Abstract

Twenty-one RAFOS floats were tracked at depths of 200-1000 meters in and around several North Brazil Current Rings between November 1998 and June 2000. This was part of an experiment to study the role of these current rings in transporting upper level South Atlantic water across the equatorial-tropical gyre boundary into the North Atlantic subtropical gyre. The float trajectories in combination with surface drifters and satellite imagery reveal the sometimes complex life histories of several rings and their fate as they collide with the Lesser Antilles Islands. This report describes the float trajectories, the velocity, temperature, and depth time series, and a preliminary analysis of the float data.

1. Introduction

North Brazil Current (NBC) rings are large (400 km diameter) anticyclones that pinch off from the NBC retroflection in the western tropical Atlantic near 8N and translate northwestward along the coast of South America toward the Caribbean (Johns et al., 1990; Didden and Schott, 1993; Richardson et al., 1994; Fratantoni et al., 1995). NBC rings have been proposed as one of several important mechanisms for the transport of South Atlantic upper-ocean water across the equatorial-tropical gyre boundary and into the North Atlantic subtropical gyre. Such transport is required to complete the meridional overturning cell in the Atlantic forced by the high-latitude production and southward export of North Atlantic Deep Water. The mechanisms that contribute to NBC ring formation and the structure and dynamics of the rings themselves are not well understood. The purpose of this study is to obtain, for the first time, comprehensive observations of the NBC retroflection, the NBC ring formation process, and the physical structure and properties of NBC rings as they translate northwestward along the low-latitude western boundary. The goal is to understand the process of NBC ring generation and to quantify the role of NBC rings in cross-equatorial and cross-gyre transport within the Atlantic meridional overturning cell. Specific objectives of the float component of the NBC Rings Experiment are to:

Measure and describe the physical structure of NBC rings after separation from the NBC, including both near-synoptic and time-evolving temperature and velocity characteristics.
Determine characteristic rates of translation, mixing, and decay as NBC rings move northwestward toward the Caribbean Sea.
Identify the long-term fate of South Atlantic water trapped within the ring core, and determine the effective intergyre ring transport in several temperature/density watermass classes.

2. Float Deployment

Four cruises were conducted during the North Brazil Current Rings Experiment. On each cruise, shipboard ADCP, XBT, and CTD-LADCP surveys were used to locate NBC rings and to measure their physical properties (Fleurant et al., 2000a, b, c). Subsurface floats were deployed during the first three cruises (see Table 1) and during a non-NBC cruise following cruise 2. Two acoustic sound sources (S1 and S2) were deployed on the first cruise and retrieved on the fourth. An ALFOS float was also launched on the first cruise in order to monitor the sources.

Table 1. Number of RAFOS floats launched during each cruise.

Cruise	Beginning Date	Ending Date	RAFOS Floats Launched
Cruise	Beginning Date	Ending Date	Total	Into Rings
NBC 1	11-07-98	12-11-98	14	13
NBC 2	02-06-99	03-09-99	11	9
NBC 3	01-29-00	02-24-00	2	2
NBC 4	07-06-00	23-06-00	0	0

Note: All cruises were on the R/V Seward Johnson.

The floats equilibrated, on average, deeper than their target depths. We have divided them into three main depths: a shallow layer near 250 m, a mid-depth layer near 550 m, and a deep layer near 900 m. North Atlantic Deep Water (NADW) is located below roughly 1000 m; it flows southward on average, counter to the translation direction of the rings. The swirl velocity of several rings extended into the NADW, but because one of the goals of the NBC Ring Experiment was to measure the northward flux of South Atlantic Water, the floats were confined to depths less than 1000 m.

3. Description of the Floats

The North Brazil Current Rings Experiment used three different kinds of subsurface floats. One was an ALFOS (ALACE-RAFOS) float, which surfaced once every 10 days and transmitted like a very short-mission RAFOS float. There were six RAFOS floats of the "standard ROM" type, and there were 22 of the latest kind of RAFOS float, the DLD2 (see Appendix A for a discussion of DLD2 technology).

Eighteen DLD2 floats were purchased from Seascan Corporation of Falmouth, Mass., and assembled, calibrated (temperature and pressure), and ballasted at WHOI. Four other DLD2 floats, prototypes provided earlier by Seascan that had been used for testing in the WHOI float laboratory, were also used. Six additional floats were assembled from spare parts and parts of floats retrieved from earlier experiments. The ALFOS float, assembled at WHOI, was used to monitor the sound sources.

The floats recorded temperature, pressure, and times of arrival (TOAs) of sound signals transmitted by moored sound sources. At the end of their missions, the floats dropped ballast weights, rose to the ocean surface and transmitted data to WHOI via the Service Argos Inc. satellite system.

Four of the 22 DLD2 floats were never heard by Service Argos, and one (float 037) was not tracked because it went deeper than its depth limit and surfaced after only 36 hours. Two of the six older-design floats were not tracked because of poor-quality TOAs. Overall we obtained 21 trajectories, although some were shorter than planned (Figure 1). Considering that we purchased 18 floats, this seems like a good record.

The majority of floats was deployed on two cruises roughly two months apart. In order to have most of the floats complete their missions in the same time period, the later-deployed floats were scheduled for shorter missions than the earlier ones. The DLD2 floats were set to repeat their listening schedule every 12 hours. Two of the standard-ROM RAFOS floats were also set to repeat listening every 12 hours; the other four repeated every 24 hours.

Figure 1. Float duration chart. Float numbers are shown to the left of the line that indicates submerged time. The dark part of the line indicates when the float was tracked. Floats are listed in order of launch date.

The floats were generally set to listen starting at 0100, 0130, and 0200. Nothing was gained by including the 0200 window, and little was lost by omitting the 0030 window. However, the tracking program is hard-wired for 0030, 0100, 0130, and for 17 columns per record. This meant that an extra step was needed, splitting the longer records, and shifting all TOAs and correlations right four columns, inserting a "null" 0030 window.

Columns before conversion:

1 | 2 3 4 5 | 6 7 8 9 | 10 11 12 13 | 14 15 16 17

count | 0100 | 0130 | 0200 | T&P

Columns after conversion:

1 | 2 3 4 5 | 6 7 8 9 | 10 11 12 13 | 14 15 16 17

count | 0030 | 0100 | 0130 | T&P

(null)

Neither kind of RAFOS float stored the complete temperature or pressure values. This means that a target value must be assigned, and the appropriate rollover value applied until the result is within range of the target. The DLD2s used the last-recorded pressure and temperature of the mission as a default target (for earlier floats, the target was assigned). These values were often at the maximum or minimum. The floats also experienced a wide range of temperatures (especially those reaching the Caribbean or going ashore). For these reasons, a number of floats (18 of 21) needed second and sometimes third targets to get the pressure and temperature correct, before they could be processed.

The floats were ballasted to drift at a preset depth. This involved weighing the floats in air and in water at different pressures (see Tables 2 and 3). Two floats (037 and 038) surfaced early due to exceeding their pressure limits. On average the floats equilibrated 73 m deeper than their ballast depths. RAFOS floats 431 and 432 were always shallower by 25 to 75 m than their target depths. Floats 156 and 193 had mean pressures close to their targets. The DLD2s were all deeper than their target depths throughout their missions.

Table 2. RAFOS float summary – launch and end positions

	Launch				Surface				Target	Equil	Mean	Mean
Float	Date	Lat	Lon		Date	Lat		Lon	Depth	Depth	Temp	Depth	Ring	CTD
	yymmdd	˚ N	˚ W		yymmdd	˚ N		˚ W	m.	m.	˚ C.	m.
	DLD2 Floats
022	990219	9.179	52.439	000513		12.633	52.287		450	550	7.38	530	R3	2-42
023	990219	9.179	52.439	000513		7.981	47.296		800	910	5.21	907	R3	2-42
024	990219	8.780	52.481	000513		13.063	74.758		200	250	11.90	268	R3	2-41
025	990219	9.179	52.439	000513		15.086	70.016		200	250	13.90	282	R3	2-42
029	981208	8.841	56.242	000530		10.018	77.339		200	275	11.70	269	R1
030	981206	6.405	50.521	000527		8.756	48.164		450	555	7.44	531	R2	1-52
031	990213	10.266	57.203	000507		16.681	52.500		450	540	8.34	512	R2	2-23
032	981209	9.218	56.373	000531		19.544	66.109		450	620	7.29	592	R1	1-64
033	981206	6.405	50.521	000527		12.544	58.603		200	320	10.24	311	R2	1-52
034	981208	8.841	56.242	000530		13.571	64.200		450	600	7.38	527	R1
035	981206	7.112	50.771	000527		12.177	60.228		450	540	7.56	502	R2	1-53
036	990219	8.780	52.481	000513		12.209	52.102		800	860	5.31	843	R3	2-41
037	990302	5.498	48.321	990304		5.298	48.366		800	1000	Not tracked		R4	2-68
038	981206	7.112	50.771	991104		1.685	23.879		200	240	15.03	248	R2	1-53
039	990305	2.671	47.336	000814		1.367	35.457		800	880	4.87	879	NBC	2-80
040	981208	8.841	56.242						800		Never heard		R1
041	981206	7.112	50.771	000527		6.782	46.751		800	890	4.98	860	R2	1-53
042	981206	6.405	50.521	000527		12.881	50.788		800	870	5.22	860	R2	1-52
043	990411	8.314	52.608						200		Never heard		NBC
044	990219	8.780	52.482						450		Never heard		R3	2-41
045	990213	10.266	57.203						200		Never heard		R2	2-23
046	990213	10.266	57.203	000507		9.547	55.230		800	840	5.36	891	R2	2-23
Older-design RAFOS Floats
156	000220	10.020	57.103		000714	13.779	63.230		200	220	12.95	223	R5	3-52
193	000220	10.020	57.103		000714	14.478	61.948		450	520	9.79	491	R5	3-52
430	981209	8.901	56.177		000607	9.936	84.087		200		Not tracked		R1	1-63
431	981205	6.231	49.509		000528	4.795	1.957		200	180	18.65	148	R2	1-50
432	981206	7.112	50.771		000528	11.110	49.987		1000	950	4.98	935	R2	1-53
433	981209	9.218	56.373		000227	14.360	68.639		200		Not tracked		R1	1-64

Floats 022 - 046 were DLD2 floats that listened and recorded T/P every 12 hours. Float 43 was deployed by John Whitehead from R/V Oceanus. Floats 156 and 193 were older, re-powered RAFOS floats that listened every 12 hours and recorded T/P every 24. Floats 430 - 433 were older-design RAFOS floats that listened and recorded T/P every 24 hours. R1 refers to Ring 1, etc. These designations are internal to this report. The station number in the last column denotes the cruise and the CTD taken at the time that float was launched (Fleurant et al., 2000a, b, c).

Table 3. RAFOS float summary – performance

	Startup	Expected		Actual		Offsets (secs)		Messages		Track-able	Sources	Comments
Float	Date	Surface		Surface		Initial	Final	Received	%	Months	Heard
037	990221	000814	990304					2/2		0/0		surfaced after 36 hrs out-of-range to east never heard "exponential"pressure ended in Caribbean ended in Caribbean never heard TOAs end at surface ended in Caribbean ended in Caribbean never heard ended in Caribbean battery failure? on surface, no TOAs never heard ended in Caribbean ended in Caribbean
038	981204	000527	991104		0.0		0.23	329/334	98	6/11	4
045	990213	000508
031	990213	000508	000507		0.0		0.68	450/450	100	15/15	5
046	990213	000508	000507		0.0		-2.02	450/450	100	15/15	6
024	990218	000513	000514		-1.0		1.19	432/450	96	5/15	3
025	990218	000513	000514		-1.0		1.45	446/450	99	9/15	4
022	990218	000513	000514		-1.0		1.58	445/450	99	15/15	5
044	990218	000513
023	990218	000513	000514		-2.0		2.76	440/450	97	15/15	5
036	990218	000513	000514		-1.0		1.55	341/540	63	15/15	6
030	981204	000527	000527		0.0		2.07	537/540	99	18/18	5
033	981204	000527	000527		0.0		0.86	539/540	99	18/18	4
035	981204	000527	000527		0.0		2.00	540/540	100	18/18	6
041	981204	000527	000527		0.0		4.04	481/540	89	18/18	5
042	981204	000527	000527		0.0		3.76	539/540	99	18/18	6
431	981205	000528	000528		-23.0		-206.4**	215/271	79	2/18	2
432	981205	000528	000528		-24.0		-258.0	246/271	91	18/18	5
029	981208	000531	000530		0.0		2.67	538/540	99	3/18	2
034	981208	000531	000530		0.0		-0.32	538/540	99	16/18	6
040	981208	000531
032	981209	000601	000531		0.0		3.30	466/540	86	15/18	4
433	981209	000601	000227		-24.0		-194.8	212/223	95	0/18	2
430	981209	000602	000601		-22.0		-158.5	67/271	25	0/18	0
043	990409	000702
156	000215	000714	000714		4.0		0.19	151/151	100	3/5	4
193	000215	000714	000714		2.0		-6.77	150/151	99	4/5	4
039	990221	000814	000814		0.0		1.22	441/540	82	18/18	6

All dates are year-month-day.

* The first Argos transmissions for these two floats were on land.

** The final offset for 431 was reduced to -67.31 to match the TOA end date of 990414.

One advantage of RAFOS floats over the previous SOFAR floats is that two end points are available to check the acoustic tracking: the launch location and the surfacing location, obtained from Service Argos' calculation of the locus of the transmitter. The surfacing location was not useful in two cases because the floats were on land before they began transmitting. Float 430 stayed submerged for 100 days; it was probably picked up around day 250. Float 430 first transmitted from Costa Rica. Float 431 stayed submerged for 120 days, and was picked up after 200 days. It first transmitted from Point Three Points in Ghana (so-called because it is the land closest to the Equator, the Greenwich Meridian and sea-level).

Before we received all messages from 432, it was picked up and taken to Martinique. From there it was returned to Woods Hole.

Five floats (029, 031, 032, 034, 046) shared identical zeroeth messages (roughly one-half of their first output records). This had the effect of making them look like they had recorded 450 messages in 449 days. In order to produce the correct result, the start dates had to be moved one day earlier than shown in the launch logs. It is not clear why these floats failed to overwrite the default information in the zeroeth message.

Several floats (see 031, 033, and 035) show a clumping or quantizing in their temperature records. This is a result of being located in layers of nearly constant temperature and salinity (Schmitt et al., 1987)

Float 433 had no useable TOAs after 80 records, and only heard one source. It stopped collecting data and began transmitting after 223 days and gave its final status as "on the surface." The TOA record is typical of the case where the battery voltage drops.

RAFOS floats can detect when they are on the surface by recrossing a minimum pressure and also exceeding a maximum temperature. Because of past problems, for the NBC experiment the minimum pressure was set to 1000 counts, very close to (and sometimes above) the surface. This explains why 430 and 431 continued "in mission" after reaching the surface, and suggests that the "on surface" status of 433 was caused by low voltage.

The pressure record of float 046 repeatedly sweeps the same 150-meter range. No oceanographic explanation has been posited and these results are attributed to instrument problems.

The DLD2 floats can store message-numbers up to 2048, but the earlier floats can only store values up to 256. The checksum is used to decide whether a message-number is greater than 256 or not. This creates uncertainty when you want to fill in gaps in the received data with messages with bad checksums. This was an issue for floats 431 and 432.

4. Description of Sound Sources

The NBC floats used sources deployed by scientists from three different countries for five different experiments (Table 4, Figures 2, 3). The oldest source used, 185, was deployed October 19, 1994, as part of the Deep Western Boundary Current Experiment (BOUNCE) (Hunt and Bower, 1998). Source 185's clock had drifted at a faster rate than was deduced at the end of BOUNCE, so an expedient offset was applied the one time it was used.

Table 4: Sound source summary

	ID	Pong	Corr	Rep	Start	End	Offset	Drift	Depth	Lat	Lon	S.V.
Exp.		Time	(min)	Rate	Date	Date	(sec)	(sec/rep)	(m)	˚ N	˚ W	(m/sec)
BOUNCE	185	0100	0	12	941019	none	0.561	0.0057	1500	36.687	58.263	1484
French	B1	0100	0	24	950101	none	0.0	0.005	1000	7.028	49.925	1480
DBE	191	0100	0	24	950212	none	0.0	0.0	1000	-0.327	32.643	1482
NBC	69	0100	0	12	981108	000622	0.0	0.0043	650	13.000	57.886	1484
NBC	77	0130	-30	12	981110	000620	0.0	0.0113	700	12.996	51.107	1484
MOVE	403	0035	25	24	000117	none	0.0	0.0	1100	21.938	62.570	1484
MOVE	404	0100	0	24	000129	010103	-3.0	0.0	1110	15.324	51.526	1484
MOVE	406	0130	-30