CMO Dye Studies - Sept '95 Cruise 


summary plot of 1995 rhodamine experiment

Introduction

A pilot dye study cruise was conducted aboard R/V Oceanus from September 9-14, 1995. Rhodamine dye was released in a single streak near the planned CMO central mooring site (moorings were not deployed until summer 1996), and sampled twice over a period of 3 days. Three ARGOS drifters with holey sock drogues tethered at 40 m depth were deployed at regular intervals during the injection as an aid to tracking the movement of the dye patch. A summary of the injection and sampling surveys for the 1995 experiment is shown to the right. An inset to the figure shows mean vertical profiles of dye as measured during the two surveys.


Hydrography

temperature section salinity section
sensity section hydrographic stations

A hydrography section extending across the shelf was made at the beginning of the cruise which consisted of ten CTD stations between approximately the 35 m to the 310 m isobath. These stations were occupied over a 29 hr period between September 9-11, 1995.

Contour plots (right) show the (a) temperature, (b) salinity, and (c) potential density, with station locations marked by vertical lines. The approximate location of the dye injection relative to the hydrography is shown as bold ellipse. Station locations are shown in plan view (d) (small circles) along with the planned CMO mooring locations (large circles). (Moorings were not deployed until summer 1996.)

The shelf-slope front extended upward and off-shore from approximately the 100 m isobath and presumably outcropped beyond the offshore extent of our survey. The cold pool was evident in both temperature and salinity, distinct from the shelf-slope front, although it was not a pronounced feature in this transect. (For comparison, see the September 1996 hydrography.

An additional feature visible in this transect, although it is more transient than the shelf-slope front or the cold pool, is a warm saline surface intrusion emanating from off-shore. Such intrusions, presumably fed by Gulf Stream water, are not uncommon to the region (Linder, 1996). However, due to their extreme temporal variability, these features are not well described by long term means and are poorly documented by climatological studies.

Buoyancy frequencies computed from this section ranged from 0 to 20 cph, with typical values of 10 cph in the region near the dye injection site (indicated by a red dot within each section). Based on this section and subsequent dye surveys, stratification remained fairly constant over the course of the experiment, with typical values of 12 cph at the level of the dye.




The Pilot Dye-Release Experiment

injection profile Rhodamine dye was released in a single streak near the planned CMO central mooring site (moorings were not deployed until summer 1996), and sampled twice over a period of 3 days. Three ARGOS drifters with holey sock drogues tethered at 40 m depth were deployed at regular intervals during the injection as an aid to tracking the movement of the dye patch.

The dye injection was performed along a target density surface of sigmatheta = 25.2 kg m-3, or approximately 48 m depth. A total of 100 kg of dye released over a period of about 1 hour while the ship was under way at approximately 1 kt. Analysis of the CTD data indicated that the standard deviation of the density during the injection was 0.031 kg m-3. Based on the average density vs. pressure relationship computed from the same CTD data, this corresponded to a standard deviation in pressure of 0.25 dbar.

In dye experiments, the exact dimensions of the initial streak are difficult to determine. Laboratory studies by {\it Lin and Pao} [1979] of dye dispersal in the wake of a towed vehicle showed that after a few buoyancy periods the width is approximately 5 times the height. Applying this guideline to this experiment gives minimum vertical and horizontal thicknesses of 4*sigmaz=1.0 m


Survey 1

survey 1 stick plot

The first survey of the pilot dye experiment consisted of six meridional tow-yo transects performed between 3 and 10 hours after injection. Of these transects, three transects yielded significant concentrations of dye. The images to the right show plan views of Survey 1 of the 1995 pilot study. The first contour plot is a plan-view kriging map (kg km-2), and the second a plan-view stick plot (g m-2) of vertical integral of tracer after correcting to a common time using the ADCP data. (In the stick plot, a vertical line is drawn at the location of each vertical profile with the length of the line indicating the magnitude of the vertical integral of tracer at that location.)

dye transect M32 dye transect M32 dye transect M36

Integration of the shipboard ADCP data as described above suggests that these three transects sampled nearly the same water; hence they provide little information regarding the zonal extent of the patch. In this respect, the first survey was somewhat incomplete. However, despite the poor coverage, useful information can still be gleaned from these data. The patch was fairly well homogenized and somewhat Gaussian in shape both in the horizontal and vertical directions. The significant vertical tilting of the patch between the first and third transects also suggests the possible importance shear dispersion as a mechanism of lateral dispersion. Note, however, that dispite the presence of vertical shear, Sundermeyer (1998) has shown conclusively that shear dispersion is in fact not the dominant mechanism of lateral dispersion in this expiriment.


Survey 2

survey 2 stick plot
The second survey for the pilot experiment was performed between 42 and 61 hours after the injection and consisted of twelve zonal transects through the dye patch.
dye transect Z01 dye transect Z04 dye Transect Z08
Transects for this survey show that the patch remained fairly homogeneous in both the horizontal and vertical directions. However, superimposed on the broader-scale dye distribution some variability in the concentration can also be seen with scales of a few hundred meters horizontally and a few meters vertically. Significant vertical tilting of the patch can again be seen over the course of the survey, this time in the zonal rather than meridional direction, consistent with the direction of the transects. Horizontal elongation of the patch is also evident, with the major axis oriented in approximately the north/south direction (see plan view maps).


Vertical Diffusivity

mean vertical dye profiles Bounds on the vertical dispersion for this experiment were obtained from the estimated initial condition and the above vertical variances by Ledwell et al. (1998). If the first survey is used as the initial condition, then the zero growth in variance between the first and second surveys yields a lower bound on the vertical diffusivity of zero. However, since the large vertical variance estimated for survey 1 may have been an artifact of the poor spatial coverage of that survey, the vertical diffusivity might better be estimated without those data. Assuming that the initial condition was a delta function in z, and considering the variance from the second survey, a Fickian diffusion model yielded an upper limit on the diffusivity of Kz <= 1.5 x 10-5 m2 s-1.

Horizontal Diffusivity

Horizontal diffusivities were estimated in two ways. First, the overall dispersion Ktot was estimated based on the vertically-integrated tracer. This measures the overall (total) horizontal spreading of the dye patch, and can be thought of as a the dispersion due to the advective effects of vertical shears as well as the effective diffusion due to small-scale mixing and stirring processes.

A second measure of lateral dispersion is the effective horizontal diffusivity, Kirrev, estimated from the dye concentration along the target density surface. This diffusivity takes into account the depth-dependence of the tracer and represents the irreversible dispersion of the dye patch in the sense that it includes only those processes that spread the dye along isopycnal surfaces. For example, Kirrev does not include the advective effects of vertical shears since these represent relative motions of isopycnals. However, it does include horizontal diffusion due to vertical shear dispersion, i.e., the interaction between vertical shear and vertical diffusion, since this involves spreading of dye along isopycnals.

Based on these two measures, the horizontal diffusivities inferred from the above two surveys were Ktot=9.2 (5.8 to 12.6) m2 s-1 and Kirrev=4.9 (2.9 to 7.3) m2 s-1.