Optical Oceanography Project

The first objective of this project was to characterize the variability of colored dissolved organic matter (CDOM) in the Optical Oceanography cruise data set. A second objective was to evaluate instrument closure between the ac9 and the spectrophotometer for measurement of CDOM absorption. The final objective was to evaluate the bio-optical models of HYDROLIGHT for local waters.

1. CDOM absorption was measured by spectrophotometry for four depths at nine stations. An average CDOM absorption coefficient (m^-1) at 400 nm and 440 nm was calculated for the water column at each station. The slope coefficient (S) was calculated for all four depths for two hydrocasts (BARNES 8/4/98-Strait of Juan de Fuca and 8/5/98-East Sound, Orcas Island) and all surface samples by fitting an exponential curve to the CDOM absorption data. CDOM absorption coefficients at 400 nm and slope coefficients were plotted against salinity to discern any correlation between CDOM and salinity.
2. CDOM absorption data from the discrete water samples (referred to as in situ) was compared with dissolved ac9 data from associated instrument package casts. This was accomplished by biasing the ac9 curve to the in situ curve at 676 nm. The in situ data was previously corrected at 690 nm or above for temperature effects and 715 nm could therefore not be used to bias the ac9 curve. This method may have resulted in a slight underestimation of CDOM absorption measured by both instruments. Slope coefficients were calculated for the ac9 data at the same depths as the discrete water samples. The ac9 data was plotted against the in situ data for regression analysis.
3. Three bio-optical models in HYDROLIGHT 4.0 (Mobley, 1998) were compared using cruise data from BARNES 8/4/98 for the Strait of Juan de Fuca. The first model (referred to as HL Chl, from Morel, 1991 and Gordon and Morel, 1983) was a Case 1 two-component model (pure water and pigmented particles) that used the input chlorophyll profile to estimate a and b. The second model (referred to as HL Chl + CDOM) was the same as the first but had a third component for CDOM and used the input chlorophyll profile and CDOM absorption coefficient at 440 nm and slope coefficient averaged over the water column. The third model (referred to as HL ac9) was actually direct input of ac9 data from which a and b were calculated.

1. The CDOM absorption coefficient at 400 nm varied from 0.22 to 0.36 m^-1 with a mean of 0.29 m^-1. The CDOM slope coefficient varied from 0.013 to 0.020 with a mean of 0.014. Neither of these two parameters showed significant variability with salinity over the measured depths. The Strait of Juan de Fuca and Griffin Bay had the lowest surface CDOM absorption coefficients at 400 nm, West Sound and Park’s Bay had the highest, and East Sound fell in between. An interesting feature of CDOM slope coefficient v. salinity was that the Strait showed a higher slope coefficient at the surface and East Sound showed a higher slope coefficient at depth.
2. The regression analysis of ac9 data plotted against in situ data resulted in an R² 0f 0.68 for the Strait and 0.91 for East Sound. Although it is based on only four data points per regression this result was good considering the long storage time of the discrete water samples prior to analysis.
3. Comparisons of the bio-optical models in HYDROLIGHT 4.0 were as expected for local (case 2) waters. The run with direct input of ac9 data was considered correct for comparison purposes due to good correlation with Rrs measured by Satlantic. The total absorption coefficients for the first two models varied from the ac9 run at most by a factor of 2. The total absorption coefficient calculated for the discrete water samples between 400 and 500 nm varied from the second model by less than 0.04 (absolute magnitude). The backscattering coefficient for the first two models varied from the ac9 run by factors of 2 to 3. Rrs for the first two models varied from the ac9 run by factors of 3 to 11. Rrs was also calculated with a and bb outputs by Rrs = 0.51(bb/a + bb). The calculated Rrs was slightly lower than HYDROLIGHT 4.0 Rrs.

1. Variability in CDOM absorption coefficient and slope coefficient did not appear to be a function of salinity in the cruise data set.
2. Acceptable instrument closure between the ac9 and the spectrophotometer was obtained for measurement of CDOM absorption in local waters.
3. The bio-optical models of HYDROLIGHT are not adequate for local waters due to high absorption by CDOM and relatively low particulate backscatter.