Feature ArticleAutonomous pH and pCO2 Measurements in Marine Environments
By Reggie Spaulding
Sunburst Sensors LLC
Department of Chemistry and Biochemistry
University of Montana
The carbon dioxide (CO2) released to the atmosphere by humans is expected to increase global temperatures by 1.5° to 6.0° C in the next 100 years. The oceans have absorbed 25 to 30 percent of this anthropogenic CO2, reducing its effects on climate, but also increasing the partial pressure of CO2 (pCO2) in the oceans.
CO2 reacts with water to form carbonic acid, decreasing the pH of the oceans, a process known as ocean acidification. Surface ocean pH is projected to decrease by 0.4 pH units relative to preindustrial levels by the year 2100, representing a 150 percent increase in the concentration of hydrogen ions. This acidification shifts the carbonate equilibrium, leaving less carbonate available and reducing carbonate saturation states for calcifying organisms such as corals and shellfish, but also increasing the CO2 available to photosynthetic organisms. The overall effect on marine ecosystems is therefore complex and not well understood.
Tracking the impacts of increased CO2 in the oceans requires widespread and sustained measurements of the inorganic carbon system, which can be described by the equations for total dissolved inorganic carbon (DIC), total alkalinity (Aτ) and the equilibrium expressions for dissolution of CO2. A combination of the equations results in two independent equations and four unknowns. Therefore, measurement of two inorganic carbon parameters can be used to quantify all of the others. Calcium carbonate saturation states (Ωaragonite and Ωcalcite) can also be calculated from this information. With saturation states greater than one, calcium carbonate is supersaturated and calcifiers can more readily form skeletons, whereas saturation states less than one indicate undersaturation and dissolution of calcium carbonate structures such as shells and corals.
Shipboard systems are routinely used to measure pCO2, pH, DIC and Aτ. In most studies, two or more of these four parameters are measured, making it possible to calculate the others and compare measured and calculated values. Shipboard studies only capture brief periods of variability, however, and miss remote and stormy regions that are important to the ocean carbon cycle, such as the Southern Ocean.
To improve data coverage, autonomous sensors for measurement of inorganic parameters would ideally be available for deployment on moorings, drifters and other unmanned platforms. Developing the autonomous technology for these complex chemical measurements has been challenging, but there has been some success, and steady progress continues to be made. For example, autonomous pCO2 instruments have been commercially available from a number of sources for more than a decade, and autonomous pH instruments have more recently become available.
The Submersible Autonomous Moored Instruments (SAMIs) for pCO2 and pH are two examples of autonomous sensors designed specifically for marine research that have been commercially available for a number of years. These instruments, the SAMI-CO2 and SAMI-pH, were recently deployed together for almost two months to collect data on the Oregon shelf. This deployment, representing one of the first times pH and pCO2 data have been collected simultaneously over an extended time period, demonstrates the application of SAMI data, along with salinity data, to define the inorganic carbon system.
Additionally, new models of these instruments, the SAMI2-CO2 and SAMI2-pH instruments, were recently introduced. The SAMI2s are smaller, consume less power and are more user-friendly than the original SAMIs. However, although many of these instruments have been deployed within the past year, results from simultaneous collection of pH and CO2 data with the SAMI2 instruments are not yet available.
pCO2. In the SAMI-CO2, a pH-sensitive indicator solution (bromothymol blue) is pumped into a gas-permeable membrane that is submersed in seawater. CO2 diffuses across the membrane until the pCO2 in the indicator solution is in equilibrium with the pCO2 in the surrounding seawater. Carbonic acid, formed from dissolution of CO2, changes the pH of the indicator solution, which in turn changes the relative concentrations of [I²-] and [HI-], the unprotonated and protonated forms of the indicator, respectively, as shown in Equation 1, where Ka' is the acid dissociation constant of the indicator and [H+] is the hydrogen ion concentration. The peak absorption wavelengths of [I²-] and [HI-] are 620 nanometers and 434 nanometers, respectively. The pCO2 in the seawater can be related to the ratio of the absorbances of the indicator at these wavelengths.
The upper graph shows a comparison of pCO2 measured in a test tank by the SAMI2-CO2 and by an infrared analyzer. The lower graph shows water temperature.
To calibrate the SAMI-CO2, the instrument is submersed in water with varying pCO2 and the indicator absorbances are measured. A calibration equation is determined by the relationship of the absorbance ratio and the pCO2 measured by an infrared analyzer. Long-term drift-free performance is obtained by renewing the indicator for each measurement and by periodically measuring optical transmittance with a blank solution.
pH. The SAMI-pH operates on the same theory as the SAMI-CO2, using a pH-sensitive indicator solution. In this case, the indicator is metacresol purple, with peak absorption wavelengths of [I²-] and [HI-] at 578 nanometers and 434 nanometers, respectively. Seawater is mixed with indicator solution and the absorbances of the indicator-seawater mixture are measured at these wavelengths. The pH of the mixture is calculated using Equation 2, where the ratio of the absorbances at 578 nanometers and 434 nanometers and the molar absorption of the indicator at these wavelengths is related to the ratio of [I²-] and [HI-].
The SAMI-CO2 and SAMI-pH are both colorimetric, reagent-based sensors and have very similar designs. Both use a solenoid pump and a solenoid valve for fluid control. The solutions are pumped to a fiber optic flow cell where absorbance measurements are made. Photodiodes detect the light signals, and a thermistor measures temperature in both designs. Time-stamped data is logged to internal memory for each measurement and downloaded when the instrument is retrieved. Some users have set up the instrument to telemeter data in near real time.
In the SAMI-CO2, equilibration between seawater and indicator solution takes place through a small-diameter, tubular, gas-permeable silicone membrane, and indicator is renewed to the membrane at the beginning of each measurement cycle. A nanopure water absorbance blank is recorded approximately every 3.5 days. In the SAMI-pH, seawater is flushed through the optical cell and light transmitted through the pure seawater is used as the absorbance blank at the beginning of each measurement cycle. Indicator solution is then pumped into the seawater, followed by additional pumping of seawater. A static mixing cell between the pump and the optical cell mixes seawater and indicator solution as they are pumped toward the optical cell, absorbances of the mixture are measured, and the pH of the mixture is calculated using Equation 2.
SAMI-CO2 and pH Deployment
The recent availability of autonomous pH instruments makes it possible to gather simultaneous pH and pCO2 data on moorings. SAMI-CO2 and SAMI-pH instruments were deployed on the Oregon Coastal Ocean Observing System's NH10 mooring from March 24 through May 31, 2009, and took measurements at 30-minute intervals. Two weeks of this data, March 27 through April 10, are presented here. Temperature and salinity were also measured on the mooring. The dataset is being used to estimate diurnal and seasonal pH and pCO2 variation as well as calcium carbonate saturation states in this region.
Initial inspection of the data reveals that pH and pCO2 are inversely correlated, as expected, because higher pCO2 indicates higher levels of carbonic acid and therefore lower pH. Large swings in pH and pCO2 were found due to daily cycles of photosynthesis and respiration in addition to water movement such as tides and upwelling. Aτ, which is strongly correlated with salinity in certain regions of the ocean, was calculated using a known relationship between salinity and Aτ for California coastal waters.
The salinity-derived Aτ from this dataset was used with SAMI-pH data to estimate pCO2, DIC, Ωaragonite and Ωcalcite). Additionally, the pH and pCO2 data were used to directly estimate DIC, Ωaragonite and Ωcalcite). The data show that Ωaragonite and Ωcalcite) calculated using the two different methods are in very good agreement (the average difference was 0.012±0.002), and that pCO2 measured by the SAMI-CO2 and calculated by SAMI-pH and AT are in good agreement (the average difference was 3±5 microatmospheres).
These data indicate good in-situ accuracy of both instruments and consistency with carbonate system models, suggesting a beneficial growing environment for organisms such as shellfish, which form calcite or calcite-aragonite shells. However, DIC calculated using SAMI-pH with SAMI-CO2 data is noisy compared to DIC calculated from salinity-derived AT with SAMI-pH data (the average difference was 18±2 micromoles per kilogram of seawater). Researchers have found similar results with shipboard CO2 and pH data. However, more research is needed to determine to what extent in-situ pH and pCO2 can be used to estimate DIC.
These data show that calcite and aragonite are both well above saturation off the Oregon coast, suggesting a beneficial growing environment for calcite and calcite/aragonite-forming organisms such as shellfish during this sampling period. However, continued measurements through other seasons and years will determine whether the saturation states differ by season or are changing with increasing oceanic CO2.
While the original SAMIs have made significant progress in autonomous pH and CO2 sensing, five years of research and development has culminated in the SAMI2, which features the same data-collection capabilities in a smaller package that consumes less power and is easier to use.
One difference in the design is that the old SAMI used a single tungsten light source, fiber optic splitter and optical bandpass filters, while the new SAMI2 uses high-output light-emitting diodes for a better signal-to-noise ratio and lower power consumption.
The accuracy and precision of the SAMI2 (pH and pCO2) are similar to the SAMI. The average accuracy of 15 SAMI2-pH instruments, using Tris buffer solutions with pH certified by Scripps Institution of Oceanography, was 0.0003±0.0016 pH.
SAMI2-CO2 accuracy was determined by submersing the instrument in a test tank and comparing CO2 measurements to those from an infrared analyzer. A gas-liquid membrane contactor was used to equilibrate carrier gas with the tank water prior to infrared analysis.
Throughout the 50-day test, the temperature and pCO2 ranges were approximately 10° to 27° C and approximately 200 to 425 microatmospheres, respectively. Large temperature changes were controlled by a water bath and small temperature oscillations were caused by changing room temperature. pCO2 measured by the infrared analyzer and SAMI2-CO2 closely tracked each other, with an average difference between the two methods of -1.7±3.6 microatmospheres (averaged from 3,009 measurements).
There were disparities at times due to response-time differences between the two systems, but there was no long-term drift in the data and no recalibration or offsets were applied to this dataset.
The SAMI-CO2 and SAMI-pH instruments demonstrate the power of using pH and Aτ data to quantify the entire carbonate system. In areas of the ocean where salinity and Aτ relationships are known, only pH and salinity measurements are needed to fully characterize the inorganic carbon system. Similar results were seen with pCO2 and salinity data. Additionally, the combination of pH and pCO2 data can accurately quantify carbonate saturation states and is useful for assessing data quality.
The ability of the SAMI instruments to operate autonomously for long time periods allows oceanographers and marine biologists to study the effects of increasing CO2 on marine organisms throughout the year. The introduction of the new SAMI2 instruments, which have been shown to be equally capable while using less reagent, space and power, will only increase the benefits that these systems offer.
The authors thank Burke Hales of Oregon State University for mooring logistics and deployment of the SAMIs, Cory Beatty and Jenny Newton for testing and validation of the SAMI2 instruments, and Jim Beck for help in preparing this article. They also thank the National Oceanographic Partnership Program, National Science Foundation-Ocean Sciences and NOAA-Global Carbon Cycle programs for funding.
Reggie Spaulding obtained a Ph.D. in environmental chemistry from the University of California at Davis in 2002. She was a post-doctoral scholar at the University of Montana before joining Sunburst Sensors in 2004, and she has worked on the development of autonomous instruments for measurement of total alkalinity, dissolved inorganic carbon, the partial pressure of carbon dioxide and pH.
Mike DeGrandpre obtained a Ph.D. in analytical chemistry in 1990 from the University of Washington. He was a post-doctoral scholar and research associate at Woods Hole Oceanographic Institution from 1990 to 1995 before becoming a chemistry professor at the University of Montana in 1996.
Katherine Harris received a B.Sc. in chemistry from the College of William and Mary in 2008. She is currently a Ph.D. student at the University of Montana. Her research is focused on oceanic inorganic carbon system dynamics in coastal upwelling zones.