Kartlegging og testing av metodikk for telling av lakselus og beregning av luseforekomst

Solberg, Ingrid and Finstad, Bengt and Berntsen, Henrik Hårdensson and Diserud, Ola H. and Frank, Kevin and Helgessen, Kari Olli and Jeong, Jaewoon and Kristoffersen, Anja Bråthen and Nytrø, Ane Vigdisdatter and Revie, Crawford W. and Sivertsgård, Rolf and Solvang, Torfinn and Sunde, Leif Magne and Thorvaldsen, Trine and Tor Atle Mo, Ingebrigt Uglem og (2018) Kartlegging og testing av metodikk for telling av lakselus og beregning av luseforekomst. Norwegian Institute for Natural Research, Trondheim, Norway.

[thumbnail of Solberg-etal-2018-Kartlegging-og-testing-av-metodikk-for-telling-av-lakselus-og-beregning]
Preview
 Text. Filename: Solberg_etal_2018_Kartlegging_og_testing_av_metodikk_for_telling_av_lakselus_og_beregning.pdf
Final Published Version
Download (808kB)| Preview

Abstract

All Norwegian fish farms are required to count and report the presence of sea lice on salmon in accordance with the regulations on the control of sea lice in aquaculture farms also known as the sea lice regulation. The registration of lice incidence in the fish farms is used both to assess whether measures are necessary and to assess the impact on wild salmonids. It is therefore very important to have reliable figures on the number of sea lice in fish farms. In addition to a minimum number of fish to be counted in individual net pens (10-20 fish), the current regulations for the implementation of lice counting on farmed fish also allow different methods for catching and the out-take of fish and counting lice on these fish. Since different catch and counting methods can cause a different degree of uncertainty as to whether the fish is a random selection from the net pens and if all lice on the fish are recorded, it can be difficult to directly compare the lice occurrence between sites and production areas. It is therefore important to know how lice counting practices vary among farms and how this may affect the precision of the counting results. To ensure that selection and results from lice counting are as representative as possible and that reporting from different sites and actors becomes comparable, it is therefore very important to know the characteristics of the different methods used. In this project we have conducted a survey of today's methods for counting sea lice in the Norwegian salmon industry through an interview survey. The results of this study confirm that there are different interpretations of the regulations and different implementations of lice counting, but also that the counting of lice on farmed fish is generally carried out in a thoughtful and systematic manner. We also conducted a field survey in which farmed salmon were taken with large dip-nets (orkastnot) at different capture time and time of day, where lice were counted after one protocol used for regular lice counting at the net-pen and in a more thorough control counting. Capture time and time of day turned out not to affect the counting results, and the results from the ordinary counting and control counting were in general almost the same. However, there was a deviation between the counting results of the ordinary counting and control counting at a relatively low number of fish, and individual lice numbers could be both higher and lower when using the ordinary counting. Very few sexually mature female lice were overlooked in the ordinary counting when compared to the control. The weather conditions were exceptionally good during the field trial, so it may have been the case that any variation due to environmental conditions or human factors became too small to allow for relationships associated with the different counting methods to be modeled. In addition, we looked at various issues solely related to the statistical residual uncertainty ("sampling uncertainty") which will always be present in a counting results based on the relatively small selection from these salmon populations; given that the lice are not evenly distributed among the fish in the farm. We have shown how this statistical uncertainty can be calculated, identified factors that affect this and come up with suggestions on how it can be handled by the government in the context of the set lice limits. It is here, among other things, proposed a single action system for how one can use multiple lice counts in a row to minimize the chance of failing to conclude an overrun of the lice limit ("false alarm"). A system that unilaterally focuses on minimizing the chance of a "false alarm" will at the same time increase the chance of wrongly concluding that the limit is not exceeded, when it is ("false acquittal"). How one chooses to balance the possible consequences of these two types of errors will be crucial for assessing whether measures are necessary. In addition, we have explored the calculation of lice occurrence and proposed an approach for calculating the actual release of infectious sea lice larvae from aquaculture farms where lice biology and water temperatures are considered. In this project we have investigated some factors that may affect lice counting results, and our results indicate that these factors will not significantly increase uncertainty in the estimated lice incidences, under the prevailing lice density and environmental conditions that prevailed while conducting the investigations. However, there are a number of other factors that may lead to uncertainty in estimates around the occurrence of lice in salmon farms, and there is a need to gain additional knowledge about these in order to assess the need for suggesting changes to any revised lice counting standard. To obtain good estimates of lice occurrence, it is critical that the sampling methods provide representative selection and that the counting methods do not introduce systematic errors in the results. In this report we have discussed and summarized elements that may be included in an industry standard for lice counting, if establishing such a standard is deemed necessary. Some of these elements are already standardized in the sea lice regulation, while other factors require more knowledge. Development of a standard counting method involves synthesis of all types of knowledge, both research and experience-based, thus requiring interactive dialogue with the industry.

CORE (COnnecting REpositories)