Optical Plankton Counters (OPCs) have been in operation for approximately 12 years and while there has considerable research studies by the user community, some operational issues have emerged over this time, such as the its operational limitations in high densities, the lack of accompanying flow measurements and other measurement limitations. Reported here is the next generation of this device, the Laser-Optical Plankton Counter or LOPC which was designed at the Bedford Institute of Oceanography to address the future needs of optical plankton counting. Using a narrow laser beam and new sampling geometry, the LOPC is now capable of working in plankton densities of 106 /m3 nearly 100x greater than it predecessor, the OPC and is also capable of providing shape profiles of plankton >1.5 mm. Other new features include; i) the measurement of flow speeds through the tunnel by making a statistical estimate of particle time-of-transit, ii) a lower detection limit of 100 microns, iii) high speed towing up to 8 m/sec, and iv) overall smaller physical size. Data is presented from an LOPC mounted on the inside of 0.5 m plankton net (75 micron mesh size) showing good linear correlation between net samples (eg. copepod eggs, nauplii) and LOPC measurements. The LOPC-measured shape profiles from the same LOPC/Net tows clearly show copepods with antennae and in some case euphausiids. By increasing the tunnel width, the volume sampled by the LOPC can be increased by 5x. Other towing platforms that have been tested to-date are the Batfish vehicle towed at 8 kts and the Moving Vessel Profiler towed at 12-14 kts.

The Optical Plankton Counter (OPC) was originally designed at the Bedford Institute of Oceanography as a remotely-towed sensor providing continuous real-time information on zooplankton (Herman (1988); Herman (1992); Herman et al. (1993)). The OPC was intended to complement the information obtained from net tows by providing information overlap and higher resolution measurements. In retrospect from the author=s original objectives, the OPC was designed as a tool to separate zooplankton by size distributions in which peaks could be related to taxa. However additional information provided by the OPC became equally valuable such as its ability to measure a wide range of size distributions continuously, rapidly, in real-time and simultaneously with other environmental parameters. Moreover the OPC could provide high spatial and temporal resolution while being towed over large areas without the need of recovery or servicing.

The OPC has been deployed on a wide variety of platforms such as SeaSoar, AquaShuttle, Batfish, Scanfish, MOCNESS, BIONESS, V-fins, ARIES, LHPR, ROVs, and integrated with BONGO and ring-nets. There have been many oceanographic studies utilizing the OPC, such as: 1) investigations of macrozooplankton in the California Current from retrospective data analyses of CALCOFI data (Checkley, 2001; Mullin et al., 2002; Beaulieu et al., 1999), 2) studies in the North Atlantic to survey the vertical and horizontal distributions of overwintering C. finmarchicus (Heath and Jónasdóttir, 1995; Heath et al., 1999), 3) studies of plankton biomass size spectra in freshwater systems (Sprules et al., 1998; Sprules, 2002), 4) modelling of normalized zooplankton biomass size spectra and the study of size-structured zooplankton populations dynamics (Zhou and Huntley, 1997; Zhou et al., 2001), 5) measurement of zooplankton size distributions along CALCOFI stations east of Monterey Bay (Hopcroft et al., 2002), and 6) measuring zooplankton abundance and biovolume in water with high detritus (Zhang et al., 2000).

The OPC measures and transmits the cross-sectional area of each particle passing through its beam. As with all instruments, there are inherent limitations in its measurement capabilities. The most dominant is that of the "coincidence limit", that is, the density at which there a significant probability of two or more particles occurring in the beam simultaneously. As a result, the multiple particles are counted as "one count" and the size is presented as the sum of cross-sectional areas of all particles. In the case of the OPC, coincidence became a problem at plankton density of about 10,000/m3. Another limitation for the user is the lack of direct measurement of flow through the OPC tunnel. Even with recent technology advances, there is still not a suitable flowmeter small enough in size that would easily adapt to the existing OPC tunnel. As a result users have adapted larger flow meters in the vicinity of the OPC or have simply used tow speed as a proxy for flow.

The new laser-OPC (LOPC) has been under development and field testing at the Bedford Institute of Oceanography (Ocean Physics Section) since 1996 and first reported in Herman et al., 1998. It was designed to resolve the limitations of the original OPC and also to provide further zooplankton identification capabilities by measuring the shape profile of plankton larger than 1 mm. This paper describes the principles of operation of the LOPC, its mechanical configuration, the zooplankton parameters (eg. size , shape, etc.) that it measures, and its output format. Data sampled regionally from Scotian Shelf waters and the Gulf of St. Lawrence will be shown consisting of size distributions, shape profiles of large zooplankton such as, Calanus spp. and euphausiids, and flow measurements. The deployment of the LOPC on various platforms and vehicles will be discussed.

The operational principle of the LOPC is demonstrated in Fig. 1. A combination laser diode and line generator produces a diverging linear beam 1 mm in width. The beam is focussed by a cylindrical lens producing a parallel beam of 1 mm x 35 mm, subsequently reflected 900 by a mirror and directed through an air-water interface window into the sampling volume. At a selected distance from the window, a mirrored prism is used to redirect the beam back to the window on a parallel path directly below the emerging beam. Based on a beam height of 35 mm, the total frontal area occupied by the beam is 70 mm x 70 mm. The microprocessor shown in Fig. 1 does not represent a single component but rather a system consisting of a number of microprocessors. The central unit is comprised of; i) a digital signal processor (DSP - Texas Instruments, 200 Mhz, 1600 MIPS) processing data and algorithms, ii) a programmable logic device (PLD - Cypress Semiconductor, in-circuit programmable, 64 macrocell, 125 Mhz) digitizing and multiplexing data from the photodiode, and iii) a microprocessor (Persistor Instruments, CF1 module, 32 bit, 16 Mhz) operating as a system manager (ie. sensor integration, telemetry and data handling/management.

Figure 2 shows the components of Fig. 1 housed in a pressure case and attached sampling tunnel used for directing water flow through the light beam. In Fig.2, the standard tunnel is 70 mm wide resulting in a beam area (frontal cross-section) of 49 cm2. The tunnel can be replaced by other similar tunnels with a reasonable range of widths enabling the LOPC to sample higher water volumes with lower sensitivities. The example shown in Fig. 2 ( dashed line) is 5X the standard width or 350 mm and provides better sampling statistics for larger zooplankton, such as euphausiids, typically present in lower concentrations.

The light beam and optics of the LOPC were designed to address the issue of reduction of "coincidence counts" inherent in the OPC, that is, the probability of two or more particles occurring in the beam simultaneously and being counted as one particle. The principle used for coincidence reduction is depicted in Figs. 3(a,b) where a high density of particles is shown in Fig. 3a crossing a rectangular cross-section of the detector similar to that of a conventional OPC with dimensions 4 mm in width. With the extremely high density of particles shown in Fig. 3a, the detector would measure the sum of 13 particles yet only count them as one particle. In the example of Fig. 3b, the new detector is shown to have individual detection elements of 1 x 1 mm. It can be seen that the vertical array has already isolated 6 particles separately and in time, as these particles exit the detection area and as new ones arrive, these will be isolated also. The reduction in coincidence probability is essentially the ratio of the two smallest detection areas. In the case of the OPC, the cross-sectional measurement area is 80 mm2. The LOPC utilizes a laser diode line generator producing a narrow beam 1 mm in width and 35 mm in height (vertically-aligned as shown in Fig. 1) which impinges on a 35-element linear array photodiode 4 mm in width and 35 mm height. The "effective" cross-sectional beam area measured by a single photodiode element is therefore is 1 mm x 1 mm or 1 mm2 resulting in a reduction of the concidence by a factor of 80X as compared to the OPC. This effectively increases the operational capability of the LOPC in particle concentrations of approx. 106/m3.

The basic principle of measurement of the LOPC remains unchanged from the OPC and that is, the occluded cross-sectional area of the particle in the beam is measured relative to the area of the photodiode element detecting that occlusion. The OPC measures the cross-sectional area of any particle in its beam (depending on orientation) over the size range of about 250-25000 microns (esd). The LOPC splits the size range measurement into two parts by: 1) measuring the cross-sectional area of particles in its beam over a size range of about 100-1500 microns in what is termed as "single element plankton" or SEP, and 2) measuring the shape profile (2-dimensional) of particles over the size range of about 1500-35000 microns in what is termed as "multi-element plankton" or MEP. In the latter case, the components of each shape profile can be summed to yield an "esd" measurement and thus added to the size distribution described in the former case yielding a size distribution over the full range of 100-35000 microns.

The upper detection limit of the each photodiode element of the LOPC is approximately 1 mm (esd). If the particle size passing through the beam is less than 1 mm and occludes only 1 element, then it will constitute what will be defined as a "single-element plankton" or SEP. An example of SEPs is shown in the case of 2 particles on the right-hand side of Fig. 4(a).

If the particle size passing through the beam exceeds 1 mm (esd), then it will occlude more than 1 element at a time constituting a "multi-element plankton" or MEP. An example of MEPs is shown in Fig. 5(a). While each of 35 elements are scanned at a rate of 1 MHz, software algorithms residing within the DSP microprocessor analyze and separate an SEP from an MEP. One criteria was established for determining whether a particle is an MEP and, that is, if any pair of elements, eg. 2 & 3, are "active" (ie. detecting an occlusion) and simultaneous in time, either partially or totally, then the algorithm assumes that it is measuring an MEP. If the next pair of elements scanned, eg. 3 & 4, are "active" and simultaneous, then it assumes that elements 2, 3 & 4 are part of a single particle/animal. An example of the measurement principle of MEPs is shown in Figs. 5(a,b,c) illustrating the occlusion signals measured as zooplankton pass through the beam and over the photodiode elements. In order to reconstruct the shape profile of each MEP, a number of measured parameters are required. These are: i) each element number, ii) the start time while entering the each element/ beam, iii) the time-of-transit in each element, and iv) the area occluded by each element. All these parameters are transmitted to a surface computer for logging and reconstruction (non real-time). A preliminary or unprocessed shape outline of each animal is also displayed real-time by the acquisition software of the logging computer.

The separation between SEPs and MEPs is not precisely defined within the LOPC. Consider, for example, a particle with an esd of 500 microns which can also straddle the boundary separating 2 elements. Such a particle would be measured as an MEP, however, it would be difficult to build a shape profile based on the information acquired from only 2 elements. As a general rule, the greater the number of elements, the better the definition of the animal, ie., a euphausiid would be better-defined than a small copepod. The LOPC algorithms are set, therefore, to sum the measured areas of each 2-element MEP and treat it as an SEP. All SEP measurements, therefore, will consist of 1- or 2-element plankton and will extend to an upper size range of about 1500-1700 microns. By default, therefore, all MEPs consist of 3 elements or more.

There are situations where the algorithm may be changed to measure 2-element MEPs and not sum the measured areas, for example, when measuring fish eggs using the laboratory version of the LOPC. With the foreknowledge that the measured particles are spherical, the measured parameters of 2-element MEPs can be reconstructed to yield reasonable measurements of a "spherical" esd.

An MEP is assumed to consist of multi-elements being "active" within a simultaneous time period and therefore be caused by a large zooplankter passing over more than 1 element. It is also possible, however, that in extremely high density plankton regions, several small particles may cross the elements simultaneously and result in an MEP measurement. Such events are considered less probable, however, and more easily discernible by the LOPC data. High densities of small particles will yield characteristically low values of size and time-of-transit within the MEP measurements while large particles yield characteristically higher values of size (nearly full occlusion of each element) and time-of-transit.

In both cases of the LOPC and OPC, the ratio measured (particle area to photodiode element area) is approximately proportional to the particle area, however, some deviation occurs since the light beam distribution across the photodiode possesses i) a near-Guassian shape, and ii) some non-uniformities. The major change results from utilizing a smaller sensing element of 1 mm2 in the LOPC as compared to the OPC whereby a lower size detection capability is obtained with the LOPC. Currently the lowest detection threshold is set at 100 microns (esd - equivalent spherical diameter) for optimum performance with the LOPC, however, lower thresholds of 50 microns have been tested also and are expected to become operational in the near future. In comparison, the lowest detection limit of the OPC is 200-250 microns.

The LOPC size measurement of SEPs differs from that format of the OPC, in that the OPC presented a size distribution linear with particle area whereas the LOPC presents a size distribution linear with particle diameter (esd). The equivalent spherical diameter (esd) refers to that diameter derived by equating the area of a circle to the area measured of an irregularly-shaped (ie., a copepod) particle. The calibration of esd vs. LOPC-measured area is first performed by using a range of near-spherical beads which are passed by the light beam and measured. The resultant calibration of esd vs. area then becomes internally resident within the LOPC microprocessors. Following each measurement of a particle area, the esd is calculated by the LOPC and a "count" incremented within a size interval or "bin" having a size range encompassing that measured esd.

Figure 4 (a) shows a example of small particles of approx. 0.5 mm esd passing through the beam and over the photodiode elements. The two particles on the left comprising of a circular and ellipsoidal shape are shown to be encompassed by 1 element only. However, there is also a reasonable probability that the same particles may straddle 2 elements each, as shown in the example of the two particles on the right in Fig. 4(a). Normally the result would be an MEP measurement, however it is very difficult to reconstruct a shape profile from only a 2-element measurement. Therefore the particle areas from each element are added to form a single particle and the esd is subsequently calculated and "binned". The SEP measurement therefore is not strictly that of a single element but is meant to connote a "small" size range in which i) the cross-sectional areas are measured resulting in esd measurements, and ii) the shape profile of particles is not easily measured.

The SEP size distribution is comprised of 128 bins in total with a size range of 15 microns each and extending to a maximum of 1920 microns. All 128 bins and their appropriate counts/bin are telemetered to surface every 0.5 sec. designed to provide reasonable spatial resolution while towing at most speeds. Each transmission of 128 bins is immediately followed by a data line consisting of auxiliary instrument data, such as CTD, providing the depth information required by the user to locate vertically the size distributions. Examples of SEP or "small plankton" distributions measured by the LOPC are shown in Fig.6.

In considering the cross-over region of SEP and MEP measurements of the LOPC, an example is presented of spherical particles ranging in size between 1-2 mm as shown in Fig. 4b). In examining the simple geometry of such particles passing over the multi-element photodiode, there is a reasonable probability that the spherical particle of 1-2 mm esd will "occupy" 3 elements during its passage. For sizes on the high-side of the 1-2 mm range (ie. 1.8 mm as shown in Fig. 4(b)), the probability (of occupying 3 elements) is higher while for sizes of the low-side of the 1-2 mm range( ie. 1.2 mm, Fig. 4(b)), the probability is lower.

There is a low probability, therefore, that a particle with an esd of 1800 microns will end up as a count in the 120th bin of SEPs but more likely in becoming classified as an MEP measurement. By the same token, it is also unlikely that a particle with an esd of 1200 microns will be classified as an MEP since there is a low probability that it will occupy more than 2 elements in passage. As an approximation, therefore, it is estimated that the cross-over between SEP and MEP is about 1500 microns.

The parameters measured for each MEP necessary for reconstruction of the shape profile are transmitted asynchronously to surface by the LOPC in real-time as they are measured. The MEP data are inserted in between the 0.5 sec. intervals of the 128 size bins and provide the user with the same time/space resolution as the SEP data. Examples of processed MEP shape profiles are presented in Fig. 7(a) representing a composite of a number segments of data sampled from Scotian Shelf waters. With some foreknowledge of the contents of a simultaneous net sample, the profiles in some cases indicate clearly the shapes of a euphausiid (Meganyctiphanes norvegica), large copepods (Cal. finmarchicus or Cal. hyperboreus) and many shapes which remain unidentified at this time. Figure 7(b) represents data from the Emerald basin at 250m depth as sampled by the wide-tunnel LOPC mounted on a Batfish vehicle. Further analyses of the shape profiles (post-processing) is required to obtain a quantified size distribution similar to Figs. 6. By summing the signal "areas" of each of the elements of an MEP, an esd can then be estimated and finally a size distribution generated. An example of an MEP size distribution is shown in Fig. 8.

By combining the data of Figs. 6 & 8, the complete size distribution can now be generated describing the full size range measurable between 100-35,000 microns. The purpose in generating the MEP shape profiles ( Fig. 7) before calculating the size distribution (Fig. 8), is to provide the images necessary for the development of identification of zooplankton taxa. Although the images can only be used qualitatively at this time, future development of some form of statistical analyses of taxa using image identification software is foreseen.

The necessity for measuring simultaneous water flow through the LOPC tunnel was addressed by programming the LOPC to perform a statistical measurement of the average time it takes a particle to pass through the light beam of a fixed distance (1 mm approx.). The inherent problem with such a measurement is its dependency on the particle size. Consider, for example, a small particle of approx. 0.20 mm esd passing through the LOPC beam at a constant speed as shown in Fig. 9(a) and producing a time pulse corresponding to its time-of-transit. A larger particle of 0.60 mm passing through the beam at the same speed (Fig. 9(b)) would produce a time pulse wider than that of the smaller particle . Therefore the condition imposed on the accurate measurement of particle speed or flow speed, is that it be made only with particles of size much less then the beam width or << 1 mm. This particle size range was selected at 0.10-0.30 mm. The LOPC estimates the average time based on the number of particles measured in the required size range in a 0.5 sec. period and transmits the data attached to the SEP binned data. The reliability of the flow speed estimate will depend on the number of particles used to make the estimate. Therefore data corresponding to the number of particles used in the flow speed estimate are also transmitted along with the average time enabling the user to attach a confidence level to the estimate.

The relationship between the average time measured by the LOPC and the actual speed when calibrated is non-linear and for reasons of brevity will not be discussed here. A calibration curve will provide the user with the corresponding flow speed. An example of flow speed measured by the LOPC is presented in Fig. 10 where each measurement of speed represents a 3 second average. The LOPC, a mini-CTD and a TSK flowmeter were installed within the mouth (0.5 m) of a plankton net with a 75 micron mesh size and deployed in a vertical haul mode. The ascent speed as measured by the CTD was 1 m/sec as shown in Fig. 10 while the TSK flowmeter and LOPC correlated well while measuring flow speeds slightly less 1 m/sec and much lower flow speeds in the upper 50 m. The decrease in efficiency was attributed to partial clogging of the net by algal material or other small particles.

The response of the LOPC to actual zooplankton populations must be thoroughly measured by each user for their specific region by comparing the LOPC output to samples collected simultaneously with a plankton net.. The most convenient configuration for this sampling is obtained by mounting the LOPC inside a plankton net (described in following section) and performing vertical tows. An example of such intercomparisons is shown in Figs. 11(a,b) for LOPC data and net samples collected over a 18 stations in the Gulf of St. Lawrence. Figure 11(a) shows the intercomparison of the size range 100-200 microns which were found to correspond to copepod eggs and Ceratium sp.. The linear regressions showed good correlation (Figs. 11(a,b)) between the LOPC and net samples. The total counts in the net should exceed those of the LOPC by ~30x based on; i) the area ratio of net mouth to LOPC tunnel opening, and ii) measured flow in the net. The slope of 2.1 in the case of Fig. 11(a) is quite low, however, this is often the case for the smallest size range near the mesh size at which point there occurs considerable extrusion of plankton. At larger size ranges where the capture by the net mesh is more efficient, slopes ~20 are measured and are more representative of the expected ratios.

It is estimated that given the lower detection limit of the LOPC at 100 microns, the correct mesh size for intercomparison would be approx. 30 microns. Using such a small mesh size net would not be practical such operations due to clogging problems and the resulting sampling inefficiencies and uncertainties encountered. The use of a 75 micron net appears to be the reasonable compromise at this time.

The range of tow speeds for the LOPC is 0.25-8.0 m/sec which places its maximum tow speed somewhat higher (2X ) than the OPC. This enables the deployment of the LOPC with a variety of platforms ranging from slow net-tow to any platform being towed from most vessels steaming at full speed. Several examples are presented here, the first is that of an LOPC mounted inside a plankton net with a 0.5 m opening and a 75 micron mesh as shown in Fig. 12. The LOPC sampling tunnel is centered within the net mouth while a TSK flowmeter is mounted at a position of 1/4 of its diameter and level with the mouth of the LOPC sampling tunnel. A small CTD is mounted behind the LOPC. The mean tow speed used during vertical ascent was ~ 1 m/sec. This LOPC-Net configuration was used in the collection of data shown in Figs. 6, 8, and 9 and is most useful for the intercomparison of net samples and LOPC measurements collected simultaneously.

The Moving Vessel Profiler (MVP; Herman et al., 1998) was developed at the Bedford Institute to provide high speed sampling capability with vessels operating at their maximum speeds of 12-14 kts. A towed multi-sensor fish (shown in Fig. 13) is released from the stern of a ship and "free falls" vertically (3.5 m/sec.) to a maximum preselected depth of 300 m while the host vessel is steaming at 12-14 kts. Once the descent of the fish is halted, it is recovered by a multi-conductor cable to a position near the vessel=s stern and re-released into its "free-fall" mode once again. While operating in this mode, profiles to 300 m can be obtained indefinitely without stopping the vessel. The towed fish shown in Fig. 12 is instrumented with a small CTD, fluorometer and the LOPC.

The BATFISH vehicle towed in a sawtooth undulating pattern at 8 kts. was used to test a modified version of the LOPC adapted to measure more efficiently the euphausiid populations in the deep basins of the Scotian Shelf. By extending the beam path by 5X with a wider tunnel as shown by the outline of Fig. 1, the volume of water sampled has been increased substantially thereby improving sampling statistics. Such a modification now improves our capability of assessing more accurately populations of larger animals such as Calanus spp. and euphausiids by measuring their shape profiles as shown in Figs. 7(a,b). The wide-tunnel LOPC mounted of a BATFISH vehicle is shown in Fig. 14. This adaptation of the LOPC is not yet available commerically as it is still being tested and assessed on a number of issues, ie., i) differences in the lower detection limit as compared to the standard LOPC, ii) intercomparison of small plankton SEP size distributions with the standard LOPC and, iii) the validity and standardization of flow speed measurements.

The wide tunnel LOPC shown in Fig. 14 was also mounted on the mouth opening of the BIONESS as shown in Fig.15 towed at 2-3 kts. The BIONESS contains 10 opening and closing nets allowing discrete layer sampling and intercomparison to the LOPC measurements.

The advances and advantages of the new generation LOPC can be summarized by the following:

1. an substantial increase in the coincidence limit which enables the LOPC to be used in high plankton concentrations of up to 106 /m3.
2. a decrease in the lower detection limit to ~100 microns.
3. the provision of shape profiles of larger zooplankton sizes > 1.5 mm enabling improved taxa identification capability.
4. operation at higher towing speeds of up to 8 m/sec.
5. measurement of flow speeds within the tunnel.
6. overall smaller physical size.

With the onset of new data form the LOPC, we see new development work in data interpretation beginning to arise. For example, the algorithms used to reconstruct shape profiles as those shown in Fig. 7 need further development and data interpretation. Antennae and appendages of copepods shown in Fig. 7 are greatly overestimated in their thickness and, as a result, biomass determined from these shape profiles are also overestimated. With LOPC data somewhat analogous to video imaging, there is also the need for the development of image recognition software that would provide some taxa identification capabilities to the user. The LOPC shape images are, of course, only in 2-dimensions while animal orientation will result in a projected area and not the full shape. However there are certain geometrical features that can be used, for example, when considering the case of a ellipsoidal shape similar to that of copepods, the body length may not be properly oriented for a correct measurement but the body diameter will always be projected fully. Using algorithms with such boundary conditions would enable reasonable measurements and perhaps a statistical "best-guess" of the taxa.

Another area of data interpretation will be the measurement and identification of larger animals of a gelatinous or more transparent nature. Large zooplankton (MEPs) occlude many elements simultaneously and spend a much longer time traversing the LOPC light beam. Euphausiids, for example, tend to be reasonably opaque and occlude nearly 100% of the light beam during passage. However, many large plankton have also been measured yet result in measured signals indicating only a fraction of the light beam, eg. 15-40%, has been attenuated only and this is interpreted as a measurement of a large animals with significant transparencies. The ability to categorize taxa and their transparency potentially opens a whole new area of research and data interpretation with the LOPC.

There are other areas of LOPC data interpretation which are yet unresolved, particularly in the case of data sampled at the thermocline which often shows high plankton concentrations and also shapes of a "stringy" nature. The question that arises here is whether such particles measured at the thermocline are a result; i) of attenuance gradients causing false measurements or ii) large numbers of particulates which normally aggregate at such density gradients. This data interpretation remains ongoing.

The technology for the LOPC has been transferred to private industry and has been commercialized by Brooke Ocean technologies Ltd. (www.brooke-ocean.com). Currently only the standard tunnel version of LOPC is commercially available. Although the wide tunnel is more practical for sampling larger zooplankton, its intercomparability with the standard tunnel at the lower size ranges is still under data evaluation.