Résumé : Recent field observations of Phaeocystis globosa colonies (number, size and cell content) and historical ecosystem model simulations have been used to define and diagnose undesirable eutrophication in the Belgian coastal zone (BCZ) as recommended by the EU Water Framework Directive and OSPAR. The proposed methodology included an initial definition of a reference abundance of Phaeocystis cells from which the disturbance could be scaled. This number was based on the Phaeocystis attribute that creates ecosystem disturbance: the ability of Phaeocystis colonies to grow and reach sizes unmanageable for the current copepods. The threshold was set at 4 × 106 cells L− 1 and corresponded to the maximum Phaeocystis cells contained in the grazable colonies (< 400 µm) recorded in the BCZ between 1988 and 2000. Interestingly, this cell number was similar to the maximum number of Phaeocystis colony cells simulated for pristine time by the coupled RIVERSTRAHLER-MIRO (R-MIRO) model, suggesting that natural Phaeocystis ecosystems are well balanced and efficiently transfer their production to higher trophic levels. However, both the field observations obtained during the 1988–2000 period and the 1950–2000 historical reconstruction of eutrophication in the BCZ simulated by the R-MIRO model showed Phaeocystis colony cell maxima well above the threshold value. Considering that a return to the pristine reference value is not achievable, another methodology was then proposed based on historical R-MIRO simulations of annual primary (PP) and copepod (CP) productions where the CP:PP ratio can be considered an indicator of trophic efficiency. The analysis of changing CP:PP ratios in relationship with simulated winter concentrations and annual inputs of nutrients showed that the decrease in trophic efficiency was related to the imbalanced DIN and DIP inputs (N:P > 25) explained here by the reduction in DIP loads after 1985 while DIN remained elevated. On this basis, a critical DIN load estimated at 60 kT y− 1 was proposed to re-equilibrate the N:P balance in the BCZ and to favour diatoms over Phaeocystis colonies. More generally we conclude that N:P ratios in nutrient loads (or winter stocks in the coastal area) in excess of 25 in Phaeocystis ecosystems are indicative of Phaeocystis colony dominance and weak trophic efficiency.