mainpage.txt

/*! \mainpage Routines to model the ocean carbonate system (mocsy v2.0)

Details of individual routines are available under the "Files" tab
above, while the overall description, implementation and use of
<em>mocsy</em> is described below.

- \subpage synopsis 
- \subpage authors
- \subpage bkg 
- \subpage fort
- \subpage pyth
- \subpage refs
*/

//-----------------------------------------------------------

/*! \page synopsis Synopsis

<em>mocsy</em> is a Fortran 95 package designed to compute all ocean
carbonate system variables from DIC and total Alk, particularly from
models.  It updates previous OCMIP code, avoids 3 common model
approximations, and offers the best-practice constants as well as more
recent options. Its results agree with those from CO2SYS-MATLAB within 0.005%.

*/

//-----------------------------------------------------------

/*! \page authors Authors

<em>James Orr</em>, LSCE/IPSL, CEA-CNRS-UVSQ, Gif-sur-Yvette, France

<em>Jean-Marie Epitalon</em>, Geoscientific Programming Services, Toulouse, France

*/

//-----------------------------------------------------------

/*! \page bkg Overview

The <em>mocsy</em> package is designed to accurately and efficiently
compute all carbonate system variables given input for dissolved
inorganic carbon (DIC), total alkalinity (Alk), temperature, and
salinity as well as concentrations of total dissolved inorganic
phosphorus and silicon concentrations. By default, the thermodynamic
equilibrium constants used in <em>mocsy</em> are those recommended by
the Best-Practices Guide (Dickson et al., 2007). <em>mocsy</em> also
includes options to use more recent formulations for the constants
\f$K_1\f$ and \f$K_2\f$ from (Millero, 2010) and for total boron
concentrations as a function of salinity (Lee et al., 2010), although
these are not recommended at this time (Orr and Epitalon, 2015).

<em>mocsy</em> is designed for users with programming experience,
preferably in either Fortran or python; others may well prefer the
facility of packages such as ODV or the Excel version of CO2sys.
However, <em>mocsy</em>'s greater computational speed (Orr et al.,
2014) makes it well suited for working with large data sets and ocean
models. Furthermore, <em>mocsy</em> replaces 3 typical model
approximations with exact formulations to improve model accuracy.
Whereas the equations describing carbonate system thermodynamics
require input of pressure, in situ temperature, and concentrations in
mol/kg, modelers have used more accessible quantities: depth,
potential temperature, and concentrations in mol/m3 converted to mol/kg
with a constant density.  <em>mocsy</em> avoids these approximations
by calculating the 3 exact quantities from other input
variables. These approximations lead to significant errors in computed
variables in the deep ocean (Orr and Epitalon, 2014).

With none of the approximations, input from 3-D gridded data products,
namely GLODAP's DIC and Alk (Key et al., 2004) and World Ocean Atlas
(2009) fields for T, S, and nutrients (Antonov et al., 2010;
Loocarnini et al., 2010; Garcia et al., 2010), output from
<em>mocsy</em> agrees with that from CO2SYS-MATLAB (van Heuven et al.,
2011), within 0.005% for the three inorganic carbon
species. Corresponding pH differences are less than 0.00002, while
differences between packages for \f$\Omega_C\f$ and \f$\Omega_A\f$ are
offset by constant 0.018%. These small differences and their causes
are discussed in Orr and Epitalon (2014).

*/

//-----------------------------------------------------------

/*! \page fort The <em>mocsy</em> Fortran code

The fortran code for <em>mocsy</em> is available as a series of
Fortran 95 routines: vars.f90, constants.f90, depth2press.f90,
f2pCO2.f90, gasx.f90, p2fCO2.f90, p80.f90, phsolvers.f90, rho.f90,
rhoinsitu.f90 , sw_adtg.f90, sw_ptmp.f90, sw_temp.f90, test_mocsy.f90,
tis.f90, tpot.f90, and varsolver.f90. Most users will simply call
vars.f90, which calls other routines.  Extensive documentation and the
code listing for each of these routines is accessible by clicking on
their links above.

\section download Download it

If <em>git</em> software is installed, the most recent version of the
mocsy package can be downloaded using typical git commands. Go to a
directory under which you want install a new subdirectory <em>mocsy</em>
containing the code.  Then type

\verbatim
$ git clone git://github.com/jamesorr/mocsy.git
\endverbatim

Otherwise, just go to the main web page on github
https://github.com/jamesorr/mocsy and click on the link "Download
ZIP".

\section compile Compile it

Once the download is complete, compile the code in the normal fashion
by issuing the following command:

\verbatim
$ make test_mocsy
\endverbatim

If successful, you will have just produced an executable
<em>test_mocsy</em>. Otherwise, if the gfortran (alias f95) compiler
is not installed on your system, install it or try another one that is
already installed.  So far, <em>mocsy</em> has been tested and run
successfully with two compilers: gfortran and ifort.  To change the
compiler name, edit the <em>makefile</em> and change "f95" to your
Fortran 95 compiler command.  Then issue the following 2 commands:

\verbatim
$ make clean
$ make test_mocsy
\endverbatim

\section run Run test program
To run the test program, simply execute it as 

\verbatim
$ ./test_mocsy
\endverbatim

\section output Output
If successful, 2 tables will be printed to your screen, as below:

\verbatim
Table 1: Typical MODEL output
---------------------------------------------------------------------------------------------------------
                                                                               in situ         in situ
  pH     pCO2   fCO2     CO2*       HCO3-       CO32-      OmegaA OmegaC  R    Density Press  Temperature
(total) (uatm) (uatm)  (mol/m3)    (mol/m3)    (mol/m3)                        (kg/m3) (db)      (C)
---------------------------------------------------------------------------------------------------------
 8.0592  374.9  373.3  0.2234E-01  0.2086E+01  0.1061E+00   1.55   2.47  14.45 1027.97    0.0   2.000
 8.0169  652.3  423.3  0.2201E-01  0.2089E+01  0.1038E+00   1.24   1.95  14.51 1032.68 1012.2   2.062
 7.9741 1139.4  481.4  0.2172E-01  0.2091E+01  0.1015E+00   1.00   1.55  14.56 1037.30 2029.1   2.142
 7.9308 1997.5  549.1  0.2146E-01  0.2094E+01  0.9926E-01   0.81   1.24  14.61 1041.83 3050.6   2.240
 7.8872 3513.6  627.9  0.2124E-01  0.2096E+01  0.9712E-01   0.66   0.99  14.66 1046.28 4076.9   2.354
 7.8431 6199.0  720.0  0.2105E-01  0.2098E+01  0.9505E-01   0.54   0.80  14.70 1050.65 5108.0   2.483
---------------------------------------------------------------------------------------------------------
\endverbatim


\verbatim
Table 2: Typical DATA output
---------------------------------------------------------------------------------------------------------
                                                                               in situ         in situ
  pH     pCO2   fCO2     CO2*       HCO3-       CO32-      OmegaA OmegaC  R    Density Press  Temperature
(total) (uatm) (uatm)  (mol/kg)    (mol/kg)     (mol/kg)                       (kg/m3) (db)      (C)
---------------------------------------------------------------------------------------------------------
 8.0546  382.8  381.1  0.2219E-04  0.2028E-02  0.1042E-03   1.57   2.49  14.38 1027.97    0.0   2.000
 8.0146  661.8  431.5  0.2182E-04  0.2030E-02  0.1021E-03   1.27   1.99  14.45 1032.63 1000.0   2.000
 7.9746 1145.7  489.3  0.2148E-04  0.2033E-02  0.9998E-04   1.03   1.59  14.51 1037.19 2000.0   2.000
 7.9345 1986.2  555.5  0.2117E-04  0.2035E-02  0.9797E-04   0.84   1.28  14.57 1041.65 3000.0   2.000
 7.8944 3448.2  631.4  0.2090E-04  0.2037E-02  0.9602E-04   0.69   1.04  14.62 1046.02 4000.0   2.000
 7.8544 5994.4  718.7  0.2066E-04  0.2039E-02  0.9414E-04   0.56   0.84  14.68 1050.29 5000.0   2.000
---------------------------------------------------------------------------------------------------------
*Table 2 differs slightly from Table 1 because input DIC, Alk, pressure, and temperature differ.
*Concentration units differ between Tables 1 and 2 because input options differ
\endverbatim

Comparison of the 2 tables reveal in part differences due to avoiding
the 3 approximations that are commonly used in model calculations of
the ocean carbonate system. For details on the specific options, see
the program <em>test_mocsy.f90</em> as well as the documentation of
routines under the "Files" tab.

\section args Calling the main routine

In a standard Fortran 95 program, <em>mocsy</em>'s routine that
computes all carbonate system variables by calling the <em>vars</em>
routine

\verbatim
call vars(ph, pco2, fco2, co2, hco3, co3, OmegaA, OmegaC, BetaD, rhoSW, p, tempis,                  &  ! OUTPUT
          temp, sal, alk, dic, sil, phos, Patm, depth, lat, N,                                      &  ! INPUT
          optCON='mol/m3', optT='Tpot', optP='m', optB='u74', optK1K2='l', optKf='dg', optGAS='Ppot')  ! INPUT OPTIONS
\endverbatim

where arguments on the 1st line are output variables, those on the 2nd
line are input variables, and those on the 3rd line are input
options. Extensive documentation on the arguments, units, and options, 
is accessible by clicking on the "Data Types List" tab and then on
"mvars" (the module that contains the <em>vars</em> routine). The
options above would be the common choice for typical ocean models (see
also <em>test_mocsy.f90</em> for typical options used with
observations.

Other <em>mocsy</em> routines are available, for example 
to get the entire set of equilibrium constants (constants.f90), 
to convert between potential and in-situ temperature (tpot.f90 and tis.f90) , 
to convert between depth and pressure (depth2press.f90), 
to compute in situ density (rhoinsitu.f90),
to convert between fCO2 and pCO2 (f2pCO2.f90 and p2fCO2.f90), and
to compute atmospheric pCO2 and xCO2 as well as air-sea CO2 fluxes (gasx.f90)

The <em>mocsy</em> Fortran 95 routines can also be called from
<em>python</em> as described in the next section.

*/

//-----------------------------------------------------------

/*! \page pyth Using <em>mocsy</em> in python
This page is for users that want to access the vast capabilities of python, e.g. reading & writing netCDF files easily.

The mocsy package, written in Fortran, is also designed to be called
within a python script. Using mocsy from within python offers a simple
and more flexible interface to handle input and output files and to
integrate with other packages.  For example, the python interface
makes it easy to read and write netCDF files as well as to interface
with R using <em>rpy2</em>.  

\section shareobj Make it useable

The first step, before importing mocsy in python, is to make a shared
object module <em>mocsy.so</em> using the utility <em>f2py</em>.  This
fortran-to-python lilbrary converter is part of the python's widely
used NumPy library. Secondly, you will need a Fortran 95 compiler
(e.g., GNU's gfortran).  With both <em>NumPy</em> and gfortran
installed, just go to the directory where you downloaded the mocsy
routines, and type

\verbatim 
make mocsy.so
\endverbatim 

If successful, output from that command will end in "Removing build
directory ..." and the new file <em>mocsy.so</em> will have been
produced.  If unsuccesful, test to see that you have gfortran
installed by typing

\verbatim
f2py -c --help-fcompiler
\endverbatim 

If <em>--fcompiler=gnu95</em> does not show up just under the line
"Fortram compilers found:", then try again to reinstall it.
Otherwise, if you already have another Fortran 95 compiler installed,
simply edit the <em>makefile</em> and change the name of the
FC and F90 variables, accordingly. In addition, change the
corresponding part of the last line <em>--fcompiler=gnu95</em> to the
name of your compiler, as listed under the "Fortran compilers found:"
line from the above command.

\section import Import it

After a successful build, the shared-object file <em>mocsy.so</em>
will be found in the local directory. It is then imported into a
python script like any other python library with

\verbatim
import mocsy
\endverbatim

\section pydoc Documentation in python

Once <em>mocsy</em> is imported, you can see the basic routines that
are available, with the standard python library documentation command:

\verbatim
print mocsy.__doc__
\endverbatim

For details about the arguments of each routine, just add the name of
the "module", e.g.,

\verbatim
print mocsy.mvars.__doc__
\endverbatim

One could in addition add the name of the "routine", e.g.,

\verbatim
print mocsy.mvars.vars__doc__
\endverbatim

but that is usually unnecessary. Typically the documentation is
identical because there is but 1 routine per module, except for the
<em>gasx</em> routine.  More extensive documentation (on arguments,
units, options, etc.) is available in this manual by clicking on the
"Data File Types" tab and then the name of the module. The routines in python
have the same names as in Fortran. The module names are the routine
names preceded by a "m" (except for the gasx module, which contains
several routines).

\section use Use it

As an example, here is the call in python of the main routine <em>vars</em> 

\verbatim
    pH, pco2, fco2, co2, hco3, co3, OmegaA, OmegaC, BetaD, DENis, p, Tis = (
    mocsy.mvars (t, s, alk, dic, sio2, po4, Patm, depth, lat, optcon='mol/kg', optt='Tinsitu', optp='m', optb="u74", optk1k2='l', optkf="dg", optGAS="Ppot")
    )
\endverbatim

Input variables and options are on the right side of the equal sign,
whereas output variables are on the left. All input and output
arguments must be 1-D vectors (not multidimensional arrays); the
conversion, both ways, is illustrated in the script
<em>mocsy_GLODAP.py</em> as are other finer points for using
<em>mocsy</em> in python. For instance, that script also uses the
python library <em>cdms</em> to read input files (for DIC, ALk, T, S,
\f$\hbox{PO}_4^{-3}\f$, and \f$\hbox{SiO}_2\f$) and write the output
files, all in netCDF format. Moreover, to save disk space for the
user, input files are read remotely, using OpenDAP. This script is
executed with the following command:

\verbatim
python mocsy_GLODAP.py
\endverbatim

Other subroutines needed by <em>mocsy.mvars</em> are also available to
the user.  For example, to get the equilibrium constants that were
used in the above call to <em>mocsy.mvars</em>, the command is

\verbatim
    Kh, K1, K2, Kb, Kw, Ks, Kf, Kspc, Kspa, K1p, K2p, K3p, Ksi, St, Ft, Bt = mocsy.mconstants (t, s, Patm, depth, lat, optt='Tinsitu', optp='m', optb="u74", optk1k2='l', optkf="dg", optGAS="Ppot")
\endverbatim

Other routines are available to compute pressure, in-situ density,
potential temperature, and in-situ temperature as well as to convert
between pCO2, fCO2 and xCO2.

\verbatim
# Compute pressure (db) from depth (m) and latitude (degrees north) following Saunders (1981)
pfd = mocsy.mdepth2press(depth,lat)   

# Compute in situ density (kg/m3) from salinity (psu), in-situ temp (C), and hydrostatic pressure (db)
rhois = mocsy.mrhoinsitu(s, t, pfd)  

# Compute potential temperature (C) from salinity (psu), in situ temperature (C), and hydrostatic pressure (db)
Tpot = mocsy.mtpot(s, Tis, pfd, 0)     

# Compute in situ temperature (C) from salinity (psu), potential temperature (C), and hydrostatic pressure (db)
Tis = mocsy.mtis(s, Tpot, pfd, 0)

# Compute oceanic pCO2 (uatm) from arrays of fCO2 (uatm), in situ temp (C), atm pressure (atm), & hydrostatic pressure (db)
pCO2 = mocsy.mf2pCO2(fCO2, temp, Patm, p)

# Compute oceanic fCO2 (uatm) from arrays of pCO2 (uatm), in situ temp (C), atm pressure (atm), & hydrostatic pressure (db)
fCO2 = mocsy.mp2fCO2(pCO2, temp, Patm, p)

# Compute atmospheric xCO2 (ppm) from arrays of atmospheric pCO2 (uatm), surface ocean T (C), S (psu), & atm pressure (atm)
xCO2atm = mocsy.mgasx.pCO2atm2xCO2(pCO2atm, temp, salt, Patm)

# Compute atmospheric pCO2 (uatm) from arrays of atmospheric xCO2 (ppm), surface ocean T (C), S (psu), & atm pressure (atm)
pCO2atm = mocsy.mgasx.x2pCO2atm(xCO2, temp, salt, Patm)

# Compute vapor pressure of seawater (atm) following procedure from Weiss & Price (1980)
vpsw = mocsy.mgasx.vapress(temp, salt)

# Computes air-sea CO2 flux & surface-ocean carbonate system vars (see mocsy doc & OCMIP2 Abiotic HOWTO)
co2flux,co2ex,dpco2,ph,pco2,fco2,co2,hco3,co3,OmegaA,OmegaC,BetaD,rhoSW,p,tempis = \
mocsy.mgasx.flxco2(temp, sal, alk, dic, sil, phos, kw660, xco2, Patm, dz1,         \
optCON, optT, optP, optB, optK1K2, optKf, optGAS)

\endverbatim

\section examples Examples

1) Simple scaler variables:

For DATA input:
DIC and ALk in mol/kg, in situ temperature, pressure.

\verbatim
$ python
import mocsy
pH,pco2,fco2,co2,hco3,co3,OmegaA,OmegaC,BetaD,DENis,p,Tis = \
mocsy.mvars (temp=18, sal=35, alk=2300.e-6, dic=2000.e-6, sil=0, phos=0, patm=1., depth=100, lat=0, optcon='mol/kg', optt='Tinsitu', optp='db', optb="u74", optk1k2='l', optkf="dg", optgas='Pinsitu')
\endverbatim

Compute the CONSTANTS with the same scalar input for S, T, and P (as above) but
change to the newer options published since the best-practices guide:
Lee et al. (2010) for total boron and Millero (2010) for \f$K_1\f$ and
\f$K_2\f$:

\verbatim
Kh,K1,K2,Kb,Kw,Ks,Kf,Kspc,Kspa,K1p,K2p,K3p,Ksi,St,Ft,Bt = \
mocsy.mconstants (temp=18.,sal=35., patm=1., depth=0.,lat=0., optt='Tinsitu', optp='m', optb="l10", optk1k2='m10', optkf="dg", optgas='Ppot')
\endverbatim
 
For MODEL input:
DIC and Alk in <em>mol/m3</em>, potential temperature, and depth. 
Also do NOT specify last 4 options, thus using defaults (optb='l10', optk1k2='l', optkf='pf', optgas='Pinsitu').

\verbatim
pH,pco2,fco2,co2,hco3,co3,OmegaA,OmegaC,BetaD,DENis,p,Tis = \
mocsy.mvars (temp=18, sal=35, alk=2300*1028e-6, dic=2000*1028e-6, sil=0, phos=0, patm=1, depth=100, lat=0, optcon='mol/m3', optt='Tpot', optp='m')
\endverbatim

2) Simple arrays (numpy):

\verbatim
$ python
import numpy as np
import mocsy

#Make dummy array with 11 members
one = np.ones(11,dtype='float32')

sal   = 35*one
temp  = 2*one
depth = np.arange(0,11000,1000,dtype='float32') # units in 'm'

# Compute in situ density (3 steps using 2 mocsy routines):
# a) Specify latitude = 60°S for depth to pressure conversion formula
lat = -60*one
# b) Compute pressure (db) from depth (m) and latitude following Saunders (1981)
pfd = mocsy.mdepth2press(depth,lat)  # units in 'db' 
# c) Compute in situ density (kg/m3) from salinity (psu), in-situ temp (C), and pressure (db)
rhois = mocsy.mrhoinsitu(sal, temp, pfd)  

# Specify surface concentrations typical of 60°S (from GLODAP and WOA2009)
alk   = 2295*one # umol/kg
dic   = 2154*one # umol/kg
sio2  = 50.*one  # umol/L
po4   = 1.8*one  # umol/L

# Convert nutrient units(mol/L) to model units (mol/m3)
sio2 = sio2 * 1.0e-3   
po4  = po4  * 1.0e-3 
# Convert Alk and DIC units (umol/kg) to model units (mol/m3)
dic = dic * rhois * 1e-6
alk = alk * rhois * 1e-6

# Compute carbonate system variables, using 'model' options
pH,pco2,fco2,co2,hco3,co3,OmegaA,OmegaC,BetaD,DENis,p,Tis = \
mocsy.mvars (temp=temp, sal=sal, alk=alk, dic=dic, sil=sio2, phos=po4, patm=1, depth=depth, lat=lat, optcon='mol/m3', optt='Tpot', optp='m')

# Print out depth and carbonate ion concentration (converted to umol/kg)
print depth, co3 / (DENis * 1e-6)

\endverbatim

3) Read input from a .csv (spreadsheet) file:

\verbatim
$ python
import numpy as np
import mocsy

# Read .csv input file with 7 columns: flag, Alk(mol/kg), DIC(mol/kg), salinity(psu), temp(C), press(db), PO43-(mol/kg), SiO2(mol/kg)
infilename = "DIC-Alk-P_vary_input.csv"
indata = np.loadtxt(infilename, delimiter=',', dtype='float32', skiprows=1)
n = len(indata)
alk = indata[:,1]
dic = indata[:,2]
sal = indata[:,3]
temp  = indata[:,4]
depth = indata[:,5] * 10 #Convert from bars to decibars
phos  = indata[:,6]
sil   = indata[:,7]

# Specify that latitude = 45°N for all samples (for depth to pressure conversion)
# note: "lat" is required input for mocsy, but results are only weakly sensitive
lat  = np.zeros([n,], dtype='float32')
lat.fill(45.)

# Recover computed carbonate system variables
ph, pco2, fco2, co2, hco3, co3, OmegaA, OmegaC, BetaD, rhoSW, p, tempis = \
mocsy.mvars (temp, sal, alk,dic, sil, phos, depth, lat, optcon='mol/kg', optt='Tinsitu', optp='db', optb="u74", optk1k2='l', optkf="dg")

# Save results to another .csv file
results = np.dstack ((sal, temp, p, ph, pco2, fco2, co2, hco3, co3, OmegaA, OmegaC, BetaD, dic, alk))
outfilename = "DIC-Alk-P-vary-from-mocsy.csv"
outfile = open(outfilename, "w")
outfile.write ("S, T, P, pH, pCO2, fCO2, CO2, HCO3, CO3, OmegaAragonite, OmegaCalcite, BetaD, DIC, ALK\n")
np.savetxt(outfile, results[0], delimiter=',', fmt="%1.6g")
outfile.close()

\endverbatim

4) Example with netCDF files for input and output:

MODEL example: See the python script mocsy_CMIP5.py to compute carbonate
chemistry variables from CMIP5 model input.

DATA example: See the python script mocsy_GLODAP.py to compute carbonate chemistry
variables from GLODAP DIC and Alk along with T, S, PO43- and SiO2 from
the World Ocean Atlas (2009)

*/

//-----------------------------------------------------------

/*! \page refs References

Antonov, J. I., Seidov, D., Boyer, T. P., Locarnini, R. A., Mishonov,
A. V., Garcia, H. E., Baranova, O. K., Zweng, M. M., and Johnson,
D. R.: World Ocean Atlas 2009, Volume 2: Salinity, S.  Levitus, ed.,
Atlas NESDIS 69, NOAA, U.S. Government Printing Office, Washington,
D.C., 184 pp., 2010.

Dickson, A. G., Sabine, C. L., and Christian, J. R.: Guide to best
practices for ocean CO2 measurements, PICES Special Publication 3,
IOCCP Report No. 8, http:// aquaticcommons.org/1443/, 2007.

Frankignoulle, M.: A complete set of buffer factors for acid/base CO2
system in seawater, <em>J.  Mar. Sys., 5,</em> 111–118, 1994.

Garcia, H. E., Locarnini, R. A., Boyer, T. P., Antonov, J. I., Zweng,
M. M., Baranova, O. K., and Johnson, D. R.: World Ocean Atlas 2009,
Volume 4: Nutrients (phosphate, nitrate, silicate), S. Levitus, ed.,
Atlas NESDIS 71, NOAA, U.S. Government Printing Office,Washington,
D.C., 398 pp., 2010.

Lewis, E. and Wallace, D. W. R.: Program Developed for CO2 System
Calculations, ORNL/CDIAC- 15 105, Carbon Dioxide Inf. Anal. Cent., Oak
Ridge Natl. Lab., Oak Ridge, Tenn., 38 pp., 1998.

Lee, K., Kim, T.-W., Byrne, R. H., Millero, F. J., Feely, R. a., and
Liu, Y.- M.: The universal ratio of boron to chlorinity for the North
Pacific and North Atlantic oceans, <em>Geochimica et Cosmochimica Acta,
74,</em> 1801--1811, 2010.

Locarnini, R., Mishonov, A., Antonov, J., Boyer, T., Garcia, H.,
Baranova, O., Zweng, M., and Johnson, D.: World Ocean Atlas 2009,
Volume 1: Temperature, S. Levitus, ed., Atlas NESDIS 68, NOAA,
U.S. Government Printing Office, Washington, D.C., 184 pp., 2010.

Millero, F. J.: Carbonate constants for estuarine waters, <em>Marine and
Freshwater Research, 61</em>, 139--142, 2010.

Orr, J. C. and Epitalon, J.-M.: Improved routines to model the ocean
 carbonate system: mocsy 1.0, <em>Geosci. Model Dev. Disc., 7,</em>
 2877–2902, doi:10.5194/gmdd–7–2877–2014, 2014.

Orr, J. C., Epitalon, J.-M., and Gattuso, J.-P.: Comparison of seven
packages that compute ocean carbonate chemistry, <em>Biogeosciences
Discuss., 11,</em> 5327–5397, doi:10.5194/bgd-11-5327-2014, 2014.

Saunders, P: Practical conversion of pressure to depth,
<em>J. Phys. Oceanogr., 11</em>, 573-574, 1981.

van Heuven, S., Pierrot, D., Rae, J. W. B., Lewis, E., and Wallace,
D. W. R.: MATLAB Program Developed for CO2 System Calculations,
ORNL/CDIAC-105b, Carbon Dioxide Inf. Anal. Cent., Oak Ridge
Natl. Lab., Oak Ridge, Tenn., 2011.

*/