MU _. NAVAL POSTGRADUATE SCHOOL Monterey , California THESIS ENVIRONMENTAL INELLENCES ON THE PRO- DUCTION OF ARCTIC HALOCLINE AND DEEP WATER by • James A. Hill f * * June 198S Thesis Advisor Albert J. Scmtncr Approved for public release; disiribmion is unlimited. T238976 Unclassified securily classification o( this page Rp:r()Ri D()( IMl NTAIiON P \(,1. a Report Security Classirication I'nclassified 2a "security Classirication Authority 2b Declassification; Downgrading Schedule 4 Performing Organizalion Report \uniher';s) 6a Name of Perforniing Organisation Naval Postgraduate School i^h Office Symbol ( if appHrahln) 35 lb Restrictive Markings ■) Distribution Availability (^{ Report Approved for public release; distribution is tinlitnitci. 5 Monitoring Organizalion Report Number's) 7a Name of Monitoring Organisation Naval Postgraduate School 6c Address (citv. slalp, and ZIP mdf / Monterey. CA 93943-5000 /b /Xddiess i rity, siaiP. and 7.11' mdr; Monterey, CA 93943-5000 !5a Name of Funding Sponsoring Organisation 8h orncc Symbol ( if applicable > !>-c Address (dry, statp. and 7.1 P mdp) '■I Pn^cLivemiml InslrLimcnl Idcntificalion Number 10 Source of Funding Numbers Program Fllement No Project No Task No Work l/iiit /\cccssion 11 ra\e (include ^pruriivcla.s.u/iratinn) I'NVIRQNMr.N'IAl INFI (['NCT'S HN TUF PROOrCi ION or /\Rk HALOCI INE AND DFJ-,P WATER IC 12 Personal Autho;v,>; .lames A. IIUl 1,5 a Type oT Report Master's I'hesis 13b Time Covered From To 14 Dale of Report i yrar. month, day) June 1* 1 5 Pige Count 54 16 Supplementary Notation The vievvs expressed in this thesis are those of the author and do not reflect the official sition of the Department of Defense or the U.S. Government. policy (^r po- 17 Cosali Codes Field Group Subgroup IS Subjccl Terms (mnlinur •'/; reverse if necessary and identify hy hlr.k number Arctic, halocline. deep water, plume flow, polyna 19 Abstract (continue on reverse if necessary and identify by block number) Pease (1987; related the effects of atmospheric forcing, mainly temperature and wind fields, to the size of coastal polynas. Using Pease's formulation and Killworth's (1977) plume model as applied by Melling and Lewis (1982), the effects of atmo- spheric forcing on brine injection into the Arctic pycnoclinc are investigated. 'T'his paper will discuss the likelihood o/coa^t.al polynas as a source for denser abyssal waters. A standard ca,se was developed for the model with initial conditions taken from Melling and I ewis (1982) and Pease ( 1987) for comparison v. ith individual sensitivity experiments. Ten environmental parameters were individually examined for their influence on ihe plume depth after 90 days. T"he standard case resulted in a 90-day plume depth of 436 meters. A submarine canyon case was simulated, resulting in plume penetration to over 1300 meters in 90 days. Further experiments used actual T-S soundings from Aagaard et al. (1981) ,nnd Ostlund et al. ( 1987;. Finally, a 20 kilometer wide plume is shown to penetrate to a!—i ist 600 meters in 90 davs. 20 Distribulion/Availabiiiiy of Abstract S unclassified/unlimited D same as report D DTIC u.scr 21 Abstract Security Classification Unclassified 22a Name of Rcspot^siblc individual Albert .F Semtner 22b Telephone (include .irea rode) (408) 646-2768 22c OlTicc Symbol 54Ss ^D FORM 1473.84 MAR ?:?> APR edition may All oth.-r ediln le used until exhausted ins are obsolete security classification of this page [ nclassificd Approved for public release; dustribunon is unlirruted. Environmental Influences on the Production of Arctic lialocline and Deep Water by James A. Hill Lieutenant. United States Na\y B.S.. Sam Mouston State University, 1979 Submitted in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE IN METEOROLOGY AND OCEANOGR.\PHY from the NAVAL POSTGR.-\DUATE SCHOOL June l^SS ABSTRACT Pease ( 19S7) related the effects of atmospheric forcing, mainly temperature and wind fields, to the size of coastal polynas. Using Pease's formulation and Killworth's (1977) plume model as applied by Melling and Lewis (I9S2), the elTects of atmosoheric forcing on brine injection into the .Arctic pycnocline are investigated. This paper will discuss the likelihood of coastal polynas as a source forxienser abyssal waters. .A. standard case was developed for the model with initial conditions taken from VIelling and Lewis (1982) and Pease (1987) for comparison with individual sensitivity experiments. Ten environmental parameters were individually examined for their influ- ence on the plume depth after 90 days. The standard case resulted in a 90-day plume depth of 436 meters. A submarine canyon case was simulated, resulting in plume pene- tration to over 1300 meters in 90 days. Further experiments used actual T-S soundings from Aagaard et al. (1981) and Ostlund et al. (1987). Finally, a 20 kilometer wide plume is shown to penetrate to almost 600 meters in 90 days. Ill ■. ,^.'.,f f TABLE OF CONTEiNTS I. INTRODUCTION I I. NUMERICAL METHODS AND PAR.\METERS 6 A. THE NUMERICAL MODEL 6 The coastal polyna , 6 Salinity distribution within the water column 7 Equation of state S The streamtube model S The coupled model II B. MODEL PAR.AMETERS 12 1. Environmental variables 12 2. Environmental conditions for the standard case 14 III. RESULTS 16 A. STANDARD CASE 16 B. SENSITIVITY EXPERIMENTS 16 C. THE SUBMARINE CANYON" CASE 2S D. RESULTS USING OBSERVED TEMPER.ATURE AND SALINITY FIELDS 30 IV. DISCUSSION 37 REFERENCES 40 INITIAL DISTRIBUTION LIST 43 M< LIST OF TABLES Table 1. LLST OF EWl RONMFN I.AL INPLTS 1.; Tjblc :. IMTLAL \-ALLFS LSED FOR TIIF STANDARD CASE 1^ l.;bie 3. SALINFFY CONTRIBLTIOX OF FOLYNA WITH DECREASING AIR TEMPER.-\TLRE 1q Tabic 4. TMIS TABEE COMPARES SOME RESUETS FROM LSING AC- TUAL T-S RECORDS ' 30 Tabic 5. INITEAL VALUES USED FOR THE -SUBMARINE CANYON' CASE 3l LIST OF FIGURES Figure 1. The streamtube coordinate system as given by Smith (1975) 10 Figure 2. Schematic representation oftlie coupled model 13 Figure 3. The standard case plume's velocity is shown with depth 17 Figure 4. The standard case plume How path in the x-y coordinates 18 Figure 5. EHect of decreasing air temperature on the plume's 90-day depth 20 Figure 6. Efiect of varying oiTshore wind speed on the 90-day plume depth 21 Figure 7. The efiect of initial shelfwater salinity on the 90-day plume depth .... 22 Figure 8. Effect of varying initial shelfwater velocity on 90-day plume depth .... 23 Figure 9. This figure shows the efiect of the initial shelf slope on the 90-day plume depth 24 Figure 10. The elTect of varying the secondary slope on the 90-day plume depth . . 25 Figure 1 1. The elTect of varying the shelf width (after the 45 m isobath) on the 90-day plume depth 26 Figure 12. The elTect of varying the initial pycnocline strength on the 90-day plume depth 27 Figure 13. The elTect of varying Layer 2 Briint-Vaisalii frequency on the 90-day plume depth 28 Figure 14. Decreasing Layer 2 thickness results in a deeper penetrating plume ... 29 Figure 15. The "submarine canyon" plume's fiow path 32 Figure 16. The velocity of a "submarine canyon" plume with increasing depth .... 33 Figure 17. The normalized mass fiux of the "submarine canyon" plume 34 Figure IS. The normalized mass fiux of the standard case plume 35 Figure 19. Velocity shown with depth of a 20 kilometer wide plume 36 VI ACKNOWLEDGEiVIENTS I would like to state my appreciation to Mr. Steve Ackley for his assistance and encouragement: Mr. Humphrey Melling. for his advice and for providing the numerical code of the streamtube model; the Oceanography research assistants (mainly Vis. Arlene Bird and Mr. .Mike McCann) for their inexhautable resourcefulness thoughout my studies; and Professor Albert J. Semtner for, his unflappable patience. Lastly, I would like to thank my lamily, Juanita, Patrick, and Corey for their support. vu I. INTRODUCTION The Arctic Ocean, a small and remotely located part of the world's ocean, has de- manded the attention of those who wish to understand the Earth's climate. The renewed emphasis in recent years given to understanding the Earth's climate has underscored the role of the Arctic Ocean in the scheme of the global energy balance. The Arctic Ocean is also a key player in the world oceans' thermohaline circulation. Because of the im- portance of the Arctic Ocean in the global picture, understanding the nuances o[ the Arctic Ocean's physical oceanography and its sensitivities to environmental change is essential before climatologists, oceanographers. and meteorologists can develop a more reahstic and longer range global climatological concept. Two puzzling aspects o[ the Arctic Ocean are its isothermal halocline and the more saline deep waters. How is the halocline maintained in a nearly horizontally independent state'? What is the mechanism behind the formation of the Arctic Ocean's deep waters? These topics will be discussed in this thesis. Throughout the Arctic Ocean, the characteristics of the upper layers are remarkably independent of the horizontal location of the sample. The water column consists of a mixed layer, halocline, thermochne, and deeper waters. The mixed layer is a relatively fresh water layer approximately 30 meters deep. River run-off and the summer cycle of melting ice are considered the main contributors to the fresh water layer. The halocline is the layer of water beneath the mixed layer. In this layer, salinities rise from 31 to as high as 34.6 ppt at the 200 meter depth level (Killworth and Smith, 1984). The halocline IS. in essence, an isothermal layer with temperatures just above the freezing point. The thermocline is a layer below the halocline. in which temperatures rise to approximately 0^' C at 250 meters (Melling and Lewis. 1982) and continues to rise to 0.5" C at 400 meters due to the layer of warm Atlantic water. Below this depth, temperatures decrease with increasing depth. The deep waters of the .Arctic Ocean are among the most dense waters found in the world's oceans and have been suggested as being an additional source for ventilating the world's oceans' deep thermohaline circulation (Aagaard,1981). Although the physical oceanography of the Arctic Ocean is well described in literature, the exact mechanism(s) for the maintenance of the halocline and deep water formation are not fullv understood. The halocline is generally isothermal except in the Canadian Basin, where warm Pacific waters entering via the Bering Strait causes a warm layer at about the 75-100 meter depths, and presents a paradox when examining its origin. Sandwiched between a cold, relatively fi-esh layer and a warm, saline layer, an obvious solution to the main- tenance of the halocline might be simple \crtical mixing between the two Uuers. How- ever. \ertical mixing of the cold, less saline surface water with the warmer, more saline Atlantic water is not considered to be a source since heat is not mixed along with salt. Aagaard et al. (1981) showed a discontinuity existed in the T-S correlation slope at about 150 meters. Either salinization of surface waters by freezing (i.e. brine rejection) or cooling of Atlantic water could be two possible mechanisms to produce halocline water. The authors (Aagaard. et al., 1981) demonstrate mixing between the two layers cannot be a source of halocline water by examining the required salinity of the surface water. Using a T-S diagram, and assuming linear mixing at a ratio of 3:2 (surface water to Atlantic water), the surface water would need to have a salinity of 34.7 ppt. Salinity this high has not been observed in the Arctic surface layer. Therefore, the maintenance of the Arctic halocline is generally thought to be attributable to lateral adveciion of cold, saline water from the continental shelves. The shelf waters are made more saline due to brine rejection associated with ice grouih. Aagaard (1981) proposed that continental shelf areas where an annual growth of ice less than 2 meters is necessary to modify summertime salinity profiles to wintertime profiles may be the source areas of the halocline waters. Upwelling of Atlantic water onto the shelves and subsequent cooling has also been suggested as a possible source. Upwelling along submarine canyons has been reported as early as 1929 (Sverdrup.1929). As further evidence eluding to the halocline's possible shelfwater origin, water on the continental shelves has been shown to have temperature and salinity characteristics similar to the waters found ofTshore at a depth of 150 meters (Melling and Lewis, 1982). In their paper, Melling and Lewis (1982) discuss the lateral advection of low tem- perature high salinity shelfwaters into the halocline, with particular attention paid to the efiects of brine rejection due to the growth of ice. Aagaard (1981) used two methods of computing the necessary production rate of high salinity shelfwater to maintain the halocline. By considering the production rate required to renew the halocline in 10 years and necessary mixing ratios to achieve desired T-S characteristics, Aagaard suggests a 2.5.vlO* to 4.0.vl0''m^— sec"' for a production rate of cold saline shelfwater, assuming a halocline renewal period of 10 years. In order to increase the ambient shelfwater salinity to values which would form sufTiciently dense water to descend the slope into the halocline, Melling and Lewis (1982) emphasize the important role of ice dynamics, spe- cifically the role of divergence. Areas of large wintertime ice cover divergence allow rapid ice grovnh hence increased salinization of the underlying water column. Leads and polynas are locations oi^ rapid ice growth due to the divergence of the insulating ice cover. Schumacher et al. ( 1983) examined the degree of salinization of the underlying water column beneath a polyna. In their study, measurements of salinity and water current velocity beneath a polyna located at St. Lawrence Island indicated polynas to be an area of high ice growth rates resulting in density driven currents due to increased salinization (Schumacher, et al., 1983). Coupling the salt source, i.e. polyna, with the interleaving plumes of dense shelfwater found by Melling and Lewis (1982) may provide a possible mechanism for the maintenance of the Arctic halocline. The polyna provides an atmo- spheric dependent source of salt into the water column possibly forming plumes which descend down into the halocline, as described by VIelling and Lewis (1982). .Along with the mechanism associated with the halocline's maintenance, the formation of the Canadian Basin's deep water has also been attributed to having a source of low tem- perature, high salinity shelfwater. The dense waters which ventilate the world's oceans flow out of the Arctic Ocean over the Greenland-Scotland ridge system at a depth of about 2600 meters. A single source was advocated by Helland-Hansen and Nansen (1909) in which deep water is formed in the Greenland Sea. Nansen ( 1906) had written earlier of the possibility of the Greenland Sea Deep Water being modified by cooling in the Barents Sea. This theor\' omitted the possibility of deep water formation within the two basins of the Arctic Ocean, namely the Eurasian and Canadian Basins. In 1981, Aagaard postulated that deep water is also formed within the Arctic Ocean due to the salinity distribution. Greenland Sea Deep Water has a salinity of approximately 34.90 ppt, which is the least saline of Arctic deep waters. Therefore, at least two sources of deep water formation are thought to exist. One source is the traditional Greenland Sea formation area and the new source proposed by Aagaard, is within the Arctic Ocean's two basins. The characteristics of water mass properties found in the Canadian and Eurasian Basins are similar until about the depth of the Lomonosov Ridge. At about 1500 meters depth, the Canadian Basin becomes more saline and is slightly warmer than the Eurasian Basin Deep Water (Aagaard, 1981). Lnder the traditional viewpoint of Greenland Sea Deep Water filling the two basins through lateral advection, salinities should not exceed 34.9 ppi. However, during the I.ORHX (Lomonosov Ridge rxpenment; Weher. 1979) salinities in the I" ins were detcrmuicd to he greater than 34.9 ppt. "Ihis suggests an additional source of salt if not a completely new source of deep water. Current meters, placed o^ :r the Lomonosov Ridge, gave evidence to the existence of some tidal and other episodic surges carrying salini/ed Greenland Sea Deep Water over the ridge into the Canadian Basin. Fvcn with the surges over the Lomonosov Ridge into the C\anadian Basin, water from the Eurasian Basin must have an additional salt source to obtain the characteristics of water found in the (\anadian Basin (Aagaard, 1981). Residence times can be calculated by comparing the mixing ratios of shelf water to Atlantic Intermediate water necessary to achieve the L-S characteristics found in the (Canadian Basin Deep Water, (.'onsidering a small flux of shelfwater (.003 to .06 Sv) necessary to ventilate the Canadian Basin over several hundred years may also explain the warmer temperatures found in the Canadian Basin (Aagaard et al., 1985). By using tracers such as O'^O and '"C, Ostlund et al. (1987) agree with the magnitudes of the renewel time scales for both basins presented by Aagaard et al.(l9S5). Ostlund et al. (1987) and Aagaard et al. (1985) place the residence time of the I'urasian Basin in the order of decades \ice several hun- dred years (700 years, Ostlund, et al.. 1987) for the Canadian Basin. The requirement remains of an additional salt source at small fluxes to ventilate the Canadian Basin over several hundred years point to an episodic mechanism. Killworth and Smith (1984) write that shelfwater sinking at the basin edges may occur but a source of waters sufficiently dense to penetrate the .Arctic Ocean's stable mid-layer remains a problem. Again, using a coastal polyna as a salt source, could shelfwater be modified under extreme conditions to produce such dense water which would then descend to the deeper depths via gravity How plumes? This thesis will combine Pease's numerical formulations relating atmospheric forc- ing, mainly wind and air temperature, to the opening and closing of coastal polynas. with a streamtube gravity flow model developed by Smith (1975). With simplified hydrography and suitable environmental conditions, this model can produce events of deep plume penetration along with a more consistent source of high salinity, low temperature water for the maintenance of the Arctic Ocean's halocline. Chapter 2 will describe the numerical model used in this study along with an examination into drawbacks associated with Smith's streamtube model as described by other studies, i.e. Smith (1975). Killworth (1977), and Melling and Lewis (19S2). Chapter 3 will describe results obtained by altering environmental inputs into the model and Chapter 4 will discuss the result^ and im.plications. II. NUMERICAL METHODS AND PARAMETERS A. THE NUMERICAL MODEL With the increased desire to understand the nature of the Arctic halocUne and its origin, there existed suITicient pieces of the puzzle to develop a coupled model. The 3-D turbulent entraining plume model developed by Smith (l'^T5) has been used by several studies ( KiUworth (1977). KilKvorth and Carfnack ( 197S). and Melling and Lewis ( 1982)) to gain insight into gravity flow currents arising from brine rejection into ^he water col- umn due to ice growth. Melling and Lewis (1982), compute the initial density of the plume by estmiating the average ice grown annually in a shallow shelf and and the re- sultant salinity contribution. In this study. Pease's (1987) formulation relating a polyna's size to ofTshore wmd speed and air temperature, provides a specific source of brine which is then coupled with Smith's plume model, as used by Melling and Lewis (1982), in order to study the possible contribution of the polyna produced denser waters to the .Arctic halocline and possible formation of Canadian Basin Deep Water. L The coastal polyna Pease (1987) describes the polyna's ma.ximum width as an equilibrium between frazil ice production and its advection rate. The frazil ice production rate is a function of the heat exchange at the air-sea interface. Therefore, with wind speed and air tem- perature as independent variables, one can compute the size of the polyna and its frazil ice production rate. The polyna size can be found by solving: /: ' //; where — 7- is the chanse in polvna width with time, V, is the advection rate of frazil ice. X is the polyna width, F^ is the frazil ice production rate, and //, is the collection depth of grease ice or nilas. The solution of equation (1) assumes the meteorological condi- tions are steady throughout the polyna's opening time, thus allowing treatment as a linear differential equation. Furthermore, if the meteorological event is assumed to be of sufTicient duration so that the polyna reaches maximum width, equation (1) can be written as: XpUnaximum) = — - — . (2) This allows solving equation (2) for the maximum width after computing the frazil ice production rate. F. In this study, meteorological events of interest are 4-10 days duration, therefore this assumption allowing equation (2) is valid. Ou (1988) de- scribes a more detailed coastal polyna model and concludes the ice edge is less aflected by higher frequency atmospheric disturbances than longer period, synoptic type vari- ations. Although Ou (19SS) includes additional physics to model the temporal changes of the polyna's ice edge, the steady state solution is still that as given by Pease (1987) and shown above in equation (2). The frazil ice production rate is given by Pease (1987) as: PiL where the evaporative heat flux has been neglected due to its small contribution relative to the uncertainty of the sensible heat flux. Q,^ is the upward longwave radiation. ae,T^, is the downward longwave radiation, L is the latent heat of freezing for salt water, p, is the density of young sea ice, and p.Ci^C^VXT, — T^) is the sensible heat flux. 2. Salinity distribution within the water column The salinity contribution to the water column from frazil ice grovnh is assumed to be uniform with depth and is computed using: S„^^, = S H « (4) 'new { qdz] from Kilhvorih and Smith (1984). 5„,„ is the new salinity value of the water column after brine rejection. S is the average salinity of the shelfwater. S. is the salinity of the ice, A. is the effective area of salinization. and J qaz is the areal flux of the underlying water. Using F, from the polyna model, one can compute the salinity of shelfwater exiting a polyna. 3. Equation of state The density of the current is computed using the International one atmosphere equation of state developed by Millero and Poisson ( 1981). The equation is given below. p = p^ + AS + BS+ CS (5) where, A = 8.24493 x 10"^ - 4.0899 x 10"^: + 7.6438 x 10"^r^ - 8.2467 x 10"'/^ + 5.3875 x 10"^* B = -5.72466 x 10"^ + 1.0227 x \0"^t - 1.6546 x 10 - 6/^ C = 4.8314 x lO"'^ and p, is the standard density of seawater. This allows the calculation of the currents relative buovancv usina: A. = -I(^i^). (9) In this fashion, the results from the polyna model can be used as inputs to the streamtube model. 4. The streamtube model Smith's (1975) model of boundar>' currents includes the effects of friction, entrainment, and Coriolis forces. The 3-D plume was modified by Killworth (1977) by altering the entrainment rate. Killworth (1977) related the entrainment rate to the product of the plume's area and velocity vice only the velocity (Smith, 1975). Melling and Lewis (1982) used Kilhvorth's version of the plume model with an entrainment co- elTicient dependent on the Richardson number. The model used in this paper is the Melling and Lewis (1982) adaptation with the inclusion of several stratified layers to al- low a single model initialization. Smith's plume model uses two coordinate systems. The Cartesian system is aligned with the bottom so the X-axis lies along the shore and the Y-axis lies along the slope. The second coordinate system is the curvilinear system with c and ;/ axes. The position of the plume axis can be defined by the value of c . The angle which the plume crosses the isobaths is given by the angle fi . The coordinate system is illustrated in Figure 1 on page 10 and includes gravitational, g, and rotational. Q., vectors. The equations for the path fiow are:' 4^ = cos/? (10) and 4r = sm/?. (11) The plume is considered to be a well-mixed, broad, thin layer of high density fluid adja- cent to the bottom. Pressure gradient forces pull the plume down the slope while CorioUs forces bend the flow to the right. If frictional forces were not considered, the flow would move horizontally along the isobaths (Killworth, 1977) in geostrophic bal- ance. Friction is included, using quadratic drag laws to relate frictional resistance to the square of the mean velocity (Smith, 1975). The plume is thus allowed to flow down the slope, constantly entraining ambient fluid and losing its negative buoyancy. Once the plume reaches neutral buoyancy, the plume interleaves with the ambient waters. The equations of the plume's flow from Killworth (1977), are: -Jt{AV^ = E,A^V, (12) ^4^[AV") = A^sm^s\np-K.Al v', (13) d^ -jr {A VA) = -A VN^ sin 9 sin /?, (1-4) Figure 1. The streamlube coordinate system as given by Smith (1975). F'44- = Asin^cos/?-yT'. 15) These equations are, the along-stream derivatives of the mass flux (12), the momentum flux (13), the buoyancy flux (14) and the cross stream momentum balance (15). In the above equations, A is the cross-sectional area of the plume, V is the plume's velocity, A is the plumes buoyancy, d is the slope angle, /? is the angle at which the plume crosses 10 the isobath, N is the Briint-Vaisala frequency, A', is the drag coefTicient./is the Coriolis parameter, and £, is the entrainment coeHicient. The equations were integrated using a modified Adams method. Predictor: J/+, =yi + 3r (5^ - ^Vli + m-2 - %-2) (16) Corrector: JWi =>•/ + 77 (9/;+, + i9y; - 5y;_, +/,_,), (i6a) with the initial steps provided with the Runge-Kutta method, This model uses Killworth's (1977) entrainment and frictional parameterizations, which has the entrainment rate and frictional force porportional to the product of the area and velocity (velocity squared for friction). Thus, —^ = —rr = ( -7- )t , where w is the width of the plume and b is the depth of the plume. L, A., b Also used in this model is Bo Pedersen's (1980) entrainment parameterization for small slopes. Melling and Lewis (1982) included this in their version of Smith's model. The entrainment coefficient is given as E =■ 0.072 sin ^ sin /? = 0.072( -r- ) where Ri = hA cos 0/ J-'- is the bulk Richardson number of the fiow. 5. The coupled model The model takes the form illustrated in Figure 2 on page 13. Underneath the polyna, a plume with a depth, b. and a width, w, is salinized due to brine rejection. The plume has an initial velocity, V, . The efiects of varying the initial velocity will be ex- amined in Section 3, however, the standard case value of V„ = .OA/nIs is used. This value 11 is close to the values used by Melling and Lewis (1982) and observed values of currents adjacent to polynas reported by Schumacher et al. (1983). The value of l'\ determines the areal flux used in calculating the salinity mcrease in the water column. The initial velocity is increased after the non-entraining, non-rotating rudiment plume travels 10 km straight downslope. Effects of rotation are neglected at this early stage of the plume to minimize the exposure to brine rejection under the polyna. Increasing the velocity after 10 km ensures more realistic salinity values of the plume, and would, in any case, occur in the first steps of the streamtube model. The non-entraining, non-rotating rudiment plume continues straight downslope until a depth of 45 meters is reached. This depth was chosen due to its use by Melling and Lewis (I9S2) as the initial depth of the streamtube model. .At 45 meters, the plume descends the slope according to the dynamics set forth and established by Smith (1975), Killworth (1977). and Melling and Lewis (1982). Three stratified layers are defined in this coupled model. Layer 1 lies between 45 and 55 meters depth. Layer 2 lies between 55 and 300 meters, and Layer 3 lies below 300 meters. The plume is assumed to interleave with ambient waters when the velocity is less than .01 m s; however, all sensitivity studies will be carried out using 90 days as the maximum number of days the plume can descend while retaining its integrity. B. MODEL PARAMETERS 1. Environmental variables In order to examine the environment's influence upon possible polyna- related production of halocline and deep waters, certain environmental variables were individ- ually varied during model simulations. The sensitivity of the model to a variable's al- teration is defined by the dependence of the plume's depth after 90 days upon the particular variable, relative to a standard case. The relationship between the depth of the plume and a single environmental variable will be reported in Chapter 3. The vari- ables to be investigated are listed in Table 1 on page 14. 12 ZO - i 10 .M Y _ n 1 f. «. A \ m ■ -- ■ '\ — r- IT^ ■ - - \V, ~~- -~ ^_ -~- -— -. '•^»V, . , . .__ 1 -Avrp 1 '^~'^^~-^. ■ -A^ER 2 \ 30O- -A'-ER Figure 2. Schematic representation of the coupled model. As the shelfwaier is ex- posed to the polyna, a high density plume is formed, resulting in the plume's descent down the slope until it is neutrally buoyant. 13 Table 1. LIST OF ENVIRONMENTAL INPUTS: The dependency of the plume depth after 90 days on each of the variables below will be examined in Section 3. \'ariable Definition. T.. air temperature ['. wind velocity .? shelfwater salinity r. initial sheliwater velocity .S;v shelf width, from the A5 m Isobath ^^ initial shelf slope 0. secondary slope ■\ Brlint-Vaisala frequency, layer 1 .\\_ Brunt-Vaisala frequency, layer 2 AZ, thickness of layer 2 2. Environmental conditions for the standard case To achieve consistency with pre-existing works, many of the environmetal con- ditions for the standard case were borrowed from Melling and Lewis (1982), Killworth and Smith (19S4). and Pease (I9S7). The standard case initial values are listed in Table 2 on page 15. The most significant departure from an initial value given by previous work, is the H, value. In this paper, the frazil ice collection thickness is taken to be .2 meters. Pease (1987) shows a higher collection thickness results in longer polyna opening time and larger polyna width for the same wind speed and air temperature. Weeks and Ackley (CRREL Monograph 82-1. 1982} give higher values for frazil ice collection thicknesses in a wind driven scenario vice .01 to .1 meters in quiescent conditions. In addition to the frazil ice collection thickness, the salinity of the ice is set to 7 ppt (Cox and Weeks, 1974) instead of 5 ppt used in Killworth and Smith (1984). 14 Table 2. INITIAL VALUES USED FOR THE STANDARD CASE. Variable \'alue Definition T:r.< -15. air temperature T... C -l.S water temperature r. .. ni sec"' 15. wind velocity \ . HI sec"' 3 ^) r;. ice tloe velocity //, . m T frazil collection thickness (7. ll'nr-ik'g-^ 5.67.V10-* ' Stephan-Bolt/mann constant t\ 0.95 emissivity of the air p,kgm~- 1.3U air density p,. kg/ir^' l.OKxxHP seawater density p.kgnr^ 0.95.vl(V ice density c, 2.0.vl()-^ sensible heat coetTicient C. Jdeg-'kg-' 1004 specific heat of air 0,. IVnr- 301 longwave radiation upward L. J kg-' 3.34.vl0^' latent heat of fusion h. . m 10. flow thickness vv„ . m 1000. How width V^. ni sec"' .04 flow speed /? . degrees 29 flow direction /. sec-' 1.38.vlO-^ Coriolis parameter .V. sec-' 0.0316 Briint-Vaisala frequency, layer 1 \ 0.0077 Briint-Vuisala frequency, layer 2 N 0.001 Briint-Vaisala frequency, layer 3 iSZ. . meters 195 Layer 2 thickness ^, 0.5.vlO-^ initial bottom slope ^: 5.V10-3 secondary bottom slope K 0.01 drag coetlicient E 0.072 sin 6 sin /? entrainment coetTicient J.ppt 32.5 initial shelfwater salinity 5, . ppt 7.0 Salinity of frazil ice 15 III. RESULTS A. STANDARD CASE In order to examine the environmental intluence on the possible polyna production ofhalocline and deep waters, a standard case was defined with values which are typical of Arctic Ocean observations. The olTshore wind \elocity requires a considerable meteorological event providing winds higher than the monthly mean winds. However, the \alue of 15 m s for otTshore wind \'elocity is within obser\'ed u'ind speeds for .Arctic storms reported in previous papers (Schumacher, et al. (1983) and .Aagaard. et al. (1985)). The standard case plume obtains a salinity of 34.98 ppt after transversing the polyna. which corresponds to a density of 1032.93 gmjni^ and a negative buoyancy of .0643 nils- . It is therefore not suprising when the plume penetrates the same Arctic pycnocline used in Melling and Lewis (1982) to a depth of 436 meters. Figure 3 on page 17 shows the velocity of the plume as it descends the slope. Due to the plume's high negative buoyancy, the plume accelerates slightly on the initial slope then rapidly in- creases in speed after the shelf break at 55 meters. The maximum velocity is 0.30 m s reached at a depth of 60 meters. The mass flux, .-W, increased by over 1600'" o empha- sizing the dependence of the entrainment rate on the velocity of the plume. At the end of 90 days, the standard case plume has a velocity of 0.017 m s and a negative buoyancy of 0.000482 ml sec'. The plume has travelled 253 kilometers along the shoreline and 96 kilometers down slope. The flow path of the plume can be seen in Figure 4 on page 18. Overall, the plume in the standard case behaves similarly, although with different magnitudes, as the plumes in Melling and Lewis (1982). With this in mind, sensitivity studies of the impact of individual environmental parameters on the 90-day plume depth can be examined. Sensitivity experimental results typically include 10-20 data points with increased resolution in areas of abrubt change. B. SENSITIVITY EXPERIMENTS The polyna's maximum width is dependent on the growth rate of ice. the advection rate of ice, and the collection thickness of ice. Air temperature and the wind speed af- fects the growth rate of ice, while the advection rate is proportional to the wind speed. Lowering the air temperature results in a monotonic increase in the plume's depth at 90 days as shown in Figure 5 on page 20. This is expected due to the increased production of frazil ice resulting in higher salinization of the water column. Although Pease (1987) 16 ICC.O 2C0.0 3jG.O 40G.0 PLUME DlPTH (M) 0.0 Figure 3. The standard case plume's velocity is slioun \vith depth: Note the rapid increase in velocity as the plume crosses the shcllbreak at the 55 meter isobath to a steeper slope. has shown the polyna size decreases with lowering air temperature at a given wind speed, the plume's salinity {hence, density) increases most rapidly in the first 10 kilometers of downslope travel. Here the plume will be subjected to significantly higher salinization due to higher ice gro\nh rate and lower areal flux rate. Although the polyna size is smaller with lowering air temperatures, the majority of the salinization has already oc- curred in the first 10 kilometers. Indeed, the polyna shows a decreasing salt contribution 17 2 ° >- a CJ CO a LJ Q_ o CD JH _1 en I CD O _J cn o c CO l.O 2.0 RLONG-SHORE DISTRNCE 3.0 Figure -J. The standard case plume flon path in the x-y coordinates: The efTects of Coriohs can be seen as the plume's angle p decreases during the plume's descent. IS after the first 10 kilometers with decreasing air temperature. This can be seen in Table 3 on page 19. This decreasing salt contribution to the water column is due to the polyna's smaller width. Table 3. SALINITY CONTRIBUTION OF POLYNA WITH DECREASING AIR TEMPERATURE: The table shows the initial salinity of the plume. S„„ after the first 10 kilometers of downslope travel, the final salinity of the plume. S., after travelling the entire width of the plume. A'., and the de- creasing salt contribution of the polyna during the distance A'^ — lOkni with decreasing air temperature. T..." S;,.r . ppt S, . ppt A5 , ppt T:,r. ° C 1-1 -* / 34.36 1.0 -5.0 33.75 34.6" 0.92 -11). 34.14 34.9S 0.S4 -15. 34.54 35.30 0.76 -20 34.93 35.61 0.6S -25 35.32 35.92 0.60 -30 35.72 36.23 0.51 -35 On the other hand, if air temperature is held constant, and offshore wind speed is increased, the polyna grows in size due to increased ice advection rate. The initial con- tribution of salt during the first 10 kilometers of the polyna does not vary however, the increased polyna size exposes the water column to further salinization. This results in a denser, deeper penetrating plume with increasing wind speed. (See Figure 6 on page 21). The dependence of plume depth upon the wind velocity and air temperature illus- trate the possible atmospheric influences upon the Arctic Ocean's salinity characteristics. The frequency and severity of Arctic storms which would be conducive to high density plume formation is a consideration in the calculation of production rates, but cannot be discussed quatitatively using this simphfied model. The effect of varying the initial salinity of the shelfwater on plume depth is shown in Figure 7 on page 22. The results of this experiment show deeper penetration with higher initial salinity. This is not suprising due to the significant role of salinity upon water density near the freezing point. The plume depth after 90 days is greater than the depths reported by Melling and Lewis (1982). even with relatively low salinity values. Initial shelfwater velocity defines two parameters within the model, the areal fiux used in computing the new salinity of the water column and the initial velocity of the 19 o o o o , — . LO o Q_ LO LJ CD o LJ o_ > \ Vh CD _ i \ a a o O o -; 1 50.0 -40.0 1 1 -30.0 -2G.0 -IG.O 0.0 FilR 'T' p' r^' "^ r^ ~~. ^> "^ 1 1 (^ i— ' (C) Figure 5. Effect of decreasing air temperature on the plume's 90-day depth: With decreasing air temperature, the increased ice growth rate results in higher initial salinization of the water column. plume. The higher the initial velocity, the less saline the plume. However, when using the plume depth at 90 days as a standard criteria to examine the results, the higher ve- locity plumes reach deeper depths at 90 days than the denser, but lower velocity plumes. If a speed criteria were used, then there would be a negative correlation between the plume depth and initial shelfwater velocity. Instead, Figure S on page 23 shows a pos- itive correlation between plume depth and initial shelfwater velocity. 20 o o _ c UD y CD X ^-^ IT! ~ z: ^ / DEPTH 0 500.0 1 ■9 jT LJ o -J a. o / ■ a 1 a ^ CD o CO o a 1 ( 2.0 5.0 10.0 1 1 1 1 1 15.0 2C.0 25.0 3C.0 35.0 1 4C.0 CFPSHGRE WIND VELOCITY (M/S) Figure 6. Effect of varying offshore nind speed on the 90-day plume depth: The polyna's increased width uith increasing wind speed results in a higher density plume due to longer exposure of the water column to the polynas brine rejection. This results in a deeper penetrating plume. Similarly, the parameters of shelf slope and secondary slope appear to have contra- dictor}' results. Figure 9 on page 24 shows as the initial shelf slope is lessened, the plume penetrates to deeper depths after 90 days of travel. In this experiment, the shelf width was fixed at 20 kilometers after the 45 meter isobath. This deeper penetration appears to be due to a lower entrainment parameter, which is a function of the shelf 21 o o o ' CD o y^ DCPTH 500, * j/^ PLUME 400.0 1 / >-> Q / 1 o gg- ^^ ' a o o 30.0 31.0 32.0 33.0 34.0 1 35.0 INITIRL SfiLINITY [PPT) Figure 7. The effect of initial siielfwater salinity on the 90-day plume depth: With increasing shelfwnier salinity the plume attains a higher density due to increased salt rejected by ice growth added to the shelfwater. slope. In contrast. Figure 10 on page 25 shows deeper penetration with an increased secondary slope. Although the entrainment parameter still increases with slope, the 90-day plume depth is limited by the slope itself. Since 90 days was considered a rea- sonable time scale for the plume to maintain its integrity, deeper plume penetrations would most likely occur with steeper slopes after the shelf break, i.e. in submarine T) G.Ol 0.02 0.03 0.04 0.05 0.06 INITIRL SHELF/IRTlR VELCCITY (.^/S) Figure 8. Effect of varying initial slielfnater velocity on 90-day plume depth: An increased initial velocity results in a deeper penetrating plume at 90 days. canyons. Varying the shelf width after the 45 meter isobath has little effect on the 90-day plume depth. This is shown in Figure 1 1 on page 26. Examining the stratification of the Arctic Ocean is accomphshcd by var>ing the Briint-Vaisala frequency of Layers 1 and 2. The deep Layer 3 remains constant at a value used in Kilhvorth (1977) for deep waters. Layer 1 is the initial pycnocline and re- presents the transition from the fresh surface layer to the saltier halochne. Figure 12 on page 27 shows the 90 day plume depth is little changed even with a substantially 23 ( o LO DEPTH (M) 0 550.0 ■• M o _I K >-• cn c 1 LO \ •■ o o o ^_ 1 ^"^^--^_^__ 1 1 1 1 1 1 2.0 3.0 4.0 5.0 6.0 7.0 8.0 1 9.0 INITIAL SHELF SLOPE (M/M) KlO'' Figure 9. This figure shows the effect of the initial shelf slope on the 90-day plume depth: Due to the entrainment parameter being a function o[ slope, a steeper slope results in higher entrainment of ambient water, thus more rapid loss in negative buoyancy. weakened mitial stratification with .V = 2.0.vlO-* sec^ The short length of time and space which the plume is exposed to this layer is probably the cause of its diminished impact on the 90-day plume depth. 24 a o c in ^ o Q o PLUfiC 400. 1 d a cn c o o r^ _ 2.0 1 1 I 1 3.0 4.0 5.0 6.0 7.0 8.0 1 9.0 SlCondrry slope (M/M) KlO'' * Figure 10. The effect of varying tlie secondary slope on the 90-day plume depth: Even with a higher entrainment rate, the plume reaches a deeper depth with a steeper slope, due to the physical boundary of the ocean bottom. On the contrary. Figure 13 on page 28 shows the dramatic difTerence in the 90-day plume depth with small changes in the Briint-Vaisala frequency in this 2-^5 meter thick layer. Additionally, if the second layer's thickness is decreased, the plume will penetrate 25 Figure 11. The effect of vaning the shelf width (after the 45 m isobath) on the 90-day plume depth; Varying the shelf width has little cllcct on the 90-day plume depth. to deeper depths. This is illustrated in Figure 14 on page 29. Note the discontinuity in the slope o[ the curve is caused when the plume no longer pentrates the second layer. Next, the standard case is compared in detail with a "submarine canyon" simulation. 26 o o o " - o o ,-. LP " ^_ o LJ Q o - ♦ LJ c 7~ LH ZD ^ _J CI- c - az ^ Q 90- 350.0 -• o o o 1 .0 2.0 ' 1 1 1 1 i 3.0 4.0 5.0 6.0 7.0 8. LRYlR 1 N VRLUE 0 9.0 1 10.0 ^10"' Figure 12. The effect of varying the initial pycnocline strength on the 90-day plume depth: The initial pycnoclmc. between 45 and 55 meter depths, has a small effect on the 90-day plume depth due to the small thickness of this layer. C. THE "SUBMARINE CANYON" CASE The "submarine canyon" case is a modification of the standard case using a steeper secondary slope, lower Layer 2 Brlint-Vaisala frequency value, and a decreased Layer 2 27 a o CD ~ CD DEPTH (M) 0 700.0 1 ■^- \; r, 1 ^ _1 Q_ \. CZ a 1 o cn \ — - — a o o 1 1.0 2.0 1 i 1 1 : 3.0 4.0 5.0 6.0 7.0 LflYLR 2 N VRLUE 8.0 1 9.0 Figure 13. The effect of varying Layer 2 Briint-Vaisala frequency on the 90-day plume depth: Due to the plumes longer exposure time in this layer, changes in Briint-Vaisala frequency tend to have a more dramatic cflcct on the vertical motion of the plume. thickness. In addition, the angle /?. is forced to the initial value of 29 degrees to simulate the steermg of the plume by topography. The initial values used in the "submarine canyon" case are shown in Table 5 on page 31. Forcing the angle p towards its initial 28 o o _ o CO o 5 - \^^ DCPTH 600.0 \ ■ PLUME 500.0 1 \' >-• d \ CD d_ ^ 2 \ o v_ o o f^ _ 50.0 1 1 1 1 1 1 IGC.O 130.0 2C0.G 250. 0 :00.0 353.0 LnY[:R 2 THICKNESS (ri) 1 4C0.0 Figure 14. Decreasing Layer 2 thickness results in a deeper penetrating plume: As the plume can no longer penetrate Layer 2, due lo its thickness, the 90-day plume depth becomes constant. value results in the plume travelling down the Y axis further and less down the X axis. Of course, the value chosen for forcing /? will aher the X-Y flow path of the plume. The How path of the "submarine canyon" case is seen in Figure 15 on page 32. The plume, descends to a depth of over 1300 meters in 90 days, has moved 207 kilometers along- shore and 160 kilometers ofT-shore. In Figure 16 on page 33. the "submarine canyon" plume is seen to reach a maximum velocity of 0.36 m s whereas the standard case plume 29 has a maximum velocity of 0.30 m s (Figure 3 on page 17). The eflect of increased slope, the angle /?. and velocity on entrainment is seen when comparing the volume flux of the "submarine canyon" plume (Figure 17 on page 34) and the standard case plume (Figure 18 on page 35). D. RESULTS USING OBSERVED TEMPERATURE AND SALINITY FIELDS The values for this experiment are the same as the standard case except the Brlint- Vaisiila frequency is computed using observed values of temperature and salinity. The temperature and salinity records are from Ostlund (19S7) and Aagaard et al. (1981) which were taken in the Canadian Basin. The results of the experiment are shown m Table 4. The 90-day plume depth and maximum velocities are very similar to the results from the Layer 2 Briint-Vaisala frequency sensitivity experiment. In an attempt to model a current as observed in Aagaard et al. (1985). a plume width of 20 kilometers was used together with a steep secondary slope and the temperature saUnity fields reported in Aagaard et al. (1981). The plume's velocity with increasing depth is shown in Figure 19 on page 36. Aagaard et al. (1985) reported a 25 kilometer wide. 15 meter deep plume with a mean velocity of 0.45 m, s. The simulation yields a velocity of 0.30 m s before the shelf break and 0.92 m, s while descending the steep secondary slope. Table 4. THIS TABLE COMPARES SOME RESULTS FROM USING ACTUAL T-S RECORDS: The AI\VEX-2 and AIWEX-8 stations are from Ostlund( 1987) and the third record is taken from Aagaard et al. (1981). Variable AI\VEX-2 AIWEX-8 Aagaard f'n,3x m s .397 .389 .395 90-day depth, m 217 213 305 X distance, km 311 312 363 ^' distance, km 380 375 478 30 Table 5. INITIAL VALUES USED FOR THE "SUBMARINE CANYON" CASE: The highlighted values are those parameters which difier from the standard case. \'ariable \ alue Definition 7"..... C -15. air temperature T..,.'C -l.S water temperature r'.,. ni sec"' 15. Wind \elocity r,. m sec"' 3 "''^ y„. . ice tloe velocity // . m ■> frazil collection thickness a. U'fn~-(icg~-^ 5.67.vl()-^ Stephan-Boltzmann constant t'. 0.95 emissivity of the air p,kgm~- 1.30 air density p .. kgnr' 1.026.vlO^ seawater density p kgnr^ ().95xny ice density c, 2.0.vlO-^ sensible heat coelTicient C. Jdcg-'kg'' 1004 specific heat of air Q^. ll'nr' 301 longwave radiation upward L. J kg-' 3.34.vl(>' latent heat of fusion h. . m 10. How thickness '►V, . in 1000. ilow width r,, ni sec~" .04 (low speed P . degrees 29 How direction /. sec-' i.38.vlO-^ Coriolis parameter .V. see-' 0.0316 Brtint-Vaisala frequency, layer 1 N 0.006286 BriJnt-Vaisala frequency, layer 2 AZ, . meters 145 Layer 2 thickness N 0.001 Briint-Vaisala frequency, layer 3 f^ 0.5.rl0-^ initial bottom slope 0: 9.rl0-5 secondary bottom slope K 0.01 drag coelficient E 0.072 sin 6 sin /) entrainment coelTicient S .ppt 32.5 initial shelfwater salinity S, . ppt 7.0 Salinity of frazil ice 31 2 ° 'x ^' ^ ^ >-> •. c az 1 — ( ^^^ a ^^ 1 -SLOPE 1.0 ^^ 1 CD O -J / o y a _L y^ G-0 I'.G 2 0 F.LONG-SHCRE DI5TR,\'CE 'X' [\ 1 3.0 1) HlO' ■- ■ Figure 15. The "submarine canyon" plume's flon path: The plume s How paih is forced to the initial /> value to simulate the topographic steering within a submarine canvon. 32 Figure 16. The velocity of a "submarine canyon" plume with increasing depth: The plume has the characteristic "jump" in velocity after the shelf break where it reaches a maximum velocitv of 0.36 m s. 33 o LH - X =D —\ c un m CD cr: z: ;/ a / ' — ' o a -21 / -• o / o 1 1 1 CO 1.0 2.0 1 3.0 fiLGNG-SHGRE DISTnNCE 'X' CD >^iO' Figure 17. The normalized mass flux of the "submarine canyon" plume: The plume increases in mass by over 6800° o indicating only 1 5°o of the water reaching the deeper depths is shelf water. 34 o CO - MRSS FLUX 12.0 1 y^ / 6.0 1 / o o _ 0-0 1.0 2.0 RLQiNG-SHGRE DISTnNCE 'X' [fi] 1 3.0 Figure 18. The normalized mass flux of the standard case plume: The standard case plume . 35 r\] — ~ o — ~ , — > * CD \ \ \ ^ CO \ \ >H \ f— \ X^^ _J \s^ M N<^^ :> N . TT N. ^ °" \^ ID N. ^ / \ ^ -^ \ -■ o ~ \ a \ o _. ' CO 200.0 400.0 1 500.0 eoo.o Dl 1 ir/ip 1 Lui IL nrPTu u L^ 1 i n (M) Figure 19. \ elocity shown N\ith depth of a 20 kilometer uide plume: This figure shows a wider plume can penetrate deeper than the narrower 1 kilo- meter plumes used in the previous experiments. 36 IV. DISCUSSION This paper has attempted to provide a mechanism in which high saUnity, low tem- perature water is produced and transported oil the continental shelves into the Arctic halocline or deeper depths. The environmental parameters which alTect the production and transport of this cold, salty water were examined, resulting m the conclusion that a polyna can produce water with a salinity greater than 34.5 ppt near the freezing pomt which can be transported via gra\ity flow plumes. The results show deeper penetrations to depths greater than 700 meters is possible depending on the set of environmental pa- rameters. This indicates polyna-produced high density plumes may also be responsible for the ventilation of the deep Arctic Basins, specifically the Canadian Basin. Ostlund et al. (1987) placed the residence time of Canadian Basin Deep Water be- tween 500 and 800 years. Lsing the mixing ratio of 2:1 (shelfwaler to intermediate wa- ter) from Aagaard et al. (1985), shelf water with a salinity of 35.1 ppt requires an otT-shelf flu.x of approximately .0063 Sv to ventilate the Canadian Basin in 800 years. The results from the 20 kilometer wide plume experiment give an annualized flux of 0.0035 Sv. In other words, an average of about two such events per year where polyna-produced plumes penetrate below the temperature maximum over the past 800 years could account for the ventilation of the Canadian Basin. Furthermore, Ostlund et al. (1987) states that the deep water probably consists of much less than lOSo shelfwater and would have to be a salinity of 36.15 ppt. The "submarine canyon" case shows the plume reaching a depth of over 1300 meters consisting of only 1.3'^/'o shelfwater at this depth. Table 3 on page 19 shows under extremely cold conditions, a salinity value of 36.23 ppt can be achieved with this model. Therefore, it is reasonable to expect that polyna-produced saline water may significantly contribute to the venti- lation of the Canadian Basin. Climatological variations which would aflfect the number of deep penetrating plumes are worthy of discussion. Besides the major periods of global ice ages and their counterpart warm eras, there have been recorded several shorter cold and warm epochs. These shorter periods coincide with weak and strong wind circulation, respectively. In the 20th century, both a warm period lasting about 20 years and a colder period begin- ning around 1950 have been recorded. In each of these two eras, there has been a sud- den outflow of surface lavers from the Arctic Ocean. Increased ice flow was recorded 37 in 1938 around Iceland during the warm period, while an increased flow of cold, low salinity water reached Iceland from the Arctic in the 1960's. Removal of the colder, fresher surface layer from the Arctic Basin would reduce the vertical temperature and saUnity gradients resulting in a less stable vertical structure for the ocean. In elTect. an environment similar to reducing the model's Layer 2 Brlint-Viiisala frequency and re- ducing the Layer 2 thickness may occur, enhancing deep plume penetrations. In addi- tion to possible weakening of the vertical structure, removal of the ice cover leaves large areas of open water where heat is more readily remo\ed from the ocean. In these large areas of open water, frazil ice grows rapidly and abundantly to increase the brine rejected into the water column. In the warm period of the early 20th century, the extent of the Arctic's semi-permanent ice pack was greatly reduced (up to \0%) between 1920 and 1938. The reduction in the semi-permanent ice pack is attributed to higher ice removal rates (increased advection) due to the strong wind circulation of the period. In Ostlund et al. (1987). it is noted that the deepest waters in the Canadian Basin may be the re- mains of the warm climate which occurred around 1000 A.D. vice any cold epochs such as the "Little Ice Age" in the late 17th and early 18th centuries. In summar>". periods when Arctic wind circulation is strengthened resulting in open water due to rapid ice divergence coupled with a weaker vertical Arctic Ocean stability seems to be the most conducive environmental scenario for deep plume penetrations to provide the additional salt source in the Canadian Basin. In Killworth and Smith (1984). the heat diffusivity constant, k , required to achieve steady state balances requires some peculiar changes in magnitude with increasing depth. A value of k.2i3 x 10-'^m'5"' is established to achieve a steady balance between the non- turbulent shelfwater plume used in Killworth and Smith (1984) and the region where the plume interleaves (note: the density of this plume is based on the brine rejection due to a single year's ice growth over a continental shelf of 50 meters average depth). To properly supply the East Greenland current outflow, k was required to be ^3 X 10--m^ sec-'. And finally, to balance the infiow of warm Atlantic water and upward advection of heat, k at this level was required to be :i:1.2 x lO-^m^ sec' If the normalized mass flux of the turbulent plume model was to be used as a gauge to understand the turbulent diffusion, this thesis would present some interesting results. In Melling and Lewis (1982). the normalized mass flux was of order ^10°. In the standard case, the normalized mass flux was of order ^10'. The "submarine canyon' case had a mass flux of order ::ilO-. Could it be that the three above cases provide a insight into the missing 38 physics so stated in Killvvorth and Smith ( 19S4)? Some other missing physics may be the double dilTusion process as outhned in Carmack. and Aagaard (1973) and the breaking of internal waves (Perkin and Lewis. 1978). Aagaard et al. (19S5) suggests the incorpo- ration of turbulent plume dynamics into the "filling box" model of Killworth and Smith {19S4). A mechanism providing a specific low temperature, high salinity water source has been developed in this paper. Although the mechanism is episodic, as suggested Aagaard et al. {19S5}, {Killworth and Smith (1984) had earlier used the term "spasmodic"), due to the dependence on environmental variables. llu.\ rates of this highly saline water can be estimated from renewal times of the Canadian Basin. The problem of incorporating the turbulent plume in the "filling box" model may be tackled, but is beyond the scope of this paper. The maintenance of the Arctic halocline by the mechanism presented in this study alone would require many times the number of active polynas producing saline shelfwater. Using the 3:2 mixing ratio proposed by Aagaard et al. (1981\ a minimum of 2.5 Sv production rate of cold, saline shelfwater was determined necessary to replenish the halocline over a period of 10 years. With the tlux rate of the 20 km wide plume (.0035 Sv), the individual mechanism modelled in this paper would require 200-300 coastal polynas producing 2-3 high salinity, low temperature plumes per year to maintain the Arctic halocHne. Although 200-300 such polynas may be thought prohibitive, areas of open water which exist during the onset of freezing temperatures in the late fall, as well as leads which continue through winter over as much as lO'^o of the ice pack, could provide a rapid source of salt to help overcome the summertime salinity gradient. In times of strong vertical structure, these polyna-produced plumes can transport low- temperature, high salinity water into the Arctic halocline. Overall, the polyna seems to be an important source of brine which may help maintain the Arctic halocline and ven- tilate the deep basins. 39 REFERENCES Aagaard. K.. 1981. On the deep circulation in the Arctic Ocean. Deep-Sea Research. 28, 251-268. Aagaard, K., L. K. Coachman, and E. Carmack. 1981. On the halocline of the Arctic Ocean. Deep-Sea Research. 28, 529-545. - Aargaard. K., J. H. Swift, and E. C. Carmack. 1985. Thermohaline circulation in the Arctic .Mediterranean seas. Journal of Geophysical Research. 90, 4833-4846. Bo Pedersen. F., 1980. Dense bottom currents in a rotating ocean. Journal of Hydraulics Division. 106. 1291-1308. Carmack, E. and P.D. Kilhvorth, 1978. Formation and interleaving of abyssal water masses ofT Wilkes Island, Antarctica. Deep-Sea Research, 25, 357-369. Carmack, E. and K. Aagaard, 1973, On the deep water of the Greenland Sea. Deep-Sea Research, 10, 6S1-1\5. Cox, G. F. N. and W. F. Weeks. 1974, Salinity variations in sea ice. Journal of Glaciology, 22, 853-873. Helland-Hansen, B., and F. Nansen, The Norwegian Sea: Its physical oceanography based upon the Norwegian researches, 1900-1904. Rep. Xorw. Fish. Mar. Invest., 2(1). 1909. Kilhvorth, P. D., 1977, Mixing on the Weddell Sea continental slope. Deep-Sea Re- search, 24. 427-448. Kilhvorth, P. D.. and J. M. Smith, 1984. .A one-and-a-half dimensional model for the Arctic halocline. Deep-Sea Research. 31, 271-293. 40 Lamb. H. H.. Climate: Past, Present, and Future, Volume 1, pp, 256-263, Methuen and Co.. LTD. 1972. Melling. H.. and E. L. Lewis, 19S2. Shelf drainage flows in the Beaufort Sea and their elfect on the Arctic Ocean pycnocline. Deep-Sea Research. 29, 967-9S6. Millero. R. J. and A. Poisson. 1981. International one-atmosphere equation of state of scawater. Deep-Sea Research, 28, 625-629. Nansen, F.. Northern waters: Captain Roald Amundsen's oceanographic observations in the arctic seas in 1901, Xor. I'idensk. Acad. Kl. Skr. Mat. Xaiurvidensk. KL, 1(3). 1906. Ostlund. H. G.. G. Possnert. and J. H. Swift, 1987, Ventilation rate of the deep Arctic Ocean from Carbon 14 data. Journal of Geophysical Research, 92, 3769-3777. Ou, H. W., 19SS. A time-dependent model of a coastal polyna. Journal of Physical Oceanography, 18. 584-590. Pease, C, 1987. The size of wind driven coastal polynas. Journal of Geophysical Re- search. 92, 7049-7059. Perkin, R. G. and E. L. Lewis, 1978, Vlixing in an Arctic fjord. Journal of Physical Oceanography, 8, 873-80. Schumacher, J. D., K. Aagaard, C. Pease, and R. B. Tripp, 1983, Effects of a shelf polyna on How and water properties in the Northern Bering Sea. Journal of Geophysical Research, 88, 2723-2732. Smith, P. C, 1975, A streamtube model for the bottom boundary currents in the ocean. Deep-Sea Research, 22, 853-874. Sverdrup, H. U., 1929, The waters on the north Siberian shelf The Sorwegian Sorth Polar Expedition with the "Maud", 191S-1925. Scientific Results, , 4(2).. Bergen. 41 Weber, J. R., 1979, The Lomonosov Ridge Experiment: 'Lorex 79'. £05 Transactions of [he American Geophyisical Union. 60. 715-720. Weeks, W. F.. and S. F. Ackley. 19S2, The growth, structure, and properties of sea ice. CRREL Monograph S2-1. U.S. Army Cold Regions Research And Engineering Laboratory.', pp. 130. 42 INITIAL DISTRIBUTION LIST No. Copies nclensc Technical Inrormntici Center • 2 Cameron Siation Alexandria, VA 22304-6145 2. Library, (^ode 0142 Naval Postgraduate Schooi ^^^r^te■',-v (,\ 9394,1-5002 3. Chairman (Code oSC/o) Department of Oceanography Naval Postgraduate School Monterey. CaliTornia 93943-5000 4. Chairman (Code 68Rd) Department of Meteorology Naval Postgraduate School Monterev. California 93943-5000 5. Dr. Knut Aagaard NOAA/PMEL/MSRD 7600 Sandy Point Way , NT: Seattle, Washington 98115 6 Mr. Steve Ackley L'. S. Arrny Cold Regions Research and Fuginecring Laboratory 10 Lyme Road Hanover. New Hampshire 7. Maj. WilriamC. Hill (Ret.; Route 3 liox 420 Sealy. Texas 77474 8. ^"r. Humphrey Melling histitute of Ocean Sciences 1'. O. Box 6000 9860 West Saanich Road Sidney, British Columbia C\anada V8L 4B2 9. Dr. H. Gote Ostlund Tritium Lab L'nivcrsity of Miami Rosentiel School of Manne and ,\rmospheric Science 4600 Rickenbacker Causeway Miami, i'lorida 33149 43 10. Dr. Albert .1. Scmtner (Code 6SSc) l)cpt. of Oceanography Naval Postgraduate School Monterey. California 'T3'^43 1 .. i)r. F",ddy (?armack HKtitute of Ocean Sciences P.O. Box 600() 9860 West Saanich Road Sidney, British (Columbia Ca,-.avia VSf. 4B2 I ?-. Dv. William Hibier {haver Scltnol of Fnginccring Darmouth l;ni\'ersity Hanover. New Ilamp'Jhire 1 3. Or. (^aroi Pease NOAA/PMBL^MSRi:) 7600 Sandy Point Way, NF Seattle, \^''ashington 9811.^ 14. Dr. .lames Swift Scripps institute of Oceanosraphy MLR (iroup, A030 : ■; ; ^lln. (A i.S. Dr. Peter Killworth Robert Hooke Institute Dept of A":^nspheric Physics Parks Rd. Oxford OXl 3 PI: England 16. Dr. Peter Wadhams Scott Polar Research Institute Can:bridge. CB2 lER, imgland :7. Or. Robert Bourke (Code 68B0 Department of Oceanography Naval Postgraduate School Monrerey. California 9";943-.5000 18. Dr. D. C. Smit:., IV (Code 6SSi) Department of Oceanography Naval Postgraduate School Monterey, CA 9394^ ^'onr, 19. Sciciific Liaison Office Olfice of Naval Research Scripps Institute of Oceanography La folia, (\alifornia 92037 44 ^0. Dr. Bernard i.ettau Office of Poiar Research National Science Foundation '.Vashmgton. DC 205^0 2i. Dr. C. \. K. Mooers, Director Institute of Navai Occanographv Bldg. 1100. Room 3i! XS'I'K Station Bay St. I.vyuis, Mississip-^- V)52'^ 22. Mr. Michael Steele Polar Science (^enter \'~'' 1013 NH 40th Street Seattle. Washington 9810'^ 2"^ Commading Offcer Naval Ocean Research .i id Development .Xctivitv NSTL Station Bay St. I.ouis, Mississippi 3Q.'^22 ?4. OlTice of Naval Resern.h ((^odc 1 I 22P) 800 N. Quincy Street Arlington, Virginia 22217 25. Chairman, Oceanography Department I;. S. Naval Academy Annapolis, Maryland 21402 45 %\!S M^^ \\ < lueticess -■ Arc- ep Thesis H5293 c.l Hill EnviT-onmental influences on the production of Arc- tic halocline and deep water. ^^B'NOfJ v£ft«