A rapid, precise potentiometric determination of total alkalinity in seawater by a newly developed flow-through analyzer designed for coastal regions

Abstract

A flow-through analyzer has been developed for the rapid, potentiometric determination of total alkalinity (TA) in seawater in order to capture large temporal and spatial variations of the CO₂ system in coastal zones, where the carbon cycle is still not well understood despite the potential large contribution of such areas to the global carbon cycle. This analyzer requires small amounts of seawater (3 ml min⁻¹), and has realized continuous measurements of TA with a response time of 4–5 min, with long-term precision of about 0.1% (ca. 2 μmol kg⁻¹), and accuracy of a similar level over a range of TA (more than 200 μmol kg⁻¹) when appropriate standard solutions are used. This analyzer is fully automated and can achieve stable measurements (within 0.15%; 3 μmol kg⁻¹) over 1–2 days without regular recalibration for drift, which will enable us to carry out continuous, in situ measurements of TA in coastal waters.

Introduction

The world ocean is estimated to be taking up 1.7 GtC per year, which is almost 30% of the CO₂ released anthropogenically into the atmosphere (Prentice et al., 2001). Much attention has been paid to the open ocean in many programs such as the Joint Global Ocean Flux Study (JGOFS) and the International Geosphere Biosphere Programme (IGBP) because of its large area. CO₂ dynamics in the coastal ocean, however, have generally received less attention (Chen et al., 1994).

Several studies have suggested the importance of the CO₂ dynamics in the coastal ocean despite its small fraction (8%) of the total ocean area. Between 15% and 50% of the oceanic primary production is now attributed to coastal ocean Walsh, 1991, Muller-Karger, 2000, and recent studies have concluded that some continental shelves, in general the zone shallower than 200 m, act as a sink for atmospheric CO₂ Tsunogai et al., 1999, Frankignoulle and Borges, 2001, of up to 0.6 GtC per year worldwide (Yool and Fasham, 2001), which is about 30% of the oceanic CO₂ uptake. Some indirect studies also support the concept of excess production in coastal ocean and its transportation to the deep ocean as dissolved organic carbon (DOC) and particulate organic carbon (POC) Duarte and Agustı́, 1998, Bauer and Druffel, 1998. Smaller scale studies are needed to elucidate the mechanisms by which CO₂ is absorbed in coastal ocean in conjunction with nutrient and organic matter cycling. This requires high time-resolution measurements of the CO₂ system in these coastal seawaters.

There are four measurable parameters of the seawater CO₂ system: pH, total alkalinity (TA), total dissolved inorganic carbon (DIC), and the fugacity of CO₂ (fCO₂). Measuring two of these parameters allow us to fully characterize the CO₂ system in seawater, i.e., the rest of the parameters can be calculated from thermodynamic relationships.

TA measurements act as a water mass tracer. For instance, salinity-normalized TA is generally constant in subtropical gyres except in areas of upwelling, where it increases (Roche and Millero, 1998). Calcification by coccolithophorids, corals, coralline algae, etc., decreases TA in a predictable way, so that TA data can be used as a measure of the ecosystem calcification as well. For detecting these changes in coastal areas, the combination of manual sampling and the conventional titration methods (e.g., Millero et al., 1993, Haraldsson et al., 1997), which use a stepwise addition of HCl to a known volume of sample, are not rapid enough.

For rapid and simple determination of TA, the single-addition technique is used (e.g., Anderson and Robinson, 1946, Culberson et al., 1970, Culberson, 1981, Perez and Fraga, 1987). This method is fairly simple compared with the conventional titration method. A certain amount of HCl solution is added to a seawater sample, air is bubbled through the solution to remove CO₂, the pH of the acidified solution is measured, and TA is calculated from the pH value (Culberson et al., 1970). This technique is simpler and thus easier to automate than the conventional titration technique.

Roche and Millero (1998) developed a flowing total alkalinity system (FLTA) with an adaptation of the spectrophotometric method to determine TA based on the single addition technique. The FLTA realized reasonably rapid sampling (ca. every 15 min) and good precision (2–3 μmol kg⁻¹). This system, however, needs precise calibration for the pipette, the titrator, and the sample loop, and also requires an accurately prepared HCl solution. The FLTA is not a completely flowing system but rather measures discrete samples sampled from a continuous seawater supply. This does not provide sufficient time resolution for the in-situ measurement of TA in coastal environments.

Recently, we developed a flow-through analyzer for total alkalinity in seawater, based on the continuous potentiometric measurement of pH (Kimoto et al., 2001). This system, also based on the single addition technique, continuously mixes seawater and HCl solution in a particular proportion using valveless piston pumps and a mixing cell and measures the pH of the acidified solution using a combination glass electrode. Pontentiometric pH measurements are chosen here, because this method can be as precise as spectrophotometric pH measurements when used properly (Byrne et al., 1988), and the pH electrode is much cheaper than the spectrophotometer. The system automatically introduces two standard solutions of different concentrations at a certain time interval, using measurements on these to assign values to the sample and also to compensate for drift. This system was deployed on a small boat in a coral reef, and succeeded in the quasi-in situ monitoring of TA with high time resolution (30 s for 90% response) and good precision (±2 μmol kg⁻¹) Kimoto et al., 2001, Kayanne et al., 2002. This system, however, had input lines for only two standard solutions, which did not allow us to obtain information on the non-linearity of the measurement and made it difficult for us to assign analytical figures of merit.

Here, we report work in which we have improved the prototype flow-through analyzer developed by Kimoto et al. (2001) by providing three lines for input of standards, and made measurements for TA in the laboratory to evaluate the system performance parameters such as the response time, noise level, precision, and accuracy. These results provide us with useful information on the automated operation and, in turn, on the probability for in situ monitoring of TA with high resolution, precision, and accuracy in coastal areas in the near future.