Fish spawning phenology is a major concern for conservation and fisheries management. New intensive data sources such as GPS-based tracking data or high resolution catch declaration data are progressively becoming available in the field of marine ecology. These benefit from high spatio-temporal resolution and open new research avenues to investigate inter-annual and seasonal variability of phenology. In this paper, we illustrate how catch declarations modeling coupled with spatio-temporal dimension reduction methods known as Empirical Orthogonal Functions (EOF) can be used to synthetize spatio-temporal signals in fish distribution; Specifically, we address the following questions; (1) can we identify spatio-seasonal patterns that can be interpreted in terms of seasonal migration between essential habitats? (2) can we identify changes in the phenology? (3) are those changes related to environmental drivers? The analysis is illustrated through the analysis of the reproduction phenology on three key commercial species in the Bay of Biscay (Hake, Sole and Sea Bass). The EOF analysis on these species emphasizes strong seasonal spatio-temporal patterns that correspond to migration patterns between feeding areas and reproduction areas. Based on this methodology, we identify seasonal variations in the timing of the reproduction and we relate these to Sea Surface Temperature, a key driver of fish reproduction.