NOAA's September forecast gives Chilean salmon producers a reason to pay attention to the coming summer. El Niño is strengthening, and NOAA estimates a greater than 90% chance that it will become very strong during the Northern Hemisphere autumn and winter of 2026–27. (NOAA's September ENSO discussion) We looked back at local surface temperatures during every completed El Niño since 1982 to see what actually occurred around salmon farms in Los Lagos and Aysén.
SRS, or salmonid rickettsial septicemia, is one reason local temperature changes during El Niño matter for Chilean salmon health. The disease was first identified in 1989 during a mortality outbreak in Chilean coho salmon, where roughly 1.5 million fish died at an estimated cost of $10 million. By the following year it had also appeared in Atlantic salmon, with mortality rates reaching as high as 90% at some sites and an annual industry cost near $50 million by the mid-1990s. Today the disease accounts for over half of Chile's infectious-disease-related salmon mortality and an estimated $700 million in annual losses. It has since been reported in Norway, Canada, and Scotland as well, but at far lower severity than in Chile.
The difference comes down to the bacterium's biology. Piscirickettsia salmonis was the first rickettsia-like organism recognized as a fish pathogen, adapted to an intracellular lifestyle rather than the extracellular one most bacteria use. It is not, strictly speaking, a true rickettsial disease, since the bacterium can survive both inside and outside host cells rather than exclusively inside them. That dual capability cuts both ways: once inside a cell, the bacterium is largely out of reach of antibiotics, which work best before the bacteria establish themselves intracellularly, and that biology is central to why reducing antibiotic use industry-wide has proven difficult. It also limits standard vaccines, since antibodies triggered by a vaccine can't reach bacteria once they're already inside a cell. Researchers at the Pontificia Universidad Católica de Chile have pointed to cell-mediated immunity, rather than antibody response alone, as the more promising target for next-generation vaccines against the disease, precisely because of this intracellular niche. (Fish Farming Expert, interview with Dr. Alexis Kalergis, PUC Chile)
The bacterium's ability to survive outside a host is also what makes water itself part of the transmission picture, not just individual fish health. P. salmonis can persist in seawater for several weeks at temperatures between 5°C and 20°C, long enough to spread between sites without any direct contact between fish. Vertical transmission through eggs is also considered increasingly plausible, tied to an adhesion complex that lets the bacterium attach to egg surfaces. How much each route actually contributes at the farm level is still not fully settled. (The Fish Site, disease guide: salmonid rickettsial septicaemia)
Our surface-temperature record spans 44 years and roughly 21 million site-days across salmon-farming areas in Los Lagos and Aysén. We compared local temperatures against their 1991–2020 seasonal norms during 13 completed El Niño events since 1982.
How local temperatures responded during 13 El Niños
Los Lagos was warmer than normal during eight of those events; Aysén was warmer during six. Across all 13, the average anomaly was +0.018°C in Los Lagos and +0.010°C in Aysén.
During the 1997–98 El Niño, Los Lagos averaged about +0.26°C above normal and Aysén about +0.16°C. The similarly strong 1982–83 event coincided with much smaller local anomalies, around +0.07°C and +0.09°C respectively. In 2014–16, Los Lagos warmed modestly during the event, then averaged roughly +0.34°C in the following 12 months. During the strong 2023–24 El Niño, both regions were colder than their seasonal norms: about −0.16°C in Los Lagos and −0.22°C in Aysén.
The warming that followed 2014–16 in Los Lagos led us to check the lagged relationship month by month. In this series, the correlation between the ENSO index and local temperature anomalies is about 0.05–0.06 in the same month. It rises when ENSO leads local temperature, peaking around six months at 0.21 in Los Lagos and 0.20 in Aysén. Even at that peak, a simple linear comparison accounts for only about 4% of the variation in the regional monthly anomalies.
What regional temperatures leave unresolved for SRS
Winds, currents, freshwater inputs and the shape of each channel affect the water reaching a farm. This record also measures surface temperature; it has not been corrected for the difference between the surface and the depths where fish are held.
Not every El Niño produces this kind of warming along the Chilean coast. Chile's farming regions sit well south of the equator, and the cooling influence of Antarctic winds is one plausible reason the effect doesn't show up consistently, though isolating that specific mechanism is beyond what this dataset can confirm on its own. When a strong event does reach the coast, though, the effect can persist: Los Lagos and Aysén both saw temperature anomalies following the 1997-98 El Niño, and the 2015-16 event brought an extended period of elevated coastal temperatures, roughly 30 weeks by our data.
Our temperature record runs 44 years back, but Chile's SRS-specific monitoring data does not reach nearly that far. The industry's disease-reporting systems only go back roughly a decade, which is why 2015-16, rather than 1997-98 or 1982-83, is the strongest comparison available: it's the most recent event where both the extended coastal warming and the SRS mortality data sit inside the same window we can actually verify.
Within that window, the seasonal pattern is visible. During the 2015-16 event, the monthly share of salmon mortality attributed to SRS climbed to over 30%, against a more typical peak closer to 20% in the years since, though part of that decline may also reflect more granular reporting of other diseases rather than a real drop in SRS pressure alone.
Sernapesca's own 2025 reporting shows the disease's current scale in a different, non-seasonal snapshot: infectious causes accounted for 25.7% of reported Atlantic salmon mortality that year, with piscirickettsiosis responsible for 52.8% of that infectious share, and for 97.51% of antimicrobial use by disease among production cycles closed in 2025. Neither figure describes the share of all mortality caused by SRS, but together with the 2015-16 pattern, they point to a disease that dominates both the mortality and treatment side of the ledger, in an El Niño year and outside one. (Sernapesca's 2025 health report; 2025 antimicrobial report)
A farm's temperature history alone still leaves open whether its fish actually encountered the pathogen. In Los Lagos, researchers combined disease records with hydrodynamic modeling and found that the number of infected farms in upstream waters, and the severity of infection at those farms, were among the strongest determinants of modeled piscirickettsiosis prevalence, with temperature and salinity contributing as well. The model does not prove exactly how bacteria moved between farms, but it shows why a site's own temperature reading may need to be read alongside what's happening in the water connected to it. (Bravo and colleagues' study)
At a site where exposure is plausible, lice pressure may affect how fish respond to P. salmonis, though the size and mechanism of that effect remain uncertain. Weekly farm data from southern Chile found that egg-bearing female Caligus rogercresseyi levels predicted both greater SRS mortality risk and an earlier first reported outbreak. The researchers proposed that lice, P. salmonis, and rising temperatures may place combined stress on fish; their observational analysis cannot establish that mechanism. (Diethelm-Varela and colleagues' study)
That's the precedent worth tracking as this next El Niño event develops: not just whether the coast warms the way it did in 2015-16, but whether an extended warm period again coincides with a rise in the mortality share attributed to SRS specifically, rather than to the industry's other, better-controlled diseases.
Temperature-data note: Manolin site-level surface-temperature analysis, 1982–2026, 1991–2020 climatological baseline. The charted tropical index is ONI; the 2026 NOAA strength forecast uses RONI.
