Multivariate curve resolution alternating least squares is shown to be a powerful chemometric tool for investigation of main surface water contamination patterns affecting a particular geographical area over a period of time. When environmental monitoring data tables are analyzed using this method, the identification of the main contamination patterns, and the description of their geographical and temporal distribution profiles, can be obtained. To show the potential of the proposed method, the investigation of the pesticide contamination affecting the Ebro River delta (Catalonia, NE Spain) during the rice-growing season in 2005, is selected as a case study in this work. Three different contamination patterns of pesticides with different spatial end temporal behaviors were identified. A method validation using simulated data is then proposed to evaluate the suitability of the proposed multivariate curve resolution method for the analysis of the different possible date structures currently occurring in environmental monitoring studies. In particular, different data structures previously encountered in the experimental study of the Ebro River have been investigated in detail. The importance of using either raw or scaled data is contrasted using the simulated data sets. Possible propagation of noise on resolved profiles is also investigated to establish the difference between its effects and the possible ambiguities inherent to multivariate curve resolution methods.