I~ 5- 9ij;~i~ 14: -
3
lN
•D~ix\TE B;OIIOOL OF OCEANOGRAPHY 7TUNIVERSIrPY OF RHODE ISLANDX-,I+UGANSFTTU IOI-TSLAND
in tl-e TVertical Motionin the SYNOP Central Array
z ~~~ - - 7i•m•06 a -~ eprtr
;J' 4304
15J$n 733am 103Gam 23an
Thl -s tfwc:1 ane tee Qppovo (3V)•I k-°° xb-- •t-e+'s- w- s +,-e; it - -Q--_"_Scott " .- - -- -Ldr a D Rnl Watts
-I Direction -. 000 ++,+, ,
I~ I
Sc93 J i2 Ja 23atIsla M NOO49J-•an d I nOOO3.9Jazlg
IAbstract
As part of the SYNOP (ynoptic Qcean Prediction experiment) field program, twelvetall moorings measured the Gulf Stream's temperature and velocity fields with currentmeter (CM) at nominal depths of 400 m, 700 m, 1000 m, and 3500 m for two years,from May 1988 through August 1990. Simultaneously, 24 inverted echo sounders (IES)
monitored the thermocline topography. A third observational component of the ex-periment was the release of isopycnal RAFOS floats; 70 such floats traversed the areamonitored by the CM and the IES. This report documents the methods used to com-pute vertical motion for each data source, and the differences and similarities betweenthe three methods. Typical velocities during 'strong' events, as observed by or inferredfrom all three instruments, was 1 - 2 mm s-' in regions near the center of the GulfStream. The comparison of RAFOS vertical motions and vertical motions disanosedfrom CM data showed excellent agreement; furthermore, CM vertical motions and UESvertical motions are statistically coherent for periods longer than 12 days. We concludethat we may map mesoscale fields of w(z, y, t); the fields mapped ae consistent with3 quasi-geostrophic dynamics.
II
Accesion For
NTIS CRA&IDTIC TAB 0Uiao, ouiicedJustification
By ........................
Distribution IAvailability Codes
Avail aidl/orI Dut Special
,. iA-(I Q , ,,
il'x ux~ 5 E~
I
III£III
Ii
.1aIII
IIII
Contents
Abstract ii
Table of Contents inl
List of Figures v
5 List of Tables vin
1 Introduction 1
2 Instruments and vertical motion calculations 12.1 RAFOS floats ................................... I
2.1.1 Methodology ............................... I2.1.2 Errors in Float vertical motion estimates .................. 1
2.2 Current meters ......................................... 22.2.1 Methods .......................................... 22.2.2 Application ............ ............................ 3
2.3 IESs ......... ........................................ 4
3 Case Study Comparisons 53.1 Case 1: Float 123, CM H4, 4 July 1988, 0000 UTC ........... 6
3.1.1 Overview ................................. 63.1.2 Verticalmotion results .......................... 6
3.2 Case 2: Float 129, CM 14, 23 September 1988, 0800 - 1600 UTC. 103.2.1 Overview ................................. 103.2.2 Vertical motion results .......................... 10
3.3 Case 3: Float 136, CM 14, 8 December 1988, 1600 UTC ....... 143.3.1 Overview . ................................. 143.3.2 Vertical motion results .......................... 14
3.4 Case 4: Float 141, CM 14, 2 November 1988, 1600 UTC ....... 173.4.1 Overview ................................. 173.4.2 Vertical motion results .......................... 17
3.5 Case 5: Float 175, CM 14, 20 January 1989, 1600 UTC ........ 203.5.1 Overview ................................. 20.3.5.2 Vertical motion results .......................... 20
3.6 Case 6: Float 176, CM 15, 7 February 1989, 0000 UTC - 8 February1989, 0000 UTC ................................ 233.6.1 Overview ................................. 233.6.2 Vertical motion results .......................... 23
3.7 Case 7: Float 194, CM 15, 15 November 1989, 0000 UTC ..... 293.7.1 Overview ................................. 293.7.2 Vertical motion results .......................... 29
3.8 Case 8: Float 194, CM 14, 16 November 1989, 0000 UTC ..... 323.8.1 Overview ................................. 323.8.2 Vertical motion results .......................... 32
3.9 Case 9: Float 199, CM 15, 23 November 1989, 0000 UTC ..... 353.9.1 Overview ................................. 35
I Hfi
U3.9.2 Vertical motion results .......................... 35 I
3.10 Came 10: Float 199, CM I1, 28 November 1989, 1600 UTC ..... 383.10.1 Overview ................................. 383.10.2 Vertical motion results .......................... 38
3.11 Case 11: Float 201, CM I1, 23 November 1989, 0800 UTC ..... 413.11.1 Overview ................................. 413.11.2 Vertical motion results .......................... 41
3.12 Case 12: Float 207, CM 12, 4 September 1989, 0800 UTC ..... 443.12.1 Overview ................................. 443.12.2 Vertical motion results .......................... 44
3.13 Case 13: Float 209, CM 15, 13 September 1989, 0800 UTC .... 473.13.1 Overview . ................................. 473.13.2 Vertical motion results .......................... 47
3.14 Case 14: Float 210., CM A4, 6 October 1989, 0800 UTC ....... 503.14.1 Overview ................................. 503.14.2 Vertical motion results .......................... 50
3.15 Case IS: Float 210, CM 15, 21 October 1989, 0000 UTC ....... 533.15.1 Overview ................................. 533.15.2 Vertical ,,lotior. results .......................... 53
3.16 Case 16: Float 211, CM MI1, 8 October 1980, 0000 UTC ..... 563.16.1 Overview ................................. 563.16.2 Vertical motion results ........................... 56
3.17 Case 1T: Float 216, CM I1, 26 November 1989, 0000 UTC ...... 593.17.1 Overview . ................................. 593.17.2 Vertical motion results .......................... 59
3.18 Case 18: Float 221, CM 12, 6 January 1990, 0800 UTC ........ 623.18.1 Overview ................................. 623.18.2 Vertical motion results .... ...................... 62
3.19 Case 19: Float 224, CM H4, 20 January 1990, 0800 UTC ...... 653.19.1 Overview ................................. 653.19.2 Vertical motion results .......................... 65
4 Comparing the three different methods 684.1 Statistics ..................................... 684.2 Case Studies . ................................... 684.3 Coherence between tuw, and wt,,, ...... ...................... 77 1
5 Summary 82
6 References 83 1
IUivi
List of Figures
1 Case Study 1: Track of RAFOS float 123 ........................ 72 Case Study 1: IES z12 data and diagnosed vertical motion 0000 UTC 4 July
1988 .... .................................... 83 Case Study 1: CM H4 data and diagnosed vertical motion ............ 94 Case Study 2: Track of RAFOS float 129 ........................ 105 Case Study 2: IES z12 data and diagnosed vertical motion 0800 UTC 23 Sep
1988 ........ ........................................ 116 Case Study 2: IES Z12 data and diagnosed vertical motion 1600 UTC 23 Sep
1988 .................................................. 127 Case Study 2: CM 14 data and diagnosed vertical motion .......... ... 138 Case Study 3: Track of RAFOS float 136 ........................ 149 Case Study 3: IES Z12 data and diagnosed vertical motion 1600 UTC 8
December 1988 ...................................... 1510 Case Study 3: CM 14 data and diagnosed vertical motion ............. 1611 Case Study 4: Track of RAFOS float 141 ........................ 1712 Case Study 4: IES Z12 data and diagnosed vertical motions 1600 UTC 2
November 1988 ......................................... 1813 Case Study 4: CM 14 data and diagnosed vertical motion .......... ... 1914 Case Study 5: Track of RAFOS float 175 ........................ 2015 Case Study 5: IES z12 data and diagnosed vertical motion 1600 UTC 21
January 1989 .......................................... 2116 Case Study 5: CM 14 data and diagnosed vertical motion .......... ... 2217 Case Study 6: Track of RAFOS float 176 ........................ 2318 Case Study 6: IES z12 data and diagnosed vertical motion 0800 UTC 7
February 1989 .......................................... 2419 Case Study 6: IES Z12 data and diagnosed vertical motion 1600 UTC 7
February 1989 .......................................... 2520 Case Study 6: IES Z12 data and diagnosed vertical motion 0000 UTC 8
February 1989 ......................................... 2621 Case Study 6: IES z12 data and diagnosed vertical motion 0800 UTC 8
February 1989 ...................................... 2722 Case Study 6: CM 15 data and diagnosed vertical motion ........... .* 2823 Case Study 7: Track of RAFOS float 194 ........................ 2924 Case Study 7: IES 212 data and diagnosed vertical motion 0000 UTC 15
November 1989 ......................................... 3025 Case Study 7: CM 15 data and diagnosed vertical motion .......... ... 3126 Case Study 8: Track of RAFOS float 194 ........................ 3227 Case Study 8: IES Z12 data and diagnosed vertical motion 0000 UTC 16
November 1989 ......................................... 3328 Case Study 8: CM I5 data and diagnosed vertical motion .......... ... 3429 Case Study 9: Track of RAFOS float 199 ........................ 3530 Case Study 9: IES Z12 data and diagnosed vertical motion 0000 UTC 23
November 1989 ......................................... 3631 Case Study 9: CM I5 data and diagnosed vertical motion .......... ... 3732 Case Study 10: Track of RAFOS float 199 ........................ 38
v
33 Case Study 10: IES Z12 data and diagnosed vertical motion 1600 UTC 28November 1989 ......................................... 39
34 Case Study 10: CM 15 data and diagnosed vertical motion .......... .. 40 335 Case Study 11: Track of RAFOS float 201 ........................ 4136 Case Study 11: IES z12 data and diagnosed vertical motion 0800 UTC 23
November 1989 ......................................... 42 537 Case Study 11: CM I1 data and diagnosed vertical motion .......... .. 4338 Case Study 12: Track of RAFOS float 207 ........................ 4439 Case Study 12: IES Z12 data and diagnosed vertical motion 0800 UTC 4
September 1989 ......................................... 4540 Case Study 12: CM 12 data and diagnosed vertical motion .......... .. 4641 Case Study 13: Track of RAFOS float 209 ........................ 47 I42 Case Study 13: IES z12 data and diagnosed vertical motion 0800 UTC 13
September 1989 ......................................... 4843 Case Study 13: CM 15 data and diagnosed vertical motion .......... .. 49 s44 Case Study 14: Track of RAFOS float 210 ........................ 5045 Case Study 14: IES z12 data and diagnosed vertical motion 0800 UTC 6
October 1989 ....................................... 51 j46 Case Study 14: CM H4 data and diagnosed vertical motion ......... ... 5247 Case Study 15: Track of RAFOS float 210 ........................ 5348 Case Study 15: IES z 12 data and diagnosed vertical motion 0000 UTC 21 "5
October 1989 .......................................... 5449 Case Study 15: CM 15 data and diagnosed vertical motion .......... .. 5550 Case Study 16: Track of RAFOS float 211 ........................ 56 I51 Case Study 16: IES Z12 data and diagnosed vertical motion 0000 UTC 8
October 1989 ....................................... 5752 Case Study 16: CM M13 data and diagnosed vertical motion ........ ... 58 i53 Case Study 17: Track of RAFOS float 216 ........................ 5954 Case Study 17: IES Z12 data and diagnosed vertical motion 0000 UTC 26
November 1989 ......................................... 60 I55 Case Study 17: CM I1 data and diagnosed vertical motion .......... .. 6156 Case Study 18: Track of RAFOS float 221 ........................ 6257 Case Study 18: IES Z12 data and diagnosed vertical motion 0800 UTC 6
January 1990 .......................................... 6358 Case Study 18: CM 11 data and diagnosed vertical motion .......... .. 6459 Case Study 19: Track of RAFOS float 224 ........................ 65 I60 Case Study 18: IES ZI2 data and diagnosed vertical motion 0800 UTC 20
January 1990 .......................................... 6681 Case Study 19: CM H4 data and diagnosed vertical motion ......... ... 67 I62 Mean Z12 and wuss for June 1988 - August 1990 ................... 6963 Time series of ws.. and w., at G2 and I1 ....................... 7064 Time series of w,,, and wc" at G2 ...... ...................... 70 I65 Time series of ws,, and wCU at G3 ...... ...................... 7166 Time series of ws,, and 1 1 cm at H2 ...... ...................... 7167 Time series of wu,, and wCU at H3 ........................... 71 .T68 Time series of will and wCM at H4 ...... ...................... 7269 Time series of wis, and twCM at H5 ...... ...................... 7270 Time series of ws,, and wCU at H6 ......................... 72.
Ai
71 Time series of wis$ and wCM at 11 ............................. 73
72 Time series of wj, and wcu at 12 ............................ 7373 Time series of w,,, and wcpC at 13 ............................ 7374 Time series of w... and wcw at 14 ............................ 7475 Time series of w15 S and wJC at 15 ............................ 7476 Time series of wis, and w., at M13 ........................... 7477 Scatterplots of wAA, Vs. w,,, and wCM ......................... 7678 Coherence as a function of frequency for selected sites on the 'R' line . . .. 7779 Coherence as a function of frequency for CM G2 ................... 7880 Coherence as a function of frequency for CM G3 ................... 7881 Coherence as a function of frequency for CM H2 ................... 7882 Coherence as a function of frequency for CM H3 ................... 7983 Coherence as a function of frequency for CM H4 ................... 7984 Coherence as a function of frequency for CM H5 ................... 7985 Coherence as a function of frequency for CM H6 ................... 7986 Coherence as a function of frequency for CM 11 .................... 8087 Coherence as a function of frequency for CM 12 .................... 8088 Coherence as a function of frequency for CM 13 .................... 8089 Coherence as a function of frequency for CM 14 .................... 8090 Coherence as a function of frequency for CM 15 .................... 8191 Coherence as a function of frequency for CM M13 .................. 81
v
II
I vii
U
List of Tables iVertical Motion Statistics: Mean, maximum, and minimum vertical motion(in mm s-) with standard deviation and number of (not all independent)
observations for CM, IES, RAFOS floats in the Central Array only, and forall RAFOS floats ........................................ 68
2 Vertical Motion Comparisons: Vertical motions from RAFOS float, CM, andIES data (WRAP, t wce, and w,,,, respectively), with units of mm s-1 fordate/time shown and for RAFOS float and CM indicated. Depth of theRAFOS float at the comparison time (in meters) is also shown ........ . 75
II
IBIiUI
IIU
viii I
1 Introduction
The S=Xoptic Ocean Erediction experiment (SYNOP) was a multi-institutional studyof the Gulf Stream from 1987-1990. Designed to further the understanding of Gulf Streammeanders, and to facilitate the modeling of the meanders, the SYNOP experiment wascomprised of three different arrays: an Inlet Array near Cape Hatteras, a Central Arraynear 68°W, and an Eastern Array near the Grand Banks at 55°W. Data collected in theCentral Array to be considered in this report came from 12 tall, high-performance mooringson which were four current meters (CM) (at nominal depths of 400, 700, 1000, and 3500 m),from 24 inverted echo sounders (IES) which acoustically monitored the depth of the mainthermocline, and from approximately 70 isopycnal RAFOS floats which sampled pressureand temperature along an isopycnal surface as they moved downstream through the GulfStream.
Vertical motion, w, in the Gulf Stream has been observed in previous studies (e.g.Bower and Rossby 1989) and inferred in others (Hall 1986). Here, we wish to compare thevertical motions observed by the isopycnal floats to those inferred from CM observationsand those inferred from IES observations. The three methods of computin/determiningvertical motion are detailed in the following section, with 23 case study events following.
2 Instruments and vertical motion calculations
2.1 RAFOS floats
2.1.1 Methodology
Bower and Rossby (1989) show how vertical motion is related to RAFOS float motion,namely that upward vertical motion occurs as the float moves downstream from trough tocrest, and downward motion accompanies motion downstream from crest to trough. TheRAFOS float recorded temperature, pressure and location at eight-hour intervals as it movedthrough the stream. The method for determining vertical motion from these variables isstraightforward. The computed wjj, (a D) is estimated using a second-order centeredfinite difference. Because 6t = 8 hours, the vertical motion computed is an average overa 16-hour interval. Of the 72 RAFOS floats that were released as part of the SYNOPexperiment and returned useful data, more than 90% passed directly through the CentralArray, providiag a large dataset of observed vertical motion.
2.1.2 Errors in Float vertical motion estimates
Errors introduced by the float not being perfectly isopycnal should be insignificant for thescales of motion we consider (Rossby, personal communication). However, because floatsmove through features, the vertical motion measured by the float, a sixteen-hour averagevalue, may differ considerably from the point values measured by CM and IES. This ismost obviously the case if the float passes through an up/down couplet in sixteen hours, inwhich case the average vertical motion may be near zero, although the values of upwealgor downwelling within the couplet (as sampled by a current meter mooring, perhaie) willnot be. This bias error may underestimate several of the largest w events by as much as30%.
1
U
To understand this bias, consider a pattern of vertical motion that along a fluid particle Utrajectory varies sinusoidally such that the pressure measured by a RAFOS float and thevertical motion will be
P = A sin wt and wt,,,. = A w coswt, (1)
respectively. How is the computed vertical motion affected by discrete, rather than contin- Iuous, sampling? The estimated vertical motion, 6, is
t= Asin [w(t + At)] - sin [w(t - At)] =in wAt (2)2 At = W• t
The ratio .'i approaches unity for all motions that have period T much longer than the#rs.Isampling time, At, but it crosses zero for T = 2At.
The IRAFOS floats that generated the data used in this study sampled data at At = 8hours. For T ? 3 days, for example, > > 0.92; however, for a period T of 36 hours, theestimated vertical motion is 70% of true. Vertical motions in the inertial frequency rangw I(or higher frequencies) are significantly underestimated by the 8-h sampling period. Fora period of 19 hours, for example, the estimated vertical motion is only 18% of the true.However, this is desirable, because we are using the floats to estimate vertical motions on Imeso- and synoptic scales.
The largest vertical motions that we investigate in this paper are typically associatedwith float displacements with period three days or less. Several of the float trajectoriesthrough the central array show strong up and down couplets, with the float moving up anddown > 200 m in less than three days. Equation (2) indicates that the vertical motion asmeasured by the float may underestimate the true vertical motion in these cases by around30% if the dominant period of the float is about 36 hours.
2.2 Current meters !
2.2.1 Methods
Bryden (1976, 1980) was among the first' to use the backing (turning counterclockwise with4ecreasing depth) and veering (turning clockwise with decreasing depth) of current metersto infer vertical motion in the ocean. Following his notation, and that of'Hall (19$6), thetemperature equation,
87 T 87 T 87 dO0B--T + UE + VT + -L0=0, (3)Tt O: Oy dz
can be transformed, using the thermal wind equation, to
we -= , (4)
or, using 1-tan-' to I
WCM = tr(5)]P
'First in oceanography. Arnason (1942) and Panofiky (1944) developed similar methods for theatmosphere. I
2g
Here, R is magnitude of the current velocity (u2 + v2)J, Po is mean density (1035 kg M- 3 ),g is gravitational acceleration (9.8 m s-2), f is the Coriolis parameter, a. is the effective3 thermal expansion coefficient (- f + - 1.2 x 10-1 kg m- 3 K- 1 ) appropriate forthe T-S relationship in mid-thermocline in the Gulf Stream (Hall 1985), * is the meanstratification (0.02 K m-i), 0 is the angle of the current with respect to east, and u and vare the eastward and northward components of the flow. The constants used in this workare consistent with a Brunt-Vaisala period of about twenty minutes. We note that wCMcould be biased slightly by this choice of constants.
For isopycnal motions, vertical motion is proportional to the two terms in the numeratorof (4) or (5). The first describes the motion of the isopycnals themselves: for instance, inthe absence of horizontal motions, if the temperature at a current meter is increasing,isopycnal surfaces, and the water parcels on them, must be descending. The second term inthe numerator is the heat advection in the presence of sloping isopycnal surfaces, which isproportional to the cross-frontal component of velocity in the presence of sloping isopycnalsurface, as in (3), or equivalently to the turning of the current vector with height, V, as in(5). Where currents back or veer with height, vertical motion must exist, if the temperatureis constant. Consider if the flow at 400m is towards the southeast at 80 cm s-'and the flowat 1000m is towards the northeast, also at 80 cm s-1. For this case the thermal wind willpoint due south. Cold water is to the east, and warm water is to the west, of this vecta,i.e. isotherms slope up to the east. The mean eastward flow encountering this upward slopeto the isotherms therefore generates upward vertical motion.
2.2.2 Application
The CM data used in this study have been smoothed. A 40-hour low-pus filter has beenapplied to the data to remove high-frequency signals (diurnal and semi-diurnal tides, forexample). In addition, the CM data have been adjusted to compensate for mooring motion.When strong currents surround a current meter mooring, it will tilt over more than whenin an environment of weaker currents. Sensors therefore do not always remain at the samelevel. The method of Hogg (1991) interpolates (or extrapolates) the current meter dataat variable levels to constant horizons, in this case at 400, 700, and 1000 m below the seasurface. We computed the vertical motion using data with and without the mooring motioncorrection and noted negligible differences. Because the corrected data also yield verticalmotions at the same level throughout the domain we used in this study data corrected formooring motion.
The vertical motion wCM was computed using (4), with the vertical derivatives (J.1 andj) approximated as centered finite differences using data at 400 and 1000 m. Currentvelocities (u and v) were averages of values at 400 and 1000 m. Temperature tendencies,
(),were computed using a second-order centered-in-time finite different scheme, withAt = 12 hours, and data at 700 m. The vertical motion computed can be considered as arepresentative value about halfway between the 400 and 1000 m. We have used the formulaincluding vy - u' for two reasons. The data includes u and v at three levels, so less data
processing is necessary than if we used the formula including R 2 # (although interpretationof the results is easier using the turning current vector with height). Furthermore, the R 2
term introduced spikes into the computed vertical motions when current speed profile wasnon-linear.
Because the current speeds used in (4) are at two levels, 400 and 1000m, we originallycomputed the temperature tendency term, S, using the average of the values at 1000 and
*3
400m to estimate the water column average. This introduced an error, however, for thosecases in which the 400-m CM was above the main thermocline and the 1000-m CM wasbelow. Both current meters then recorded small temperature tendencies, even though themovement of the main thermocline was large (as measured by the temperature change at700 m). For an example, see Case 1, section 3.1. Although using the averaged temperaturetendency caused an error as just described, no similar error is evident from using the averageof u and v values at 400 and 1000 m to represent the flow at 700 m.
2.3 IESs 3Inverted echo sounders sit on the ocean floor and acoustically monitor the depth of the mainthermocline, denoted as the depth of the 12? isotherm, i.e. z 12 (Watts and Roisby 1977).Vertical motion is computed from IESs by using objectively mapped IES data (Tracey andWatts 1991) to determine a streamfunction and therefore a vorticity field as in Kim (1991).As with CM data, all IES data are 40-hour low-pass filtered to remove high-frequencysignals. This smoothing is done before the objective mapping. The IES 1elds ar objectivelymapped daily for 26 months, yielding estimates not only of streamfunction and vorticity,but of the time tendencies of both. if the motion is assumed to be quasi-geostrophic, themthe vorticity equation as in, for example, Holton (1979),
&4 Vu(C + f) + f.~ 3Z(can be rearranged to determine vertical stretching. Vertical motion in the Gulf Stremhas primarily a first baroclinic mode structure (Hall 1986; Rossby 1987). The maximumvertical motion occurs below the main thermocline (Gill 1982, fig. 6.14); even for veryshallow main thermoclines (i.e., z 1 2 around 200 m - much shallower than for any of the23 cases considered here), the methods used by Pickart and Watts (1990) show that themaximum in vertical motion is below 500 m. Vertical motion in the the upper ocean atdepth z can be approximated as
S= ZoO + v, + n)] (7)
using data at 400 m.We have used this method of computing vertical motions instead of the quasi-geostrophic U
w-equation as used by Leach (1987), Tintor6 et al. (1991) and Pollard and Rqgier (1902)because it is more suited to our data. In the other studies, roughly synoptic CTD data wereavailable to provide data at different levels as is necessary for Q-vector computations. Wehave a streamfunction map at just one level, but at a sequence of times, making (6) a morenatural way to compute vertical motion. Our assumption of primarily first-baroclink modestructure is consistent with the definition of the geostrophic streamfunction (Kim 1991),namely
=*7, x Vz12, (8)
fowhere we have used for 9*(E_ g4 g) a value that is a weak function of :12. The gosatrophic 3baroclinic current from which we compute S is valid at 400m, and is referenced to 3500m,where no motion is assumed. 3
41
To generate the field of vertical motion, the time tendency •'It=t. was estimated using
a'centered finite difference ((C@Q±A&t1 Q t), with At = 24 hours. Vorticity advectionwas computed at time t. with horizontal derivatives computed using second-order centeredfinite-differencing. We have smoothed the IES streamfunction fields with three applicationsof a Shapiro second-order filter (Shapiro 1970) to reduce the amount of noise in the highlydifferentiated vertical motion field. Such smoothing reduces values of extrema in the resul-tant vertical motion field without changing their location. The filter wavelength cutoff is-6 dB at 133 kin; features with lengthscales = k-1 > 37 km are passed with greater than80% amplitude. The gridded values of w,_, were interpolated to each CM location so thatdirect comparisons could be made between wcM and w,,,.
The vertical motion as diagnosed by (7) is forced by two effects, the local change and
the advective change (-L and ;7, V(C + f), respectively). Advection of vorticity causesvertical motion because, as explained in Holton (1979), the vertical motion is required tokeep the ocean hydrostatic and geostrophic. Specifically, vertical motion acts to depressthe thermocline (for the case of anticyclonic vorticity advection in which case a crest isapproaching) or to raise the thermocline (for the case of cyclonic vorticity advection, asso-ciated with an approaching trough). If the vorticity locally is changing in a positive Sease,i.e., becoming more cyclonic with time, then we should expect downward motion: vorticityis most easily produced locally by convergence, which will be accompanied by downwardmotion. Note that if the advection at a point is cyclonic and the local tendency is positive(i.e., becoming more cyclonic with time), the two effects compete. If the local tendency isexactly equal to the advective tendency, then there is no vertical motion (and, inckddutalby,the phase speed and the float parcel speed are equal). Normally, the advective term islarger; that is the case for 75% of the 23 cases considered here.
Finally, note that the mapped IES fields are a refined version of those described inTracey and Watts (1991); in those fields, the first guess field (i.e., the so-called "mean"field) was a broad diffuse Gulf Stream (see their Figure 6). The fields in Tracey and Watts(1991) contained measurable bias that could be traced to this first-gues mean field. Thefirst-guess field used for this study is a 31-day running mean field produced from the data inTracey and Watts. Because the first-guess field is closer to the actual field, the objectivelyanalyzed fields are more accurate. A further refinement, that used pseudo-IES data fromCM data to fill in gaps in the IES data, and for which IES z 1 2 data was recomputed to reflectnew B-intercepts based on a different Q curve (see Watts and Rossby 1977 for definitions),has not been incorporated into this study.
I 3 Case Study Comparisons
3 To compare wcM with writ, We selected those floats that passed within 10 km of a mooringon which all of the top three CMs were functioning. The 10-km cutoff was chosen asa compromise between increasing error (due to the highly sheared environment) as thedistance increased from CM to float versus obtaining a useful number (23) of comparisons.The requirement that the three top CM functioned was to minimize error in this preliminarystudy: as previously noted, the vertical motion could be calculated using data at ouly 2 CMlevels, but this introduces an error; furthermore, requiring the presence of three levels ofdata minimizes the error in the mooring motion correction algorithm. Note that the dataare chosen based on the CM; these sites/times chosen may not have optimal 1ES coverage.We will note where this may cause errors.
1* 5
I
We carefully scrutinized the CM data records to make certain that submesoscale coherent Ivortices (SCVs), as described in Bane et aL. (1989), were not present when comparisonswere made. Because SCVs typically affect only one CM on a mooring, causing up to a3600 rotation in the current vector as well as a temperature perturbation, their presenceobviously would adversely impact this primarily quasi-geostrophic study. No evidence ofobvious SCVs was found in any 5-day interval centered on the 23 closest-passing float casesstudied here. It is possible, however, that there could be cases in which subtle submesoscale Ifeatures partially detected in CM data and not easily identifiable as SCVs (and thereforepresent in wcm) are filtered out of the objectively mapped IES streamfunction fields (andtherefore not present in wt,,,), i.e. submesoscale eddies could causeerrors as shown inPanofsky(1951).
For each of the case studies that follow, several figures are presented. The track of theRAFOS float through the Central Array, as well as the pressure record of the float willbe shown. Mapped fields of w,,, will also be shown, as will mapped fields of W &adVs. V(C, + f), i.e., the two components of w.,,,. We will also show a figure of the termscontributing to wcN.
3.1 Cam 1: Float 123, CM H4, 4 July 1988, 0000 UTC
3.1.1 Overview
The time surrounding float 123'4 passage by CM H4 (Figure 1) is one during which a serdaof crests and troughs propagated steadily through the Central Array. This is rdected In Ithe float's motion through several crests and troughs (made obvious by the large vrtial ex-cursions associated with changes in path curvature), in the IES Z12 topography (Figure 2a),which shows a crest (at x = -100) and a trough (at x = 40) in the Central Array, madin the CM records, which show a distinct oscillation in current direction and temperature(Figure 3a,b) caused by propagating features. 33.1.2 Vertical motion results
WP= -0.58 mm s , WtoS = -1.42 mm s-I , IWC = -0.98 mm a-
Float 123 moved from 532 to 550 to 565 db (Figure 1) for the three times closest to Itspass by CM H4, a mean vertical motion of -0.58 mm s- for those 16 hours. When Soat123 was closest to CM H4, the CM data suggests that vigorous downwelling was occurriag.Temperatures were increasing as isotherms descended (Figure 3b) in advance of the ap-proachdng crest (at -1.13 mm s-1 ) concomitant with a slight veering in the currents withdecreasing depth (Figure 3a) (suggesting weak warm advection and upward vertical motionof 0.15 mm s-1 (Figure 3e)). Note that for this case, using the average of the tempeuitrchange at 400 and 1000m to estimate the temperature change at 700m yields aa incauectanswer (Figure 3d$). w,,,, for the same time (0000 UTC 4 July 1988) was -1.00 mms-1.Based on advection alone (Figure 2d), vertical motion should be downward at 1.67 mma-' ;strong anticyclonic vorticity advection forces the local vorticity to become more an momenegative. Aaticyclonic spin-up would be accompanied with downward motion to force dowmthe thermocline. The local tendency at the H4, however, contributed to upwwd verticalmotion (0.25 mm s- ). Upward motion is forced by divergence, which destray* vorticity.This is an example of the advective part of the vorticity equation being conteracted bythe local tendency. 3
6
I
620
0246810 12 14 16 18 20 22 24 26 28 30II~460
i (~d• i.t ae~thom •uRt-o AM OS lof123 1 M • wll
3a 50
I.
3660
39
77 76 75 74 73 72 71 70 69 68 67 6U 65Laongtude (W)
F~gur 1: Pressure record in db (top) and horizontal track (bottom) - points after the Ame(launch) paint awe eight hours apart - of LAFOS float 123. CM locaion withis the CeatralArray awe indicated by open circles. The three closest ILAFOS fixes to CM H4 awe band.Launch date of float: 1037 UTC, 29 June 1968.
1 7
I
Iso so
4
0 0
-4074
-120 -120
-0e4 -100
-160-120-60-40 0 40 80 -16-120-60-40 0 40 80Tm (tendency) wm (total) I
Figure 2: (A) Objectively mapped depth of 1?° isotherm ('12) as measured by IES darta Jota0O00OUTC 4July 1968, the tlme when float l23 was closest toC 11H4. Contourmiterval Uevery lO0m. Aiso, deep current stick vectors plotted at CM locations, with a hey (value= 0.25 m -s') plotted at z = -15,5, p = -180. (B) Objectively mapped :,a topography J
(bold sold lines, contours atl50, 350, 580,750, and 9s0m) and portlom ofD forcud byvorticity advection (thin solid lines, contour interval 0.50 mm s-' , negative values daubed).(C) Objectively mapped z12 topography (bold solid lines, contours at 150, 350, 550, 7MI
sad 950.) and portion of D) forced by local vorticity tendency (thin linme, contour ltermidU0.25 mm e- , neatv values dauhed). (D) Objectively mapped al, topogrphylD (held reidlimes, contours at 150, 350, 550, 750, and O50m) and total vertical motion computed he.
(7) usig !ES dat (thin lines, contour interval 0.5 mmn s- , negative vaue dashe). ib ,an four plots, the track of RAFOS fLost 1231 is indica~ted by dots, sad the location ofC314 Is maked by a circle.
IU
8!
90. 'l - 16.0
.0 12.0 To•,t•70. 80ArSN~re I "O -:E I +- e-,-
IIaI o. •,.,W
70. --. 00 i- . - --------
U.0 1.001.20o -1010= "0.0 - - 1-- ...--.003 ~ ~~0.40 i.."--(
SOA-- ........- --I 1.00.00' " -- "in inM AMV
I 1.00 &3d/. 1.00
3.00 f•n ' ---. 00{. w #td} 7•i-1.00 Wallv ---. 0 •"•_ O w
-l' l""IZ'T ' --- ... -,- I'' - I--' ' '
A foe ASW Av WAl
Figure 3: CM H4 data for 5 days surrounding Boat 123's closest approach to N4, wkhi timeis indicated by a vertical line. (A) Direction (in * ) towards which the currmt is Lo g(solid line: 400m; long dashed line: 700m; short dashed line: 1000m). (B) Tmpuratumat CM site In degrees celsius (solid line: 400m; long dashed lime: 700m; short dashed hIe1000m). (C) Current Speed in m s-(solid line: 400m; long dashed fne: 700s; short dashadline: 1000m). (D) Vertical motion (in mma-' ) forced by local temperature tedmbcy (saidline: using the observed temperature tendency at 700m; dashed Aime: uwlng the aveap of400 and 1000 m temperature tendencies to estimate the 700m temperatum teadcy). (E)Vertical motion (in mm s-' ) forced by backing/veesing curreat (dashed Us. eamputudas in (4) usaingtv - aft; solid lne: computed as in (5) usingR . (F) lbtagveamotion in mm.s-?, i.e. the sum of E (dashed line) and E (solid anddashed•a asnl D).
* 9
m
3.2 Cae 2: Float 129, CM 14, 23 September 1988, 0800 - ISM UTC I3.2.1 Overview
ILAFOS Ut 129 spent considerable time in the Central Array (lqpm 4). Although itentered moving steadily on an anticyclonically curved path, it then apparently escapedtemporarily from the stream and moved only slowly before a meander trough moved throughand started to steer the float again steadily to the northeast. The IES topography for thedates when the loat was closest to CM 14 (Figures 5, 6) do show a comsiderable trough neathe center of the Central Array. CM currents (FIgure 7) are weak (les than 0.5 m s-1).3Temperatures are slowly Increasing.
450In
750 a. .I*S
0 2 4 6 8 10 12 14 16 18 20 22 24 26 26 30 IDAVIS AM UUS
41RAFOS 129 3
40
,39
]37 .0* I
35
7776 75 74 73 72 71 70 09 66 67 66 06
ongitude (W)
Figure 4: As in Figure 1, but for toat 129. The four closest RAFOS bw to CM 14 areboed. Launch date: 0638 UTC, 31 August 1988. 3
3.2.2 Vertical motion resultsI
w a019MM#- , C = 0.53 mm &- = 0.04 mm a-' (000 UTC)
10 1
-40 -40
-120 -120 •
3-160 -160-- -1j2Q-80--40 0 40 80 -160-120-• 0 -40 0 40 80
Zia wm (adveoaton)
1120 120| so s o .,
40 60
-120 -120
-160 a a a A-16Goa0-16O-120-W0-40 0 40 60 -160-120-60.-40 0 40 80
wm (tendency) vm (total)
SFigure 5: As in Figure 2, except for 0800 UTC 23 September 1988. The trak of fset 120is indicated by dots, and the location of CM 14 is marked by a circle..
WIUP = 0.72 mnm s-1 , CWM = 0.48 mm s-1 , win& = 0.05 mm s-i (1600 UTC)
For the four closest loat locations to CM 14 (Figure 4), the oat upwdled frmn 717 to711 to 706 to 669 db (between 00 UTC 23 Sep and 00 UTC 24 Sep); the two mum. valuesof vertical motion ar 0.19 and 0.72 mms-1 . Clearly, the floUt is beginning to acederateupwards as It moves out of its position at the base of a trough. CM data also smq t tigmotion. Temperatures are cooling, suggesting that isotherms are rising (11gm 7d,e). Atthe same time, there is significant veering to the currents from which we can isn tingmotion (Figure 7a-c). The vertical motio as estimated by the thermodize topography (E3data) is considerably smaller, about 0.04 mmua' for both times (Figures 5,6). That partdue to advection is nearly zero and the part due to local tendency is about 0.04 s- s .We note, however, that a lare area of ascent is located just to the north of 14.
UI
III
120 120 i
80 4 -0
-40 .40
-120 -120Iftu 040s IW2-0 0 408
-120 -1203
-160 -160
-160-120-80-40 0 40 80 -160-120-M0-40 0 40 80wm (tendency) wm (total) 3
Fqgure 6: As in Figure 5, except for 1600 UTC 23 September 1988.
1IUU
TIfilI
3 0.8 Spe irm 14.00
50. 400 --.00
600.00~1.0t,',%" *'&1.001 toU
0.01m-.00 1 T
I1.0• -{.0o-O 1. -1 :0.40 - p ed" ,n' --.O'O W ---- j.[,
i -- --- " - -- I.o 1--0
m-1.00 to 14,bothofwh t--. 0 0 Wcteb a--- --- - ... M . -
Figure 7: As in Figure 3, except for CM 14 for the live days surrounding los 129'8 desistapproach~es to 14, both of which times are indicated by vertical lines.
13
3.3 Case 3: Float 136, CM 14, 8 December 1988, 1600 UTC I3.3.1 Overview
The Gulf Stream in the period during which Rost 136 (Figure 8) moved through the CentralArray was c ae by a succession of crests and meanders. Float 136 initially movedfairly slowly through the Central Army along the southern edge of the stream. Halfwaythrough the army, however, the float accelerated to the northeast, apparently influenced by Ian approaching propagating trough (Figure 9).
II~I5 5~I U~5 5 *~I I I I oi l III I 1 1 * 1 1*
3. 30 2 4 -72 1 1 0 12 24 6 30
MAIN~d (W)LAF'-e 136s
boe.I~nhdte 917T,1 Noeme 18..*
34 , -c : 0.4 mm - I , I I I- I,
77o 73 6 75ve 74o 739 7o68 o2 71 70in 6he 68m 67 w66 66m oCMI F
40 |boxed. 7a6c date 092 UTC 19 November 9 88. 6 6
3.3.2 Vertical motion nmults
WM = 0.81 mms81 , wc = 0.43 mum s-I , w,,,,=0.17 mm s 1 3Float 136 moved from 669 to 668 to 622 db during the time it was closest to CM 14 (F*g
ure 8), reflecting the acceleration up and northeastward out of a meandgr trough. CM data3
14
120k '\N. 120
* 80 so8
40 40 C
140 r40-400
-120 -120
-160 -160-M159Q-80 -40 0 40 80 -160-120-80 -40 0 40 80
Z12 wm (advection)
12 120
*40 40
0 0
I-40 4
3-120 -120
-160 -160W-160120-80-40 0 408 -160-120-80-40 0 40 00I n (tendency) wM (total)
Figure 9: As in Figure 2, except for 1600 UTC 8 December 1988. The track of float 136 isindicated by dots, and the location of CM 14 is marked by a circle.
1 at the same time show currents that awe veering with decreasing depth, suggesting upwardmotion (Figure l0a,c~e). At the same time, CM temperatures are steady (Figure 10b); theI ~majority of vertical motion diagnosed from CM data Is related to the champe of the cumetvector with height. Vertical motion determined from IES data shows weak upward modoecforced by vorticity advection (Figure 9b), and weaker still downward motion related to theI ~vorticity teadency (Figure 9c); the net motion (Figure 10d) is upward.
3 15
IIII
D ectic -M:, 18.0~90. 3- - B
7.10.0 TempeuwIo.o.-Speed / r-._ 000,~o0.40:--.1,.oo._ .T_,- .
60.0
50" ,M-I : . -
1.00 " ..... i• ]"•----dv/d 1.00 Aj ..... ..... •
.00 - - -- ----- -- - - -- --- --- --
iw y -1.00D
Figure 10: As in Figure 3, except for CM 14 for the five da~y. surrounding fost 136'.osest
approaches to 14, which time is indicated by vertical line.
IIIl
1. I
3.4 Case 4: Float 141, CM 14, 2 November 1988, 1600 UTC
3.4.1 Overview
RAFOS float 141 (Figure 11) moved steadily through the Central Array at a time duringwhich flow was from the northwest to the southeast, suggesting a large ridge upstream of theCentral Array. However, there are wiggles in the float path suggesting smaller-scale crestsand troughs are propagating through the larger-scale ridge, and this view is confirmed bythe IES z12 topography (Figure 12).
1~450I .. ,
7500 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30DAY3 ArMi LUNCK
41 FO 141
i4 -400
37 0
36
* 35
3477 76 75 74 73 72 71 70 69 68 67 66 65
Longitude (W)
Figure 11: As in Figure 1, but for float 141. The three closest RAFOS fixes to CM 14 areboxed. Launch date: 0921 UTC, 28 October 1988.
I3.4.2 Vertical motion results3P = -0.87 mm s- , WCM = -0.94 mm s-I , Wto# = -0.43 mm
RAFOS float 141 downwelled from 589 to 603 to 639 db (Figure 11) in the period closest3 to CM 14, which Is a mean motion of -. 87 mm s-. CM data for the time of closest ap-
17
I
120 • ,,Jo,'-• 120
80- 40
0 0~~~
-80 °*-80
-120 -1.20
-160-1.IZQ-80-40 0 40 80 -160IW120-80-40 0 40 801
Z1, wris (advection)
ao o so800
-40 -40 - ,
-80 ' 80 '
-120 -120
-160 -160-160-120-80-40 0 40 80 -160-120-80-40 0 40 80
wM (tendency) WI (total)
Figure 12: As in Figure 2, except for 1600 UTC 2 November 1988. The track of float 141 3is indicated by dots, and the location of CM 14 is marked by a circle.
proach also suggests downwelling; although the currents are veering, suggesting upwilling U(Figures 13a,ce), temperatures are warming rapidly enough that the accompanying down-welling dominates (Figures 13b,df). Note that here, again, using the average of 400 sad1000 temperature tendencies to predict the tendency at 700 m leads to an underpredictiomof the magnitude of the vertical motion caused by the local temperaure tendency, sadhence a sign error in the resultant total vertical motion (Figures 13df). JES data showsthat CM 14 is downstream of a crest; anticyclonic vorticity advection associated with thiscrest is forcing downward motion (Figure 12).
I
IIiiI
110.- l w :M 14.0
go. :10.0
1.60:1 ----------- 400M, -- -----"x.- -'-- /-- -
0.40 : - -0.00o -
0.00wg 19je 1.00 Wj~d00 I n l I-"'-- i- VT '
asM UM ,U11.00 4 U *WA 1.00 F
0.00 r-Ao
I
Figure 13: As in Figure 3, except for CM 14 for the five days surrounding float 141's cloestapproaches to 14, which time is indicated by vertical line.
iI
19
I
3.5 Case 5: Float 175, CM 14, 20 January 1989, 1600 UTC I3.5.1 Overview
As was the case for float 141, RAFOS float 175 traversed the Central Array moving fromnortheast to southwest on a track consistent with an upstream crest (Figure 14). Therewere wiggles in the track that suggest smaller scale features propagating through the largerscale crest. This is confrmed by the map of 12 (Figure 15).
250 .............................4I
25W
%15 I
0 2 4 6 a 10' 12 14 16LAEi udM (W)
41
RAfOS 175 3
39e.Luc dt:105UT,1 aur 19 0 I
400
*39
z.
5 37 I36I
35I
77 76 75 74 73 72 71 70 69 66 67 66 65Longitude (W)3
FlIgure 14: Asin Figure1, but for Boat 175. The three closest RAFOS fixes to CM 14aweboxed. Launch date: 1005 UTC, 14 January 1989.
3.5.2 Vertical motion results
WAP= -0.58 mm s-1 , We,, = -0.05 mm 9-1 , U,13, = -0.63 mms-
RAFOS float 175 moved in the vertical from 328 to 364 to 361 db during the time whenUit was closest to CM 14 (Figure 14), a mean downwifling of -0.56 mm s-1 . Note that thefloat is near the bottom of a trough and is about to commence upwelling dowastream of ft.3
20
II
120 120
40 40
0 0 0 .,\--40 • -40 4-',
-80 -80
-120 -120i ~-160 r I-IGO
-IXLM-80-40 0 40 80 -10-120-80-40 0 40 80
Zia wnM (adveotion)
120 0.00 120
80 80
* 40 .40
-120 - -120
-160 -160-160-120-80-40 0 40 80 -160-120-80-40 0 40 80
uwM (tendency) wM (total)
I Figure 15: As in Figure 2, except for 1600 UTC 21 January 1989. The track of float 175 isindicated by dots, and the location of CM 14 is marked by a circle.
CM vertical motion is also flipping sign (Figure 16) as the float passes by 14, rdecting thepropagation of the meander crest. Vertical motion is still slightly downward, bowever, asmeasured by the CM when the float is close by. Vertical motion as diaposed by verticalstretching data at 400 m is also downward; CM 14 is downstream of a crest (Figure 15).
I
* 21
i
i
I
II
130-. ----- - - - --
120. L AE -1B 16.0110. " 12.0
100. ie~ 8.0~ Temnperur F9 0.ArI -l-O 4 .0 1 1
~a amoo Ul
1.20 IM~ m .0
0.80 sdrm.-0*.000.40 --ed 1.'00 w4a I0.00 . . -I- --
toam U an
a• #AHs 1,.0, , -. ,. -. 00- '-l
.0 ,,,, . 1.0 , U
Figur 18: As in Figure 3, except for CM 14 for the five dayp srrouading Goat 175's doomtapproce to 14, which time is indicated by vertical line.
IIIU
I
l 3.6 Case 6: Float 176, CM 15, 7 February 1M89, 0000 UTC - S February 1M89,0000 UTC
3.6.1 Overview
Float 176 moved slowly through the Central Array at Am (Figure 17), tracing out anaticyclonic path, before being expeed from the stream, after which time it drhfe veryslowly southward. IES data (Fiqgure 18- 21) show the crest oa the eastern edge of theCenmtral Arrmy, and also suggests a sharp southern edge to the stream: loat 176 is movingvery slowly in a region that is close to the main part of the Gulf Stream.
-5W0
4 6W
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30RATEl APIMLAý
RAFOS 176
II39 7
38 -8
Z 37 * 6
* 36
35
77 76 75 74 73 72 71 70 69 68 67 66 65Longitude (W)
Figure 17: As in Figure 1, but for float 176. The seven cosest R.AFOS txea to CM IS aeboaed. Launch date: 1131 UTC, 25 January 1989.
3 Z3.6.2 Vertical motion results
W"P = -0.04 mm s- WC, = 0.02 mm a- WtS = 0.25 mm - (0000 UTC TFbruary)
23
120 A'120I
-40804 408 •°•e-2-o40I oo
40
-40 -40
' "~
-. 80 -80
-120 -120
-1-0 0 W -•m- - (o-•-e4o I120 120
40 40
40 I8~ (tI1iD )w II-0 -80
-120 -160%-160 10-12V0.-W -40 40 80 10-IW100 4 08
Wm (teudducy) =Od,
Fig•re 18: As in Figre 2, except for 0800 UTC 7 February 1989. The track of fLoast 176 1.inda& by dots, and the location o( CM 15 is marked by a circle.
P= -0.05 m s-5 , wac =0.03mm s- ,WIS, 0.26mma' (0SW UTC7IFbruary) IP M0-0.04MM 2-1 , wvm = 0.02. s- ,I w,,= 0.27 mm - (1600 UTC 7 Fbay)
W,&P = -0.03 mm s-1 , WCM = -0.05. s- , WS = 0.26 mm s- (0000 UTC 8 Jebu-mry)
As the RAFOS lost drifts southward in the vicinity of CM 13 (It remains within 10 ki d15 for aday), it stays between 761 and 7Mdb (Figure 17), so vertical.motioa is nsegia. Is 3addition, CM data diagnose small vertical motion (Figure 22); there Is yeerin and backlongof curmnts, but the curents are extremely weak. Similarly, local toepatur, chain" wmvery small. JES data (Figures 18- 21), however, diagnose upward vertical motloa1M P I eI by
qdncvorticity adveetlon (local tendency of vorticity is wear sue). wj,,, is narlyt OM6MM8- for the dayloog period when the Bloat is close to 15; however, 15 is near the eWV of
the IES mapping riegion, and the derivatives required to compute vorticity and Its adveetom
24Lm
I
1 120 .'120
so .80
40
-40-4
*8 -60
-120 -120
- I160o -160-ljfrj3-80 -40 40 80' I10W4 08
1~Zi v12mw (advedtou)
120 120
I 40 U' qu e* I*-
6-40 - 140
1-120 -610
-160 -160-160-120-80 -40 40 600 -W2- -0i 0
Wm (tand~ney) (o,
3 ~ ~FIgur 19: As in FIgur 18, except for 1600 UTC 7 Febrwy 1989.
may not be accurIe
2,5
IIII
120 120I
so so340 4
-40 -40-80 -80I
-120 -2o0
-t .9-M-40 40 8o -1W 120-80-40 406s0Z•wi (adve On)
120 120
go so
"Sm
40 'I40 S
40-403-g0
-120 ___120_ _
WM10-04 (40ud 8c -160-120-80 -40 4060o
FipMn 20: As in Figur 18, except for 0000 UTC 8 February 1989.3
263
120 120
so so
40 40
I u 0
-40 -40
*8 -60
-120 -120
-15Q12Q-80 -40~ 40 80 -10108-40 40 SOZia w, (adve~lon)
1120 120
80 80
-40 I 40
I-80 -8
-120 -2
*6 - 160- 120-W 0-40 4o SO -160120-So-40 4o SO*wm (tendducy) o(&'
F~pr 21: As in F14ur 18, except for 0800 UTC 8 Februawy 198.
1 27
140 . - ---- -- --- -- 99 -- - - - - -. . , ,- --.. . . . -S...• •-M 18.0
120. 1oo4.0 •ilm• •100,. ". 1 -,--• I --- V"
1.110 ------ ------ -... . -.-.- ---- - -- --- --
0.8l IIII fm-"-' "°°- ,,,, -I0.40 LSOla.&1 -" -. 0 w
0.00 , J 1o ,• JI ---., Iil -- n-G ftbtro Ill '-1.00 &/ 1 -1.00
Im
1e0. e 4.0 I
-1.00 Wa& -Z6.00 to
Ftiure n: As in Figure 3, except for CM 15 for thiedays surroundingBot16sdM
" V.I
"apr w e to1,wihtmswe Iniae y veria lines
28
i -I
I 3.7 Case 7: Float 194, CM 15,15 November 1989, 0000 UTC
3.7.1 Overview
RAFOS lost 194 moved through the central array during a time of significant stream-ring interaction. This is reflected in its unusual south-to-north path through the array(Figure 23), moving around a cyclonic cold eddy in the southern part of the domain, ^'around a crest in the northern part of the domain (Figure 24).
I
I . . \ .4. 0.
0347 7 6 8 10 12 714 8 20 26 23
Figure ~ ~ ~ ~ AT 23 Ai igr ,bt forfl 194 h he coetRFS MoC 5a
40 u
437.*
366
I .3534
77 76 75 74 73 72 71 70 69 68 67 66 65Lonigitude (W)
Figure 23: As in Figure 1, but for float 194. The three closest RLAFOS ANNe to CM Is aneboxed. Launch date: 1843 UTC, 5 November 1989.
I 3.7.2 Vertical motion results
WAAP = 0.71 mm s-1 , WC = 0.23 mm s- , = -0.22 mm
As float 194 moved around the cold eddy past I5, it upwelled from 663 to 662 to 622db (Figure 23), a mean value of .71 mm s-1 . This is just as expected from the work oBower(1989); the float has emerged from a trough and is moving towards a crest. The path
29
I
1,20 1. 20• "80- ..soal
40 40 I0 01
-40 -4
-12012-160-8G.1 A 40 I8a -160 1
G-40 0 40 80 160-120:-00-40 0 40 80
Zia w= (advection)
120 - 120 I80 *80
40 aa40
.o I
0 0-40 - ,0
-a-S
-120 -120-120 -. "
-160 -160-1I W-I20•400-40 0 40 80 -160120:-I0-40 0 40 80i
w= (tendency) wU (total)
Figure 24: As in Figure 2, except for 0000 UTC 15 November 1989. The track of Boat 194is indicated by dots, and the location of CM 15 is marked by a circle.
is more convoluted than normal, however. CM data at the same time at 15 also suggest iupwelling (Figure 25): Currents are veering, and temperatures are cooling slightly. Bothof these will force upward vertical motion. IES data (Figure 24) for this time, however,suggests downward motion. Both vorticity advection and local tendency are forcing down-ward motion. A possible cause for the discrepancy in diagnosed vertical motion is discumedIn the next section.
I
30
i
I
I----,- -.--o--.
4,0.: sooB
30. 14.0 -- -20. 10.0 Temper tu'e10. l''' 1 6.0-ir' 'WYTY J
1.60 - ------- '-.- -4W = - - -
1.20 - a1.000
I 0.40 a0.00 - ag" ' ----O.O0 -
i o3om,•• i ,t2itw 18 8 M4
o~oo-- ---- ,, ----- - -: -- • ---- -- '-- --1.00 &/E Ua 3 1.00
0.00. I jm-0.00 - aEru~L
-i~~"&W fm -. -- r0r0i -- ~1.00 Wadv --- 1.00 2Wtotid ',
,!IN• '1 or IBM gag;
II
Figure 25: As in Figure 3, except for CM I5 for the five days surrounding float 194's closest
approaches to I5, which time is indicated by vertical line.
IIII* 31
I
3.8 Case 8: Float 194, CM 14, 16 November 1989, 0000 UTC I3.8.1 Overview
RAFOS float 194's unusual path (Figure 26) south to north from ring to crest (Figure 27) Iis described in Section 3.7. I
550
5 i1: I
~ 50
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 iDAS AFIU LAUIICK
Figureýý F'S''9 26....Fgre..T..heecoes.AOS..s oC 1 r bxd Luc
40
36 -.
35I
77 76 75 74 73 72 71 70 69 68 67 66 65
Longitude (W)
Figure 26: As in Figure 23. The three closest RAFOS fixes to CM 14 are boxed. Launchdate: 1843 UTC, 5 November 1989.
I3.8.2 Vertical motion results
W)UP = 0.86 mms-1 , WCU = 1.20 mm s , w,, = -.1.06 mm, i-
Float 194 continued upwelling as it passed CM 14, moving from 619 to 584 to 570 db(Figure 26). This is a mean upwelling of 0.86 mm s-I . CM data at 14 also diagnosedupward motion (Figure 28), forced mostly by strong cooling observed at the current metersite (Figure 28b,d). Vertical motion caused by currents veering with height is smaller insize. In contrast to the upwelling measured by the float and diagnosed by the CM, EES data
32
I
, 1020
40 4
0 0
-40 -40
-120 -120I -16
- -IkjC)8G --40 0 40 80 -160-120:-80-40 0 40 80
Zi wM (advectIon)
40 80I e•i,:• I. •. •
I 0
-40 ***-~*40
-120 -.- - -°120
I 6G- I20ý-80 -40 0 40 80 -16 0-1-120:-80 -40 0 40 80I w (tendency) wma (total)
Figure 27: As in Figure 24, except for 0000 UTC 16 November 1989. The track of Rlos 194is indicated by dots, and the location of CM 14 is marked by a circle.
(Figure 27) suggests downwelling. Weak downwelling is being forced by vorticity advectiam,supplementing the stronger downwelling forced by the local change in the vorticity. Thisconsiderable disagreement between tow,, and the other two methods is discussed in the nextsection.
33
IiIII
365.- - - - - -- - - - - - 4 - -- - - - - - - - -
-Tun" l 1.0008.0 ---------- - - - -M -O.O 0 ---- ---- ---- ---
C - -MM 1.0
1.20 v" IM'-niO ta
0.40 . - 1.00 w-. .
0.00I E-1 "IIm.-• .f. . "l
Ii
Figure 28: Au in Figure 3, except for CM 14 for the five days surrounding float 194'. closestlapproaches to 14, which time is indicated by vertical line.
III
34I
3.9 Case 9: Float 199, CM I5, 23 November 1989, 0000 UTC
3 3.9.1 Overview
The circuitous path that lLAFOS float 199 (Figre 29) took through the Central Arraywas unique for SYNOP floats. The thermocline topography (Figure 30) shows two distinctKm features: an east-to-west Gulf Stream to the north of a cyclonic cold eddy. Float 199 leftthe stream and was entrained into the cold eddy, circling three times before being expelled.
6500-
* *700
0 2 4 6 8 10 12141618'20 2224 26 6 30DAYN ArMI uMM
41
2AeRAFOS 199
I 39
*378
36
1 3534 p p
77 76 75 74 73 72 71 70 69 68 67 68 65Longitude (W)
Figure 29: AsinFigure1, but for float 199. The three closest RAFOS faneto CM15 (Amrapproach) are boxed. Launch date: 1258 UTC, 10 November 1989.
3.9.2 Vertical motion results
wja,= = 0.24 mm s- 1 , :~ =0.23mm.' , wigs = -0.13 mm#-I
As float 199 circled the cold eddy and passed close to 15 (the first time), it was movedfrom 588 to 593 to 574 db, a mean upwelling of 0.24 mm s-I (Figure 29); at thi time, thefloat was in the middle of a vertical displacement of some 70 m. CM data (Figure 31) for
35
Ii
40 40
0801
-1 W120-801-40 0 40 80 -160-120-80-40 0 40 80- Z1 wI (advecton)
80 so
-Q 10-40
_aO _W. •-120 312-160 _6
-16W120-80-40 0 40 80 -160-120-0-40 0 40 80wm (tendency) Wm (total)
Figure 30: As in Figure 2, except for 0000 UTC 23 November 1989. The track oafoat 199 9is indicated by dots, and the location of CM 15 i marked by a circe.
the same time also suguets upward vertical motion: although wrests are backig withdecreasing depth (suggatve of downwieling) (Figure 31a,c,e), the Is a slmultaaes 4e-crease is temperaftur associated with a rise in isothermal surfaces that owuwheha thisdescent (Figure 31b~df). Vertical motion as diagnosed firom KES data (figre 30), how-ever, suggests weak downwelLing as a result of both local vortcity tendency and vortid tyadvectioua.3
II36!
I
I
I11 -...-.----------------- ..... --------- ;
1 ~100. 1A0
0. _- 10.0 Tempe turo
1.6 0 " ---. . . . . . . ."---- - -- - I . - -- -. . . . . .- - . . . . ..- - ---C 16.00
1.• -"a
0.40Z
0f--1.00 - TA
0I0 1,1 11.00 -- O-s-. 1.00 ,-
_,,,. ;; ...-_o, =,, 1 ---Oft I' I'ia film I lia'I
I|Figure 31: As in Figure 3, except for CM 15 for the five days surrounding loat 199's desist(flirst) approach to 15, which time is indicated by vertical line.
37
I
3.10 Cam 10: Float 199, CM I1, 28 November 1989, 1600 UTC i
3.10.1 OverviewSee Can 9. At this time, the goat is still cirding around the cold eddy (Figure 32) thatSpree prominently in the IES topography (Fqgure 33).
7500I
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
4 1 - 1" , 1 1 0 " m - 1, - 9w - 1
41 RAFOS 19940
30.
Sz
537I.1 136
35 134
77 76 75 74 73 72 71 70 69 68 67 6 65LonGitude (W)
Figure 32: As in Figure 29, but for second clod approach.
3.10.2 Vertical motion reiults IwAAz = -0.13 mm s-I , =0.04 mm s' ,w,,, =0.54mms'
As RAFOS float 199 passed by CM 15 for the second time, it moved aum 613 to 627to 621 db (Figure 32), a mesa downwelling of -0.13 m sm ; hoswee, immodiatuly aft"
pasing 15, the lost upwelled about 75 m in 2 days. CM data (Figure 34) shom upqmdng,caused mostly by local temperat changus (F'igure 34bdf); cumrrts ae switching bsetwbacking and veering when the foat passed by, so vertical motion from this elect was war
"38
120
40 4
-00
-80 -
-120 -
-160ja-1-8-40 0 40 80 -160120-80-40 0 40 60Zia (tneny m (aotiota)
Fiue3:1s20 igr 0 ecp o 1600 8Nvme 99
zer (Fgrsoc) E etclmto Fgue3)a h aetm ustduwr
mo4o 0.5 -40)fre otl yccou oriiyavci
pp4
-120 -32
I
18.0
I_ II I ~ L
14.0 I1.• - • ... ... .. ... .. • O~ - - . ....-
00. D 0 •-.0"00-6.0
0.o -_ _ rm ,.t--o1.oo'w .Io1 o0 " --- t--- o--- .o------- ---. .--- w-- --- ,------
0.00 3- 11 i P w"--'0.00
00 . - I l 1 -- ----I= m I ft s_ --
IFgu 34: As in Fgure 31, except for the second cle approack by lost 1N to 15, W kWtime is IniAte by vertical line.
III
* a|
3 3.11 Case 11: Float 201, CM I1, 23 November 1989, 0800 UTC
3.11.1 Overview
Float 201 moved though the northern part of the Central Army (Figure 35) at a time whenthe thermocline topography (Figure 36) indicated a vigorous cold ring in the southern partof the array (In fact, float 199 was circling this eddy as float 201 moved through the array).Z12 data also show the sharp trough in the center of the Central Array which float 201pased just before being expelled from the stream to the north of the Central Array.
3 ~50 ..... rr~. pqrr~-.r-.*rr.rrr.r-
-150
*250I ~3503 1450it550
650- - -0 2 46 8 101214161820222426283032343638404244
rAY8 APMR IAýf. 411 1 1 1 1 " I I I I .41 3
3 3
34
77 76 75 74 73 72 71 70 69 68 67 66 65Longitude (W)
Figure 35: As in Figure 1, but for float 201. The three closest RAFOS fixes to CM II areboxed. Launch date: 1550 UTC, 15 November 1989.
3.11.2 Vertical motion results
wRAP= 0.97 mms 1- , wC¢ = 1.51 mms-1 , wISS : 1.66 mm s-1
As float 201 passed by CM II, it was just beginning an ascent that would take the floatfrom 600 db up to 150 db (Figure 35). For the three times that the Boat was closest to II,
41
-40 4
-0 3-120 -2
-160 1 G --1.120-80 -40 0 40 80 -160120-60-40 0 40 600
Zia wm (advecion)
ItIIt~
0 0P8
-10-1201-160 -160
-1I6S-1I20-W0-40 0 40 80 -160-20-60-40 0 40 80
Tm (tondency) WM (total)
Figure 36: As in Figure 2, except for 0800 UTC 23 November 1989. The track of float 2013is indicated by dots, and the location of CM 11 is marked by a circle.
it moved from 575 to 50 to 519 db, amean upweliing of 0.97 mma-' . CM data for thistime also suggest strong upward motion (Figure 37). Not only are temperatures cooling
signifcantly (suggestive of strong upwelling), but there is also signifcant veewing. Note thatIthis Is a case where the computation of womC using R2# leads to an (assumed) erroneouslylarge component of vertical motion due to backing and veering currents (Figure 37a,CGe)that does not occur if 10Cjw is computed using vpjk - uft. EES topography also suggstsstrong upward motion, forced entirely by the cyclonic vorticity advection downstream ofthe sharp trough at x = -40 (Figure 36b,d).I
425
16 . ... . .. . .. . .. • 400 - - - ---- -" . . . . . . . . ..-- -
-0.8 8 pe m --. 0° -tur -'
310 12.0j
'jS•I so• 881um
I
1 .60 *- ------- -- - - - - 41 m -- - - - -- - - - -
1.20- lo a
IIindiate y Spertca line.__.0
I 33
0.0 fo - -Yi 1.00 Wd0.00 _ _ ---.0 r t--
aaw
IFigure 37: As in Figure 3, except for closest approach by float 201 to 11, which time isindicated by vertical line.
34
I
3.12 Case 12: Float 207, CM 12, 4 September 1989, 0800 UTC U3.12.1 Overview iDuring float 207's lengthy passage through the Central Array (Figure 38), the CentralArray was characterized by unusually intense eddy activity (Figure 39). As a result, thefloat followed a complex route, entering the Central Array four times, twice from the easternedge. At the time of passing 12, the float was moving to the west on the southern edge ofan anticyclonic (warm) eddy.
100 .3
. 150
4200 3
II
~250
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30DAYS AtM 2 AeN CA
41 540 139
*38 I
36
35134oe a d 8 U 2 A I 989. L - -L
77 76 75 74 73 72 71 70 69 68 67 66 65Longitude (W)I
Figure 38: As in Figure 1, but for float 207. The three closest RAFOS fixes to CM 12 areboxed. Launch date: 1858 UTC, 23 August 1989.
3.12.2 Vertical motion results
WRAP= -0.25 mm s- , CM = -0.06 mm s-1 , WIS5 = -0.24 mm s-I
Float 207 (Figure 38) moved from 158 to 164 to 172 db during the time it was closest toCM 12, at a level in the ocean that was the shallowest of all the RAFOS floats considered in 3
44 1
I
£12 -A 120
-so sog..08o- o..•
80 • 0
-80 -80
-120 -120 " " -
o16 0 40 Go!-160-120 - 0 40 -160-120-80- 04
w1( wns ( ---- ,
1 9 120 U
0 so e
40 i 40
I0
*wS (tendency W : 0
5 Figure 39: As in Figure 2, except for 0800 UTO 4 September 1989. The track of float 207is indicated by dots, and the location of CM 12 is marked by a circle.
I this report (Float 201 moved to a shallower level after exiting the Central Array, however).Current meter data at the same time also suggested downwelling - albeit very weak -
(Figure 40), forced by backing currents. The effect of the backing currents dominatesupwelling diagnosed as a result of cooling temperatures. IES data suggests downwellingforced mostly by local vorticity changes (Figure 39c).
4II1 4
IaII
270. - -. - 40 - -- ----
260. 0 -M= 14.o0 B250.
10.0
240. 6.0 .U !230. 2.01 3 -1 11
1.60 --- - --- - - --------- -•
1.20: I ." *dl 1.00 0 MD"0.8-0 -0.0 --
0.40 - f,.r 00.00-U
1 .0 0 ]E &"/ds. 1.00 " Ft-
0.0 E -- 00
-1.00 avg.0fW=.tI
- t- - ~~ -1 -- I*-f-'1-- A-
II
Figure 40: As in Figure 3, except for closest approach by float 207 to 12, which time isindicated by vertical line.
41
46 1
£ 3.13 Case 13: Float 209, CM 15, 13 September 1989, 0800 UTC
3.13.1 Overview
RAFOS float 209 moved steadily through the Central Array (Figure 41) on a path thatsuggests the float moved around a large meander trough in the eastern part of the array.This trough is obvious as well in the z12 topography (Figure 42), which also shows a warmeddy moving through the slope waters to the north of the stream.
~500
600
8000 2 4 6 8 101214161820222426283032343638404244
I.
1AYN AF!U L&IAIM
411 RAFOS 209
39
I 06.,38 0*
I boe. 3 7 S
* .36
3 3534
77 76 75 74 73 72 71 70 69 68 67 66 853 Longitude (W)
Figure 41: As in Figure 1, but for float 209. The three closest RAFOS fixes to CM 15 wre5 boxed. Launch date: 1034 UTC, 7 September 1989.
3.13.2 Vertical m'otioni results
WRA= 1.3 m sI WM =-21 4mm s- , WIES = 0.22 mm s-
Float 209 (Figure 41) moved from 584 to 629 to 672 db as it moved closest to CM 15. Themean downwelling from this motion is among the largest observed in the Central Array. At3 the same time, CM data (Figure 43) shows currents are backing, which is consistent with
* 47
I
120 120 ~ * -40 40
0 0-4b -4
-80 38-120 f
-160 -160-1I.Z-80 -40 0 40 80 -160-120-80-40 0 40 W
z,2 wnM (advection)
so I o80 8
40 40
0 0
-80 -80
-120 * -120 I- i o -- - -16o0 -
-160-120-80-40 0 40 80 -16-120-80-40 0 40 80was (tendency) WM (total) I
Figure 42: As in Figure 2, except for 0800 UTC 13 September 1989. The track of float 209 3is indicated by dots, and the location of CM 15 is marked by a circle.
the strong downward motion. Furthermore, isotherms are descending (Figure 43d), also 3suggesting downward motion. IES data (Figure 42) for this time, however, suggests weakupward motion, a result of cyclonic vorticity advection and local tendency. Site 15 is nearthe edge of the IES mapping region; the derivatives of the Z12 field necessary to computevorticity and vorticity advection may not be accurately estimated. A second possible reasonfor the startling disagreement is considered in Section 4. 5
IUI
48
95. 10.0 -e r -,
8 5 . ---r • 6 .0 T a m• i e- ", ;-7 ý' l " 1*M - 7A ------- -- - .g-4 n-- -- ---- 1. I -am -14.0
o5.o 10.0 Deo . ;U.1.60 lo
0.20 -.. 0 D1-
0.00 I t" 1 "I" 1i1" 1 ---.III
o- --o-- .- - - --__- -- -- -
49~b
-1.00 w vo/S10
3 ~Figure 43: As in Figure 3, except for closest approach by float 209 to 15, whiich time is
indicated by vertical line.
*1 49
m
3.14 Case 14: Float 210, CM H4, 6 October 1989, 0800 UTC
3.14.1 Overview
Float 210 passed through the Central Array twice, suggesting significant interaction withnearby rings (Figure 44). Indeed, z12 data (Figure 45) suggests that float 210 is circling adeveloping ring. i
0- m
1650
750- ' A I m
0 2 4 6 8 10121416182022242628303234363864042
DAI nltd()
41 1a miuRAFOS 210
39z0
i~37
36 .*
351
34 377 76 75 74 73 72 71 70 69 66 67 66 65
Longitude (W)
Figure 44: As in Figure 1, but for float 210. The three closest RAFOS fixes to CM H4 areboxed. Launch date: 1326 UTC, 26 September 1989.
3.14.2 Vertical motion mults IWRA,? = -0.02 mm s-1 , wct = -0.19 mm s- , wSS = -0.84 mm s-1
As float 210 moved by CM H4 (Figure 44), it moved only from 636 to 644 to 637 Tdb, despite the anticyclonic curvature to the path that might suggest strong downweUlling.Current meter data at the same time also reflects the slight downward motion (Figure 46).Temperatures are warming, which suggests downwelling of isotherms, but this motion is I
"50
120 A12
40 40
-15 -160LL--15Q:.I.Z-W -40/0 -OD 80I0- 00 D8
Zi w,*(adw~cton)
40-160-1004 0U win' (tndencyl J (ot5 ~ ~ 8 Fiue4:A-nFgr ,ecp o 800 UT coe 99 hetako a 1
(Figure 46). nFgr ,ecp fr00 T coe 99.Tetako ot20i
* 51
i
III
IS --- -- -- -- -- R------ ----I--l
160. B140. , 14.0 I - " I120. 10.0 "mpe .itu e
10 brci N6.0 WITS,Vo --- 1, -• - I-----,I---To , _ • _, _
A. - 4s%, % '-" "m"1.60 -4 I- A m Il.olp. dIot .-nfu - .oo- jD E50."80 .. .. .--. oo '0.400m;& mY'I.000.00- --- - 3
we 4A SWl -L---- ~ - ~ -- - - - - .
1.001 3 0 or/do 1.003
0.00 i'. -0.00 Z
-1.00 ftd --"1.00 Se
Ai
Figure 46: As in Figure 3, except for closest approach by flost 210 to H4, which time isindicated by vertical line.
5II
52 1
3.15 Case 15: Float 210, CM 15, 21 October 1989, 0000 UTC
3.15.1 Overview
Float 210's second passage through the central array occurred as the float (Figure 47),circulating around a cold eddy (Figure 48), entered the southern edge of the Central Arrayand then turned to the west.
S750 ....... .. ............... ... .... . . . . . . . ..
0 2 4 6 8 1012141618202224262830323436W3404244
I 1550~,A
*41
RAFOS 210£ 4039
0
~37
36
77 7 6 75 74 73 72 71 70 69 68 67 66 6
Longitude (W)Figure 47: As in Figure 44. The three closest RAFOS fixes to CM 15 are boxed. Launch
date: 1326 UTC, 26 September 1989.
3.15.2 Vertical motion results
WRAF = 0.36 rmt s-1 , Wm = 1.07 mm s-1 , wSs = -0.33 mm s-1
As float 210 passed by CM 15, the float upwelled from 571 to 557 to 550 db, (Figure 47)consistent with the float's cyclonically curved path. Indeed, the float was in the middle of athree-day upward excursion of I00 m as it passed 15. CM data also suggest strong upwefling(Figure 49): there is significant veering to the currents and isotherms are rising, as indicated
53
t2I
82 Cso&. 16
-120 M.h j\ - 60 .
-7012 00-00 (60 -0120 -- 4 0 0ONO so40(edecYwg ttl
Fiue 9~ si iue2 xetfr00 T 1Otbr18.Tetako ot20iinicte bydtadtelc4ino0M1 smakdb ice
thtFitue 415 is ino wiuel lo capte for 0000 cmuta 1Otions. 399 h rc o ot20i
inicte b dts ad heloato o C 1 i mrkd y crce
54I
I
I
I3 2 0 . --- - - ------ - -.. . . . --. , ,- -. . -. . . . . . . . . .* ,o
mm 18.0-310. I/ f l o im-
iN'. 12.0290. & Temperl Lto"I280.' .0 -- •- v I-I
1.o0 I m a V1 ,,o.0 -------------- ----------- D1.20 C - . .
0.80 fi -I~-0.00 p-:;1.0.40 f .--so0.00 -V "1 ' ---V
9l IW to 9"111Vno
0.00o
,--'.OO/",F19 1i -I~.0o0 We&, "•-'-.00 Wtw, tt
I tM 3M tIM 3M
3 IFigure 49: As in Figure 3, except for closest approach by float 210 to 15, which time isindicated by vertical line.
Il
•I 55
I
3.16 Case 16: Float 211, CM M13, 8 October 1989, 0000 UTC U3.16.1 Overview
Float 211 (Figure 50) moved on an anticyclonically curved path the entire time it was withinthe Central Array. This was because of a crest/trough structure in the z12 topography(Figure 51) that was directing the main Gulf Stream flow out the southern Central Arrayborder.
450 11550 1~650
SI
750
0 2 4 6 8 10 •2 14 161820 2224 26 28 3032343638DAY$ £FM lUMcH
41 . . v
RAFOS 211 3
39 Ue r.0
*38010
~37.0
363
3477 76 75 74 73 72 71 70 69 68 67 66 65
Longitude (W)3
Figure 50: As in Figure 1. The three closest RAFOS fixes to CM M13 are boxed. Launchdate: 1243 UTC, 30 September 1989.
3.16.2 Vertical motion results IW RAP = -0.44 mm s- , wC• = -0.68 mm s-1 , wits = -0.78 mms-1
During the three closest positions to CM M13, RAFOS float 211 was at pressure levels I613, 659, and 638 db, for a mean downward motion of -0.44 mm s- (Figure 50). CMdata (Figure 52) also suggested downward motion: temperatures at M13 were increasing, 3
58
II
I 8~10 12-40 0 -B,,
3 80 so -8 0
40 40
0 0
-40 -40
-80 -80
-120 -120
--160 o 0 -08•-§E&.12o80-40 0 40 80 -80o- 2oo-40 0 40 80
2 w wnts (advecton)
112012so ~80am
0.
*-40 00 4
-80 -80
-1 1a8-00 40 80 -160-120u80-40 0 40 80Iwmý (tendency) WMu (total)
3 Figure 51: As in Figure 2, except for 0000 UTC 8 October 1989. The track of float 211 isindicated by dots, and the location of CM M13 is marked by a circle.
reflecting downward-moving isotherms, which dominated the vertical motion diagnosed bythe veering currents at M13. IES data also suggests downward motion (Figure 51), as3should be expected given the strong anticyclonic vorticity advection over the site.
* 57
I3I
210 .- 1 .0 - --------
20019.: A ~ 1.
190. low\a 10.0
180. "0."
10 Direct~ojil 6.0 TiegUe170.60 W I
1.60 - - ---- ---- -
1.20: C IM *SM 1.00 Wjy~dt D
0.40 - ! i---0.00 - _,-
0.0• tO
.00 Vic 1a W at•/ 1#6 tta ;
0.00 -m -- .0 YJ 0 --s-
-1.00 -wady "•-io. ; -. .. . - .--. .. .....W . .--. M
II
Figure 52: As in Figure 3, except for closest approach by float 211 to M13, which time isindicated by vertical line.
i
I
i
I58
3.17 Case 17: Float 216, CM I1, 26 November 1989, 0000 UTC
3.17.1 Overview
Float 216 (Figure 53) moved through the Central Array at a time when a cold eddy wasin the southeast corner of the array (Figure 54); however, the float's motion through thearray was evidently unaffected by this cold ring.
I -r -lu,-r "r ur "l.. -l..r ..-.. l.r.-i.i" -.r *--l " " r "ri -i". - r 1 "i" r- "
.550
I I,,*650
h*750
850 I..I.II.I.I..I..I. I . .. I..I .I.I .I..I .. II .I.I...I.I. I.I..I..I..I.II.I I .
0 2 4 6 8 101214161820222426283032343638404244DAYS AT= IAUNCHI 41
I RAFOS 216I-40
39
date: 1034U ,18 Noebr 99
0*
U 37 0
S3635
77 76 75 74 73 72 71 70 69 68 67 66 65Longitude (W)
Figure 53: As in Figure 1. The three closest RAFOS fixes to CMA 11 are boxed. Launchdate: 1034 UTC, 18 November 1989.
3.17.2 Vertical motion results
WRAF = -0.02 mm s- , WCAI = 0.14 mm s-I , wis = 0.86 mm
3 Float 216 passed by CM 11 just before downwelling more than 200 m in a week (Fig-ure 53). Close to I1, however, the float moved from 567 to 576 to 568 db on a path thetrajectory of which was straight. IES data (Figure 54) shows significant curvature to the Z12field, however, with cyclonic vorticity advection forcing upward motion at 11 (in fact, the
59
II
120 120 °
S40 ----O' 0 .
-160 *. . ,, .-." -160
-1V 2..-80-40 0 40 80 -160-120-80-40 0 46 80
Z12 Wis (adveoUon)
0 05
-40 " -40I
-80 ,;- -80
-120 r,,.....--o•-120
-160 -160-160-120-80-40 0 40 80 -160-120-80-40 0 40 80
wln (tendency) wM (total)
Figure 54: As in Figure 2, except for 0000 UTC 26 November 1989. The track of float 216 5is indicated by dots, and the location of CM 11 is marked by a circle.
fields look very similar to those associated with float 201). CM data for the same time alsosuggest a Gulf Stream that is not straight (there is time variability in the direction (Fig-ure 55a)). Upward motion is forced by veering currents, overwhelming the weak downward Imotion associated with the isotherm's descent (Figure 55b,d).
IIII
60 5
1 2 0 ------.-.- -.-- ----. ...80. JA - 5m , 6.0 B
Im
40. 12.0
-40. D
4.0-
C MM 1.00 1.605 ~1.20 Io-a.1.0iA I 100{SpeedL ~~100
0.40 -- .- 1.',00" Wdgdt"0.00 I T T T"" .. ' l I" "
-1.00 EWaq & d" 1.00 W~ I0.00 • •ym.......0i ,f
-1.00 W&&v . v 1 "•.00 Wtota
Figure 55: As in Figure 3, except for closest approach by float 216 to I1, which time isindicated by vertical line.
61
I
3.18 Case 18: Float 221, CM 12, 6 January 1990, 0800 UTC I
3.18.1 Overview 3Float 221 (Figure 56) moved quickly through the Central Array on a straight path. z12
data (Figure 57) also shows a mostly straight Gulf Stream path (in which are embeddedminor troughs/crests) with a cold eddy far to the south of the stream.
. . .420 i
50
0 2 7 747 72 41U70 6 a 10 12
41 DY P AM
RAFOS 221b d
390
35
3477 76 75 74 7372 71 70 69 68 67 66 65
Longitude (W)
Figure 56: As in Figure 1. The three closest RAFOS fixes to CM 12 are boxed. Launchdate: 0253 UTC, 30 December 1989. 3
3.18.2 Vertical motion results 3WRAP = 0.18 mm s-1 , wCM = 0.21 mm s-1 , wtSs = 0.45 mm s-1
Float 221 moved from 490 to 483 to 480 db in the time surrounding its closest passage to 3CM [2 (Figure 56). [ES data suggest this upward ascent is forced both by cyclonic vorticityadvection and by local vorticity tendencies (Figure 57). CM data (Figure 58) also reflectthe upward motions, as there is significant veering motion and slight cooling. U
623
12 120180 s0
7000
*-40 -40
112 -120
-160 -160-5Ift -80 -40 0 40 80 -160-120-80 -40 0 40 80
Z12 wa (advection)
1120 120
40 80
-40 -40
-120 -- 120
-160 - 160-160-120-80 -40 0 4.0 80 - 1SO- 120-80 -40 0 40 8
wM (tendency) WM (total)
Figure 57: As in Figure 2, except for 0800 UTC 6 January 1990. The track of float 221 is
indicated by dots, and the location of CM 12 is marked by a circle.
63
III
7 Ma12.01
1.20 : Spe : -1. 0- ,•0.80- f,." •,.--Q.00 - • .
0.40:- ; ... • f' -- "-1.oo 'J .:. I
8.0 Te p. tu. Oli
0.80 "� • --------- •------ --"_ ---o. 4opo4..ins .oo•w- -- j-i I
.0 /t1.00 F*
-1.00 Wad- ams. T' -==1.00 WtoI
. .. I '1 " "e1 'e . .. ..1- ' I ' ' '
UI
Figure 58: As in Figure 3, except for closest approach by float 221 to 12, which time is iindicated by vertical line.
III[
1 3.19 Case 19: Float 224, CM H4, 20 January 1990, 0800 UTC
3.19.1 Overview
Float 224 (Figure 59) passes through the Central Array at a time when a cold eddy isapparently interacting with the Gulf Stream (Figure 60). This interaction, however, does3 not seem to affect the float.
~450
S1..550k650
I0 2 4 6 8 10 12 14 16 I8 20 22IRAFOS 224
*40
39 00
38 -s...
76 0 57 37 17 L 68 67 66 65
IFigure 59: As in Figure 1. The three closest RAFOS fixes to CM H4 are boxed. Launchdate: 1200 UTC, 15 January 1990.
3.19.2 Vertical motion results
WRAP = -0.18 mm s-1 WC = -0.10 mm 5-1 W,,S = -0.30 mms-
As it passed close to CM H4, float 224 moved from 477 to 497 to 487 db. The pressuretrace (Figure .59) suggests that the float downwelled until 0800 UTC 20 January 1990 (whenit was closest to H4), and then it leveled off, or upwelled slightly. IES data suggests this3 may also be true; although downwelling is diagnosed in the vicinity of H4, upwelling is
65
120 120
s0 80 140
-40 - -40 '
-80 - -0
-120 -120
-160 -160l-19SI29-M0-40 0 40 80 -160-120-80-40 0 40 80
Z12 wm (advecuon)
120 120
60 60
0I-40 -40
-120 -120 1-160 -6
-160-120-80 -40 0 40 80 -16G-120-80-40 0 40 60wM (tendency) WM (total)U
Figure 60: As in Figure 2, except for 0800 UTC 20 January 1990. The track of float 224 isindicated by dots, and the location of CM H4 is marked by a circle.
occurring just downstream, both forced primarily by vorticity advection (Figure 60). CMdata also suggest downwelling; veering currents are forcing upward motion (Figure 61a,ce)that is offset by the greater downward motion associated with local warming/descending Iisotherm (Figure 61b,d,f). This is another example where the temperature tendency at 700m is much stronger than the average of the 400 m and 1000 m tendencies. The 1000 m and400 m sensors are respectively below and above the main thermocline. I
III
68!
I
1 680. ----------- iso70.-- 14.0-
60. 60.040 10.0
5Hi 1 ~ an~ur1.60- ----------
S8peed C - *Ya 1.00D
0.40: f --1.00 wea&
00 - -- 1 --______- .
.004 I s -U Wdo 1.00 I Ft1 ~0.00 fm-I', -- 0.00*
-1.00 -~Wad, JU 1.1 0 0 wft
Figure 61: As in Figure 3, except for closest approach by float 224 to H4, which time is
indicated by vertical line.
67
I
4 Comparing the three different methods I4.1 Statistics 3What are the general statistics of vertical motion in the Central Array from the three datasources? Table I shows the means, standard deviations, maxima, minima, and the numberof observations for the two-year period for each data source. All data sources indicate meandownwefling, which is consistent with expectations from quasi-geostrophic motion and thetime average of equation (7) given the the observed crest just upstream of the Central Array(nea the Gulf Stream 'node' at 70* W) and a trough near the eastern boundary of the array Iin the mean. The larger mean value of wa,, may reflect a sampling bias: floats remained
mostly within the strongly baroclinic structure of the central Gulf Stream when traversingthe Central Array, whereas CM and IES measured vertical motions throughout the region Ispanning also the Slope and Sargasso waters where large-scale vertical motions are smaller.Note that WRAP downwelling averaged over all tracks, including those that escaped theGulf Stream, is an order of magnitude smaller than within the Central Array. We are Uexamining a segment of the Gulf Stream that, for these 26 months at least, experiencedmean downwelling. 5"Table 1: Vertical Motion Statistics: Mean, maximum, and minimum vertical motion (inmm s-1) with standard deviation and number of (not all independent) observations forCM, IES, IAFOS floats in the Central Array only, and for all RAFOS floats
Source Mean Std. Deviation Maximum Minimum # obsCurrent Meters -0.015 0.347 3.676 -2.297 25419
Mapped IES -0.027 0.440 4.006 -3.794 188640RAFOS in central array -0.048 0.617 2.10 -2.66 947
All RAFOS -0.006 0.532 3.30 -3.34 5939 1
To test the statistical significance of mean dowuwelling in Table 1, we computed theuncertainty of the mean for timeseries of wC. at each CM site, and at the ten extrema inthe mean w,,, field in Figure 62 using techniques in Bendat and Piersol (1973) and Dewarand Bane (1985). For each of the ten W... extrema, the mean value differed from zero byat least one standard deviation (and usually two). Mean values of wcm also were usually Iat least one standard deviation from zero, except for several sites where 10 cm was verynearly zero. The mean downwelling is associated with the fact that the IES instrumentspreferentially sampled a crest-to-trough segment of the Gulf Stream. It follows from the Uwork of Bower (1989) that downwelling should occur. How representative of the long-termmean are our results? Lee (1993) shows a trough in the Central Array for 6 out 8 yearsusing infrared satellite data. We do believe the values for mean vertical motion in Table I Ihave some statistical significance.
4.2 Case Studies IWe now present several Figures that give evidence that the vertical motion determined fromall three data sources is consistent. Figure 63 shows two time-series of Wcu and wvs, at Itwo unrelated sites and times, G2 and II. There are clear similarities between wCM andw,,,: both show events of up- and downwelling that develop and pass coherently with somehigh-frequency noise, extrema are approximately colocated, signal strength is similar in the I
68
a, e.asavsi.m 120
I m _-_2 s-g ..... *
..4.US.151~4•M -120
1 -680
-160 -120 -80 -40 0 40 80
IFigure 62: Mean Z12 topography (bold lines, contoured every 100 m) and mean verticalmotion from (7) (thin lines contoured every 0.05 mm s-1, negative values dashed) for theperiod 15 June 1988 - 7 August 1990 within the Central Array. Point values (:k standarddeviation) of extrema as indicated. z- and y- axis labels are in kilometers.
two lines, and the episodic nature is obvious. We have made a similar comparison for all CMsites in the two-year deployment period; the similarity evident in Figure 63 characterizes allsites: large vertical motion in one data source usually has a corresponding extremum in theother; however, some counterexamples will be shown. Examples of the tw records along the'G' line of current meters (G2 and G3 for both Year 1 and Year 2), along the 'H' line (H2,H3, H4, and H6 for both Year 1 and Year 2, H5 for only Year 1), along the 'I' line (I1, 12,13, 14, 15 for both Year 1 and Year 2), and at M13 (Year 2) are shown in Figures 64- 76.In each of these time series, the characteristics highlighted in Figure 63 are evident. In eachtime series, a large peak in vertical motion in one method generally has a correspondingpeak in the other method, although some time series (e.g., Figure 69) are better than others3(Figure 75). There is inter-time series variability as well. Compare, for instance, Year 1in Figure 68, which shows little coherence, with Year 2, which shows considerably more.Despite occasional short-period discrepancies between methods, we show below that time
I series of wcU and wt,, are coherent at periods longer than 12 days.
69
Ii
I
N0.0
~ 2.0 6 E_________________A _____
15 Feb 1 r I 6t 9 16 Apr0 9 6 May 9
2.o.-:1110.0-
21 Oct 80 23 No 09 27 Dc89 29 Jan 90
IFigure 63: Time series of vertical motion. (A) CM G2 and (B) CM I1 with dats as indi-cated on the z-axis. Vertical motion from CM data (wC,) using equation (4) in indicatedwith the dotted line; that from IES data (w,,,) and equation (7) is indicated with the solidline. I
I
1.0-
0.0i
-1.0
-2.0 I15 Jun6 5 DeO 67 2mayas i ftb90 6 A" 90
SO Aug IS
Figure 64: Time series of w,,, (solid line) and wcM (dashed line) for CM G2 for Year 1and Year 2. 3
70I
2.0- cUm....
Ui
1.0
* 0.0
-1.0
* -2.0
ishas a SD" aU 7 K" yso 16 Feb 90 Avg 903 UAugN
Figure 65: As in Figure 64, but for G3.
2.0- cm....
S1.00.0- jFd¼A -1A
-1.0- ' -
3 Figure 66: As in Figure 64, but for 112.
-2.0 -
1.0O
I
:
3 -2.0
15 Jun 9 6 D"o a 2 1" 09 16 Feb 90 6 Aug0
39Aug N
Figure 67: As in Figure 64, but for H23.
I.1
* 71
FIIn
-2.0
J•s in 6 I5 Do 2?ayso 16 Feb 90 6 Ago29 Aug 89U
Figure 68: As in Figure 64, but for 114.
~II
2.0- cm .... _______ _____ __ I1.0 -_ _ _ _ _
-1.0-
I I1i Jmu a 5 Doe as rMay6p 16 90 1 Aug 90
Figure 69: As in Figure 64, but for H5, Year I only. 3
1.0 I0.0 -
-1.0 , -... :1 .
-2.0- oA-
15 Jun as 6 Doc 1 a? May 1o is m9o 6AuM N
Figure 70: As in Figure 64, but for H6. 9 Aug W9
U
ii
2.0- cm....
1.0- .A- -. ,I. -A !-1.0- 4 o-2.0 OA--
16 Junas 5 D N 27 May 89 16 Feb 90 6A ug90
29 Aug 9
Figure 71: As in Figure 64, but for I1.
2.0- Cm....
11.0-0.0 - -ab
* -1.0,
-2.0 OA-
15 Jun 8 5e6 N 27 MayN 16 Feb 90 6 Aug9029 Aug 09
Figure 72: As in Figure 64, but for 12.
2.0- Cm .... _
1.0-
0.0 -I,-1.0-
15 Jun6a 5 Deo6 18 7 Iy7 89 16 Feb 90 6 Aug 90N9 Aug 8g
Figure 73: As in Figure 64, but for 13.
73
ii
2.0 " -r-
-1.01
-2.0- oA-
is um as 5 DeW 27 my a 19 Feb 9 SaAm9
Figure 74: As in Figure 64, but for 14. 2Aug N
I2.0- -- _i
1.0 - -1AO
0.0-
-2.0-
is JuSn 6 De 86 27"may so 16 Febi0 6 AUgN
29 Aug 80
Figure 75: As in Figure 64, but for 15.
I
1.0 c --1.0- J 0 4A . ... I- A
0.0 --1.0-2.0 - OA , rI7
29 Aug 8 16 Feb 60 Aug6 90
Figure 76: As in Figure 64, but for M13, Year 2 only. UI
Table 2 presents comparisons of vertical motion for all 23 cases in which a float passedwithin 10 km of a CM mooring with working instrumentation at 400, 700, and 1000 m.These are the cases discussed in Section 3. Figure 77a is a scatterplot of RAFOS verticalmotions versus CM vertical motions, with data taken from Table 2. A linear relationshipis clearly evident with correlation r 2 = 0.82. That the slope of the line is greater thanunity could arise from our choices of values for ao, or !, , or from WRAF being biased low.However, the plot underscores the excellent agreement between wtAP and wcm.
Table 2: Vertical Motion Comparisons: Vertical motions from RAFOS float, CM, and IESdata (wRAF, wcAI, and wES, respectively), with units of mm s-I for date/time shown andfor RAFOS float and CM indicated. Depth of the RAFOS float at the comparison time (inmeters) is also shown.
Float CM date/time to P WM W Es DepthRAF123 H4 4Jul88/00z -0.58 -0.98 -1.42 550.3 mRAF129 14 23Sep88/08z 0.19 0.53 0.04 710.7 mRAF129 14 23Se,,8/16z 0.72 0.48 0.05 706.4 mRAF136 14 8Dec88/16z 0.81 0.43 0.17 668.4 mRAF141 14 2Nov88/16z -0.87 -0.94 -0.43 603.0 mRAF175 14 20Jan89/16z -0.56 -0.05 -0.63 363.7 mRAF176 15 7Feb89/00z -0.04 0.02 0.25 761.5 mRAF176 15 7dFeb89/08z -0.05 0.03 0.26 763.0 mRAF176 15 7Feb89/16z 0.04 0.02 0.27 764.5 mn
RAF176 T5 8Feb89/00z -0.03 -0.0.5 0.26 760.7 mRAF194 15 15Nov89/00z 0.71 0.23 -0.22 662.2 mRAF194 14 16Nov89/00z 0.86 1.20 -1.06 584.3 mRAF199 15 23Nov89/00z 0.24 0.23 -0.13 593.2 mRAF199 15 28Nov89/16z -0.13 0.04 0.54 627.4 mRAF201 11 23Nov89/08z 0.97 1.51 1.66 560.4 mRAF207 12 4Sep89/08z -0.25 -0.06 -0.24 164.0 m_RAF209 15 13Sep89/08z -1.53 -2.14 0.22 629.2 mRAF210 114 60ct89/08z -0.02 -0.19 -0.84 642.4 mRAF210 15 210ct89/00z 0.36 1.07 -0.33 557.4 m_RAF211 M13 80ct89/00z -0.44 -0.68 -0.78 658.9 mAAF216 II 26Nov89/00z -0.02 0.14 0.86 576.2 mRAF221 12 6Jan9O/08z 0.18 0.21 0.45 482.9 mRAF224 H4 20Jan9O/08z -0.18 -0.10 -0.30 496.7 m
The comparison between wRAF and to IES for the same 23 cases (Figure 77b) showsmore scatter, in part because the cases were selected for floats passing CM sites; severalof these cases are not in the best IES mapping region. Nevertheless, after excluding thetwo obvious outliers that are discussed below, the correlation r 2 = 0.32, which is significantat a confidence level of greater than 99%. That the slope of the line is less than unitysuggests there may be a bias towards underestimating WEs that would result from thespatial filtering of the IES maps.
Careful scrutiny of Table 2 shows that tow•ES consistently is of the wrong sign at I5.The computation of vorticity and vorticity advection at site I5 is adversely impacted by
75
II
that site's close proximity to edge of the domain. After excluding all Wz, from site I5 inTable 2, we plot the remaining values in Figure 77c, from which we compute a correlation r 2
of 0.51, significant at a confidence level exceeding 90%. To understand the causes of errorsin Figures 77b,c, we have closely investigated the vertical motion fields associated with theone remaining outlier in Figure 77c. Two possible error sources are discussed below.
4 30 C" 3.0 4ý'3:0 C &&L~J2.0 B + C +E 1.0 +oE 1.0.-.0 +0.0 A0.0 ' 00
C -1.0 1.0 + 1.0 +S-."3•,. -3.0 0.•"qt0 -1.5 -0 ':~ y irr
+ -2.0 -2.L0" 9--3.0o-. 0. -3.0 3:-&-1.5-I0.0 1.5 3.0 -3.-1* -'.30.0 1.5 3.0 --3.0--1.10.0 13'~ &) •
Wpr (10-3 m '-1) WIar (10- m ) W (10-3 m e)
IFigure 77: (A) Scatterplot of wR4F versus UCM for data as listed in Table 2. (B) Scatterplotof WRAF versus w,,S as listed in Table 2. (C) As in (B), but excluding wles at site 15. The
box point in (B) and (C) is discussed further in text.
The point in the lower right quadrant of Figure 77b,c is from the case when RAFOSfloat 194 passed by CM 14 at 0000 UTC 16 November 1989 (See Section 3.8). Both CMand float vertical velocities are upward; however, w,,, is strongly downward. This may becaused by poor IES data at site 14 during winter of Year 2, when this IES suffered frommany bad data returns, necessitating frequent interpolations to compute thermocline depthat that site (Fields and Watts 1991). Such data degradation would of course have a negativeimpact on the computed streamfunction field and all quasigeostrophic calculations near 14for winter of 1989-1990 data.
It is far more likely that errors were introduced into w,,, at this time because significantcurrents existed at deep levels. IES measurements determine the barodinic component ofthe flow only. IF there is a significant barotropic flow, then the vorticities and advectionscomputed in (7) using only IES data will not be representative of the trvp flow. Figure 27presents the IES topography and w,,, for this time, as well as the track of float 209 as ittraversed the central array. The deep currents at the CM sites are also plotted. Note theunusually strong flow at 3500 m at 14. In this area, strong upward motion associated withthe cyclonically curved onshore barotropic flow overwhelms the downward motion associatedwith the baroclinic flow. In fact, this event is associated with the strongest deep currentsfor the WJEs cases plotted in Figure 77c: It35oo[1 = 0.19m s-, which is a value nearly twostandard deviations above the mean speed at 14 during Year 2. Note that strong deepcurrents do not necessarily mean poor values of w,,,: Case 11 (Section 3.11) also had deepvelocities in excess of 0.15 m s-1 and there is good agreement between wRA, and wos,. Forthis case, the deep currents flow parallel to the Gulf Stream; for the case above, the deepflow crosses under the Gulf Stream. Where the barotropic flow is significant, especially if itis spatially non-uniform, errors in wIEs are quite likely. Because it is possible to account forthe deep currents, however, (for example by combining a deep barotropic streamfunction -
76
I
measured by current meters or perhaps by electric field measurements - with the baroclinicstreamfunction) we do not consider this an unbearable problem. Indeed, this is the subjectof an ongoing followup study.
4.3 Coherence between wCM and w,,,
The time-series estimates of wCM and wtS (e.g., Figures 63 and 64- 76) show smoothlyvarying signals with varying degrees of high-frequency noise superimposed. We spectrallydecomposed each pair of records and computed their statistical coherence at all sites alongthe central mooring H-line (whose time series are shown in Figures 66- 70) at those CMsites that had all three upper CMs functioning and that were surrounded by functioningIESs (these were H3, H5, and H6 from Year I and H4 and H6 from Year 2). These coherenceresults were averaged for the wholt line and are shown in Figure 78. The average coherencebetween the two independent estimates of vertical motion is greater than 0.5 (at far abovethe 99% confidence level) throughout all periods longer than 16 days. At periods shorterthan 12 days, the coherence is insignificant. This confirms the visual impression fromFigures 63- 76 that the large amplitude, low-frequency variations of w from both CM andIES techniques are very consistent, while the smaller-amplitude, high-frequency variationsare noise-dominated. In Figures 79- 91, we show the coherence for all of the sites, calculatedseparately for Year 1 and Year 2 because the instrument arrays changed between the twoyears. In particular, the mapping ability of the IES array to estimate terms in the vorticityequation could have changed between the two years - as it appears to have done at somesites like G2 (Figure 79) and H2 (Figuire 81).I
0.5
* 0.4 Confidence Levof_99Z
. -onfidence Lovel=95X
l 0.1
32 16 8 4Period (days)
Figure 78: Coherence as a function of frequency for w,,, and wCM for selected sites on the'H' line.
Given these statistical coherences, how well can we use equations (4) and (7) to infervertical motion using data from the CM and IES arrays that were designed for mesoscale
I 77
0.22S0.0
-. I32. 18 a
days days dfr 4&y
Figure 79: Coherence as a function of frequency for wigs and wCM for CM G2 Year 1 (solidline) and Year 2 (dotted line).
0 o.4s I
0.0_=I ° -' ,0 , - .,days days dadm days
Figure 80: As in Figure 79, but for G3.
sampling? Both techniques involve temporal and spatial differentiation of measured quan-tities, an analysis which accentuates measurement and sampling errors. The errors in wCMfrom (4) are not dominated by measurement error but rather by submesoscale features forwhich the signal is incoherent between our moored u, v, T sensors. Errors in woi,, from(7) are dominated by instrument/analysis error in the streamfunction fields. The objectivemapping and measurement technique filters out submesoscale features, including some thatmay be sampled at individual CM moorings. Nevertheless, with either technique, the majormesoscale features in w are dear. The coherence between woigs and w.tH shown in Figure10 and the particularly good agreement between WRAF and wcM illustrated in Table '2 andFigure 77a, i.e. agreements between three independent techniques for measuring w,, showthat the toCu(t) and wxs,(z, y,t) fields do characterize mesoscale vertical motions in theGulf Stream.
I
S0.40.0
32 16 aI
daysda" TWO6days
Figure 81: As in Figure 79, but for H2.
78
0.8
S0.4 sB ... o-------
0.2 e-0.0 :
day dys d'$vurwa ayI Figure 82: As in Figure 79, but for 113.
1.0 •
o.aI *.6
o.4
0.20.0 "-
5days days daeiddaysFigure 83: As in Figure 79, but for H4.
1.O
0.6
8 0.4
0.2
32 16 a 4days days a..•od days
Figure 84: As in Figure 79, but for H5 (Year 1 only).
0.0/
32 16 a 4days days- -eiod days
Figure 85: As in Figure 79, but for H6.
79
1.0*
0a.60.4S
0.2
0.0
32 16 a 4days days ":eiddy
Figure 86: As in Figure 79, but for 11.
1.0
days days d~
Figure 87: As in Figure 79, but for 12.
10.8J0.60.4
- . '-
- --S- - - - - -
days days des daysFigure 88: As in Figure 79, but for 13.
0.2 \ * * *
days days de¶ro days
Figure 89: As in Figure 79, but for 14.
80
IIIIII
0. I0o.4
days days 'I:bVeiod days
Figure 90: As in Figure 79, but for 15.
IIIIIi .0 " I I , I I I I. I I I
0.5 8
I ~0.4O.
80.2
0.0 1 I I I32 16 a4days days-daeo days
Figure 91: As in Figure 79, but for M13.
III
i
5 Summary IWe have computed and compared vertical motions from three different sources in the GulfStream downstream of Cape Hatteras. Magnitudes of vertical motion as measured byRAFOS floats and as estimated from CM data and IES data are 1-2 mm s-1, with rarevalues of up to 3 mm sa-. Both the CM and IES vertical motions are smoothly varyingfields not dominated by the submesoscale and measurement noise that is present. It is easy Ito track features from day to day in both datasets; evolution of the fields occurs only slowlyas mesoscale jet features evolve.
CM vertical motions agree closely with observations from RAFOS floats. There is also Iconsiderable agreement between IES and RAFOS float or CM vertical motions, and verygood coherence at long timescales between wiu$ and w¢• timeseries. Indeed, for timescalesof greater than 16d, wcu and wigs show coherence > 0.5 at far above the 99% confidence Ilevel, but such coherence is absent for timescales of less than 12d. This lack of coherence maybe caused in part by errors related to the presence of strong deep (barotropic) velocitiesunder the meandering Gulf Stream. By combining the IES data with an independent Imeasure of the barotropic velocity, it should be possible to account for the vertical velocitydriven by deep currents.
To our knowledge, this is the first study in which vertical motions associated with Imesoscale processes in the ocean have been observed and verified by independent measure-ments and independent dynamical methods. The spatial and temporal structures of the wfield can be diagnosed with both CM and IES arrays; the Gulf Stream clearly exhibits sec- Iondary circulations on the mesoscale that are consistent with quasi-geostrophic dynamics. I
Acknowledgements
The SYNOP Experiment was supported by the Office of Naval Research under contractnumbers N00014-90J- 1568 and N00014-92J-4013 and the National Science Foundation undergrant number OCE87-17144.We thank Meghan Cronin at URI for her considerable work in applying the mooring motion Icorrection algorithm to the CM data. This study benefitted greatly from fruitful discussionswith Professor Tom Rossby (at URI) on RAFOS float technology, with Karen Tracey at
URI on objective mapping of IES data, and with Dr. Melinda Hall at Woods Hole on Icurrent meter estimates of vertical motion.
III
I
I
6 References
Anderson-Fontana, S. and H. T. Rossby, 1991: RAFOSfloats in the SYNOP ezperiment 1988-1990. Technical Report 91-7, University of Rhode Island Graduate School of Oceanography,Narragansett, RI 02882-1197. 155pp.
Arnason, G., 1942: Distribution of mass variations in atmospheric air columns. MeteorologiskeAnnaler, 1, 10, pp. 255-279.
Bane, J. M., L. M. O'Keefe, and D. R. Watts, 1989: Mesoscale eddies and submesoscalecoherent vortices: Their existence near and interactions with the Gulf Stream. MesoscalelSynoptic Coherent Structures in Geophysical Turbulence, J. C. J. Nihoul and B. M. Jamart,eds. Elsevier Science Publishers, Amsterdam. pp. 501-518.
Blanton, S. L., I11, 1991: Computations of vertical velocity in the Gulf Stream northeast ofCape Hatteras, North Carolina. M.S. Thesis, University of North Carolina Marine SciencesProgram, Chapel Hill, NC, 27599-3300. 84pp.
Br den, H. L., 1976: Horizontal advection of temperature for low-frequency motions. DeepSela Res., 23, 1165-1174.
Bryden, H. L., 1980: Geostrophic vorticity balance in midocean. J. Geophys. Res., 85C,2825-2828.
Bower, A. S., 1989: Potential vorticity balances and horizontal divergence along particletrajectories in Gulf Stream meanders east of Cape Hatteras. J. Plys. Oceanogr., 19,
I 1669-1681.Bower, A. S., 1991: A simple kinematic mechanism for mixing fluid parcels across a mean-
dering jet. J. Phys. Oceanogr., 21, 173-180.Bower, A. S. and H. T. Rossby, 1989: Evidence of cross-frontal exchange processes in the Gulf
Stream based on isopycnal RAFOS float data. J. Phys. Oceanogr., 19, 1177-1190.
Fields, E. and D. R. Watts, 1990. THE SYNOP EXPERIMENT: Inverted Echo Sounder datareport for May 1988 to Aug 1989. GSO Technical Report No. 90-2, Graduate School ofOceanography, University of Rhode Island, Narragansett, RI 02882-1197. 231 pp.
Fields, E. and D. R. Watts, 1991. THE SYNOP EXPERIMENT: Inverted Echo Sounderdata report for Jun 1989 to Sep 1990. GSO Technical Report No. 91-2, Graduate Schoolof Oceanography, University of Rhode Island, Narragansett, RI 02882-1197. 255pp.
Gill, A. E., 1982: Atmosphere-Ocean Dynamics. Academic Press. 662pp.
Hall, M. M., 1985: Horizontal and vertical structure of velocity, potential vorticity and energyin the Gulf Stream. Ph.D. thesis, MIT/WHOI, WIIOI-85-16, Woods Hole, MA, 02543,165pp.
Hall, M. M., 1986: Horizontal and vertical structure of the Gulf Stream velocity field at 68W. J. Phlys. Oceanogr., 16, 1814-1828.
3 Hall, M. M., 1989: Energetics of the Kuroshio extension at 350 N, 152? E. J. Phys. Oceanogr.,21, 958-975.
Hogg,.N.G., 1991: Mooring motion corrections revisited. J. Atmos. Ocean. Technol., 8,
Holton., J. R., 1979: An Introduction to Dynamic Meteorology. Academic Press, 391pp.
Kim, H.-S., 1991. An observational streamfunction in the Gulf Stream. M.S. thesis, Univ. ofRhode Island Graduate School of Oceanography, Narragansett, RI, 02882-1197, 126pp.
Leach, H., 1987: The diagnosis of synoptic-scale vertical motion in the seasonal thermocline.Deep-Sea Res., 34, 2005-2017.
* 83
I
Lee, T., 1993: Variability of the Gulf Stream path from Cape Hatteras to 45 west observed from Iinfrared imagery. Ph.D. thesis, Univ. of Rhode Island Graduate School of Oceanography,Narragansett, RI, 02882-1197, 95pp.
Osgood, K. E., J. M. Bane, Jr., and W. K. Dewar, 1987: Vertical velocities and dynamicalbalances in Gulf Stream meanders. J. Geophys. Res., 92C, 13029-13040.
Panofsky, H. A., 1944: The effect of vertical motion on local temperature and pressure ten-dencies. Bull. Amer. Meteor. Soc., 25, 271-275. I
Pickart, R. S. and D. R. Watts, 1990: Using the inverted echo sounder to measure verticalprofiles of Gulf Stream temperature and geostrophic velocity. J. Atmos. Ocean. Tech., 7,146-156.
Pollard, R. T. and L. A. Regier, 1992: Vorticity and vertical circulation at an ocean front. J.Phys. Oceanogr., 22, 609-625.
Rossby, H. T., 1987: On the energetics of the Gulf Stream at 73W. J. Marine Res., 45, 59-82.
Rossby, H. T., D. Dorson and J. Fontaine, 1986. The RAFOS system. J. Atmos. OceanicTechnol., 3, 672-679.
Shapiro, R., 1970: Smoothing, filtering and boundary effects. Rev. Geophys. Space Phys., 8,359-387.
Shay, T. J., S. Haines, J. M. Bane and D. R. Watts, 1992: SYNOP Central Array currentmeter data report: Mooring period May 1988-September 1990. University of North CarolinaTechnical Report.
Tintori, J., D. Gomis, S. Alonso and G. Parrilla, 1991: Mesoscale dynamics and verticalmotion in the Alborn Sea. J. Phys. Oceanogr, 21, 811-823.
Tracey, K. L. and D. R. Watts, 1991: THE SYNOP EXPERIMENT: Thermocline depthmaps for the Central Array October 1987 to August 1990. GSO Technical Report No. 91-5,Graduate School of Oceanography, University of Rhode Island, Narragansett, Rhode Island,02882-1197. 193pp.
Watts, D. R. and H. T. Rossby, 1977: Measuring dynamic heights with inverted echo sounders:Results froth MODE. J. Phys. Oceanogr., 7, 346-358.
IIIIiI
I
I
I fECLRIrY CLASSIFICATION OF rimisR-AG
* REPORT OOCUMENTATION PAGEl a. REOTSCRT LSIIAINlb. RESTRICTIVE MARKINGS
Za. SECURITY CLASSIFICATION AUTHORITY 3. OISTRIBUTiON/IAVAILASILITY OF REPORTI Distribution for Public Release-Zb. OEC-.ASSIFICATION/IDOWNGRAOING SCHEDULE Distribution is unlimited
I 4. PERFORMING ORGANIZATION REPORT NUMBER(S) S. MONITORING ORGANIZATION REPORT NUMBER(S)Univ. or RI, Graduate School of Oceanography,GSO TechnicalReport_93-2 _______________________________
m 6a. NAME OF PERFORMING ORGANIZATION 6b. OFFICE SYMBOL 7a. NAME OF MONITORING ORGANIZATION* Univ. of RI (If 80011cable)
- Graduate School of Oenga 1122 P0I 6r- ADDRESS (City. $tat*, and ZIP Code) 7b. ADDRESS (City, Staft, and ZIP Codle)
South Ferry RoadNarragansett, RI '02882-1197
I So. NAME OF FUNDING iSPONSORING 8b. OFFICE SYMBOL 9. PROCUREMENT INSTRUMENT IDENTIFICATION NUMBERORGANIZATION Of fice Naval Resea ch('y 1 '~bk
NatinalScinceFoundation I_____________________________
m Sc. ADDRESS (City. State. and ZIP Code) 10. SOURCE OF FUNDING NUMBERS* 100 N.Quincy St., Arlington, VA 22217 PROGRAM IPROJECT ITASK -WORK UNIT- 1800 G St., N.W., Washington, DC 20550 ELEMENT NO. 1NO. jNO. JACCESSION NO.
I1i. TITLE (Includo Security Classification)
Vertical Motion in the SYNOP Central Array
I12. PERSONAL AUTHOR(S) Scott S. Lindstrom & D. Randolph Watts
113a. TYPE OF REPORT 13b. TIME %%VEjE 114. DATE OF REPORT (Year. Moneft Day) IiS. PAGE COUNTI Technical IPROM ____To 8L/90 ~Oct. 93 I 8416. SUPPLEMENTARY NOTATION
I 17. COSATI CODES 18I. SUBJECT TERMS (Continue on 'everse if necessary ard ident f by blode number)FIELD -GROUP SU"GROupII Gulf Stream, SYNOP, Vertical Motion
I 1 ABsTRC (Corte of the SeifP avyendpi Oeani Predictionbr experiment) field program, twelve tallImoorings measured the Gulf Stream's temper~ature and velocity fields with current meters (CM)at nominal depths of 400 mi, 700 m, 1000 m, and 3500 m for two years, from May 1988 throughAugust 1990. Simultaneously, 24 inverted echo sounders (IES) monitored the thermoclineI ography. A third observational component of the experiment was the release of isopycnalRAOS floats; 70 such floats traversed the area monitored by the CM and the IES. This report
dtocuments the methods used to compute vertical motion for each data source, and thedifferences and similarities between the three methods. Typical velocities during 'strong'Ievents, as observed by or inferred from all three instruments, was 1-2 mm :-I in regions nearthe center of the Gulf Stream. The comparison of RAFOS vertical motions and vertical motionsdiagnosed from CM data showed excellent agreement; furthermore, CM vertical motions and IESIvertical motions are statistically coherent for periods longer than 12 days. We concludethat we may man me oscaje fields of w(x,y,t); the fields mapped are consistent with
juas~geotrinfil d~gjc520. DISTRIBUTION /AVAILABILITY OF ABSTRACT 21. ABSTRACT SECURITY CLASSIFICATION
CUNCLASSIFIEOIUNUMITEO CO SAME AS RPT. OrC o USERS22a. 'JAME OF IIESPONSIALE NOIVIOUAL 22b. TELEPI4ONE (Include Area Codu) I- 2c. OFFICE SYMBOL
00FR 47.4MR33 AP'R eaition May o0 uSk until exnausteo. SECURITY CLASSIFicAr;ON OF 7HiS 2&GiI 00FOR 147.a4MARAll other editions are obsolete.
