Measurements made during the Ocean Abyssal Carbon Experiment (OACE, a joint project with investigators from Japan, Norway, and the USA, sponsored by NEDO in Japan) first revealed the problem with the under-estimation of total CO2 when determined by pH measurements alone in a flowing seawater environment immediately down-stream from a pool of liquid CO2.  When instruments (recording CTDs with integral pH sensors) were placed at various distances downstream from the benthic flume, negative pH spikes were recorded when the currents were favorable for the plume of CO2 enriched seawater to pass by the sensors. These low pH excursions were frequently accompanied by negative spikes in conductivity.  Normally, one would expect that if sufficient CO2 dissolved to depress the pH, a proportional increase in the conductivity would occur due to the presence of the additional H+ and HCO3- ions.  It was soon realized that the negative spikes in conductivity were due to various quantities of unreacted, dissolved CO2 "diluting" the seawater, lowering the salinity and hence the conductivity.  See the figure below for an example of this effect.