WHP Ref. No.: AR11
Last updated: January 4, 1993

Cruise Report  AR11 Subduction 3

Chief Scientist: Terrence M. Joyce, WHOI
Ship :  R/V Oceanus OC254/4 (EXPOCODE: 32OC254/4)
Ports of Call: Las Palmas to Azores
Cruise dates: Nov. 24 - Dec. 17, 1992

Introduction

This was the third in a series of 4 cruises in the Subduction Experiment 
devoted to Seasoar sampling in the upper ocean. On the first cruise in May 
1991, a total of 18 Bobber floats were released south of the Azores front in 
the general vicinity of 30o N, 23o W. These were put at three different depth 
intervals (shallow, medium, deep) defined by bounding isotherms between which 
to 'bob'. At the deployment sites, mesoscale-resolving spatial surveys were 
carried out by towing the Seasoar undulating CTD in a 'radiation pattern' 
between depths of 0 and 400m: spanning the depth range of the Bobbers. Tracer 
samples were collected from a conventional profiling CTD for tritium/helium-3 
dating of the upper ocean. Some limited tracer and Seasoar sampling was also 
carried out near the Azores front. The second cruise in February 1992 focused 
on Seasoar and tracer studies of the Azores front; no measurements were made 
over the Bobbers. This third survey was planned  to concentrate on the 
measurement of the mesoscale variability following the Bobbers in order to 
define the nature of the change since deployment. The fourth and final cruise 
is planned for summer 1993. All of these cruises were/are aboard the R/V 
Oceanus.

Of critical importance to goals of this cruise was the location of Bobber 
positions. The main array of Autonomous Listening Stations (ALSs) deployed for 
Bobber float tracking is still in the water and the data are (hopefully) safely 
still in the instruments. In order to discover float locations for the cruise, 
several 'real time' tracking options were available. The first of these was a 
Drifting Sofar Receiver (DSR) consisting of a surface float, 500 m of cable and 
hydrophones at the bottom. Telemetry signals received from the Bobbers were 
processed by the DSR and relayed via satellite (ARGOS) to shore. We deployed 
(and later recovered) one of these drifters on the cruise (it did not work). 
The second mode was an ALFOS float: a modified ALACE float built by D. Webb 
(with much input and support from R. Davis) and modified at WHOI to carry a 
small hydrophone (like a RAFOS) and signal detector. It was normally floating 
at a depth of 1000m (???), but every 10 days would come to the surface, and 
transfer its Bobber telemetry data and its position to shore via ARGOS. One of 
these worked prior to and during our cruise: two others did not. The third and 
last resort for tracking was by heaving to and lowering a hydrophone array over 
the side of the ship to a depth of 1200m (sound channel axis) and 'listening'. 
A total of 4 listening stations were necessary to 'locate' the shallow Bobbers 
selected for tracking. 

In addition to the above work, deep CTD stations were made over Bobbers to 
collect water samples for calibration of both CTD and Seasoar sensors (all 
SeaBird), and to fill in gaps in a large-scale tracer survey of tritium/helium-
3 in the region made in October 1992 aboard the R/V Charles Darwin. 

Cruise Narrative

The Oceanus left Las Palmas, Canaries, at 0910Z on 24 November and steamed to 
the southwest for approximately 1 day before deployment of the DSR (serial 
#02), an initial listening station (LS#1), and test CTD station (CTD#1). The 
time window for Subduction Bobber telemetry was 0-600Z and 12-1800Z every day. 
Just prior to LS#1 it was discovered that the hydrophone array was 'dead' with 
no signal at all. We deployed the backup cable, but our LS results showed very 
low signal levels at telemetry times and thus did not provide hard evidence for 
Bobber locations. It was later discovered that the penetrator pin was bad on 
the main array-this was subsequently fixed and used for the remainder of the 
listening stations. The main problem with the backup array was a broken wire in 
the cable which cut off 3 of the 6 hydrophones and thus limited the signal gain 
of the array. The DSR launch went well, but the first CTD station showed a lot 
of unbelievable vertical structure in the salinity and oxygen profiles. Water 
samples were collected at the bottom (4500m) of the cast and at the oxygen 
minimum layer (app. 900m). The test station was quite useful in teaching us the 
importance in 'turning on' the SeaBird CTD system after it is in the water to 
properly 'prime' the pump and associated plumbing, and to identify 3 bad 
(leaky) bottles out of the 24 place 1.2 liter General Oceanic rosette. 

Because we knew before the cruise that the Bobbers were all to the south of the 
working ALFOS (because they had been tracked by an equatorial ALS array 
recovered in September 1992!), data from the DSR or the LS was needed to 
determine float position. Unfortunately the LS#1 was not successful and the 
first data back from the DSR were not good either. For this reason, we steamed 
south and made LS#2 on 27 November. Data return from this station was much 
better and gave us tentative locations for a few of the Bobbers. One of these 
was Bobber 26, which was the furthest south of the lot. We went to this 
location and made a CTD station #2 followed by  listening station #3 and our 
first 'radiation pattern' mesoscale survey with the Seasoar, which lasted 
approximately 27 hrs. towing the Seasoar at 7.5 kts. Because we experienced 
problems getting the servo controls to properly fly the Seasoar at first, we 
repeated legs 1 and 2 of the pattern before departing the region. The problem 
with the control of the servo was quickly fixed and did not re-occur for the 
remainder of the cruise. The LS#3 did not give good signals for Bobber 25, the 
next one to the north. For this reason, we towed the Seasoar over the suspected 
location and then headed for one of the two possible locations for Bobber 15 
which were both possible based on the data from ALFOS and LS#3. During this run
the main acquisition computer (a hybrid SUN workstation) for the Seasoar 
crashed and could not be re-booted once a bad fuse wasa replaced. Some quick 
shuffling put the data acquisition back on a PC with the navigation data logged 
to a laptop as backup (ADCP data were collected for the entire cruise including 
navigation data at 5 min intervals). A total of about 5 hrs. of data were lost 
due to the crash. Fortunately, the Seasoar deckunit and command system was 
unaffected and the fish contiued to fly and collect engineering data on pitch, 
roll, tension, etc., during the CTD data loss. 

Following a second mesoscale pattern (refered to as 'star' in the log), CTD #3 
was made on 28 November. This was the first station where tritium/helium-3 
samples were collected (from all water samples - subsequently samples were 
collected only from the 12 surface-most bottles). We then re-deployed the 
Seasoar and towed it to the site of Bobber 19 where star #3 was made and thence 
to the second possible site for Bobber 15 (we later determined that this was 
the most likely one) where star #4 and, after recovery of the Seasoar, CTD #4 
were made. Our final listening station (#4) was then made at the site of the 
CTD station and we steamed to Bobber 21 with the Seasoar deployed. We reached 
this site on 8 December and carried out star # 5, followed by CTD #5. 

As at this point of the cruise it was clear that the DSR was not working and we 
were at our closest point to its position (ARGOS positions were OK and relayed 
out to the ship by fax), we steamed to the DSR and recovered it on 10 December. 
After steaming back over Bobber 20, we did CTD #6, deployed Seasoar, and made 
our last mesoscale survey (star #6).  We then began a two day steam to the 
north, ending just to the west of the sites where the Bobbers were set in May 
1991. Seasoar was then recovered, and a final CTD (#7) was made to the south of 
the Azores front, after which we steamed for Ponta Delgada, Azores where we 
arrived on 16 December.

Instrumentation Notes

The Seasoar, manufactured by Chelsea Instruments, Ltd., is a towed fish 
equipped with a propeller-forced wing that can be adjusted to fly the fish. A 
winch with 750m of faired cable is used with a multi-conductor cable. The fish 
was loaded with a pair of Seabird  temperature and conductivity sensors and 
pressure and dissolved oxygen probes, all of which telemeter data up the cable 
at 24 hz. In addition, engineering sensors measured pitch, roll, propellor 
rotations, pressure, wing angle; and tension on deck; they are sampled at 2.5 
hz and the data displayed and stored on a separate computer on board the 
vessel. Just prior to the final recovery of the Seasoar, the fish became 
'stuck' at the surface for extended periods of time; this was due to the wings 
sticking in the up position -possibly due to loss of hydraulic pressure in the 
Seasoar. On the profiling CTD, pairs of temperature and conductivity sensors 
and a single oxygen probe have water pumped through tubing. During the course 
of the cruise, the secondary sensors on each instrument were exchanged, as well 
as the Seasoar oxygen probe on the final CTD station. Problems were encountered 
with the oxygen probe on station 2, which led us to change sensors. When the 
problem recurred on station 3 (an abrupt shift in the oxygen current at a 
pressure of 2500 dbar on the down trace), we were able to isolate and correct 
the problem, which was with a faulty cable connection. 

Preliminary Results

Various results have been tabulated at the end of this report. Table 1 is a 
complete event log of the cruise, while Table 2 extracts  locations of CTD, 
listening, and DSR 'stations'. Table 3 gives the estimated positions of the 
Bobbers tracked on the cruise along with the sources for the estimate. The 
cruise track is shown in Figure 1 with station locations indicated. 

When the Bobbers were initially deployed in May 1992, they were in or slightly 
deeper than winter mode water of the Madeira type. Our preliminary findings 
show that none of these winter modes remain in the water tagged by the Bobbers; 
they have been replaced by and embedded within a smooth pycnocline. At the time 
of deployment, the Bobbers were grouped in two regions and Seasoar mapping was 
carried out with 'radiator patterns' with diameters of approximately 120 km. It 
was discovered subsequently that the mesoscale variability had scales of order 
10 km. leaving large areas with the pattern that were poorly mapped. For this 
reason the diameter of the pattern was reduced to 80 km. for this cruise. This 
was fortuitous because the smaller pattern could be carried out in 1 day as 
opposed to 2 days, thus permitting more patterns over the now dispersed Bobbers 
to be done. Data from the first pattern or star will be shown for illustration 
of our results. This pattern is chosen not because it is typical, but because 
the loss of the main acquisition computer and switch over to the backup 
required a change in data format and loss of the ability to provide the same 
advanced level of processed data at sea. 

The Seasoar sensors data at 24 hz. was edited and averaged into a 3 second time 
series as the instrument 'flew' between the mixed layer and a depth of about 
400 dbar. Data were interpolated onto a uniform grid in depth/distance along 
track using a Gaussian filter with vertical and horizontal scales of 5 dbar and 
4 km, respectively. One of the long legs of star 1 (leg 6) is shown in figure 
2. The mixed layer had a temperature and salinity of 24.5 oC and 37.1 psu 
respectively. A subsurface salinity maximum in excess of 37.2 can be seen below 
the base of the mixed layer over most of the section. The density structure 
shows no deeper mode below the mixed layer. The locations of the interpolated 
data from the entire survey (figure 3) show the spatial mapping pattern. Data 
were then mapped onto density surfaces (figure 4) at intervals of 0.05 sigma 
theta. The thickness of on each density surface is the difference between 
pressures at bounding density surfaces. Below the mixed layer, thickness 
monotonically increases with depth. The 'thin' layers of the pycnocline show a 
large degree of potential temperature variation over the survey. Two density 
levels have been selected and objectively mapped using a spatial correlation 
scale of 10 km in figures 5, 6. The T/S variability can be seen to have short 
spatial scales of  10-20 km in the north-south and 40 km east-west which are 
'coherent' in the vertical. Thicker layers tend to be warmer (and saltier). The 
error map assumes a noise to signal ratio of 50 (0.5) and shows the ability of 
the array to 'map' the variability, which is much improved over the larger 
patterns of the first cruise. We have not examined the variability scales for 
the entire cruise yet, but these data suggest a much smoother and somewhat 
larger scale than the deployment cruise with a clear relationship between the 
thickness and the potential temperature of the layers. The oxygen data appear 
rather good and once the probe data are 'calibrated' should yield oxygen maps 
as well. The pair of conductivity sensors on the Seasoar appeared stable over 
the cruise and the preliminary salinity data shown in the figures should be 
correct to at least 0.01.

Seasoar conductivity, temperature and oxygen sensors were exchanged with  those 
on the cable-lowered CTD and we should eventually be able to calibrate the 
conductivity and oxygen sensors used on the cruise by reconciling the results 
with the bottle data. Sampling depths for the bottle data were at 500 dbar 
increments from the bottom of the stations (4000 to 5000 dbar) up to 1500 dbar. 
Intervals were reduced to 50 dbar between 400 dbar and the surface. Sampling 
for tritium/helium-3 was done for the shallowest 12 bottles (700 dbar to the 
surface) on all stations after #2, except for station 3, where all levels were 
sampled with the copper tube 'clangers'. Deep CTDs were done near Bobber 
locations except for the last station (#7) which was done to the west of the 
original deployment sites for the Bobbers and south of the Azores front in 
order to 'fill in' sampling gaps from the Darwin mooring and tracer cruise of 
Weller and Jenkins in October 1992. The front itself  was crossed on 15 
Decmber, as can be seen from the ADCP temperature sensor (figure 7) and the 
ADCP velocities (contoured vs. latitude, figure 8a,b). 

In total, approximately 4100 km of Seasoar profiling was obtained during the 
cruise.

Personnel and Acknowledgments

Six members of the scientific party carried out all of the work and are listed 
below together with general areas of responsibility; all except Szelag, who is 
from URI, are from WHOI.

	Terrence Joyce	Chief scientist, oxygen analyses
	Julie Pallant		Seasoar, CTDs
	Frank Bahr		Seasoar, ADCP
	Jerry Dean		Float tracking, Seasoar
	Paul Fucile		Float tracking, electronics troubleshooting
	Jan Szelag		CTDs, salinity analyses

We wish to thank Captain Howland and the crew of the R/V Oceanus for their 
assistance and the financial support of the Office of Naval Research for the 
Subduction ARI program, which sponsored this WOCE process study.

Terrence M. Joyce
15 December 1992

