INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI films the
text directly from the original or copy submitted.
Thus, some thesis and
dissertation copies are in typewriter face, while others may be from any type of
com puter printer.
The quality of this reproduction is dependent upon the quality o f the copy submitted.
Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleedthrough, substandard margins, and improper alignment
can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript and
there are missing pages, these will be noted. Also, if unauthorized copyright
material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning
the original, beginning at the upper left-hand comer and continuing from left to
right in equal sections with small overlaps. Each original is also photographed in
o n e exposure and is included in reduced form at the back of the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6” x 9” black and white photographic
prints are available for any photographs or illustrations appearing in this copy for
an additional charge. Contact UMI directly to order.
UMI
Bell & Howell Information and Learning
300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA
800-521-0600
The Buoyancy Forcing and Dynamical Response of
the Red Sea
by
Eleni-A nthi Tragou M.Sc.
A Thesis su b m itte d in P artial Fulfillm ent of the R equirem ents for the Degree of
D o c t o r o f P h i l o s o p h y
[N THE Sc h o o l o f Ea r t h a n d Oc e a n Sc i e n c e s
We accept this thesis as conform ing to th e required stan d a rd
Dr. C. G a rre tt, Supervisor (in th e School of E a rth an d O cean Sciences)
---Dr. I. Fung, D epajLm ental M em ber (in the School of E arth and Ocean Sciences)
Dr>^j0fM ||ie(^, D e p a rtm e n ta l M em ber (in th e School of E a rth an d O cean Sciences)
Dr. A. W eaver, D ep artm en tal M em ber (in th e School of E a rth and Ocean Sciences)
Dr. R. S tew art, O utside M em ber (D ep artm en t of Physics)
Dr. P. M alanotte-R izzoli, E jiternal E xam iner (D ep artm en t of E a rth , A tm ospheric, and P la n e ta ry Sciences, M IT)
@ Eleni-A nthi Tragou, N ovem ber 9, 1998
Un i v e r s i t y o f V i c t o r i.a.
All rig h ts reserved. This thesis m ay not be reproduced in whole or in p a rt, by photocopy or other m eans, w ithout th e perm ission of th e author.
A b s tr a c t
T h e buoyancy forcing of th e R ed Sea and its dynam ical response are ex am in ed . Buoyancy transports th ro u g h th e S tra it of Bab el M andab, th e m ajo r oceanic exchange point of the Red S ea w ith th e open ocean, provide a stro n g con s tra in t on th e surface buoyancy fluxes. H ydrographic d a ta and cu rren t records at th e S tra it require th e annual m ean surface heat flux to be — 8 ib 2 W m~^. For the a n n u a l m ean freshw ater fluxes th e conservation of volume and salt give the net e v ap o ratio n ra te as 1.60 ± 0 .3 5 m y ~ ’^.
T h e surface fluxes estim ated fro m th e heat and freshw ater tra n sp o rts at the S tra it are com pared to the annual m ean surface fluxes estim ated from stan d a rd m eteorological d a ta sets and form ulae used on a global scale as in th e revised Com prehensive O cean-A tm osphere D a ta Set (U W M /C O A D S ). T he difference betw een th e surface heat fluxes and th a t im plied by th e exchange th rough th e S tra it is large a n d close to 100 W m~^. A large p o rtio n of this difference is explained by the o v e re stim a te d solar irradiance due to th e neglect of sp atial and seasonal variations o f aerosol concentration, and m isap p licatio n of a sta n d a rd form ula for insolation. A n o th e r p o rtio n of th e difference com es from th e u n d erestim ated longwave ra d ia tio n d u e to th e use of a bulk fo rm u la which is adequate for th e open ocean b u t in ap p ro p ria te for th e Red Sea. T h e evaporative losses are also found to be u n d e re stim a te d , probably because of u n d e re stim ate d wind speeds. T h e net evap o ra tio n is th e m ain contributor to th e an n u al m ean buoyancy loss ap p ro x im ately of 2 X 10"®m^s~^.
T h e annual m ean surface buoyancy flux, which is com patible w ith th e oceanic buoyancy tra n sp o rt, is used w ith P h illip s’ sim ilarity m odel to investigate
I l l
th e buoyancy driven flow of th e upper 140 m of th e Red Sea. T h e observed s tr a ti fication of th e Red Sea can be achieved only w ith a very large eddy viscosity in th e re tu rn flow. It is possible th a t this high vertical viscosity could be a proxy for pro cesses neglected by this m odel such as b o tto m friction on the sloping boundaries. T h e effect of wind stress is sm all, but a southw ard wind com bined w ith the b o tto m friction of a m odified m odel w ith dep th -d ep en d en t basin w idth could account for th e viscous force required by a model.
T h e effectiveness of the bottom friction in retard in g th e flow depends on th e m ag n itu d e of th e la te ral diffusion of m o m en tu m . To explore th e possibility of m easuring th e horizontal m om entum fluxes above a sloping boundary in a channel, we perform ed an experim ent in the S trait of G eorgia w ith two A coustic D oppler C u rre n t Profilers. A lthough further investigation of such m easurem ents is required and several issues rem ain to be resolved, it is show n th a t an estim ate of the hori zo ntal ed d y viscosity acting on the tidal cu rren ts is possible w ith this m ethod an d gives a b o u t 50 m^ s~^.
O verall, th e dynam ics of the Red Sea ap p ea rs to be determ ined m ainly by th e surface buoyancy fluxes and internal a n d la te ra l frictional forces. A ccurate m odelling of th e R ed Sea requires im proved know ledge of th e forcing and of an a p p ro p ria te p aram eterizatio n of the friction.
Exam iners:
Dr. C. G a rre tt, Supervisor (in the School of E a rth an d Ocean Sciences)
Dr. I. Fung, D ep artm en tal M em ber (in the School of E a rth and Ocean Sciences)
Dcj^^kT Lueck, D ep artm en tal M em ber (in the School of E a rth and O cean Sciences)
Dr. A. W eaver, D epartm ental M em ber(in the School of E a rth and O cean Sciences)
Dr. R. S tew art, O utside M em ber (D epartm ent of Physics)
Dr. P. M alanotte-R izzoli, E x tern al Exam iner (D e p a rtm e n t of E arth , A tm ospheric, and P la n e ta ry Sciences, M IT)
T able o f C o n te n ts
A b s t r a c t ... ii
Table of C o n te n ts ... v
List o f T a b l e s ... viii
List of Figures ... x
A cknow ledgm ents... xix
C h a p te r 1. I n tr o d u c tio n ... I 1 . 1 M otivation an d thesis outline ... 1
1 . 2 G e o g ra p h y ... 3
1.3 C lim ate c o n d i t i o n s ... 5
1.4 H ydrographic s t r u c t u r e ... 7
1.5 C irc u la tio n ... 13
1.5.1 Shallow c i r c u l a t io n ... 13
1.5.2 Deep circulation: fo rm a tio n a n d renewal t i m e ... 15
C h a p te r 2. T h e S trait of Bab el M a n d a b ... 17
2.1 I n tr o d u c tio n ... 17
2.2 Oceanic heat and freshwater tra n sp o rts through the S tra it of Bab el M a n d a b ... 18
2.2.1 Review of annual e s ti m a t e s ... 18
C h ap ter 3. Surface Buoyancy Fluxes of th e Red S e a ... 30
3.1 I n tr o d u c t i o n ... 30
3.2 Surface h eat flux e s t i m a t e s ... 31
3.2.1 Solar i r r a d i a n c e ... 34
3.2.2 Longwave r a d i a t i o n ... 44
3.2.3 L atent an d sensible heat flu x e s ... 48
3.3 Im plications from th e w ater b u d g e t ... 49
3.4 T otal heat f l u x ... .51
3.5 Buoyancy f l u x ... 53
3.6 S um m ary an d d i s c u s s io n ... 56
C h a p ter 4. T h e Buoyancy Driven C irculation of th e Red S e a ... 61
4.1 I n tr o d u c t i o n ... 61
4.2 A sim ple m odel of the Red S e a ... 65
4.2.1 Governing equations and sim ilarity f o r m s ... 65
4.2.2 Do observations su p p o rt th e sim ilarity s c a lin g ? ... 6 8 4.3 P a rtic u la r s o lu t i o n s ... 73
4.3.1 Sensitivity to m ixing c o e fflc ie n ts... 77
4.3.2 C irculation for th e default solution an d com parison w ith o b s e r v a tio n s ... 81
4.4 T h e effect of ro tatio n an d cross-channel f l o w ... 84
4.5 Is b o tto m friction responsible for th e a p p are n t high viscosity? 8 6 4.6 Solution for a basin of dep th -d ep en d en t w i d t h ... 8 8 4.7 T h e effect of w i n d ... 91
4.8 M atching th e solution to th e S tra it of Bab el M a n d a b ... 93
4.9 C om parison w ith lab o rato ry e x p e r im e n ts ... 97
vu
C h a p te r 5. M easurem ents of H orizontal Eddy M om entum F l u x e s ... 104
5.1 M otivation an d i n t r o d u c t i o n ... 104
5.2 S trait of Georgia: m om entum balance and s c a l i n g ... 107
5.3 T h e e x p e r i m e n t ... 112
5.4 D ata a n a l y s i s ... 114
5.4.1 D erivation of velocity com ponents an d d a ta quality . . 114
5.4.2 L ong-term m ean c u r r e n ts ... 116
5.4.3 H arm onic a n a ly s is ... 118
5.4.4 Reynolds stress e s tim a te s ... 125
5.4.4.1 Ensem ble averages of —u'v' ... 128
5.4.4.2 Ensem ble average c o s p e c t r a ... 135
5.4.5 E stim a te of th e horizontal eddy v i s c o s i t y ... 136
5.5 C o n c lu s io n s ... 140
C h a p te r 6. Concluding R e m a r k s ... 143
List o f T ab les
2.1 M ean volum e tra n sp o rts, tem p eratu res an d salinities of the two layer exchange system during M arch 1924 (Verceili 1927), and the re su ltin g heat and salt transports, ad ju stin g th e volume tran sp o rt
of th e outflowing layer (upper) and inflowing layer (lower)... 25 2.2 V olum e tran sp o rts, tem peratures an d salinities of the three layer
exchajige system during sum m er 1982 [M aillard and Soliman (1986); Souverm ezoglou et al. (1989)], and th e estim ated heat and salt
tra n s p o rts ... 26 2.3 M ean volum e tra n sp o rts during N ovem ber 1964 from Siedler (1968),
a n d estim ates of heat and salt tra n sp o rt using clim atological tem p e ra tu re s and salinities (Levitus et al. 1994), adjusting th e volume tra n s p o rt of the outflowing layer (upper) an d inflowing layer (lower). 27
3.1 L ist of estim ates for th e mean surface solar radiation of th e Red Sea. 37 3.2 List of m ethods used to calculate net longwave radiation, and the
a n n u a l Red Sea estim ates using various d a ta sets... 46 3.3 List of exchange coeflhcients (in n e u tra l conditions) used to calcu
la te la te n t and sensible heat fluxes an d th e ir m ean values in the
R ed S ea... 49
4.1 L ist of th e p aram eters th a t determ ine th e flow in the tan k experi
IX
4.2 List of th e sim ilarity solutions for th e buoyancy and cu rren t, and th e actu al eddy coefficients halfway along the basins required by P hillips’ m odel to m atch th e observations from the two ta n k ex
1 . 1 B ath y m etric chart of th e Red Sea... 4 1 . 2 C lim atological m onthly m ean w ind stress (45-year averages from
1945 to 1989 from U W M /C O A D S )... 6 1.3 M onthly m ean clim atology of th e Red Sea for January, A pril, Ju ly
an d O ctober (d a ta from U W M /C O A D S ). (a) Sea surface te m p e r
a tu re in °C, (b) Sea surface salinity, an d (c) in kg m “ ^... 8 1.4 H ydrographic sections along th e axis of th e Red Sea from th e G ulf
of Suez to the S tra it of Bab el M andab using d a ta for th e poten tial te m p e ra tu re (upper panel), salinity (m iddle panel), and (lower panel) from th e cruise of Cdt. R. Giraud during Jan u ary -F eb ru a ry
1963... 10 1.5 H ydrographic sections along th e axis of th e Red Sea using d a ta for
th e p o ten tial tem p e ra tu re (u p p er panel), salinity (m iddle panel), an d (Tg (lower panel) from th e cruise of R.V. Atlantis during July-
A ugust 1963... 12
2 . 1 E stim ates of h eat and salt advected through th e S tra it of Bab el M andab. Crosses m ark results using P a tz e rt’s volum e tra n sp o rt, circles m ark Vercelli’s results, diam onds m ark M aillard and Soli- m a n ’s, an d squares m ark Siedler’s. T he annual m ean heat tra n s p o rt is equivalent to a surface heatin g of the R ed Sea of 8 W m~^.
X I
3.1 L ong-term averaged net shortw ave solar radiation estim ated from e q u a tio n (3.1) (w ith th e cloud correction) and cloudiness from
U W M /C O A D S ... 36 3.2 Net shortw ave radiation at three statio n s. Crosses m ark the es
tim a te d from equation (3.1) (w ith th e cloud correction), solid circles m a rk th e observed and open circles m ark th e corrected
using th e transm ission anom aly Tr* from satellite d a ta cali
b rate d to ground stations... 39 3.3 (a) Transm ission anomaly coefficient estim ated from com parison
of th e insolation at g ro u n d -tru th statio n s with the calculated (w ith th e cloud correction) during th e sam e period of tim e {Tr~ =
Q s^/Q s)- T h e dashed line is the m ean Tr" from th e GEBA s ta
tions, th e dashed-dotted line is T r" from the statio n a t Jeddah, the solid line is th e average betw een th e results at G EBA and Jeddah (assum ing 50% weights to G EBA an d Jeddah, i.e. 25% to each G EBA s ta tio n and 50% to Jed d ah ). T h e thick solid line is the sm oothed average Tr". T h e sm oothing is obtained with a 10th order polynom ial fit on th ree rep etitiv e annual cycles of the m ean
T r" an d keeping the second armual cycle, (b) C alibration factor
for th e sa te llite transm ission coefiScient index e stim ated from com parison of th e insolation from satellite d a ta with th e g ro u n d -tru th
records; th e line-style corresponds to th a t of (a )... 41 3.4 Transm ission anom aly coefficient T r " for the Red Sea from the
optical ex tin ctio n index calibrated to ground observations, (a) sea
3.5 Insolation corrected for the a tte n u a tio n due to aerosols, (a) Sea sonal cycle, and (b) sp atial d istrib u tio n ... 4 4 3.6 (a) M onthly m ean wind speed averaged over the Red Sea and fil
tered w ith a 23-point filter to rem ove seasonal variability, (b) La te n t heat flux estim ated from U W M /C O A D S ... 51 3.7 (a) Seasonal cycle of th e heat flux com ponents: th e dashed line is
th e insolation, the dash-dotted line is th e net longwave rad iatio n , th e th in sohd line is the laten t heat flux, th e d o tted line is the sensible heat flux, and the thick solid line is th e to ta l heat flux.
(b) S patial distrib u tio n of th e to ta l heat flux ... 54 3.8 T im e series of the heat flux com ponents; the line-style corresponds
to th a t of Figure 3.7a... 55 3.9 S patial d istribution of the m ean surface buoyancy flux, (a) T h e r
m al buoyancy flux, (b) Haline buoyancy flux, (c) Total surface buoyancy flux... 57
4.1 Schem atic of the buoyancy driven circulation of a semi-enclosed basin of length L, separated from the open ocean by a sill a t d ep th
h ... 66 4.2 (a) Sym bols m ark the stations of th e R.V. Atlantis along th e m ain
axis of th e Red Sea in M ay and J u n e 1958. (b) (Surface buoyancy)^'^^ as a function of horizontal distance. T he intersection w ith th e hor izontal axis is th e v irtu al origin of the basin, (c) Buoyancy profiles. (d) Buoyancy sim ilarity profiles according to P hillips’ scaling, i.e.
g(jl) = bh{BQx)~^^^ using h = 140 m and Bq — 2 x 10~®m^s“ ^. T h e
solid line is th e m ean profile and th e non-dim ensional vertical axis
X I I I
4.3 T he sam e as F igure 4.2 for statio n s of the Cdt. Robert Giraud in Jan u a ry an d F ebruary 1963. (a) Stations along th e m ain axis of the R ed Sea. (b) (Surface buoyancy)^^^ as a function of distance
from Suez, (c) B uoyancy profiles, (d) Scaled buoyancy profiles. . . 72 4.4 Solution of th e m odel for profiles of (a) fC and (b) VV given by
equations (4.25) and (4.26). T h e sim ilarity functions of buoyancy and velocity are shown in (c) an d (d). T h e actu al scales of the dim ensional variables a t th e m iddle of the basin (x = lO^m) are given a t th e top of each d iag ram (ac and u are given in m ^ s '\ buoyancy in m s “ ^, and velocity in m s “ ^), and the non-dim ensional vertical axis corresponds to a d ep th range of 0 — 140 m which is the sill d ep th . T h e stream lin es (continuous) and th e isopycnals
(dashed) are shown in (e )... 75 4.5 (a) Profiles of th e v o rticity eq u atio n (4.34) term s show a balance be
tween viscous an d buoyancy term s at all depths, (b) T h e buoyancy equation (4.35) term s show a depth-dependent balance betw een the
te rm s... 76 4.6 Solution for th e special case of a well-mixed layer (large vertical
eddy viscosity and diffusivity) above a perfect fluid (/C^ = = 0). (a) Diffusivity profile, (b) S tream function, (c) S im ilarity profile of th e velocity, aud (d) B uoyancy sim ilarity profile, (e) Stream lines
4.7 Contours of ^ (1) in th e space of eddy viscosity an d e d d y diffusivity of th e lower layer, ajid IC^ respectively. T h e o th e r param eters are A7q = 10, jVq = 100, H = 0.7, e = 0.1, and 7 = 0. T h e default solution is m arked w ith a s ta r (see te x t for th e a c tu a l values of the
dim ensional eddy coeflScients)... 79 4.8 A f{R i) for various profiles Af{T]) chosen by varying (a) th e transition
layer depth H , (b) th e thickness of th e tra n sitio n layer a, and (c)
p aram eter 7 in (1 2). O th e r p aram eter values are as for Figure 4.4. 80 4.9 (a-c) cTj sections from th e cruise of Cdt. R. Giraud in January
an d February of 1963 (d). (e) T he geostrophic velocity for section C is calculated from th e average relative velocity betw een each successive p air of sta tio n s assum ing th a t th e level of no m otion is
th e deepest m easu rem en t of each p a ir... 81 4.10 Buoyancy sections for th e default solution a t (a) 500km , (b) 1000km,
an d (c) 2000km from th e v irtu al origin... 83 4.11 Schem atic diagram of th e cross section of a basin of average width
W and sides w ith slope s ... 87 4.12 Solution of th e m odel for depth-dependent w idth of th e basin. The
values for th e lower layer of tC and A/* profiles (first two panels) are
/Cj = 0 .5 and A/^ = 12 r e s p e c tiv e ly ... 90 4.13 Solutions for (a) buoyancy, and (b) horizontal velocity, under the
influence of w ind stress. T h e continuous line rep resen ts th e default solution, the dashed line is th e solution for a n o rth w ard wind, and
X V
4.14 Schem atic of th re e possible states of circulation of a sem i-enclosed sea forced by surface buoyancy loss. S ta te (a): slow circulation, strongly stratified. S ta te (b): rapid circulation at th e overm ixed lim it. S tate (c): even m ore rapid flow, so th a t recirculation is req u ired... 9 4 4.15 Solutions for eddy diffusivity = 0 . 1 (solid line) and K,^ = 0.5
(dashed line), (a) norm alized buoyancy profiles w ith respect to th e surface value ^ (1), (b) horizontal velocity profiles (not norm alized),
an d (c) buoyancy profiles (not norm alized)... 96
5.1 Schem atic diagram of two possible d istrib u tio n s of m om entum across a channel for th e sam e w idth-averaged tra n sp o rt of volum e. The p arab o la on the left is th e solution to th e problem w ith large hori zontal viscosity b u t no intern al friction, an d th e profile on th e right hand side is the solution including in tern a l friction away from th e
boundaries for { r / A fjŸ ^ ^ L = 1 0... 106 5.2 M ap of th e southern end of Vancouver Island (u p p er panel). T he
area of th e experim ent is m arked w ith a border and expanded in th e lower panel, w hich also shows th e m ooring locations an d th e and th e fram e of reference ro ta te d to th e direction of th e tid al
currents at the two locations... 108 5.3 Cross-section of th e channel near th e sloping boundary, u'v' is th e
z-m o m en tu m tran sfer in th e ^-direction tow ards th e boundary, and
C ^ U ^ / h is the m o m en tu m lost in b o tto m frictio n ... 109
5.4 a^ profiles estim ated from six CTD casts tak en 14 days a p a rt near th e A D C P m oorings. Each panel shows th e profile on th e day
5.5 A a exam ple of th e th re e velocity com ponents in e a rth coordinates, derived from th e A D C P records in beam coordinates. T hese tim e series were e x tra c te d from a bin at 2 0 m below th e sea surface at
th e shallow m ooring... 116 5.6 Sam e as in Figure 5.5, but for a bin a t 40 m below th e sea surface
for th e deep m o oring... 117 5.7 D irection of th e m ean current estim ated from th e full 14-day record
a t th e shallow (left panel) and deep m ooring (right panel). Arrows in th e left panel in d icate the m ean-current direction a t each 1-m
bin, an d in th e right panel a t each 2-m b in ... 118 5.8 D ep th dependence of th e m ean current com ponents estim a te d from
th e full 14-day record a t the shallow (upper panels) an d deep (lower panels) m oorings. T h e fram e of reference has been ro ta te d —40°
w ith respect to th e east-w est axis... 119 5.9 Exam ples of th e tid a l fit (red line) com pared to the low-pass filtered
recorded com ponents (blue line) during th e 14-day record. T h e two u p p er panels correspond to th e tim e series a t 2 0 m d e p th for the shallow m ooring, an d th e two lower panels to th e tim e series a t 40
m d e p th for th e deep m ooring... 1 2 2 5.10 T h e co n trib u tio n of each tid al constituent to th e am p litu d e of the
tid al cu rren t a t various depths for th e shallow m ooring, as esti m a te d by th e harm o n ic analysis. T he vertical scale is correct only for th e deepest b in (33 m) and has been offset by a m u ltip le of
0.2 m s~^ for th e shallow er d e p th s... 123 5.11 Sam e as Figure 5.10, b u t for th e deep m ooring... 124
X V ll
5.12 D epth dependence of th e tid al c u rre n t am p litu d e (left panels) and th e inclin atio n of the tid al ellipse (rig h t panels) for th e four m ost im p o rta n t tid a l constituents (M^: solid line, O^: dashed line, K^: do t-d ash ed line and Sg: d o tted line) for th e shallow (u p p e r panels)
and deep (lower panels) m oorings... 126 5.13 E stim ates of Reynolds stress w ith d e p th a t th e shallow mooring
from ensem ble averages of th e correlatio n u'v' averaged over short blocks of d a ta which have been e x tra c te d around th e sam e tidal speed (in d ic a te d in m s“ ^ a t th e to p of each profile) an d during th e sam e tid a l phase (accelerating, m axim um , and decelerating in d icated a t th e top of each d iag ram ). T he horizontal scale is correct for th e first profile and has been offset by a m ultiple of 0.005 m^ for the subsequent profiles. T h e dashed lines indicate
6 8% confidence intervals... 130 5.14 Sam e as F ig u re 5.13, but for the deep m ooring. T h e horizontal scale
is correct for th e first profile and hcis been offset by 0 . 0 1 m^ s“ ^ for
th e subsequent profiles... 131 5.15 D ep th -m ean tid a l current a t th e shallow m ooring (solid line), th e
deep m ooring (dashed line) an d th e ir difference (dot-dashed line)
for one d ay of th e recorded tim e series... 133 5.16 Reynolds stress estim ates from ensem ble averages aro u n d m axi
m um tid a l speeds as a function of th e averaging period. Each line represents th e Reynolds stress a t a c erta in d epth. R eynolds stress values have been norm alised to th e m ean value for all averaging
5.17 E nsem ble averages of variance preserving c o sp ectra of the fluctuat ing p a rts u an d v' around the mean tid al c u rre n t a t various depths for th e shallow m ooring. D ata have been e x tra c te d from succes sive tid a l cycles in 2-h blocks (a) around a positive tidal current of 0 . 1 m s“ ^ at th e accelerating tidal phase an d (b) around a negative tid a l cu rren t of —0.1 m s” *^ at the deceleratin g tid al phase. The
vertical scale has been offset by —0.5 x 10"^ m^ s~^... 1.37 5.18 Sam e as F igure 5.17, b ut for the deep m ooring. T h e vertical scale
has been offset by —0.3 x 10“^ m^ s“ ^... 138 5.19 S c a tte r plot of th e depth-m ean estim ates of th e ensem ble averages
{u'v') arou n d various tidal speeds and phases, versus the horizontal
sh ear d U /d y of th e depth-m ean tidal c u rre n t. T h e panel on the left corresponds to th e estim ates at the shallow m ooring, and the panel
X IX
A ck n o w led g m en ts
I am m ost g ratefu l to m y thesis advisor, Prof. Chris G a rre tt, for the guid ance, patience, an d encouragem ent he has shown me at every stag e of my research. I am especially in d eb ted for th e financial support which he has provided to me over th e years. I am also g ratefu l to Prof. Inez Fung, for m any helpful suggestions and enjoyable discussions. I th a n k m y other com m ittee m em bers, Drs. Rolf Lueck, A ndrew Weaver, a n d Bob Stew art, for th eir insightful com m entary on m y work. I am also grateful to m y ex tern al exam iner. Prof. Paola M alanotte-R izzoli, for her insightful and co n stru ctiv e com m ents on m y thesis.
Thajiks are d u e to R ichard O uterbridge for his help w ith th e program m ing for th e analysis of th e U W M /CO A D S d a ta set. Prof. Jim Bishop is th anked for providing d a ta for th e surface solar irradiance and for helpful advice. T he field experim ent in th e S tra it of Georgia was conducted thanks to Dr. R ichard Dewey’s skillful design of th e m oorings, and help w ith the deploym ent an d recovery. I also th a n k him for p a tie n tly answ ering m y num erous questions ab o u t th e d a ta analysis. I am grateful to Dr. K a te Stansfield for her friendship and enjoyable discussions and helpful suggestions o n a variety of topics of m y work. I th an k D rs. A m it Tandon an d M ing Li for inspiring discussions. T hanks are due to Rosalie R utka for her help w ith the graphics and for proofreading p arts of my thesis. I would also like to th a n k m y office-mates and friends, X iaojun Jiang, K onstantin Zahariev, Michael O tt an d Keir Colbo, an d all m y friends and colleagues at SEOS for stim ulating discussions, help an d su p p o rt through the years and for m aking life at Gordon H ead Complex in tere stin g and enjoyable.
Special th an k s are due to all my friends here in V ictoria. In particu lar, I am m ost grateful to S m aro K am boureli for her friendship and never failing support during my years in V ictoria, and for m aking Canada my second home. Last, b u t not least, I w ould like to th an k m y fam ily for their love, encouragem ent and su p p o rt.
C h a p ter 1
In tr o d u c tio n
1 .1 M o tiv a tio n a n d t h e s is o u tlin e
Im proving our u n d ersta n d in g of the oceanography of sm all, semi-enclosed basins is im p o rtan t for th e countries surrounding th e m , but m ay also be of consid erable scientific in terest, especially if the processes encountered th e re have global relevance. T he Red Sea is a n exam ple of such a w ater body. Because of its small size an d sim ple geom etry, th e Red Sea provides an ideal n a tu ra l laboratory for th e stu d y of problems of general interest. These problem s include; (i) T h e buoy ancy driven circulation of an inverse estuary w ith tu rb u le n t convection, (ii) T he evaluation of m eteorological d atasets and bulk p aram eterizatio n s for th e surface buoyancy fluxes th a t are used on a global scale, exploiting th e sem i-enclosed n atu re of th e R ed Sea. (iii) P ro b lem s related to the hydraulic control ex erted by a stra it an d a sill an d its im p o rtan ce on th e circulation and stratificatio n of a semi-enclosed basin, (iv) The role of h o rizo n tal eddy viscosity above sloping topography on the dynam ics of the flow in a channel.
A p art from these problem s of broader relevance, the R ed Sea is an interest ing basin in its own right because of the unique phen o m en a tak in g place there. T he exceptionally strong ev ap o ratio n ra te of the Red Sea (of alm ost 2 m y ear~ ^ ) is the cause for th e very high surface salinity of more th a n 40 a t th e n o rth e rn end - the highest surface salinity in th e world ocean a p a rt from very sm all lagoons. This surface buoyancy loss com bined w ith the geom etry of th e basin leads to th e form a tio n of deep waters w ith th e highest tem p eratu re a n d salinity in th e world. Salty
I. In tro d u ctio n
ia te rm e d ia te an d deep w ater pro d u ced in the Red Sea flows into th e In d ian ocean and could possibly aiFect its therm o h alin e circulation; Red Sea water can be tra ced as far as th e so u th ern hem isphere. Finally, th e Red Sea is an exam ple of th e first stage of oceanic basin form ation due to th e spreading of tectonic plates.
D espite th e fascinating oceanography of the Red Sea, it is a very poorly surveyed basin. T he m ain reason for th e lack of synoptic observations is th a t th e Red Sea is surrounded by developing or th ird world countries w hich are in hostile relatio n s. These countries face severe problem s from lack of freshw ater and th e e x a m in atio n of the hydrological cycle of th e area, w ith the Red Sea being an integral p a rt of it, could provide valuable inform ation towards solutions to th e problem . For th e w estern world, th e stu d y of th e Red Sea is interesting because of its geographic location; the R ed Sea is in an area of high oil p roduction an d it provides an im p o rta n t route betw een th e M editerranean and tlie Indian O cean.
In this thesis, some in tro d u cto ry inform ation ab o u t the geography, forcing, hydrography a n d circulation p a tte rn s of th e Red Sea will be provided in C h a p te r 1. In C h a p te r 2, a detailed review of the observations of th e heat and freshw ater fluxes th ro u g h th e S trait of B ab el M andab is presented. These fluxes provide a strong c o n stra in t for th e surface buoyancy fluxes estim ated from m eteorological p aram eters a n d bulk form ulae, exam ined in C h ap ter 3. T he dynam ical response of th e R ed Sea to th e buoyancy forcing using a sim ple sim ilarity model is in v estig ated in C h a p te r 4. Questions ab o u t th e dynam ical balance in the Red Sea lead to th e ex am in atio n of horizontal Reynolds stress in C h ap ter 5, where an a tte m p t to d eterm in e horizontal Reynolds stress from A D CP records is presented. G eneral conclusions from this thesis are presented in C h ap ter 6.
1.2 G e o g r a p h y
T h e Red Sea is a deep and narrow semi-enclosed basin th a t sep arates A frica from A sia (Figure 1.1). In th e south, it is connected to the G ulf of A den and the open ocean through th e S tra it of Bab el M andab. T he Suez C anal a t th e north provides a navigational connection of th e Red Sea to th e M ed iterran ean , b u t is not considered an im p o rta n t oceanographic exchange route. T h e R ed Sea has a charmel-like shape w ith a length of 1932 km and an average w idth of 280 km (Morcos 1970). T h e S tra it of Bab el M andab is the narrowest point w ith a w idth of 26 km a t P errim Narrows, but th e shallowest depth (137 m) of th e m o u th occurs ab o u t 150 km northw est of Bab el M andab at the Hanish Sill. A t th e n o rth end, th e R ed Sea bifurcates into two narrow gulfs: the shallow G ulf of Suez (of about 40 m d ep th ) w ithout a sill a t th e m outh, and the deep G ulf of A qaba (w ith a dep th of m ore th a n 1800 m ) which is separated from the Red Sea by a sill of less th an 300 m (Laughton 1970).
T h e average d e p th of th e Red Sea is 491m (Morcos 1970), b u t abyssal depths of m ore th a n 2500 m have been recorded along th e axial trough of th e basin (Tom czak and G odfrey 1994). Topographic sections across the Red Sea (A llan 1970) show very wide shallow shelves (less th a n 50 m deep) in th e so u th ern portion occupying m ore th a n two th ird s of the basin width and there are broad banks of subm erged reefs th a t form a com plicated bath y m etry (Ross 1983). T h e w idth of th e shelves becom es narrow er towards th e central and n o rth ern p arts. Large island system s exist close to th e southern end, off the coasts of E ritre a and Yem en.
T h e Red Sea is considered a m odel for sea-floor spreading as a m echanism for th e form ation of an oceanic basin (Ross 1983). T h e rich g eo th erm al and ge o m agnetic activ ity on th e b o tto m of th e Red Sea has been associated w ith the
I. Introduction S a a S u e z 30°N Gulf Gulf S u e z SAUDI ARABIA EGYPT o> *o 3 J e d d a h SUDAN 20' Port Sudan jJ ERITREA YEMEN 15' Hanlsh SOI strait o f Bab al Mandat G u lfd f I Ada* DJIBOUTI / SOMALIA 45' 40' Longitude
sea-floor spreading, an d it is the first region in th e world where h o t brines were discovered on th e sea floor (M iller 1964).
1 .3 C lim a te c o n d itio n s
T h e clim ate of th e Red Sea a rea is extrem ely arid. E vaporation rates are exceptionally high and of th e order of 2 m eters per year (Yegorov 19-50; N eum ann 1952; P riv e tt 1959). P recip itatio n is very sm all and never exceeds 2 0 cm per year even in th e convergence zone (Pedgley 1974). T here is no river runoff in th e Red Sea.
R egarding th e h e a t flux in th e R ed Sea, it has long been recognized th a t th e re is a sm all annual overall surface heat loss of less th a n 10 W m~^ (Yegorov 1950), an d th a t the oceanic heat tra n sp o rt through th e S tra it provides a strong co n strain t on th e surface fluxes (B unker et al. 1982). A review of th e available m easu rem en ts of th e oceanic heat tra n sp o rt will be given in C h a p te r 2 an d a d e tailed exam ination of each com ponent of th e surface heat and freshw ater budgets will be presented in C h a p te r 3. For exam ple, it has been sp ecu lated th a t the extensive desert areas surrounding th e Red Sea are sources of aerosol particles, such as sand and d u st, th a t could be responsible for th e depletion of incom ing solar ra d iatio n (N eum ann 1952), and. th is will be examined, in detail.
T h e w ind stress clim atology is influenced by the high m ountains a n d plateau x th a t b order th e Red Sea, while th e m onsoons are responsible for th e seasonal changes in th e wind direction. As show n in Figure 1.2, m onthly m ean winds are, on average, directed along th e m ain axis of th e basin; th e d a ta are o b ta in e d from th e revised version of th e C om prehensive O cean-A tm osphere D a ta Set (COADS) by d a Silva et al. (1994) a t th e D ep artm en t of Geoscience of th e U niversity of W isconsin-M ilw aukee (henceforth referred to as U W M /C O A D S). N o rth of ab o u t
1. In tro d u ctio n
UWM/COADS Wind Stress climatology (max arrow = 0.08 Nm ) Ja n u a ry 30 25 20 30 35 40 45 February 30 25 20 30 35 40 45 March 30 35 40 45 30 35 40 45 30 25 20 15 April 30 25 20
May Ju n e July A ugust
30 20 30 35 40 45 30 25 20 30 35 40 45 30 25 20 10 30 35 40 45 35 40 45 30 35 40 45 30 25 \ \ V 20 35 40 45
S e p te m b e r O ctober N ovem ber D ece m b e r
30 25 20 30 35 40 45 30 25 20 10 30 35 40 45 30 35 40 45 30 35 40 45 30 25 20 15 30 25 20
Fig. 1.2. C lim atological m o n th ly m ean w ind stress (45-year averages from 1945 to 1989 from U W M /C O A D S ).
19°N th e winds are from th e north-northw est throughout th e year. S outh of 19°N th e winds change direction u n d er the influence of th e monsoon system . From Oc to b e r th rough Decem ber th e convergence zone of th e winds is a t around 19°N, but after D ecem ber th e convergence zone moves southw ard, so th a t by Ju n e and until S eptem ber th e winds are from th e northw est over th e entire basin. O verall, winds are stro n g er during w inter m onths.
1 .4 H y d r o g r a p h ic s tr u c tu r e
T h e R ed Sea is very poorly surveyed. M ost of the w ater p ro p erty m ea surem ents are tak en along th e m ain axis of th e basin, whereas very few synoptic cruises contain any cross sections. From all th e surveys carried o u t in th e Red Sea du rin g th e last 50 years only 4 cruises contain extensive cross-section m ea surem ents. T hese are th e cruises of the R.V. Atlantis in M ay 1958, th e Cdt. R.
Giraud during Jan uary-F eb ruary 1963, th e R.V. Sonne in Janu ary-M arch 1984,
and m ore recently, the NAVOCEANO survey in Ju n e 1993 an d A ugust 1994. The n o rth -so u th hydrographic stru c tu re is thus relatively b e tte r surveyed, while there is very little confidence in th e cross-axis property distribution.
Clim atological m onthly m ean values for th e sea surface te m p e ra tu re (SST), salinity (SSS), and potential density (erg) in January, April, July, an d O ctober are presented in Figure 1.3 (th e d a ta for tem p eratu re an d salinity have been obtained from th e U W M /C O A D S clim atology). T he SST (Figure 1.3a) increases from the n o rth e rn end tow ards m axim um values in th e southern Red Sea an d decreases again close to th e S tra it of Bab el M andab. T he zone of m axim um te m p e ra tu re , which in late sum m er occurs in th e so u th a t around 14°N (w ith SST higher th a n 31°C), moves north w ard up to 19°N in th e w inter (w ith SST higher th a n 26°C). This phenom enon is a ttrib u te d to th e reversal of the winds th a t pile up w ater towards
I. In tro d u ctio n (a) (b) (c) 30 25 20 IS 10
January April July O ctober
30 30 30 30 25 25 25 25 T V 20 20 20 20
I n
IS 15 15 15 10 10 10 10 22 24 26 28Sea Surface Temperature [° C]
30
Ja n u ary April July
32 O ctober
■
36 January 37 38 39 40 41AprilSea Surface Salinity [psu]July O ctober
23 e e e z x 26 . 27 24 25 __ , Sea Surface a [kg m~^ 28 30 35 40 45 30 35 40 45 30 3 5 40 45 30 35 40 45 30 30 30 30 25 25 25 25 20 20 % 20 20 15 15 15
T
l 15 Qy
10 - 10 10 ...^
10 --- --- h : 30 35 40 45 30 35 40 45 30 35 40 45 30 35 40 45 30'X
l
30 30\
25 25 25%
20TL
20 20 15 10 15 10 qy
15 10 qy
30 35 40 45 30 35 40 45 30 35 40 45 30 35 40 45Fig. 1.3. M onthly m ean clim atology of th e R ed Sea for Jan u ary , A pril, Ju ly and O ctober ( d a ta from U W M /C O A D S ). (a) Sea surface te m p e ra tu re in °C, (b) Sea surface salinity, an d (c) erg in kg m~^.
th e c en tral zone of th e Red Sea d u rin g winter. In su m m er, th e north-northw estern w inds d o m in ate th ro u g h o u t th e basin causing the sou th w ard tra n sp o rt of the tem p e ra tu re m ax im u m (Morcos 1970). T h e seasonal variation of th e m onthly m ean SST varies w ith position an d is higher at the n o rth e rn p a rt (w here th e seasonal change is a b o u t 8°C) an d sm aller in th e area of B ab el M andab (w here the seasonal change is a b o u t 5°C).
T h e d istrib u tio n of SSS (F igure 1.3b) reveals th e m ost strik in g feature of th e R ed Sea: th e highest surface salinity in the w orld oceans. T his occurs as a resu lt of th e strong evaporation rate. Sea-surface salin ity increases from about 36.5 n ear B ab el M andab S tra it to m ore th an 40 a t th e n o rth e rn en d of the basin. Seasonal variations show th a t th e SSS is higher in fall th a n in spring a t all latitudes. A t th e n o rth e rn end, salinity ranges from 40 to 40.5 in w inter, b u t in sum m er it reaches values of m ore th a n 41. T h e SSS near Bab el M andab S tra it increases slightly during su m m er m onths as th e north-northw estern winds carry salty Red Sea surface w ater into th e G ulf of Aden.
Figure 1.3c shows th a t thro u g h o u t the year th ere is a stro n g north-south p o te n tia l d en sity difference of up to 4 k g m “ '^. M axim um surface densities are observed du rin g w inter and th e n o rth end density reaches cTg = 28 kg m ” ^ or more.
T h e vertical d istrib u tio n of th e hydrographic p roperties in th e Red Sea is show n in F igures 1.4 an d 1.5. T hese are hydrographic sections along the m ain axis of th e basin for w in ter an d sum m er from th e cruises of Cdt. R . Giraud in Ja n u a ry -F e b ru a ry 1963 an d th e R.V. Atlantis in A ugust 1963, respectively. It is ev id en t th a t m ost of th e stratifica tio n occurs in th e top 150 m above th e sill depth.
D uring w inter (F igure 1.4), a 50 m thick surface inflow of w arm and low- salin ity w ater {T ~ 25°C, 5 ~ 36.5, ag ~ 25 kg m "^) th ro u g h th e S tra it of Bab el M an d ab is observed. T his surface layer becomes sa ltie r as it advances through the
I. Introduction 10 Potential T em perature [“ C]
I
•a 300 0 .2 0 .4 0.6 0 .8 1 1.2 1 .4 1.6 1.8 Salinity •E-200 •5 300 0 .2 0 .4 0.6 0 .8 1 1 2 1 .4 P otential Density [Og]p-200
•5 300
d ista n c e from S u e z x I O m
F ig. 1.4. H ydrographic sections along th e axis of th e R ed Sea from th e G ulf of Suez to th e S trait of B ab el M andab using d a ta for th e potential te m p e ra tu re (upper p an el), salinity (m iddle panel), a n d ag (lower panel) from th e cruise of Cdt. R.
R ed Sea un til it reaches a m axim um salinity of more th an 40 at th e n o rth e rn part of th e Red Sea. W ater of higher density is thus produced due to evap o ratio n . It sinks a n d flows o ut of th e basin in a lower layer of salinity between 37 an d 40. It is likely th a t th e outflowing interm ed iate Red Sea w ater is produced th ro u g h o u t th e basin on th e shallow continental shelves of th e Red Sea and not only a t th e far n o rth end (M aillard 1974). T he outflowing salty Red Sea interm ediate w ater spreads into th e Indian Ocean, m uch as M editerranean w ater spreads in to th e A tlantic, and significantly affects th e stru c tu re of th e northw est Indian O cean (S hapiro and M eschanov 1991).
D uring sum m er th e wind changes direction and blows out o f th e R ed Sea into th e G ulf of Aden (F igure 1.5), causing an outflow of the top 20 m of warm and salty Red Sea water ( T ~ 30°C, S ~ 37.5, CTg ~ 2 3 k g m “ ^). Below this shallow outflow, a cool low -salinity inflow from th e Gulf of A den ( T ~ 1S°C,
S ~ 36) between 30 and 80 m d ep th is observed (M aillard and S olim an 1986).
T his intrusion propagates into th e R ed Sea at a speed of 0.06 m s"^ (S m eed 1988) and is d etected as far n o rth as 19°N until October when it vanishes. Salty Red Sea w ater continues to outflow into th e G ulf of Aden in the b o tto m layer, but the volum e flux is m uch less (M urray and Johns 1997).
Below th e top 200 m th e Red Sea is weakly stratified and th e w ater prop erties are relatively uniform down to th e bottom , w ith T ~ 21.5°C, S — 40.6, and
Œg ~ 28.6 kg m "^. Siedler (1968), however, found th a t the Red S ea deep w ater
undergoes notable seasonal variations and appears colder by 0.1°C a n d saltier by 0.05psu in sum m er than in w inter. Finally, it is worth m entioning th a t, as a result of th e basin geom etry and th e local buoyancy forcing, the deep w a te r of th e Red Sea has a higher tem p eratu re th a n any o ther p art of th e world ocean a t th is dep th (M orcos 1970).
1. In tro d u ctio n 12 P o ten tial T e m p e ra tu re ( CJ ■5 300 4 0 0 -Salinity •a 300 0 ^ 0 .4 0.6 0.8 1.2 1.4
P o ten tial D ensity [c y
râ 300
0 .8 1 1 .2 1.4 d ista n c e from S u e z x 10® m
F ig. 1.5. H ydrographic sections along th e axis of th e R ed Sea using d a ta for th e p o te n tia l te m p e ra tu re (u p p er panel), salin ity (m iddle panel), an d Cg (lower panel) fro m th e cruise of R .V . Atlantis during Ju ly -A u g u st 1963.
1.5 C i r c u l a t i o n
1 .5 .1 S h a llo w c ir c u la tio n
C u rren t m easurem ents in th e Red Sea are sparse and th e few available are in th e vicinity of the S trait of B ab el M andab (Vercelli 1927; S iedler 1968; M aillard an d Solim an 1986; M urray a n d Johns 1997), or a t th e m outh of G ulf of A qaba (M u rray et al. 1984). Thus o u r present knowledge of th e cu rren ts of th e Red Sea p ro p er is based on geostrophic cu rren t estim ates (M aillard 1974; Q uadfasel and B au d n e r 1993), on d a ta com piled from ship drift observations (P a tz e rt 1974a), on sa te llite a ltim e try d a ta (Eshel et al. 1995), and on m atching O G C M results to hydrographic observations (E shel an d Naik 1997; Clifford et al. 1997).
T h e near-surface cu rren t system above th e sill dep th of a b o u t 150 m, is believed to be the result of th e com bined effects of wind and b u oyancy forcing, while th e effects of the b o tto m topography are rarely m entioned. A large-scale, an n u al-m ean picture of the shallow circulation of th e Red Sea is th a t of an inverse estuaxy w here th e w ater mass deficit due to evaporation causes a n ear-su rface inflow of fresher w ater, from th e G u lf of A den, which undergoes an increase in salinity an d becom es denser as it flows tow ards th e n o rth ern end of the b asin . T h is causes vertical convection into a lower layer which leaves th e basin as a n u n d ercu rren t over th e sill.
T his first-order buoyancy-driven circulation of th e Red S ea was first ex am ined by P hillips (1966) who used th e sim ple geom etry of th e basin to in tro d u ce an elegant sim ilarity m odel based on scaling argum ents. Recently, M ax w o rth y (1997) proposed a modified sim ilarity scaling for the two-layer circulation of a buoyancy- driven basin an d a tte m p te d to include th e effects of a contraction a t th e entrance. Solutions of P h illip s’ m odel for th e Red Sea were calculated by T rag o u an d G a rre tt
I. Introduction 14
(1997) and will be discussed and com pared w ith Majcworthy’s solutions in d etail in C h a p te r 4.
T he wind also influences th e near-surface circulation p artic u la rly in th e so u th ern end of th e basin; according to P a tz e rt (1974b) the wind in th e n o rth e rn p a rt of th e Red Sea has a southw ard direction th a t opposes the buoyancy-driven circulation during th e en tire year, b u t affects th e circulation only du rin g th e su m m e r (Siedler 1969). In th e so u th ern p a rt the wind changes direction seasonally (F ig u re 1.2) enhancing th e buoyancy-driven circulation in the w inter (n o rth w ard w ind) and opposing it in the sum m er (southw ard wind). Thus the tw o-layer sys te m of inflow-outflow at th e S trait of Bab el M andab becomes a weaker th ree-layer sy stem during th e su m m er m onths from Ju n e to S eptem ber (M aillard a n d Soli man. 1986) with a shallow surface outflow above an interm ediate inflow an d deep outflow. T he wind clearly affects th e circulation in the Red Sea but it seem s th a t th e buoyancy-driven circulation dom inates the large-scale flow p a tte rn th ro u g h o u t th e year.
Eshel et al. (1994) also suggested th a t th e therm ohaline driven circu latio n is m ore im portant th a n th e wind driven. T hey estim ated the circu latio n from hydrographic and ^He d a ta w ith an inverse calculation which allows for advective circulation w ithout m ixing. This is essentially a calculation of the m ass an d tra ce r budgets; the relative dynam ical im p o rtan ce of therm ohaline and w ind d riv en cir cu latio n cannot be determ in ed from th e ir model.
Two OGCM studies of th e dynam ics of th e Red Sea (Eshel an d N aik 1997; Clifford et al. 1997) show a com plex circulation p a tte rn , seasonally variable, com posed of a series of eddies, je ts and subgyres. A part from the cyclonic m otion confined to the far n o rth end of th e basin (which is aissumed to be asso ciated w ith
deep-w ater form ation), there are no co n sisten t patterns betw een th e se m odel re sults. Nevertheless, bo th studies a g ree th a t th e eddy circulation of th e R ed Sea is m ainly due to wind stirring and th e th e eddy activity is m ore im p o rta n t when th e winds axe cross-axis. C urrent speeds of th e eddies reach 0.5 m s “ ^ or higher, whereas th ere are no estim ates for th e clim atological-m ean speed. E d d y m otion w ith cu rren t speeds of up to 0 .4 m s ~ ‘ were found from geostrophic cu rre n ts es tim a te d by M aillard (1974) and Q uadfasel and Baudner (1993). T h e la tte r also suggested th a t the eddies are wind d riven, b u t topographically tra p p e d a t the w ider areas of the basin. T he forcing an d th e results of the two O G C M studies will be discussed in C hapter 4.
1 .5 .2 D e e p circu la tio n : f o r m a t io n a n d ren ew al t im e
Since there are no direct deep cu rren t observations our know ledge of the deep w ater circulation is based on tra c e r budgets. The deep w ater te m p e ra tu re an d salinity are nearly hom ogeneous, b u t th e strongly stru c tu re d d istrib u tio n of n u trien ts, dissolved oxygen, and geochem ical tracers are used to e s tim a te th e deep circulation.
It is generally believed (C em b er 1988; Eshel et a i 1994; VVoelk an d Q u ad fasel 1996) th a t deep w ater fo rm atio n occurs a t the far n o rth end of th e basin in two processes: continental shelf fo rm a tio n in th e shallow G ulf of Suez, a n d open- ocean deep convection at th e n o rth e rn en d of th e Red Sea. Some c o n trib u tio n also comes from th e outflow of th e G ulf o f A qaba into the Red Sea, alth o u g h this is considered sm all com pared to the first two sources (Cem ber 1988).
W hile there is general ag reem en t on th e form ation site, th e re is a variety of estim ates for th e annual m ean p ro d u c tio n rates th a t range from 0.04 to 0.16 Sv (W yrtki 1974; M aillard 1974; C em ber 1988; Eshel et al. 1994), an d for th e renew al
I. Introduction 16
tim e from 36 to 200 years. T h ese p ro d u ctio n rates suggest a ra th e r weak deep circu latio n com pared to th e shallow th erm o h alin e cell of 0.3 Sv (Siedler 1969) of th e to p 150 m. Moreover, the deep circu latio n is considered to be ra th e r independent of th e shallow circulation (W y rtk i 1974) an d it is possible th a t th e deep w ater flows o u t th ro u g h the S trait of Bab el M andab independently of th e u p p er therm ohaline circulation.
It is believed th a t all th e a c tiv ity in th e deep basin is lim ited to th e w inter m o n th s (W yrtki 1974; WoeLk a n d Q uadfasel 1996) when deep w ater is produced. D uring sum m er, higher te m p e ra tu re s and lower evaporation rates prevent deep- w ater form ation an d th e deep basin is considered sta g n a n t. W oelk an d Q uadfasel (1996) also em phasize the im p o rta n c e of in teran n u al variability in th e deep-w ater fo rm atio n rates.
C h a p te r 2
T h e S tra it o f B a b e l M a n d a b
2 .1 In tro d u ctio n .
T h e existence of a contraction (in th e form of a sill a n d /o r narrow s) at th e point connecting a semi-enclosed b asin to th e open ocean can be cru cial for th e oceanography of th e basin. F irst, it provides a convenient m easuring point for the surface heat a n d freshw ater budgets of th e basin, an d second, th e d en sity differences betw een th e inflow and outflow, as well as the d e p th of th eir interface allows for conclusions on th e dynam ics in th e interior of th e basin.
In a pioneering study, Bryden an d Stom m el (1984) showed how th e vol um e and p ro p erty exchange rates at th e s tr a it m ay influence th e stratificatio n an d circulation of a buoyancy-driven sem i-enclosed inverse estu a ry such as the M ed iter ranean. G a rre tt et al. (1990) em phasized th e im portance of resolving th e qu estio n w hether the exchange a t th e stra it is m ax im al or subm axim al in order to d e te rm in e th e overm ixed or not-overm ixed state of th e interior, and to assess th e role of th e s tra it as an a p p ro p ria te m onitoring point for changes in th e interior.
Recent studies of th e S trait of B ab el M andab suggest th a t th e exchange th e re is subm axim al (M axw orthy 1997). T h e outflow occurs in a th in b u t hy- draulically controlled layer as there is stro n g u p stream /d o w n stream a sy m m e try of th e flow. P r a tt et al. (1998), however, based on cu rren t m easurem ents a t th e sill and narrows, found th a t hydraulic control m ay not exist throughout th e year, especially at th e narrow s. A t any rate, th e strong seasonal signal in th e outflow
2. T he S tra it o f B ab el M andab IS
rates (M urray an d Johns 1997) strongly implies th a t the exchange a t th e S trait is subm axim al a n d th e Red Sea is not over mixed.
Sub m ax im al exchange at th e S trait suggests th a t the exchange is not con trolled by th e S tra it alone, but processes in th e interior of the Red Sea m ay be im p o rtan t. T h e subm axim al n atu re of th e exchange also indicates th a t there is a much m ore rap id response of the Strait to conditions in the Red Sea, so th a t it is a convenient m onitoring point for changes in th e interior.
In fact, even sta n d a rd records of w ater properties and currents a t th e S trait of Bab el M andab m ay provide valuable inform ation about th e interior of the bcisin. A useful s ta rtin g point involves estim ates of heat and freshw ater budgets which can be used as constraints for the surface buoyancy fluxes of th e basin. In the following section we will exam ine the cap ab ility of m easurem ents a t the S trait of Bab el M andab to provide constraints for th e heat and freshw ater fluxes and present u p -to -d ate estim ates of these constraints.
2.2 O c e a n ic h e a t a n d fresh w a ter tr a n s p o r ts th r o u g h th e S tr a it o f B ab e l M a n d a b
2 .2 .1 R e v ie w o f a n n u a l e s tim a te s
H eat an d freshw ater transports through a section at the entran ce of the Red Sea m ay be calculated from direct in situ m easurem ents of te m p e ra tu re and salinity w ith sim ultaneous observations of th e currents, but the results m ust be com patible w ith conservation of volume and salt (th e K nudsen form ulae). For the Red Sea these are
- P , + F , - F , = B „ ,A (2.1)
w here (i = 1 ,2 ,3 ) are th e annual m ean volum e fluxes (in Sv = lO^m^s"^) of
th e surface outflow (present during sum m er m o n th s only), th e in term ed iate (su m m er only) or surface inflow and th e b o tto m outflow F^. A is th e surface area of th e Red Sea (0.45 x lO^^m^), and = E — P , where E is the evaporation
and P th e p recipitation (b o th in m s “ ^); th ere is no runoff in th e Red Sea. p,- and S',- are th e average densities an d salinities of th e corresponding layers.
M onthly volum e fluxes for th e near surface layer (i.e. F^ or F^) have been e stim a te d by P a tz e rt (1974a) using ship drift observations [from th e K oninklijk N ederlands M eteorologisch In stitu u t (KNM I) A tlas (1949)], an d su m m ertim e vol um e fluxes for the in te rm e d ia te layer [F^] are available from m easurem ents by M aillard and Solim an (1986). M onthly m ean te m p e ra tu re and salinity profiles are tak en from Levitus et al. (1994) at the grid point closest to th e S trait (15.5°N, 41.5°E), b u t 250 km to th e northw est of Bab el M andab where th e d ep th is ab o u t 400 m . Since the sill a t H anish Island is about 150 km northw est of Bab el M andab narrow s, th e hydrographic d a ta are about 100 km aw ay from th e sill. T he sensitiv ity of o u r results to th is effect was checked by allowing for a tim e lag of one m onth for th e inflowing w ater to reach th a t point, which is th e speed of propagation of the inflowing w ater according to Sm eed (1988). No significant differences were found.
T h e evaporation rates for th e Red Sea have been e stim a ted by Yegorov (1950) (2 .3 0 m y “ ^), N eum ann (1952) (2.15 m y~^), a n d P riv e tt (1959) (1.83 m y~^) using a bulk form ula for th e w ater vapor transfer, as described by Morcos (1970) in his thorough review of the R ed Sea. T here are considerable differences am ong these estim ates not only for th e annual ra te b u t also in th e seasonality; P riv e tt rep o rts a higher evaporation ra te during w inter, w hile Yegorov and N eum ann found m ax im u m evaporation rates during sum m er.
2. T h e S tra it o f B ab el M andab 20
To provide an e stim a te for th e evaporation ra te th a t would satisfy the Knud sen form ulae, we consider b o th and as unknow ns and solve th e system of equations (2.1) an d (2.2). T his gives th e annual m ean = 1.60 ± 0.35 m y “^, sm aller th a n previously claim ed. T he overall u n c e rtain ty is e stim a te d from the ± 2 0% u n c e rta in ty in th e volum e flux and ±1 0% u n certain ty in th e salinity dif ferences betw een inflow and outflow. T he la tte r is e stim ate d from th e average d ev iatio n of th e salinity difference from the sm oothed annual cycle of th e salin ity difference w ith a 5 th degree polynom ial fit. A sim ilar e stim a te of usiag th e sam e clim atological salinities a n d up d ated volum e fluxes for th e lower layer from M u rray and Johns (1997), and considering F^ as unknow n, also gave
^net = 1.6 0m y - \
T h e e s tim a te d m o nthly E^^^ (assum ed co n stan t) balances th e m onthly vol u m e bu d g et (2.1) to provide th e unknown volum e tra n sp o rt F^ (seasonal changes in th e ev ap o ratio n ra te and volum e changes due to sea level changes are neglected), so th a t we m ay e stim a te the an n u al m ean h eat tra n sp o rt Fj.. T his is expressed as th e equivalent flux across th e surface of th e R ed Sea and is e stim a ted from the form ula
+ ^2 ^ 2 - % - (2.3)
w here p is th e m ean w ater density (1025 kg m "^), is th e h eat capacity of water (3986 J K ~ ^ kg "^), an d T- are th e m ean tem p eratu res of th e inflowing an d outflow ing layers. T h e last te rm in eq u atio n (2.3) represents th e h eat tra n sp o rt due to th e volum e th a t leaves th e basin through its surface and it should be included w hen ^ 2 7^ 0: otherw ise th e n et heat tra n sp o rt a t th e S tra it depends on the choice of te m p e ra tu re scale (i.e. Celsius or K elvin). T h e te m p e ra tu re of th e evap o ra te d volum e of w ater is th e sea surface te m p e ra tu re , therefore we have used the m o n th ly m e a n te m p e ra tu re clim atology from U W M /C O A D S for com bined with
th e an n u al evaporation ra te E^^^A of 0.023 Sv to estim ate th e net heat tra n sp o rt
Fj- as approxim ately 8 ±2VV m~^ (u n c e rtain ty in th e tem p eratu re differences be
tw een inflow an d outflow is 10%). T his com pares well w ith the original e stim a te by P a tz e rt (1974) of 7 W m~^, who used historical m onthly tem p eratu re profiles from th e N ational Oceanographic D ata C en ter (N ODC) combined w ith th e afore m en tio n ed m onthly volume tran sp o rts from KNM I. O ur m onthly e stim a te s for th e h eat tra n sp o rt are presented in F igure 2.1a. We note th a t, in principle, th e last te rm in equation (2.3) should be —E A T ^ -f P A T ^ where P is th e ra te and Tp th e te m p e ra tu re of the precipitation, b u t th e effect of this is sm all (less th a n 0.5 W m~^). T h e m onthly heat tra n sp o rt e stim ate d from equation (2.3) is ra th e r insensitive to uncertainty and seasonal changes in E^^^ so th a t use of th e an n u al average E^^^ in (2.3) is adequate.
O u r estim ates for th e m onthly sa lt tra n sp o rts are shown in F igure 2.1b. W hile we have assumed a m onthly co n stan t net volum e flux, the salt tra n sp o rt from Fg = — + P2S2F2 — PzSsF^ allows for a m o n th ly salt transport. T h e an n u al
m ean is zero. We note th a t in our e stim a tes for the heat and salt tra n sp o rts we have not tak en into account the correlation term s betw een the volume flux an d th e te m p e ra tu re and salinity fluctuations in tim e scales sh o rter than one m o n th as these are not possible to estim ate w ith th e pressent d a ta sets.
A second estim ate of th e annual h eat exchange through th e S tra it was pro\dded by A hm ad and S ultan (1989), who used th e sam e hydrological d a ta s e t as P a tz e rt, b u t unfortunately, th e source of th e ir volum e flux d a ta was not clearly s ta te d in th e ir paper which makes com parison w ith P a tz e rt’s results diflScult. T h e m ost strik in g diff^erence betw een th e two analyses is th a t, according to A h m ad an d S u ltan , th e R ed Sea gains heat through th e S tra it during the sum m er in ste ad of losing it as suggested by P a tz e rt’s original resu lts (—17 W m~^). Ahm ad an d S u lta n
2. T h e S tra it o f Bab el Mandab 2 2
Heat Transport through Bab el Mandab
(a) 4 0 3 0 5 10 ■c - 1 0 -2 0 - 3 0 - 4 0 10 12
Salt Transport through Bab el Mandab (b) 1 .5 oi % 0 .5 - 0 . 5 12 10 Month
Fig. 2.1. E stim ates of heat and salt ad v ected through the S trait of B ab el M andab. Crosses m ark results using P a tz e rt's volum e tra n sp o rt, circles m ark Vercelli's re sults, diam onds m ark M aillard an d S o lim an ’s, and squares m ark Siedler’s. T h e an n u al m ean h eat tra n sp o rt is equivalent to a surface heating of th e R ed Sea of 8 W m~^. T he annual m ean salt tra n s p o rt is zero.
