Environmental Engineering Reference
In-Depth Information
K
z
=K
h
, where
K
z
is the eddy diffusivity for tracer species, and
K
h
is the eddy
diffusivity for heat. The
Kz
equations for the various stability regimes of the surface
layer and layers above that and below the PBL top are repeated in (Lee et al.,
2009).
2.2. Use NAM predicted Kz for CMAQ vertical mixing
The NAQFC had an important improvement in the vertical grid alignment between
NAM and CMAQ in 2006. Both models are now using a common hybrid sigma-
P
vertical coordinate. NAM uses 61 interface levels and CMAQ in NAQFC selects a
subset of 23 levels from them with coarser spacing near the model top at 100 hPa.
In the NAM,
Kz
is defined at these interface surfaces from the Mellor-Yamada
Level 2.5 turbulence closure scheme (Janjic, 2001). In accordance with Janjic, the
Turbulent Kinetic Energy (TKE),
2
2
2
r
∂
q
q
∂
∂
q
(1)
(
)
+
V
•
∇
−
[
Kz
(
)]
=
P
s
+
Pb
−
ε
∂
t
2
2
∂
z
∂
z
2
2
'
2
'
2
'
2
where
u
+ ;
V
is the mean wind;
Ps
is the shear production;
Pb
is production by buoyancy;
and ε represents rate of dissipation of turbulent energy.
Kz
is given by
q
is the sum of square of the wind turbulence fluctuations,
v
w
(2)
Kz
=
qS
q
l
where
l
is the master length scale for turbulence, and
S
is an empirical constant
for which the numerical value of 0.2 was found (Mellor and Yamada 1982) to
optimize agreement between model results and observed data.
2.3. Use hourly averaged NAM predicted Kz for CMAQ vertical mixing
NAQFC is an offline system. NAM feeds meteorological data to CMAQ hourly. A
total of 12 3-D fields and 33 2-D fields were thus past. Most fields are instantaneous
values, except for the following time-averaged 2-D fields: precipitation rate,
shortwave and long-wave radiation fluxes, and latent and sensible heat fluxes.
Within CMAQ, the Models-3 Input/Output Applications Programming Interface
(I/O API) temporally interpolated all these 2-D, and 3-D fields between the full
hours into the advection time steps which are typically 5 min. Ideally all fields fed
should be hour-centered hourly averaged fields to match the temporal interpolation.
To understand the impact of such considerable restructuring of the coupling
interface,
Kz
, one the most influential fields in governing surface O
3
concentration
is chosen to be studied. A β-version of NAM is used to feed CMAQ with hourly
averaged
Kz
output at full hours for the duration of the previous hour. Therefore it
