Dissertation Visualizations
The ocean can act as a sink of carbon dioxide from the atmosphere. One of the most important mechanisms of ocean carbon uptake is the biological carbon pump. This pump describes the process in which phytoplankton converts CO2 into organic carbon in the surface ocean. Particulate organic carbon can then sink out of the mixed layer into the mesopelagic zone where it could be remineralised back into CO2 and escape back to the atmosphere. Alternatively it can continue to sink eventually entering the deep ocean. Once carbon has reached the deep sea, it can stay trapped there for hundreds of years, effectively hiding from the atmosphere. This process mitigates anthropogenic climate change. Our understanding of the dynamics of carbon is hindered by logistical and technological hurdles of quantifying the intrannual variability of carbon dynamics in the mesopelagic. Optical proxies are a promising new way to measure the seasonal evolution of carbon in the water column. In this dissertation, high resolution optical data collected on cruise DY086 in November and December 2017 is used to determine the temporal evolution of interior carbon stocks in South Georgia. In the surface layer, chlorophyll a and POC concentration decreased over the time period of the cruise, with evidence of a flux event occurring at the end of November. Backscattering spike frequencies, used as a proxy for large particles in the water column, had attenuation coefficients ranging from 0.51-1.77 (95% C.I.) depending on phase of the bloom, methodology, and platform used. Particulate organic carbon flux was estimated to range from 144-285 mg C m-2 d-1 (95% C.I.) at 200 m. Comparisons of optical spikes obtained using different platforms and methods suggested that the magnitude of optical spikes is sensitive to either platform type, velocity of the optical sensor or both. Conversely, optical spike frequency attenuation with depth appeared to be less sensitive to methodological differences. This suggested that optical spike data is not comparable across different methodologies, depending on analysis type. Nevertheless, attenuation coefficients and flux estimates compare reasonably well with literature values from the same region or similar methods, suggesting it is a robust method of describing interior carbon dynamics. Future work should focus on fully understanding the impact of platform type and sensor velocity on optical spike data, as well as incorporating sediment trap data from cruise DY086 into the calculations provided in this project.
Objective: use high resolution optical data collected on COMICS research cruise to determine the temporal evolution of particulate organic carbon (POC) in South Georgia.
-
Describe the temporal evolution of chlorophyll a and POC at station P3.
-
Quantify particle attenuation and flux below the surface layer using optical spike data.
Why is this exciting? This is the first time high-resolution optical backscatter from ship-based measurements have been used to estimate POC flux in South Georgia.
Figure I. Chlorophyll a calibration. Two Model II regressions were run to calibrate chlorophyll optical data to chlorophyll bottle data (collected on the CTD). (a) Corrected CTD chlorophyll data was calibrated with collocated ECO Triplet data. (b) This same equation was used to calibrate the ECO Triplet profiles from the RCF and collocated RBR chl data was calibrated. (c) An example of before and after calibrating chl values. Raw chl (orange; ‘before’), calibrated chl a (black; ‘after’), and the ECO Triplet data used to calibrate (blue) on one RCF cast.
Figure II. Separation of optical spikes from baseline. _Baseline (red) is representative of background concentration of POC, while spikes (black) are representative of larger particles and aggregates. Baseline and spikes are separated using a running median.
Figure III. Summary diagram of the temporal evolution of salinity, temperature, POC concentration and chlorophyll a concentration in the upper 250 m at P3. P3A (earliest visit) is denoted with light blue, P3B with dark blue, and P3C (latest visit) in magenta. (a) Mean temperature profiles become warmer in the surface layer over time, developing a step-like profile by P3C. Winter Water (cold mass around 150 m) slightly shallows with time. (b) Mean salinity profiles increase slightly at approximately 150 m with time but stay relatively constant at the surface. (c) POC concentration decreases in the surface layer, with increased concentration at approximately 150 m visible during P3B. (d) Chl concentration decreases in the surface layer over time with slightly increased concentrations below the surface layer at P3B and P3C compared to P3A
Figure IV. Oceanographic setting at P3 from November-December 2017. Over time, temperature increased in the surface layer (upper 50 m) while interior temperature and salinity stayed relatively constant. (a) A T-S diagram shows warming in the surface layer, and a cold and fresh Winter Water (WW) layer sitting beneath. (b) Temperature scattered over time (x-axis) and depth (y-axis). Here, a general warming pattern including thermal stratification in the surface layer is visible, with a cold WW layer located beneath (approximately 150 m). Coloured boxes show time periods associated with visits P3A, P3B and P3C.
Figure VI. Temporal evolution of chlorophyll a and POC concentration in the upper 800 m at P3. Both chl and POC are plotted in log space with time on the x-axis and depth on the y-axis. Coloured boxes at the middle of the figure indicate time periods that each visit occurred (P3A= cyan; P3B= dark blue; P3C= magenta). (a) Chlorophyll a concentration decreased in the surface layer during the time period of the cruise, with a signal of a flux event occurring at the end of November visible at approximately 150 m (P3B). Concentrations under the surface layer remain elevated for the duration of the cruise. (b) Patterns in POC concentration are similar to those in chlorophyll. POC concentrations decrease in the surface layer over time, with elevated concentrations seen under the surface layer starting the end of November (P3B).
Figure VII. Comparison of spike frequency attenuation derived from different platforms and noise thresholds. Data combined across all visits to P3. (a) Spike frequency attenuation derived from platform-specific noise thresholds. Here, the b-value from the RCF is significantly higher than the b-value from the CTD. Binned spike frequencies from the RCF are overall slightly lower than those from the CTD for the same depth bin. (b) Spike frequency attenuation derived from equal noise thresholds. Although the b-value from the RCF was slightly higher, b-values from the RCF and the CTD were not significantly different (95% C.I.s do not overlap). Spike frequencies from the RCF are higher than the CTD at every depth bin in which the RCF collected data.