Introduction Disinfection is a very important part of recirculation aquaculture systems (RAS). RAS allows for controllable environments in which main variables relevant to animal welfare and a successful production can be manipulated as required to improve efficiency and profitability. Common disinfection methods include chemical disinfectants, antibiotics, biocides, UV radiation and ozone. They can be used to treat disease outbreaks, or to reduce the bacterial load of the system which otherwise could lead to the overgrowth of potential pathogens or opportunistic bacterial groups competing with biofilter bacteria for space and resources. Ozone is the most used disinfection method requiring expensive technology and trained stuff. Hydrogen peroxide (H2O2) has been on focus as a “green” alternative. High H2O2 doses are associated with acute toxicity symptoms for some species. Low doses are harmless, offer additional system oxygenation and contribute to water quality improvement. After testing the use of low doses in a small research RAS, the present study aims to describe the first case study up-scaling a continuous hydrogen peroxide application to commercial fish production in RAS with focus on the determination of required concentrations, application monitoring and variations on microbiome composition. Material and Methods The present study was performed at RAS facilities of Danish Salmon A/S in Hirtshals, Denmark, one of the European pioneers in rearing salmon in land based aquaculture facilities and producing about 1.200 metric tons/year of Atlantic salmon (Salmo salar). Two identical Grow-out RAS with own water treatment elements and eight pre-grow tanks were used as treatment and control systems to compare the effects of continuous H2O2 application in combination with ozone to common operational practice. Oxygen Cones and additional aeration stones ensured the basic Oxygen supply in the tanks. Defined H2O2 quantities were applied with a dosing lance connected to a peristaltic pump and an International Bulk Container with PERSYNT® 50, (EVONIK Industries) into the distribution pipe feeding a collection tank from where the water was evenly distributed to all tanks. Based on previous experiments, a final dosing of about 20 l/h was projected. Water samples were collected at the start (REF) and after slowly enhancing the dosage over time (24h and 30d), for the determination of the total microbial count (certified chromogenic Compact Dry TC plates from R-Biopharm), microbial viability (BacLight Viability Kit) and bacterial community composition (FISH). Water parameters (Ammonia, Nitrite, Nitrate, Phosphate, COD, Turbidity and H2O2 concentration) and production related information (feeding rate, fish biomass, oxygen consumption, ozone production) were also regularly evaluated. Results and Discussion H2O2 application started on August 06, 2020 with 1 l/h (0.51mg/L) and was increased over time up to 14 l/h (7.09 mg/L) (Fig1). On September 30, 2020 the application was stopped due to detected changes on feeding behavior of the fish as well as incorrect redox measurements in the treatment system which could probably be attributed to accumulation of oxidative species not having enough organic material to react. The test and control systems had similar biomass during the experimental period (test: 34.9-52.5 tones and control 36.4-51.4 tones) and the feed intake was accordingly adjusted (mean feed intake test: 496 kg/day; control 450 kg/day). The oxygen demand registered on the treated system (82 l/min - 90 l/min) was lower than the control (104 l/min - 115 l/min). In general, there was a reduction of turbidity and decreased nitrogen species and phosphate in the treated system. System maintenance (biofilter cleaning and backwashing) might had influenced the COD and total microbial counts measurements. Total microbial counts reflected a steady increase in the number of CFU/ml in both systems (Fig.2) and evidenced microbial accommodation. There was an increased rate of microbial mortality according to the viability results with higher values in the treated system reflecting the longer exposure of this system to oxidative stress (Fig 3). The community composition varied according to the treatment (Fig.4) and changes in the abundance of the different bacterial groups analyzed could be attributed to variable vulnerability of members of these groups to the disinfection potential of H2O2 or its combination with ozone.