The kinetics of ferrate(VI) decomposition in natural water were assessed in the absence and presence of natural organic matter (NOM) (pH = 7.50, [Fe(VI)] = 54 mu M, DOC = 0.00-10.00 and 1.00-8.57 mg/L for a simulated natural water and six real natural waters, respectively). Without NOM, Fe(VI) decomposition in simulated natural water exhibited a biphasic kinetics pattern, i.e. a 2nd-order reaction with respect to Fe(VI) concentration followed by a 1st-order reaction. However, an additional instant Fe(VI) loss was observed at the onset in the presence of NOM for both simulated and real natural waters, thereby rendering Fe(VI) decay with NOM a unique three-stage kinetics pattern. The initial instant Fe(VI) loss was caused due to the homogenous Fe(VI) self-decay and the rapid reactions between Fe(VI) and NOM. The latter accounted for a major fraction of the initial Fe(VI) loss and was in direct proportion with the initial DOC (DOC0) (Delta[Fe(VI)](0)(DOC)= alpha(NOM) DOC0; alpha(NOM) = 1.45 mu M Fe (VI)center dot L/mg DOC for the simulated natural water, and 0.66-1.35 mu M Fe(VI)center dot L/mg DOC for the six natural waters). Fe(VI) decomposition experiments with different Suwannee River NOM isolates revealed that hydrophobic NOM fractions (i.e. fulvic acid (FA) and humic acid (HA)) caused a more significant initial Fe(VI) loss than the hydrophilic group (HPI), suggesting that Fe(VI) preferentially reacted with hydrophobic NOM molecules rather than hydrophilic compounds. Furthermore, an approach developed for the estimation of aNOM revealed a linear correlation between alpha(NOM) and specific UV absorbance (SUVA) (alpha(NOM) = 0.27SUVA+ 0.18, R-2 = 0.71). This study provides essential information regarding Fe(VI) decomposition for the determination of Fe(VI) dose and exposure for effective water