Abstract
Oceanic buffering in response to the increase in atmospheric pCO2 due to anthropogenic emissions has led to a decrease in seawater pH and carbonate ion concentration ([CO32 ]) in a process known as ocean acidification (OA). The decrease in [CO32 ] also results in a reduction in the saturation state (Ω) of calcium carbonate (CaCO3) minerals. As CaCO3 plays an important role in the global carbon cycle and the health of many marine species, the effects of OA are predicted to cause unprecedented challenges and alterations to the function, structure, and distribution of marine ecosystems. Ω provides a convenient means by which to summarise the changing carbonate chemistry onto a biologically and geochemically relevant scale that can be used as an indicator of the progression of OA in the open ocean and coastal waters.
Accurate estimation of Ω is crucial to monitoring the environmental progress of OA, as well as for research into biological processes such as calcification and dissolution kinetics. However, Ω cannot currently be measured directly. Instead, it must be calculated from the calcium ion concentration ([Ca2+]), [CO32 ], and the stoichiometric equilibrium solubility product, "K" _"sp" ^"*" . [Ca2+] is routinely calculated using a conservative relationship with salinity, while [CO32 ] is determined by carbonate system characterisation. At present, the relatively large uncertainty in the "K" _"sp" ^"*" values, defined most recently in 1983, is the limiting factor in the accuracy of Ω estimates. Thus, the overarching objective of this thesis is to increase the accuracy of the CaCO3 saturation state assessment, on which our ability to monitor and understand the impacts of anthropogenically driven climate change on the marine environment are predicated. To this end, the thesis is separated into two parts.
The first part (Chapters 2-6) explores the implementation of a new optical fibre based sensor into three new sensor systems designed for the purposes of 1) improving the accuracy of "K" _"sp" ^"*" estimates, 2) re examining the pressure coefficients currently used to account for pressure effects on "K" _"sp" ^"*" , and 3) in situ monitoring of marine CaCO3 saturation conditions. The new sensor, termed the Ω probe, is sensitised to the seawater saturation conditions via functionalisation of the exposed core by deposition of CaCO3 of a known morphology (i.e., calcite or aragonite). Changes in the CaCO3 coating due to dissolution or precipitation are detected as changes in the monitored intensity of a propagating light signal.
The first new sensor system, developed for the purpose of improving the accuracy of "K" _"sp" ^"*" estimates, is presented in Chapter 4. This system involves the integration of the Ω probe (Chapter 2) into an analytical procedure designed to predictably manipulate seawater CaCO3 saturation conditions across typical seawater conditions (salinity and temperature ranges of 15 35 and 3 35 °C, respectively). This procedure, developed as part of this study and termed the Ω titration (Chapter 3), was demonstrated to be capable of reproducibly manipulating [CO32 ] of a seawater sample to a given target value with an accuracy of < 5%, which is less than the current uncertainty in "K" _"sp" ^"*" estimates. The response of the Ω probe, characterised through use of the Ω titration, was shown to be sensitive to the saturation conditions of the surrounding seawater sample. However, the observed response of the Ω probe was not consistent with that expected based on the demonstrated solubility behaviour of CaCO3 in seawater. A thorough investigation into experimental parameters that could have led to this large discrepancy was unable to successfully correct or account for the disparity between the expected and observed behaviour. It is proposed that the observed response of the Ω probe is complicated by the combined and competing effects of CaCO3 precipitation and Ostwald ripening. The understanding of the Ω probe response with respect to the surrounding seawater saturation conditions is currently insufficient to be used to determine, and thereby improve, estimations of "K" _"sp" ^"*" . Nevertheless, as the Ω probe response may be sufficiently characterised in the future, the two other implementations of the Ω probe were still explored.
The second new sensor system was developed to re examine the "K" _"sp" ^"*" pressure coefficients (Chapter 5). This project involved the design and fabrication of a high pressure reaction cell and a new method for the manipulation of carbonate chemistry through changes in pressure. Termed the pressure Ω titration, the method was demonstrated to extend the potential "K" _"sp" ^"*" determination capabilities of the Ω titration system to pressures up to 1000 bar (corresponding to water depths of up to 10 km).
The third sensor system developed in this work (Chapter 6) is a fully self contained and autonomous deployable Ω sensor system based on an embedded controller. The system was tested, and a prototype developed. Potential applications of this technology include the in situ detection of undersaturation events in highly variable coastal waters and direct evaluation of the oceanic saturation horizon during depth profiling. The newly developed system is fully functional and validated in suitable test environments. As soon as the Ω probe response is more clearly understood, this device can be used to monitor Ω both in the laboratory and in the field.
The second part of the thesis focuses on the examination of potential impacts of variable [Ca2+] on data quality collected by current and future coastal OA monitoring programs. The key goals of this work included 1) the development of the uncertainty propagation software (through modification of the R based seacarb library) to include propagation functionality for the uncertainty in [Ca2+], 2) extend the Global Ocean Acidification Observing Network’s (GOA ON) quality goals in the development of a framework by which to assess the potential impacts of non conservative [Ca2+] on OA monitoring data quality, and 3) the demonstration of the developed methodology in a case study of [Ca2+] salinity behaviour in in a range of different New Zealand coastal sites.
The general implications of a non conservative [Ca2+] salinity relationship for OA monitoring data quality were explored using the developed software and framework in Chapter 7. This work was subsequently applied in a case study of the NZOA ON monitoring program (Chapter 8), where it was determined that variability in the [Ca2+] salinity relationship is unlikely to affect data quality due to the high accuracy and precision of NZOA ON’s carbonate system analysis. The case study is intended to provide a pilot example for other regional OA monitoring programs to determine if routine [Ca2+] analysis is required to improve or maintain data quality.
Overall, this thesis has provided substantial insight into the response of the Ω probe to typical marine conditions, marking the first step towards new methods for accurate "K" _"sp" ^"*" determination and in situ monitoring of marine saturation conditions. Furthermore, the methods developed for the assessment of the impacts of non conservative [Ca2+] will aid in the collection of reliable OA data in coastal and estuarine waters.