seacarb::carb() and shiny app results differ

[chrisdane]

Hi

With flag=15, ALK=2200 µmol kg-1 and DIC=2000 µmol kg-1 the shiny app yields

Sample	flag	S	T	Patm	P	pH	CO2	fCO2	pCO2	fCO2pot	pCO2pot	fCO2insitu	pCO2insitu	HCO3	CO3	DIC	ALK	OmegaAragonite	OmegaCalcite
NA	15.00	35	25	1.00	0	7.87	17.62	620.53	622.52	620.53	622.52	620.53	622.52	1836.73	145.66	2000.00	2200.00	2.31	3.51

When I do it directly in R the result is

library(seacarb)
seacarb::carb(flag=15, var1=2200/1e6, var2=2000/1e6, S=35, T=25) # flag=15 --> var1 = Talk in mol kg-1; var2 = DIC in mol kg-1

flag	S	T	Patm	P	pH	CO2	fCO2	pCO2	fCO2pot	pCO2pot	fCO2insitu	pCO2insitu	HCO3	CO3	DIC	ALK	OmegaAragonite	OmegaCalcite
15	35	25	1	0	7.870532	1.739097e-05	612.5331	614.4945	612.5331	614.4945	612.5331	614.4945	0.001835282	0.0001473273	0.002	0.0022	2.338158	3.547316

I am aware that the shiny app returns different units than the carb() function, e.g. for ALK and DIC.

However, pCO2_app = 622.52 while pCO2_carb = 614.4945. Could you please explain these differences?

Thanks for any help and kind regards, Chris

packageVersion("seacarb")
[1] ‘3.3.0’

[martens73]

The carb() function of the seacarb package uses the boron concentration of Uppstrom (1974) as default, while the newer reference of Lee et al. (2010) is more widely used in the community. You can test it by selecting Uppstrom (1974) from the dropdown menu in the Additional choices box in ScarFace. Alternatively, apply the Lee et al. (2010) boron concentration in the original carb() function:

carb(flag=15, var1=2200/1e6, var2=2000/1e6, S=35, T=25, b="l10")

You should then get identical results between seacarb and ScarFace.

[chrisdane]

Thanks a lot for your answer!

carb(flag=15, var1=2200/1e6, var2=2000/1e6, S=35, T=25, b="l10")$pCO2 equals 622.5467 which is quite similar as the ScarFace value 622.52 but still not equal?

[martens73]

I can not reproduce this issue. I always get exactly the same numbers, no matter whether I use seacarb or ScarFace. In fact, there is even no way that they can be different as ScarFace is using seacarb, and there are no conversions involved. Please check and compare carefully the parameters set in each program.

[chrisdane]

Screenshot from ScarFace: --> pCO2_scarface = 622.52

library(seacarb)
round(seacarb::carb(flag=15, var1=2200/1e6, var2=2000/1e6, S=35, T=25, b="l10")$pCO2, 2)
[1] 622.55

--> pCO2_seacarb = 622.55

You cant reproduce this?

This is the flag=15 case in ScarFace:

if (!input$pHscale == "NBS") {
    carb(flag=as.numeric(input$pair), var1=input$first/1000000, var2=input$second/1000000, 
         S=input$salc, T=input$tempc, Patm=1, P=input$presc, 
         Pt=input$pho/1000000, Sit=input$sil/1000000, k1k2=input$k1k2, 
         kf="x", ks="d", pHscale=input$pHscale, b=input$b, gas=input$gasm,
         warn="y", eos="eos80", long=1.e20, lat=1.e20)
} else {
    carb(flag=as.numeric(input$pair), var1=input$first/1000000, var2=input$second/1000000, 
         S=input$salc, T=input$tempc, Patm=1, P=input$presc, 
         Pt=input$pho/1000000, Sit=input$sil/1000000, k1k2=input$k1k2, 
         kf="x", ks="d", pHscale="SWS", b=input$b, gas=input$gasm,
         warn="y", eos="eos80", long=1.e20, lat=1.e20)
}

Since by default pHscale = "T", the first case applies and using the same arguments yields

round(seacarb::carb(flag=15, var1=2200/1e6, var2=2000/1e6, S=35, T=25, 
                    Patm=1, P=0, Pt=0/1e6, Sit=0/1e6, k1k2="x", kf="x", ks="d", 
                    pHscale="T", b="l10", gas="potential", warn="y", 
                    eos="eos80", long=1.e20, lat=1.e20)$pCO2, 2)
[1] 622.55

, i.e. pCO2_seacarb = pCO2_myscarface = 622.55 != pCO2_scarface = 622.52.

What is the seacarb package version used by ScarFace?

[chrisdane]

Yes its the package version:

packageVersion("seacarb")
[1] ‘3.2.16’
round(seacarb::carb(flag=15, var1=2200/1e6, var2=2000/1e6, S=35, T=25, b="l10")$pCO2, 2)
[1] 622.52

but

packageVersion("seacarb")
[1] ‘3.3.0’
round(seacarb::carb(flag=15, var1=2200/1e6, var2=2000/1e6, S=35, T=25, b="l10")$pCO2, 2)
[1] 622.55

Thanks and cheers, Chris