This report describes the validation exercise conducted on MERIS thanks to the MUMM data base. Two codes for the correction of the adjacency effect were used: SIMEC from VITO and ICOL from ADRINORD. The evaluation of these two codes is performed by comparison to in situ measurements collected off shore of Oostende. The results confirm the sensitivity of the MERIS L2 to the adjacency effect at least for the traditional scheme. The second algorithm based on a neural network which analyses all the reflectance spectrum appears to be robust to the adjacency effect and, at least on the 8 matchups were analyzed, quite accurate in the retrieval of the water reflectance.