Some Micrometeorological Heresy

1
Some Micrometeorological Heresy Confronting the
Issues of Urban Areas and Complex Terrain Bruce
B. Hicks, Director Air Resources Laboratory 1315
East West Highway Silver Spring, MD 20910
2

• Some history of air-surface exchange
• The origins, driven by gas warfare.
• Smokescreen studies in the US in the 1930s and
  1940s.
• Forest meteorology starting in 1941.
• Theory of exchange, starting in the 30s and still
  contentious in the 60s (Deacon, Swinbank,
  Priestley, Monin, Obukhov, Panofsky, etc.)
• Eddy correlation developments from the 30s to the
  70s. Heat budgets in the 1960s.
• Flux gradient relationships first documented in
  1961 (Deniliquin), then improved in subsequent
  studies at Kerang, Hay, Tsimlyansk, Gurley,
  Kansas, Minnesota, and Mt. Gambier.

3
The search was for an outdoors laboratory, to
test the theories then being developed. e.g.
Swinbanks exponential wind profile
(1962/3) McIlroys belief that Kh ? Kw
(1963/64) (everybody accepted that convection
implied Kh ? Km) Priestleys free convection
Monteiths stomatal-control theories Monin,
Obhukov, Yaglom, Businger, Dyer, Zilitinkevitch,
etc. Edithvale pasture Deniliquin, Hay and
Kerang short grass Gurley bare soil Kansas
grass (sonic anemometry) Mt. Gambier pine
plantation (helicoid propellers) Wangara,
Minnesota extension to the PBL,
grassland Sangamon extension to the PBL,
cropland Prairie Grass, CAPTEX, ANATEX, METREX
4
The effects of shared variables confound much of
micrometeorology. Tables of random numbers,
constrained to give values within observed
bounds, can yield surprising results. Random
numbers yielded this plot, which mirrors
published relationships. The flux-gradient
relations accepted by convention are strongly
affected. F1. That we understand the way
turbulence intensity scales with stability.
5
The multiple resistance approach of
agrometeorology offers considerable
advantage. Consider free convection, working in
parallel with mechanical mixing.. The free
convection resistivity is Rfc
A/(-z/L)1/3.(Kmn) The neutral mechanical
resistivity is Rmn 1/(kuz) 1/Kmn The
total resistivity is then Reff 1/(Kmn).(1
A(-z/L)1/3)
6
Hence, we expect to find ? 1/1
A(-z/L)1/3 In practice, the empirical
forms are well approximated by this form.
Common-variable effects might well explain
departures. Consider turbulence
intensities ?(w)/u Fn(-z/L) In fact, u
-r(u,w).?(w).?(u)0.5, so that both u and ?(w)
can be considered shared variables.
F2. In air, the flux gradient relationships are
well described.
7
The effects of surface vegetation irregularity
and topography combine to generate a highly
uncertain picture of the areal averaged air
surface exchange.
The vegetation distribution across a test area of
the Adirondacks, based on satellite data.. A
field survey indicated the need for considerable
ground truthing. The original estimate between
1 and 10 of the pixels. F3. For vegetation
effects, satellite data are a sufficient answer
8
Ozone deposition velocities computed using a
multiple resistance approach, allowing for
complex terrain effects.
Mid Adirondacks, July F4. Existing surface
exchange formulations describe fluxes adequately
(perhaps true for H, LE).
9
(No Transcript)
10
F5. The effects of surface complexity are well
described by current models.
11
Todays heresies To what extent are we fooling
ourselves? F1. We understand the way turbulence
intensity scales with stability. F2. In air,
the flux gradient relationships are well
described. F3. For vegetation effects,
satellite data are a sufficient answer F4.
Existing surface exchange formulations describe
fluxes adequately (perhaps true for H, LE). F5.
The effects of surface complexity are well
described by current models. F6. We cannot
measure eddy fluxes in complex terrain. F7. We
know enough about the surface heat balance. And
we have not even considered the fallacies about
soil moisture!
12
Fort Peck Reservation, NE Montana
13
In collaboration with other agencies and
universities, we set up surface stations, and
conduct aerial transects to assess spatial
variability. The matter of severe terrain
complexity remains to be addressed.
14
Todays challenge Getting data in real time for
urban use (DCNet and UrbaNet)
Meteorology by NOAA Rad and chemical sensors by
DOE Communications by DOE and
others Dispersion forecasts by NOAA. Bio threat
danger by NOAA
15
Initial DCNet Met Site DOC (Hoover) Building
Measurements are made at 20 Hz. Summaries of the
data are transferred every 15 minutes.
16
DCNet Locations The intent is to set up 10 to
20 stations in the downtown DC area.
The area covered will be expanded as funding
permits.
17
NAS Maximum vector speed 32 m/s (62 knots) (72
mph)
Maximum u/v component is the maximum speed
recorded during the 15-minute averaging period
Wind Speed is the 15-minute average velocity
measured at the station
SSMC3 Maximum vector speed 37 m/s (72 knots)
(83 mph)
18
DCNet already shows why standard airport data are
not appropriate for downtown dispersion
applications. The wind roses show where winds
are from, for different locations.
19
Sample DCNet Data
Turbulence differs greatly from place to place,
although the general patterns seem to be shared.
These data are for DC -- Commerce Building, Natl
Arboretum, and NOAA Silver Spring Building III.
20
Sample DCNet and UrbaNet Data
Data continue to yield credible sensible heat
fluxes. Note values close to zero for the two DC
sites (left), but (as expected) often continually
positive for the New York site (below) due to
enhanced heat storage. F6. We cannot measure
eddy fluxes in complex terrain.
21
In 1972, studies over a pine plantation in Mt.
Gambier, Australia, verified that heat storage in
the canopy vastly affected the surface energy
budget at dusk and dawn. The error could easily
be 100 W/m2. F7. We know enough about the
surface heat balance.
Mt. Gambier
Current data from the NOAA forest meteorology
facility in Oak Ridge confirm the expectation.
Much of the eastern US is forested. Omission of
canopy storage will doubtlessly lead to errors in
the predicted growth of the daytime mixed layer
and the onset of stratification at night.
Oak Ridge
22
The modeling systems are now being tested at a
number of locations. A default-source forecast
initiates at the click of a button. The answer
appears in about one minute. Once source-term
information is available, this can be entered to
re-initiate the model run. Outputs are
GIS-compatible.
Red dots show trajectory end points at hourly
intervals. This particular output uses HYSPLIT.
Other models will be integrated. At small
scales, local DCNet data are used.
23

• In Nevada, the NOAA dispersion forecast system
  is used to predict susceptible areas following a
  release into the atmosphere at the Nevada Test
  Site.

The same system is used to forecast plume
dispersion from the Las Vegas strip. ? Note the
effects of changing meteorology.
24

• We are now moving towards forecasting personal
  exposure weather, air quality, and risk
  following releases of hazardous materials.
• Urban test beds are being set up. The start is
  in Washington, DC, and in New York City. Other
  focal areas
• Houston the site of the multi-agency 2004 air
  quality study
• Oklahoma City where the Joint Urban 2003
  study was conducted
• Las Vegas where an urban test bed is already
  being constructed by CIASTA (a NOAA Cooperative
  Institute)
• Los Angeles for development in an area
  dominated by mesoscale meteorology
• The basics
• -- surface towers for state variables, velocity,
  and turbulence/fluxes
• -- sodar, lidar and radar for winds aloft
• -- data assimilative fine-grid models (how
  fine???)
• -- probabilistic approaches to address issues of
  street canyons.