Abstract In this paper we investigate the spatial extent and biogeochemical properties of a known CO2 plume using the pelagic transport-biogeochemical model BROM (Bottom RedOx Model). The model consists of a biogeochemical module, a 2-dimensional vertical transport module and gas bubble fate module, parameterizing bubbles rising and dissolution according to existing approaches. A controlled CO2 release experiment was carried out in the Horten Inner Harbor, Norway, in September 2018. This isolated bay is characterized by limited water mixing and anoxia in the bottom layer. CO2 was released at a water depth of 18 m either in a gas phase or dissolved in seawater at leak rates ranging from 0.1 l/min to 15.8 l/min. The chemical response to the release events relative to background variations was measured using chemical sensors mounted on two seabed templates located 4 m and 15 m from the release point, respectively, and compared to the values predicted by the model. The measurements show elevated levels of pCO2 and simultaneously decreased values in pH corresponding to the controlled release experiments. The model's simulations were in good agreement with the baseline observations and the measured changes forced by the experimental leak. The model predicts that after a continuous leak of this magnitude in stagnant conditions of anoxic bottom water, a 2–3 weeks long restoration period occurs, after which the disturbances disappear. This work confirms that the footprint of a potential CO2 leak is localized in the vicinity of the source (tens of meters) where it can be detectable with available chemical sensors.