||In this work, a coupled biological-physical sea ice model was developed to investigate the influence of transient changes in environmental conditions (e.g. light, temperature and nutrient dynamics) on the sea ice biological communities. To simulate this complex environment, the biological model uses self-adapting physiological schemes and variable cellular N:C and Si:C quotas, including co-limitation, to decouple biomass accumulation and inorganic nutrient availability. Forcing terms (ice temperature, brine salinity, light and nutrients) are controlled by the thermodynamic sea ice model, with emphasis on the light and heat conduction through the sea ice and desalination processes (flux of brine). The model indicates that thermodynamic processes controlling ice formation and growth play a key role in the establishment and vertical distribution of biological sea ice communities. The coupled biological-physical model is described in a Lagrangian manner, where the time-dependent position of simulated ice floes extracted from ice velocity fields are used to compute forcing parameters (e.g. air temperature, oceanic heat flux, solar radiation). A good agreement between model results and field observations was found, indicating that the model well represents the physical and biological processes in sea ice. The model was also used to estimate the total productivity of sea ice in the Weddell Sea, resulting in an annual carbon production of 11 Tg C with a strong seasonal variability. The most productive months are between December and February, when light and temperature conditions in the sea ice are considerably enhanced as compared to the other months. Between May and September, the sea ice productivity ranges from 0.16 to 0.6 Tg C month-1 contributing 17% to the total annual production.