Graduation Year


Document Type




Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Marine Science

Major Professor

Robert H. Byrne, Ph.D.

Committee Member

Kristen N. Buck, Ph.D.

Committee Member

Richard A. Feely, Ph.D.

Committee Member

Pamela Hallock-Muller, Ph.D.

Committee Member

Lisa L. Robbins, Ph.D.


m-cresol purple, thymol blue, cresol red, indicator impurities, internal consistency


Nearshore and estuarine environments play a vital role in the cycling of carbon, but the effects of ocean acidification in estuarine waters have not been studied as extensively as in the open ocean. One reason for this is the limitation of pH measurement capabilities in low-salinity waters. Typically, pH in these environments has been measured using potentiometric methods that are subject to uncertainties on the order of 0.01. Spectrophotometric methods for measuring pHT offer precision and accuracy superior to those of potentiometric methods. However, previous characterizations for purified sulfonephthalein indicators, used for marine spectrophotometric measurements, are not applicable to estuarine salinities. Some estuarine datasets using unpurified indicators exist, but the presence of dye impurities affects the accuracy of these characterizations. Colorimetric impurities are known to interfere with absorbance measurements and can cause errors in pH on the order of 0.02.

In this work, a mathematical model has been developed to correct spectrophotometric pHT determined with unpurified m-Cresol Purple (mCP), the indicator used most widely for these measurements. The model accounts for absorbances of colorimetric impurities that interfere with absorbance by mCP. This corrective approach brings measurements made using unpurified mCP in synthetic solutions of 0.7 M NaCl into better agreement with those made using purified mCP: within ±0.004 pH units for all six indicators tested at pHT ≤ 8.0. The model is useful for both (a) research groups currently using unpurified mCP to measure pHT, and (b) retrospective correction of historic pHT datasets collected using unpurified mCP. The correction requires only that a small sample of the unpurified mCP is saved for a single-point test at high pHT (~12), and that historic absorbance measurements are archived for subsequent correction.

The principles of the corrective model were applied to an historic calibration of the mCP dissociation constant (KI) at 0 ≤ S ≤ 40 and T = 298.15 K using unpurified indicator. After correction of absorbances for dye impurities, recalculation of KI was performed, and the recalculated values were combined with mCP KI data for freshwater and seawater. The combined dataset was then refitted as a function of S and T. The resulting model is representative of mCP behavior across 0 ≤ S ≤ 40 and 278.15 ≤ T ≤ 308.15 K and produces p(KIe2) values that are within ±0.004 of p(KIe2) values from previously published purified mCP calibrations.

This refitting approach was also applied to pHT determinations made with Thymol Blue (TB) and Cresol Red (CR), two sulfonephthalein indicators that have been previously used in waters outside the indicating range of mCP. The models, which were of the same form as the estuarine p(KIe2) model for mCP, performed approximately as well as the mCP model: with the exception of one high-salinity, high-temperature TB datum, all residuals were within ±0.0043 of the previously published TB and CR calibrations.

Finally, an internal consistency analysis was performed using carbon chemistry data collected during two recent coastal ocean acidification research cruises. For pHT measurements performed during both cruises, purified mCP was used, and corresponding measurements of total alkalinity (TA) and dissolved inorganic carbon (DIC) were conducted. Both cruises included excursions into the Columbia River, where low salinities prevent usage of the marine p(KIe2) model for purified mCP. The Columbia River samples provided the opportunity to evaluate the internal consistency of pHT measurements made in low-salinity waters using the refitted estuarine p(KIe2) model. Although internal consistency agreement in the estuarine range is poor compared to marine measurements, pHT calculated using the new estuarine model compared well with pHT calculated using the previously published estuarine mCP model. The poor internal consistency in the estuarine range, even when making state-of-the-art pH measurements, points toward the need for a more robust characterization of the carbonic acid dissociation constants in the estuarine salinity range. This characterization should take into account the contributions of organic acids to total alkalinity in nearshore waters.