412 The Atmospheric Boundary Layer
such as the logarithmic wind profile, are founded on
similarity theory, which basically ignores physics and
dynamics and focuses on the dimensions of the key
variables. The nature of turbulence and entrainment
of ambient air into clouds is also incompletely under-
stood, as well as the feedbacks between clouds and
the fluid motions in the boundary layer.
Despite these formidable obstacles, significant
advances are being made in many areas. Boundary-
layer models are being coupled with larger-scale
models to capture the important effects of multiscale
advection. Remote sensing is used to observe the
eddies and thermals that cause the turbulence, and
very fine resolution computing is being used to simu-
late boundary-layer turbulence, as explained later.
Horizontal advection of heat, moisture, and momen-
tum by larger-scale mean wind nearly always domi-
nates over turbulent effects in the boundary layer, even
during periods of fair weather and light winds. Also,
vertical advection by the large-scale motion field is
usually the same order of magnitude or larger than the
turbulent entrainment at the mixed-layer top. To prop-
erly include these larger-scale advective effects, bound-
ary-layer models are often embedded in mesoscale
models, which, in turn, are nested in large-scale numer-
ical weather prediction models.
Remote sensing has been used increasingly by
researchers since the 1970s to observe the boundary
layer. Clear-air radar emits microwaves to observe
boundary-layer eddies, as made visible by humidity-
related microwave refractivity fluctuations. Lidar (light
detection and ranging) transmits a beam of radiation at
visible or near-visible wavelengths to illuminate
aerosols carried aloft by thermals (Fig. 9.39) and uses
large telescopes to capture the photons scattered back
to the lidar. Sodar (sound detection and ranging) emits
loud pulses of sound and uses sensitive microphones to
detect faint echoes scattered from regions of strong
temperature gradients in and around surface-layer
eddies and plumes. Profilers monitor the Doppler shift
of pulses of off-vertically propagating radio waves,
from which the wind profile can be inferred, while
RASS (radio acoustic sounding systems) use radio
waves to measure the speed of sound waves, from
which the temperature profile can be inferred.
Realistic numerical simulation of boundary-layer
phenomena is inherently more computationally inten-
sive than numerical weather prediction. The range of
horizontal scales is larger (in a logarithmic sense), so
more grid points are needed. In addition, motions in
the boundary layer tend to be fully three dimensional
and nonhydrostatic so additional layers are required
in the vertical as well. The timescales of microscale
turbulence are orders of magnitude shorter than the
timescale for phenomena such as baroclinic waves so
much shorter time steps are required.
Despite these formidable computational require-
ments, some notable advances have been made in
recent years. A particularly productive approach has
been large eddy simulation (LES), in which numeri-
cal weather prediction models are run on high-per-
formance computers with 10-m grid spacing, which
is sufficient to resolve the larger, more energetic
eddies. Finer resolution direct numerical simulation
(DNS) computer models have been run over very
small domains with grid spacings as fine as millime-
ters, which is sufficient to resolve all eddy sizes down
to scales on which molecular conduction and diffu-
sion dominate. As computer power continues to
increase, DNS and LES are merging to provide
direct computation of turbulent flow over arbitrarily
complex terrain.
As more of these new observing and numerical-
modeling capabilities come on line, it is becoming
increasingly possible to transcend the limitations of
statistical approaches and similarity theory by pursu-
ing a more phenomenogically based approach to
boundary-layer studies. Part of the thrill of boundary-
layer research is that there are still many important
phenomena and processes waiting to be discovered.
Relative backscatter intensity (dB)
at 1543 nm
33.5 34.5 35.5 36.5 37.5
90.0
38.5
90.0
90.0
EL
AZ
2.5
2.0
1.5
1.0
0.5
0.0
MT09.3309.3209:31
2.5
2.0
1.5
1.0
0.5
0.0
Height above ground (km)
Fig. 9.39 Time-height cross section showing boundary-layer
thermals (red and yellow colors) as they drift over a vertically
pointing lidar. Pollution particles in the rising thermals, which
backscatter light back to the lidar telescope, are rendered in
red and yellow colors. Black, purple, and blue indicate air
that is relatively clean in the free atmosphere. [Courtesy of
Shane Mayor, National Center for Atmospheric Research.]
P732951-Ch09.qxd 9/12/05 7:48 PM Page 412