seacarb::carb() and shiny app results differ

[martens73]

seacarb::carb() and shiny app results differ

[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