The Belgian Coastal Zone (BCZ) in the Southern Bight of the North Sea is a highly dynamic and optically complex area. This is due to high non-algal particles (NAP) and coloured dissolved organic matter (CDOM) content which in spring adds together with undesirable blooms of the haptophyte Phaeocystis globosa. There is a need for improving the algorithms for chlorophyll a (chl a) retrieval in these highly turbid waters and for developing algorithms for species detection in order to attempt to create an early warning bloom system. This information will contribute to the knowledge of the extent and magnitude of the P. globosa bloom in Belgian waters. In this study, pure cultures of the main taxa present in the BCZ, diatoms and P. globosa, were combined with field measurements of light absorption of total particles, phytoplankton and dissolved material, pigment determination and phytoplankton counts to address the main objectives. Sampling was performed during 8 cruises covering winter, spring, summer and late summer, and along nearshore-offshore gradients from 2003 to 2006. The area is characterised by a high spatio-temporal variability of IOPs due to the high dynamics of the area in terms of currents, salinity gradients and biological production. During spring the presence of P. globosa modulates the IOPs across all the area, the particle absorption is significantly higher than summer and there is no significant coast-offshore variability for phytoplankton and CDOM. The design of chl a retrieval algorithms assumes negligible absorption of NAP and CDOM in the near infrared (NIR) and the use of a fixed value of specific phytoplankton absorption. It is shown that neglecting the NAP and CDOM absorption in the NIR will have a significant overestimation impact in retrieval of chl a. On the other hand, the specific phytoplankton absorption was found to be highly variable (0.015 „b 0.011 m2 mg chl a-1). Both results will affect directly the retrieved chl a. The spatial variability of CDOM was significant varying between 0.20-1.31 m-1 in the marine area and between 1.81-4.29 m-1 in the Scheldt estuary. CDOM was found to be related to salinity with conservative mixing within the Scheldt estuary and during some seasons in the BCZ, however deviations from conservative mixing suggest other inputs to the CDOM pool. Analyses of the spectral slope of the CDOM absorption curve revealed two main CDOM pools in the area, an allochthonous one delivered by the Scheldt estuary and an autochthonous one associated with the phytoplankton spring bloom decomposition. Algorithms for CDOM retrieval will be affected if the variability in the relation between CDOM and salinity is not taken into account. The optical characterisation of diatoms and P. globosa from the BCZ in pure cultures revealed that the main differences in the phytoplankton absorption spectra were found at 467 and 500 nm corresponding to the absorption of the pigments chlorophyll c3 (chl c3) characteristic of P. globosa and fucoxanthin, respectively. Accordingly, both the absorption at 467 nm and the ratio 500/467 nm were successfully used to discriminate the two taxa in cultures and field samples. This latter indicator was not preserved in the reflectance signal due to degradation of the signal when passing from absorption to reflectance, and thus could not be used for algorithm development. The spectral feature at 467 nm was later used as the basis for the development of a flag-type algorithm to detect chl c3 using either absorption or water-leaving reflectance data. Also, the correlation between the algorithm¡¦s retrieved chl c3 and P. globosa cell number allowed the quantification of the bloom. The main findings of this thesis highlight the importance of the IOPs characterisation for the improvement and development of ocean colour retrieval algorithms in these highly complex waters.

Bio-optical variability of surface waters in Northeastern Gulf of Mexico (NEGOM) was examined using satellite and in situ data. Relatively high chlorophyll-a concentration (chl>=1 mg m-3) and high colored dissolved organic mater (ag443>=0.1 m-1) were generally observed inshore, near major river mouths, and in plumes of Mississippi River water that extended offshore during the three consecutive summer seasons (1998, 1999, and 2000). River discharge dominated chlorophyll-a concentration variability inshore, particularly near major river mouths. Strong interannual variability in chlorophyll-a concentration was observed inshore from Escambia to Tampa Bay region during the winter to spring transition, which was different in 1998 compared to the winter to spring transition in 1999 and 2000. This was related to higher fresh water discharge during the 1997-1998 El Niño-Southern Oscillation event as well as strong upwelling in spring 1998. The Mississippi plume extended >500 km southeast of the Mississippi delta and up to the Florida Keys was observed for the periods extending over 14 weeks between May and September every year of the study. In general, ag443 covaried linearly and inversely with salinity inshore during spring and fall, indicating conservative mixing. The NEGOM salinity-ag443 relationship of fall 1998, i.e., Salinity=36.59-29.86*ag443 (n=8771, r2=0.86; 0.01<=ag443<=0.52, 16 <=S<=36), served as the best predictor of NEGOM salinity based on in situ ag443 observations for spring and fall seasons from all years (<3% mean percentage errors; corresponding to <1.03 psu). This may help estimate salinity from satellite ocean color data, but further testing using data from multiple years is needed to improve such relationship. While river discharge was an important source of colored dissolved organic matter (CDOM), phytoplankton blooms also contributed to CDOM formation in the NEGOM. Using a pigment index of phytoplankton taxonomic groups, the variability in biomass proportion of microphytoplankton explained up to 76% of the variability of the average of normalized phytoplankton absorption coefficients (545, 625, and 673 nm). The clorophyll-specific absorption coefficient, a*ph(440), varies by a factor of 7 (0.02-0.15 m2mg-1). Particle size and pigment composition played important roles in determining a*ph(440) variability. This must be accounted for in chlorophyll-a concentration algorithms based on aph.

The objective of this dissertation is to gain a better understanding of the environmental and climatic effects on the temporal and spatial variability of phytoplankton biomass along the West Florida Shelf. Chapter 1 examines temporal and spatial patterns in chlorophyll concentrations using satellite data collected between 1997 and 2003. Chlorophyll data derived from the SeaWiFS sensor are validated with in-situ data and analyzed. Wind, current, sea surface temperature, river, and rain data are used to better understand the factors responsible for the patterns observed in the satellite data. My question is whether the standard OC4 algorithm is adequate for studying short-term variability of chlorophyll concentrations along the WFS. I will examine temporal and spatial trends using the OC4 and compare them to the Carder semianalytical algorithm which uses remote sensing reflectances at 412nm, 443nm, 490nm,and 555nm to estimate chlorophyll concentrations separately from CDOM estimates. In Chapters 2 and 3 the potential problems due to CDOM and bottom reflectance are examined. In Chapter 2 I analyze the influence of riverine induced CDOM. Water leaving radiances are analyzed in an effort to discriminate true chlorophyll patterns from CDOM contaminated signals. Chapter 3 examines the impact of bottom reflectance on the satellite signal by using the percentage of remote sensing reflectance at a wavelength of 555 to differentiate between optically shallow waters and optically deep waters. Optically shallow waters are defined as those with the percentage of Rrs at 555 due to bottom reflectance greater than or equal to 25 percent, while optically deep waters have percent bottom reflectance less than or equal to 25 percent. These analyses will help assess the validity of the temporal and spatial patterns ofchlorophyll concentration observed with the SeaWiFS data described in Chapter 1.

The Belgian Coastal Zone (BCZ) in the Southern Bight of the North Sea is a highly dynamic and optically complex area. This is due to high non-algal particles (NAP) and coloured dissolved organic matter (CDOM) content which in spring adds together with undesirable blooms of the haptophyte Phaeocystis globosa. There is a need for improving the algorithms for chlorophyll a (chl a) retrieval in these highly turbid waters and for developing algorithms for species detection in order to attempt to create an early warning bloom system. This information will contribute to the knowledge of the extent and magnitude of the P. globosa bloom in Belgian waters. In this study, pure cultures of the main taxa present in the BCZ, diatoms and P. globosa, were combined with field measurements of light absorption of total particles, phytoplankton and dissolved material, pigment determination and phytoplankton counts to address the main objectives. Sampling was performed during 8 cruises covering winter, spring, summer and late summer, and along nearshore-offshore gradients from 2003 to 2006. <p>The area is characterised by a high spatio-temporal variability of IOPs due to the high dynamics of the area in terms of currents, salinity gradients and biological production. During spring the presence of P. globosa modulates the IOPs across all the area, the particle absorption is significantly higher than summer and there is no significant coast-offshore variability for phytoplankton and CDOM. <p>The design of chl a retrieval algorithms assumes negligible absorption of NAP and CDOM in the near infrared (NIR) and the use of a fixed value of specific phytoplankton absorption. It is shown that neglecting the NAP and CDOM absorption in the NIR will have a significant overestimation impact in retrieval of chl a. On the other hand, the specific phytoplankton absorption was found to be highly variable (0.015 „b 0.011 m2 mg chl a-1). Both results will affect directly the retrieved chl a. The spatial variability of CDOM was significant varying between 0.20-1.31 m-1 in the marine area and between 1.81-4.29 m-1 in the Scheldt estuary. CDOM was found to be related to salinity with conservative mixing within the Scheldt estuary and during some seasons in the BCZ, however deviations from conservative mixing suggest other inputs to the CDOM pool. Analyses of the spectral slope of the CDOM absorption curve revealed two main CDOM pools in the area, an allochthonous one delivered by the Scheldt estuary and an autochthonous one associated with the phytoplankton spring bloom decomposition. Algorithms for CDOM retrieval will be affected if the variability in the relation between CDOM and salinity is not taken into account. <p>The optical characterisation of diatoms and P. globosa from the BCZ in pure cultures revealed that the main differences in the phytoplankton absorption spectra were found at 467 and 500 nm corresponding to the absorption of the pigments chlorophyll c3 (chl c3) characteristic of P. globosa and fucoxanthin, respectively. Accordingly, both the absorption at 467 nm and the ratio 500/467 nm were successfully used to discriminate the two taxa in cultures and field samples. This latter indicator was not preserved in the reflectance signal due to degradation of the signal when passing from absorption to reflectance, and thus could not be used for algorithm development. The spectral feature at 467 nm was later used as the basis for the development of a flag-type algorithm to detect chl c3 using either absorption or water-leaving reflectance data. Also, the correlation between the algorithm¡¦s retrieved chl c3 and P. globosa cell number allowed the quantification of the bloom. The main findings of this thesis highlight the importance of the IOPs characterisation for the improvement and development of ocean colour retrieval algorithms in these highly complex waters.<p> / Doctorat en Sciences / info:eu-repo/semantics/nonPublished