Primary productivity in the Arctic Ocean is vastly driven by single-celled phytoplankton. Within a comparably short growing season, they provide most of the carbon and energy for higher trophic levels and simultaneously influence biogeochemical cycles and the global climate. With the Arctic Ocean being one of the few regions that are expected to increase their productivity with future global change, it is especially important to understand how phytoplankton there is going to adapt. The potential for this adaptation to future climate change is often extrapolated from studies using single strains of a representative species that were cultured in laboratories for years. Since several decades, however, it is known that phytoplankton species and even local populations can exhibit large intraspecific diversity. During a field campaign on Svalbard, we isolated different strains of the Arctic diatom Thalassiosira hyalina from community incubations, which resembled present and future climate conditions. We then exposed these freshly isolated monocultures again to a matrix of CO and temperature. The results revealed that even within a single species, response patterns can differ greatly, comparable to variations seen between diatom species. Moreover, while only minor reactions of the communities were observed, the strain responses corresponded strongly with the previous selection environment. A strain isolated from future-like treatments, for instance, had its growth optimum at a higher temperature and pCO than another strain isolated from more present-like conditions. This suggests that intraspecific variability and the selection between coexisting ecotypes may be an underestimated source of species’ plasticity under changing environmental conditions and could ‘buffer’ functional species shifts. Adaptation of phytoplankton assemblages may therefore occur also by selection within rather than only between species, and species-wide inferences from single strain experiments should be handled with great care.