Introduction

We have implemented a modification in the towed OPC which enables the measurement of average flow through the tunnel. This is accomplished internally in the OPC by measuring the transit time of small particle passing through the light beam of a know width of approx. 4mm. Only moving particles that fall within the size range of (250-350 µ ) are selected by the OPC microprocessor and an average is estimated over 5-6 particles. The average flow is transmitted to surface at the repeating 1/2 second data stream of the OPC.

This same circuitry has been applied to the lab OPC for measuring the lengths of animals > 1cm which are in turn greater than the beam width. For a measured transit time and a known flow speed, an accurate estimate/calibration of the length can be made. The measurement is enabled by 2 factors; 1) laminar flow through the flow tube and , 2) forced alignment along the flow path by the narrow flow tube (2 x 2 cm cross-section). Measurements of lengths ranging from 1 - 4 cm were made on 2 species; Antarctic krill Euphausia superba supplied by University of California San Diego and Meganytiphanes norvegica from our local Scotian Shelf waters.

Methodology and Calibration

In order to measure >average = flow, the OPC microprocessor will process all particles and before applying its filter algorithm. In this version used for the Lab OPC, we acquired the time measurement (count) of every animal and attached it to the size and subsequently sent both parameters to a computer for processing. The data format is a simple addition to the existing file generated by the FOCAL OPC-DAC program. A sample file is presented here:

Tue Mar 02 09:54:00 1999	( Normal Focal data header)
2,492			(id=2, light attenuance=492)
3,232			(id=3, 1/2 sec time counter=232)
7,93			(id=7, average flow - time count=93)
1,493			(id=1, digital bug size= 493)
6,1528			(id=6, transit time (count)= 1528 for previous bug)
2,492			(id=2, light attenuance)
3,233			(id=3, 1/2 second time counter)

In the case of time counts denoted by id=6 & 7, the actual time is derived by multiplying the time count by 0.01736 yielding time in milliseconds.

The flow schematic of lab circulator used in measuring euphausiids is shown in Fig. 1.

The pump in the reservoir tank delivered water to a sample tank which was kept filled to a constant level. By maintaining a constant >head= in the sample tank, a constant flow was maintained through the flow tube of the Lab OPC situated below the tank. By adjusting the water level in the sample tank, 2 rates of flow were tested at 1.15 m/sec ( 385 cc/sec) and 0.81 m/sec (274 cc/sec). A cod-end bucket placed in the reservoir tank contained a 100 µ mesh filtration window.and was primarily used to collect samples from the effluent of the OPC. The bucket also served to filter and clean the water of particles prior to running experiments. Euphausiids added to the sample tank flowed through the OPC and thus be recovered in the bucket without damage.

Using a constant flow rate of 1.15 m/sec, the OPC was calibrated using 1/8" nylon rods cut to specific lengths of 1, 2, 3, 4, 6, & 8 cms. Each rod was passed through the circulator/OPC approximately 20 times and the (transit) time count measured from data with id=6. Figure 2 shows the frequency histogram of digital time counts for the 2, 4, & 8 cm nylon rods for a total of 20 measurements or "passes" through the OPC. The mean and standard deviation for each of the 3 rods was 827 +/-71, 1459 +/-108, & 2899 +/-341 respectively.

The experiment was repeated for the remaining 3 rods of 1, 3, & 6 cm length. Finally, the OPC was again calibrated at a lower flow rate of 0.81 m/sec using all 6 rods. As expected the time count increased proportionately while the standard deviation remained within a range of 5-15%. It was noted in Fig. 2 that the length distribution were skewed slightly on the high side where a few measurements were found at a time count higher than the main peak. The measurement circuitry did not appear to be the cause while it appeared that the rods were actually moving through the flow tube more slowly. One possible cause is the potential for the rods making contact with the sides of the flow tube during transit.

A range of lengths of euphausiids were then selected form 1-4 cm and recycled in the OPC for measurement. The number of passes through the OPC ranged from 5-20 dependent of the condition of the animal. Repeated passes and handling often caused damage and limited our ability to achieve the maximum of 20 measurements. Moreover the damage also contributed to the variability in repeated.measurements. Figure 3 shows examples of a number of euphausiids of various lengths repeatedly measured.

These lengths were 0.9, 1.9, 3.0 & 4.0-4.2 cm. Three euphasuiids ranging in length from 4.0-4.2 cm resulted in the 3 distributions shown in Fig. 3 ranging from 1500-2000 in digital time counts. Of these 3 euphausiids, 2 distributions (denoted by dashed and heavy lines) resulted in std. errors of 5% while the remaining distribution had a larger error of 15%. In general the euphausiid measurements were comparable with those made with the nylon rods while the 4 cm animals appeared to be overestimated by about 10-15% in length. The number of passes made by each animal through the OPC Flow tube ranged from 5-20. In many cases, damage from handling limited the number of passes achievable.

All the data for length vs digital time counts from the measurements of the nylon rods and euphausiids are collectively presented in Fig. 4.

For each of the flow rates used, it was seen that the relationship between the lengths and digital time counts were reasonable linear. The measurements from all euphausiids using the high flow rate (1.15 m/sec) were superimposed on the straight lines obtained from the nylon rods and, for the most part, fell within 5%. It was seen that at short lengths of 1 cm or less the deviation increased. It was expected that at animal lengths smaller than the width of the flow tube (2 cm X 2 cm), the animal would be not be expected to be aligned with the flow. The data for the nylon rod calibration were fitted to a straight line and the resulting best-fit curves were:

L = -0.4142 + 0.002967 T		(Flow rate 1.15 m/sec)
L = -0.3639 + 0.002376 T		(Flow rate 0.81 m/sec)

where L is the length in cms and T is the digital time count obtained from the OPC. It is anticipated that for flow rates in between this range or slightly outside the range, that one could linearly interpolate or extrapolate these coefficients.

Measurements of Euphausiid Samples.

Samples of Antarctic krill were run through the lab OPC and length distribution measured. It was noted that these sample jars contained damaged animals and debris which would clearly result in measurements that were lower than the length distribution overall. Figure 5 show the length distribution of the sample resulting in a total of 223 counts. Since there were no euphausiids in the length range of 1-2 cm, these counts were clearly a result of debris or damaged animals. After manually removing the debris and any broken/damaged animals present in another sample jar, the remaining selected animals (115) were run through the circulator and counted. Figure 6 shows the resultant length distribution where we see considerably fewer counts below 3 cm. The dominant size distribution was seen to range from 3-4.2 cm.

Summary and Discussion

With the current modification to the OPC, it is feasible to measure sample lengths ranging from 1-8 cm with a standard error of " 10%. Sampling errors were primarily caused by variations in orientation along the flow axis and by variations in flow rate. Sample integrity also deteriorated resolution since animals that were fragile seem susceptible to breakup and >folding= while transiting the beam. Euphausiids with attennae intact may also prove to be problematic since the antennae will be measured and included in the total length measurement. It is estimated that for most passes through the flow tube, the antennae remain folded and unmeasured while the euphausiid transits the beam headfirst. However in a reverse direction, the antennae are outstretched and may be measured. The effect will require monitoring during during data analyses.

The Lab OPC can be used in either low or high sensitivity modes for length measurements. Ideally the low sensitivity mode should used since it renders the detection circuitry less sensitive to detecting appendage strands or antennae which may protrude and add to the length measurement. The >average flow rate= measurement identified with id=7 and transmitted every 1/2 second is more accurate and usable on the high sensitivity mode. If calibrating on the low sensitivity mode, an additional flow sensor show be utilized for intercomparison first before relying on the accuracy of this measurement.