Please change the title of this work to 'Error-space diagram for the marine CO2 system" (in the README and at the top of the page of this github archive, before the listing of files). The current version of those 2 titles are too much like the title of my manuscript that I gave you. That manuscript introduces the error-space diagram but that is not its only concern.

Cathy,
It should be indicated, perhaps in the title of the plot (or in a figure legend) that the error that is shown in the relative percent error. That is, it is 100 times the ratio of the absolute error divided by the value of the computed variable.

• provide a few contour level example R commands
• highlight a plot and explain the story it tells

When the user specifies the reference point coordinates, could you allow those boxes to take vectors e.g., c(1, 2,3) and c(2,4,6)? Currently, only 1 reference point is shown except when pH is s member (2 reference points). I just tried putting vectors into the 2 boxes that specify the Reference point coordinates, but it doesn't work. That is because in the original code I did not have this functionality in mind. There will have to be a bit of recoding to add this feature, but it should be straightforward.

For this, the two coordinate boxes for the reference point should have the same number of members, -unlike the current cases when pH is a member of the input pair.

We talked about doing this earlier. Perhaps you could add another check box "Edit input errors in pK's".

Beware though that the epK used in derivnum.R and errors.R will change so that the last error (the relative error in Bt) is taken out, reducing epK from 8 to 7 members; instead eBt will be added as a separate argument to the errors routine, with a default of 0.01.

Thus you should probably wait on this modification until the new versions of errors.R and derivnum.R are available, proabably within 30 days.

Hi Cathy,

Users would find this suggested feature useful, especially advanced ones who want to alter these vaules. The checkbox should have the label 'Change default uncertainties for constants (epK0, epK1, epK2, epKb, epKw, epKspa, epKspc, eBt)'.

If you don't have time for this, i could perhaps give it a try, make a test, and issue a Pull Request.

Thanks!

James