Full Scale Testing of a Concept for Salinity Regulation to Mitigate Sea Lice Infestation in Salmon Farming

Abstract

The large environmental and economic impact of sea lice infestation in the salmon industry has encouraged the development of non-medical methods and preventive strategies to combat sea lice infestation. Sea lice (Lepeophtheirus salmonis and Caligus elongatus) are sensitive to low salinities, and using fresh water as protection against infection may thus significantly reduce sea lice infestation of salmon while reducing the costs and impacts of traditional delousing methods. A new concept presented here is based on the manipulation of salinity within cages by adding fresh water to create an unfavourable environment for sea lice infestation. A full-scale set-up was tested in a salmon farm in northern Norway: two commercial-size cages with Atlantic salmon (Salmo salar) were enclosed with a 2 m deep tarpaulin skirt and supplied with fresh water at the centre to establish a surface layer with reduced salinity. Two reference cages had no skirt or fresh water supply. Time series of CTD-data showed that the fresh water supply caused a shallow and unstable salinity gradient, with salinities lower than 10 ppt measured for short periods in the upper 0.5 m. Despite these instabilities, significantly lower sea lice infestation in cages supplied with fresh water was observed, as infestation rates for pre-adult and adult stages of L. salmonis were reduced by 48% and 57%, respectively, in the treatment cages compared to controls. This preventive strategy is therefore very promising and deserves further development under more stable and controlled conditions. Future studies should focus on improving freshwater regulation, ensuring higher spatial resolution of salinity data in surface layers and documenting the effect on the more salinity-sensitive planktonic stages of L. salmonis. In addition, there is a need to examine the effectiveness of the technique at multiple sites and under a wide range of site conditions, especially various current rates through the site.

1. Introduction

Salmon lice (Lepeophtheirus salmonis) are a major pest of wild and farmed salmon and represent a significant threat to the fish farming industry [1,2]. Although this is mainly due to the particularly serious impact of the species L. salmonis, members of the genus Caligus are also implicated including Caligus elongatus. Together they have achieved notoriety by having the greatest economic impact of any group of parasites in salmonid fish mariculture [3] and have become collectively known as “sea lice”. For the Norwegian salmon industry, combating sea lice is one of the biggest challenges and has been estimated to cost the salmon farming industry in Norway around 9% of its income [3]. Beyond the direct treatment cost are the negative impacts on fish welfare and production outcomes [1,2,4]. The use of cleaner fish [5,6,7] is probably the most common non-medical method (NMM) not involving handling of salmon. Examples of new farm adaptations or technical concepts to mitigate sea lice infection are closed cages [8], cages semi-enclosed by skirts [9], the “Snorkel” sea lice barrier [10,11,12] and the use of plankton nets [13]. The main principle of many of these concepts is to create a barrier between the sea lice and the salmon to prevent lice infestation, with a minimum of fish handling. The most common use of NMMs to control sea lice today are exposure to fresh water, thermal and mechanical methods, where salmon are exposed to low salinities or high temperatures that are harmful to the sea lice or brushes and flushing that mechanically remove the sea lice [14,15].

The use of freshwater treatment as a delousing method has increased in recent years [14,15]. Since 2016 there has also been an increase in freshwater treatment in combination with thermal and mechanical delousing in Norway [14]. Two types of freshwater treatments are commonly used. One uses a snorkel pen where the salmon must pass through a freshwater layer when they go up to the surface for air to fill the swim bladder. A second method is freshwater treatments undertaken in well boats, and this has become a promising delousing method which the industry uses [14,16,17,18,19]. Earlier studies using freshwater treatments [14,15] have shown promising results for combating sea lice on farmed Atlantic salmon. The study of Reynolds [15] was a field study where 5200 Atlantic salmon with a mean weight of 4.4 kg were distributed into two wells (one with fresh water (0.16 ppt) and one with seawater (33.7 ppt)). The average number of attached L. salmonis was compared in both groups to a pre-count undertaken prior to the study. Salmon exposed to seawater had percentage reductions in the average number of chalimus, pre-adults and mature females of 47%, 59% and 31%, respectively, compared to the pre-count. The overall reduction of all stages of infective lice was calculated to be 51%. Salmon exposed to fresh water had higher percentage reductions in the average number of chalimus, pre-adults and mature females of 100%, 97% and 92%, respectively, compared to the pre-count. The overall reduction of all stages of infective lice was calculated to be 96%. This indicates that the freshwater treatment itself lowers the overall lice burden by at least 45%. The study of Guttu et al. [14] was also a field study that was conducted on a specially designed well boat for delousing which serves salmon farms to investigate the fate of sea lice on salmon during freshwater treatment. Crowded salmon were pumped onboard the well boat and treated with fresh water for 10 to 12 h. The authors found a pronounced variability in delousing effect for full-scale treatments by exposure to fresh water between stages of L. salmonis and C. elongatus. Adult males and females (without egg strings) of L. salmonis needed longer treatment time in fresh water than the earlier developmental stages, and adults showed a somewhat lower total reduction during freshwater treatment. The rapid reduction of lice numbers for L. salmonis and C. elongatus from the count made during the crowding of the salmon to the first count made during freshwater treatment was suggested to be a result of a combination of mechanical actions during crowding and freshwater exposure.

Barrett et al. [16] reviewed a range of potential and existing preventative methods for sea lice. They found that barrier technologies that minimise host–parasite encounter rates provide the greatest protection against lice, with a weighted median 76% reduction in infestation density in cages with plankton mesh ‘snorkels’ or ‘skirts’, and up to a 100% reduction for fully enclosed cages. Other methods such as geographic spatiotemporal management, manipulation of swimming depth, functional feeds, repellents and host cue masking can drive smaller reductions that may be additive when used in combination with barrier technologies.

The life cycle of L. salmonis involves eight stages [20], where the first two nauplii stages (nauplii I and II) are followed by the infective copepodid stage and five parasitic stages which include two chalimus stages, two pre-adult stages and the adult stage. Infections occur when the free-living copepodids attach to a host and develop into the chalimus stage. Water temperature is a key factor regulating the development time and infestation success [21], while salinity has a non-linear effect on survival [17]. Salinity may also influence development, behaviour and infestation rate [22,23], where the free-swimming nauplii and the infective copepodid stages are the most sensitive. Salinity tolerance varies between the different stages. While adult L. salmonis detached from a host reach 13–44% mortality after 4 h freshwater exposure and 44–77% mortality after 24 h [24], free-swimming nauplii exposed to 12 ppt salinity will no longer be able to osmoregulate and will die [22]. Those attached to host adults are less susceptible to reduced salinity, probably due to ingestion of hyperosmotic substances from the host [25]. Tests of salinity preference in nauplii and copepodids in columns with a salinity gradient show that both nauplii and copepodids prefer full seawater. While the nauplii avoid the brackish water layer to a large extent, the copepodids show a more gradual avoidance of the reduced salinity [26]. Bricknell et al. [22] concluded that survival, host infectivity and the ability of copepodids to remain attached to a host were compromised by short-term exposure below 29 ppt, and that the copepodids altered the swimming behavior and avoided salinities below 27 ppt. After the infective copepodid molts into a chalimus stage and further developed stages attached to the salmon, the tolerance for reduced salinity improves [27,28].

The sensitivity to reduced salinity is also reflected in the variation in infection pressure at salmon production sites in a Norwegian fjord system [28], where the innermost areas of low salinity in the surface layer show reduced L. salmonis abundance compared to the outermost sites with higher salinity. This is in line with models based on data from wild salmon in British Columbia collected over 10 years, where salinity was found to be positively associated with an increased probability of sea lice infection and intensity [29]. As the copepodids have a natural preference for the surface layer [30] and low freshwater tolerance, their infective ability is probably compromised under conditions typical for fresh-water-influenced fjord systems. Hence, a potential preventative strategy to mitigate sea lice infestation targeting the free-swimming larval stages, particularly the infective copepodids, should be aimed at modifying the surface environment for salmon aquaculture sites or within sea cages, mimicking the characteristics of a fresh-water-influenced fjord system.

A simple concept for establishing a thin brackish water layer within salmon cages was designed to create a constant osmotic stress that reduces the fitness of the L. salmonis larvae seeking the surface. The idea was to take advantage of their positive phototactic response and vulnerability to low salinity to reduce sea lice infestation. The set-up is based on a continuous flow of fresh water into the centre of the cages supplied by gravity or pumping via a small pipeline from a nearby river, and a tarpaulin wrapped around the 2 m upper depth of the cage to trap the fresh water (Figure 1). A hydrological study has shown that it is possible to establish such a water supply with sufficient water from nearby rivers based on known technology at an acceptable cost for more than 90% of the salmon sites in Troms (Norway). Probably this is also the case for a large part of coastal Norway. In contrast to many other designs providing a prophylactic environment or shelter, the design is simple and cost-efficient, does not compromise the oxygen conditions due to the shallow depth, provides a permanent salinity gradient based on a low water supply, covers the whole cage surface and may easily be adapted to enclose a whole salmon site.

The aim of this study was to study the effects of lowered salinity in full-scale salmon farming on sea lice infestation.

2. Materials and Methods

2.1. Experimental Set-Up

The commercial trial was conducted within a large-scale set-up of 11 salmon cages of 140 m circumference at the location of Uløybukt in North Troms, Norway. All cages were wrapped with a 10 m deep tarpaulin beginning at the start-up of production in 2017, which was based on spring- and autumn-released smolts. The salmon in the experimental cages and reference cages were stocked into the cages in May and June 2017 with 102,000–127,000 fish per cage averaging from 72 to 102 g. To prepare for the experiment, the tarpaulin skirt was removed from the four experimental cages on 31 May 2018. On 4 June, the two freshwater test cages (cages 3 and 4) were semi-enclosed by a 2 m deep tarpaulin wrapped around the upper part of the cages, while the two reference cages (cages 6 and 7) remained without tarpaulin and fresh water (Figure 2). The fish were at this stage averaged from 1.3 to 1.5 kg. From 4 July, the two experimental cages were supplied with fresh water from a nearby river at approx. 20 L per sec. to the sea surface in the centre of the cages through a 180 mm PE pipeline. On 12 July, a 14 m wide, T-shaped perforated strainer pipe with 62 mm openings was connected to the end of the pipeline to ensure a gentle water distribution at the sea surface. Salinity, temperature and water current were monitored continuously for 168 days from 12 June until termination of the trial on 28 November 2018.

2.2. Monitoring of the Environmental Conditions

Environmental measurements were primarily conducted to survey the development of the hydrography in both experimental groups. Due to logistic reasons, the environmental measurements were taken in two cages: in one of the test cages with fresh water supply (cage 3) to compare with control measurements in a control cage without fresh water supply (cage 6), and outside the cages at the NE corner of the fish farm. For this purpose, two types of CTDs were used: (1) a handheld CTD to provide daily salinity, temperature and density profiles with high vertical resolution, and (2) several CTD-SRDL units that logged continuously at 10 min intervals that provided time series at fixed depths. With the combination of these, the effect of the fresh water supply on the salinity gradient can be studied in detail.

The handheld CTD was an SD204 from SAIV AS (www.saiv.no, accessed on 4 April 2025) configured to sample at a rate of 1 Hz. Daily measurements were taken at 4 measurement stations within the fish farm by manually lowering the CTD to about 10–20 m and slowly pulling it up to assure high resolution profiles of the upper layers. Two of the stations were within the test cage (cage 3, centre and close to edge), one station was in the centre of cage 6 (no fresh water supply), and one station was at the NE corner of the fish farm.

CTD-SRDLs are small CTD units that were originally developed at the University of St. Andrews to be attached to seals to gather deep profiles along the paths of the seals (http://www.meop.net/meop-portal/ctd-srdl-technology.html, accessed on 4 April 2025). For this project, several CTD-SRDL units were connected in series to provide time-series data at 10 min resolution at several depths simultaneously. Each chain of units was connected to a battery pack above the surface. At the location, a total of 13 CTD-SRDL units were used, divided into three chains of units connected in series: two chains in a test cage (cage 3, centre and edge of cage) and one chain in a control cage (cage 6), as indicated in Figure 2. An overview of the two chains of CTD-SRDL units and the logging depths can be seen in

Table 1. Overview of CTD-SRDL units and depths. Chain 1 was located close to the centre of a test cage (cage 3), chain 2 was attached to the outer ring (just within cage 3), and chain 3 was located within a control cage (cage 6).

CTD-SRDL Unit	Chain Number	Depth (m) 6 June–23 August	New Depth (m) 23 August–2 November
12,857	1	0.5	0.14
12,887	1	1	0.37
12,888	1	1.5	0.60
12,882	1	2.0	0.84
12,885	1	2.5	1.07
12,881	1	3.5	2.07
12,886	1	5	3.57
12,883	2	0.5	0.1
12,856	2	1	0.6
13,069	2	2.5	2.1
13,067	3	1.0	0.1
13,071	3	2.5	1.6
12,884	3	3.5	2.6

. As the layer of brackish, low-salinity water turned out to be quite shallow, the depth of the CTD-SRDL chains was adjusted after the first logging period (Table 1).

In addition to hydrography, current measurements were undertaken using an acoustic Doppler current profiler (ADCP) of the Seaguard I type (Aanderaa Data Instruments AS, Sanddalsringen 5b, N-5225 Nesttun, Bergen, Norway). This was mounted on the NE corner of the fish farm at 16 m of depth, logging current velocity and direction at one point every 10 min.

2.3. Registration of Sea Lice Infestation

With few exceptions where only 10 fish were sampled, sea lice were counted on 20 randomly selected fish per cage on a weekly basis from week 22 (31 May) to week 40 (30 September) for a total of 19 times. Sea-lice counts were carried out according to national regulations by trained technicians. The population of caged fish was starved on the day of sampling prior to the following procedure: a 5 m dip net was introduced next to the cage net wall and lowered a few meters. Salmon were attracted to the net by hand feeding, followed by dip-netting a small number of fish. Crowding time was kept to a minimum, as crowding time may influence loss of lice and hence lead to inaccurate lice counting. A maximum of 5 fish at a time were transferred to a holding tank and anaesthetized (Benzoak 20 mL/100 L) before lice counting. Sea-lice counting was performed under good light conditions (daylight). Sea lice were registered and grouped in the following categories: “L. salmonis chalimus”, “L. salmonis preadult”, “L. salmonis mature females” and Caligus elongatus. Any lice remaining in the water containing the anaesthetic, predominantly of mobile stages, were included in the recordings. When counts were completed, individual fish were moved to a flow-through water tank and returned to the cage upon recovery (retaining swimming behaviour). Data from lice counts are presented as the mean abundance of each stage in the respective weeks.

2.4. Statistical Methods

The statistical analyses of lice data were conducted using the Statistica™ 14.1 software. A Kolmogorov–Smirnov test [31] was used to assess the normality of distributions. The homogeneity of variances was tested using the Levene’s F test [31]. Possible differences in lice infections between treatments at each sampling time (week) were tested in a one-way ANOVA. Significant differences revealed in the one-way ANOVA were followed by a Student–Newman–Keuls (SNK) post hoc test to determine differences among experimental groups. A significance level (α) of 0.05 was used if not stated otherwise.

3. Results

3.1. Hydrographic Conditions

The salinity from daily profiles from the handheld CTD profiler is displayed in Hovmöller diagrams in Figure 3 (cage 3) and Figure 4 (cage 6). With time along the x-axis and depth along the y-axis, these figures give an overview over how the freshwater input affects cage 3, compared with the control cage with no freshwater influx (cage 6). Clearly, the freshwater influx affected the salinity in cage 3, but the effect was restricted to a shallow upper layer, and there were large fluctuations in how large the effect was. The current measurements (Figure 5) indicated that the fish farm is located in an area with low current speed, with a mean velocity of 4–5 cm/s at 16 m depth with a dominant current direction from the southeast. Wind measurements were not recorded directly at the experimental site. However, data were available from the nearest meteorological station at Sørkjosen Airport, located approximately 12 km southeast of the site. These data indicate an average wind speed of 3.35 m/s and a maximum wind speed of 15.3 m/s, suggesting that conditions were generally calm during the experiment with only short periods of wind exceeding 10 m/s. Furthermore, there was no apparent correlation between the depth of the brackish layer in the test cage and the current velocity, so the large variation in the fresh water’s effect was probably in large part due to challenges with the freshwater supply.

From the salinity time series in Figure 3 (handheld CTD), there are three periods with a more pronounced effect of the freshwater supply: (1) 7 July–20 July, (2) 5 August–14 August and (3) 17 August–23 August. The time series from the CTD-SRDL units gave a more detailed picture on the development, and the uppermost measurements from the first period inside the test cage are shown in Figure 6. For only short periods, the salinity in the upper layer was below 10 psu (ppt). The salinity in the centre and at the edge of the test cage follows the same trend and does not differ much, showing that for the periods with sufficient and stable freshwater supply, a shallow upper layer of brackish water can be maintained in the entire cage. Because the uppermost CTD-SRDL units originally were located at depths of 0.5 m, some of the variability in the time series was due to the brackish water layer in some periods being shallower than that.

There was no clear correlation between the current velocity that was measured and the depth of the brackish layer in the test cage, but because the freshwater supply was very variable, any connection is hard to find. The freshwater supply was very unstable due to several technical issues in the starting process as well as several unforeseen power shortages during the whole period.

3.2. L. salmonis and C. elongatus Infestation

Between week 22 and 30 no mature female sea lice were found in either experimental group (Figure 7). Mature females of L. salmonis were observed in both the test group and the reference groups starting from week 30 (Figure 7). The infestation rates were significantly higher at weeks 36 and 37 (one-way ANOVA, p < 0.05). Due to the increased infestation rate, both experimental groups were deloused at the end of week 37 through the application of the Thermolicer method [32]. There were indications of re-infestation from untreated cages at the location, and the four experimental cages were treated again with the Thermolicer in week 41. Due to the interference these treatments might cause in the experimental design, the analyses of lice infestation rates were restricted to the period from week 22 (start of the experiment) to week 37, when the lice counting was not interfered with by treatment or other operations at the site.

For the pre-adult stages of L. salmonis, lower infestation rates were found in the test cages compared to the reference cages in weeks 36 and 37 (Figure 8b, p < 0.001, F = 5.48), whereas no differences were seen for the chalimus stage of L. salmonis (Figure 8a, p > 0.55).

For C. elongatus, lower infestation rates were found in the test group compared to the reference sea cages in weeks 34, 36 and 38 (Figure 9 p < 0.001, F = 5.38).

4. Discussion

4.1. Hydrographic Conditions

The hydrographic measurements showed that with the methods described in this paper it is possible to establish and maintain a brackish layer in the uppermost layer of a cage, with the main challenges being a sufficient and steady supply of fresh water, as freshwater pulses occurred but were quickly diluted and were not sustained for long time. Another challenge is to measure the salinity with high enough vertical resolution to fully capture the brackish layer. Even with salinity loggers at 0.5 m, the layer was at times not detected and could only be seen in the CTD profiles.

4.2. L. salmonis Infestation

The present study showed that freshwater treatment as a delousing method shows promise as a non-medical treatment for reducing the numbers of both L. salmonis and C. elongatus on farmed salmon, in agreement with findings in earlier studies [14,15]. For the whole period from week 22 to 37, the infestation rate of pre-adult L. salmonis was 48% lower in the freshwater group, mature females were 57% lower, and C. elongatus was 31% lower compared to the reference group. On the other hand, for the chalimus stage the infestation rate was 34% higher in the experimental group, likely due to the difficulty in accurately counting at this stage. The overall picture is a reduced infestation pressure of L. salmonis in cages with tarpaulin skirts and freshwater supply. The chalimus stage is known to have lower tolerance for fresh water compared to the pre-adult and adult stages [33] and would be expected to be influenced by the reduced salinity in the experimental cages. The smaller size of the chalimii, making them more difficult to detect, may have led to inaccurate counting. The tested concept with fresh water influenced cages for mitigating sea lice infestation was targeting the free-swimming L. salmonis larvae, which are the most sensitive to fresh water [11,22]. The larvae are also positively phototactic and seek the sea surface [34], where reduced salinity requires higher energy use to maintain their preferred position [22].

Moreover, in addition to the anticipated effect of low salinity, the tested concept should not ignore a possible shielding effect of the skirt or a combined effect of fresh water and shielding. Previously it has been reported that cages with 10 m deep skirts [9] or 6 or 10 m deep plankton nets [13] in normal seawater reduce L. salmonis infestation. Based on experiments with “snorkel” barriers where deeper barriers were more effective [10], it is expected that the shielding effect increases with deeper shields or barriers. It is therefore possible that the combination of the tarpaulin skirt and the reduced salinity could lead to a combined effect of those two methods leading to the overall effect found in this study. Additional control cages with only the tarpaulin skirt and no fresh water would have been ideal, but this was not possible due to practical (and production) constrains at the experimental site. Earlier trials have clearly indicated that freshwater treatment is an effective non-medical method in reducing the numbers of both L. salmonis and C. elongatus on Atlantic salmon [14,15,35], in line with present findings. Combined, these earlier studies [14,15,35] may indicate that treatment using concentrations of lower than 4 ppt, or full fresh water, is necessary to significantly reduce the survival of L. salmonis and C. elongatus, and freshwater treatment may lower the overall lice burden by 45% or more depending on the length of the treatment. Recent findings [36,37] have shed light on the salinity tolerance of both L. salmonis and C. elongatus. Borchel et al. [36] found that a 3 h treatment with fresh water had a detrimental effect on the egg strings of L. salmonis. First, the water penetrated the string, widening it, then entered the eggs and enlarged them. Further, Borchel and Nilsen [37] revealed that detached adult C. elongatus exhibit low tolerance to reduced salinity, with mortality occurring within hours at salinities below 20 ppt. In the present study we applied a brackish water layer of around 10–15 ppt, which may lend credence to the possibility that the reduction of L. salmonis and C. elongatus seen in our data is a combination of tarpaulin and low salinity effect. Salmon need to access the surface to fill their swim bladder [38], so passing through a brackish water layer (as in the present study) provides a potential self-treatment for the salmon.

4.3. Experimental Set-Up Considerations

The arrangement of the treatment and control cages within the farm site in present study has weaknesses that we acknowledge. Due to practical constraints on site, we had to co-locate the test cages in one part of the farm (cages 3 and 4) and the control cages in another part of the farm (cages 6 and 7). Interspersion of test and control cages would have been better, as it reduces any possibility of spatial differences in infection levels confounding the results. Cages within farms often experience very different infection levels, due to local hydrodynamics. However, our hydrodynamics measurements did not reveal any major differences in the local hydrodynamics between these cages, which strengthens the reliability of the current data, though we cannot rule out the possibility that the results obtained are wholly or partly due to different locations within the farm.

5. Conclusions

This study demonstrated a novel, non-invasive technological approach to reducing sea lice infestation by manipulating salinity within salmon cages through continuous freshwater supply. The concept, combining a shallow 2 m tarpaulin skirt and centrally supplied fresh water, established an intermediate brackish surface layer (10–15 ppt) with no observed operational interference or adverse effects on fish welfare. Unlike many other prophylactic methods, this technique offers a low-cost and passive treatment alternative that can be seamlessly integrated into daily operations. Despite being the first full-scale test of this unoptimized prototype and encountering operational challenges, including unstable freshwater distribution and unplanned power outages, the concept exhibited a reduction in infestation pressure of Lepeophtheirus salmonis and Caligus elongatus. Notably, infestation rates for the pre-adult and adult stages were reduced by 48% and 57%, respectively, in the treatment cages compared to controls. These effects were achieved even though salinity levels in the upper layer were only sporadically maintained below 10 ppt, highlighting the robustness and promise of the approach. Given these encouraging findings, further testing is strongly recommended under more stable and controlled conditions with improved freshwater delivery systems and real-time regulation. Enhanced environmental monitoring, particularly high-resolution salinity logging in the surface layer, is essential to optimize and document system performance. Moreover, evaluating the direct impact on the more freshwater-sensitive planktonic larval stages (nauplii and copepodids) is crucial to understanding the true preventive effect of salinity manipulation. A strong recommendation for future studies is to implement constant surveillance of the near-surface salinity in real time to enable precise regulation of freshwater input and support effective environmental documentation. Furthermore, future tests should aim to isolate the relative contribution of the tarpaulin skirt versus the salinity manipulation to fully understand the mechanisms driving lice reduction. In addition, there is a need to examine the effectiveness of the technique at multiple sites and under a wide range of various site conditions, especially various current rates through the site.