Welcome to Gaia! :: View User's Journal | Gaia Journals

Figure options
3.5. PARAFAC modeling of BEPOM in estuaries
The PARAFAC model was fit to NRE BEPOM EEM data only because most of the EEMs in this Cyt387 study came from the NRE. Five components were split-half validated in the BEPOM-PARAFAC model in rank order of their decreasing explanation of variability (Fig. 5). The components were spectrally similar to prior PARAFAC model components from a variety of freshwater, estuarine, and marine models (e.g., Fellman et al., 2011, Jørgensen et al., 2011 and Osburn et al., 2012).
Fig. 5.
Results of a PARAFAC model fit to BEPOM EEMs from the NRE. Contour plots are shown along with split-half validation results for each of five components.
Figure options
3.6. PARAFAC model component distributions in coastal waters
Fig. 6.
Component distributions (C1–C5) and Chl-a concentrations for (A) February, (B) March, and (C) August 2012 in the NRE.
Figure options
Analysis of PARAFAC component distributions from BEPOM fluorescence modeled in Charleston Harbor, inner shelf of the South Atlantic Bight (SAB), and northern GOM supported our findings in NRE (Fig. 7). Surface water BEPOM samples extending from the tidal Ashley (Station 17), Cooper (Stations 0, 6A, 6E), and Wando Rivers (Station A3), through Charleston Harbor (Station A1), and out to the shelf of the SAB (Station OS01) were strongly dominated by C1 and C4, like estuarine station NR180 in NRE (Fig. 7A). C2 fluorescence was minimal in crossing-over samples. However, the rivers had decidedly more of a terrestrial influence due to the high FMax values for C3, and to a lesser extent C2, than did the Harbor and Shelf samples.