A membrane photobioreactor (MPBR) plant was operated continuously for 3 years to evaluate the separate effects of different factors, including: biomass and hydraulic retention times (BRT, HRT), light path (Lp), nitrification rate (NOxR), nutrient loading rates (NLR, PLR) and others. The overall effect of all these parameters, which influence MPBR performance had not previously been assessed. The multivariate projection approach chosen for this study provided a good description of the collected data and facilitated their visualisation and interpretation. Forty variables used to control and assess MPBR performance were evaluated during three years of continuous outdoor operation by means of principal component analysis (PCA) and partial least squares (PLS) analysis. The PCA identified the photobioreactor light path as the factor with the largest influence on data variability. Other important factors were: nitrogen and phosphorus recovery rates (NRR, PRR), biomass productivity (BP), optical density of 680 nm (OD680), ammonium and phosphorus effluent concentration (NH 4 , P), HRT, BRT, air flow rate (F air) and nitrogen and phosphorus loading rates (NLR and PLR). The MPBR performance could be adequately estimated by a PLS model based on all the recorded variables, but this estimation 2 worsened appreciably when only the controlled variables (Lp, F air , HRT and BRT) were used as predictors, which underlines the importance of the non-controlled variables on MPBR performance. The microalgae cultivation process could thus only be partially controlled by the design and operating variables. A high nitrification rate was found to be inadvisable, since it showed an inverse correlation with NRR. In this respect, temperature and microalgae biomass concentration appeared to be the main factors to mitigate nitrifying bacteria activity.