From Wikipedia: "The density of surface seawater ranges from about 1,020 to 1,029 kg·m−3, depending on the temperature and salinity. Deep in the ocean, under high pressure, seawater can reach a density of 1,050 kg·m−3 or higher." So, if the ppm of CO2 is directly proportional in some way to water density, even seawater would only introduce a 5% error, which would be a spectacular accuracy.
Over the range of 10C to 30C the density of water varies less than 1%.
I'm inclined to believe that we can ignore water density variations in measuring ppm of CO2 in water.
Plantbrain has long been advocating measuring CO2 with the plants. In other words the absolute value of the concentration of CO2 isn't important, as long as we have determined that we have as much as the plants need at the light intensity we have. Where a sensor like this would be valuable is for monitoring changes in concentration, and even a moderately accurate sensor would work for that as long as it is sensitive enough.