Understanding the contribution of different natural water species to membrane fouling is essential for developing fouling control strategies for membrane-based drinking water treatment systems. A novel methodology capable of evaluating the fouling potential of major natural water fractions was developed. This approach identified correlations between normalized permeate flux and the accumulation of three major natural water fractions, humic substances (HS)-like, protein-like and colloidal/particulate-like matter deduced from the analysis of the principal components of the full fluorescence excitation-emission matrices collected during lab-scale cross-flow UF operations. Similar correlations were observed for HS-like matter content of the feed water, irrespective of the DOC and turbidity quality or fouling potential of feed water, while the protein-like and colloidal/particulate matter displayed different fouling behaviour according to the feed water characteristics. Based on these observations, a new component-specific fouling index, "Principal component-based Individual Fouling Index" (PIF index) was proposed for estimating the short term fouling potential of three major fractions present in natural water. The application of the PIF index for the identification of different UF membrane rejection characteristics was also demonstrated. The proposed approach is promising for monitoring fouling and understanding the impact of major foulant species present in industrial-scale UF drinking water treatment systems. (C) 2011 Elsevier B.V. All rights reserved.