Ecological modelling is widely used in the various fields of ecology but models usually require large datasets, a serious limitation to the approach for application to organisms of remote and little studied regions such as polar seas. Correlative and mechanistic modelling approaches are usually used independently in distinct studies. Using both approaches in integrative, hybrid models however can help better estimate the species realised niche, as mechanistic and correlative models complement each other very well, giving more insights into species potential response to fast changing environmental conditions. In this study, we implemented for the first time an hybrid, correlative and mechanistic model to predict the response of a marine invertebrate endemic to the Southern Ocean, the sea urchin Abatus cordatus (Verrill, 1876). We compared the respective performance of simple and hybrid models by analyzing the effect of seasonality on species distribution, a key feature of ecosystem functioning at high latitudes. Higher performances were obtained for the ‘integrated Bayesian’ approach compared to simple mechanistic and correlative models. The hybrid model more precisely predicts the effect of seasonality on habitat suitability. Such results are promising and show that hybrid approaches can be applied to case studies for which limited datasets are available.