Introduction

Marine heatwaves (MHWs) are prolonged periods (≥5 days) of extremely high ocean temperatures1. These extreme events can have major ecological and socioeconomic impacts, including shifts in species ranges, mass die-offs, ecosystem reconfigurations, declines in fishery catch rates, and increased mortality or reduced performance of farmed aquaculture species2,3,4. The frequency, duration, extension, and severity of these events have increased globally in the last decades and are projected to further increase due to global warming5,6,7.

MHWs can be generated by local and/or regional atmospheric and oceanic processes. Locally, anomalous air–sea heat fluxes (e.g., increased shortwave radiation from reduced cloud cover or suppressed latent heat loss from warm/humid air) can play an essential role in the generation of MHWs8,9,10. At a regional scale, ocean dynamics, such as current anomalies and increased advection of warmer/fresher waters, can also trigger MHWs8,9,10. The combined effects of local warming and enhanced advection of freshwater generate a stratified water column that enhances warming by restricting heat exchanges and vertical mixing in the ocean. MHWs are also associated with large-scale climate modes of variability, such as El Niño Southern Oscillation (ENSO) and North Atlantic Oscillation (NAO) that modulate ocean temperatures through atmospheric and oceanic teleconnections8.

High latitude coastal ecosystems, already under pressure from climate change amplified towards the poles, ocean acidification, and human activities, are particularly vulnerable to MHWs11. The coastal regions of northern Norway hold particular importance as they support recovering foundational kelp forest ecosystems12,13,14, include the main spawning grounds for Northeast Arctic cod15,16, and host a large salmon aquaculture industry17. This area is characterized by a distinct northward coastal current and an upper water column that is well-mixed in winter but strongly stratified during the summer months18. The Coastal Current is bordered by the northward-flowing Norwegian Atlantic Current, which carries warm and saline Atlantic Water into the area. In these northern coastal areas, surface temperatures are typically as low as 3–4 °C in winter, while in summer, they rarely exceed 12–13 °C18. Understanding the drivers of MHWs in this region is crucial for assessing their predictability and informing strategies for effective management and mitigation.

The summer of 2024 witnessed unusually high sea and air temperatures in northern Norway, linked to a sharp rise in salmon lice abundance and infestations19. The salmon louse (Lepeophtheirus salmonis), a natural ectoparasite of salmonids in the northern hemisphere, affects both farmed and wild populations, with its growth rates strongly influenced by rising sea temperatures20. Salmon lice have been identified as one of the main risk factors for further expansion of Norwegian salmon farming21, and salmon lice-induced mortality in wild salmonid populations is currently used as the key sustainability indicator in Norway’s salmon farming management system22. The unprecedented increase in salmon lice abundance and its challenges for salmon health and production motivate the present study to investigate the underlying drivers and potential ecological impacts of MHW events in the region.

We investigate the characteristics of the regional summer MHW that affected northern Norway in 2024 and identify potential atmospheric and oceanic drivers. To do this, we use ocean model hindcast data from Norkyst, atmospheric reanalysis data from ECMWF Reanalysis version 5 (ERA5), and in situ observations from coastal monitoring stations. We also study the impacts of this extreme event on salmon lice and discuss the broader implications of increasing MHWs in a warming climate for the region’s fisheries and aquaculture sectors, with a particular focus on salmon lice outbreaks.

Results

Characteristics of the marine heatwave

A strong marine heatwave was identified from August 5th to August 26th, 2024 (21-day duration) in coastal surface waters of northern Norway based on spatially averaged sea surface temperatures (SST) from the Norkyst ocean model as a reference (Fig. 1a and see Fig. 1b for the averaging area). The MHW had a cumulative intensity of 42.93 °C days with a mean intensity of 1.95 °C and a maximum intensity of 2.49 °C found on August 15th, with area-averaged SST reaching 14 °C (Fig. 1a), the highest area-averaged temperature recorded in the region in 2012–2024 (see Supplementary Table 1). In some coastal areas, however, temperatures reached up to 18 °C (Fig. 1b), exceeding 4 °C above climatology (Supplementary Fig. 1). The MHW started along the coast in the Vestfjorden and Lofoten area and then expanded throughout the entire region, reaching its greatest extent on August 15th (Fig. 1b and see Supplementary Movie 1), when the maximum intensity was observed. After August 26th, the MHW persisted in the southern parts of the Barents Sea region until the end of our time series on September 30th (see Supplementary Movie 1). From a cross-shore section starting near Eggum (68.4° N) and extending seawards, we observed that the MHW was restricted to the top 15–30 m of the water column from its onset and until August 16th at all locations. By late August, the MHW was mostly confined to depths ranging from 30–50 m, reaching the bottom (~100 m depth) at the innermost locations in September (see Fig. 1c for August 15th and Supplementary Movie 2 for temporal progression). These temperature profiles from Norkyst are consistent with observations from the monitoring station at Eggum, with strong positive temperature anomalies reaching the bottom by September (see Supplementary Fig. 2).

Fig. 1: Characteristics of the 2024 northern Norway MHW based on the Norkyst model.
figure 1

a Temporal surface MHW characteristics from SST (thick black line) averaged across the subregion shown in (b). The dashed blue line represents the 2012–2023 climatology, the solid green line indicates the 90th percentile threshold, and the MHW event (SST > threshold) is shaded in red. b Surface extent of the MHW. Empty squares mark the locations used for the temperature profile in (c), and dashed lines indicate the area used to calculate the MHW temporal characteristics (a). c MHW temperature profile (top 250 m) along a transect off Eggum shown in (b). The horizontal and vertical extent of the MHW is shown for August 15th, the date of maximum intensity. Hatching/dots highlight grid points/depth layers where sea temperatures exceed the 90th percentile threshold.

Full size image

Drivers of the marine heatwave

We first looked at the role of local ocean stratification and meteorological conditions on the MHW. Mean mixed layer depths (MLD) in the selected subregion (see Fig. 1b) were consistently shallower than the 2012–2023 average in the weeks before and after the MHW onset, with MLD shallower by more than one standard deviation from the mean from mid-July to mid-August (Fig. 2a). The relatively shallow MLD coincided with anomalously high air temperatures and below average wind speeds throughout most of July and early August as observed from a meteorological station in Tromsø (Fig. 2b–c, see location in Supplementary Fig. 3). By mid-August, wind speeds near Tromsø started to increase, enhancing mixing of the water column that led to a deeper mixed layer and the culmination of the MHW in surface waters (Fig. 2).

Fig. 2: Local oceanic and atmospheric conditions.
figure 2

Time series of mean mixed layer depth from the Norkyst ocean model averaged over the subregion shown in Fig. 1b (a), and daily air temperatures (b) and wind speeds (c) from a weather station in Tromsø (69.6537° N, 18.9368° E; https://seklima.met.no). Time series were smoothed using an 11-day running average. The mean (dashed line) and ±1 standard deviation (shaded region) of historical values (2012–2023 for mixed layer depth; 1991–2020 for air temperature and wind speed) are shown in gray, and 2024 data are represented in black. The shaded blue area marks the duration of the marine heatwave.

Full size image

Total air–sea heat fluxes across the region remained above average from early July to mid-August (Fig. 3a). The largest anomalies were observed in shortwave radiation (Fig. 3b), associated with reduced cloud cover in July and August (Supplementary Fig. 4a). Additionally, relatively high downward (warming) sensible heat fluxes, combined with low latent heat fluxes to the atmosphere (indicating reduced water evaporation from the ocean), contributed to total heat fluxes approximately 50% above average during the first half of August 2024 (Fig. 3b). See Supplementary Fig. 5 for the time series of all air–sea heat flux components.

Fig. 3: Air–sea heat fluxes from ERA5.
figure 3

a Total air–sea fluxes for 1991–2020 (gray dashed line) and 2024 (orange solid line). The shaded area indicates one standard deviation from the 1991–2020 mean, and the vertical shaded area in blue marks the 2024 MHW period. The time series was smoothed using an 11-day running average. b Mean heat fluxes for 1991–2020 (gray) and 2024 (orange) during 1–15 August 2024. Black lines indicate one standard deviation from the 1991–2020 mean. Heat fluxes were averaged over the subregion shown as a dashed box in Fig. 1b. SW shortwave radiation, LW longwave radiation, LH latent heat, SH sensible heat, TOT total heat fluxes. Positive values represent downward (i.e., warming) fluxes.

Full size image

At a larger scale, summer of 2024 was characterized by a strong sea level pressure gradient in the North Atlantic (Fig. 4a, b) that resulted in the strongest positive phase of the NAO index, defined as the pressure difference between the Icelandic Low and the Azores High, on record for July and August (Fig. 4c). This strong, positive phase of the NAO was associated with strong SW winds that likely enhanced a northward advection of warm and humid air masses that resulted in anomalously high air temperatures at 2 m (2–3.5 °C anomalies, Supplementary Fig. 4b) and positive anomalies in total water vapor in the atmosphere in the northern Norway coastal region in August (Supplementary Fig. 4c).

Fig. 4: Summer large-scale atmospheric patterns.
figure 4

ERA5 sea level pressure for July (a) and August (b), along with the normalized summer (July–August) NAO index time series relative to the 1991–2020 climatology (c).

Full size image

We also identified a northward propagation of anomalously warm and low salinity waters along the coast both at surface and 50 m depth that reached the MHW area (latitudes north of 68° N) by the time of the MHW onset (Fig. 5). Anomalously warm and low salinity surface waters reached the MHW area by mid-July (Fig. 5a, c) and by the end of June at 50 m depth (Fig. 5b, d). These model findings are supported by observational data from the monitoring stations, with low salinity waters reaching Bud (62.9° N) in May of 2024, Eggum (68.4° N) by the end of June, and Ingøy (71.1° N) by the end of August (Supplementary Fig. 2).

Fig. 5: Northward propagation of coastal water.
figure 5

Temperature anomalies (a, b) and salinity anomalies (c, d) from the Norkyst ocean model at 0 m and 50 m depth at selected locations along the Norwegian 2 nautical mile line shown in Supplementary Fig. 3. Anomalies were calculated relative to 2012–2023 climatology.

Full size image

A heat budget analysis for the mixed layer supports these findings by quantifying the relative contributions of different drivers throughout the event (Fig. 6). Atmospheric heat fluxes, particularly shortwave radiation, were the primary driver of warming during both the preconditioning phase (two weeks before the MHW) and the onset phase (from the start to the peak of the event). Advection contributed positively during the first week of the MHW but played only a minor role during the first two MHW stages. In contrast, advection became more important during the decline phase (from the MHW peak to its end), removing heat from the region through northward transport. Although ocean heat advection was not a major factor during the preconditioning phase, strong positive anomalies in freshwater transport during both the preconditioning and onset phases (Supplementary Fig. 6a) suggest that freshwater advection contributed greatly to shoaling the mixed layer (Fig. 2a), thereby enhancing atmospheric heating.

Fig. 6: Cumulative surface mixed layer budget during the different phases of the 2024 MHW derived from the Norkyst ocean model.
figure 6

The contribution from atmospheric heat fluxes (Qatm) is shown in green, with the contribution from shortwave radiation (Qsw) indicated by a dashed line. Entrainment at the base of the mixed layer (Qentr) is shown in blue, and the residual term (R), primarily including advection and other unresolved processes, is shown in purple. The net cumulative temperature change (Qnet = Qatm − Qentr + R) is shown in black.

Full size image

Impacts of the 2024 marine heatwave on salmon lice infection pressure

The anomalously high sea temperatures in the summer of 2024, greatly influenced the infection pressure from salmon lice. By mid-June, the infection pressure (i.e., number of copepodids in the water) across all Norway’s production zones (Fig. 7a) exceeded one standard deviation above the long-term average from 2012 to 2023 (Fig. 7b). While the long-term average reached its peak in mid-August, the infection pressure continued to escalate throughout August 2024 (Fig. 7b). On August 1st and 31st, infection pressures were four and six standard deviations above the long-term average, respectively (Fig. 7b), while mean sea temperatures measured at 3 m depth in production zones 9–13 reached five standard deviations above the 2012–2023 mean, peaking at 16 °C (Supplementary Fig. 7). Typically, the infection pressure has been highest in the southern regions (production zones 2–6) and lowest in northern regions (Fig. 7c). However, in 2024, this pattern was reversed, with the highest infection pressure estimated in production zones 6–10 (Fig. 7c) and record salmon lice abundances for production zones 6–13 in our study period. The greatest difference was noted in production zones 11 and 12, where salmon lice abundances in 2024 were 9 and 12 standard deviations above their 2012–2023 means (Fig. 7c).

Fig. 7: Salmon lice infection pressure.
figure 7

a Map showing locations of production zones along the Norwegian coast. b Mean number of infective salmon lice larvae in the water, averaged across all production zones in 2024 (red line) relative to the 2012–2023 mean (dashed blue line). The shaded blue area corresponds to one standard deviation from the 2012–2023 mean. Red dots indicate 4 and 6 standard deviation values based on the 2012–2023 mean. c Abundance of infective salmon lice larvae at each production zone in August for 2024 (red) and the 2012–2023 period (blue). Blue and red dashed lines represent the mean number of copepodids across all production zones for 2012–2023 and 2024, respectively. Black lines indicate the 95% confidence interval.

Full size image

Discussion

The summer MHW that affected northern Norway in 2024 was one of the most severe MHWs recorded in the area and had unprecedented impacts on salmon lice infestations and salmon aquaculture, with increased salmon mortality and financial losses. This MHW was driven by a combination of two sets of factors: 1. prevailing local meteorological and oceanic conditions, and 2. external heat and freshwater supply enhanced by large-scale atmospheric conditions in summer 2024.

The marine heatwave observed in August 2024 stands out as one of the strongest within our study period (2012–2024, see Supplementary Table 1) despite a potential underestimation due to a bias in Norkyst forcing during 2012–2020 (see “Methods” section). In August 2024, SST reached a record high, with regional averages from Norkyst measuring 14 °C and coastal areas nearing 18 °C (Fig. 2c)—more than 4 °C above the historical mean (Fig. 7 and Supplementary Fig. 1). Subsurface temperatures also reached unprecedented levels; weekly measurements at 3 m depth from aquaculture farms in production zones 9–13 recorded average temperatures of 16 °C (Supplementary Fig. 7). At the Eggum monitoring station, temperatures at 1 m depth exceeded 16 °C in the first half of August (Supplementary Fig. 8), marking the highest recorded temperature since the measurement series began in 1935. This MHW extended beyond surface waters, reaching the bottom at the innermost locations, highlighting the need to investigate subsurface signatures—an aspect addressed in very few regional studies (e.g., ref. 23).

On local and short time scales, the 2024 marine heatwave was primarily driven by a high-pressure system that resulted in anomalously high solar irradiation, high air temperatures, and relatively weak winds, all of which contributed to anomalous high heat fluxes to the sea and a shallow mixed layer depth. The relatively shallow mixed layer limited vertical mixing in the ocean and sustained warm surface waters until the end of August. Previous studies highlight the dominant role of solar radiation and latent heat fluxes in driving MHWs23,24,25,26,27. In our study region, shortwave radiation heat fluxes were anomalously high in July and August, linked to negative cloud cover anomalies. This, in combination with increased downward sensible heat fluxes and low latent heat fluxes (indicating reduced oceanic cooling via evaporation), contributed to ~50% above-average net heat fluxes into the ocean during early August. This pattern is consistent with findings from Eisbrenner et al.27, who reported that summer MHWs in the Barents Sea were strongly associated with shortwave radiation anomalies, reinforcing the role of reduced cloud cover as a critical driver of summer MHWs in northern Norway.

At a larger scale, the 2024 MHW was driven by strong sea level pressure gradients over the North Atlantic, reflected in a record-high positive phase of the NAO. This atmospheric configuration intensified southwesterly winds, enhancing the northward advection of warm, humid air and unusually fresh coastal waters, which contributed to ocean surface warming and stratification in northern Norway. Positive NAO phases are known to influence the regional climate by strengthening southwesterly winds and increasing precipitation over western Norway28,29,30. A strong positive NAO can also narrow and deepen both the warm and saline slope current of Atlantic Water and the fresh Coastal Current with reduced heat loss to the atmosphere31.

Model outputs and observations indicated a northward propagation of anomalously low-salinity surface and subsurface waters reaching the MHW region by the time of its onset. These unusually fresh waters likely originated from intense precipitation over western Norway in June–July 2024 (Supplementary Fig. 9), an episodic freshwater outflow from the Baltic Sea in early April (Supplementary Fig. 10), and reduced dilution of coastal waters, all linked to prevailing NAO conditions. Although the Coastal Current’s freshwater signal typically becomes diluted as it moves north—especially in spring, when northerly winds and offshore Ekman transport are more frequent—persistent southwesterly winds, as observed in 2024, can sustain coherent freshwater transport over long distances32. This early advection of warm and unusually fresh water likely preconditioned the region for increased stratification, facilitating and intensifying the MHW. Figure 8 summarizes the contributing factors to the 2024 MHW.

Fig. 8: Summary of contributing factors to the development of the August 2024 marine heatwave in northern Norway.
figure 8

The color scale indicates sea surface temperature anomalies (SSTA) from ERA5 for August 15th. NAO North Atlantic Oscillation, CC coastal current, MLD mixed layer depth.

Full size image

The strong marine heatwave in northern Norway triggered unprecedented increases in salmon lice abundance. It is well established that higher temperatures promote salmon lice larvae production, shorten the development of non-infective nauplii to infectious copepods, and increase the infectivity of copepodids, which intensifies infection pressure on both farmed and wild salmonids20,33. A 2 °C increase in temperature can double infection pressure20, as higher temperatures shorten the generation time from copepodid to adult female louse, with the most pronounced reduction occurring at lower temperatures— decreasing by 27 days between 6 and 9 °C compared to just 6 days between 15 °C and 18 °C34. Since salmon lice larvae predominantly inhabit the upper 20 m of the water column35, and salmon farms typically extend to ~30 m depth36,37—both parasites and hosts are exposed to surface and subsurface temperature anomalies. These conditions highlight the heightened vulnerability of northern Norway’s salmon farming industry to MHWs.

The 2024 outbreak was further exacerbated by limited delousing capacity and a growing number of hosts due to aquaculture expansion in northern Norway over the past decade, intensifying lice transmission potential38 (Supplementary Fig. 11). During the MHW, lice counts exceeded the legal threshold of 0.5 adult female lice per fish, with counts reaching 1–4 adult female lice per fish by mid-August (Norwegian Ministry of Fisheries and Coastal Affairs 2017). A sharp increase in the number of farms exceeding the threshold was observed from late July onward (Supplementary Fig. 12). Exceeding the threshold triggers regulatory requirements for treatment, and these elevated infestations prompted large-scale delousing efforts. Persistently high lice levels into late summer suggest that delousing capacity may have been insufficient to bring infestations under control in a timely manner. Such treatments not only compromise fish welfare but also impose substantial economic losses on farmers39,40.

Beyond its impact on salmon lice outbreaks, elevated sea temperatures pose additional challenges to salmon aquaculture. While moderate warming has been linked to accelerated growth in northern and central Norway, prolonged exposure to temperatures exceeding 17–18 °C can lead to reduced feed intake and increased fish mortality41. Warmer waters also heighten the risk of harmful algal blooms and decrease oxygen availability, further contributing to fish mortality and stressing farmed populations42,43. Since temperature is a key factor in the aquaculture industry, it is imperative to consider both climate change and the likelihood of MHWs in management and future operational planning.

Understanding the drivers of MHWs in coastal Norway is critical, both due to their impact on salmon aquaculture and because the region hosts foundational species like kelp forests and key spawning grounds for Atlantic cod. Kelp forests, which have shown signs of recovery in northern Norway following declines associated with overfishing14, are highly vulnerable to MHWs44. Kelp farming has recently expanded across Europe, and with active cultivation sites along the Norwegian coast and seaweed now included in the Norwegian Fish Export Act, signaling an intent to expand production45, it is crucial to account for the potential impacts of MHWs in the planning and expansion of kelp aquaculture in Norway. Additionally, with potential temperature extremes exceeding 1.4 °C superimposed on projected temperatures of up to 8.2 °C at grazing and spawning depths along the northern Norwegian coast46 in winter/spring, the thermal tolerance limit of 9.6 °C for Atlantic cod during oocyte maturation and spawning will be exceeded during MHW events. Such conditions could lead to a northeastward shift in the spawning grounds of Northeast Arctic cod, potentially moving these critical habitats along the Norwegian coast and into Russian coastal waters16.

The observed interactions between oceanic and atmospheric processes across local to basin-wide scales in the development of MHWs in northern Norway underscore the need for advancements in regional coupled ocean-atmosphere modeling to enhance the possibility of predicting these extreme events and inform mitigation strategies. With MHWs projected to intensify and occur more frequently in northern Norway under ongoing global warming47, it is crucial to assess their cumulative impacts alongside other climate stressors, such as ocean acidification and deoxygenation, on marine ecosystems and economically important species in the region.

Methods

Datasets

Sea temperature and salinity for the Norwegian coast were obtained from a hindcast simulation from January 2012 to September 2024 using the hydrodynamic model system Norkyst. The Norkyst model is an implementation of the Regional Ocean Modeling System (ROMS; http://myroms.org, Shchepetkin and McWilliams48; Haidvogel et al.49) for the entire Norwegian coast with an 800 m × 800 m horizontal resolution and 40 vertical topography-following layers. The Norkyst domain is shown in Supplementary Fig. 3, and Asplin et al.50 provide several details about the model system and external forcing sources. However, our simulation is built up slightly differently and comprises a new, upgraded version. Our high-resolution atmospheric fields used to force Norkyst were retrieved from both a 3 km simulation of the Weather Research and Forecasting model (WRF), same as explained in Asplin et al.50, for the years 2012–2016, and the non-hydrostatic 2.5 km AROME MetCoOp regional atmospheric model51 shared by the Norwegian Meteorological Institute and accessible from https://thredds.met.no for the years 2017–2024. The atmospheric data used includes wind at 10 m, air temperature at 2 m, relative humidity at 2 m, and total precipitation. In addition, our Norkyst setup is forced with downwelling shortwave and longwave radiation fluxes directly from an atmospheric model. Note that while Norkyst applied these fluxes from AROME MetCoOp for 2021–2024, we obtained these fluxes from the regional atmospheric hindcast NORA352 for 2012–2020, as AROME MetCoOp was not available for the first years of the simulation period. River runoff was based on daily measurements from the Norwegian Water Resources and Energy Directorate. Input along the open boundaries consisted of both tidal forcing from the global TPXO953 and daily averaged currents, hydrography and water level from the E.U. Copernicus Marine Service Information (ARC MFC; doi: 10.48670/moi-00007) with the exception of the southern part of the eastern boundary, where we used boundary conditions from the BAL MFC (doi: 10.48670/moi-00013). We applied daily averaged fields from the Norkyst model in our analysis.

Through a thorough validation of forcing data and near-surface temperatures in the Norkyst archive, we identified a slight overestimation of solar radiation in the NORA3 archive52, likely due to cloud cover effects, which also influence longwave radiation. Since the Norkyst hindcast applied these fluxes for the years 2012–2020, we assessed the impact of this bias by rerunning 2021—the only year with both NORA3 and AROME MetCoOp radiation data available—with NORA3 radiation forcing. Our results indicate that the NORA3 fluxes introduced a summer warm bias of approximately 0.7–0.8 °C compared to radiation from AROME MetCoOp. Additionally, solar radiation data from ERA5 suggest that 2021 experienced lower-than-average solar radiation (i.e., increased cloud cover), indicating that the observed warm bias represents an upper limit. Importantly, this bias did not affect our analysis of the 2024 event in terms of absolute sea temperatures and their drivers; however, it suggests that the intensity of the MHW would have been greater relative to ocean temperatures from 2012 to 2020.

Modeled oceanographic data was complemented with temperature and salinity measurements from the Norwegian Institute of Marine Research (IMR) fixed coastal monitoring stations Bud (62.9333° N; 6.7833° E), Skrova (68.1167° N; 14.6500° E), Eggum (68.3667° N; 13.6333° E), and Ingøy (71.1332° N; 24.0166° E) (Supplementary Fig. 3). Vertical temperature and salinity profiles at these stations are measured 2–4 times per month by local observers and time series are available from the 1940s (http://www.imr.no/forskning/forskningsdata/stasjoner). A comparison of SST from the Norkyst model and the monitoring stations Skrova, Eggum, and Ingøy during the MHW period is shown in Supplementary Fig. 13.

Large-scale, long-term atmospheric data, including total column water vapor, cloud cover, air temperature at 2 m, mean sea level pressure, precipitation, and air–sea heat fluxes, were obtained from the ERA5 global atmospheric reanalysis product available at the Copernicus Climate Change Service54. ERA5 products have a spatial resolution of 0.25° by 0.25° and cover the period from January 1940. All variables, except air–sea heat fluxes, were obtained at a monthly resolution, and monthly anomalies for 2024 were calculated relative to a 1991–2020 climatological mean, following the World Meteorological Organization (2017) recommended reference period. Air–sea fluxes data, including mean surface shortwave and longwave radiation, sensible heat, and latent heat, were obtained at a daily resolution and averaged over a subregion within the MHW onset area shown in Fig. 1b and Supplementary Fig. 3.

Local air temperature and wind speed data were obtained from the meteorological station SN90450 (69.6537° N, 18.9368° E; https://seklima.met.no) located in Tromsø. Daily wind and air temperature values from 2024 were smoothed using an 11-day moving average to remove high-frequency variability. The 11-day window was chosen for consistency with MHW calculations.

Time series of the NAO index were obtained from https://psl.noaa.gov/. The NAO corresponds to the pressure difference between Iceland and the Azores. The NAO exerts a strong influence on northern European rainfall, temperature, and cloudiness and is of key importance in generating summer climate extremes, including flooding, drought, and heat stress in northwestern Europe55,56.

The dispersal and density of infectious salmon lice larvae (copepodids) along the Norwegian coast were estimated using a bio-hydrodynamic salmon lice model50,57, in which a particle-tracking model (LADiM), including an individual-based behavioral model for lice larvae, is coupled with the Norkyst ocean model system. Data on the number of salmon lice larvae released into the water column were derived from reports on lice counts per fish, fish abundance per site, and water temperature at all active aquaculture facilities58,59. Weekly nationwide estimates of salmon lice density, including two-day forecasts, are published on lakselus.no.

To ensure consistency across data sources, we also evaluated all variables using a uniform climatological period (2012–2023) matching the period available from Norkyst. The resulting anomalies and main patterns remained unchanged, indicating that differences in baseline periods had minimal influence on our conclusions. We therefore retained the original baselines specific to each dataset to reflect the most robust datasets available for each variable.

Marine heatwave definition

A MHW event is defined here following Hobday et al.1 as a period of ≥5 days with sea temperature above the 90th percentile of its daily climatology. We used daily averages of ocean temperatures from Norkyst from 2012 to 2023 to construct a seasonally varying climatology. Using a recent 12-year climatology enables us to distinguish an extreme event in relation to current ‘normal’ conditions. The climatology and the 90th percentile threshold were calculated on every grid point and depth layer of the Norkyst domain between 66° N and 72° N, and for each day of the year, using an 11-day window centered on each date. A 31-day running mean was then applied to the daily climatological and threshold temperature time series following Hobday et al.1 recommendation. The analysis was also performed over the spatially averaged SST for the subregion shown in Fig. 1. We described the surface 2024 MHW using common MHW metrics such as start and end dates, duration, and maximum, mean, and cumulative intensities. Duration is the number of days between the start and end of the MHW, and the intensities are calculated as anomalies relative to the daily climatology. MHW metrics were calculated using the marineHeatWaves Python module (https://github.com/ecjoliver/marineHeatWaves).

To assess the sensitivity of our results to the choice of climatological baseline, we repeated the MHW analysis using ERA5 SST data for the same region. We compared MHW characteristics derived using a recent 12-year baseline (2012–2023)—which aligns with the baseline used in our Norkyst analysis—with those obtained using a longer 30-year baseline (1991–2020). The key features of the 2024 MHW event, including its onset timing and intensity, were consistent across both baselines. However, the event duration was longer under the 30-year baseline, reflecting the lower threshold associated with a cooler historical reference period. This comparison, included in Supplementary Fig. 14, confirms that our results are robust to the choice of baseline.

Hydrographic characteristics

The mixed layer depth (MLD) was estimated as the depth where the potential density difference with respect to that at the surface first exceeds 0.03 kg m−360,61. The MLD was averaged over a subregion within the MHW area shown in Fig. 1b and Supplementary Fig. 3.

Volume, heat, and freshwater transports for the upper 50 m of the water column were calculated at a section located just south of the MHW onset area shown in Supplementary Fig. 3, extending from the coast to the seaward boundary of the coastal current. Volume, heat, and freshwater transports were calculated as:

$$\int {vdxdz},\,{c}_{p}{\rho }_{0}\int \theta {vdxdz},\,\int \frac{(S0-S)}{S0}{vdxdz}$$
(1)

where v is the velocity perpendicular to the section, θ and S are the potential temperature and salinity, interpolated to the location of v, and S0 is a reference salinity of 34.8. CP is the heat capacity of water (4008 J kg−1 °C−1) and \({\rho }_{0}\) is the reference density (1027 kg m⁻³).

Surface mixed layer heat budget

We considered the surface mixed layer as fully mixed and used a bulk formula for the MLD heat budget, with an entrainment term following62.

$$\frac{{dT}(t)}{{dt}}= \frac{{Q}_{{SW}}\left(1-\left(0.78{e}^{\frac{{-h(t)}}{{1.4}}} + \left(1 - 0.78\right){e}^{\frac{{-h(t)}}{{7.9}}}\right)\right)}{{\rho }_{0}{C}_{p}h(t)}+\frac{{Q}_{{LH}}{+Q}_{{SH}}+{Q}_{{LW}}}{{\rho }_{0}{C}_{p}h(t)} \\ -\epsilon \frac{T\left(t-1\right)-{T}_{{dh}}\left(t-1\right)}{h\left(t\right)}\frac{{dh}\left(t\right)}{{dt}}+R$$
(2)

T(t) is temperature integrated over the MLD at day t, QSW is shortwave flux at the surface, \(\big(0.78{e}^{\frac{{-h(t)}}{{1.4}}} + \left(1 - 0.78\right){e}^{\frac{{-h(t)}}{{7.9}}}\big)Q_{SW}\) is shortwave flux at the bottom of the MLD, according to the ROMS parameterization for water type III used in Norkyst. \({Q}_{{LH}}{+Q}_{{SH}}+{Q}_{{LW}}\) are latent heat, sensible heat, and longwave radiative heat flux, respectively, h is MLD (m), \({T}_{{dh}}\left(t-1\right)\) is the temperature averaged over the entrained layer (h(t − 1) to h(t)) (°C), ϵ is a Heaviside function: 0 when \(\frac{{dh}}{{dt}}\) < 0 and 1 when \(\frac{{dh}}{{dt}}\) > 0. \(\frac{{dh}}{{dt}}\,\) is the change in MLD from t − 1 to t. R is the remaining terms (including advection). We calculated each term at every grid point on a daily basis and averaged them over the selected area shown in Fig. 1b.