Modeling Local Urban Meteorology for Smoke Obscuration

1
Modeling Local Urban Meteorology for Smoke
Obscuration Donald Hoock Saba (Lou)
Luces Ronald Cionco U.S. Army Research
Laboratory Computational and Information Sciences
Directorate Battlefield Environment
Division AMSRL-CI-E White Sands Missile Range, NM
88002-5501
See Paper 02F-SIW-071 2002 Fall Simulation
Interoperability Workshop, 8-13 September 2002,
Orlando, FL
2
Army Transformation to an Objective Force
Continues
Emphasis on light force operations in more
complex terrain
By 2020 over 60 percent of the worlds
population will live in cities. Complex urban
environments, ranging from modern skyscraper
jungles, to huge shantytowns are therefore an
increasingly predominant feature of the
operational environment. These environments will
challenge the Unit of Action in a complex 3D
fashion - elevated, surface and
subsurface. (TRADOC PAM 525-3-90 OO, "The
United States Army Objective Force Operational
and Organizational Plan for Maneuver Unit of
Action, 22 JUL 2002)
3
Urban Target Acquisition is No Longer a
Range-Limited Problem
Target acquisition performance modeling will be
dominated by signatures (complex illumination
environments) and visibility-reducing obscurants

Complex illumination
Extensive use of obscurants to conceal movement
High urban clutter
Short lines of sight, smaller optics and larger
sensor fields of view
4
COMBIC - an idealized, deterministic obscuration
model
The Combined Obscuration Model for Battlefield
Induced Contaminants (COMBIC) is an EO/IR
transmission model for battlefield smoke and dust
It considers the 10 minute average of fine scale
fluctuations and plume meander COMBIC is a mean
plume obscurant model represented as a smooth
Gaussian concentration about a uniform wind
direction centerline.
5
Obscuration Modeling
Yesterdays simple COMBIC obscurant model
designed for rural terrain and uniform winds
Our challenge is to develop an urban non-uniform
obscurant model suitable for new generation Army
combat simulations ------- Somewhere within the
spectrum of SNE models and simulations
Todays new urban combat simulation model non
uniform wind transport and dispersion based on
Urban Morphology
Detailed Urban Databases
Tomorrows very high fidelity Urban CFD and Large
Eddy Simulations
6
Urban Terrain Zone (UTZ) Classification and
Morphology
Developed by Richard Ellefsen, San Jose State
Univ. - Classify 16 domains with common
characteristics
7
Urban Terrain Morphology Databases (50 100 m)
8
Example San Francisco
IMETS BFM Forecast Model Input
High Resolution Wind Model Effects of Urban
Morphology
RIMPUFF Dispersion Model
9
New algorithms for obscuration in urban domains
and complex terrain

• The Urban Obscurant Propagation Model
• non-uniform wind transport
• puff dispersion model
• Technical Challenges
• Efficiently compute transmission through the
  puff ensembles
• Consolidate clouds to reduce the number of puffs
• Building and urban domain boundary intersection
  with larger puffs

10
The Problem Compute line of sight propagation
efficiently for puff ensembles

• T transmission (non-dimensional 0 - 100)
• a obscurant bulk optical property mass
  extinction
• coefficient at the relevant wavelength and
  aerosol
• size distribution in (m2 g-1)

CL concentration length (g m-2) C(x)
concentration at location x (g m-3) xt target
location xo observer location
11
Each puff contributes its obscurant to the Line
of Sight
Obscurant CL sums the individual integrated mass
concentration contribution from each sub-cloud on
the Line of Sight (LOS)
12
Example - puffs in an urban setting - effects of
obscuration (darkening transmission loss)
Four small puffs as they might obscure in an
urban setting. (Transmission loss is always a
darkening effect. Back-lit scattering, a
lightening effect, is only being approximated
here.)
Sigmas for these small clouds range from 1 to
1.5 m with center CLs ranging from 1 to 2 g m-2
and a 3 m2 g-1
13
Computing CL efficiently along an LOS through a
gaussian ellipsoid - basic geometry
Every point on the LOS satisfies
For some s in the range
With direction cosines defined by
And target-observer range
14
Computing CL efficiently along an LOS through a
gaussian ellipsoid - solution
Cumulative CL for observer and target distances
from closest point D along the LOS in units of
W sigmas
15
Computing CL efficiently along an LOS through a
gaussian ellipsoid - simple analytic solution

• Rt and Ro are in units of effective W standard
  deviation sigmas
• Ro ltlt -3 observer outside cloud F(Ro) 0.
• Rt gtgt 3 target outside cloud F(Rt) 1.

16
Simple Surface Reflection Alternatives
Cloud Mirror Reflection in the Surface Boundary
A virtual mirror image of the cloud is created.
The reflected obscurant is the superposition
of the real and virtual clouds. Obscurant tends
to hug the surface.
Side View
Cloud Mass Renormalization The mass of smoke on
the other side of the surface is totaled up and
used to re-adjust the entire real cloud by
total mass so that none is lost.
In both cases it is possible to reduce the
adjustment to account for surface deposition
factors, if any.
17
Four Methods to Correct for Building Interaction
Effects on Cloud Concentration

• Rescale Using the Urban Morphology Building
  Density
• - applicable to clouds larger than the morphology
  resolution

Buildings cover 20 of the domain, so increase
the puff concentration by 1/(1-0.20) gt 25
increase
(2) Ignore the wall extend the Line of Sight to
catch all the obscurant
18
Surface Mirror Reflection Can Involve Multiple
Walls

• Surface Mirror Reflection is straightforward for
  simple urban geometries
• The number of clouds is doubled for each mirror
  reflection surface.
• Concentrations and CL values require summing
  over each sub-cloud
• Corners begin to show the increasing
  complication, however.

(Assume this is a top-down view of an urban
building corner)
(Top View)
Building
Cloud Reflection Method
Cloud
19
Problem Surface Mirror Reflection Method Rapidly
Becomes Ambiguous in Street Canyons
Surface Mirror Reflection becomes ambiguous and
likely incorrect as one introduces street canyons
between building surfaces.
Street Canyon
Simple reflection clouds
Top View
20
How much modeled smoke is inside the building?
- Rescaling transformation for the cloud
renormalization method
Divide all x,y and z coordinates by cloud sx, sy
and sz as for the CL calculation
Top View

• The procedure is the same as for CL
• Define the wall boundary as a plane or as a
  horizontal line boundary at one point Xo with
  direction W (a, b, 0) direction cosines.
• Compute D D as shown earlier
• Then

21
Cloud Renormalization Handles street canyon
geometries better and requires fewer calculations
than reflection clouds
Puff principal axes do not need to align with the
walls. The equations work for all orientations
(transform scaling by sigmas distorts the
building corners and intersection angles, but the
wall surfaces remain planar)
22
Summary and Conclusions

• We are currently upgrading the uniform wind
  COMBIC model to urban non-uniform wind domains.
• An efficient approach is to use the Urban
  Morphology statistics in place of individual
  building effects on the flow.
• The CL required for any path through an
  ellipsoidal gaussian puff can be easily computed
  in closed form.
• The observer or target can be inside the cloud.
• There are at least 4 ways to account for the
  urban wall interactions with the clouds, ranging
  from ignoring the walls to a closed form
  correction for individual buildings.