Towards high and low pH the mCP absorption spectra are dominated by the deprotonated and the monoprotonated form of the pH dye and the spectrophotometric measurement approach reaches its application limits. The applicable pH range was investigated theoretically by IOW, performing a proper error analysis according to Gauss’ law of error propagation. This law was applied to uncertainties associated to the absorption measurement, which control the decreasing precision at high and low pH. An example of the outcome of this procedure is displayed in Fig. 1.
The same approach was also applied to the propagation of uncertainty associated to other parameters that have to be assessed (salinity, temperature, extinction coefficients). A clear finding with importance towards the instrument development is that the projected error of even a specific instrument cannot be represented by a single number, but shows a strong dependency on salinity, temperature, and the actual pH.
Figure 1: pH precision that can be obtained by spectrophotometric pH measurements with m-cresol purple over a salinity range 0-40 and a pH range 5-10. The dashed lines represent isoprecision lines. The optimum conditions for spectrophotometric pH measurements correspond to an absorption ratio R = 1, which is roughly at pH levels 0.3 units below pK (mCP).
Furthermore, the potential uncertainty of pH measurements performed at 25°C was assessed by a chemical speciation model for realistic conditions in the Gulf of Bothnia, the Gulf of Riga, and the Baltic Proper, in the pCO2 range 100-600 ìatm, reflecting the maximum range of the seasonal cycle of the carbonate system in surface waters in the Baltic Sea. These results show that where the sample pCO2 is at atmospheric or supersaturated levels, significant uncertainties will be encountered only in the low salinity waters of the Gulf of Bothnia (and also in the Gulf of Finland). Strongly undersaturated samples present more of a problem, most particularly in the summer period where the water temperatures are high.