The phytoplankton absorption cross-section is a fundamental quantity in biogeochemical ocean models that alters the underwater spectral light field and the photosynthetic response of phytoplankton. Phytoplankton taxa are characterized by absorption spectra with defined absorption bands in the visible region of the light spectrum that govern the capability of different taxa of phytoplankton to exploit light as a resource for growth. The interplay between the spatial and temporal gradients of underwater spectral light and the light absorption characteristics of different phytoplankton types contribute to the interactions and selection among groups and hence can be a determinant in shaping phytoplankton diversity and biogeography. In this work, we used a biogeochemical model of the Mediterranean Sea to simulate nine optically different phytoplankton functional types (PFTs) and coupled it to a radiative transfer model to simulate the spectral underwater light field where the PFTs interact. We investigated the competitive advantage provided by the different absorption spectra and the possibilities of coexistence emerging from the availability of different light habitats. By considering non-spectrally resolved optical differences among PFTs, our model results led to the dominance of Synechococcus sp. within picophytoplankton, coccolithophores within nanophytoplankton and dinoflagellates within microphytoplankton. By including the spectral dependency of photosynthesis, we observed how the availability of different spectral light habitats led to the coexistence of picophytoplankton groups in clear waters, whereas for nano- and microphytoplankton PFTs, the extent of coexistence permitted by light fluctuations was minimal. By combining optical and functional differences in the PFT description, we obtained PFT distributions that were in relative agreement with observed phytoplankton distributions in the Mediterranean Sea estimated from satellite and in situ high-performance liquid chromatography sampling. This result suggests that combining spectrally resolved optical traits with functional traits in the description of the autotrophic community can be a valuable tool to improve simulations of the diversity and biogeography of primary producers in the ocean.