 45.5 47.4 49.2 52.8 52.2 49.4 49.0 47.5 46.8 46.5 46.5 46.9 44.7 47.4 48.4 50.4 51.3 49.3 48.0 47.2 42.0 51.2 50.3 45.7 45.9 46.1 48.0 46.1 46.6 47.0 47.1 48.9 48.6 52.3 51.3 53.4 51.9 46.4 45.7 51.4 49.0 47.8 49.6 53.1 42.9 47.2 52.0 50.1 48.7 48.4 45.7 45.4 46.9

260.6 244.9 245.6 244.1 245.6 248.0 250.1 247.1 246.4 249.0 249.4 249.4 256.6 257.3 261.7 262.2 262.1 261.9 261.8 251.7 255.2 256.5 255.5 258.8 258.7 260.8 258.5 259.8 258.3 258.1 249.8 256.7 250.9 245.7 249.7 249.9 250.2 250.2 250.8 251.2 251.5 252.5 251.4 259.1 260.9 259.0 260.8 253.5 257.1 264.1 258.7 257.7 263.1 261.3 256.7 250.9 262.0 261.7 262.3 263.6 249.2 247.8 251.7

1.0 1.0 0.7 3.3 2.0 3.2 0.8 0.4 0.8 0.7 0.7 2.8 1.9 1.5 2.2 1.4 0.8 0.4 0.7 4.3 2.6 2.7 1.7 0.9 2.0 0.8 0.6 1.1 2.9 1.0 1.5 0.4 1.1 1.5 2.1 1.2 2.4 1.2 3.7 3.3 4.0 2.1 3.4 2.9 2.6 2.7 1.9 3.3 2.6 2.0 3.1 1.4 1.8 2.2 1.7 3.1 3.4 1.4 4.7 1.4 2.0 2.3 1.8

FOREMAN ET AL.' TIV•.s ANyRESONANCE, NORTHCOASTOFBRITISHCOLUMBIA

2516

TABLE 3a.

RMS Elevation Differencesat All the Sites

M2

Amp.,cm 4.0 Phase,deg 1.7

K1

$2

O1

N2

P1

some of the observed values do not seem to be consistent

with their neighbours.

Listed in Table 2 and Shown in Figure 2

K2

Q1

1.6 1.9 0.9 1.3 0.7 0.7 0.7 1.9 2.1 1.7 2.2 2.0 2.9 6.0

Table 4 summarizes the rms differences between the

model and observedM2 and K1 currents. As was found with the FW model for the southwestcoast of Vancouver

Island,thesecurrent differences are muchlargerthan the elevation differences. Again this is probably due to a combination of the same three reasons discussed

TABLE 3b. I•MS ElevationDifferences at 34 SitesCommon

to the Flather[1987]modelandThisModel,and Interpolationto the Flather Locations This Model

Amp.,cm Phase,deg

in FW (pages277-278),namely,(1) crudeaveraging of the observationsat specificdepths in order to obtain a

Flather's Model

M2

K1

M2

K1

2.5 1.5

1.6 2.0

3.4 2.9

2.2 2.9

barotropiccurrent;(2) baroclinicprocesses presentin the observations but not represented in the model; and (3) largervariabilityin the observedtidal currents. Grid resolutionmay be an additional reason that was not discussed in FW.

Whereas the current observations are

pointmeasurements thatcanreflect verylocalfeatures in

140

M2 co-ampl]tudes[cm)

and co-phases[OMT)'

Fi•. •. •ode•½o-•mp•i•.d• (½m)(•o•id½..•) •-d •o-ph• (d.•r•. UT) (d•h•d •.r•)

•or(•) •= •.d (•)

FOREMANET AL.' TIDES ANDRESONANCE, NORWHCOASTOF BRITISHCOLUMBIA

2517

b co-ampltudes[cm]

and

co-phases[¸P1T]"

Fig.4 (continued)

the bathymetryand/or coastline,the modelsimulationcan only hope to represent such features if their scale is larger than the grid resolution. Figure 6 servesto illustrate the degree of baroclinicity in four representative regions of the model domain. It shows model and observed M2 and K1 current ellipses versus depth. At site D04 in the middle of Dixon Entrance south of Cape Chacon, the M2 ellipsesare seento rotate clockwise with depth while the K1 ellipsesrotate counterclockwise. Although the model K1 ellipse is a reasonably good representation of the averaged observation, the fact that the shallowest M2 ellipse is fatter than the deeper ellipses

suggeststhe presenceof a larger cross-straitcomponent.As there can be significant stratification in Dixon Entrance in

At site W02 in the center of easternHecate Strait, both constituentsare seento be predominantly barotropic and in good agreement with the model. At site G04 in the trench south of GooseIsland Bank and QC2 off the west coast of the Queen Charlotte Islands, M2 internal tides also seem to be presentas variationsin the ellipseangle of inclination with depth suggesta componentthat is perpendicular to both the bathymetry contours and the inclination of the

model (barotropic)ellipse. The fact that the K1 currents are comparable in magnitude to the M2 currents at QC2 (though the K1 elevation amplitudes are less than half

the M2 amplitudes)is similar to the situation(FW) off Vancouver Island and suggeststhat diurnal shelf waves may exist along the narrow continental shelf to the west

the summermonths(whentheseobservations were made) of the Queen Charlotte Islands. An extensionof Figure due to freshwater input from the Skeena and Nass l•ivers, 5b northwardsuggests(as in F87) that diurnal shelfwaves it likely that this component is an internal tide.

are also present along the Alaskan shelf north of Dixon

\

FOREMAN ET AL.' Tress ANy RESONANCE, NORTI-ICOASTOF BRITISHCOLUMBIA

2518

'2

H2 major

semi-axis

[cm/s]

Fig. 5. Major semi-axismagnitudes(cm/s) for (a) M2 and (b) K1. Asterisk and number denoteobservationsite and observed depth-averaged value.

Entrance; however, there are no current meter observations Figure 4a). As M2 is dominantin this region,this means for confirmation.

that at floodtide, the currents(and thuspassiveparticles)

Figure 7 showsM2 current ellipsesfor northeast Hecate Strait. It is interestingto note that the anglesof inclination

will converge along the northeast coast of the Queen Charlotte Islands, while at ebb tide they will diverge from

convergeto the region of maximal tidal amplitude (see

that

location.

FOREMANET AL.' TIDES ANDRESONANCE, NORTHCOASTOF BRITISHCOLUMBIA

2519

•t.O

o/

0 ß

;

;, /o

ß

1.4

,•,

.,-

'...

.-



,'

.-.

_-. '-"

,

.,

:..• '.•,. ,..•

,"'- ,,.

R ! maj oc sem •- ax • s ( cm/s ) Fig. 5. (continued) TABLE 4. RMS Differences for Ellipse Parameters at All Current Meter Sites Shown in Figure 2

Figure 8 shows the barotropic residual tidal velocities calculated by the model. For this computation, the sea level was set to zero along all eight open boundaries(i.e., ßonly generationwithin the modeldomainis assumed),and

Major semi axis,

Minor semi axis,

inclination,

cm/s

cm/s

deg

deg

M2

7.5

5.3

23.1

26.4

a smalllateralviscosity (Ah= 10m2/s)wasaddedto the

K1

3.2

1.7

24.8

47.5

momentum equations in order to assist convergencein some

S2 O1

3.0 2.1

2.1 0.9

22.1 30.5

31.3 61.1

of the narrow passagesand fiords and in the region around Cape St. James. Clockwise eddies are evident around

phase lag,

M2

K1 Mode 1 I>

Observed :

2

Observed

Mode ]

cm/s

:

b

20

cm/s

(• (•

cI

M2 Mode l

13bserved

Mode ]

Observed

c)



-

2

cm/s

-

20

cm/s

c)

Fig.6. •1•2andK1current ellipses at sites (a)D04inthemiddle ofDixon Entrance south ofCapeChacon, (5) •V02in center ofeastern Hecate Strait,(c)G04in thetrench south ofGoose IslandBank,and(•) (•C2midway alongthecontinental slopewestof the(•ueenCharlotte Islands.Thelinewithineachellipse denotes O, the positionof the currentvectorat the timeof maximumtidal potential,andthe arrowdenotes the sense of rotation.

FOl•EMAN ETAL.' TIDESANDRESONANCE, NOl•THCOAST OFBI•ITISHCOLUMBIA

K1 Mode I

•2 Observed

Mode I

Observed

c

K1 Mode ]

I)

M2 Observed

-

2

cm/s

d Fig.6 (continued)

Mode 1



Observed

-

2

cm/s

2521

2522

FOREMANET AL.: TIDES ANDRESONANCE, NORTHCOASTOF BRITISHCOLUMBIA

M2 current

ellipses

Fig.7. M2 modelcurrent ellipses fornortheast HecateStrait.Thelargest majorsemi-axis is71cm/s.Theline

within eachellipsedenotesG, the currentvectorpositionat the time of maximumtidal potential.Solidline ellipses denoteclockwise rotationof the currentvector,dashed ellipses denotecounterclockwise rotation.

GooseIsland and Learmonth Banks, while counterclockwise the cape and a setup along the coast to north and west eddies are seen off Cape Chacon and south of Learmonth of the cape. The former feature is consistent with the

eddy of Figure 8, while the latter feature Bank. (The complicatedflows near Learmonth Bank counterclockise suggest the needfor furthergrid refinement in this region.) would be consistent with a clockwise eddy were it not Strong clockwiseflows are also evident to the northwest landfast. Although it is possible that better resolution of Rose Spit, but they do not closeto form an eddy. In of the coastline and the narrow continental shelf might addition, there is no evidenceof the large, counterclockwise, produce a clockwise eddy consistent with Thomson and RoseSpit Eddy studiedby Bowmanet al. [1992].A strong Wilson[1987],both stratificationand a three-dimensional

counterclockwise eddy (the maximumspeedis 69 cm/s)

model are probably necessaryto accurately reproduce both the Cape St. James Eddy, and the Rose Spit Eddy. The final interesting feature of Figure 8 is a southward there is no evidence of the clockwise eddy west of the cape that was calculatedand observedby Thomsonand residual flow along the western side of Hecate Strait that Wilson[1987]. However,the elevationcontoursassociated is accompanied by a maximum setup along the coast of with these residual flows show both a depressioneast of approximately 2 cm.

is also predicted to the east of Cape St.

James but

FOREMANET AL.' TIDES ANDR•so•^•c•,

NORTHCOASTOF BRITISHCOLUMBIA

-•

-

2523

10.0

.

ß

o

.

ß

ß

o

,

, .

..

.

. o

ß

Fig. 8. Barotropic residual tidal velocities.

.

cm/s

2524 4.

FOREMANET AL.: TIDESANDRESONANCE, NORTHCOASTOF BRITISHCOLUMBIA SHALLOW-WATER CONSTITUENTS AND RESONANCE

Figures 4a and 4b show that the amplitude increase northward

in Hecate

Strait

and eastward

in Dixon

Entrance

is greater for M2 than K1. These results lead to the natural question: What is the resonant frequency of the system?

The

tidal

model

confirms

these

conclusions.

When

run with radiation (normal flow) conditionson all eight sea boundaries(see Figure 3) except number I where the elevation is set to zero (it is necessaryto specify at least one boundary elevation in order to avoid an

ill-conditioned problem), the M4 phases indicated the

ChoosingQC14 (becauseit was the longesttime seriesof

expected outward radiation from northwest Hecate Strait all the tide gauges at the entrance to Queen Charlotte and the M4 amplitudes were reasonably accurate. For Sound), and Cape Ball (becauseit was situated closeto example, the amplitudes at Cape Ball and QC14 were 4.3 the largest M2 amplitude in Hecate Strait) as reference and 0.3 cm, respectively. However, the $K3 model solution sites, Table 5 shows the observed amplitude ratios and did not agree with the observations as it calculated a value phasedifferencesfor severaltidal constituents.(Both these of only 0.6 cm for Cape Ball and phases that increase sites are shown in Figure 2.) A resonantfrequencyin away from that general region. In order to confirm that the terdiurnal or quarter-diurnal frequency band is clearly reasonably accurate $K3 values could be obtained, the suggested. model simulation was repeated with specified elevations However the situation is complicated by the fact that alongboundaries3 and 4 (seeFigure3) that are consistent these so-called shallow water constituents also have conwith the observations from Bowie and Union Seamounts, tributions that arise from their local nonlinear generation. and radiation conditions on the other six outer boundaries. For example, the relatively large signals for SK3 and M4 As shown in Figure 9c, the $K3 amplitudes and phasesat at Cape Ball will have contributions from the nonlinear both QC14 and Cape Ball now agree reasonably well with interactions(primarily via the secondterms in equations the observed values, and the phases indicate a net energy (1) and (2)) of $2 and K1, and M2 and M2, respectively. propagation into Hecate Strait. Figures 9a and 9b show the observed amplitudes and Having demonstrated that the values in Table 5 should phases at selected sites for these two constituents. As the not be used to estimate the resonant frequencies of the smaller amplitudes are comparable to the background noise region, accurate estimates were then calculated by making level, it is not surprising that the harmonic analysis has the following modifications to the finite element model. produced some values that appear to be inconsistent with The tidal potential and Earth and loading tide forcing neighbouring values and trends. However, general patterns were turned off, and periodic forcing with frequency c•0 are apparent. For both constituents, the amplitudes was specifiedon boundaries3 and 4 (see Figure 3). The increase uniformly to maxima in the northwest region of amplitude and phase of this forcing varied linearly from Hecate Strait. However, whereas the M4 phases increase 10 cm and 90 ø in the southwest to 18 cm and 180 ø in away from that region, suggestingoutward radiation, the the northwest. The associated amplitude ratios and phase $K3 phases increase toward that region and suggestinward differences are close to those chosen for the $K3 simulation radiation. The M4 signal at Cape Ball would therefore illustrated in Figure 9c and represent propagation to the appear to arise largely from local nonlinear generation northwest. All the other open boundaries were given a

(consistentwith both the elevationand velocity gradients normal flow condition (i.e., the computed velocity was apparentin Figures4a and 5a), while the $K3 signalwould restrictedto be normalto the boundary)in order to permit appear to have a significant contribution from •he resonant growth of energy propagating into Hecate Strait. TABLE 5.

Freq.

Constituent

(cph)

the radiation of incident waves.

Following a procedure

similarto that employedby Walters[1988]in SanFrancisco

Tidal Harmonics at Cape Ball and QC14

QC 14

Amp., cm

Cape Ball

Phase

Amp., cm

Phase

Amp.

Phase

Ratio

Diff.

Q1

0.0372

5.30

230.9

6.01

238.5

1.13

7.6

O1 P• K• N2 M2 $2 K2 M03 M3 MK3 $K3 MN4 M4 M$4 2M K5 2MN6

0.0387 0.0416 0.0418 0.0790 0.0805 0.0833 0.0836 0.1192 0.1208 0.1223 0.1251 0.1595 0.1610 0.1638 0.2013 0.2400

27.90 14.02 45.22 22.16 106.80 31.87 8.66 0.12 0.23 0.12 0.22 0.11 0.34 0.22 0.10 0.11

232.7 242.4 246.9 223.7 248.2 274.6 265.7 134.7 309.0 239.3 90.0 100.2 110.4 127.1 130.6 116.4

30.91 15.41 49.72 39.26 191.50 59.60 15.72 1.09 2.20 1.37 2.58 1.91 4.92 2.45 0.39 0.18

242.5 257.2 260.3 240.6 265.2 297.6 292.2 6.3 326.7 19.3 147.3 16.5 40.4 93.9 41.0 19.6

1.11 1.10 1.10 1.77 1.79 1.87 1.95 9.08 9.59 11.42 11.72 17.36 14.47 11.14 3.90 1.64

9.8

Phases are degrees UT.

14.8

13.4 16.9 17.0 23.0

26.5 231.6 17.7

140.1 57.2 276.3 290.0 326.8 270.4 263.2

FOREMANET AL..' TIDES AND RESONANCE, NORTH COASTOF BRITISH COLUMBIA

I .0

2525

149

0.4•114

"Q

)•147 0.!

1

! .9)• .9•154

!32

I .6•1

0.5•E123

0.4•120

0.2•E99

0.8 •98

0.7)•97 0.2•90

0.6•9%

0.2)•52 ß

0.!•67

(]

sk3

observed

amplitudes

and

phases

Fig. 9. Observed amplitudes(cm) andphases(degrees, UT) at selecttide gaugesites(asterisk)for (a) SK3 and (b) M4. Amplitudes are shownto the left of the siteand phasesto the right. (c) Modelco-amplitudes (cm) (solid)andco-phases (degrees, UT) (dashed)for $K3 with nonlineargeneration andspecified elevations alongthe western and northwestern outer boundaries. Asterisk and associated numbers denote select tide gauge sites and observed amplitudes and phases.

Bay, coowas then steppedthrough a wide range of values geometry is not that of a classicalharbor. In this case, and the amplitude ratio and phase differencesbetween resonanceoccurs when the wave travelling northward into Cape Ball and QC14 were calculated. (It should be QueenCharlotte Soundand the wavetravellingeastwardin mentioned that this response-versus-frequency procedure is much easier to explore with a harmonic model than a

time-steppingmodel.) The resultsare shownin Figure 10. Although a classical quarter-wavelength resonance is evident in Figure 10 when coo= 0.132 cycles/hour,the

Dixon Entrance, meet in phase along the western shoreline of Hecate Strait. Figure 11 showsthe co-amplitudesand co-phasesfor this resonant frequency. Notice the phase shifts of 90 ø •om

the entrances of Hecate Strait

and

Dixon Entrance to Cape Ball, and a factor of 14 increase

2526

FOREMANET AL.: TIDES ANDRESONANCE, NORTHCOASTOF BRITISHCOLUMBIA

265

0.6•226

0.6 •1•4

1.0•3



1.3

41

t.5•3

O.

ß2 •(268

2.6•(76

I .8•9

0.5 •(108

0.7•(113

I .5•94

0.7•(74 ••'•

0.5•(147 0.3•110

0.2•213 b

m4 observed

amp ] i tudes

and

phases

Fig.9 (continued)

in amplitude from QC14 to Cape Ball.

Figure 10 also

The

Table

5 observed amplitude

ratios and phase

shows3/4-wavelengthresonance at coo= 0.255cycles/hour, differences for individual tidal constituents are also shown and an interestingpeak near 0.310 cycles/hour. A closer in Figure 10. Considering the simple boundary forcing, inspection of the latter reveals amphidromes east of both Rose Spit and Cape St. James, and amplitude growth both near Cape Ball and to the northeast of Vancouver Island. Cape Ball amplitudes at these frequencies are about 120 cm and 130 cm, respectively. Were Figure 10 continued for larger frequencies,shorter wavelengthresonanceswould also appear.

the diurnal and semi-diurnal constituents Q1, O1, P1, K1, N2, M2, 52, and K2 are seen to agree reasonably well with the model curves. Amplitude ratios for the terdiurnal constituents also agree reasonably well but the observed quarter-diurnal amplitude ratios are seen to be much larger than predicted by the model. As was demonstrated with $K3 and M4, this is because the

FOREMANET AL.' TIDES ANDRESONANCE, NORTHCOASTOF BRITISHCOLUMBIA

2527

ß ,

i

ß

ß

ß

,

ß ß

,



--. /0.2

,'•90

..' ,"-

'

ß

SK3 co-amplitudes[cm)

and

co-phases[O-MT]

Fig.9 (continued) energy associated with these shallow water constituents does not arise solely through inward radiation from the deep ocean. In particular, the extent to which the large observed amplitude ratios for MN4, M4, and M$4 are not reproduced by the model is an indication of the local nonlinear generation of these constituents near Cape Ball.

thatthe$K3 signal is largerthantheMK3 signal.As the M2 elevations and velocities are approximately three times larger than their $2 counterparts, we would expect approximately the same proportionality for MK3 and $K3 if nonlinear generationwere the only energy source. It should be noted that although all eight major tidal

A furtherindicationthat the terdiurnalsignalat Cape

constituents

Ball is not primarily due to local nonlinear generation is

that

were included

in the bottom

friction

the forced waves in the resonance calculation

term

so

had an

2528

FOREMANET AL.' TIDESANDRESONANCE, NORTHCOASTOF BRITISHCOLUMBIA

Tidal

Harmonics'

(Z) observed

ampl•_tude



phase

observed

Cape

Ball

rat•_os

d:•fferences

vs.

QC14

model

amplltude

model

phase

ratlos

dlfferences

Or)

/ /

H(

/

I

/

................. .......... / ........................ '.................. ;' I I

/

I

/ /

" /

/ • '••,// \\/ /

/

I

I

I

I l

/

\/'

!I

I

I

I

_• ¸ e-

I

o '•-

I I I



I

o•

I I

! I I !

o('0

/ /

/

..... • .....

! ...............

ii

iI

I

ß

'

/

/ / /

ß

.oo

I

0.05

I

O.lO

I

0.15

I

I

I

I

0.20

0.25

0,30

0.35

0

frequency

Fig.10. Amplitude ratios andphase differences (degrees) between CapeBallandQC14asa function offrequency(cycles/hour). Modelamplitude ratios (solid), model phase differences (dashed), observed amplitude ratios (circles), observed phasedifferences (asterisks).

Defining2•r/Q as the fractionof the energydissipated appropriatedrag, their inclusiononly affectedthe results slightly. Also increasingthe magnitudeof the forcing in onecycleand usingthe sameformulaeas Garrett[1972] producedessentially the samecurvesas thoseshownin for determiningresonancein the Bay of Fundy and Gulf of Figure 10. So, at least for the purposeof resonance,the Maine, we calculatedQ for this system. Using the M3 and system is essentiallylinear.

SK3 observationsof Table 5, Q was computedto be 11.3,

whereasusing the numericalmodel resultsat frequencies for-medwith different values of the amplitudes and phases 0.12 and 0.13 cycles/hour,Q was foundto be 9.5. Unlike the Bay of Fundy and Gulf of Maine region alongthe westernboundariesin order to mimic radiation from another source. For example, when our calculations where the primary resonantfrequencyis very closeto the wererepeatedwith constantvaluesof amplitudeand phase N2 frequency[Garrett, 1972],resonancein Hecate Strait (in order to representradiation from the southwest),the has not been previously noticed because there is little Obviously other resonancecalculationscould be per-

results differed slightly from those shown in Figures 10 and oceanicenergy at the peak frequenciesshownin Figure 10. 11. Thus the direction of wave propagation does influence However,this is not to say that resonancein Hecate Strait resonance.

is not important. With the imminent possibility of an

FOREMAN ET AL.' TIDES ANDRESONANCE, NORTHCOASTOF BRITISHCOLUMBIA

2529

IO0

70•

Resonance

co-ampltudes{cm)

and co-phase

Fig. 11. Co-amplitudes (cm) (solid)andco-phases (degrees) (dashed) for the resonant frequency cv0-- 0.132 cycles/h.SitesQC14 and Cape Ball are denotedby asterisks.

of Vancouver Island and other intermittent earthquakes around the North Pacific Ocean, there is a constant threat

Further investigationsshould therefore be carried out to determine the shorter resonant periods for Hecate Strait and the possibilitythat they might enhancethe danger of

of tsunamis along the west coast of British Columbia.

future

earthquakealongthe Juan de Fuca Plate off the west coast

tsunamis.

2530

FOREMAN ET AL.: TIDESANDRESONANCE, NORTHCOASTOFBRITISHCOLUMBIA 5.

SUMMARY AND DISCUSSION

Bowman,M. J., A. W. Visser,and W. R. Crawford,The Rose

The preceding presentation has detailed the verification of a barotropic tidal model for the northern coastal waters

Spit Eddy in Dixon Entrance: Evidence for its existenceand

underlyingdynamics,Atmos.Ocean,30(1), 70-93, 1992.

Cartwright,D. E., Oceantides. Int. Hydrogr.Rev. Monaco,

60(œ),35-84, 1978. of British Columbia. Although the model has on average, W. R. and R. E. Thomson, Physical oceanography reproduced the elevation amplitudes and phases of the Crawford,

eight largest constituents to within 4.0 cm and 6.0ø when compared with observations at 63 sites, it has not been

as successfulwith the tidal currents. This is primarily because in most regions of the model, the true currents contain significant baroclinic effects that the barotropic model cannot reproduce. The

shallow

water

constituents

in Hecate

Strait

were

shown to have significant contributions from the constructive interferenceof signalspropagating into Dixon Entrance and Queen Charlotte Sound. By forcing the model from

the western boundarieswith a wide range of frequencies, the longestresonantperiod of the systemwas calculatedto be 7.6 hours. This estimate was confirmedwith analyses of observationsfrom tide gaugesin the middle of Queen Charlotte Sound and near the peak responsepoint in Hecate Strait. The energy dissipation parameter Q was estimated to be 9.5 for the primary resonantfrequency. The effects of tidal potential, Earth tide, and loading tide forcing were not found to be significantfor this model. In particular, the inclusion of these forcingsincreasedthe largest M2 and K1 amplitudesin Hecate Strait by 2.5%, and lessthan 1%, respectively. With referenceto the P ERD project goal of providing a better understanding of surfacecurrents in northern British Columbia waters, it is clear that a descriptionof the tidal currentswith a barotropicmodel is not sufficient.Although

of the westernCanadiancontinental shelf,Continental Shelf Res., 11, 669-683, 1991.

Crawford,W. R., W. S. Huggert,and M. J. Woodward, Water transportthroughHecateStrait. Atmos. Ocean,26(3), 301320, 1988.

Crean,P. B., Physicaloceanography of DixonEntrance,British Columbia,Bull. Fish. Res. BoardCan., 156,66 pp., 1967. Dodimead,A. J., A generalreviewof the oceanography of the QueenCharlotteSound-Hecate Strait-DixonEntrance region, Can. Manuscr.Rep. Fish. Aquat.Sci. 157•, 248pp., Dep. of Fish. andOceans, Resour.Serv.Branch,Pac. Biol. Sta., Nanaimo, B.C., 1980.

.Flather,R. A., A tidal modelof the northeastPacific,Atmos. Ocean,25(1), 22-45, 1987.

Foreman, M. G. G., Manualfortidal heightsanalysis andprediction. Pacific Mar. Sci. Rep. 77-10, 101 pp., Inst. of Ocean Sci., Patricia Bay, Sidney,B.C., 1977.

Foreman,M. G. G., Manualfor tidal currentsanalysis andpre-

diction. PacificMar. Sci. Rep. 78-6, 70 pp., Inst. of Ocean Sci., Patricia Bay, Sidney,B.C., 1978.

Foreman, M. G. G., andR. F. Henry,The harmonic analysis of tidal modeltime series,Adv. WaterResour.,12, 109-120, 1989.

Foreman,M. G. G., and R. A. Walters, A finite-elementtidal model for the southwestcoast of Vancouver Island. A tmos. Ocean,28(3), 261-287, 1990. Francis, O., and P. Mazzega, Global charts of the ocean tide

loading effects, J. Geophys. Res.,95(C7),11,411-11,424, 1990. Freeland,H. J., W. R. Crawford,and R. E. Thomson,Currents alongthe Pacificcoastof Canada,Atmos.Ocean,22(2), 151172, 1984.

Garrett, J. G. R., Tidal resonance in the Bay of Fundyand Gulf

of Maine, Nature, 238, 441-443, 1972. observedtidal currentsare predominantlybarotropicin HecateStrait wheremost of the oil leasesare located, Hannah, C. G., P. H. LeBlond, W. R. Crawford, and W. P. Budgell,Wind-drivendepth-averaged circulationin QueenCharevidenceof vertical variationsin other regionsof the

model necessitates a full three-dimensional model that includesdensity variations and mean flow currents. Such a

model(employing techniques similarto thosedescribed by Lynchet al. [1992]and Walters[1992])is presentlyunder developmentand its successwill be reportedin future manuscripts.

lotteSoundandHecateStrait, Atmos.Ocean,29(•), 712-736, 1991.

Hendershott, M. C., The effects of solid earth deformation on

globaloceantides, Geophys.J. R. Astron. Soc.,29, 389-402, 1972.

Henry, R. F., and R. A. Walters, A geometrically-based, automatic generator for irregular triangular networks. Commun. Appl. Numer. Methods,in press,1992.

Kinney, P. J., J. C. H. Mungall, C. E. Abel, R. O. Reid, A. C. Acknowledgments. WethankBill Crawford, MikeWoodward, Vastano,and R. E. Whitaker, Oil spill movementpredictions, RickThomson, andBodode LangeBoomfor helpfulcomments, Kitimat Pipeline Project, Kitimat Pipe Line, Ltd., Termpol submission re marineterminalat Kitimat, B.C., vol. V, aptidal data,and/orharmonicconstants; CharlesHannah,Zygmunt pendix XII, Oil Spill ContingencyPlan, Attachment3, VanKowalik,and Tad Murty for helpfulcomments; Do Kyu Lee for couver, B.C., 1976. assistancein constructingthe grid; Roger Flather for the use Lynch, D. R., F. E. Werner,D. A. Greenberg,and J. W. Loder, of his model;PatriciaKimberfor draftingsomeof the figures; Diagnostic model for baroclinic,wind-driven and tidal circulaRosalie Rutka for typesetting assistance;and the reviewersfor theirconstructive criticismof an earlierversionofthemanuscript. tion in shallowseas,ContinentalShelfRes. 12, 37-64, 1992. P. W., Introductionto Map Projections,174 pp., This projectwassupportedby the CanadianPanelof Energy McDonnell, Marcel Dekker, New York, 1979. Researchand Development,project67146,and by the Canadian Pekeris,C. L., and Y. Accad, Solutionof Laplace'sequations Department of Fisheries and Oceans. for the M2 tide in the world oceans,Philos. Trans. R. Soc. London, 265(1165), 413-436, 1969. REFERENCES

Bell, W. H., and N. Boston,The Hecatemodel,Manuscr. Rep. Set. (Oceanogr.and Limnol.) 159, 27 pp., Fish. Res. Board

Perry, R. I., B. R. Dilke, and T. R. Parsons,Tidal mixingand summerplankton distributionsin Hecate Strait, British Columbia, Can. J. Fish. Aquat. Sci., •0, 871-887, 1983. Ray, R. D., and B. V. Sanchez, Radial deformation of the earth

by oceanictidal loading,NASA Tech. Memo., 100743,1989. Schwiderski, E. W., Globaloceantides,Part I. A detailedhydroBell, W. H., and N. Boston,Tidal calibrationofthe Hecatemodel, dynamicalinterpolationmodel, Rep. NSWC/DL TR-3866, Can., Ottawa, 1962.

J. Fish. Res. Board Can., 20, 1197-1212, 1963.

Bennett,A. F., and P. C. Mcintosh,Open oceanmodellingas an inverseproblem: Tidal theory, J. Phys. Oceanogr.,12, 10041018, 1982.

Naval SurfaceWeaponsCenter, Dahlgren,Va., 1978.

Thomson,R. E., Oceanography of the British ColumbiaCoast, Can. Spec. Publ. Fish. Aquat. Sci., 56, 291 pp., Dep. of Fish. and Oceans, Ottawa, 1981.

FOREMAN ET AL..' TIDESANDRESONANCE, NORTHCOASTOF BRITISHCOLUMBIA Thomson, R. E., and R. E. Wilson, Coastal countercurrent and mesoscaleeddy formation by tidal rectification near an oceanic

cape, J. Phys. Oceanogr.,17(11), 2096-2126, 1987. Wahr, J., Body tides on an elliptical, rotating, elastic and oceanless earth, Geophys. J. R. Astron. Soc., 6•, 677-703, 1981. Walters, R. A., A finite element model for tidal and residual cir-

culation,Commun.Appl. Numer. Methods,2, 393-398, 1986. Walters, R. A., A model for tides and currents in the English Channel and southern North Sea, Adv. Water Resour., 10, 138-148, 1987. Walters, R. A., A finite element model for tides and currents with field applications, Commun. Appl. Numer. Methods, •, 401-411, 1988.

2531

Walters, R. A., A three-dimensional, finite elementmodel for coastal and estuarine circulation, Continental Shelf Res., 12, 83-102, 1992.

V. A. Ballantyne, M. G. G. Foreman, and R. F. Henry, Department of Fisheries and Oceans, Institute of Ocean Sciences, P.O. Box 6000, Sidney, British Columbia, VSL 4B2, Canada. R. A. Walters, U.S. GeologicalSurvey, 1201 Pacific Ave., Suite 901, Tacoma, WA 98402. Received March 9, 1992; revised August 31, 1992; accentedSeptember29, 1992.)

Suggest Documents