AIMAP -2010-G-05: Innovations in farmed halibut handling for improved production efficiencies

Final Report
March 31, 2011

Halibut PEI Inc.

Table of contents

• EXECUTIVE SUMMARY
• INTRODUCTION AND GOALS OF PROJECT
• A: PASSIVE GRADER
  • METHODS AND RESULTS
    1. Measurement of fish morphological characteristics for grading ring design
    2. Design of grading devices
    3. Incorporation of grading rings into modified crowding seine
    4. Test population selection
    5. Placement and adjustment of passive grader
    6. Determination of effect of grading on population characteristics
    7. Determination of effect of grading on population growth
  • DISCUSSION
• B: FISH TRANSPORTER
  • METHODS AND RESULTS
    1. Auto Cad virtual design of conveyor
    2. Design critique
    3. Prototype design and construction
    4. Prototype testing
    5. Construction of the commercial size unit
  • DISCUSSION
• CONCLUSIONS AND RECOMMENDATIONS
• ACKNOWLEDGEMENTS

EXECUTIVE SUMMARY

Halibut PEI Inc. grows halibut from 5g to market size in a modified lobster plant that is supplied with geothermally modified moderate salinity well water that maximizes the growth potential of this species without incurring heating or cooling costs. The company has been growing Atlantic halibut in this land based facility for three years.  This experience defined two problems with feedlot production of Atlantic halibut: namely variable growth rates within cohorts of the same size class and a long grow-out time from egg to plate. This project is one of many steps to address these problems by improving stock management, in this case through the application of innovative fish handling techniques.

The objective of the project was to develop improved (less stressful) fish handling mechanisms to allow gentle, frequent grading and low stress fish movement within facilities and between facilities. 

Halibut PEI Inc. collaborated with a manufacturer of passive grading systems for other species, EzeeGrader, to design a passive grader for flatfish to separate fish of different sizes in an attempt to optimize growth within the entire population.  The passive grader was effective at separating fish by size, as indicated by size distributions of graded and ungraded populations, and incurred minimum stress in the fish, as indicated by days off feed.

In conjunction with Atlantech Companies, Halibut PEI Inc. also designed and tested a prototype belt conveyor for halibut.  The testing results led to the design of a commercial scale conveyor that is effective at moving halibut without damaging the fish once they are on the belts.  Loading of the conveyor remains a key issue to increasing its efficiency and effectiveness. 

INTRODUCTION AND GOALS OF PROJECT

Current stocks of halibut produced for aquaculture are only one or two generations removed from wild populations.  Although a breeding program is in place for this species in Canada, it is still in its early stages and has not, as of yet, had a significant impact on the stocks produced. Use of such an undomesticated stock results in variable growth rates within cohorts of the same size class and a long grow-out time from egg to plate for slower growing fish.

The large variations in the size distribution of populations make frequent grading and movement of the fish between tanks imperative for proper stock management. Although mortality from these procedures is typically not an issue with current methods, loss of feeding days post handling can be an issue for larger fish (>750g) and results in extended time to market. Reducing the days to market is key for profitability in a grow-out operation. For every day to market that is reduced, several dollars are added to the bottom line of a halibut feedlot. That one factor contributes more to the margins involved in growing this fish than any other factor.

To our knowledge no one has developed a passive grading system for halibut and yet with the diverse rates of growth currently experienced, a successful adaptation of such a technology to this species can make a profound difference in the efficiency of moving the fastest growing fish to market. This should also disrupt the established social hierarchy in the tanks and promote growth of the remaining population. In the case of Halibut PEI Inc., passive grading would begin early in the life cycle and continue through to market. Fish concentrated on the big fish side of the passive grade would be moved into a tank with other fish from other tanks to ensure equality of size and moderate aggressive behaviour. This will require both a low stress (passive) grader and a low stress method of moving fish between tanks.

The objective of the project was to develop improved (less stressful) fish handling mechanisms to A) allow frequent, gentle grading and B) low stress fish movement within facilities and between facilities. 

In order to achieve the first part of the objective (frequent, gentle grading), we collaborated with EzeeGrader of Scotland and adapted an existing passive grading system used in salmonid cages for use in our land based system for halibut. The shallow concrete tanks and consistent temperatures of the MorningStar facility made it ideal for development and testing of the machinery as well as measuring growth characteristics in the halibut.

For achieving the second objective (low stress fish movement), we worked with engineers, fish culturalists, transporters and aquaculture equipment manufacturers to devise a belt driven fish conveyor system. This would allow fish in the shallow tanks to be crowded onto a soft, moist, covered conveying device to lift them out of the tanks, either to a weight table to be separated/counted, onto a similar device to be delivered to another culture tank or into a truck box for delivery to another site.

A: PASSIVE GRADER

The development of a passive grader for halibut was a collaborative project between Halibut PEI Inc. (HPEI) and Ezee Grader, a company belonging to Aqua Mar Ltd of Scotland (http://www.ezeeproducts.co.uk). 

METHODS AND RESULTS

1. Measurement of fish morphological characteristics for grading ring design

HPEI gathered morphological data necessary for designing a grading ring for halibut of a target size of 1.0 to 2.5kg. A measuring board was fabricated to ensure replicable measurements. Body thickness was determined anterior to the pectoral fin with an improvised but accurate homemade calliper device. 

2. Design of grading devices

Using the fish morphometric data supplied by HPEI, the grid design for the grading device was determined by EzeeGrader using their proprietary knowledge on fish space perception. A description of the design of the grid and its construction, as it appears on their website follows: 

“The device consists of a series of rigid grids made from injected Polyamide, circular in profile and with apertures of a predetermined width, which smaller fish swim through. The larger fish are harmlessly contained. The smooth flowing surface of the molded Polyamide grids, and the circular profile prevents any possible scaling or snagging of the fish passing through. This eliminates physical damage and stress to the fish, ensuring a higher quality end product for the farmer. The lightweight Polyamide material ensures that the product is sturdy and durable, while still being light and easy to use. The rigidity of the grid stops distortion of the apertures preventing larger, stronger fish from pushing through or the apertures collapsing to trap smaller fish.

Each rigid grid is connected to its neighbours by our unique loose connectors. Located on each side of the grid, they align with the neighbouring grid allowing the EzeeGrader to be made to any size (m2). These loose connectors give the whole grader flexibility allowing for easier handling during use, transportation and storage. “ (Excerpt from http://www.ezeeproducts.co.uk/ezeegrader_flat_fish_grading_system.html)

There were some setbacks encountered during manufacture of the plastic grading rings due to technical difficulties in the mould making. This caused a time delay in the fabrication and increased the cost of the product. After obtaining additional funding, the rings were produced and shipped to HPEI.

3. Incorporation of grading rings into modified crowding seine

The aluminum skeleton for a crowding seine, previously designed and developed for halibut by HPEI, was manufactured to provide the framework for the grading rings. These were installed by staff of HPEI along with Wilbert Johnson of Ezee Grader. The grading ring panel was 0.75m high by 5.0m wide and was installed in the center portion of the crowding seine. Sides and the bottom of the modified seine were constructed similar to the HPEI original seine design which allows movement of the seine through the water and accommodates the uneven bottoms and side walls typical of concrete tanks. The seine also allows extension of the sides to allow it to accommodate the varying widths of concrete tanks at HPEI.

4. Test population selection

In order to determine the effectiveness of the grader, an appropriately sized test population was required. Of the three populations available for use, the population with the highest mean weight (1kg) was selected. This was at the lower end of the range for grading of fish using the size of panel designed.

5. Placement and adjustment of passive grader

The passive grader was installed in the selected tank on December 1, 2010 using standard methods for crowder seine installation. It was placed 1/5th from the rear wall of the tank and all fish behind the grader were removed by hand dip-netting. Within minutes of securing the grader in the tank, fish were observed calmly swimming through the grading rings. A video of this can be seen at http://www.ezeeproducts.co.uk/ezeegrader_flat_fish_grading_system.html.

Distribution of the fish in the tank was noted on a daily basis. In a week, the density of fish on either side of the grader appeared about equal so the grader was moved through another 1/5th of the tank. Again, after a week, the density of fish on either side of the grader appeared about equal so the grader was moved through another 1/5th of the tank. After an additional week, it was evident that the fish were not going to redispurse themselves equally on either side of the grader so that the grader was no longer moved but remained at this position until a population distribution assessment was completed.

Because of the relatively rapid dispersion of fish through the grader, none of the expected required incentives to distribute the populations (e.g. water broom, aeration, selective population feeding) were used.

6. Determination of effect of grading on population characteristics

In order to determine the effect of the grader on the size distribution of the fish populations, 300 fish from each side of the grader were individually measured for weight. The size distributions clearly differed although the separation of fish was not as distinct as we had anticipated. 

7. Determination of effect of grading on population growth

It is too soon to tell the effect of the grader on population growth based on fish size. However, the effect of grader placement on fish feeding behaviour was observed. Feeding the fish before grader placement or move was avoided in order to ensure a high oxygen level in the tank during these manipulations.  Other than this voluntary reduction of feed at the discretion of the biologist, no change in feeding behaviour was observed with amounts fed and feeding aggression being consistent before and after grader placement and adjustment.

HPEI will continue to assess the effect of the grader on fish performance, including its long term effects on population growth. 

DISCUSSION

The grader was simple to assemble, structurally sound and easy to manipulate. Grader installation in the tank was also very straightforward and followed the standard protocols already in use at HPEI for installing crowder seines. Similarly, movement of the grader was met with no great challenges. 

Based on observations of the fish population, distribution across the grader took one week or less and movement of the grader could be conducted after this time period without significantly crowding the fish on the front of the grader. Of most importance was the fact that no incentives were required to coerce the fish through the rings. They simply placidly swam through with no physical damage apparent on any fish due to grader placement or subsequent passive grading.

Fish feeding behaviour did not decrease on the day of or the day after grader placement, or during subsequent moving of the device, indicating a minimal impact on fish stress levels and days to market. 

Separation of fish by size was apparent although it was not as distinct as we had anticipated. It was noted in the methods section that the population used for this trial was at the lower end of the range for which the grading device was intended, simply because of a lack of availability of larger fish. It is likely that an improved separation will result when this device is used in a population with a greater mean weight. This will be available at HPEI in upcoming months and the grader will be reassessed at that time.

Our overall impression of the grader is that it had great potential for use within our tank systems. The passive grader allowed the halibut to be crowded initially. All sizes of fish look for better “real estate” that is less competitive. The smaller fish see the opening in the rigid grader and swim directly through the opening. The larger fish swim up to the grader, put their rostrum through it and then back off to try again until they give up. It appears to be a natural response and shows no stress related behaviour. Those close to market size may attempt to squeeze through indicating their natural tendency to burrow under things as they also do with the pipes and other objects in the tanks. The only problem we see is that there are small fish in the remaining part of the tank that refuse to move. There are only a few that are easily noticed by size and can be captured and removed by net.

The effect of grading on time to market is yet to be determined and will require repeat use of the grader in numerous populations.

B: FISH TRANSPORTER

The live fish conveyor prototypes were developed in a multi step process in association with Atlantech Companies. Their description as well as an image gallery and video of the product can be seen on their web site: http://www.atlantech.ca/public/project_aquaculture.html.

METHODS AND RESULTS

1. Auto Cad virtual design of conveyor

Ample discussions on the special biological considerations for the conveyor initiated the design of the conveyor. Several site visits were conducted by the project manager of Atlantech Companies in order to assess physical space constraints within the building and with respect to the live fish hauling methods used. A 3D model of a live fish conveyor was built on Solidworks software in order to conceptualize the discussion. This was reviewed in full by all involved parties.

Materials construction was also of paramount importance in order to ensure comfort of the fish and usability of the final unit. Samples of potential belt and flight components were proposed and samples were obtained for examination.

2. Design critique

Before proceeding with construction, the project manager from Atlantech presented a review of the project to date (Appendix A). The areas examined included a restatement of the design objectives, the design criteria, the design considerations and assumptions, and the current design direction. This was subsequently reviewed and critiqued. 

3. Prototype design and construction

A prototype was proposed and priced. The length of the unit was increased from 1.2m to 2.2m in order to make sure rigorous testing of inclines could occur. The width of 0.6m, as planned, was considered adequate. Materials selection was completed. It was decided to use the prototype to test two different types of belt top coverings. These coverings had to ensure sufficient friction to hold the fish in place,  rapid dewatering  and yet be gentle enough to not cause damage to the fish, despite their expected constant movement on the belt. After agreement on all factors, a small prototype was constructed.

4. Prototype testing

The prototype was tested by setting it in a 1.8m X 3.7m shallow (0.46 cm depth) fibreglass tank which served as both the delivery and recipient tank for the fish. For the initial test of the conveyor, fourteen fish, averaging 1100g, were repeatedly moved along the conveyor at four different speeds and three inclines. Staff of Atlantech and Halibut PEI assisted with this assessment. The height of the flights (5cm) was deemed insufficient since the fish sometimes skipped over them, so the prototype was sent back to the manufacturer for modification. 

For the second test using higher flights (7.5cm), fish of three size classes, representing the broadest size range available at the facility, were selected. Nine to twelve fish of each of three size classes (mean weights of 518g, 1040g and 1383g) were used for the testing. The fish moved along the conveyor well regardless of the speed or incline of the prototype. However, getting the fish onto the conveyor belt was a challenge.  Fish introduced onto the belt above the waterline, whether individually or en masse, were easily transported to the top of the conveyor.  Fish introduced to the conveyor below the waterline, however, would repeatedly try to avoid the belt, swim off and away from the conveyor. A flatter surface beneath the water was suggested in order to aid the loading of the conveyor.

After each of the trials, the fish that underwent repeated transports on the conveyor were sequestered in a shallow fibreglass tank for several weeks following the trials in order to closely observe the fish for signs of damage. No physical damage was evident on any of the fish and feeding resumed within a day or two.

5. Construction of the commercial size unit

After additional discussion and consultation, the commercial size unit was designed. Conceptual designs and costing were presented by Atlantech to Halibut PEI.
To accommodate the increased size and need for a flatter surface in the water, the transporter became longer and with the additional hydraulics, much heavier. The motors and controls were placed in a consul that is on wheels; and the conveyer platform extends out from this base unit which is on wheels. The conveyer itself has double positioning hydraulics that allows a more horizontal portion to get the fish onto the belt as well as the capability to fold the conveyer for transport so that it fits through all our doorways and corners. The length of the conveyor unit, when open is 5.4m, with a width of 0.66m. The Intralox Series 900 Square Friction Top was chosen for the belt as it was assumed to have a greater ability to cushion the fish. The belt speed is adjustable from 0 to 0.3m/sec.

The conveyor was delivered to Halibut PEI for testing. Upon receipt, it became obvious that moving the unit through the building would be difficult. Removal of side components (motor, control panel) is necessary to do so.  

The unit was installed in one of the concrete tanks at Halibut PEI for the initial operational testing by the staff of HPEI and Atlantech. No significant issues were encountered with the mechanical operation of the unit. Slightly larger fish (mean size of 1200g) were available for this testing.

As with the prototype conveyor, the fish moved well along the belts once they were out of the water. They did not flip off the conveyor and no physical damage was evident from the handling.  However, getting the fish onto the conveyor was a huge challenge as they clearly avoided the unit and swam off of the belt before being lifted out of the water. 

DISCUSSION

The conveyor is relatively simple to operate once in position but requires care when moving the unit through the building. This is not unexpected however, due to the physical constraints of pipes running along floors and overhead and tight packing of fibreglass tanks in the nursery.  

The conveyor does successfully move fish from one point to another with no physical damage evident on the fish. This was true even for fish that were moved on the prototype conveyor repeatedly (5 times or more). The belt itself creates a platform that is soft and smooth enough to provide a non damaging surface even when the halibut are thrashing out of water. There is sufficient water cushion retained for the entire transport time to minimize any injuries to fins, and eyes. Return to feed was quite rapid (one to two days) for the fish moved on the prototype conveyor. A good range of sizes of fish were used for the testing with no difference noted in the ability of the conveyor to move the fish. The conveyor clearly meets the biological needs for moving halibut.

Loading the fish onto the conveyor in a low stress, efficient way remains a key issue. There is reluctance of the halibut to use the conveyor, which, of course, is expected. Moving the fish toward the conveyor is difficult with their tendency to stay on the bottom. When prompted or forced to move they initially resist the slope of the moving conveyor. With the final prototype that surface has been made to be adjustable to be less of slope making their approach to the conveyor easier. It is also anticipated to make it more expedient for a person in the tank to net them onto the belt when necessary. Until the issue of loading the conveyor is adequately addressed however, the speed of fish move (kg/hour) and manpower required for operation cannot be accurately assessed.

The original design called for the conveyer to be covered by a clear cover to prevent fish from jumping over the sides. This was not utilized in this prototype but it may be required as we refine the techniques of moving fish onto the conveyor. The conveyer speed is adjustable so that as techniques are refined we will be able to determine the optimal loading rate to provide the maximum number of fish moved per hour that is least stressful for the fish.

The final modifications and testing will continue as the prototype is used to move fish before and after grading. Procedural manuals have been developed and will continue to be edited as experience is gained with the unit.

CONCLUSIONS AND RECOMMENDATIONS

The passive grader developed within this project was effective at separating fish by size, as indicated by size distributions of graded and ungraded populations, and incurred minimum stress in the fish, as indicated by days off feed. It is likely that an improved separation will result when this device is used in a population with a greater mean weight. This will be available at HPEI in upcoming months and the grader will be reassessed at that time. The only issue we noted was that there are small fish in the ungraded part of the tank that refuse to move. The effect of grading on time to market is yet to be determined and will require repeat use of the grader in numerous populations.

The manufacturing of the prototype commercial scale conveyor was completed. It proved to be effective at moving halibut without damaging the fish once they are on the belts. Loading of the conveyor remains a key issue to increasing its efficiency and effectiveness. Atlantech and HPEI will work together on optimization of the machine and the behavioural biology of moving halibut. The next steps in adapting this technology will be to develop an effective loading system for the halibut followed by its use it for a variety of fish movements including transfers between tanks, weighing for market selection and determining growth ranges within tanks.

ACKNOWLEDGEMENTS

HPEI would like to express its thanks to AIMAP and especially to the people within the program that we work with. We expect that both projects will help move the halibut industry in Atlantic Canada ahead with more efficient procedures that are less stressful and fully adapted to Atlantic halibut.