AN INTER-COMPARISON OF COLOURED DISSOLVED ORGANICMATERIAL AND PHYTOPLANKTON ABSORPTION COEFFICIENTS:IMPLICATIONS FOR MERIS DATA VALIDATION

摘要

Understanding differences between laboratories in the measurement of absorption coefficients ofrnphytoplankton, detrital and coloured dissolved organic material is important if we are to accuratelyrnvalidate MERIS data. Phytoplankton absorption coefficients (aphyto) and the absorption coefficient ofrncoloured dissolved organic material (arnys) were determined by three laboratories on natural seawaterrnsamples taken from three stations off Plymouth Sound between 50 20.02 °N, 4 10.02 °W, during arnsecond inter-comparison exercise (PlymCal 2). The absorption coefficient of an artificial CDOMrnsample composed of a coloured humic acid salt suspended in buffered MilliQ water was alsorndetermined by the three laboratories. For the natural seawater samples, good agreement was foundrnbetween arnphyto at 442 nm at three stations where a aphyto phyto varied between 0.066 and 0.145 m-1 . OverrnMERIS bands (412 to 680 nm) there was a 14 % difference between laboratories and the standardrndeviation was small (mean of all labs at MERIS bands ～0.003 rn M-1 ). The range in coloured dissolvedrnorganic material absorption coefficient at 442 nm was from 0.1 at Station L4 (50 15 °N, 04 12.5 °W)rnwhen there was a 6 % difference between laboratories to 1.0 rn M-1 -at Barnpool (50 21.3 °N, 04 10.1 °W),rnwhere there was a 48 % difference between laboratories. For the artificial CDOM sample the differencernwas 13 % at 442 nm and 20 % over MERIS bands.rnUsing a aphyto phyto and a ays ys values at Station L4 for the three laboratories during PlymCal 2, abovernwater surface reflectance and water leaving radiances were calculated and then SeaWiFS OC4rnswitching band ratios was applied to calculate corresponding SeaWiFS Chlorophyll a (Chla)rnconcentrations. The results showed that the variation in arnys at L4 produced on average an 11 %rnvariation in Chla and that the variation in rnaphytoresulted in an 18 % variation in Chla. The variation inrnChla using in situ absorption coefficients at L4 was less than the accuracy of SeaWiFS Chla using thernOC4 algorithm (～35 %), which suggests that once methodologies have been fixed, the betweenrnlaboratory variation in rnaphyto phyto and ays at L4 is acceptable and will not significantly affect the validationrnof MERIS data.