638
R. Feistel et al.: Oceanographic application and numerical implementation of TEOS-IO: Part 1
Ocean Sci., 6, 633-677, 2010
www.ocean-sci.net/6/633/2010/
equation without the cancelling terms increases speed and ac
curacy. In such cases a function is more naturally (and easily)
implemented by calling the separate functions (Eqs. 2.3-2.5)
rather than their combination in g s (S A , T, P).
The expansion terms gi(T,P), Eqs. (2.3-2.5), together
with their partial derivatives are available from the library
function sal_g_term_si.
2.4 Humid air
For a correct description of the thermodynamic properties at
the ocean-atmosphere interface a thermodynamic potential of
humid air is required and available from the literature (Feis
tel et al., 2010a). A related document is in preparation by
IAPWS (2010). The Helmholtz function for dry air of Lem
mon et al. (2000) has the form of the molar Helmholtz en
ergy, / A,mo1 (T, p mo1 ), depending on absolute temperature T
(ITS-90) and molar air density, p mo1 . For its conversion to
the specific Helmholtz energy, / A , depending on the mass
density, p,
f A (T,p) = -L/ A ’ mol (V,—
M A J V M A
(2.6)
the molar mass of air, Ma=28.965 46 g mol -1 , is computed
from the recent highly accurate air composition model of Pi
card et al. (2008). The dry-air part (Eq. 2.6) can be combined
with the vapour part, / V =/ F (IAPWS-95, Sect. 2.1), in
volving the second virial coefficient B A w(T) of Harvey and
Huang (2007) and the third virial coefficients C AA w(T) and
Caww ( 7 ) of air-vapour interaction reported by Hyland and
Wexler (1983), to obtain the Helmholtz function of humid
rAV
air, / , as
/ AV (A, T, p) = ( 1 - A) / v (T, ( 1 - A) p) + A/ a (T, Ap) (2.7)
+2A(1-A)p
RT
M A M W
B A w(T) + -p
A (1—A)
-rr- C A aw(T)-\———C A ww(T)
M A «W
Here, p is the density of humid air, A is the mass frac
tion of dry air in humid air, q—\—A is the specific hu
midity, (1—A)p is the absolute humidity, and r—(\—A)/A
the humidity ratio or mixing ratio (van Wylen and Sonntag,
1965; Gill, 1982; Emanuel, 1994). R=8.314 51 J mol -1 K -1
is the molar (or universal) gas constant used by Lem
mon et al. (2000), rather than the most recent value
of R=8.314 472 J mol -1 K -1 (Mohr et al., 2008), and
Mw=0.018015268kgmol -1 is the molar mass of pure wa
ter (IAPWS, 2008b). The effective molar mass of humid air
Mav depends on the mass fraction A in the form
1
M AV = (2.8)
(1 - A)/M w + A/M a
The mass fraction A of air is computed from the mole frac
tion x A of dry air as
x A M A
x A M A + (1 - x A )M w
(2.9)
_ XA
~~ 1 - (1 — x A ) (1 — M w /M A ) ’
and it follows that the mass fraction of vapour is given by
_ ^ _ 1 x A M A
x A M A + (1 - x A )M w
1 ~ * A
1 — Xa( i _m a /M w )‘
(2.10)
The inverse function of Eq. (2.9), i.e., the mole fraction of air
as a function of the mass fraction of air, is
= x _ (1 - A)/Mw
(1 - A)/M w + A/M a
A(M w /M a )
1 - A(1 - M W /M A )
(2.11)
and the related mole fraction of vapour is
(1 - A)/My
(1 - A)/M w + A/M a
1 - A
(2.12)
1 - A(1 - M W /M A )'
The Helmholtz potential (Eq. 2.7) is formally symmetric in
the fractions of air and of water vapour. We note that the
Helmholtz functions / v and f A that we have chosen to use
in Eq. (2.7) are complete expressions rather than truncated
expansions in terms of powers of density. Consequently,
they include contributions corresponding to higher powers
of density than included in the cross-virial terms represented
by the third term in Eq. (2.13), / mlx =/ AV —A/ a —(1—A)/ v .
Equation (2.7) is thus an inhomogeneous approximation for
mula with respect to the powers of density and the related
correlation clusters. However, its validity is not restricted to
small specific humidity, q={\—A), such as some 1-3% of
ten assumed for empirical equations used in meteorology. It
can even be applied to physical situations in which air is the
minor fraction, such as condensers of desalination plants or
headspaces over subglacial lakes. The mass fraction A rather
than the specific humidity q is chosen as a composition vari
able of humid air for its analogy to Absolute Salinity; the two
describe the amount of natural mixtures, gases or salts, con
tained in ambient water in either gaseous or liquid form. This
leads to thermodynamic equations that are formally similar
in A and 5 a (Feistel et al., 2010a).
The range of validity is bounded by the simultaneous va
lidity of the vapour formula (IAPWS-95), of the dry-air for
mula (Lemmon et al., 2000) and of the cross-virial expan
sion. The dry-air function correctly describes reliable ex
perimental data for pressures up to 70 MPa and for tempera
tures from 60 to 873 K; the maximum air density in this re
gion is 1035.8 kg/m 3 . The temperature range where all three