United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park, NC 27711
EPA-454/R-92-017
May 1993
Air
& EPA
DEVELOPMENT AND TESTING
OF DRY DEPOSITION
ALGORITHMS
-------
EPA-454/R-92-017
Development and Testing
of Dry Deposition
Algorithms
U.S. Environmental Protection Agency
Region 5, Library (PL-12J)
77 West Jackson Boulevard, 12th Floor
Chicago, IL 60604-3590
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Technical Support Division
Research Triangle Park, NC 27711
May 1993
-------
This report has been reviewed by the Office of Air Quality
Planning and Standards, U.S. Environmental Protection Agency, and
has been approved for publication. Any mention of trade names
and commercial products is not intended to constitute endorsement
or recommendation for use.
EPA-454/R-92-017
-------
ACKNOWLEDGEMENTS
This report has been prepared by Sigma Research Corporation
and funded by the U.S. Environmental Protection Agency under
Contract No. 68-D90067, with Jawad S. Touma as Work Assignment
Manager.
-------
PREFACE
The ability to accurately estimate deposition due to
atmospheric releases is important to the modeling community.
Limitations of the Industrial Source Complex (ISC2) models (dated
92273) for estimating deposition of small particles have been
known for some time. An improved algorithm for estimating
deposition for a wide range of pollutants has been developed and
tested and is described in this report.
The Environmental Protection Agency must conduct a formal
and public review before the Agency can recommend for routine use
this new algorithm in regulatory analyses. This report is being
released to establish a- basis for reviews of the capabilities of
this methodology in routine dispersion modeling analyses. This
report is one part of a larger set of information on the ISC2
models that must be considered before any formal changes can be
adopted.
-------
Table of Contents
Section Page
1. Introduction 1-1
2. Dry Deposition 2-1
2.1 Dry Deposition of Particulate Matter 2-5
2.1.1 ISC Method 2-7
2.1.2 CARS Model 2-13
2.1.3 ADOM Model - Particles 2-21
2.1.4 UAM-V Model 2-25
2.2 Dry Deposition of Gaseous Pollutants 2-26
2.2.1 RADM Model 2-27
2.2.2 ADOM Model - Gases 2-28
2.2.2 NOAA/ARL Models 2-31
3. Plume Depletion Techniques 3-1
3.1 Source Depletion 3-1
3.2 Surface Depletion 3-2
*
3.3 K-theory Approach 3-2
3.4 Modified Source Depletion 3-4
4. Calculation of Meteorological Variables . . .' 4-1
4.1 Unstable/Neutral Conditions 4-1
4.2 Stable Conditions 4-5
5. Model Evaluation Protocol 5-1
5.1 Evaluation Approach ." 5-1
5.2 Stratification of Deposition Velocity Datasets 5-2
5.3 Comparison of Predictions and Observations 5-3
5.4 Scoring Model Performance Using Composite Measures 5-6
6. Model Evaluation Data Bases 6-1
6.1 Particle Data Sets 6-1
6.2 Gas Data Sets 6-2
-------
Table of Contents - Continued
7. Results of Model Evaluation 7-1
7.1 Particle Deposition Models 7-1
7.1.1 Full Data Sets 7-1
7.1.2 Stratification by Particle Diameter 7-12
7.1.3 Stratification by Roughness Length 7-15
7.1.4 Stratification by Leaf Area Index 7-15
7.1.5 Stratification by Day vs Night 7-18
7.1.6 Stratification by Friction Velocity 7-20
7.1.7 Stratification by Temperature 7-20
7.1.8 Estimation of CPM from Tables 7-23
7.1.9 Selection of the Recommended Deposition Model 7-23
7.2 Sensitivity Calculation 7-27
7.3 Summary of Results 7-30
8. Summary of Recommendations 8-1
9. References 9-1
Appendix A Estimation of ISC Deposition Velocity A-1
Appendix B Supplemental Graphics B-1
Appendix C Observational Particle Deposition Velocity Data Sets C-l
Appendix D Predicted Deposition Velocities vs Particle Diameter D-1
Appendix E Implementation of the Modified Source Depletion Method in ISC2 E-1
-------
List of Figures
Figure 2-1. Summary of observed SO2 deposition velocities 2-3
Figure 2-2. Observed deposition velocities (vd) as a function of particle size for 2-4
1.5 g/cm3 density particles.
Figure 2-3. Predicted deposition velocities for u. = 100 cm/s and particle 2-6
densities of 1, 4, and 11.5 g/cm3.
Figure 2-4. Illustration of the reflection coefficient scheme used in ISC . 2-8
for reflection coefficients of 0.0, 0.5, and 1.0.
Figure 2-5. Reflection coefficient as a function of the gravitational settling 2-12
velocity
Figure 2-6. Deposition velocity as a function of particle diameter as predicted 2-18
by the Sehmei/CARB model for three different values of ambient
temperature (0°, 60°, 100° F).
Figure 2-7. Deposition velocity as a function of particle diameter as predicted 2-19
by the Sehmel/CARB model for surface roughness lengths of 0.001,
0.01, 3, and 10 cm.
Figure 5-1. Schematic illustration of division of deposition velocity data into subsets. . . . 5-4
Figure 7-1. A summary of absolute value of the fractional bias using all data 7-6
(n = 168).
Figure 7-2. Summary of absolute value of the fractional bias averaged over all 7-7
12 subsets.
Figure 7-3a. Co-plot of fractional bias of the standard deviation for each of 7-8
the deposition models.
Figure 7-3b. Co-plot of fractional bias of the 10 largest deposition velocities 7-9
for each of the deposition models.
vn
-------
List of Figures - Continued
Figure 7-3c. Co-plot of fractional bias of the 10 smallest deposition velocities 7-10
for each of the deposition models.
Figure 7-3d. Co-plot of fractional bias of the 10 largest and 10 smallest 7-11
deposition velocities for each of the deposition models.
Figure 7-4. A summary of the total CFB of the six types of data subsets 7-14
Figure 7-5. A ranking of the models by CPM 7-25
Figure 7-6. A summary of the MCM for each unique model pair 7-26
Figure B-la. Scatter plot of observed deposition velocity (cm/s) versus model B-l
predicted deposition velocity (cm/s) for the complete small
particle data set.
Figure B-lb. Scatter plot of observed deposition velocity (cm/s) versus model B-2
predicted deposition velocity (cm/s) for the complete small
particle data set.
Figure B-lc. Scatter plot of observed deposition velocity (cm/s) versus model B-3
predicted deposition velocity (cm/s) for the complete small
particle data set.
Figure B-ld. Scatter plot of observed deposition velocity (cm/s) versus model B-4
predicted deposition velocity (cm/s) for the complete small
particle data set.
Figure B-le. Scatter plot of observed deposition velocity (cm/s) versus model B-5
predicted deposition velocity (cm/s) for the complete small
particle data set.
Figure B-lf. Scatter plot of observed deposition velocity (cm/s) versus model B-6
predicted deposition velocity (cm/s) for the complete small
particle data set.
via
-------
List of Figures - Continued
Figure B-lg. Scatter plot of observed deposition velocity (cm/s) versus model B-7
predicted deposition velocity (cm/s) for the complete small
particle data set.
Figure B-lh. Scatter plot of observed deposition velocity (cm/s) versus model B-8
predicted deposition velocity (cm/s) for the complete small
particle data set.
Figure B-li. Scatter plot of observed deposition velocity (cm/s) versus model B-9
predicted deposition velocity (cm/s) for the complete small
particle data set.
Figure B-lj. Scatter plot of observed deposition velocity (cm/s) versus model B-10
predicted deposition velocity (cm/s) for the complete small
particle data set.
Figure B-2a. Cumulative probability plot of deposition velocity (cm/s) using B-ll
the complete small particle data set.
Figure B-2b. Cumulative probability plot of deposition velocity (cm/s) using B-12
the complete small particle data set.
Figure B-2c. Cumulative probability plot of deposition velocity (cm/s) using B-13
the complete small particle data set.
Figure B-2d. Cumulative probability plot of deposition velocity (cm/s) using B-14
the complete small particle data set.
Figure D-l. Predicted deposition velocity for the CARB-based models for D-l
u. = 10 cm/s, z0 = 10 cm, LAI = 1.0, p = 1.0 g/cm3, and neutral stability.
Figure D-2. Predicted deposition velocity for the ADOM-based models for D-2
u. = 10 cm/s, z0 = 10 cm, LAI = 1.0, p = 1.0 g/cm3, and'neutral stability.
Figure D-3. Predicted deposition velocity for the UAM-based models for D-3
u. = 10 cm/s, z0 = 10 cm, LAI = 1.0, p = 1.0 g/cm3, and neutral stability.
IX
-------
List of Figures - Concluded
Figure E-l. Illustration of the depletion factor FQ and the corresponding profile E-2
correction factor P(x,z).
Figure E-2. Vertical profile of concentration before and after applying FQ and E-3
P(x,z) shown in Figure E-l.
-------
List of Tables
Table 1-1
Table 2-1
Table 2-2
Table 2-3
Table 2-4
Table 2-5
Table 2-6
Core Models to be Evaluated in the Study 1-4
Factors Influencing Dry Deposition Rates 2-2
Typical Surface Roughness Lengths for Various Land Use Types 2-17
Values of Chemical Input Parameters Required by the RADM 2-29
Deposition Model
Summary of Input Requirements of RADM Deposition Module for 2-30
Gases
Leaf Area Index Values as a Function of Land Use Type and Season 2-32
Summary of Input Requirements of ADOM-type Deposition Modules . .
for Gases
2-33
Table 4-1 Values of Net Radiation Constants 4-3
Table 4-2 Minimum Values of Monin-Obukhov Length During Stable Conditions for . . 4-7
Various Land Use Types
Table 7-1 A Summary of the Model Designations 7-2
Table 7-2 A Summary of the Stratifications Made to the Small Particle Data Set 7-3
(N = 168)
Table 7-3 A Summary of the Fractional and Composite Statistical Measures for . .
Each of the Models Examined
Table 7-4 A Summary of the Fractional and Composite Statistical Measures for . .
Each of the Models Examined
Table 7-5 A Summary of the Fractional and Composite Statistical Measures for
Each of the Models Examined
. . 7-16
XI
-------
List of Tables - Concluded
Table 7-6 A Summary of the Fractional and Normalized Statistical Measures for
Each of the Models Examined
. . 7-17
Table 7-7 A Summary of the Fractional and Normalized Statistical Measures for
Each of the Models Examined
. . 7-19
Table 7-8 A Summary of the Fractional and Normalized Statistical Measures for . .
Each of the Models Examined
. . 7-21
Table 7-9 A Summary of the Fractional and Normalized Statistical Measures for . .
Each of the Models Examined
. . 7-22
Table 7-10 Summary of Composite Statistical Measures that Illustrate how the CPM . . 7-24
Arises for the CARB 1 Model
Table 7-11 Aerosol Mass Fraction as a Function of Size Distribution for Two
Assumed Aged Sulfate Distributions
7-28
Table 7-12 Summary of Composite Model Performance Statistics for Two
Assumed Sulfate Size Distributions
7-29
Table C-l Observational Particle Deposition Velocity Data Sets C-l
xn
-------
1. Introduction
The intermedia transfer of pollutants from the atmosphere to land, water, and vegetation
is an increasingly important concern in many regulatory environmental impact analyses. The
process of dry deposition is a critical transport route for pollutant movement across the
air/surface interface. However, current regulatory modeling tools have limitations in their
ability to evaluate dry deposition. For example, the Industrial Source Complex (ISC) model
contains an empirical dry deposition algorithm based on reflection coefficients which is
appropriate for large particles dominated by gravitational settling (i.e., particle diameters larger
than approximately 20 um). However, this algorithm is not designed for small particles or
gaseous pollutants, both of which are of concern for many analyses involving criteria pollutants
and toxic air pollutants. Recognizing the need for a generalized, scientifically-credible dry
deposition algorithm capable of estimating deposition fluxes for a wide range of pollutants, the
U.S. Environmental Protection Agency (EPA) has sponsored a study to develop such an
algorithm for use in the ISC model. The primary objectives of the present study are to identify
deposition models suitable for regulatory use, evaluate and intercompare several of the
techniques, and implement the most appropriate approach into the ISC model.
The dry deposition flux (F) of a pollutant at a reference height above the surface can be
defined as the product of the ambient concentration of the pollutant, (x), and a "deposition
velocity" (vd), i.e., F = % vd- Therefore, a model to predict deposition fluxes requires an
estimation of both the ambient concentration and deposition velocity. The EPA has established
procedures for estimating ambient pollutant concentrations, which can be used for deposition
modeling as well. However, an approved procedure for calculating deposition velocities is not
established. The focus on this study is the testing and evaluation of algorithms to compute
deposition velocities for particulate matter.
The following criteria are being applied in the selection of the dry deposition velocity
algorithms for evaluation.
(1) The algorithm should parameterize the important physical/chemical processes
known to determine the rate of deposition.
(2) The scheme should require only routinely and/or readily available
meteorological, chemical, and physical input data.
(3) The algorithm should have modest computational requirements.
1-1
-------
(4) The formulation should be general enough to accommodate, with the proper
input parameters, a wide variety of pollutants.
Although the number of models to be tested must be limited, hybrid approaches, consisting
of the best components of two or more models, have been constructed for some of the particle
deposition velocity models. Various components of these models have been modified or
enhanced in order to provide increased generality or to eliminate weaknesses in the approach.
Table 1-1 lists the base (core) particle deposition models which have been evaluated in this
study along with the gas deposition, plume depletion, and meteorological processing algorithms
which have been reviewed.
Because pollutants that are deposited are removed from the plume, there is some
feedback between upwind plume depletion and the ambient concentration at a particular
receptor. Therefore, a second component of deposition which is reviewed in this study is the
method used to track and deplete a pollutant from the plume as it travels downwind and
deposits onto the surface. Among the methods which have been proposed to incorporate plume
depletion into the Gaussian plume framework are source depletion, surface depletion, modified
source depletion, and K-theory methods. Based on a review of the literature, the advantages
and disadvantages of these techniques are discussed in Section 3.
Most dry deposition algorithms require as input certain micrometeorological parameters
such as the surface friction velocity and Monin-Obukhov length. In order to provide a complete
system for modeling deposition, a revised ISC modeling system must include a technique for
estimating these parameters. A review of some previous studies which have tested various
methods for the computation of friction velocity and Monin-Obukhov length from routinely
available meteorological observations is discussed in Section 4.
The model evaluation protocol used in the present study, described in Section 5, is based on
the Cox-Tikvart protocol. It represents EPA's recommended approach to the evaluation of
model performance. Section 6 contains a summary of the observational particle deposition data
bases used in the model evaluation efforts (described in Section 7.1) as well as some gas
deposition data for possible use in future evaluation work. A summary of the recommendations
and conclusions reached to-date is contained in Section 8.
1-2
-------
Table 1-1
Core Particle Deposition Models and Other Algorithms Included in the Study
Type
Model
References
Deposition velocity models
(particles)
CARB/Sehmel model
ADOM/CALPUFF/CALGRID
models
UAM-V model
ISC reflection coefficient scheme
Sehmel and Hodgson (1978)
Sehmel (1980)
Pleim et al. (1984)
Scire et al. (1990)
Yamartino et al. (1992)
Gray et al. (1991)
Dumbauld et al. (1976)
Overcamp (1976)
Bowers et al. (1979)
Deposition velocity models
(gases)
RADM model
ADOM/CALPUFF/CALGRID
models
NOAA/ARL models
Wesely (1989)
Walcek et al. (1986)
Shieh et al. (1986)
Pleim et al. (1984)
Scire et al. (1990)
Yamartino et al. (1992)
Meyers and Baldocchi (1988)
Baldocchi et al. (1987)
Meyers (1987)
Hicks et al. (1987)
Plume depletion techniques
Source depletion
Surface depletion
K-theory approach
Modified source depletion
Chamberlain (1953)
Horst (1977)
Rao (1981)
Horst (1983)
Meteorological preprocessor
Holtslag and van Ulden
technique
Holtslag and van Ulden (1983)
van Ulden and Holtslag (1985)
Hanna and Chang (1992)
1-3
-------
2. Dry Deposition
Many complex processes are involved in the transfer and deposition of pollutants at the
surface. Sehmel (1980) compiled a list of the factors known to influence dry deposition rates
(see Table 2-1). Although it is not practical to incorporate all of the factors listed in the table
into a routinely-applied regulatory deposition model, algorithms with only modest data
requirements exist which parameterize many of the processes which typically dominate
deposition rates. Among the most important factors included in Table 2-2 are pollutant
properties such as the size and density of particles, and the solubility, reactivity, and diffusivity
of gases; surface characteristics such as the surface roughness, vegetation type, amount, and
physiological state; and atmospheric variables including the stability, friction velocity, and
turbulence intensity. Hicks (1982) noted the important differences controlling the deposition of
larger particles (e.g., gravitational settling, inertia! impaction) and those controlling gases (e.g.,
turbulence, molecular diffusion). Deposition of small particles is complicated by the fact that
because of their intermediate size, they may be influenced by the processes affecting gases and
large particles.
Due to the number and variability of the factors influencing dry deposition rates,
measured values of deposition velocities exhibit considerable variability. For example, SO2
deposition velocity observations summarized by Sehmel (1980) range over two orders of
magnitude (Figure 2-1). Particle deposition velocities (Slinn et al., 1978) show an even greater
variability. This is illustrated in Figure 2-2, which shows the deposition velocity as a function of
particle size as measured in two wind tunnel experiments. The deposition velocity ranges from
over 10 cm/s (for > 10 um diameter particles) to less than 0.01 cm/s (0.25 urn diameter
particle). Although it is not practical to include in the deposition model all of the variables listed
in Table 2-1, it is possible to parameterize many of the most important effects known to control
deposition rates in terms of routinely-available variables which describe the state of the
atmosphere, surface conditions, and pollutant properties.
The treatment of particle resuspension is not being treated in the current study. The
problem of particle resuspension is very complex, involving highly variable and site-specific
factors. Sehmel (1984) notes that experimentally-determined resuspension factors vary over
eight orders of magnitude (i.e., 10'10 to 10"2 m"1). Among the factors influencing resuspension
rates are soil/surface moisture, mean wind speed, gust intensity, particle/soil/surface properties,
and the nature, frequency, and magnitude of mechanical disturbances of the surface.
Development of a generalized particle resuspension model would involve a substantial effort
beyond the scope of the present study.
2-1
-------
Table 2-1
Factors Influencing Dry Deposition Rates
(from Sehmel, 1980)
Microraeteorological
Variables
Aerodynamic roughness
- Mass transfer
(a) Particles
(b) Gases
-Heat
- Momentum
Atmospheric stability
Diffusion, effect of:
-Canopy
Diurnal variation
- Fetch
Flow separation:
- Above canopy
- Below canopy
Friction velocity
Inversion layer
Pollutant concentration
Relative humidity
Seasonal variation
Solar radiation
Surfflcc hcittmff
Temperature
Terrain
- Uniform
- Nonuniform
Turbulence
Wind velocity
Zero-plane displacements
- Mass transfer
(a) Panicles
(b) Gases
-Heat
- Momentum
Depositing Material
Panicles
Agglomeration
Diameter
Density
Diffusion
-Brownian
Eddy equal to
() Panicle
(b) Momentum
(c)Hemt
- Rffcct of canopy on
Diffusiophoresu
Electrostatic effects
- Attraction
- Repulsion
Gravitational sealing
Hygroscopiary
Impacaon
Phvncftl nmiMTtiflc
fUjf»Uiml pIU£^IUC4
Resuspeniion
Shape
Size
Solubility
Thennophoresu
Gases
Chemical activity
Diffusion:
- Brownian
-Eddy
Parade pressure in equilibrium with
surface
Solubility
Surface Variable
i.-.-j-.^1-.i.--,
* Eodates
-Trichomes
-Pubescence
-Was
Bktie surfaces
Canopy growth:
- Dotmant
-Expanding
Senescent
Canopy structure!
- Aral density
-Bad:
-Bole
-Leaves
-Porosity
- Reproductive structure
-Soils
-Stem
-Type
Leaf-Vegetation:
- Boundary layer
. Change at high winds
- Flutter
- Stomatal resistance
Non-biotic surfaces
pH effects on:
-Reaction
-Solubility
PoOmant penetration and distribution in
canopy
Prior deposition loading
Water
2-2
-------
ST. LOUIS - 1975
ST. LOUIS - 1973
HEDGE
WATER, LAPSE ATM.
, MAX. RATE
GRASS, D STABILITY
ALFALFA
GRASS, NEUTRAL ATM.
CEMENT, MAX RATE
GRASS, LAPSE ATM
GRASS
WATER, NEUTRAL ATM.
GRASS
CEMENT, MAX RATE
FOREST
GRASS, MEDIUM
STUCCO, MAX RATE
GRASS, D STABILITY
CEMENT, MAX RATE
SNOW, LAPSE ATM
GRASS
GREAT BRITAIN
SOIL, CALCAREOUS
WATER, B STABILITY
SOIL, ADOBE CLAY-MAX
STUCCO, MAX RATE
WATER. B STABILITY
STUCCO, MAX RATE
WHEAT
GRASS, D STABILITY
GRASS, B STABILITY
SOIL, ADOBE CLAY-MAX
SOIL, SANDY LOAM-MAX
SOU, SANDY LOAM-MAX
FOREST. 17 m
WATER, D STABILITY
GRASS, SHORT
SNOW, NEUTRAL ATM
GRASS, STABLE ATM
WATER, FRESH
SNOW
ICE
SNOW, LAPSE ATM
SNOW, STABLE ATM.
ASPHALT, MAX. RATE
X
A
o-a
o
a
A
a
x
a
XX
a
A "MAXIMUM" RATES
D GRASS
X WATER
V SNOW
O OTHER
,!
1(T2 10'1 1 ' 10
DEPOSITION VELOCITY, cm/sec
Figure 2-1. Summary of observed SO2 deposition velocities (Sehmel, 1980).
2-3
-------
u*
(cm s
-1
10
o-
11
44
117
-40 -a 05
zo 0 l~10cm)
(cm)_ (m s'M
a 002
a 02
ai
2.2*
7.2*
13.8*
-8**
*SEHMELAND SUTTER (1974)
**MOLLER AND SHUMANN (1970)
e
<_>
>~ 1
o
o
^ . -1
(/>
o
Q.
10
10
-2
\
V
-2/3
PARTICLf DIAMETER, pm
Figure 2-2. Observed deposition velocities (vd) as a function of particle size for 1.5 g/cm3 density
particles. Measured by Sehmel and Sutter (1974) and Moller and Schumann (1970).
Figure from Slinn et ai. (1978). The gravitational settling velocity (Vg) is also shown.
2-4
-------
2.1 Dry Deposition of Participate Matter
Many models of dry deposition express the deposition velocity as the inverse of a sum of
"resistances" plus, for particles, gravitational settling velocity terms. The resistances represent the
opposition to transport of the depositing material from a reference height through the turbulent
atmospheric surface layer, and through a quasi-laminar layer just above the surface to the surface
itself. The major processes are briefly described below.
Gravitational Settling. The gravitational settling velocity is a function of the particle diameter,
shape, and density. Figure 2-3 shows the gravitational settling velocity (Vg) as a function of particle
size for several values of the particle density (p) for solid spherical particles. In many cases, effective
or aerodynamic particle diameters are reported which express the size and shape of the particle as an
equivalent diameter for a spherical particle of unit density, which simplifies the input to the deposition
model. Note that the gravitational settling velocity represents a lower limit to the deposition velocity.
It can be seen that for larger particles, in the range of 20-40 um diameter and higher, the deposition
velocity approaches the gravitational settling velocity, which indicates that the rate of deposition is
dominated by the gravitational settling mechanism. The gravitational settling velocity decreases with
decreasing particle size. However, for particles smaller than about 20 um diameter, the deposition
velocity curve shows larger and larger deviations from the gravitational settling curve as the particle
size decreases. This is due to the effect of other mechanisms, discussed below, in enhancing the
deposition rate of smaller particles above that predicted by gravitational settling alone.
Atmospheric Diffusion. The rate of deposition can sometimes be limited by the transfer of
pollutant material to the vicinity of the surface by atmospheric turbulence. For example, atmospheric
turbulence-limited deposition situations would typically occur during very stable conditions for an
elevated plume of material composed of small-sized particles with small gravitational settling. In the
lowest layer of the atmosphere, the aerodynamic resistance is used to parameterize the rate of mixing
in terms of the wind speed, atmospheric stability, and surface roughness length. The aerodynamic
resistance generally decreases (i.e., the deposition velocity will increase) with increasing wind speed
and/or surface roughness.
Quasi-laminar Layer. Over smooth surfaces, a thin non-turbulent sublayer develops that can
be a significant obstacle to the transfer of the pollutant onto the surface. For rough, real-world
surfaces, this sublayer is constantly changing and is likely to be intermittently turbulent. For this
reason, Hicks (1982) calls this layer the "quasi-laminar" layer. It is also known as the deposition layer.
Small particles (< 0.05 um diameter) are transported through the quasi-laminar layer primarily by
Brownian diffusion. However, Brownian diffusion becomes less efficient as the particle size increases.
Particles in the 1-20 um diameter range tend to penetrate the quasi-laminar layer by inertial
impaction. Since particles larger than 20 jim diameter are less efficiently captured, the inertial
impaction mechanism is most effective in the 1-20 \im diameter size range. Because particles in the
0.1-1.0 um diameter size range are not efficiently transported across the quasi-laminar layer by either
Brownian diffusion or inertial impaction, particles of this size have the lowest deposition velocities.
2-5
-------
10
*
1
O
a.
10
,-1
10
.-2
10
,-3
'""i p. 11.5 / '/'
STABLE ATMOSPHERE WITH
ROUGHNESS HEIGHT, cm
! . I V I / I I I/ I 1 I
ill I t r i
10
i 10
PARTICLE DIAMETER, urn
Figure 2-3. Predicted deposition velocities for u. = 100 cm/s and particle densities of 1, 4, and
11.5 g/cm3. Also shown is the gravitational settling velocity (vj. Figure from
Sehmel (1980).
2-6
-------
Based on the discussion above, it is considered essential that the final generalized model
selected be able to parameterize, at a minimum, the effects of gravitational settling, which is a
dominate effect for large particles (> ~ 20 jim diameter), inertial impaction (dominates in the
size range 1.0 to 20 \im diameter), and Brownian motion (important for small particles less than
about 0.1 um diameter).
2.1.1 ISC Method
The basis for the present ISC deposition algorithm is found in Dumbauld et al. (1976)
and Overcamp (1976). In the ISC approach, the particles are allowed to move toward the
ground by the combined processes of atmospheric turbulence and gravitational settling. At the
surface, a portion of the plume determined by a user-specified reflection coefficient, Yn» is
assumed to be reflected from the surface and the remainder (1 - yj is assumed to be retained
by the surface. The reflection coefficient scheme is illustrated in Figure 2-4 for yn = 0.0 (no
reflection), 0.5 (partial reflection), and 1.0 (full reflection).
The effect of gravitational settling is assumed to result in a tilted plume with an angle, 9, to
the horizontal given by:
8 = tan1 (vg/u) (2-1)
where vg is the gravitational settling velocity and u is the wind speed.
The concentration in ISC is given by:
KQD.V
X = exp
where K is an units scaling factor,
Q is the pollutant emission rate,
Dt is the decay term,
V is the vertical term,
oy is the standard deviation .of the concentration distribution in the
horizontal,
oz is the standard deviation of the concentration distribution in the
vertical,
2-7
(2-2)
-------
(9
{H-vtx/u)
5O% REFLECTION
(Y-Q.S)
ZERO REFLECTION
-------
u is the wind speed, and
y is the crosswind distance from the plume centerline to the receptor.
The user must subdivide the pollutant size distribution into N particle settling categories.
A maximum of 20 particle settling categories is allowed. The vertical term for a ground-level
receptor involves a summation over all N particle categories.
= E
(2-3)
where
i-l
(2-4)
= Yn' exp
- 0.5
(2-5)
A2 = Yn'+1 6XP
- 0.5
(2-6)
A3 = Yn' exP
- 0.5
(2-7)
2-9
-------
_ i-1
exp
(2iHm-H+H\2
- 0.5 = 2 ^
(2-8)
and, 4>n is the mass fraction of the pollutant emitted in the nth particle settling category,
yn is the reflection coefficient for particles in the nth particle settling category,
Hy is the height lost at the downwind distance x due to the gravitational settling
[Hv = (v *)M
vg is the gravitational settling velocity,
Hm is the mixing height, and,
He is the effective plume height (stack height + plume rise).
The ISC manual suggests the use of the following equation (McDonald, 1960) for
calculating the gravitational settling velocity, vff in cm/s.
(2-9)
where p is the particle density (g/cm3),
g is the acceleration due to gravity,
dp is the particle diameter (cm), and
\i is the absolute viscosity of air (ISC manual suggests p. ~ 1.83 x 10"4 g/cm/s).
The total deposition, D, is summed over all N particle categories, and is given by:
(2-10)
(2-11)
2-10
-------
where the vertical term for deposition, V, is defined as:
V' = [ b Ht + (1 - b)Hv] exp
-0.5
i-l
- H
' (2M. - H<] - (1 '
1 [ b (2iHm + H.) + (1 - E)#v] exp
exp
tUHm - Ht . HA
( °< J
J2iH_
(2-12)
I, ~ Hv]2
where Qt is the total pollutant mass emitted during the time period over which the deposition is
_ calculated, and
b is the average value of the exponent b of the oz equation (oz = ,axb) for the interval
between the source and the downwind distance x.
The ISC deposition method requires that the user specify the following:
(1) mass fraction of pollutant emitted in each particle settling category
(2) reflective coefficients for each particle settling category.
The reflection coefficient is given by Dumbauld et al. (1976) as a function of the particle
settling velocity (see Figure 2-5). The settling velocity, which depends on the particle diameter
and particle density, must be calculated by the user for each size category and used to determine
the reflection coefficient, which is then entered into the model. The reflection coefficients are
assumed to be spatially and temporally constant in ISC.
The reflection coefficient approach used to model dry deposition in the ISC model
suffers several drawbacks for applications to small particles (< 20 jim diameter) and gases (e.g.,
see Scire and Wojichowski, 1987). The ISC deposition algorithm was developed for use only
with large particles dominated by gravitational settling effects. The empirical reflection
coefficients used in ISC are based on data collected and analyzed by Dumbauld et al. (1976).
Their experiment consisted of several aircraft releases of two spray carriers: Duphar and No. 2
2-11
-------
0.30
i t t » i r
0.2 0.4 0.6 0.8
REFLECTION COEFFICIENT 7||
1.0
Figure 2-5. Reflection coefficient as a function of the gravitational settling velocity (from
Dumbauld et al., 1976).
2-12
-------
fuel oil. The size distributions of the particles were heavily weighted with large particles. The
mass mean diameter of the particle distributions were 66 um (for the Duphar releases) and
55 um diameter (fuel oil releases). Less than 1% of the mass of the Duphar was in particles
with diameters < 30 um. For the fuel oil distribution, less than 1% of the mass was in particles
smaller than 20 um diameter. Particles of this large size are nearly completely controlled by
gravitational settling, which, because it is a characteristic of the particle and does not depend on
factors such as meteorological conditions and surface characteristics, is compatible with the use
of spatially and temporally constant reflection coefficients. However, the deposition of small
particles is controlled by other physical processes such as inertial impaction and Brownian
motion, which are affected by both meteorological and surface conditions. These effects are not
easily parameterized within the framework of a reflection coefficient model.
For small particles, the ISC reflection coefficient curve (Figure 2-5) was derived based
on an extrapolation of the gravitational settling-dominated data for larger particles (Dumbauld
et al, 1976). The extrapolation led to the assumption that particles with settling velocities less
than 0.1 cm/s (approximately 5.7 um diameter for unit density particles) are completely
reflected (i.e., experience no deposition). However, this assumption is not consistent with
measurement studies of small particles, as discussed previously, which show significant
deposition velocities for particles at and below 5.7 ^m diameter (e.g., see Figures 2-2 and 2-3).
It was concluded that the ISC scheme could significantly underestimate small particle deposition
because several important physical processes are not parameterized in its approach. Therefore,
the ISC method is not considered appropriate for small particles. However, it was tested and
intercompared with the other approaches to serve as a reference point for the performance of
the other models.
2.1.2 CARB Model
Sehmel and Hodgson (1978) and Sehmel (1980) proposed a model for predicting
deposition velocities of particles above smooth surfaces. The basis of the model is a set of wind
tunnel observations of deposition for monodispersed particles to surfaces such as gravel,
artificial grass, brass shim stock, and water. The model consists of empirical equations for
transfer resistances derived from a least-squared empirical fit of deposition velocity as a function
of particle size, density, surface roughness, and friction velocity. The equations were converted
into a computer code by the California Air Resources Board (CARB) and is widely known as
the CARB model. It is also used in the Fugitive Dust Model (FDM) (Winges, 1990).
In the CARB model, integrated resistances to mass transfer are computed within two
layers. The first layer extends from a reference height of one meter above the surface down to
2-13
-------
one centimeter above the surface. In this layer, atmospheric turbulence dominates mass
transfer. Eddy diffusivities are used to describe the transfer rate. The second layer is the
deposition surface layer within one centimeter of the surface. The integrated resistance within
the deposition surface layer is derived from a statistical fit of the wind tunnel particle deposition
data. Sehmel and Hodgson express the deposition velocity as:
v,
1. - exp-v^ + /3y«.]
where, vd is the deposition velocity (cm/s),
vg is the gravitational settling velocity (cm/s),
Iu is the atmospheric diffusional resistance (dimensionless),
13 is the surface resistance integral (dimensionless), and,
u. is the surface friction velocity (cm/s).
The gravitational settling velocity (cm/s) in the CARB model is calculated as:
v*= \i * "V" (p" pA«x^ -^
( ^; ist
18u
(2-14)
where dp is the particle diameter (um),
p is the particle density (g/cm3),
PAIR is tne density of air (CARB assumed p^ = 1.2 x 10"3 g/cm3),
T is the air temperature (°K),
H is the absolute viscosity of air (CARB used \i = 1.78 x 10"4 g/cm/s), and
the constants q and c^ were assigned values of 9.73 X 10"3 and 1.0 x 10"8,
respectively.
The atmospheric diffusion resistance used by Sehmel and Hodgson (1978) is based on
the flux profile relationships of Businger et al. (1971). For neutral or stable conditions,
Iu= [Inz./z,) + 4.7(2l - z,)/L]/k (2-15)
where, zl is the upper limit of the atmospheric diffusional resistance integral (i.e., 100 cm),
Zj is the lower limit of integral (i.e., 1 cm),
L is the Monin-Obukhov length (cm), and,
2-14
-------
k is the von Karman constant (Sehmel used a value of 0.35).
For unstable conditions, the atmospheric diffusional resistance integral is:
- 1.)]}
(2-16)
2.[tan-1(Azl) -
(1. - 15.Zl/L)°* (2-17)
(1. - 15.Z2/L)0-25 (2-18)
The surface resistance integral is an empirical relationship based on the wind tunnel
observations. For particles with a diameter £ 0.01 jim,
73 = exp{-378.051 + 16.498 ln(5c) + ln(f*)[-11.818 - 0.2863
0.3226 ln(> )]
(2-21)
where T is the temperature (deg. K).
2-15
-------
The data requirements of the model are relatively simple (particle size, density, surface
roughness, and routine meteorological parameters to compute the friction velocity and
Monin-Obukhov length).
The plots of deposition velocity in Sehmel (1980) (e.g., see Figure 2-3) show a
reasonable variation of deposition velocity as a function of model parameters (i.e., density, size,
surface roughness, friction velocity). The predicted deposition velocity is close to the
gravitational settling velocity for large particles (e.g., greater than about 20 jim diameter), and
decreases with decreasing particle size to about 0.1-1.0 um, where it reaches a minimum. The
deposition velocity then increases with decreasing particle size for smaller sized particles. This
behavior is consistent with the importance of Brownian motion in enhancing deposition rates for
very small particles. The CARB scheme produces increased deposition rates for increased
particle density, surface roughness length, and friction velocity, which is expected based on
physical considerations.
The main concern about the CARB model deals with the generality of the highly
empirical relationship for the surface resistance integral (Eqn. 2-19). I3 is based on wind tunnel
data for relatively smooth surfaces conducted under a limited range of conditions. For example,
in order to avoid extrapolation, the CARB implementation of the model does not allow the
surface roughness length used in the algorithm to exceed 10 cm, even though most real world
surfaces have significantly greater roughness lengths (e.g., see Table 2-2). In addition, sensitivity
testing of the model has shown it exhibits some non-physical behavior when the inputs are
varied beyond the range of conditions tested in the wind tunnel. For example, the CARB model
showed a very strong sensitivity to temperature which is not exhibited by other deposition
models. For particles in the 0.1-1.0 um diameter size, a change on nearly an order of magnitude
in deposition velocity was predicted for a temperature change from 0°F to 100°F (see Figure
2-6). This behavior is probably an artifact of the regression equations used to fit the surface
resistance integral to the wind tunnel data. The original wind tunnel tests were performed at a
constant temperature, and, as a result, application to a realistic range of atmospheric conditions
involves extrapolation outside the range on which the model was developed.
A second problem noted from the sensitivity testing is that the deposition velocities show
a kink in the curve at about 0.03 jim particle diameter. For particles smaller than about 0.1 \im
diameter, the deposition velocity increases with decreasing particle diameter. However, the
CARB model shows this trend only to about 0.03 jim diameter, beyond which the deposition
velocity is predicted to decrease or level off with decreasing particle diameter (see Figure 2-7).
The kink in the curves appears to be another artifact of the regression equations which occurs
2-16
-------
Table 2-2
Typical Surface Roughness Lengths for Various Land Use Types
(From Hjelmfelt, 1982)
Typical
Land Use Roughness Length (cm)
Urban - Commercial/Industrial 200
Common residential - single family dwellings 20
Compact residential - multi-family dwelling 50
Metropolitan natural (parks, golf courses) 15
Agricultural - rural 20
Semi-rural 20
Undeveloped, wasteland 5
Forest 100
Bottomland agricultural 15
2-17
-------
10J
K 10
,-2
H
1-4
CO
10
-3
10
-4
SENSITIVITY RUN FOR TEMPERATURE
PARTICLE DENSITY = 1 g/cm3
0.01 0.10 1.00 10.00 100.00
PARTICLE DIAMETER (MICRONS)
SENSITIVITY RUN FOR TEMPERATURE
PARTICLE DENSITY = 4 g/cm3
n
10J
10
H -1
810
§
-2
10
-3
10
-4
Figure 2-6.
0.01 0.10 1.00 10.00 100.00
PARTICLE DIAMETER (MICRONS)
Deposition velocity as a function of particle diameter as predicted by the
Sehmel/CARB model for three different values of ambient temperature (0°, 60°,
100° F).
2-18
-------
io F
10
-2
10
-1
10 10
PARTICLE DIAMETER (MICRONS)
10
Figure 2-7. Deposition velocity as a function of particle diameter as predicted by the
Sehmel/CARB model for surface roughness lengths of 0.001, 0.01, 3, and 10 cm.
2-19
-------
as a result of extrapolation outside the range of conditions on which the model is based.
In addition to evaluating the basic CARB model, several modifications to the model
were made in an attempt to address some of the concerns raised above. In the model
evaluation (Section 7), the basic CARB model described by Eqns. (2-13) to (2-21) is designated
CARB 1. Three variations of the model were developed and evaluated (CARB 0, CARB 2, and
CARB 3). The modifications to each version of the model are described below.
CARB 0 - same as the CARB 1 except the portion of the code which limits the value of z0
to 10 cm was eliminated. Thus, the CARB 0 revision uses actual surface
roughness lengths which are allowed to be greater than 10 cm.
CARB 2 - same as CARB 1, except in Eqn. (2-21), a constant temperature of 75°F is
always used rather than the actual air temperature. This change was intended to
eliminate the temperature sensitivities of the model noted above.
CARB 3 - contains changes to eliminate the temperature dependency (as in CARB 2) and
an adjustment to the surface resistance integral, I3. The value of I3 used in CARB
3 is that given by Eqn. (2-19) divided by (1.0 + LAI), where LAI is the leaf area
index. In addition, a fixed reference length of 0.5 cm is used in CARB 3 in the
calculation of the LAI-adjusted I3 instead of z0.
The Sehmel and Hodgson (1978) empirical relations are based on simple surfaces. For
compound surfaces which are aerodynamically independent, it seems reasonable to expect the
deposition velocity to be increased by the additional surface area. The LAI adjustment in
CARB 3 is made to test this assumption.
Also, the roughness length in the surface integral resistance, I3, differs from the
traditional roughness length defined by other researchers such as Plate and Quraishi (1965).
Rather, it is related to the thickness of the laminar sublayer, which is proportional to the ratio
of the molecular diffusivity to the surface friction velocity. The constant of proportionality for
an established sublayer is approximately 20. Thus, for a friction velocity of 0.1 m/s and a
diffusivity of 2 x 10"s m2/s, the resulting sublayer depth is .4 cm, which is close to the largest
roughness length used by Sehmel and Hodgson (1978). Thus, for the relatively smooth surfaces
tested by Sehmel, z0 may be a good approximation of the sublayer depth length scales, but for
the LAI-adjusted I3, it may be more appropriate to use a constant reference length. A constant
value of 0.5 cm is tested in CARB 3.
2-20
-------
2.1.3 ADOM Model - Particles
The Acid Deposition and Oxidant Model (ADOM) contains a dry deposition algorithm
(Pleim et al., 1984) which has served as the basis for the deposition scheme in several other
models, including CALPUFF (Scire et al., 1990) and CALGRID (Yamartino et al., 1992).
The deposition flux, F, is calculated as the product of the concentration, x> at a reference
height and a deposition velocity, vd.
F = X ' vd ' (2-22)
The general approach used in the resistance methods such as in the ADOM model is to
include explicit parameterizations of the effects of Brownian motion, inertial impaction, and
gravitational settling. The deposition velocity is written as the inverse of a sum of resistances to
pollutant transfer through various layers, plus gravitational settling terms (Slinn and Slinn, 1980;
Pleim et al., 1984).
v = + v
r + r. + r rv *
a a a a g
(2-23)
where, vd is the deposition velocity (cm/s),
vg is the gravitational settling velocity (cm/s),
ra is the aerodynamic resistance (s/cm), and,
rd is the deposition layer resistance (s/cm).
Note that for large settling velocities, the deposition velocity approaches the settling
velocity (vd - vj, whereas, for small settling velocities, vd tends to be dominated by the ra and rd
resistance terms.
The lowest few meters of the atmosphere can be divided into two layers: a fully turbulent
region where vertical fluxes are nearly constant, and the thin quasi-laminar sublayer. The
resistance to transport through the turbulent, constant flux layer is the aerodynamic resistance.
It is usually assumed that the eddy diffusivity for mass transfer within this layer is similar to that
for heat. The atmospheric resistance formulation used in ADOM is based on Wesely and Hicks
(1977).
2-21
-------
'TIT
where, i|rH is a stability adjustment factor,
u. is the surface friction velocity (cm/s),
k is the von Karman constant (0.4),
ZR is reference height (m) (- 10 m), and,
z0 is the surface roughness length (m).
(2-24)
-5 zJL 0 < zIL < 1
0 ' zIL = 0
(2-25)
The approach used by Pleim et al. (1984) to parameterize the deposition layer resistance
terms is:
(2-26)
where, Sc is the Schmidt number (Sc = u/D) (dimensionless),
u is the viscosity of air (- 0.15 cm2/s),
D is the Brownian diffusivity (cm2/s) of the pollutant in air,
St is the Stokes number [St = (Vg/g)(u,2 /u)] (dimensionless), and,
g is the acceleration due to gravity (981 cm/s2).
The gravitational settling velocity (cm/s) in the ADOM model is calculated as:
v _ dl
(2-27)
where p is the particle density (g/cm3),
PAIR ^s tne air density (- 1.2 x 10"3 g/cm3),
dp is the particle diameter ((im),
2-22
-------
|i is the absolute viscosity of air (- 1.81 x 10"* g/cm/s),
Cj is air units conversion constant (1 x 10"* cm2/urn2), and
So? is the slip correction factor, which is computed as:
SCF =
10-4 dp
(2-28)
and, x2, a.v a2, a3 are constants with values of 6.5 x 10"*, 1.257, 0.4, and 0.55 x 10"*, respectively.
The diffusivity of the pollutant (in cm/s) is computed from the following relationship.
D = 1.38 x 10'16 T]Scp
(2-29)
where T is the air temperature (°K).
The first term of Eqn. (2-26), involving the Schmidt number, parameterizes the effects of
Brownian motion. This term controls the deposition rate for small particles. The second term,
involving the Stokes number, is a measure of the importance of inertial impaction, which tends
to dominate for intermediate-sized particles in the 2-20 nm diameter size range.
The data requirements of the ADOM-type deposition model are identical to those of the
CARB model. Particle size, particle density, surface roughness length, friction velocity, and
Monin-Obukhov length are input parameters used by the model.
The ADOM approaches produce patterns of deposition velocities as a function of
particle size, density, etc. similar to those of the CARB model. However, the ADOM algorithm
tends to predict somewhat higher deposition velocities in the 5-15 um diameter size range than
the CARB model and lower values in the 0.1-5.0 (im diameter range, although the general shape
of the deposition velocity curves are similar. The minimum deposition velocity tends to occur
closer to 1.0 \im than 0.1 \im diameter.
The parameterizations of the resistance models for Brownian motion and inertial
impaction effects involve empirical factors derived from field studies and wind tunnel
experiments. The sensitivity analyses with the ADOM model showed no unusual response to
temperature variations over the range from 0° F to 100°F. The resistance models all showed a
2-23
-------
steady increase of deposition velocity with decreasing particle diameter in the small particle
range (< 0.1 |im diameter).
The basic (core) ADOM model defined by Eqns. (2-23) to (2-29) is designated as
ADOM 1. Two sets of changes to the ADOM formulation, described below, were also tested.
The -3/St exponent in the inertial term of Eqn. (2-26) may represent too sharp a cutoff
as pointed out by Slinn (1982). Slinn suggests that a better function for the inertial impaction
term, E^, for canopies is:
(2-30)
where St is the Stokes parameter. In addition, the power law dependence on the Schmidt
number for diffusive interception is likely to be a function of the surface type, with established
laminar boundary layers over smooth surface having an exponent closer to -0.5, while for more
complex surfaces, the exponent 'is likely to be nearer -0.7.
Therefore, in the version designated ADOM 2, the deposition layer resistance was
computed as:
St
tfj «
(2-31)
instead of using Eqn. (2-26), with n = -0.5 for z0 i 10 cm and n = -0.7 for z0 > 10 cm.
A second change in ADOM 2 is an allowance for a small adjustment to the deposition
rates to account for possible phoretic effects. Some examples of phoretic effects (Hicks, 1982)
are:
THERMOPHORESIS: Particles close to a hot surface experience a force directed
away from the surface because, on the average, the air
molecules impacting on the side of the particle facing the
surface are hotter and more energetic.
DIFFUSIOPHORESIS: Close to an evaporating surface, a particle is more likely to
be impacted by water molecules on the side of the particle
2-24
-------
facing the surface. Since the water molecules have a lower
molecular weight than the average air molecule, there is a
net force toward the surface, which results in a small
enhancement of the deposition velocity of the particle.
A second effect is that the impaction of new water vapor
molecules at an evaporating surface displaces a certain
volume of air. For example, 18 g of water vapor
evaporating from 1 m2 will displace 22.4 liters of air at
standard temperature and pressure (STP) conditions
(Hicks, 1982). This effect is called Stefan flow. The Stefan
flow effect tends to reduce deposition fluxes from an
evaporating surface. Conversely, deposition fluxes to a
surface experiencing condensation will be enhanced.
ELECTROPHORESIS: Attractive electrical forces have the potential to assist the
transport of small particles through the quasi-laminar
deposition layer, and thus could increase the deposition
velocity in situations with high local field strengths.
However, Hicks (1982) suggests this effect is likely to be
small in most natural circumstances.
Phoretic and Stefan flow effects are generally small. However, for particles in the range
of 0.1 - 1.0 urn diameter, which have low deposition velocities, these effects may not always be
negligible. Therefore, the ability to specify a phoretic term to the deposition velocity is added
ta ADOM 2 (i.e., vd' = vd + vd(phor), where vd' is the modified deposition velocity and vd(phor) is
the phoretic term). Although the magnitude and sign of vd(phor) will vary, a small, constant value
of 4- 0.01 cm/s is used in ADOM 2 for testing purposes.
Another version of the model, ADOM 3, was tested which included the ADOM 2
changes, and a leaf area index (LAI) adjustment to rd. In ADOM 3, rd is that given by Eqn.
(2-31) divided by (1.0 + LAI).
2.1.4 UAM-V Model
The deposition formulation variable grid version of the Urban Airshed Model (UAM-V)
(Gray et al., 1991) is similar to that in ADOM. The deposition velocity is expressed in terms of
an aerodynamic resistance (rj, deposition layer resistance (rd), and a gravitational settling
2-25
-------
velocity (Vg) according to Eqn. (2-23). The formulation for r. in UAM-V is identical to that in
ADOM (Eqn. 2-24). The form of the stability adjustment factor, i|rH, is given by Eqn. (2-25),
except the coefficient for positive L is
-4.7 in UAM-V instead of -5.
The main difference is in the formulation of the inertia! impaction term of the
deposition layer resistance.
(2-32)
where the parameter Q is computed each hour to force the Sc and St terms to be equal at a
critical diameter of 0.3 urn.
The form of the gravitational settling velocity equation in UAM-V is identical to that in
ADOM (i.e., Eqns. 2-27 and 2-28). The values of the constants in the v equations are slightly
different. The values of x2, \i, and pMR used in UAM-V are 6.53 x 10"6, 1.83 x 10"4 g/cm/s, and
1.0 x 10"3 g/cm3, respectively.
The data input requirements of the UAM-V dry deposition model (particle size, particle
diameter, roughness length, friction velocity, and Monin-Obukhov length) are the same as those
required by the CARS and ADOM models.
Tfie UAM-V dry deposition algorithm described above is designated as UAM 1 in the
model evaluation tests. A second version, called UAM 2, was also evaluated. UAM 2 was
modified to include a LAI adjustment to the deposition layer resistance (i.e., rd in UAM 2 is
that given by Eqn. (2-32) divided by (1.0 + LAI)).
2.2 Dry Deposition of Gaseous Pollutants
Many of the available models for predicting dry deposition of gaseous pollutants have a
similar structure. The deposition velocity is expressed as the inverse of a sum of resistances.
(2-33)
where ra is the aerodynamic resistance (s/m),
rd is the deposition layer or quasi-laminar layer resistance (s/m), and,
rc is the bulk canopy resistance (s/m).
2-26
-------
The gravitational settling velocity is not a factor for gases, and therefore does not appear
in Eqn. (2-33). The aerodynamic resistance for gases is the same as discussed previously for
particles, see Eqn. (2-24). It is determined from meteorological and surface parameters and
does not depend on the pollutants characteristics. The deposition layer resistance involves
transport by molecular diffusion through the thin quasi-laminar sublayer which intermittently
exists just above the surface. The deposition layer resistance is a function of the properties of
the pollutant (i.e., molecular diffusivity in air) and meteorological conditions. The canopy
resistance involves the physical capture or chemical reaction of the pollutant within the
vegetative canopy or at the surface. The canopy resistance, r0 is often the controlling resistance
determining deposition flux. Therefore, considerable effort is devoted to estimating r: in many
of the deposition models.
There are many models based on this simple resistance concept. Three widely-used
techniques appropriate for deposition of gases have been selected for future further evaluation.
In some cases, various components of the algorithms are very similar or identical to each other
(e.g., the formulation of rj. However, other important variations exist, especially in the
parameterization of r0 which will be intercompared, tested and evaluated. Because the goal of
this study is to produce an algorithm which can be used on a routine basis for regulatory
applications, top consideration has been given to simpler parameterizations which capture the
most significant features of the physical system while requiring modest, readily-available data as
input
The major features of the algorithms selected for the current study are described below.
The original references provide a detailed description of the algorithms.
2.2.1 RADM Model
The dry deposition algorithm used in the Regional Acid Deposition Model (RADM)
meets all of the model selection criteria listed in Section 1. It has undergone considerable
testing and refinement over the past five years. The RADM deposition scheme is described in a
series of reports and papers, including Sheih et aL, 1986; Walcek et al, 1986; and Weseiy, 1989.
It has also been incorporated into the latest version of the UAM-V model (Gray et ai., 1991).
The early version of the RADM deposition model used fairly standard formulations for
ra and rd, and look-up tables for rc based on land use and season. However, the difficulty with
the look-up table approach is that detailed rc tables must be developed for each pollutant of
interest, which can be difficult for many pollutants. The latest version of the model computes rc
2-27
-------
based on the bulk surface resistance along several parallel pathways of mass transfer (Weseiy,
1989). The equation for rc is:
-I
(2-34)
where, r, is the bulk canopy stomatal resistance,
r,,, is the leaf mesophyU resistance,
r,u is the leaf cuticle resistance,
r#. is the within-canopy aerodynamic resistance to the lower canopy,
rd is the exposed surface resistance in the lower canopy,
rx is the within-canopy aerodynamic resistance to the ground, and,
r9 is the ground resistance.
The data requirements of this more sophisticated module include chemical parameters
such as solubility, reactivity, and diffusivity. Table 2-3 contains values of these parameters for
several gaseous species. Values for other toxic pollutants of concern can be obtained in the
literature or estimated based on the properties of the pollutant. Other parameters required
include meteorological factors such as solar radiation, friction velocity and Monin-Obukhov
length (which can be estimated from routinely-available data), and geophysical data such as
surface roughness and land use type. The data requirements of the RADM scheme are
summarized in Table 2-4.
The stomatal resistance, rs is computed as a function of air temperature and solar
radiation. The resistance, r.,,. is specified by Weseiy (1989) as a function of the land use type.
The resistance, r^ is often small It can be conservatively estimated based on solar radiation.
The other resistances in Eqn. (2-34) are computed by the modei using values of solubility and
reactivity of the pollutant as shown in Table 2-3.
222 ADOM Model - Gases
The dry deposition modules in the Acid Deposition and Oxidant modei (ADOM) (Pleim
et al, 1984), CALPUFF (Scire et ai., 1990), and CALGRID (Yamartino et al., 1992) models use
the same basic approach, although the CALPUFF/CALGRID models contain a few enhanced
features. The parameterization of the aerodynamic resistance is the same as for particles (Eqn.
(2-24)). The deposition layer resistances is parameterized in terms of the Schmidt number, Sc.
rd = dt Scr'3/ (k u.) (2-35)
2-28
-------
Table 2-3
Values of Chemical Input Parameters Required by the RADM Deposition Model
(From Wesely, 1989)
Gaseous species
Sulfur dioxide
Ozone
Nitrogen dioxide
Nitric oxide
Nitric acid vapor
Hydrogen peroxide
Acetaldehyde
Formaldehyde
Methyl hydroperoxide
Peroxyacetic acid
Formic acid
Ammonia
Peroxyacetyl nitrate
Nitrous acid
Symbol
SO2
03
NO2
NO
HN03
H202
ALD
HCHO
OP
PAA
ORA
NH3
PAN
HNO2
1.9
1.6
1.6 "
1.3
1.9
1.4
1.6
1.3
1.6
2.0
1.6
1.0
2.6
1.6
Solubility(2)
IxlO5
0.01
0.01
2 x lO'3
1 x 1014
Ix 105
15
6xl03
240
540
4x 106
2xl04
3.6
IxlO5
Reactivity(3)
0
1
0.1
0
0
1
0
0
0.1
0.1
0
0
0.1
0.1
(1)
(2)
(3)
Ratio of molecular diffusivity of water to that of the pollutant
Effective Henry's law coefficient (m/atm)
Pollutant reactivity parameter as defined by Wesely (1989)
(0 = nonreactive, 0.1 = slightly reactive, 1 = highly reactive)
2-29
-------
Table 2-4
Summary of Input Requirements of RADM Deposition Module
for Gases
Pollutant Characteristics:
- solubility (effective Henry's law coefficient)
- molecular diffusivity
- reactivity class parameter (nonreactive, slightly reactive, or highly reactive)
Geophysical Data:
- surface roughness length
- land use category
Meteorological Variables:
- friction velocity
- Monin-Obukhov length
- air temperature
- solar radiation
2-30
-------
where d: is an empirical constant (~ 2). The Schmidt number is defined as the ratio of the
viscosity of air (~ 0.15 cm2/s) to the molecular diffusivity of the pollutant in air.
The canopy resistance is computed by considering three pathways for uptake of the
pollutant within the vegetation or at the surface:
(1) transfer through the stomatal pore and dissolution or reaction in the leaf
interior,
(2) reaction with or transfer through the leaf cuticle, and,
(3) transfer into the ground or water surface.
This is expressed as three resistances in parallel.
rc = [LAI/rf +LAI/rcut + 1/rJ'1 (2-36)
where LAI is the leaf area index (ratio of leaf surface area divided by ground surface area).
The LAI can be estimated from land use type and season (e.g., see Table 2-5),
rf is the internal foliage resistance,
r,.,,, is the cuticle resistance, and,
rg is the ground resistance.
SO2 is used as a reference species because considerable data are available for it. The
model inputs for other species are computed by scaling the reference values for SO2 by the
relative adjustment factors for the other compounds. For example, the cuticle resistance for a
pollutant twice as reactive as SO2 is assumed to be one half the reference SO2 cuticle resistance.
Typical input values for many chemical species are described in the ADOM User's Guide.
The data requirements of the ADOM-type deposition models are very similar to those of
the advanced RADM deposition model. Table 2-6 lists the required input parameters. All of
the geophysical parameters can be derived from a classification of the land use in the area (e.g.,
urban, agriculture, forest, etc.). The meteorological variables are either observed routinely (e.g.,
temperature) or can be computed from routinely-available data. The pollutant characteristics
can generally be obtained from the literature.
2.2.3 NOAA/ARL Models .
A hierarchy of dry deposition models developed by NOAA/ARL are described in papers
by Meyers and Baldocchi (1988), Baldocchi et al. (1987), Meyers (1987), Hicks et al. (1987), and
2-31
-------
Table 2-5
Leaf Area Index Values as a Function of Land Use Type and Season
(From Scire et ai, 1986)
Vegetative Growing Season"
Land Use Type I H ffl IV
Water
Deciduous forest
Coniferous forest
Swamp
Cultivated
Grassland
Urban
Desert shrubland
0.00
6.00
7.00
2.00
3.00
2.00
0.30
0.10
0.00
6.00
7.00
1.50
1.00
1.50
0.20
0.10
0.00
0.50
7.00
1.00
0.20
1.00
0.05
0.10
0.00
0.40
6.00
0.50
0.01
0.50
0.01
0.05
* Definitions of season categories:
I = peak growing season
n = early growing season
HI = non-growing season without snow
IV = non-growing season with snow
2-32
-------
Table 2-6
Summary of Input Requirements of ADOM-type Deposition Modules
for Gases
Pollutant Characteristics:
- solubility (Henry's law coefficient)
- molecular diffusivity
- pollutant reactivity
-aqueous phase dissociation constant
Geophysical Data:
- land use category
- surface roughness length (derived from land category)
- leaf area index (derived from land use category)
Meteorological Variables:
- friction velocity
- Monin-Obukhov length
- air temperature
- solar radiation
2-33
-------
others. There are four basic models of increasing complexity and sophistication (as well as
increasing data requirements and computational cost) referred to as model I through model IV.
The simpler two models (I and II) were reviewed in the current project. Model I is known as the
"big leaf1 model because it treats the plant canopy as a single surface partitioned into shaded
and sunlit portions to account for differences in stomatal resistances. The aerodynamic
resistance in model I is approximated (Hicks et al., 1987) by:
neutral, stable conditions
unstable conditions
(2-37)
where u is the wind speed (m/s), and,
oe is the standard deviation of the horizontal wind direction fluctuations (deg.).
The level II model is similar to model I, except that a somewhat more refined estimate is
made of the aerodynamic resistance and canopy resistance. The formulation of ra in model II
may be better suited for regulatory applications because it does not need an estimate of oe,
which is not routinely measured at airport meteorological stations. Model II computes ra based
on the surface friction velocity and Monin-Obukhov length.
The deposition layer resistance is parameterized in models I and II as:
\V3
(2-38)
where, Dt is the thermal diffusivity of air (crn2/s), and,
Dx is the molecular diffusivity of the pollutant (cm2/s).
The canopy resistance is expressed as a composite resistance composed of several
parallel pathways, including transfer to the leaf stomata into the plant tissue, transfer to the leaf
cuticle, or transfer directly to the ground or water surface.
The variables used by the model to compute the stomatal resistance include detailed
information on vegetation which would not be available for many routine applications. The data
and computational requirements of the level III model (K-theory model) and level IV
(higher-order closure model) are considered too extensive for routine applications. Therefore,
the NOAA/ARL models are not being used as core models in'the current study, although the
2-34
-------
parameterization of some individual resistances may be included in future hybrid versions of the
core gas deposition models to be evaluated.
2-35
-------
3. Plume Depletion Techniques
Several different approaches to account for the depletion of the pollutant from the plume
due to dry deposition processes have been reviewed. The schemes examined include the source
depletion method (Chamberlain, 1953), the surface depletion method (Horst, 1977), the
K-theory approach (Rao, 1981), and a modified source depletion technique (Horst, 1983).
These models determine how the pollutant which is deposited at the surface is removed from
the plume.
3.1 Source Depletion
The source depletion technique is the simplest method for removing pollutant material from
a Gaussian plume. At each downwind distance, the source term (emission term) of the
Gaussian equation is adjusted (decreased) to simulate the effect of increased pollutant removal
with distance. At any distance, x, the source strength, Q is:
f X v _
QW = Q0 exp - / ^ D (x) dx
I « ,
(3-1)
where, Q0 is the initial source strength (g/s),
u_ is the transport wind speed (m/s), and, _
D(x) is a cross-wind integrated diffusion function (D = u ^(/Q0) with
units of (1/m), and,
X is the cross-wind integrated concentration (g/m2).
Although the source depletion method is computationally efficient, it has the drawback that
it immediately redistributes any pollutant loss at the ground throughout the entire depth of the
plume. This results from making the adjustment for plume depletion only in the source term
while retaining the original Gaussian shape of the vertical distribution. This redistribution of
plume mass in the vertical results in an overestimation of deposition in the near-field but
underestimation at large distances. The source depletion technique would probably be
acceptable for determining deposition at the point of maximum impact, but is not recommended
for use in the current study, because of the requirement for applicability to a wider range of
conditions.
3-1
-------
-\
3.2 Surface Depletion
The surface depletion technique (Horst, 1977) allows the pollutant to be depleted in the
vicinity of the surface rather than throughout the vertical extent of the plume. Thus, the
resultant plume is allowed to assumed a non-Gaussian shape. This is done by treating the
ground surface as "negative" sources, representing a sink for the pollutant. The surface
depletion equation is:
(3-2)
where D is the diffusion function (D = ux/Q0),
X is the concentration at spatial coordinates (x,y,z), and
the other terms have been defined previously.
The surface depletion method is generally considered to be the most accurate solution of
plume depletion and is frequently used as a reference method for other techniques. However,
the numerical solution of the complex integral requires between 10 and 100 times more
computer time than the source depletion technique. The large computational cost of the surface
depletion method makes it impractical for routine use. Another drawback of the surface
depletion technique for the current application is that its applicability is restricted to pollutants
with negligible gravitational settling velocities.
3.3 K-theory Approach
Rao (1981) describes a K-theory model for estimating concentrations in the case of
gravitational settling and deposition. The concentration is given by:
. Q0 Si si
x(*,y,z) = -i -i
« L Lt
(3-3)
3-2
-------
where
gi fey) 1S ^e crosswind diffusion function,
g*2 (x,y) is the vertical diffusion function modified for deposition effects, and
Ly, Lj are the length scales of the concentration distribution in the y and z
directions.
(3-4)
(3-5)
& exp
-vg (z-H)r _ vgr*
exp
-(z+Hf
-fiZ (vd - vgf2) r exp(
(3-6)
T -
2x
(3-7)
-v
,/2)r
v/2
(3-8)
(3-9)
The K-theory model was included in the evaluation study of Doran and Horst (1985). They
noted that the K-theory model is valid only when the vertical dispersion coefficient, oz, varies as
x1/2. For more general forms for oz (e.g., such as those used in regulatory models such as ISC),
the K-theory model does not conserve mass. Doran and Horst developed a method to estimate
the missing mass, and then add it back into the plume. Although this correction improved the
3-3
-------
performance of the model, it still did not perform quite as well as the modified source depletion
technique.
Winges (1990) describes the numerical integration scheme used to compute the mass
conservation correction factors for the K-theory method in the Fugitive Dust Model (FDM). A
least squares fit to computed values of the correction factors was calculated for the various
combinations of 6 wind speeds, 6 stability classes, 6 particle size classes, and 5 release heights
(i.e., 1080 values), and entered into the FDM code in data statements. Winges (1992, personal
communication) has indicated that in a new release of FDM, an external disk file with a much
larger number of the correction factors is supplied with FDM to cover a wider range of wind
speeds, size classes and release heights than the previous version of the model.
The lack of mass conservation in the unmodified K-theory model or alternatively, the need
to store large numbers of pre-computed mass conservation correction coefficients are viewed as
the major drawbacks to the K-theory scheme, especially since alternatives without these
liabilities are available.
3.4 Modified Source Depletion
Horst (1983) describes modifications to the source depletion technique to account for the
change in the vertical distribution of the pollutant due to deposition at the surface. In the
modified source depletion method, a profile correction factor, P (x,z) is defined and applied to
the diffusion function to account for plume depletion effects.
C(x,z) = Q(x)D'(x,z,h)
= Q(x)D/0(x,z,h)P(x,z)
(3-10)
where C(x,z) is the crosswind integrated concentration,
D'(x,z,h) is the crosswind integrated diffusion function,
D0'(x,z,h) is D' for a nondepositing pollutant, and
P(x,z) is the profile correction factor.
For oz = ax,
(3-H)
3-4
-------
I"1
1 I [ in^/z,) - 1 ]
(3-12)
where the resistance, R, is defined such that
C(x,z) = C(x,zd) [1 + vd
(3-13)
and zd is the reference height.
Horst (1983) provides equations for other forms of oz as well, including ax1",
ax(l + bx)"1/2, and ax(l + bx)"1. However, he does not provide solutions for the form
ax(l + bx)1/2, which is needed for McElroy-Pooler coefficients under A and B stability
conditions.
The modified source depletion method has been shown to produce results in close
agreement with the exact surface depletion method (Horst, 1984), but only required a small
fraction of the computer resources. In fact, its computational requirements were found to be
comparable to the source depletion method. In the dual tracer study of Doran and Horst
(1985), the modified source depletion model produced the best agreement with observations of
any of the plume depletion models tested. They used dual tracer field observations of a
depositing aerosol (ZnS) and a non-depositing gas (SF6) in their evaluation.
In summary, the modified source depletion method has been shown to produce accurate
results compared to both the exact surface depletion method and field observations. In addition,
its formulation is well-suited for use in a Gaussian model, it can treat gravitational settling
effects and plume tilt, it is computationally efficient, and it conserves mass exactly. For these
reasons, the modified source depletion method of Horst (1983) was selected as the approach to
use in the deposition module for assessing plume depletion.
The Horst (1983) method was implemented in ISC2 during a later phase of this study.
Extensions to the method that were required are described in Appendix E.
3-5
-------
4. Calculation of Meteorological Variables
Nearly all of the deposition models require estimates of the surface friction velocity (u.),
Monin-Obukhov length (L), and solar radiation. The methods of Holtslag and van Ulden (1983)
are widely-used to estimate solar radiation and surface sensible heat flux from
routinely-available meteorological and surface (land use) data. The Holtslag-van Ulden scheme
has been implemented into the HPDM model (Hanna and Chang, 199 la), extensively compared
and tested with field data (e.g., Hanna and Chang, 1992), and has been shown to produce
reasonable results. Therefore, the techniques used in the HPDM meteorological preprocessor
are used to produce the micrometeorology variables required in the deposition calculations.
4.1 Unstable/Neutral Conditions
The energy balance at the surface can be written as:
. + Qf = Qh - Qe - Qg (4-1)
where Q« is the net all-wave radiation,
Qf is the anthropogenic heat flux,
Qh is the sensible heat flux,
Qe is the latent heat flux, and
Qg is the ground/storage heat flux.
Holtslag and van Ulden (1983) provide the following parameterization of the net
radiation term:
Q - i (4.2)
(4-3)
4-1
-------
c3 =0.38^ -"*""* (4.4)
where T is the measured air temperature (deg. K),
A is the albedo,
o is the Stefan-Boltzmann constant (5.67 x 10"8 W/m2/deg. K*),
N is the fraction of the sky covered by clouds,
<|> is the solar elevation angle (deg.),
a is an empirical surface moisture parameter, and,
S is the slope of the saturation enthalpy curve [$=S/Y], where
s=8(qs)/a(T)andY=cp/L,
X is the latent heat of water vaporization,
qjj is the saturation specific humidity, and,
cp is the specific heat at constant pressure.
The four terms in the numerator of Eqn. (4-2) account for absorption of short-wave
radiation at the surface, incoming long-wave radiation from gaseous components of the
atmosphere (e.g., water vapor and carbon dioxide), incoming long-wave radiation due to clouds,
and outgoing long-wave radiation from the surface, respectively. The factor in the denominator
(l+Cj), results from the use of air temperature rather than the more difficult-to-determine
surface radiation temperature in the equation. The term in the first set of parentheses in Eqn.
(4-3) represents short-wave solar radiation in the absence of clouds. The second term
(l+b1Nb2), accounts for the reduction of incoming solar radiation due to clouds (bt is negative).
The values for the empirical constants clt Cj, a^ a2, bt, and b2 suggested by Holtslag and van
Ulden (1983) are shown in Table 4-1.
The flux of heat into the ground or storage in surface materials, Qg, is usually
parameterized during the daytime as a fraction of the net radiation (e.g., DeBruin and Holtslag,
1982; Oke, 1978).
Q8 - cgQ- (4-5)
where cg is an empirical coefficient which depends on the properties of the surface. Holtslag
and van Ulden (1983) obtained a value of cg of 0.1 for a grass covered surface in the
Netherlands. Oke (1982) indicates that typical ranges for cg are 0.05 to 0.25 in rural areas, 0.20
to 0.25 in suburban areas, and 0.25 to 0.30 in urban regions and suggests that typical values of cg
are 0.15, 0.22, and 0.27 for rural, suburban, and urban areas, respectively.
4-2
-------
Table 4-1
Values of Net Radiation Constants
(Holtslag and van Ulden, 1983)
Constant Value
at . 990 W/m2
a2 -30 W/m2
-0.75
3.4
5.31xlO-13W/m2/deg. K6
60 W/m2
4-3
-------
The anthropogenic heat flux, Qf, can usually be neglected, except in highly urbanized
areas. Hanna and Chang (1991b) contains a table listing typical values of Q{ for various cities.
The sensible heat flux, Qh and latent heat flux are determined by Holtslag and van Ulden
(1983) as:
* = F£F «?.+/-,)-«*xuo (4-6)
(4-7)
where P' is an empirical coefficient (-20 W/m2).
Typical values of a, based on empirical data of Holtslag and van Ulden and summarized
by Hanna and Chang (1991b) are:
« = 0.2 (arid rural areas)
a = 0.5 (urban areas, some parks, crops and fields during mid-summer .
when rain has not fallen for several days)
a = 0.8 (crops, fields, or forest with sufficient moisture)
a = 1.0 (normal wet grass in a moderate climate)
In neutral and unstable conditions, the following relationship developed by Wang and
Chen (1980) is used in HPDM and other models such as MESOPUFF II to compute the friction
velocity.
u. = [1 H- d. ln(l + d. dj } (4-8)
-l i I "2 3H
where
4-4
-------
0.128 + O.ff051n(z0/z) , z0/z <: 0.01 (4-9)
10.107 z0/z > 0.01 (4-10)
d2 = 1.95 + .32.6(z0/z)°-45 (4-11)
d3= (4-12)
The term d1ln(l + d2d3) represents the correction due to instability, u.n = ku/[ln(z - d)/z0], k is
the von Karman constant (~ 0.4), and d is the displacement height.
Hanna and Chang (1990, 1992) tested the analytical formula against values produced by
the iterative solution of u. and L. They found that the Wang and Chen (1980) expression
produced values within 10% of the results determined by the iterative solution for z = 10 m, d
= 0, Z0 = 1m, and a large value of Qh (400 W/m2). Better agreement was found for smaller
roughness elements and smaller sensible heat fluxes. In addition, the analytical solution was
computationally significantly faster.
The Monin-Obukhov length can then be computed directly from its definition once u, is
determined from Eqn. (4-8) and Qh from Eqn. (4-6).
L-
4.2 Stable Conditions
The Weil and Brower (1983) method for estimating u. is applied in HPDM during stable
conditions. A first estimate of the scaling temperature, 0., is calculated using Holtslag and Van
Ulden's (1983) equation:
6,! = 0.09(1 - 0.5N2) (4-14)
where N is the total fractional cloud cover and 6. has units of °K. Another estimate of 6. is
made from the profile equation for temperature:
4-5
-------
e.2 =
18.8 zg
(4-15)
where the neutral drag coefficient C^ is defined as k/ln[(z - d)/z0].
Then, 8. is set equal to the smaller of 8n and 8n.
The sensible heat flux, Qh, is defined during stable conditions as:
Qh = - pcpu.8.,
(4-16)
For large values of u (or u.), 8n (which depends only on cloud cover) is smaller than 8.2, but an
additional check on the product u.8. must be made, since Qh does not keep increasing
indefinitely with higher wind speeds. In HPDM, the value of 8. is not allowed to exceed
0.05/u., where the numerator has units of °K m/s and denominator has units of m/sec. This
limit is estimated from observations of heat fluxes during high-wind, stable conditions.
The friction velocity, u., can be calculated from:
V/2
u. = -=- 1 + 1-| ^- | (4-17)
where u0 - (4.7 z g 8./
Because 6. is set equal to the smaller of 8.t and 6.2, the following condition is always met:
(4-18)
During stable conditions, Hanna and Chang (1992) suggest a lower limit on L in
recognition of the fact that the atmosphere is less stable over urban areas than over rural
surfaces. Their suggested values for use for the various land use categories defined in the Auer
(1978) scheme are shown in Table 4-2.
4-6
-------
Table 4-2
Minimum Values of Monin-Obukhov Length
During Stable Conditions
for Various Land Use Types
(From Hanna and Chang, 1992)
Auer (1978) Class Description Minimum L
Cl Commercial > 40 story buildings 150 m
10-40 story buildings 100 m
< 10 story buildings 50 m
II, 12 Industrial 50 m
R3 Compact Residential " 50 m
Rl, R2 Residential 25 m
A Agricultural 2 m
4-7
-------
5. Model Evaluation Protocol
The major objective of the evaluation exercise is the development and implementation of
an objective model evaluation and scoring methodology that allows good-performing and
poor-performing models to be distinguished, so that an appropriate dry deposition technique
representing the current state-of-the-science may be selected. This protocol outlines an
evaluation approach which allows, to the extent possible, an objective model performance
scoring and selection process which results in the selection of a deposition model for gases and
another for particles. We have adopted the EPA's model statistical approach to model
evaluation as described by Cox and Tikvart (1990).
The accuracy of a model performance evaluation is dependent upon the accuracy and
representativeness of both the observations used for comparison and the model input data used
to produce predictions. Estimates of dry deposition rates presently suffer from numerous '
sources of uncertainty (Hosker and Lindberg, 1982) leading to the situation where there is a
wide spread of overlap for sets of deposition velocity observations even under similar
experimental conditions. As a consequence, it is expected that, given the small size of the
deposition velocity data sets and uncertainties present in the measurements and model input
data, the resulting uncertainties present in both observations and predictions of deposition
velocity will make it difficult in many cases to distinguish between the statistical performance of
some models. An exact quantitative description of the uncertainties is not possible for many of
the data sets due to limited information on the reported observed data.
5.1 Evaluation Approach
The model evaluation focused on a comparison of observed and predicted deposition
velocities. Since the ISC model uses reflective coefficients instead of deposition velocities, the
effective deposition velocity for the ISC scheme had to be estimated. Appendix A contains a
description of the method used to convert the reflective coefficients into effective deposition
velocities for use in the model evaluation effort.
In the present study we combined all the particle deposition data from the various
observational studies, many of which had less than 10 observations, into one large data set of
168 data points that have sufficient concurrent data for inputs to the deposition models. Several
subsets of deposition data were created based on one or more stratification criteria. On each
subset model performance statistics were calculated. These statistics were used to estimate a
composite measure of performance where the most accurate model produces the smallest
5-1
-------
composite measure based on the EPA's Cox-Tikvart approach utilizing fractional bias. In
summary, the model evaluation exercise consisted of the following steps:
1) Stratification of the observed deposition velocity data into subsets based on
several criteria such as physical and chemical properties, surface characteristics,
meteorological conditions, etc.
2) Direct statistical and graphical comparison of observations versus predictions of
deposition velocity for each subset using the EPA's Cox-Tikvart protocol (Cox
and Tikvart, 1990).
3) Ranking of model performance using a composite performance measure (CPM)
and selection of the best performing model based on rank and uncertainty in the
CPM.
The models which were evaluated and scored were discussed in Section 2. In addition to
core models, several hybrid models were created for performance evaluation based on our
review of the physics represented by the models, and on the outcome of the sensitivity analysis.
The technical review provided indications of which components of the various core models are
superior, and which may have some unrealistic characteristics. We created several hybrid
models from the core models in order to offer the U.S. EPA the best performing model (instead
of several that perform about the same and which may have one or more flaws that can create
problems in the context of the wide ranging applications seen by ISC). The hybrid models were
exercised in the same manner as the core models and received the same statistical treatment
and scoring as the core models.
5.2 Stratification of Deposition Velocity Datasets
The complete data set of observed deposition velocities includes observations for several
chemical species plus particles made over a variety of surfaces and under a wide range of
meteorological conditions and sampling times. Model performance was expected to vary
significantly between chemicals or particle sizes, and for different surfaces and meteorological
conditions. As a result we divided the complete data set into subsets based on the following
criteria:
1) physical and chemical characteristics
2) surface conditions
3) meteorological conditions
5-2
-------
Figure 5-1 illustrates these subdivisions that were made in the data. A primary division is made
depending whether the substance is a particle or a gas since the physical deposition processes
are different for gases than particles. Further subdivision of the data was made in such a
manner that a minimum number of observations was lost to datasets with sample sizes less than
10. The stratification criteria that appear to provide large enough sample sets and which are
informative and nonredundant are the following:
> Physical characteristics
a) Particles
1) 2 size ranges (diameters less than 0.1 micron and between 0.1 and
20 microns),
b) Gases
1) Chemical species (SO2 and O3 are only two species with sample
sizes in the hundreds)
> Surface Characteristics
a) Particles and Gases
1) rough/smooth surface based on surface roughness length
2) dense/sparse vegetation density based on leaf area index
> Meteorological Conditions
a) Particles and gases
1) day/night based on time of day
2) large/small atmospheric diffusion based on friction velocity
3) hot/cold based on ambient temperature
Division criteria were selected in order to (1) provide a logical break between two
conditions (e.g., night = 000-600 1800-2400, day = 600-1800) and (2) provide a breakdown of
the data set into subsets of approximately the same size.
The evaluation effort of particle deposition has focused on smaller particles ( < 20 urn
diameter) because this size range is where the most significant differences among the deposition
models occurs. For particles above 20 jim diameter, the deposition velocity predictions for all of
the models tend to quickly approach the gravitational settling velocity. Appendix D contains
plots of deposition velocities as a function of particle diameter which illustrates this point.
5-3
-------
Deposition velocity dataset
PARTICLES
GASES
|
size
range
physical
rough/
smooth
dense
/sparse
surface
met
night/
day
hot/cold
eorologic
I 1
rough/
smooth
dense
/sparse
night/
day
hot/coldi
i
, surface meteorologic
d 1
characteristics type conditions typc conditions
Figure 5-1. Schematic illustration of division of deposition velocity data into subsets.
5-4
-------
5.3 Comparison of Predictions and Observations
For paired comparisons, the performance measures for the previously described subsets
are based on an analysis of deposition velocity residuals either paired in time, paired in space,
or paired in both space and time. Since deposition velocities are dependent on local surface
conditions, pairings only by time would not be appropriate for the present study. Likewise, due
to (1) limited simultaneous monitoring and (2) variations in sampling time between data sets,
the use of pairing only by space does not seem to be appropriate either. Consequently the focus
of the present study is on fully paired (space and time) data.
A number of traditional performance measures such as bias, variance, gross variability,
and average absolute residuals suffer from the lack of ability to intercompare across the
deposition velocity data subsets. Furthermore, other statistical measures utilized by the EPA
(e.g., Cox and Tikvart, 1990) possess similar information on bias and precision, and are
normalized in a manner that allows data set intercomparison and composite scoring of model
performance. We utilized four types of normalized statistical measures including:
1) fractional bias of arithmetic averages (FBA)
2) fractional bias of arithmetic standard deviations (FBSD)
3) fractional bias of robust extreme statistic for the n smallest values (FBSE)
4) fractional bias of robust extreme statistic for the n largest values (FBLE)
These four measures all possess the same metric (e.g., have the same range bounding
and limits) and consequently are directly intercomparable. The fractional bias of the average
(FBA) is defined as:
FBA
(P + P)
(5-1)
where the average is given by:
N
N
1
o = - o.
(5-2)
The fractional bias of the standard deviation (FBSD) is defined as:
(<3n - a\
FBSD = 2^-2 ft-
. (5-3,
5-5
-------
where the standard deviation is given by:
r < N w
(5-4)
The FBA is useful to quantify the degree of model over or underprediction. FBSD is useful to
determine if model predictions have more or less variability than observations.
Deposition velocity minimums are important to maximum ambient concentrations, while
maximums are important in maximum exposure estimates. Thus the robust extreme statistic for
both the n largest and smallest deposition velocities are important.
The robust extreme statistic (RES) is defined for the maximum/minimum extreme as:
O(RES) = O(n) + ( O - O(n) ) In
i z i
(5-5)
where the average O is over the n - 1 (largest/smallest) values, O(n) is the nth (largest/smallest)
value and n is the number of values in the extreme set.
The fractional bias of the extreme statistic is given by:
Z7D/T E* «». CFY *> \\J\K**&) "~ * V*^****/)
rD\L+c* or o/*/ = z.
(0(RES) - P(RES))
(5-6)
The FBLE provides an indication of the degree of over- or underprediction in the largest values
while FBSE does the same for the smallest values.
5.4 Scoring Model Performance Using Composite Measures
A composite performance measure (CPM) is calculated for each model as a weighted
linear combination of the individual absolute fractional bias components. Each of the four
fractional bias measures previously defined are combined as a weighted linear sum. The
compensatory effect of sign (over- or underprediction) is removed by taking the absolute value
of each bias measure. The resulting sum which serves as a composite fractional bias (CFB) is
defined for both the overall data set
CFB0 = Wl\FBAo\ + w2\FBSDa\ * W3|FBSEJ * *4\FBLE0\
. . (5-7)
5-6
-------
and for each of the stratification subsets. The CFB for kth subdivision is defined as:
CFBt = wjFa4t| + w2\FBSDt\ + w3\FBSEk\ + w4\FBLEt\
(5-8)
All measures are assumed to be equally important and the weights serve merely as an average,
e.g.,
(5-9)
The CFBk is averaged over all subsets (assuming each subset is equally important).
,
CFB = - £ CFBk
n t-i
(5-10)
where in the present case there are 12 subsets (M = 12). The average is then combined with
the CFB0 resulting from the overall data in order to produce the composite measure (CPM).
CPM = - CFB + - CFB
2 2
(5-11)
The CPM thus contains information about both overall model performance as well as
performance under specific sets of conditions. The use of the CPM to select an appropriate
model reduces the possibility that model performance biases under specific sets of conditions
(which may be over represented in the complete data) will unduly influence model selection. A
perfect model will produce a CPM of zero. The CPM is independent of whether the model
over- or underpredicts. The model with the smallest CPM is selected as a winner if it can be
shown that the CPM is lower than other models in a statistically significant sense (e.g., different
at the 95% confidence level).
In the present study we estimate the 95% confidence interval of the CPM for each
model. The confidence interval is used to establish how robust our selection process may be
based on our single estimate of CPM. A model could have the smallest CPM but if it is not
very different from that of another model, or if the estimate is relatively uncertain, then the
model should not be selected.
5-7
-------
The difference in the CPM between all pairs of models is estimated in order to
determine if differences between models are significant. The difference between the
performance of one model and another is the model comparison measure (MCM), defined as;
MCA/043) = CPM(A) - CPM(B)
(5-12)
where CPM(A) = Composite performance measure for Model A
CPM(B) = Composite performance measure for Model B
The MCM is used to judge the statistical significance of the apparent superiority of any one
model over another. If the MCM is not significant from zero at the 95% confidence level then
the two models cannot be said to perform in a significantly different manner.
The bootstrap resampling technique is used to estimate confidence intervals on the
various measures described above. In applying the bootstrap procedure, observed and predicted
data pairs are resampled. Sampling is done with replacement, so some data points are
represented more than once. This process is repeated 1000 times so that sufficient samples are
available to calculate the standard error of each measure. The resampling recognizes the
stratifications made on each of the model input variables as individual blocks. This assures that
each of the 1000 variants of the original dataset retained the same number of samples from each
stratification. Had we not blocked the data in this way, one of the 1000 variants might, for
example, only consist of a few samples associated with say the largest friction velocity (repeated
many times). The bootstrap resampling method allows the standard deviation, spm, of any
performance measure to be estimated, from which a confidence interval can be calculated. The
actual CPM or MCM for each model or model pair is assumed to have a 95% chance of lying
within the range given as
- CSpm < CPM^ < CPM + cspm
(5-13)
where c is a multiplier for a specific percentage level of confidence.
The standard error of the estimate is simply the standard deviation of the measure over
all of the bootstrap-generated outcomes. If the measure involves a single comparison, such as
CPM for a single model, then the value of c can be set equal to the student-t parameter.
Difference measures such as the MCM require that simultaneous confidence intervals be
found for each pair of models in order to ensure an adequate confidence level and to protect
5-8
-------
against falsely concluding that two models are different. The method of Cleveland and McGill
(1984) is used to calculate c. In this method, c is found such that for 95 percent of the 1,000
bootstrap i-tuples,
(5-14)
where Ay = model comparison difference measure for model pair i,j,
Ajkl = model comparison difference measure for model pair i,j and bootstrap
replication k, and
By = standard deviation of all the Aijk values.
The 95% confidence intervals for CPM and MCM for each model and pair of models is
presented. The confidence intervals on the MCM are critical to this study since if the range
that the actual MCM may be found in does not include zero, then the two model's performance
statistics can be said to be statistically different.
Model selection was based on model rank, and if the model MCM was significantly non
zero for all other models. Models with overlapping confidence intervals (MCM that are not
significant from zero) were grouped. Models clustered within groups were treated as having the
same model performance.
Summary of Scoring Scheme
In summary, the steps taken in providing a scoring of each model analyzed are as
follows:
(1) For each model calculate the FB's for the observed and predicted deposition
velocities paired by time and location over all data and for each stratification
(block). Calculate the CPM with confidence limits and summarize the model
performance by category using CFB's. The smaller the CFB's and the CPM, the
better the overall performance of the model.
(2) Rank CPM from smallest to largest. Calculate confidence intervals for each
CPM to determine if clearly superior or inferior model performance occurs based
on ranking and confidence interval.
5-9
-------
(3) Calculate the MCM with confidence intervals for each possible model pairing.
Rank MCM's and determine which model pairs have a significantly non zero
difference.
5-10
-------
6. Model Evaluation Data Bases
A list of observational data sets of particle deposition used in the evaluation of the
particle deposition models described in Section 2.1 is presented below. The observational
deposition velocity data are reproduced in Appendix C.
6.1 Particle Data Sets
(1) Doran and Horst (1985). This paper reports on the results of a field experiment
involving dual tracers. Simultaneous releases were conducted of ZnS, a polydisperse aerosol,
and nondepositing SF6. Deposition velocities of the particles were computed and are
summarized in the paper. The study used the data to compare four different plume-depletion
models.
(2) Lorenz and Murphy (1989). The aerodynamic profile method was used to determine
deposition rates of 1.0 jim diameter particles to a rough vegetated surface. Figures showing
deposition velocities as a function of meteorological parameters are contained in the paper.
Tables summarizing the mean deposition velocity as a function of particle size contain 61 data
points.
(3) Nicholson and Davies (1987). The profile method was used to derive dry deposition
rates of particulate sulfate over a rural site in England. Approximately 170 observations were
made over a one year period. Of these, 78 data points remained after application of a set of
stringent quality criteria. Tabulated values of deposition velocities and meteorological
conditions are presented in the report.
(4) Sehmel (1980), Sehmel and Hodgson (1978). These studies contain the results of
wind tunnel measurements of particles in the size range from 0.03 jim to 29 \im diameter. Low
roughness surfaces were used in the experiments, ranging from smooth brass to gravel (z0 up to
0.6 cm). Least squared techniques were used to develop a set of equations fitting the data
points. Graphs are provided in the papers showing deposition velocities predicted by the least
squared fit equations (though not the original data points) as a function of roughness length,
friction velocity, and particle diameter. The predictions from the Sehmel and Hodgson
regression equations (as coded in the CARB model) were plotted and intercompared with the
other deposition models.
6-1
-------
(5) Hicks et al (1986). Eight data points for the dry deposition participate sulfur were
obtained from this study which also measured gaseous fluxes. Measured meteorological
parameters included temperature, wind speed, and friction velocity.
(6) Garland (1982). This paper summarizes the dry deposition rates of small particles
to grass in field and wind tunnel experiments. Deposition velocities for lead are presented as a
function of particle diameter. Additional data for an oxide of iron is presented. These data
show similar results from measurements in the field and wind tunnel.
(7) Wesefy et al (1982). Eddy correlation measurements were used to estimate the dry
deposition of particulate sulfur. Approximately 19 data points are available including
measurements of sensible heat and friction velocity. The data points were gathered over short
grass during drought conditions in eastern Texas. Summaries of eddy-correlation experiments
over other surfaces are also presented.
(8) Wesefy et al, (1983). Observations of submicron particle deposition velocities were
made in a deciduous forest during winter over the course of a week. A fine particle sensor was
used to measure particles in the 0.01 to 2 micron size range. Approximately a dozen eddy
correlation estimates of deposition velocities were reported. Micrometeorological turbulence
data was collected simultaneously with the deposition data.
6.2 Gas Data Sets
,
The observational data sets of gas deposition which have been collected for use in the
future evaluation of gas deposition models are summarized below.
(1) Meyers and Baldocchi (1988). Direct eddy correlation measurements of SO2 and O3
fluxes are summarized for both well-watered and water-stressed conditions. Deposition
velocities and meteorological data are presented for approximately 31 periods during two
experiments. The study discusses a comparison of the observations to a hierarchy of deposition
models from simple to highly complex.
(2) McMillen et al (1987). This report contains measurements of dry deposition of SO2
to a forested site in Germany. The measurements were made using the eddy correlation
method. A tabular listing of the meteorological and deposition data are presented in an
Appendix of the report. The report discusses limitations to the data and conditions under which
the results are considered usable. Basically, good results were obtained when SO2
concentrations exceeded about 1 ppb.
6-2
-------
(3) Hicks et al (1989). Data are summarized from an intensive field study of dry
deposition of SO2 to a variety of crops at a site in Pennsylvania. Conservative quality-assurance
guidelines were used to screen the raw data base and produce about 22 useful data points. The
report contains tabular listings of the meteorological data and deposition results from the study.
(4) Harrison et aL (1989). A gradient technique was used to determine deposition
velocities of HNO3 and HC1 over a variety of vegetative surfaces. Approximately 34 data points
are provided for the deposition velocity along with selected meteorological parameters, including
the surface friction velocity for each run.
(5) Meyers et aL (1989). The deposition velocity of HNO3 over a fully leafed deciduous
forest was estimated using a gradient technique. The observations were compared to the
predictions of a detailed canopy turbulence model. Heat flux and friction velocities are provided
for each experiment. Approximately 10 data points are listed in tables in the paper.
(6) Godowitch (1990). Vertical ozone fluxes were measured from aircraft over several
different land use types, including agricultural crops and forested areas. An analysis was
performed to derive ozone deposition velocities for each experiment. Measurements of selected
meteorological parameters are also provided.
(7) Wesefy et al., (1983). Eddy correlation measurements of ozone fluxes were
performed above a leafless deciduous forest. Measurements of fine particles with a diameter of
approximately 0.1 um were also performed. Ranges of deposition velocities and meteorological
parameters are tabulated in the paper. The results of 19 half-hour averaged measurements are
also provided for both ozone and particulate sulfur deposition, although the particle data contain
a low signal-to-noise ratio. However, at least the ozone measurements (10 points) appear to be
usable.
(8) Wesefy et aL, (1978). An eddy correlation technique was used to determine vertical
fluxes of ozone above both mature and senescent maize canopies. The results of approximately
26 runs are tabulated. Both deposition velocities and meteorological measurements (e.g.,
frktion velocity, heat flux) are provided.
(9) Wesefy et aL, (1981). Measurements of ozone deposition velocity over snow, wet
bare soil, and water were made using the eddy correlation technique. Over 60 data points are
listed. Meteorological data during each experiment is provided. An interesting component of
this study is the importance of evaporation (Stefan flow) in influencing the deposition rate.
6-3
-------
(10) Hicks et aL, (1986). Eddy correlation measurements of vertical fluxes of gaseous
and particulate sulfur compounds were conducted over two different surface types. Tables of
meteorological conditions and deposition velocities are provided for 26 different time periods.
Data is also presented for NOX fluxes, but it is highly variable due to a low signal-to-noise ratio
in the NOX measurement system. It appears that 8 points of SOZ deposition are usable.
(11) Fowler and Cape (1982). The eddy correlation method was used to calculate SO2
fluxes over a Scots pine forest during daytime hours. Tables of flux, deposition velocity, and
sensible heat are presented for 20 points. Dry deposition rates ranged from 0.5 to 10 mm/sec.
(12) Davies and Mitchell (1982). This paper presents dry deposition rates for SO2 based
on gradient method measurements over grass in rural eastern England during a period of "
relatively constant atmospheric conditions. Twenty-five data points are available including wind
speed, friction velocity, sensible heat flux, and Monin-Obukhov length.
(13) Huebert (1982). A modified Bowen Ratio method was used to measure nitric acid
fluxes in this study. Measurements were made over a pasture near Champaign, Illinois.
Meteorological parameters measured include wind speed and friction velocity. The results of
the study showed that dry deposition is capable of depositing HNO3 at a rate comparable to that
of wet deposition measured in previous experiments.
(14) Padro et aL (1991). In this paper, the ADOM dry deposition model was compared
to observed ozone deposition velocities taken over a fully leafed mixed deciduous forest during
July and August of 1988. Meteorological data collected during the study include wind speed,
wind direction, temperature, dew point, solar radiation, net radiation, friction velocity, sensible
heat flux, latent heat flux, and stability. Dr. Padro has kindly supplied us with the data from this
experiment on a floppy diskette.
6-4
-------
7. Results of Model Evaluation
7.1 Particle Deposition Models
For identification purposes the model designations listed in Table 7-1 identifies each of
the ten models and describes the differences between the models. The modifications adopted
for the hybrid models are summarized in the table.
The particle data set consists of 168 cases in total where nonzero observations of
deposition velocity were made. A total of twelve data subsets were developed in order to learn
if model performance biases and imprecision appeared systematically in the data. Table 7-2
summarizes the stratifications made, along with the stratification criteria. For each of the
stratifications there are greater than 10 cases. In each of the following sections we discuss the
results for each stratification.
7.1.1 Full Data Sets
One of the first steps in the model performance evaluation exercise was to examine the
scatter plots of the observed versus the predicted deposition velocities. A complete set of
scatter plots for each of the ten models is presented in Appendix B as Figures B-la through
B-lj. From these scatter plots we can note that the observed deposition velocity ranges through
nearly three orders of magnitude. Most of the models with the exception of ISC also reproduce
the wide range of variation. There appears to be a tendency for the original model algorithms
(GARB 1, ADOM 1, UAM 1, and ISC) to underpredict the deposition velocity. The scatter
plots also indicate that there is a very wide range of scatter through most ranges of observed
deposition velocity. The major exception that can be noted is a tendency for the scatter to
decrease with increasing deposition velocity, suggesting that the fractional (percentage) error
decreases with increasing deposition velocity.
The scatter plots indicate that there appears to be a significant decrease in average bias
in the hybrid model results compared with the original model formulations. For example,
comparing B-ld with B-la, or B-le with B-lf, or B-lh with B-lg we find that several groups of
observation which were originally underpredicted in a significant manner, are predicted
significantly better by the hybrid model formulation.
7-1
-------
Table 7-1
A Summary of the Model Designations
Model Name
Description of Model
CARB1*
CARBO
CARB 2
CARB 3
ADOM 1*
ADOM2
ADOM 3
UAM 1*
UAM2
ISC*
Unmodified CARB model
Removal of CARB roughness length restriction of z0 < 10 cm
Same as CARB 1 except uses fixed temperature (T = 75 degrees F)
Uses fixed temperature (as in CARB 2), a LAI adjustment to I3, and
a constant reference length of 0.5 cm in I3
ADOM dry deposition algorithm
Modified Stokes/Schmidt relations and phoretic effects term
Changes made to ADOM 2 + LAI adjustment to rd
Unmodified UAM-V model
Contains LAI adjustment to rd
ISC model with boundary layer estimator for He/x
* Core (unmodified) models
7-2
-------
Table 7-2
A Summary of the Stratifications Made To the Small Particle Data Set (N = 168)
Stratifying Variable
Particle Diameter (microns)
Roughness Length (m)
Leaf Area Index
Day/Night Insolation
Friction Velocity (m/s)
Temperature (deg K)
Definition
Threshold(s)
*0.1
0.1 < d < 20.0
sO.25
> 0.25
s3.0
> 3.0
night
day
sO.25
> 0.25
s 290.0
> 290.0
Threshold
Selection
non inertial
inertial
simple canopy
complex canopy
nonforest
forest
-
-
sample size
sample size
sample size
sample size
Sample Size
13
155
97
71
105
63
42
51
58
110
62
106
7-3
-------
The ISC model exhibits the most unique scatter plot signature. Due to the fact that
small particles with a settling velocity of less than 0.1 cm/s are assumed to have a reflection
coefficient equal to one, a majority of particles have essentially a zero deposition velocity (small
values in Figure B-lj are lower bounds that are set so that the deposition velocity does not
identically equal zero). The non-zero values of the predicted deposition velocity in Figure B-lj
are strongly underpredicted due to the fact that the deposition velocity components are
multiplied by one minus the reflection factor, a, which is generally smaller than 0.1.
All of the fractional and several composite statistical performance measures are
summarized in Table 7-3. The ADOM particle models, particularly the hybrid form AD.OM 2
appear to have, in an overall sense the lowest bias and best precision of all of the models. The
UAM line of models are second best, while the CARB line of models are third. The ISC model
is ranked as the worst performing model by a large margin of CPM. The ADOM 2 model has
relatively little bias in both the average and standard deviation.
Figure 7-1 shows the fractional biases present in the data taken as a single data set. The
results indicate that ADOM has the smallest bias in the mean and standard deviation.
According to FBSE, the smallest deposition velocities appear to be best predicted by the CARB
3 model while according to FBLE the largest deposition velocities are predicted best by the
ADOM 3 model. The confidence intervals on these measures for the entire data set are rather
large, sometimes being as large as 50% of the measure. In order to reduce this wide error bar
an average was taken over all data stratifications, reducing the confidence interval by over a
factor of 3. The resulting measures are graphically summarized in Figure 7-2. This figure
indicates that the conclusions previously drawn with the CPM are consistent across all data
subsets. In both cases the ADOM 2 model overall does the best and the ISC model does the
worst.
The scatter plots of the fractional measures of model performance are presented in
Figures 7-3a through 7-3d. A co-plot of the fractional bias in the average and standard
deviations is shown in Figure 7-3a. This plot indicates that model performance for six of the
models is clumped in a region that indicates the models underpredict on both the average and
standard deviation of fractional bias. The single model that appears clearly most accurate on
both measures is the ADOM 2 hybrid model. The ADOM 3 model appears to be the only
model which overpredicts both the average and the standard deviation. The ISC model appears
to be a second outlier, indicating a large degree of underprediction in both the average and
standard deviation. The co-plot of the fractional bias'in the average of the entire data set versus
that of the 10 largest deposition velocities in Figure 7-3b shows a strikingly similar pattern as
that of Figure 7-3a. When the average bias of the RSE of the 10 smallest deposition velocities
7-4
-------
Table 7-3
A summary of the fractional and composite statistical measures for each of the models
examined. The last column is rank based on CPM (Rank 1 = smallest CPM).
Model Name
CARBO
CARB 1
CARB2
CARB 3
ADOM 1
ADOM2
ADOM 3
UAM1
UAM2
ISC
N=168
FBA
0.99
1.07
0.98
0.80
1.01
-0.06
-0.81
0.99
0.61
1.85
N=168
FBSD
0.85
0.85
0.76
0.81
0.45
0.17
-0.34
0.60
0.54
1.72
N=10
FBLE
1.09
1.09
1.01
0.91
0.49
0.56
-0.74
0.86
0.69
1.77
N=10
FBSE
0.96
0.90
0.58
-0.02
1.03
-0.95
-1.27
0.79
0.57
1.99
CFB0
0.97
0.99
0.83
0.64
0.75
0.44
0.62
0.81
0.60
1.83
Rank
8
9
6
4
5
1
2
7
3
10
7-5
-------
2.5
CO
5
"3
c
O
1.5
O
CO
0)
J
3
"o
en
0.5
0
1
I
Carb 0 Carb 1 Carb 2 Carb 3 Adorn 1 Adorn 2 Adorn 3 UAM 1 UAM 2
ISC
Deposition model
Fba
Fbsd
Fbse
Fble
Figure 7-1. A summary of absolute value of the fractional bias using all data (n = 168).
7-6
-------
2.5
CO
CO
£
"3
C
o
_)
O
ca
-------
c
o
1-1
Is
'>
Q
$H
cd
C
(£1
-iJ
CO
2.0
l.(
1.0
Z o.o
o
w
»**
ffl
T3
fi
o
o
CO
-0.!
-1.5
-2
1 | I 1 1
G
A
O
i i
Key
CD CARB 1
O CARB 2
CARB 3
CARB 0
ADOM 1
ADOM 2
ADOM 3
UAM 1
UAM 2
ISC
A
+
X
V
Q
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0
Fractional Bias of the Average
1.5 2.0
Figure 7-3a. Co-plot of fractional bias of the standard deviation for each of the deposition
models. The sample set includes all data points. The box indicates the region
within which the predictions are within a factor of 2 of observations.
7-8
-------
«i.U
1C
.0
CO
e
%
w
"*rn r\ R
CO U.O
cc
so
J-,
CO
J
O
CO
co
"cfl
fi
.2
o 1 n
K ADOM 3
V UAM 1
© UAM 2
Q ISC
-1.5 -1.0 -0.5 0.0 0.5 1.0
Fractional Bias of the Average
1.5 2.0
Figure 7-3b. Co-plot of fractional bias of the 10 largest deposition velocities for each of the
deposition models. The sample set includes all data points. The box indicates
the region within which the predictions are within a factor of 2 of observations.
7-9
-------
2.0
1.5
W
o
6
X
M
1.0
0.5
*, 0.0
cd
*I
ffl
"3
cj
o
ii
4_>
o
(0
-O.i
-1.0
i i i I i i i j i i iji i i | i i i i i IGJJ
_p QI i i i
-2.0 -
I I I
Key
Q CARB 1
Q CARB 2
A CARB 3
+ CARB 0
X ADOM 1
O ADOM 2
^ ADOM 3
V UAM 1
© UAM 2
Q ISC
1.5 -1.0 -0.5 0.0 0.5 1.0
Fractional Bias of the Average
1.5 2.0
Figure 7-3c. Co-plot of fractional bias of the 10 smallest deposition velocities for each of the
deposition models. The sample set includes all data points. The box indicates
the region within which predictions are within a factor of 2 of observations.
7-10
-------
2.0
1.!
w
a
0>
+J
X
S 0.5
a
w
M
(0
rl
m
o.o
-O.i
-1.0
-l.i
-2
iii
X v
-o-
CD
^ ,
^2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5
Fractional'Bias of Largest Extremes
1 1 1
Key
2.0
CD CARB 2
A CARB 3
+ CARB 0
X ADOM 1
O ADOM 2
^ ADOM 3
V UAM 1
m UAM 2
Q ISC
Figure 7-3d. Co-plot of fractional bias of the 10 largest and 10 smallest deposition velocities
for each of the deposition models. The sample set includes all data points. The
box indicates the region within which predictions are within a factor of 2 of
observations.
7-11
-------
are co-plotted with the fractional bias of the average for the entire data set (Figure 7-3c) the
CARB 2 and 3 and UAM 2 models appear to best predict small deposition. The co-plot of the
fractional bias in the 10 smallest and largest deposition velocities (Figure 7-3d), indicates that
half the models underpredict both the largest and smallest deposition velocities. The only
models which lie within a factor of 2 for both extremes are the ADOM 2 and the UAM 2
models.
7.1.2 Stratification by Particle Diameter
The stratification by particle diameter is important in order to see which models perform
best for intermediate sized particles that behave neither as rapidly sedimenting particles or as a
gas. Thus we will be evaluating model performance for very small particles ( < 0.1 microns)
versus intermediate size particles of size 0.1 through 20 microns.
The results of the performance evaluation is quantitatively summarized by Table 7-4.
The most notable points to be drawn from this table are the following:
1) The CFB suggest that the ADOM 2, ADOM 3 and CARB 0 models are the most
accurate models for small particles however the CARB model variants do
perform poorer for intermediate sized particles due to significant
underproductions.
2) The overall most accurate model with the smallest CFB seems to be the ADOM
2 and 3 models which seems to treat both small and intermediate particles.
equally well.
3) All measures indicate that the ISC model consistently performs the worst of all of
the models with significant underpredictions of the deposition velocity.
While the CARB 0 model does quite well in terms of the average bias, particularly for
small particles, the variances as measured by the standard deviation (FBSD) is actually
significantly underpredicted. The ADOM 2 model underpredicts the depositions velocity for
small particles, but overpredicts the deposition velocity for the intermediate particle sizes. The
ISC model cannot predict deposition for particles smaller than 0.1 microns since the deposition
velocity is basically set equal to zero by setting the reflection coefficient equal to 1. This effect
results in the extremely large CFB in Table 7-4 for the ISC model. When the composite bias
over both diameter size ranges are considered, Figure 7-4 indicates that the ADOM hybrid
7-12
-------
Table 7-4
A summary of the fractional and composite statistical measures for each of the models examined.
The first row is for particle diameters less than 0.1 microns, the second is for particles in the range 0.1 to 20 microns.
MODEL DESIGNATION
MEASURE
FBA
FBSD
FBLE
FBSE
CFB
SAMPLB
SIZE
13
155
13
155
13
155
13
155
-
CARS 1
.90
1.08
1.83
.84
1.41
1.09
.95
1.10
1.27
1.03
CARB2
.64
.99
1.67
.76
1.16
1.01
0.68
0.75
1.04
0.88
CARS 3
1.19
.79
1.59
.82
1.45
0.91
1.24
0.17
1.37
0.67
CARBO
.10
1.04
1.44
.85
0.85
1.09
-0.24
1.10
.66
1.02
ADOM1
.97
1.01
1.62
.44
1.24
0.56
0.83
1.16
1.17
0.78
ADOM2
.61
-.08
.85
.18
0.71
0.56
0.52
0.96
0.67
0.44
A DOM 3
.12
-.83
.25
-.31
0.24
0.07
0.11
-1.26
0.18
0.62
UAM1
1.67
.97
1.89
.60
1.77
0.86
1.62
0.95
1.74
0.84
UAM2
1.50
.58
1.76
.54
1.63
0.69
1.46
0.74
1.59
0.64
ISC
2.00
1.85
2.00
1.72
2.00
1.77
2.00
1.99
2.00
1.84
-------
02
Li.
U,
«
3
I 2
u
u
"35 1
o A
D.
I
0
*l
^
ll
t
I;.
II
1^,
Carb
0 Garb 1 Garb 2 Garb 3 Adorn i Adorn 2 Adorn 3 UAM 1 UAM 2 ISC
Deposition model
Diameter
Day/Night
Type of Data Subset
Ro ugliness
Ustar
cm LAI
I I Temp
Figure 7-4. A summary of the total CFB of the six types of data subsets. The total
represents a sum over low/high categories for each subdividing variable.
7-14
-------
models perform best, while of the three core models, the CARB 0 model performs slightly but
not significantly better. The ISC model is the poorest performer regardless of size.
7.1.3 Stratification by Roughness Length
The roughness length figures prominently in most of the deposition models undergoing
evaluation. For small roughness lengths where a complex canopy is not present the models
would be expected to perform at their best. The fractional and composite performance statistics
which are presented in Table 7-5 indicate the following points:
1) According to the CFB, the ADOM 2 model performs in a superior manner
regardless of the underlying surface. The model produces slight overpredictions
on the average.
2) The statistical measures indicate that all models except ADOM 3 performs
poorer over rough surfaces than smooth ones.
3) The UAM 2 and ADOM 3 are the second best performing model with UAM 2
always tending to produce moderate underpredictions and small variability in the
deposition velocity, while ADOM 3 tends to produce overpredictions.
4) All measures indicate that the ISC model consistently performs poorest of all
models, regardless of surface, and always produces significant underpredictions as
noted from the large positive FB's.
7.1.4 Stratification by Leaf Area Index
Several hybrid models possess an explicit dependence of deposition velocity on Leaf
Area Index (LAI). The LAI for each observation was separated into complex surfaces (e.g.,
forests) where the LAI is 3 or greater and simple surface ceases where the LAI is less than 3
(e.g., grass). If LAI represents an appropriate increase in collection area, then adding a
adjustment for LAI should, in principle, improve model predictions. The normalized
performance measures summarized in Table 7-6.
7-15
-------
Table 7-5
A summary of the fractional and composite statistical measures for each of the models examined.
The first row is for roughnesses < 0.25 m while the second is for > 0.25 m.
MODEL DESIGNATION
ON
MEASURE
FBA
FBSD
FBLE
FBSE
CFB
SAMPLE
SIZE
97
71
97
71
10
10
10
10
_
.
CARB1
.90
1.49
.85
1.77
1.09
1.67
1.21
1.19
1.01
1.33
CARS 2
.81
1.38
.76
1.63
1.01
1.49
0.88
1.07
0.87
1.39
CARB3
.80
.80
.79
1.36
0.91
1.38
0.33
0.86
0.71
1.10
CARBO
.90
1.20
.85
1.26
1.09
1.27
1.21
0.93
1.01
1.17
ADOM 1
.73
1.79
.45
1.61
0.49
1.57
1.26
1.83
0.73
1.70
ADOM 2
-0.4
-.11
.20
-.12
0.56
-0.09
-0.84
0.04
0.41
0.09
ADOM 3
-.26
-1.28
.21
-1.17
0.53
-1.08
-1.16
-.41
0.54
0.99
UAM1
.73
1.76
.62
1.92
0.86
1.89
1.06
1.53
0.82
1.77
UAM2
.57
.69
.52
1.07
0.69
1.17
0.87
1.16
0.66
1.02
ISC
1.79
2.00
1.72
2.00
1.77
2.00
1.99
2.00
1.82
2.00
-------
Table 7-6
A summary of the fractional and normalized statistical measures for each of the models examined.
The first row is for LAI < 3.0 while the second is for LAI > 3.0.
MODEL DESIGNATION
MEASURE
FBA
FBSD
FBLE
FBSE
CFB
SAMPLE
SIZE
105
63
105
63
10
10
10
10
-
CARB1
.90
1.57
.85
1.87
1.09
1.84
1.10
0.94
0.99
1.56
CARB2
.80
1.50
.77
1.85
1.01
1.81
0.75
0.82
0.83
1.49
GARB 3
.83
.74
.79
1.55
0.91
1.41
0.17
-0.17
0.67
0.88
CARBO
.84
1.42
.86
1.80
1.09
1.78
1.10
0.63
0.97
1.41
ADOM1
.74
1.92
.45
1.98
0.49
1.78
1.13
1.83
0.70
1.92
A DOM 2
-.01
-.19
.19
-.10
0.56
-1.04
-1.15
-1.61
0.48
0.74
ADOM3
-.22
-1.34
.19
-1.08
.23
-1.19
-1.23
-1.62
0.47
1.31
UAM1
.76
1.76
.61
1.91
0.86
1.89
0.91
1.48
.79
1.76
UAM2
.61
.61
.52
1.24
0.69
1.20
0.74
-0.25
0.64
0.83
ISC
1.80
2.00
1.72
2.00
1.77
2.00
1.99
2.00
1.82
2.00
-------
From Table 7-6 we can note the following:
1) The ADOM 2 model produces superior composite performance according to the
CFB regardless of LAI, producing a mixture of small over and underpredictions
on the average, as can be noted from the straddling of FBA between positives for
smooth surfaces (LAI < 3.0) and negatives for complex ones (LAI > 3.0).
2) The CFB indicates that all models tend to perform poorer under large LAI
situations with a consequent increase in the average magnitude of model
residuals (e.g., FBA).
3) Of all of the CARB models, the GARB 3 version performs best under both small
and large LAI conditions.
4) The ISC model consistently performs the worst of all of the models regardless of
LAI
Figure 7-4 indicates that the ADOM 2 model performs consistently better than all other
models while the ISC model consistently performs the worst. Examination of overall model
performance as shown in Figure 7-4 for CARB 3. ADOM 3, and UAM 2 indicates the use of
the LAI adjustment factor can produce mixed results. For large LAI cases, Table 7-6 indicates
that the CARB and UAM model formulations appears to be improved, due to a reduction in
the degree of overprediction. For ADOM 2 on the other hand, the addition of the LAI
adjustment to make ADOM 3 leads to consistent overpredictions as shown by the scatter plots
in Figure B-lg.
7.1.5 Stratification by Day vs Night
The particle deposition velocity is dependent on the degree of atmospheric turbulence
which in turn is dependent on the atmospheric stability. Atmospheric stability generally
undergoes a significant diurnal variation. At night turbulent transport is generally conducted
under neutral or stable conditions. Any day-night difference in performance is likely to be
directly connected with the aerodynamic resistance formulation utilized. From Table 7-7 which
summarizes the normalized and composite performance statistics, the following points can be
noted:
7-18
-------
Table 7-7
A summary of the fractional and normalized statistical measures for each of the models examined.
The first row is for night while the second is for day.
MODEL DESIGNATION
MEASURE
FBA
FBSD
FBLE
FBSE
CFB
SAMPLE
SIZE
42
51
42
51
10
10
10
10
-
-
CARB1
.99
1.38
.99
1.41
1.33
1.51
1.70
1.34
1.25
1.41
CARB2
.89
1.24
.90
1.21
1.25
1.29
136
1.03
1.15
1.19
CARS 3
1.01
1.27
1.09
1.72
1.38
1.67
1.25
0.36
1.18
1.26
CARBO
.99
1.00
.99
.70
1.33
0.98
1.70
1.34
1.25
1.00
ADOM1
.67
136
.49
1.40
0.60
1.40
1.64
131
0.85
1.47
ADOM2
.16
.33
.33
.65
0.67
036
-0.32
-0.22
0.37
0.44
ADOM3
.07
-.34
.38
-.14
0.65
-0.15
-0.62
-0.88
0.43
0.38
UAM1
.79
1.74
.78
1.88
1.11
1.86
133
1.37
1.05
1.71
UAM2
.74
1.49
.76
1.73
1.09
1.73
1.48
0.88
1.01
1.46
IS<
1.77
2.00
1.74
2.00
1.79
2.00
2.00
1.99
1.82
2.00
-------
1) The CFB indicates that during the night the ADOM 2 model is the best
performing model while during the day the ADOM 3 model is best. The ADOM
3 model produces a mixture of small over and underpredictions.
2) Of the core models the CFB indicates that ADOM 1 model appears to perform
the best in an overall sense, however, it performs distinctly better at night than
during the day. During the day the (FB shows the CARB model as performing
marginally the best of the 5 models.
Figure 7-4 suggests that again ADOM 2 is the best overall performing model while ISC
is consistently the worst.
7.1.6 Stratification by Friction Velocity
Friction velocity is related directly to the vertical turbulent transport of momentum. In
addition, the friction velocity plays a role in determining the laminar boundary layer near the
surface. Consequently the friction velocity is a relatively important determinant of deposition
velocity. We have stratified the small particle cases into high and low friction velocity sets with
a threshold set to divide the sample into halves. The model performance statistics are presented
in Table 7-8. From this table we can make the following observations based on CFB: (1) the
ADOM 2 model is the best performing model regardless of friction velocity, while ISC is the
worst, and (2) all core models perform essentially the same. Figure 7-4 essentially underscores
these two observations. Removal of roughness length limitation on CARB 1 does not seem to
significantly affect overall performance of the CARB 0 version.
7.1.7 Stratification by Temperature
The dependence of deposition velocity model performance on temperature was
examined. The observed data was broken up into 'hot' and 'cold' subsets based on a 17°C
threshold which was applied to split the overall data set up into two large subsets. While most
particle deposition algorithm do not have an explicit temperature dependence, the original
CARB formulation (CARB 1) does. The resulting model normalized performance statistics and
performance scores are summarized in Table 7-9. The resulting fractional bias and composite
performance measures indicate that:
1) The CFB indicates that ADOM 2 is the best performing model regardless of
temperature. The FBSE shows slight overprediction for small deposition
velocities while the FBLE shows underpredictions for large deposition velocities.
7-20
-------
Table 7-8
A summary of the fractional and normalized statistical measures for each of the models examined.
The first row is for friction velocities < 0.25 m/s while the second is for > 0.25 m/s.
MODEL DESIGNATION
MEASURE
FBA
FQSD
FBLE
FUSE
CFB
SAMPLE
SIZE
58
110
58
110
10
10
10
10
-
CARB1
1.14
1.06
.77
.84
0.73
1.14
1.33
0.96
0.99
1.00
CARB2
1.04
.96
.67
.76
0.63
1.06
1.03
0.80
0.89
0.90
CARS 3
.81
.80
.77
.82
0.74
0.96
0.57
0.23
0.72
0.70
CARBO
1.12
.97
.77
.86
0.76
1.14
1.33
0.%
0.99
0.98
ADOM1
1.52
.91
1.05
.41
0.49
1.57
1.26
1.83
1.29
0.86
A DOM 2
.12
-.10
.06
.23
-0.10
0.62
-0.61
-0.%
0.22
0.48
ADOM3
-.46
-.86
-.19
-.19
-0.98
-1.26
-0.01
0.14
0.41
0.61
UAM1
1.15
.96
.60
.59
0.57
0.91
1.24
1.00
0.89
0.87
UAM2
.80
.57
.51
.55
1.48
0.88
1.07
0.62
0.71
0.62
ISC
1.93
1.84
1.80
1.71
1.84
1.79
1.99
1.99
1.89
1.83
-------
Table 7-9
A summary of the fractional and normalized statistical measures for each of the models examined.
The first row is for temperatures < 290°K while the second is for > 290° K.
MODEL DESIGNATION
MEASURE
FBA
FBSD
FBLB
FBSE
CFB
SAMPLO
SIZE
62
106
62
106
10
10
10
10
-
CARB1
1.12
1.04
1.07
.65
0.90
0.92
1.16
1.09
1.06
0.93
CARB2
.97
.98
.95
.59
0.79
0.88
0.82
1.31
0.88
0.94
CARB3
1.09
.66
1.12
59
0.92
0.72
0.25
0.46
0.85
0.61
CARBO
.99
.99
1.06
.66
1.01
0.92
1.16
1.09
1.05
0.92
ADOM 1
.81
1.13
.48
.41
0.44
0.62
1.19
1.74
0.73
0.97
ADOM 2
.26
-.21
.36
.03
0.18
0.44
-0.76
-0.43
0.39
0.28
ADOM 3
.06
-1.04
.36
-.47
0.35
-0.22
-1.09
-1.18
0.47
0.73
UAM1
1.00
.99
.82
.41
0.60
0.71
0.98
1.55
0.85
0.91
UAM2
.93
.46
.79
.34
038
OJ1
0.81
0.98
0.78
0.60
ISC
1.82
1.87
1.74
1.70
1.69
1.76
1.99
2.00
1.81
1.83
-------
2) The CFB indicates UAM 2 and ADOM 3 are the second best performing
models, with UAM 2 producing moderate underpredictions for all fractional bias
measures regardless of temperature while ADOM 3 produces overpredictions for
most measures.
3) The CFB indicates that of all the CARB models only the CARB 3 model displays
a marginal improvement over the core CARB 1 model.
4) All measures consistently show the ISC model as the worst performing model
regardless of temperature.
7.1.8 Estimation of CPM from Tables
The estimate of a CPM can be conducted from the statistics presented in the tables.
The CFB's for CARB 1 will serve as an illustrative example. Table 7-10 illustrates how one can
extract fractional bias information from the tables to estimate the CPM.
7.1.9 Selection of the Recommended Deposition Model
The previously defined composite statistical measures, the CPM and MCM and their
associated 95% confidence intervals will now be used to select the best performing models. The
CPM was used to first rank the models with the smallest CPM being designated as the best
performing model. Figure 7-5 shows the models ranked by CPM. The ADOM 2 model has the
smallest CPM while the ISC model has the largest. The 95% confidence intervals are also
plotted. These confidence intervals suggest that ADOM 2 appears to significantly out perform
ADOM 3. The confidence intervals also indicate that many of the models performances are
indistinguishable. The confidence intervals in Figure 7-5 indicate that with the exception of the
ISC model, all core models (ADOM 1, CARB 1, and UAM 1) appear likely to have the same
performance. The confidence interval for the ISC model is narrowed due to the many zero
predictions which act to reduce the effective number of degrees of freedom.
In order to actually tell if the performance of two models is significantly different the
MCM between two models is estimated. The MCM for all model pairings was estimated, along
with the 95% confidence interval. The MCM's were ranked, and all MCM's were displayed with
confidence intervals in Figure 7-6. The most important feature of Figure 7-6 is the zero line of
the absolute value of the MCM. Any models whose confidence interval crosses the zero line are
not significant, thus the two model's composite model performance is not significantly different.
7-23
-------
Table 7-10
Summary of Composite Statistical Measures that Illustrate how the CPM
Arises for the CARB 1 Model
Stratification
Source
CFB
Particle Size
small
large
Roughness Length
< 25
> 25
Leaf Area Index
< 3
> 3
Sunlight
night
day
Friction Velocity
< 25 cm/s
> 25 cm/s
Temperature
< 290°K
> 290°K
Average
CPM0
CPM
Table 7-4
Table 7-5
Table 7-6
Table 7-7
Table 7-8
Table 7-9
Table 7-3
1.27
1.01
1.01
1.53
0.99
1.56
1.25
1.41
0.99
1.00
1.06
0.93
1.17
0.99
1.08 (Table 7-3 = 1.08)
7-24
-------
cu
o
0)
en
03
(U
0)
a
G
a:
s
o
si-.
-
-------
N>
Aden 1
U»> 1
C.rk 0
U.rD 9
U.o 1
lluu 1
Cub 0
Adon 1
Uuu 1
Uu> a
C.rk a
Adam 1
Adon 3
Adorn t
C.rb 9
Uun t
U«l» >
Adorn 3
Vim 1
C.rb 0
Dun a
U.m 9
Adorn 3
C.rh 9
AJom 9
Uua 8
U.m a
Adorn a
UUD 1
Adou 9
Adon 3
Adorn 8
Adou t
Uuu 1
Adon a
Adoma
ISC
ICC
ISC
ISC
ISC -
ISC
ISC
ISC -
ISC
- c.ik a
- Cftrb 0
- C.rb a
- C»rb 9
- C.rb t
- Adotu 1
- C.rb 1
- C.ib 0
- C.rb
- Adoin
- C.rb
- C.rb
- C.rb
- C.rk
- C.rb
- Ad out
- C.rb
- Adpui
- C.rb
- C.rb
- U»u*
- C.lb
- Ad*i»
- C.rb
- C.rb
- Adoua
- C.rb
- C.rti
- Adoui
- C.rb
- C.rb
- Adam
- C.rb
- Adoui
- C.rt,
- C.rb
- C.rb
- C.rb
- Uwu
- C.rb
- Adou)
- C.rb
- U.t,
. Adwu 3
- Adou 1
-0.5
0
0.5 1
Absolute Value of MCM
1.5
Figure 7-6. A summary of the MCM for each unique model pair.
-------
The difference between ADOM 2 and all other models appears to be significant, thus ADOM 2
can be claimed as a clear cut winning model for our evaluation data base. CARS 1, UAM 1,
and ADOM do not appear to be significantly different at the 95% confidence level.
The Cumulative Distribution Function (CDF) plots presented in Appendix B illustrate
how the models tend to perform over the range of deposition velocities. Figure B-2a in the
appendix shows the performance of the core models. The curves indicate that all models
underpredict deposition velocities less than 0.2 cm/s, but did much better for larger deposition
velocities. The ISC model is the notable exception. The CDF of the CARB series of models
are illustrated in Figure B-2b. For small deposition velocities the CARB 3 hybrid model
appears to track the observed CDF most closely. However for deposition velocities greater than
0.5 cm/s there seems to be no appreciable difference between the model variants. Figure B-2c
* f
shows the closely matching CDF's of the observations and ADOM 2 and the poor match
obtained with ISC. The improvement in UAM due to the addition of the LAI influence is best
illustrated by Figure B-2d. The overprediction of the UAM is significantly lessened by the
UAM 2 model, although both models perform similarly for the largest deposition velocities.
7.2 Sensitivity Calculation
Aerosol observations such as those reported by Hidy (1984) and Richards et al. (1989)
generally exhibit a multipeaked mass distribution as a function of mean particle diameter.
Peaks have been observed in the mass fraction at 0.2, 1-2, and 6-10 microns. Each of these
peaks are associated with a particular pathway of particle emission and/or formation. In the
present study a uniform distribution of particle mass fraction between 0.1 and 1.0 microns was
utilized for sulfate.
Most of the particle experiments observed aged sulfate aerosol with a mass fraction peak
at 0.2 microns. A question was raised as to the robustness of the model evaluation statistics
with changing assumptions of the size distribution of the mass fraction. As a sensitivity test the
model performance evaluation exercise was repeated with a size distribution taken from
Richards et al. (1989) and which is summarized in Table 7-11. The mass fraction peak occurs at
0.2 microns and is nearly twice the value of the uniform distribution for a 0.1 micron size range.
The results of the sensitivity study indicates that the results do not change significantly.
The CPM's are summarized in Table 7-12. The CPM fot each model varies slightly. The
variation in CPM between the two distributions lies well within the 95% confidence interval also
7-27
-------
Table 7-11
Aerosol Mass Fraction as a Function of Size Distribution for Two
Assumed Aged Sulfate Distributions
Diameter
(Microns)
0.10
0.13
0.16
0.19
0.23"
0.29
0.36
0.44
0.54
0.66
0.81
1.00
Mass Fraction
(Uniform)
0.083
0.083
0.083
0.083
0.083
0.083
0.083
0.083
0.083
0.083
0.083
0.083
Mass Fraction
(Richards et al. 1989)
0.053
0.095
0.144-
0.189
0.161
0.111
0.053
0.046
0.053
0.062
0.035
0.000
7-28
-------
Table 7-12
Summary of Composite Model Performance Statistics for
Two Assumed Sulfate Size Distributions
Model
CARB 1
CARB2
CARB 3
CARBO
ADOM 1
ADOM2
ADOM 3
UAM1
UAM2
ISC
CPM
(Uniform
Distribution)
0.50 ± 0.09
0.42 ±0.11
0.32 ± 0.12
0.49 ± 0.09
0.37 ± 0.07
0.22 ± 0.08
0.31 ± 0.07
0.41 ± 0.11
0.30 ±0.11
0.92 ± 0.01
Rank
9
7
4
8
5
1
3
6
2
10
CPM
(0.2 urn Peak
Distribution)
0.53 ± 0.09
0.46 ±0.10
0.35 ±0.11
0.52 ± 0.08
0.35 ± 0.09
0.21 ± 0.08
0.30 ± 0.07
0.47 ± 0.09
0.37 ±0.11
0.92 ±0.11
Rank
9
6
4
8
3
1
2
7
5
10
7-29
-------
presented in Table 7-12. The sensitivity test suggests that variations in the submicron aged
suifate distribution will not alter the selection of the best model.
7.3 Summary of Results
The model performance analysis was conducted using four fractional bias measures:
FBA, FBSD, FBLE, FBSE. These measures we're combined into composite fractional bias
measures (CFB) for the entire data set and 12 subsets. The CFB's were then averaged over all
subsets and the resulting average CFB was averaged with the CFB for the entire data to provide
a composite performance measure, CPM. Differences were then taken of the CPM between .
pairs of models to produce a model comparison measure (MCM). The MCM's were tested to
determine if the model differences were significant from zero at the 95% confidence level.
Model selection was based on two factors: (1) the CPM (ranked from smallest to largest) and
(2) the significance of the MCM at the 95% level.
The composite performance measures such as the CFB and the CPM indicate that the
ADOM 2 model is the best performing model and ISC is the worst performing. The MCM's
indicate that these two models perform significantly differently from all other models at the 95%
confidence level. The MCM's also indicate that differences in model performance as indicated
by CPM's between the core models CARB 1, ADOM 1, and UAM 1 are not significant.
Modifications to the core models produced mixed results. The removal of temperature
dependence and roughness length clipping did not appear to produce any significant
modifications in model performance for the CARB models. The alteration to the CARB 1 core
model which produced the greatest improvement was the inclusion of a Leaf Area Index (LAI)
dependence in CARB 3. The inclusion of LAI dependence in UAM 1 produced a significant
improvement in performance for small deposition velocities by reducing underprediction. The
LAI modification did not improve ADOM 2 performance. Inclusion of the LAI dependence
actually produces significant overprediction of the deposition velocities from ADOM 3.
The most successful alterations to the ADOM 1 deposition model which produce
ADOM 2 are those involving the modified Stokes/Schmidt relations. The resulting ADOM 2
model shows significantly better performance as measured by the CPM and MCM over all other
models. The uniqueness of the ISC deposition model allowed no comparable types of
modifications to be made and tested.
7-30
-------
8. Summary and Conclusions
The purpose of this study was to review, refine, and test dry deposition techniques that
are suitable for use in regulatory models such as the Industrial Source Complex (ISC) model.
Dry deposition is the process by which particulate matter and gaseous pollutants are transferred
from the air to land, water, and vegetative surfaces through "dry" (i.e., non-precipitation)
mechanisms. Because indirect risk assessment pathways such as fish, food chain, and water
ingestion, commonly dominate total intake and exposure to many pollutants, an accurate
estimate of dry deposition is an important element of many regulatory analyses. Dry deposition
may also be important for a refined estimate of air quality concentrations for sources subject to
significant plume depletion.
The dry deposition flux can be written as F = x vd» where F is the flux (g/m2/s), x is the
ambient pollutant concentration (g/m3), and vd is the deposition velocity (m/s), all defined at a
reference height. Standard procedures can be used for estimating the concentration term of the
flux equation, with appropriate modifications made to account for plume depletion effects. The
main focus of this study was the testing and evaluation of various methods for computing the
deposition velocity.
A review of the technical literature identified several models that are suitable for
predicting the dry deposition velocity within the framework of a regulatory model. These
models are listed in Section 1 and described in more detail in Section 2. Three resistance-based
particle deposition models were identified which fit the required criteria of this study (i.e.,
methods of sound "technical basis that are suitable for regulatory use for both large and small
particles). The technical literature suggested that certain parameterizations in these models
could be improved. Therefore, several modifications and enhancements to the core models were
developed and tested in this study. As a result, a total of ten deposition velocity models were
evaluated (see Section 7).
A second literature review identified observational data sets which could be used to test
deposition velocity algorithms. Based on this review, eight datasets for particulate matter and
fourteen datasets for gases were assembled in a database. Although the ultimate goal is to
evaluate dry deposition for both particulate matter and gases, only particulate matter deposition
was evaluated in this study. Appendix C lists the observational particle deposition velocity
datasets. One of the recommendations of the study is that additional evaluation efforts be
conducted to test the dry deposition models for gases (see Section 2.2) with the datasets listed in
Section 7.
8-1
-------
As explained in Section 2.1, large particles (i.e., above ~ 20 um diameter for unit density
particles) tend to be dominated by gravitational settling effects. The concept of gravitational
settling is incorporated into the deposition velocity relationship described in Section 2.1.3 as well
as the reflection coefficient scheme used in the current ISC model. Particles in the size range
from 1.0 to 20.0 u.m diameter are significantly influenced by inertial effects, which enhance the
rate of deposition over that obtained by considering gravitational settling alone. The deposition
of very small particles ( < ~ 0.1 |im diameter) are dominated by Brownian diffusion. This
process increases in importance as the size of the particles decreases. Particles in the size range
from 0.1 to 1.0 um diameter show a minimum in the deposition velocity because they are not
efficiently deposited by any of the processes described above. Although the deposition velocity
database consists of particles in the range from 0.1 to 20 um diameter sizes, the resistance-based
modeling techniques tested in this study apply to larger particle as well. For all of the models,
the deposition velocity approaches the same gravitational settling velocity as the size of the
particle becomes large. Therefore, the recommended deposition model is considered to be
appEcable to the full range of particle sizes of interest as might be encountered in typical
regulatory studies.
Two related components necessary for a complete deposition model are (1) a method for
tracking mass conservation and plume depletion, and (2) a meteorological module for estimating
the micrometeorologicai parameters required by the deposition model. In Section 3, four
algorithms for computing plume depletion (source depletion, surface depletion, K-theory
method, and modified source depletion) were reviewed. As discussed in Section 3, the modified
source depletion model of Horst (1983) is recommended as the overall best approach for use in
a regulatory model. Although evaluations of plume depletion algorithms in the literature against
field data are very limited, one such study (the dual tracer study of Doran and Horst, 1985) and
intercomparisons of the various techniques with the reference surface depletion method support
the use of the modified source depletion technique. This algorithm is computationally efficient,
conserves mass, and can account for gravitational settling effects. In Appendix E,
implementation issues associated with the use of the modified source depletion method are
discussed.
Methods suitable for estimating the necessary micrometeorological parameters for the
dry deposition model are outlined in Section 4. As required for regulatory applications, these
data must be obtained from routinely available observations. In particular, the dry deposition
models require an estimate of the surface friction velocity (u.) and the Monin-Obukhov length
(L). The meteorological literature contains several techniques for estimating these input
parameters. The techniques selected here have been shown to produce reasonable results.
Although other mathematical relationships may eventually be used when the deposition
8-2*
-------
algorithm is incorporated in the ISC model, the effects of the change are likely to be minimal
since experience shows that differences among the most commonly-used techniques are small for
most conditions.
An objective model evaluation methodology was used to distinguish between the
performance of the various models for predicting particle deposition velocities. Only those
models deemed from the scientific review to parameterize the major known processes affecting
deposition of small and large particles, as discussed in Section 2, were considered for
recommendation as the preferred model. The model evaluation approach, discussed in Section
5, is based on the EPA's statistical model evaluation protocol. This approach was used because
it has been successfully demonstrated for many other regulatory model evaluation studies.
The results of the model evaluation indicates that a dry deposition algorithm based on
Pleim et al. (1984) and Slum (1982) performed the best against the observational data available
for this study. This algorithm is called ADOM 2 in Section 2.1.3. The statistical model
comparison measure (discussed in Section 7.2) indicates that this model performed significantly
better than the other models tested at the 95% confidence level.
The recommended procedures for computing the deposition velocity, plume depletion,
and meteorological variables have been implemented in a revised version of the ISC2 dispersion
model and a companion meteorological processor. Modified versions of both the short term
(ISC2ST) and long term (ISC2LT) models and the meteorological processor will be made
available through the EPA's SCRAM bulletin board system. Draft revisions to the user's guide
and model formulation documents will also be made available for the purposes of public review
and comment.
In future work, it is recommended that an analysis be made to compare the revised
version of the ISC model to the previous version of the model to determine likely changes in
modeled design concentrations. It is also recommended that some additional analysis be
conducted to examine the combined sensitivity of. the recommended deposition velocity model
and the modified source depletion model to various input variables. This can be done within
the new ISC model, since both models have been included in the revised code. For example,
the relative sensitivity of the deposition fluxes to the particle size distribution, particle density,
surface characteristics (e.g., surface roughness) and meteorological conditions should be
assessed.
8-3
-------
9. References
Auer, A.H. Jr., 1978: Correlation of land use and cover with meteorological anomalies. J. AppL
Meteor., 17, 636-643.
Baldocchi, D.D., B.B. Hicks and P. Camara, 1987: A canopy stomatal resistance model for
gaseous deposition to vegetated surfaces. Atmos. Environ., 21, 91-101.
Bowers, J.F., J.R. Bjorklund and C.S. Cheney, 1979: Industrial Source Complex (ISC)
Dispersion Model User's Guide. Volume I. EPA-450/4-79-030, U.S. Environmental
Protection Agency, Research Triangle Park, NC.
Businger, J.A., J.C. Wyngaard, Y. Izumi and E.F. Bradley, 1971: Flux-profile relationships in the
atmospheric surface layer./. Atmos. ScL, 28, 181-189.
Chamberlain, A.C., 1953: Aspects of travel and deposition of aerosol and vapor clouds. Atomic
Energy Research Establishment, HP/R 1261.
Cleveland, W.S. and R. McGill, 1984: Graphical Perception: Theory, Experimentation, and
Application to the Development of Graphical Methods. /. Am. Stat. Assoc., 79, 531-444.
Cox, W.M. and JA. Tikvart, 1990: A Statistical Procedure for Determining the Best Performing
Air Quality Simulation Model. Atmos. Environ., 24, 2387-2395.
Davies, T.D. and J.R. Mitchell, 1982: Dry deposition of sulfur dioxide onto grass in rural
eastern England (with some comparisons with other forms of sulfur deposition).
Proceedings of the Fourth International Conference on Precipitation Scavenging, Dry
Deposition, and Resuspension, Volume 2, Santa Monica, CA, 29 November-3 December.
DeBruin, H.A.R. and A.A.M. Holtslag, 1982: A simple parameterization of the surface fluxes of
sensible and latent heat during daytime compared with the Penman-Monteith concept. /.
dim. AppL Meteor., 21, 1610-1621.
Doran J.C. and T.W. Horst, 1985: An evaluation of Gaussian plume-depletion models with
dual-tracer field measurements. Atmos. Environ., 19, 939-951.
9-1
-------
Dumbauld, R.K, I.E. Rafferty and H.E. Cramer, 1976: Dispersion-deposition from aerial spray
releases. Proceedings of the Third Symposium on Atmospheric Turbulence, Diffusion, and
Air Quality, October 19-22, Raleigh, NC.
Fowler, D. and J.N. Cape, 1982: Dry deposition of SO2 onto a Scots pine forest. Proceedings of
the Fourth International Conference on Precipitation Scavenging, Dry Deposition, and
Resuspension, Volume 2, Santa Monica, CA, 29 November-3 December.
Garland, JA, 1982: Dry deposition of small particles to grass in field conditions. Proceedings of
the Fourth International Conference on Precipitation Scavenging, Dry Deposition, and
Resuspension, Volume 2, Santa Monica, CA, 29 November-3 December.
Godowitch, J.M., 1990: Vertical ozone fluxes and related deposition parameters over
agricultural and forested landscapes. Boundary-Layer MeteoroL, 50, 375-404.
Gray, HA., M.P. Ligocki, G.E. Moore, CA. Emery, R.C. Kessler, J.P. Cohen, C.C. Chang, S.I.
Balestrini, S.G. Douglas, R.R. Schulhof, J.P. Killus, C.S. Burton, 1991: Deterministic
Modeling in the Navajo Generating Station Visibility Study. Volume II. Appendix E
(Description of deposition algorithms). Systems Applications, International, San Rafael,
CA.
Hanna, S.R. and J.C. Chang, 1990: Modification of the Hybrid Plume Dispersion Model
(HPDM) for urban conditions and its evaluation using the Indianapolis data set. Vol.
HI. Analysis of urban boundary layer data. Sigma Research Corp., Concord, MA.
Hanna, S.R. and J.C. Chang, 1991a: Modification of the Hybrid Plume Dispersion Model
(HPDM) for urban conditions and its evaluation using the Indianapolis data set. Vol. I.
User's guide for HPDM-Urban. Sigma Research Corp., Concord, MA.
Hanna, S.R. and J.C. Chang, 199 lb: SIGPRO - A meteorological preprocessor for dispersion
model applications to stack plumes in urban areas. Seventh Joint AMS-AWMA Conf.
on Appl. of Air Poll. Meteor., New Orleans, LA, Jan., 1991.
Hanna, S.R. and J.C. Chang, 1992: Boundary-layer parameterizations for applied dispersion
modeling over urban areas. Boundary-Layer Meteorology, 58, 229-259.
9-2
-------
Harrison, R.M., S. Rapsomanikis and A. Turnbull, 1989: Land-surface exchange in a
chemically-reactive system: Surface fluxes of HNO3, HC1 and NR3. Atmos. Environ., 23,
1795-1800.
Hicks, B.B., 1982: Critical assessment document on acid deposition. ATDL Contrib. File No.
81/24, Atmos. Turb. and Diff. Laboratory, Oak Ridge, TN.
Hicks, B.B., D.D. Baldocchi, T.P. Meyers, R.P. Hosker, Jr. and D.R. Matt, 1987: A preliminary
multiple resistance routine for deriving dry deposition velocities from measured
quantities. Water, Air, and Soil Poll, 36, 311-330.
Hicks, B.B., D.R. Matt and R.T. McMillan, 1989: A micrometeorological investigation of
surface exchange of trace gases: A case study. NOAA Tech. Memo. ERL ARL-172, Air
Resources Laboratory, Silver Spring, MD.
Hicks, B.B., M.L. Wesely, R.L. Coulter, R.L. Hart, J.L. Durham, R. Speer and D.H. Stedman,
1986: An experimental study of sulfur and NOX fluxes over grassland. Boundary-Layer
MeteoroL, 34, 103-121.
Hidy, G.M., 1984: Aerosols: An Industrial and Environmental Science. Academic Press, Inc.,
New York, NY.
Hjelmfelt, M.R., 1982: Numerical simulation of the effects of St. Louis on mesoscale
boundary-layer airflow and vertical air motion: Simulations of urban vs. non-urban
effects. /. AppL Meteor., 21, 1239-1257.
Holtslag A.A.M. and A.P. van Ulden, 1983: A simple scheme for daytime estimates of the
surface fluxes from routine weather data. /. dim. and AppL Meteor., 22, 517-529.
Horst, T.W., 1977: A surface depletion model for deposition from a Gaussian plume. Atmos.
Environ., 11, 41-46.
Horst, T.W., 1983: A correction to the Gaussian source-depletion model. In Precipitation
Scavenging, Dry Deposition and Resuspension, H.R. Pruppacher, R.G. Semonin, W.G.N.
Slinn, eds., Elsevier, NY.
Horst, T.W., 1984: The modification of plume models to account for dry deposition. Boundary-
Layer Meteor., 30, 413-430.
9-3
-------
Hosker, R.P. and S.E. Lindberg, 1982: Review: Atmospheric Deposition and Plant Assimilation
of Gases and Particles. Atmos. Environ., 16, 889-910.
Huebert, B.J., 1982; Measurements of the dry-deposition flux of nitric acid vapor to grasslands
and forest. Proceedings of the Fourth International Conference on Precipitation Scavenging,
Dry Deposition, and Resuspension, Volume 2, Santa Monica, CA, 29 Nov.-3 Dec.
Lorenz, R. and C.E. Murphy, Jr., 1989: Dry deposition of particles to a pine plantation.
Boundary-Layer MeteoroL, 46, 355-366.
McDonald, I.E., 1960: An aid to computation of terminal fall velocities of spheres. /. Met, 17,
463.
McMillen R.T., D.R. Matt, B.B. Hicks and J.D. Womack, 1987: Dry deposition measurements
of sulfur dioxide to a spruce-fir forest in the Black Forest: A data report. NOAA Tech.
Memo. ERL ARL-152, Air Resources Laboratory, Silver Spring, MD.
Meyers, T.P., 1987: The sensitivity of modeled SO2 fluxes and profiles to stomatal and boundary
layer resistances. Water, Air, and Soil Poll, 35, 261-278.
Meyers, T.P. and D.D. Baldocchi, 1988: A comparison of models for deriving dry deposition
fluxes of O3 and SO2 to a forest canopy. Tellus, 40B, 270-284.
Meyers, T.P., B.J. Huebert, and B.B. Hicks, 1989: HNO3 deposition to a deciduous forest.
Boundary-Layer MeteoroL, 49, 395-410.
Moller U. and G. Schumann, 1970: Mechanisms of transport from the atmosphere to the
earth's surface./. Ceophy. Res., 75, 3013-3019.
Nicholson, K.W. and TD. Davies, 1987: Field Measurements of the dry deposition of
particulate sulphate. Atmos. Environ., 21, 1561-1571.
Oke, T.R., 1978: Boundary Layer Climates. John Wiley & Sons, New York, NY.
Oke, T.R., 1982: The energetic basis of the urban heat island. Quart. J.R. Meteor. Soc., 108,
1-24.
9-4
-------
Overcamp, T.J., 1976: A general Gaussian diffusion-deposition model for elevated point
sources. J. AppL Meteor., 15, 1167-1171.
Padro, J., G. den Hartog, and H.H. Neumann, 1991: An investigation of the ADOM dry
deposition module using summertime O3 measurements above a deciduous forest.
Atmos. Environ., 25A, 1689-1704.
Plate, E. and A.A. Quraishi, 1965: Modeling of velocity distribution inside and above tall crops.
/. AppL MeteoroL, 4, 400-408.
Pleim, J., A. Venkatram and R. Yamartino, 1984: ADOM/TADAP model development
program. Volume 4. The dry deposition module. Ontario Ministry of the Environment,
Rexdale, Ontario.
Rao, K. S., 1981: Analytical solutions of a gradient-transfer model for plume deposition and
sedimentation. NOAA Tech. Memo. ERL ARL-109, Air Resources Laboratory, Silver
Spring, MD.
Richards, L.W., J.A. Anderson, D.L. Blumenthal, J.A. McDonald, P.S. Bhardwaja, R.B.
Candelaria and D.W. Moon, 1989: Nitrogen and sulfur chemistry and aerosol formation
in a western coal-fired power plant plume. AWMA/EPA International Specialty
Conference. Ester Park, CO, October, 1989.
Scire J.S., F.W. Lurmann, P. Karamchandani, A. Venkatram, R. Yamartino, J. Young and J.
Pleim, 1986: ADOM/TADAP model development program. Volume 9. ADOM/TADAP
User's Guide. Ontario Ministry of the Environment, Rexdale, Ontario, Canada.
Scire, J.S., D.G. Strimaitis and R.J. Yamartino, 1990: Model formulation and user's guide for
the CALPUFF dispersion model. Sigma Research Corp., Concord, MA.
Scire, J.S. and D.L. Wojichowski, 1987: Modeling deposition and dispersion in an urban
environment. MA.S.S.-APCA 33rd Anniv. Technical Conference, November 3-6,
Atlantic City, NJ.
Sehmel, GA., 1980: Particle and gas dry deposition: A review. Atmos. Environ., 14, 983-1011.
Sehmel, G.A., 1984: Deposition and Resuspension. Chapter 12 in Atmospheric Science and
Power Production, DOE/TIC-27601. U.S. Department of Energy, D. Randerson, Ed.
9-5
-------
Sehmel, GA. and W.H. Hodgson, 1978: A model for predicting dry deposition of particles and
gases to environmental surfaces. PNL-SA-6721, Battelle Pacific Northwest Laboratory.
Sehmel, G A. and S.L. Sutler, 1974: Particle deposition rates on a water surface as a function of
particle diameter and air velocity. /. Rechs. Atmos., Ill, 911-918.
Shieh, CM., M.L.Wesely and C.J. Walcek, 1986: A dry deposition module for regional acid
deposition. EPA/600/3-86/037, U.S. Environmental Protection Agency, Research
Triangle Park, NC.
Slinn, W.G.N., L. Hasse, B.B. Hicks, A.W. Hogan, D. Lai, P.S. Liss, K.O. Munnich, G.A. Sehmel
and O. Vittori, 1978: Some aspects of the transfer of atmospheric trace constituents past
the air-sea interface. Atmos. Environ., 12, 2055-2087.
Slinn, SA. and W.G.N. Slinn, 1980: Predictions for particle deposition on natural waters.
Atmos. Environ., 14, 1013-1016.
Slinn, W.G.N., 1982: Predictions for particle deposition to vegetative canopies. Atmos. Environ.,
16, 1785-1794.
van Ulden, A.P. and AA.M. Holtslag, 1985: Estimation of atmospheric boundary layer
parameters for diffusion applications. /. AppL Meteor., 24, 1196-1207.
Walcek, CJ., RA. Brost, J.S. Chang and M.L. Wesely, 1986: SO2, sulfate, and HNO3 deposition
velocities computed using regional land use and meteorological data. Atmos. Environ.,
20, 949-964.
Wang, I.T. and P.C. Chen, 1980: Estimations of heat and momentum fluxes near the ground.
Proc. 2nd Joint Conf. on Applications of Air Poll Meteor., American Meteorological
Society, Boston, MA, 764-769.
Weil, J.C. and R.P. Brower, 1983: Estimating Convective Boundary Layer Parameters for Diffusion
Application. Draft Report prepared by Environmental Center, Martin Marietta Corp.,
for Maryland Dept. of Natural Resources.
Wesely, M.L., 1989: Parameterization of surface resistances to gaseous dry deposition in
regional-scale numerical models. Atmos. Environ., 23, 1293-1304.
9-6
-------
Wesely, ML., D.R. Cook, and R.M. Williams, 1981: Field measurement of small ozone fluxes to
snow, wet bare soil, and lake water. Boundary-Layer MeteoroL, 20, 459-471.
Wesely, M.L., D.R. Cook and R.L. Hart, 1983: Fluxes of gases and particles above a deciduous
forest in wintertime. Boundary-Layer MeteoroL, 27, 237-255.
Wesley, M.L., D.R. Cook, R.L. Hart, B.B. Hicks, J.L. Durham, R.E. Speer, D.H. Stedman and
RJ. Tropp, 1982: Eddy-correlation measurements of the dry deposition of paniculate
sulfur and submicron particles. Proceedings of the Fourth International Conference on
Precipitation Scavenging, Dry Deposition, and Resuspension, Volume 2, Santa Monica, CA,
29 November-3 December.
Wesely, M.L., JA. Eastman, D.R. Cook and B.B. Hicks, 1978: Daytime variations of ozone
eddy fluxes to maize. Boundary-Layer MeteoroL, 15, 361-373.
Wesely, M.L. and B.B. Hicks, 1977: Some factors that affect the deposition rates of sulfur
dioxide and similar gases on vegetation. J. Air Poll. Control Assoc., 27, 1110-1116.
Winges, K.D., 1990: User's guide for the Fugitive Dust Model (FDM) (revised). Volume 1:
User's Instructions. EPA-910/9-88-202 R. U.S. EPA, Region 10, Seattle, WA.
Winges, K.D., 1992: Personal communication to J. Scire.
Yamartino, R.J., J.S. Scire, S.R. Hanna, G.R. Carmichael and Y.S. Chang, 1992: The
CALGRID mesoscale photochemical grid model. Volume I. Model formulation.
Atmos. Environ., 26A, 1493-1512.
9r7
-------
Appendix A
Estimation of ISC Deposition Velocity
-------
Estimation of ISC Deposition Velocity
The basis for the present ISC deposition algorithm is found in Dumbauld et al. (1976)
and in Overcamp (1976). In this approach the particles are assumed to move towards the
ground with a total velocity equal to the sum of the gravitational settling velocity and an average
turbulent velocity which determines the rate of plume spreading. This turbulent velocity is given
by
(A-l)
where u is the stack height wind speed, He is the effective plume height, vg is the gravitational
settling velocity, x is the downwind distance, and the vertical dispersion coefficient, oz, is given
by the relation:
az = Ax1
,B
(A-2)
The coefficients A and B are stability dependent, and are treated as average values in the ISC
deposition model. By differentiating Equation A-2 and substituting into Equation A-l we have:
-
vt-B
(A-3)
The turbulent velocity is thus a function of the ratio of the plume centerline height to the
downwind distance. For small particles, the uHe/x term is much larger than the settling velocity
which can be ignored.
In nearly all of the small particle experiments there is no specific plume-receptor
information in order to directly estimate an He/x. Furthermore, in some experiments it is
possible that several sources may be contributing to the deposition fluxes. How then does one
estimate an He/x that will be appropriate and consistent with the information provided to the
other deposition velocity models?
The other deposition models estimate a turbulent velocity near the surface as being
equal to the inverse of the aerodynamic resistance, ra. Thus the He/x term can be estimated
from the relation:
A-l
-------
\
HJx = * Y"*
(A-4)
where the aerodynamic resistance is given by:
(A-5)
which is a formulation common to the ADOM and UAM models. Using the information on
Pasquill Gifford Turner stability class and the friction velocity and Monin-Obukhov length, and
assuming a reference height of 10 m, a displacement plane height of zero and using the given
surface roughness length and wind speed we estimated the equivalent He/x needed by the ISC
model.
The actual deposition velocity used in the ISC model is, following Overcamp (1976),
equal to:
1 - g 1
(A-6)
The settling velocity used in ISC, v is given by the Stokes relation:
(A-7)
where p is the particle density, g is the acceleration due to gravity, d is the particle diameter,
and u is the absolute viscosity of air (u - 1.83 x 10^* g cm"1 s"1).
The reflection coefficient, a, in Equation A-6 is the fraction of the image plume source
remaining. In the limit of a fully reflecting plume the image plume experiences no depletion
and a approaches one. The Dumbauld et al. (1976) paper indicates that when the settling
velocity drops below 0.1 cm/s the reflection coefficient is set equal to 1. As a result, ISC will
predict a zero deposition velocity for many of the cases in the small particle dataset. This limit
on deposition velocity especially affects sulfate particulate matter since the size range for such
particulate matter peaks in the submicron diameter range.
A-2
-------
Appendix B
Supplemental Graphics
-------
Series 1: Scatter plots of Observed Versus Model
Predicted Deposition Velocities
-------
Key
n CARB 1
1E-4
TT
1E-3 1E-2
Observations
1E-1
1EO
1E1
Figure B-la. Scatter plot of observed deposition velocity (cm/s) versus model predicted
deposition velocity (cm/s) for the complete small particle data set.
B-l
-------
Key
CARS 2
1E-3 1E-E
Observations
1E-1
1EO
1E1
Figure B-lb. Scatter plot of observed deposition velocity (cm/s) versus model predicted
deposition velocity (cm/s) for the complete small particle data set.
B-2
-------
W z
c
o
1I
-)->
o
3
m
i linn) 1r i i mil 1i i i mi 1i i i in
1E-3 1E-2 1E-1 1EO
Observations
Key
CARS 3
in
1E1
Figure B-lc. Scatter plot of observed deposition velocity (cm/s) versus model predicted
deposition velocity (cm/s) for the complete small particle data set.
B-3
-------
CK
f
n
Cd
Q. CA
a I
O 5
£2 n>
« -i
STL
3 o
ff
<. <
co n
s-' CL
01 5*
83 Q
3 i'
5
Q.
1E-4
Predictions
1E-3 1E-2
1EO
1E1
I
x
0)
o
o>
CL
s.
cd
o
-------
Key
H ADOM 1
E-4
i i i 1 1 1 i
1E-3 1E-2
Observations
1E-1
Figure B-le. Scatter plot of observed deposition velocity (cm/s) versus model predicted
deposition velocity (cm/s) for the complete small particle data set.
B-5
-------
Key
I] ADOM 2
E-4
1E-3 1E-2
Observations
TTT 1I MINI
1E-1 1EO
Figure B-lf. Scatter plot of observed deposition velocity (cm/s) versus model predicted
deposition velocity (cm/s) for the complete small particle data set.
B-6
-------
Key
CD ADOM 3
i i 1111 r i i i mi i
1E-3 1E-2
Observations
Figure B-lg. Scatter plot of observed deposition velocity (cm/s) versus model predicted
deposition velocity (cm/s) for the complete small particle data set.
B-7
-------
Key
m UAM i
1I I I 1111] 1I I I Illl) TT
lE-4 1E-3 1E-2
Observations
1E-1
1EO
1E1
Figure B-lh. Scatter plot of observed deposition velocity (cm/s) versus model predicted
deposition velocity (cm/s) for the complete small particle data set.
B-8
-------
Key
CD UAM 2
1E-3 1E-2
Observations
1E-1
1EO
1E1
Figure B-li. Scatter plot of observed deposition velocity (cm/s) versus model predicted
deposition velocity (cm/s) for the complete small particle data set.
B-9
-------
o
w-
I
M
O
r-1
-»->
O
I.
w
i ii rmT r r i 111 ii i i MINI i r
"1E-4 1E-3 1E-2 1E-1
Observations
I I I II I IMTIU
1EO 1E1
Key
ISC
Figure B-lj. Scatter plot of observed deposition velocity (cm/s) versus model predicted
deposition velocity (cm/s) for the complete small particle data set.
B-10
-------
Series 2: Cumulative Distribution Function (CDF) Plots for Observations
and Model Predictions of Deposition Velocity
-------
Key
D OBSERVED
O CARB 1
A ADOM 1
-f UAM 1
X ISC
1E1
Deposition Velocity
Figure B-2a. Cumulative prob'ability plot of deposition velocity (cm/s) using
the complete small particle data set.
-------
-a
1.00
0.20
Probability of
0.40
occurence
0.60
0.80
1.00
X4-0
o o o o o
> > > > CO
a a a a w
to to dd oo w
o co to ~ 5
0)
-------
W
Key
El OBSERVED
CD ADOM 1
A ADOM 2
+ ADOM 3
X ISC
?E-3
1E-2
Deposition Velocity
Figure B-2c. Cumulative probability plot of deposition velocity (cm/s) using
the complete small particle data set.
-------
Key
OBSERVED
D UAM 1
UAM 2
ISC
Sfc-3
I I r
IE-2 1E-1
Deposition Velocity
Figure B-2d. Cumulative probability plot of deposition velocity (cm/s) using
the complete small particle data set.
-------
Appendix C
Observational Particle Deposition Velocity Data Sets
-------
24 Total Number of Data Sets
ZnS 18 11
Doran & Horst (1985), AEnv, 19. 939-951.
DESERT GRASSES, 1-2 m HIGH SAGEBRUSH
3.0, 140.0 - zO (cm), zd(cm)
-999.9, 2.0 - us measurement ht. (m), temp. meas. ht.
-999.9, 2 - LAI (estimated), vegetation state
14. - no. of diameters, density(gm/caM3)
6. - parti eel diameter(nucrons)
1.
MM-DD-YY
05-18-83
05-18-83
05-18-83
05-26-83
05-26-83
05-26-83
06-05-83
06-05-83
06-05-83
06-12-83
06-12-83
06-12-83
06-24-83
06-24-83
06-24-83
06-27-83
06-27-83
06-27-83
ENDDATA
B HR E HR
(1st) (1st)
22:48 23:18
22:48 23:18
22:48 23:18
23:24 23:54
23:24 23:54
23:24 23:54
22:10 22:40
22:10 22:40
22:10 22:40
22:43 23:13
22:43 23:13
22:43 23:13
23:06 23:28
23:06 23:28
23:06 23:28
21:31 22:01
21:31 22:01
21:31 22:01
US
(m/s)
7.61
8.53
9.43
3.23
3.59
3.83
4.74
5.40
6.32
3.00
3.39
3.75
3.07
3.24
3.46
3.17
3.80
4.37
mass
TEMP
(C)
14.7
14.7
14.7
19.5
19.5
19.5
17.5
17.5
17.5
14.9
14.9
14.9
14.1
14.1
14.1
20.5
20.5
20.5
fraction
SU RAO
(U/m«*2)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
USTAR
(m/s>
0.4
0.4
0.4
0.26
0.26
0.26
0.27
0.27
0.27
0.20
0.20
0.20
0.26
0.26
0.26
0.30
0.30
0.30
HONIN
(m)
166.
166.
166.
44.
44.
44.
77.
77.
77.
34.
34.
34.
59.
59.
59.
71.
71.
71.
HEAT FLUX
(W/ra**2)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
(m)
RA
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RO
(a/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RC
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
VD
(cm/s)
4.21
4.05
3.65
1.93
1.80
1.74
3.14
3.02
2.84
1.75
1.62
1.31
1.56
1.47
1.14
1.17
1.15
1.10
S04 1 2
Nicholson and Oavies (1987), AEnv, 21, 1561-1571
BARLEY
-999.7, 12.0 - zO (cm), zd(cm)
1.0
-999.9
1.0,
-999.9,
- ws measurement ht. (m), temp. meas. ht. (m)
- LAI(estimated), vegetation state
12 1.0 - no. of diameters, density (gm/cm**3)
0.10 0.13 0.16 0.19 0.23 0.29 0.36 0.44 0.54 0.66 0.81 1.0 - particle diameters
0.05 0.10 0.14 0.19 0.16 0.11 0.05 0.04 0.05 0.06 0.04 0.0 - fractional mass
HH-DO-YY B HR E HR US TEMP SU RAO USTAR HONIN HEAT FLUX RA RO RC VD
(1st) (1st) (m/s) (C) (U/m»*2) (m/s) (m) (W/***2) (s/cm) (s/cm) (s/c«) (cm/s)
06-04-79 6:00 1.71 14.2 -999.9 0.11 -999e9 -999.9 1.49 -999.9 -999.9 0.04
ENDDATA
ZO
(cm)
0.8
Ri phim pnih
0.051 1.36 1.36
C-l
-------
S04 6 10
Nicholson and Davies (1987), AEnv, 21, 1561-1571
ROUGH PASTURE
-999.7, 11.0 - zO (cm), zd(em)
1.0, 1.0 - MS measurement ht. (m), temp. mess. ht. (m)
-999.9, -999.9 - LAI(estimated), vegetation state
12 1.0 - no. of diameters, density
0.10 0.13 0.16 0.19 0.23 0.29 0.36 0.44 0.54 0.66 0.81 1.0 - particle diameters
0.05 0.10 O.U 0.19 0.16 0.11 0.05 0.04 0.05 0.06 0.04 0.0 - fractional
MM-DD-YY
06-21-79
06-21-79
06-21-79
07-18-79
07-18-79
07-18-79
EHDOATA
B HR E HR WS
(1st) (1st) (n/s)
2:36 1.23
4:00 2.53
3:55 1.64
3:43 2.69
2:34 2.34
4:07 2.35
TEMP
(C) <
8.8
21.2
8.3
14.4
-999.9
9.9
SU RAD
:u/«*»2>
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
USTAR HONIN Y
(o/s) (m)
0.10 -999e9
0.26 -999e9
0.06 -999e9
0.29 -999e9
0.27 -999e9
0.27 -999e9
IEAT FLUX
(W/m»*2) (
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RA
;s/cm) (*/c
1.19 -999
0.38 -999
4.09 -999
0.33 -999
0.39 -999
0 32 -999
RD RC
) (s/cm)
.9 -999.9
.9 -999.9
.9 -999.9
.9 -999.9
.9 -999.9
.9 -999.9
VD
(em/s)
0.33
0.57
0.03
0.33
0.23
-0.01
ZO Ri
(on)
0.4 -0.054
1.0 -0.042
0.3 0.093
2.2 0.000
1.3 -999.9
3.2 0.005
phira
0.86
0.88
1.93
1.00
-999.9
1.02
phih
0.73
0.77
1.93
1.00
-999.9
1.02
S04 12 3
Nicholson and Davies (1987), AEnv, 21, 1561-1571
SHORT GRASS
-999.7, 8.0 - zO (cm), zd(cm)
1.0, 1.0 - ws measurement ht. (m), temp. neas. ht. (m)
-999.9, -999.9 - LAI(estimated), vegetation state
12 1.0 - no. of diameters, density (gm/cm**3)
0.10 0.13 0.16 0.19 0.23 0.29 0.36 0.44 0.54 0.66 0.81 1.0 - particle diameters
0.05 0.10 0.14 0.19 Q.16 0.11 0.05 0.04 0.05 0.06 0.04 0.0 - fractional mass
MM-OO-YY 8 HR E HR US TEMP SU RAD USTAR MONIN HEAT FLUX RA RD RC VD ZO Ri phim phih
(1st) (C) (W/w*«2) (m/s) (m) (V/»*"2) (s/cm) (s/cm) (s/cm) (ca/s) (cm)
10-08-79 6:59 3.09 -999.9 -999.9 0.22 -999e9 -999.9 0.61 -999.9 -999.9 0.29 0.4 -999.9 -999.9 -999.9
10-10-79 5:48 2.21 -999.9 -999.9 0.19 -999e9 -999.9 0.63 -999.9 -999.9 0.02 0.8 -999.9 -999.9 -999.9
10-17-79 5:20 2.90 9.4 -999.9 0.29 -999e9 -999.9 0.3S -999.9 -999.9 -0.19 1.4 -0.009 0.97 0.93
2-04-80 4:32 3.90 -999.9 -999.9 0.28 -999e9 -999.9 0.51 -999.9 -999.9 0.28 0.3 -999.9 -999.9 -999.9
2-06-80 4:50 1.44 -999.9 -999.9 0.10 -999*9 -999.9 1.41 -999.9 -999.9 0.19 0.3 -999.9 -999.9 -999.9
2-19-80 4:40 4.40 6.5 -999.9 0.33 -999e9 -999.9 0.41 -999.9 -999.9 -0.07 0.4 0.001 1.00 1.00
2-19-80 11:23 2.15 4.4 -999.9 0.17 -999e9 -999.9 0.73 -999.9 -999.9 0.02 0.9 0.016 1.09 1.09
2-20-80 4:40 3.73 5.9 -999.9 0.28 -999«9 -999.9 0.48 -999.9 -999.9 0.20 0.3 -0.010 0.96 0.73
2-20-80 11:22 2.98 -999.9 -999.9 0.22 -999*9 -999.9 0.61 -999.9 -999.9 -0.10 0.4 -999.9 -999.9 -999.9
2-21-80 9:55 3.13 5.0 -999.9 0.20 -999e9 -999.9 0.80 -999.9 -999.9 0.09 0.3 0.023 1.13 1.13
2-22-80 4:20 3.81 6.0 -999.9 0.25 -999e9 -999.9 0.62 -999.9 -999.9 0.04 0.2 0.000 1.00 1.00
2-26-80 9:58 1.79 3.0 -999.9 0.16 -999e9 -999.9 0.72 -999.9 -999.9 -0.02 1.0 0.002 1.01 1.01
ENDDATA
C-2
-------
S04 48 10
Nicholson and Davies (1987), AEnv, 21, 1561-1571
ROUGH PASTURE
-999.7, 9.0 - zO (cm), zd(cra)
1.0, 1.0 - us measurement ht. (m), temp. meas. ht. (m)
-999.9, -999.9 - LAI(estimated), vegetation state
12 1.0 - no. of diameters, density (gpi/cm**3)
0.10 0.13 0.16 0.19 0.23 0.29 0.36 0.44 0.54 0.66 0.81 1.0 - particle diameters
0.05 0.10 0.14 0.19 0.16 0.11 0.05 0.04 0.05 0.06 0.04 0.0 - fractional
MM-DO-YY
10-12-79
10-12-79
10-14-79
10-17-79
10-18-79
10-18-79
10-19-79
10-19-79
11-21-79
11-22-79
11-22-79
11-23-79
11-27-79
11-27-79
11-28-79
11-28-79
11-29-79
11-29-79
11-30-79
12-03-79
12-05-79
12-06-79
12-06-79
12-07-79
12-12-79
12-13-79
1-22-80
1-23-80
1-24-80
1-28-80
1-28-80
1-29-80
1-29-80
1-30-80
1-30-80
1-31-80
2-06-80
2-07-80
2-07-30
2-08-30
2-12-80
2-14-80
B HR E HR
(1st) (1st)
4:17
6:59
3:19
6:54
7:00
6:58
10:50
10:50.
10:53
11:02
11:01
11:09
11:12
11:20
10:50
10:59
11:03
11:00
10:42
10:49
11:16
11:16
11:22
4:41
4:38
2:55
2:48
3:34
4:02
4:07
3:55
4:31
5:52
2:35
3:18
5:04
3:58
4:07
6:25
4:00
3:48
5:03
US TEMP SU RAD
(ni/s) (C) (U/m**2)
2.55 -999.9 -999.9
2.92 -999.9 -999.9
1.96 -999.9 -999.9
3.53 7.8 -999.9
3.27 10.5
1.53 5.1
3.84 13.0
1.50 8.8
1.48 2.9
2.30 4.5
2.22 4.3
3.36 6.6
2.08 5.9
2.52 4.5
3.34 6.5
2.62 3.1
2.31 5.2
2.41 3.1
4.30 7.0
3.45 2.9
4.05 -999.9
1.47 6.8
1.78 7.6
2.31 8.0
4.72 6.7
3.29 6.3
6.03 7.7
2.07 1.7
2.19 1.4
2.75 1.2
1.06 2.1
2.22 7.7
3.47 7.3
5.02 8.2
2.40 6.7
4.18 -999.9
2.43 -999.9
3.45 -999.9
1.38 -999.9
4.31 -999.9
1.93 -999.9
3.46 5.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
USTAR HONIN HEAT FLUX RA RO
(W/s) (w) (W/ra*"2) (s/cm) (s/ca)
0.19 -999e9 -999.9 0.67 -999.9
0.24 -999e9 -999.9 0.52 -999.9
0.22 -999e9 -999.9 0.40 -999.9
0.34 -999e9 -999.9 0.31 -999.9
0.33 -999e9
0.14 -999e9
0.35 -999e9
0.18 -999e9
0.14 -999e9
0.19 -999e9
0.17 -999e9
0.27 -999e9
0.15 -999e9
0.17 -999e9
0.25 -999e9
0.17 -999e9
0.23 -999e9
0.15 -999e9
0.39 -999e9
0.23 -999e9
0.32 -999e9
0.12 -999e9
0.13 -999e9
0.20 -999e9
0.31 -999e9
0.23 -999e9
0.41 -999e9
0.11 -999e9
0.10 -999e9
0.16 -999e9
0.07 -999e9
0.14 -999e9
0.22 -999e9
0.31 -999e9
0.16 -999e9
0.26 -999e9
0.18 -999e9
0.19 -999e9
0.15 -999e9
0.28 -999e9
0.14 -999e9
0.22 -999e9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
0.29 -999.9
0.83 -999.9
0.31 -999.9
0.48 -999.9
0.71 -999.9
0.66 -999.9
0.74 -999.9
0.47 -999.9
0.97 -999.9
0.87 -999.9
0.54 -999.9
0.93 -999.9
0.61 -999.9
1.05 -999.9
0.28 -999.9
0.63 -999.9
0.39 -999.9
0.97 -999.9
1.04 -999.9
0.70 -999.9
0.49 -999.9
0.62 -999.9
0.35 -999.9
1.72 -999.9
2.00 -999.9
1.12 -999.9
2.21 -999.9
1.10 -999.9
0.73 -999.9
0.53 -999.9
0.95 -999.9
0.63 -999.9
0.76 -999.9
0.91 -999.9
0.36 -999.9
0.54 -999.9
1.01 -999.9
0.72 -999.9
RC VD
(s/ca) (cm/s)
-999.9 0.20
-999.9 0.09
-999.9 0.16
-999.9 0.04
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-0.34
0.24
0.34
0.08
0.10
0.07
-0.06
-0.01
-0.12
0.15
-0.21
-0.03
0.05
0.02
-0.28
0.32
0.09
-0.11
0.07
0.23
0.18
-0.06
0.41
0.04
0.01
0.21
-0.03
-0.12
0.22
0.22
0.04
-0.36
0.02
-0.18
0.04
0.08
0.10
0.00
ZO Ri
(cm)
0.5 -999.9
0.6 -999.9
2.6 -999.9
1.5 0.004
1.5 -0.014
2.3 0.039
1.1 -0.003
4.3 0.010
2.0 0.011
0.6 -0.006
0.7 0.010
0.4 0.003
0.3 0.000
0.6 0.034
0.5 0.006
0.6 0.042
0.4 0.014
0.6 0.044
1.3 0.007
0.4 0.020
0.6 -999.9
1.4 0.023
1.1 0.034
0.4 0.008
0.2 0.008
0.5 0.015
0.3 0.004
0.3 0.049
0.1 0.044
0.1 0.011
0.6 0.035
0.4 0.023
0.2 0.004
0.1 0.000
0.4 0.018
0.1 -999.9
0.4 -999.9
0.1 -999.9
0.5 -999.9
0.2 -999.9
0.3 -999.9
0.2 0.006
phim phih
-999.9 -999.9
-999.9 -999.9
-999.9 -999.9
1.02 1.02
0.95 0.90
1.26 1.26
0.99 0.98
1.06 1.06
1.06 1.06
- 0.98 0.96
1.06 1.06
1.02 1.02
1.00 1.00
1.21 1.21
1.03 1.03
1.28 1.28
1.08 1.08
1.30 1.30
1.04 1.04
1.12 1.12
-999.9 -999.9
1.14 1.14
1.22 1.22
1.04 1.04
1.04 1.04
1.09 1.09
1.02 1.02
1.34 1.34
1.29 1.29
1.06 1.06
1.23 1.23
1.14 1.14
1.02 1.02
1.00 1.00
1.10 1.10
-999.9 -999.9
-999.9 -999.9
-999.9 -999.9
-999.9 -999.9
-999.9 -999.9
-999.9 -999.9
1.03 1.03
C-3
-------
2-25-80
3-05-30
3-06-80
3-06-80
3-11-80
3-12-80
ENDDATA
9:59
9:56
10:03
5:16
4:36
7:03
1.96
3.12
3.23
2.70
2.52
4.94
4.6
3.2
5.3
4.7
5.2
7.4
-999.9
999.9
-999.9
-999.9
-999.9
-999.9
0.15 -999e9
0.20 -999*9
0.21 -999e9
0.17 -999e9
0.19 -999«9
0.35 -999e9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
0.92 -999.9 -999.9
0.76 -999.9 -999.9
0.73 -999.9 -999.9
0.94 -999.9 -999.9
0.67 -999.9 -999.9
0.41 -999.9 -999.9
-0.09
0.28
0.51
0.04
0.02
-0.53
0.4
0.2
0.2
0.2
0.7
0.3
-0.009
0.005
Q.OOO
0.012
0.010
0.001
0.97
1.03
1.00
1.07
1.05
1.01
0.94
1.03
1.00
1.07
1.05
1.01
S04 68
Nicholson and Davi
BARE SOIL
-999.7. 10.0
1.0, 1.0
-999.9, -999.9
12 1.0
(1987), AEnv, 21, 1561-1571
- zO (en), zd(ca)
ht. ()
ht. (m/s) (m) (U/nr*-2) (s/cm) (s/cm) (s/cm) (cm/s)
10-21-79 6:52 1.77 6.6 -999.9 0.10 -999e9 -999.9 1.62 -999.9 -999.9 0.35
ENOOATA
ZO
(cm)
1.3 0.071
Ri phim phih
1.59 1.59
S04 42
Nicholson and Davi
TALL BARLET 7T?
-999.7, 6.0
1.0, "1.0
-999.9, -999.9
12 1.0
(1987), AEnv, 21, 1561-1571
ht. (m)
- zO (cm), zd(cm)
- ws measurement ht. (m), temp, maas
- LAI(estimated), vegetation state
- no. of diameters, density (gm/oA>*3)
0.10 0.13 0.16 0.19 0.23 0.29 0.36 0.44 0.54 0.66 0.31 1.0 - particle di
0.05 0.10 0.14 Q.19 0.16 0.11 0.05 0.04 0.05 0.06 0.04 0.0 - fractional
eters
HM-DD-YY
6-09-SO
6-10-80
6-10-80
6-12-80
ENOOATA
B HR
-------
S 83
Hicks et al (1986), BLH, 34, 103-121
GRASSLAND
9.4, -999.9 - zO (cm), zd(on)
7.0, 1.0 -us measurement ht. (m), temp. meas. ht. (m)
-999.9, -999.9 - LAI(estimated), vegetation state
12 1.0 - no. of diameters, density
0.10 0.13 0.16 0.19 O.Z3 0.29 0.36 0.44 0.54 0.66 0.81 1.0 - particle diameters
0.05 0.10 0.14 0.19 0.16 0.11 0.05 0.04 0.05 0.06 0.04 0.0 - fractional mass
MM-OO-YY
09-17-79
09-17-79
09-25-79
09-25-79
09-25-79
09-25-79
09-25-79
09-25-79
ENOOATA
8 HR E HR
(1st)
15:05 15:30
15:35 16:00
11:35 12:00
12:05 12:30
12:35 13:00
13:05 13:30
14:35 15:00
15:35 16:00
US
(n/s)
1.83
1.40
1.28
1.44
1.34
1.29
1.57
1.53
TEMP
(C)
23.6
23.5
21.5
22.0
22.5
22.9
22.9
22.9
SU RAO
(U/m**2>
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
USTAR
(m/s)
0.11
0.11
0.15
0.15
0.15
0.16
0.15
0.15
HONIN
(n)
-999e9
-QQQmQ
TTTMST
QQQjMQ
TrT°Tf5T
-999e9
-999e9
-999e9
-999e9
-999e9
HEAT FLUX
(V/m*«2)
51.0
19.0
88.0
70.0
70.0
41.0
53.0
22.0
RA
(s/cn)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RD
(s/ca)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RC
(s/cm>
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
VO
(cm/s)
0.61
-0.01
0.72
0.44
0.33
0.12
0.00
0.00
Pb 1 3
Garland (1982), Conference Proceedings, 849-858
GRASS - WIND TUNNEL
2.0, -999.9 - zO estimated(cm), zd(cm)
-999.9, -999.9 - us measurement ht. (m), temp. meas. ht. (m)
-999.8, -999.9 - LAUestimated), vegetation state
1 1.0 - no. of diameters, density (gm/cm**3)
13. - diameter (microns)
1. - mass fraction
HM-OD-YY B HR E HR US TEMP SU RAD USTAR HONIM HEAT FLUX RA RD RC VD
(1st) (1st) (m/s) (C) (W/m**2) (m/s) (m) (U/«f*2> (s/cm) (s/cm) (s/cm) (cm/s)
-999.9 20. -999.9 0.35 9.0e9 -999.9 -999.9 -999.9 -999.9 1.90
ENOOATA
Pb 1 3
Garland (1982), Conference Proceedings, 849-858
GRASS - UIND TUNNEL
2.0, -999.9 - zO estimated(cm), zd(cm)
-999.9, -999.9 - us measurement ht. (m), temp. meas. ht. (m)
-999.3, -999.9 - LAI(estimated), vegetation state
1 1.0 - no. of diameters, density (gm/an**3)
10. - diameter (microns)
1. - mass fraction
HM-OD-YY B HR E HR US TEMP SU RAD USTAR HONIN HEAT FLUX RA RD RC VD
(1st) (1st) (m/s) (C) (U/m**2) (m/s) (m) (U/B**2) (s/ca) (s/cm) (s/cm) (cm/s)
-999.9 20. -999.9 0.35 9.0e9 -999.9 -999.9 -999.9 -999.9 1.40
ENDDATA
C-5
-------
Pb 1 3
Garland (1982), Conference Proceedings. 849-858
GRASS - WIND TUNNEL
2.0, -999.9 zO estimated(cm), zd(cm)
-999.9, -999.9 - us measurement ht. (m), temp. meas. ht. (m)
-999.8, -999.9 - LAI(estimated), vegetation state
1 1.0 - no. of diameters, density (gm/cm"*3)
7.5 - diameter (microns)
1. - mass fraction
MM-OD-YY B HR E HR WS TEMP SW RAO USTAR MONIN HEAT FLUX RA RD RC VD
(1st) (1st) (m/s) (C) (W/m**2) (m/s)
-999.9 0.15
Pb 1 3
Garland (1982), Conference Proceedings, 849-858
GRASS - WIND TUNNEL
2.0, -999.9 - zO estimated(cm>, zd(ca)
-999.9, -999.9 - us measurement ht. (ra), temp. meas. ht. (a)
-999.8, -999.9 - LAI(estimated), vegetation state
1 1.0 - no. of diameters, density (gm/cm**3)
1.6 - diameter (microns)
1. - mass fraction
MM-DO-YY B HR E HR WS TEMP SU RAO USTAR MONIN HEAT FLUX RA RO RC VD
(1st) (1st) (m/s) (C) (W/m**2) (m/s) (m) (W/m**2) (s/cm) (s/on) (s/cm) (cm/s)
-999.9 20. -999.9 0.35 9.0e9 -999.9 -999.9 -999.9 -999.9 0.04
ENDDATA
C-6
-------
Pb 1 3
Garland (1982), Conference Proceedings. 849-858
GRASS - WIND TUNNEL
2.0.
-999.9.
-999.8,
1 1
0.75
1.
MH-DO-YY
-999.9
-999.9
-999.9
.0
B HR E HR
(1st) (1st)
- zO estimated(cm), zd(cm)
- us measurement ht. (m), temp. meas. ht.
- LAI (estimated), vegetation state
- no. of diameters, density (gm/cm*"3)
- diameter (microns)
- mass fraction
US TEMP SU RAO USTAR MONIN HEAT FLUX
(m/s) (C) (W/m**2) (m/s) (m) (W/m**2)
-999.9 20. -999.9 0.35 9.0e9 -999.9
(m)
RA RD RC VD
(W/m**2) (s/ca) (s/cm) (s/cm) (cm/s)
'.9 -999.9 -999.9 0.02
ENDDATA
Pb 1 3
Garland (1982), Conference Proceedings, 849-858
GRASS - WIND TUNNEL
2.0, -999.9
-999.9, -999.9
-999.8, -999.9
1 1.0
0.4
1.
MM-OO-YY B HR E HR WS TEMP SU RAD
(1st) (1st) (m/s) (C) (U/m**2)
-999.9 20. -999.9
ENDDATA
- zO estimated(cm), zd(cm)
- ws measurement ht. (m), temp. meas. ht. (m)
- LAI(estimated), vegetation state
- no, of diameters, density (gm/cm**3)
- diameter (microns)
- mass fraction
USTAR HONIN
(m/s) (m)
0.35 9.0e9
HEAT FLUX RA RO RC VD
(U/m**2) (s/cm) (s/cm) (s/cm) (cm/s)
-999.9 -999.9 -999.9 -999.9 0.02
Pb 1 3
Garland (1982). Conference Proceedings, 849-858
GRASS - WIND TUNNEL
2.0, -999.9
-999.9, -999.9
-999.8, -999.9
1 1.0
0.04
1.
- zO estimated(cm), zd(cm)
- us measurement ht. (ra), temp. meas. ht. (m)
- LAI(estimated), vegetation state
- no. of diameters, density (gm/cm**3)
- diameter (microns)
- mass fraction
MM-OD-YY
ENDDATA
B HR E HR
(1st) (1st)
WS
(m/s)
-999.9
TEMP
(C)
20.
SU RAD
(U/ra«*2)
-999.9
USTAR
(m/s)
0.35
HONIN
(m)
9.0e9
HEAT FLUX
(U/m*"2)
-999.9
RA
(s/cra)
-999.9
RD
(s/cm)
-999.9
RC
(s/cm)
-999.9
VD
(CUl/S)
0.09
FECOH 2 3
Garland (1982), Conference Proceedings, 849-858
GRASS
- zO estiimted(cm), zd(cn)
2.0, -999.9
10.0, -999.9
-999.9, -999.9
1 1.0
2.3
1.
MM-OD-YY
B HR E HR
(1st) (1st)
05-21-81
09-23-81
ENOOATA
WS
(m/s)
2.5
3.0
- estimated ws meas. ht. (m), temp. meas. ht. (m)
- LAI(estimated), vegetation state
- no. of diameters, density
- diameter (microns)
- mass fraction
TEMP SU RAD USTAR HONIN
(C) (W/m**2) (m/s) (m)
20.0 -999.9 -999.9 9.0e9
20.0 -999.9 -999.9 9.0e9
HEAT FLUX RA RD RC VD
(W/m**2) (s/cm) (s/cm) (s/cm) (cm/s)
0.0 -999.9 -999.9 -999.9 0.05
0.0 -999.9 -999.9 -999.9 0.12
C-7
-------
FECCH 2 3
Garland (1982), Confer
GRASS - WIND TUNNEL
2.0, -999.9
10.0, -999.9
-999.8, -999.9
1 1.0
2.8
1.
MN-DD-YY B HR E HR
(1st) (1st)
06-01-81
08-17-81
ENDOATA
ence Proceedings, 849-858
zO estimated(cm), zd(cn)
- estimated us meas. ht. (m).
- LAI (estimated), vegetation ;
- no. of diameters, density
- diameter (microns)
- mass fraction
US TEMP SU RAO USTAR HONIN
(m/s) (C) (U/n**2) (m/s) (m)
3.5 20.0 -999.9 -999.9 9.0e9
3.5 20.0 -999.9 -999.9 9.0e9
tenp. meas.
itate
HEAT FLUX
(W/m**2)
0.0
0.0
ht. (m)
RA RD
(s/cm) (s/cm)
999.9 -999.9
-999.9 -999.9
RC VD
(s/cm) (cm/s)
-999.9 0.18
-999.9 0.12
FEQOH 1 3
Garland (1982), Conference Proceedings, 849-858
GRASS
2.0, -999.9 - zO estimated(cm), zd(cm)
10.0, -999.9 - estimated us meas. ht. (m), temp. meas. ht. (m)
-999.9, -999.9 - LAI(estimated), vegetation state
1 1.0 - no. of diameters, density
3.8 - diameter (microns)
1. - mass fraction
MM-OO-YY B HR E HR US TEMP SU RAD USTAR HONIN HEAT FLUX RA RD RC VD
(1st) (1st) (m/s) (C) (U/m**2) (m/s) (m) (U/m**2) (s/cm) (s/cm) (s/cm) (cm/s)
09-30-82 2.4 20.0 -999.9 -999.9 9.0e9 0.0 -999.9 -999.9 -999.9 0.42
ENDOATA
Fine Prt 12 4
Uesely et al (1983), BLM, 27, 237-255.
LEAFLESS DECIDUOUS FOREST IN WINTER (North Carolina)
100., -999.9 - zO estimated(cm), zd(cm)
39.0, 42.0 - us measurement ht. (m), temp. meas. ht. (m)
-999.9, 3.0 - LAI(estimated), vegetation state
6 1.0 - no. of diameters, density (gm/cm**3)
0.05 0.06 0.07 0.08 0.09 0.1 - particle diameters
0.17 0.16 0.17 0.17 0.16 0.17 - mass fraction
MM-OO-YY
01-26-81
01-26-81
01-26-81
01-27-81
01-27-81
01-27-81
01-28-81
01-28-81
01-28-81
01-28-81
01-28-81
01-28-81
ENODATA
B HR E HR
(1st) (1st)
10:00 10:30
17:30 18:00
18:00 18:30
10:30 11:00
11:30 12:00
12:30 13:00
10:30 11:00
11:30 12:00
12:00 12:30
12:30 13:00
13:00 13:30
15:00 15:30
US
(m/s)
4.0
2.1
2.5
2.5
2.6
1.8
3.6
2.7
2.9
3.3
2.8
1.5
TEMP
(C)
9.2
15.4
14.8
13.2
15.0
15.2
7.5
8.4
8.5
9.2
9.7
9.7
SU RAD
(W/m**2)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
999.9
USTAR
(m/s)
0.64
0.26
0.28
0.38
0.38
0.25
0.56
0.57
0.48
0.56
0.57
0.35
MOHIN
(m)
rtrtrt O
"TfTrWy
.QOQMa
7 TTClr
-999e9
999e9
QAA^O
yyysy
.000*0
Trycy
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
HEAT FLUX
(W/m"*2)
49.0
-16.0
-20.0
85.0
56.0
26.0
43.0
209.0
89.0
231.0
237.0
55.0
RA
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
999.9
RD
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RC
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
VD
(cm/s)
0.010
0.020
0.120
0.110
0.070
0.140
0.680
0.320
0.300
0.550
0.200
0.090
C-8
-------
S 18 3
Uesety et al (1982), Conference Proceedings, 943-952
SHORT GRASS, TEXAS
1.0, -999.9 - zO estimted(cm), zd(cm)
39.0, 42.0 - w> measurement ht. (m), temp. meas. ht. (i»)
-999.9, -999.9 - LAI(estimated), vegetation state
12 1.0 - no. of diameters, density (gm/cm**3)
0.10 0.13 0.16 0.19 0.23 0.29 0.36 0.44 0.54 0.66 O.S1 1.0 - particle diameters
0.05 0.10 0.14 0.19 0.16 0.11 0.05 0.04 0.05 0.06 0.04 0.0 - fractional
MM-DO-TY B HR E HR US TEMP SU RAD USTAR MONIN
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
ENDOATA
(1st) (1st)
11:18 11:48
11:48 12:18
12:18 12:48
12:48 13:18
13:18 13:48
13:48 14:18
14:18 14:48
14:48 15:18
15:18 15:48
15:48 16:18
16:18 16:48
16:48 17:18
17:18 17:48
17:48 18:18
18:18 18:48
18:48 19:18
19:18 19:48
19:48 20:18
(m/s)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
(0
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
(U/m**2)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
(m/s)
0.24
0.25
0.26
0.32
0.20
0.21
0.22
0.33
0.24
0.27
0.33
0.34
0.32
0.25
0.27
0.28
0.20
0.18
(m)
-999e9
-999e9
-999e9
- vWey
OOO.AO
If fW
OOO_rt
wycy
QOQmO
TTrTCV
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
(U/m**2>
144.
135.
156.
170.
173.
112.
45.
105.
102.
66.
52.
15.
-14.
-29.
-41.
-36.
-35.
-42.
Cs/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
(s/oa)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
(cm/s)
0.220
0.140
0.110
0.150
0.230
0.240
0.150
0.150
0.190
0.150
0.160
0.290
0.200
0.200
0.250
0.150
0.070
0.140
PART 61 5
Lorenz & Murphy(1989), BLM, 46, 355-366.
PINE PLANTATION
28.0, 790.0 - zO (cm), zd(cm)
9.8, -999.9 - us measurement ht. (m), temp. meas. ht. (a)
9.0 -999.9 - LAI, vegetation state
61. - no. of diameters, density (gm/cm**3)
0.5 0.6 0.7 0.3 0.9 1.0 - particle diameters (microns)
0.17 0.16 0.17 0.17 0.16 0.17 - mass fraction
MM-OD-YY B HR E HR US TEMP SU RAD USTAR MONIN HEAT FLUX RA RO RC VD
(1st) (1st) (m/s) (C) (W/m*»2) (m/s) (m) (U/m**2) (s/ca) (s/cm) (s/cm) (cm/s>
1.84 17.8 -999.9 -999.9 9.e09 0. -999.9 -999.9 -999.9 0.15
2.22 17.8 -999.9 -999.9 9.e09 0. -999.9 -999.9 -999.9 0.02
3.22 17.8 -999.9 -999.9 9.e09 0. -999.9 -999.9 -999.9 0.04
2.07 17.8 -999.9 -999.9 9.e09 0. -999.9 -999.9 -999.9 0.13
2.16 17.8 -999.9 -999.9 9.e09 0. -999.9 -999.9 -999.9 0.13
2.26 17.8 -999.9 -999.9 9.e09 0. -999.9 -999.9 -999.9 0.11
2.10 17.8 -999.9 -999.9 9.e09 0. -999.9 -999.9 -999.9 0.17
2.16 17.3 -999.9 -999.9 9,e09 0. -999.9 -999.9 -999.9 0.22
2.17 17.8 -999.9 -999.9 9.e09 0. -999.9 -999.9 -999.9 0.19
2.28 17.8 -999.9 -999.9 9.e09 0. -999.9 -999.9 -999.9 0.17
2.39 17.8 -999.9 -999.9 9.e09 0. -999.9 -999.9 -999.9 0.14
2.38 17.3 -999.9 -999.9 9.e09 0. -999.9 -999.9 -999.9 0.15
C-9
-------
2.27
2.31
2.33
2.43
2.48
2.53
2.54
2.64
2.87
2.59
2.64
2.66
2.31
3.04
3.54
2.32
2.36
2.41
2.46
2.62
2.63
2.70
2.72
2.80
2.30
2.94
3.12
3.13
3.20
3.28
3.23
3.22
3.10
3.25
3.34
3.50
3.64
3.73
3.S4
3.38
4.02
3.90
3.89
4.08
4.39
3. 85
4.41
2.81
3.39
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.3
17.8
17.8
17.8
17.8
17.8
17.3
17.3
-999.9
-999.9
-999.9
-999.9
999.9
-999.9
-999.9
-999.9
-999.9
999.9
999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
9.e09
9.e09
9.e09
9.e09
9.«09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
- 0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
-999.9 -999.9 -999.9 0.24
-999.9 -999.9 -999.9 0.23
-999.9 -999.9 -999.9 0.30
-999.9 -999.9 -999.9 0.34
-999.9 -999.9 -999.9 0.28
-999.9 -999.9 -999.9 0.32
-999.9 -999.9 -999.9 0.25
-999.9 -999.9 -999.9 0.14
999.9 -999.9 -999.9 0.14
-999.9 -999.9 -999.9 0.35
-999.9 -999.9 -999.9 0.31
-999.9 -999.9 -999.9 0.26
-999.9 -999.9 -999.9 0.23
-999.9 -999.9 -999.9 0.25
-999.9 -999.9 -999.9 0.20
-999.9 -999.9 -999.9 0.45
-999.9 -999.9 -999.9 0.44
-999.9 -999.9 -999.9 0.52
-999.9 -999.9 -999.9 0.49
-999.9 -999.9 -999.9 0.52
-999.9 -999.9 -999.9 0.41
-999.9 -999.9 -999.9 0.39
-999.9 -999.9 -999.9 0.39
-999.9 -999.9 -999.9 0.43
-999.9 -999.9 -999.9 1.17
-999.9 -999.9 -999.9 0.49
-999.9 -999.9 -.999.9 0.56
-999.9 -999.9 -999.9 0.61
-999.9 -999.9 -999.9 0.63
-999.9 -999.9 -999.9 0.75
-999.9 -999.9 -999.9 0.90
-999.9 -999.9 -999.9 0.97
-999.9 -999.9 -999.9 0.34
-999.9 -999.9 -999.9 0.39
-999.9 -999.9 -999.9 0.48
-999.9 -999.9 -999.9 0.32
-999.9 -999.9 -999.9 0.38
-999.9 -999.9 -999.9 0.44
-999.9 -999.9 -999.9 0.42
-999.9 -999.9 -999.9 0.36
-999.9 -999.9 -999.9 0.46
-999.9 -999.9 -999.9 0.59
-999.9 -999.9 -999.9 0.91
-999.9 -999.9 -999.9 0.72
-999.9 -999.9 -999.9 0.69
-999.9 -999.9 -999.9 1.32
-999.9 -999.9 -999.9 0.90
-999.9 -999.9 -999.9 0.00
-999.9 -999.9 -999.9 0.00
ENDOATA
C-10
-------
Appendix D
Predicted Deposition Velocities vs Particle Diameter
-------
CJ
LU
CO
t_J
>-
to
a
a.
1.0E+02
1.0E+01
1.0E+00
1.0E-01
1.0E-02
1.0E-03
1.0E-04
_ GARB 0, CARB 1
1.0E-05
CARB2
0.1
1.0 10.0
PflRTICLE DIflMETER (MICRONS)
100.0
Figure D-l. Predicted deposition velocity for the CARB-based models for u. = 10 cm/s,
z0 = 10 cm, LAI = 1.0, p = 1.0 g/cm3, and neutral stability. The gravitational
settling velocity is vg.
D-l
-------
C_J
UJ
CD
\
31
LJ
C_3
a
UJ
en
a
o_
1.0E+02
1.0E+01
1.0E+G
1.0E-01
1.0E-02
1.0E-03
1.0E-04
1.0E-05
ADOM 3"
0.1
ADOM2
1.0 10.0
PHRTICLE DIflMETER (MICRONS)
100.0
Figure D-2. Predicted deposition velocity for the ADOM-based models for u. = 10 cm/s,
z0 = 10 cm, LAI = 1.0, p = 1.0 g/cm3, and neutral stability. The gravitational
settling velocity is vg.
D-2
-------
o
LU
cn
C_J
>-
C_J
o
>
2
CO
a
Q-
1.0E+02
1.0E+01 -
1.0E+00
1.0E-01
1.0E-02
1.0E-03
1.0E-04
1.0E-05
~---UAM 1
0.1
1.0 10.0
PRRTICLE DIflMETER (MICRONS)
100.0
Figure D-3. Predicted deposition velocity for the UAM-based models for u. = 10 cm/s,
z0 = 10 cm, LAI = 1.0, p = 1.0 g/cm3, and neutral stability. The gravitational
settling velocity is vg.
D-3
-------
24 Total Number of Data Sets
ZnS 18 11
Doran & Horst (1985), AEnv, 19, 939-951.
DESERT GRASSES, 1-2 m HIGH SAGEBRUSH
3.0, 140.0 - zO (cm), zd(cm)
-999.9, 2.0 - ws measurement ht. (m), temp. mess. ht.
-999.9, 2 - LAI (estimated), vegetation state
14. - no. of diameters, density(gm/cm**3)
6. - particel dfameter(microns)
1.
HM-DO-YY
05-18-83
05-18-83
05-18-83
05-26-83
05-26-83
05-26-83
06-05-83
06-05-83
06-05-83
06-12-83
06-12-83
06-12-83
06-24-83
06-24-83
06-24-83
06-27-83
06-27-83
06-27-83
ENDDATA
5 HR E HR
(1st) (1st)
22:48 23:18
22:48 23:18
22:48 23:18
23:24 23:54
23:24 23:54
23:24 23:54
22:10 22:40
22:10 22:40
22:10 22:40
22:43 23:13
22:43 23:13
22:43 23:13
23:06 23:28
23:06 23:28
23:06 23:28
21:31 22:01
21:31 22:01
21:31 22:01
WS
(m/s)
7.61
8.53
9.43
3.23
3.59
3.83
4.74
5.40
6.32
3.00
3.39
3.75
3.07
3.24
3.46
3.17
3.80
4.37
mass
TEMP
(C)
14.7
14.7
14.7
19.5
19.5
19.5
17.5
17.5
17.5
14.9
14.9
14.9
14.1
14.1
14.1
20.5
20.5
20.5
fraction
SU RAD
(W/m**2)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
USTAR
(m/s>
0.4
0.4
0.4
0.26
0.26
0.26
0.27
0.27
0.27
0.20
0.20
0.20
0.26
0.26
0.26
0.30
0.30
0.30
NONIN
(m)
166.
166.
166.
44.
44.
44.
77.
77.
77.
34.
34.
34.
59.
59.
59.
71.
71.
71.
HEAT FLUX
(W/m**2)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
(m)
RA
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RD
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RC
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
VD
(cm/s)
4.21
4.05
3.65
1.93
1.80
1.74
3.14
3.02
2.84
1.75
1.62
1.31
1.56
1.47
1.14
1.17
1.15
1.10
S04 1 2
Nicholson and Davies (1987)
BARLEY
12.0
1.0
-999.9
.0
AEnv, 21, 1561-1571
-999.7,
1.0,
-999.9,
12 1.
- zO (cm), zd(cm)
- ws measurement ht. (m), temp. meas. ht. (m)
- LAI(estimated), vegetation state
- no. of diameters, density (gm/cm**3)
0.10 0.13 0.16 0.19 0.23 0.29 0.36 0.44 0.54 0.66 0.81 1.0 - particle diameters
0.05 0.10 0.14 0.19 0.16 0.11 0.05 0.04 0.05 0.06 0.04 0.0 - fractional mass
MM-DD-YY B HR E HR 'US TEMP SU RAD USTAR MONIN HEAT FLUX RA RD RC VD
(1st) (1st) (m/s) (C) (W/m**2) (m/s) (m) (U/m**2) (s/cm) (s/cm) (s/cm) (cm/s)
06-04-79 6:00 1.71 14.2 -999.9 0.11 -999e9 -999.9 1.49 -999.9 -999.9 0.04
ENDDATA
20
(cm)
0.8
Ri phim phih
0.051 1.36 1.36
-------
S04 6 10
Nicholson and Davies (1987)
ROUGH PASTURE
-999.7, 11.0
1.0, 1.0
-999.9, -999.9
12 1.0
AEnv, 21, 1561-1571
- zO (cm), zd(cm)
- us measurement ht. (m), temp. mess. ht. (m)
- LAI(estimated), vegetation state
- no. of diameters, density (gm/cm**3)
0.10 0.13 0.16 0.19 0.23 0.29 0.36 0.44 0.54 0.66 0.81 1.0 - particle diameters
0.05 0.10 0.14 0.19 0.16 0.11 0.05 0.04.0.05 0.06 0.04 0.0 fractional mass
HM-OD-YY
06-21-79
06-21-79
06-21-79
07-18-79
07-18-79
07-18-79
ENDDATA
B HR
(1st)
2:36
3:55
4:07
E HR
(1st)
4:00
3:43
2:34
US
(m/s)
1.23
2.53
1.64
2.69
2.84
2.35
TEMP
(C)
8.8
21.2
8.3
14. .4
-999.9
9.9
SU RAO
(U/m**2)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
USTAR MONIM
(m/s) (m)
0.10 -999e9
0.26 -999e9
0.06 -999e9
0.29 -999e9
0.27 -999e9
0.27 -999e9
HEAT FLUX
(U/m**2)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RA RD RC
(s/cm) (s/cm) (s/cm)
1.19 -999.9 -999.9
0.38 -999.9 -999.9
4.09 -999.9 -999.9
0.33 -999.9 -999.9
0.39 -999.9 -999.9
0.32 -999.9 -999.9
VD
(cm/s)
0.33
0.57
0.03
0.33
0.23
-0.01
ZO Ri
(cm)
0.4 -0.054
1.0 -0.042
0.3 0.093
2.2 0.000
1.3 -999.9
3.2 0.005
phim
0.86
0.88
1.93
1.00
-999.9
1.02
phih
0.73
0.77
1.93
1.00
-999.9
1.02
S04 12 3
Nicholson and Davies (1987)
SHORT GRASS
-999.7, 8.0
1.0, 1.0
-999.9, -999.9
12 1.0
AEnv, 21, 1561-1571
(m)
- zO (cm), zd(cm)
- ws measurement ht. (m), temp. meas. ht.
- LAI(estimated), vegetation state
- no. of diameters, density (gm/cm**3)
0.10 0.13 0.16 0.19 0.23 0.29 0.36 0.44 0.54 0.66 0.81 1.0 - particle diameters
0.05 0.10 0.14 0.19 0.16 0.11 0.05 0.04 0.05 0.06 0.04 0.0 - fractional mass
MM-DD-YY
10-08-79
10-10-79
10-17-79
2-04-80
2-06-80
2-19-80
2-19-80
2-20-80
2-20-80
2-21-80
2-22-80
2-26-80
ENDDATA
B HR E HR
(1st) (1st)
6:59
5:48
5:20
4:32
4:50
4:40
11:23
4:40
11:22
9:55
4:20
9:58
US TEMP SU RAD
(m/s) (C) (U/m**2)
3.09 -999.9
2.21 -999.9
2.90 9.4
3.90 -999.9
1.44 -999.9
4.40 6.5
2.15 4.4
3.73 5.9
2.98 -999.9
3.13 5.0
3.81 6.0
1.79 3.0
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
USTAR
(m/s)
0.22
0.19
0.29
0.28
0.10
0.33
0.17
0.28
0.22
0.20
0.25
0.16
MONIN HEAT FLUX RA RD
(m) (W/m**2) (s/cm) (s/cm)
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
0.61 -999.9
0.63 -999.9
0.35 -999.9
0.51 -999.9
1.41 -999.9
0.41 -999.9
0.73 -999.9
0.48 -999.9
0.61 -999.9
0.80 -999.9
0.62 -999.9
0.72 -999.9
RC
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
VD
(cm/s)
0.29
0.02
-0.19
0.28
0.19
-0.07
0.02
0.20
-0.10
0.09
0.04
-0.02
ZO Ri
(cm)
0.4 -999.9
0.8 -999.9
1.4 -0.009
0.3 -999.9
0.3 -999.9
0.4 0.001
0.9 0.016
0.3 -0.010
0.4 -999.9
0.3 0.023
0.2 0.000
1.0 0.002
ph i m ph i h
-999.9 -999.9
-999.9 -999.9
0.97 0.93
-999.9 -999.9
-999.9 -999.9
1.00 1.00
1.09 1.09
0.96 0.93
-999.9 -999.9
1.13 1.13
1.00 1.00
1.01 1.01
-------
S04 48 10
Nicholson and Davies (1987), AEnv, 21, 1561-1571
ROUGH PASTURE
-999.7, 9.0 - zO (cm), zd(cm)
1.0, 1.0 - ws measurement ht. (m), temp. meas. ht. (m)
-999.9, -999.9 - LAI(estimated), vegetation state
12 1.0 - no. of diameters, density (gnv/cm**3)
0.10 0.13 0.16 0.19 0.23 0.29 0.36 0.44 0.54 0.66 0.81 1.0 - particle diameters
0.05 0.10 0.14 0.19 0.16 0.11 0.05 0.04 0.05 0.06 0.04 0.0 - fractional mass
MM-DD-YY
10-12-79
10-12-79
10-14-79
10-17-79
10-18-79
10-18-79
10-19-79
10-19-79
11-21-79
11-22-79
11-22-79
11-23-79
11-27-79
11-27-79
11-28-79
11-28-79
11-29-79
11-29-79
11-30-79
12-03-79
12-05-79
12-06-79
12-06-79
12-07-79
12-12-79
12-13-79
1-22-80
1-23-80
1-24r80
1-28-80
1-28-80
1-29-80
1-29-80
1-30-80
1-30-80
1-31-80
2-Q6-80
2-07-80
2-07-80
2-08-80
2-12-80
2-14-80
B HR
(1st)
6:59
6:54
7:00
6:58
10:50
10:50
10:53
11:02
11:01
11:09
11:12
11:20
10:50
10:59
11:03
11:00
10:42
10:49
11:16
11:16
11:22
E HR
(1st)
4:17
3:19
4:41
4:38
2:55
2:48
3:34
4:02
4:07
3:55
4:31
5:52
2:35
3:18
5:04
3:58
4:07
6:25
4:00.
3:48
5:03
WS TEMP SU RAD
(m/s) (C) (W/m**2)
2.55 -999.9 -999.9
2.92 -999.9 -999.9
1.96 -999.9 -999.9
3.53 7.8 -999.9
3.27 10.5
1.53 5.1
3.84 13.0
1.50 8.8
1.48 2.9
2.30 4.5
2.22 4.3
3.36 6.6
2.08 5.9
2.52 4.5
3.34 6.5
2.62 3.1
2.31 5.2
2.41 3.1
4.30 7.0
3.45 2.9
4.05 -999.9
1.47 6.8
1.78 7.6
2.81 8.0
4.72 6.7
3.29 6.3
6.03 7.7
2.07 1.7
2.19 1.4
2.75 1.2
1.06 2.1
2.22 7.7
3.47 7.3
5.02 8.2
2.40 6.7
4.18 -999.9
2.43 -999.9
3.45 -999.9
1.88 -999.9
4.31 -999.9
1.93 -999.9
3.46 5.9
-999.9
-999.9
-999.9
-999.9
-999.9
;999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
USTAR MONIN HEAT FLUX RA
(m/s) (m) (U/m**2) (s/cm)
0.19 -999e9 -999.9 0.67
0.24 -999e9 -999.9 0.52
0.22 -999e9 -999.9 0.40
0.34 -999c9 -999.9 0.31
0.33 -999e9
0.14 -999e9
0.35 -999e9
0.18 -999e9
0.14 -999e9
0.19 -999e9
0.17 -999e9
0.27 -999e9
0.15 -999e9
0.17 -999e9
0.25 -999e9
0.17 -999e9
0.23 -999e9
0.15 -999e9
0.39 -999e9
0.23 -999e9
0.32 -999e9
0.12 -999e9
0.13 -999e9
0.20 -999e9
0.31 -999e9
0.23 -999e9
0.41 -999e9
0.11 -999e9
0.10 -999e9
0.16 -999e9
0.07 -999e9
0.14 -999e9
0.22 -999e9
0.31 -999e9
0.16 -999e9
0.26 -999e9
0.18 -999e9
0.19 -999e9
0.15 -999e9
0.28 -999e9
0.14 -999e9
0.22 -999e9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
0.29
0.83
0.31
0.48
0.71
0.66
0.74
0.47
0.97
0.87
0.54
0.93
0.61
1.05
0.28
0.63
0.39
0.97
1.04
0.70
0.49
0.62
0.35
1.72
2.00
1.12
2.21
1.10
0.73
0.53
0.95
0.63
0.76
0.91
0.86
0.54
1".01
0.72
RD
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RC
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
VD
(cm/s)
0.20
0.09
0.16
0.04
-0.34
0.24
0.'34
0.08
0.10
0.07
-0.06
-0.01
-0.12
0.15
-0.21
-0.03
0.05
0.02
-0.28
0.32
0.09
-0.11
0.07-
0.23
0.18
-0.06
0.41
0.04
0.01
0.21
-0.03
-0.12
0.22
0.22
0.04
-0.36
0.02
-0.18
0.04
0.08
0.10
0.00
ZO Ri ptiim phih
(cm)
0.5 -999.9 -999.9 -999.9
0.6 -999.9 -999.9 -999.9
2.6 -999.9 -999.9 -999.9
1.5 0.004 1.02 1.02
1.5 -0.014 0.95 0.90
2.3 0.039 1.26 1.26
1.1 -0.003 0.99 0.98
4.3 0.010 1.06 1.06
2.0 0.011 1.06 1.06
0.6 -0.006 0.98 0.96
0.7 0.010 1.06 1.06
0.4 0.003 1.02 1.02
0.3 0.000 1.00 1.00
0.6 0.034 1.21 1.21
0.5 0.006 1.03 1.03
0.6 0.042 1.28 1.28
0.4 0.014 1.08 1.08
0.6 0.044 1.30 1.30
1.3 0.007 1.04 1.04
0.4 0.020 1.12 1.12
0.6 -999.9 -999.9 -999.9
1.4 0.023 1.14 1.14
1.1 0.034 1.22 1.22
0.4 0.008 1.04 1.04
0.2 0.008 1.04 1.04
0.5 0.015 1.09 1.09
0.3 0.004 1.02 1.02
0.3 0.049 1.34 1.34
0.1 0.044 1.29 1.29
0.1 0.011 1.06 1.06
0.6 0.035 1.23 1.23
0.4 0.023 1.14 1.14
0.2 0.004 1.02 1.02
0.1 0.000 1.00 1.00
0.4 0.018 1.10 1.10
0.1 -999.9 -999.9 -999.9
0.4 -999.9 -999.9 -999.9
0.1 -999.9 -999.9 -999.9
0.5 -999.9 -999.9 -999.9
0.2 -999.9 -999.9 -999.9
0.3 -999.9 -999.9 -999.9
0.2 0.006 1.03 1.03
-------
2-25-80
3-05-80
3-06-80
3-06-80
3-11-80
3-12-80
ENODATA
9:59
9:56
10:03
5:16
4:36
7:03
1.96
3.12
3.23
2.70
2.52
4.94
4.6
3.2
5.3
4.7
5.2
7.4
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
0.15 -999e9
0.20 -999e9
0.21 -999e9
0.17 -999e9
0.19 -999e9
0.35 -999e9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
0.92 -999.9 -999.9
0.76 -999.9 -999.9
0.73 -999.9 -999.9
0.94 -999.9 -999.9
0.67 -999.9 -999.9
0.41 -999.9 -999.9
-0.09
0.28
0.51
0.04
0.02
-0.53
0.4 -0.009
0.2 0.005
0.2 0.000
0.2 0.012
0.7 0.010
0.3 0.001
0.97
1.03
1.00
1.07
1.05
1.01
0.94
1.03
1.00
1.07
1.05
1.01
S04 68
Micholson and Davies (1987), AEnv, 21, 1561-1571
BARE SOIL
-999.7, 10.0 - zO (cm), zd(cm)
1.0, 1.0 - us measurement ht. (m), temp. meas. ht. (m)
-999.9, -999.9 - LAI(estimated), vegetation state
12 1.0 - no. of diameters, density (gm/cm**3)
0.10 0.13 0.16 0.19 0.23 0.29 0.36 0.44 0.54 0.66 0.81 1.0 - particle diameters
0.05 0.10 0.14 0.19 0.16 0.11 0.05 0.04 0.05 0.06 0.04 0.0 - fractional mass
MM-DD-YY
1-25-80
2-15-80
2-27-80
2-29-80
3-10-80
3-13-80
ENDDATA
B HR E HR
(1st) (1st)
4:56
5:39
4:52
4:04
9:59
4:31
US
(m/s)
2.53
3.10
2.06
2.30
3.04
4.01
TEMP
(C) <
2.9
7.4
5.1
5.9
3.6
4.3
SU RAO
:W/m**2)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
USTAR MONIN \>
(m/s) (m)
0.18 -999e9
0.19 -999e9
0.16 -999e9
0.17 -999e9
0.24 -999e9
0.34 -999e9
IEAT FLUX
(U/m**2) (
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RA ' RD RC
;s/cm) (s/cm) (s/cm)
0.74 -999.9 -999.9
0.88 -999.9 -999.9
0.80 -999.9 -999.9
0.78 -999.9 -999.9
0.55 -999.9 -999.9
0.35 -999.9 -999.9
VD
(cm/s)
0.08
0.25
0.01
0.38
0.25
-0.16
ZO Ri
(cm)
0.2 -0.016
0.1 0.008
0.3 -0.024
0.2 -0.035
0.5 0.004
0.7 -0.010
phim
0.94
1.04
0.92
0.89
1.02
0.96
phih
0.89
1.04
0.85
0.80
1.02
0.93
S04 1 10
Nicholson and Davies (1987)
LONG GRASS
-999.7, 6.0
1.0, 1.0
-999.9, -999.9
12 1.0
AEnv, 21, 1561-1571
- zO (cm), zd(cm)
- us measurement ht. (m), temp. meas. ht. (m)
- LAI(estimated), vegetation state
- no. of diameters, density (gm/cm**3)
0.10 0.13 0.16 0.19 0.23 0.29 0.36 0.44 0.54 0.66 0.81 1.0 - particle diameters
0.05 0.10 0.14 0.19 0.16 0.11 0.05 0.04 0.05 0.06 0.04 0.0 - fractional mass
MM-DD-YY B HR E HR US TEMP SU RAO USTAR MOMIN HEAT FLUX RA RD RC VD
(1st) (1st) (m/s) (C) (U/ro**2) (m/s) (m) (U/m**2) (s/cm) (s/cm) (s/cm) (cm/s)
10-21-79 6:52 1.77 6.6 -999.9 0.10 -999e9 -999.9 1.62 -999.9 -999.9 0.35
ENOOATA
ZO
(cm)
1.3
Ri
phim phih
0.071 1.59 1.59
S04 42
Nicholson and Davies (1987), AEnv, 21, 1561-1571
TALL BARLEY ???
6.0 - zO (cm), zd(cm)
- us measurement ht. (m), temp. meas. ht. (m)
- LAI(estimated), vegetation state
- no. of diameters, density (gm/cm**3)
.19 0.23 0.29 0.36 0.44 0.54 0.66 0.81 1.0 - particle diameters
-999.7,
1.0, 1.0
-999.9, -999.9
12 1.0
0.10 0.13 0.16 0.
0.05 0.10 0.14 0.19 0.16 0.11 0.05 0.04 0.05 0.06 0.04 0.0 - fractional mass
MM-DD-YY
6-09-80
6-10-80
6-10-80
6-12-80
ENDDATA
B HR
(1st)
5:11
E HR
US
(1st) (m/s)
2
.45
5:05 3.58
5:10
5:12
i 1
2
.67
.28
TEMP
SU RAD
(C) (U/m**2)
11
14
10
10
.2
.7
.9
.0
-999
-999
-999
-999
.9
.9
.9
.9
USTAR
(m/s)
0
0
0
0
.17
.27
.12
.17
MONIN HEAT FLUX
(m).
-999e9
-999e9
-999e9
-999e9
(U/m**2)
-999.9
-999.9
-999.9
-999.9
RA RD RC
(U/m**2) (s/cm) (s/cm) (s/cm) (cm/s)
0.84 -999.9 -999.9
0.51 -999.9 -999.9
1.15 -999.9 -999.9
0.76 -999.9 -999.9
VD
cm/s)
0.01
0.04
-0.29
0.21
ZO
(cm)
0.3
0.3
0.3
0.4
Ri
0.006
-0.010
-0.010
0.000
phim
1.03
0.97
0.97
1.00
phih
1.03
0.93
0.93
1.00
-------
S 83
Hicks et al (1986), BLM, 34, 103-121
GRASSLAND
9.4, -999.9 - zO (cm), zd(cm)
7.0, 1.0 - us measurement ht. (m), temp. mess. ht. (m)
-999.9, -999.9 - LAI(estimated), vegetation state
12 1.0 - no. of diameters, density (gm/cm**3)
0.10 0.13 0.16 0.19 0.23 0.29 0.36 0.44 0.54 0.66 0.81 1.0 - particle diameters
0.05 0.10 0.14 0.19 0.16 0.11 0.05 0.04 0.05 0.06 0.04 0.0 - fractional mass
MM-DD-YY
09-17-79
09-17-79
09-25-79
09-25-79
09-25-79
09-25-79
09-25-79
09-25-79
ENDOATA
Pb
Garland
B HR
(1st)
15:05
15:35
11:35
12:05
12:35
13:05
14:35
15:35
1 3
(1982),
E HR
(1st)
15:30
16:00
12:00
12:30
13:00
13:30
15:00
16:00
US
(m/s)
1.83
1.40
1.28
1.44
1.34
1.29
1.57
1.53
TEMP SW RAO USTAR MONIN
(C) (U/m**2) (m/s) (m)
23.6 -999.9 0.11 -999e9
23.5 -999.9 0.11 -999e9
21.5 -999.9 0.15 -999e9
22.0 -999.9 0.15 -999e9
22.5 -999.9 0.15 -999e9
22.9 -999.9 0.16 -999e9
22.9 -999.9 0.15 -999e9
22.9 -999.9 0.15 -999e9
HEAT FLUX
(U/m**2)
51.0
19.0
88.0
70.0
70.0
41.0
53.0
22.0
RA
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RD
(s/cm)
-999
-999
-999
-999
-999
-999
-999
-999
.9
.9
.9
.9
.9
.9
.9
.9
RC
(s/cm)
-999
-999
-999
-999
-999
-999
-999
-999
.9
.9
.9
.9
.9
.9
.9
.9
VD
(cm/s)
0.61
~0.01
0.72
0.44
0.33
0.12
0.00
0.00
Conference Proceedings, 849-858
GRASS - WIND TUNNEL
2.0,
-999.9,
-999.8,
1 1
13.
1.
MM-DD-YY
ENODATA
Pb
Garland
GRASS -
2.0,
-999.9,
-999.8,
1 1
10.
1.
-999.9
-999.9
-999.9
.0
B HR
(1st)
1 3
(1982),
- zO estimated(cm), zd(cm)
- ws measurement ht. (m), temp
. meas. ht
. (m)
- LAI(estimated), vegetation state
- no. of diameters, density (gm/cm**3)
E HR
(1st)
US
(m/s)
-999.9
- diameter (microns)
- mass fraction
TEMP SU RAO USTAR MONIN
(C) (U/m**2) (m/s) (m)
20. -999.9 0.35 9.0e9
HEAT FLUX
(U/m**2)
-999.9
RA
(s/cm)
-999.9
RD
(s/cm)
-999.9
RC
(s/cm)
-999.9
VD
(cm/s)
1.90
Conference Proceedings, 849-858
WIND TUNNEL
-999.9
-999.9
-999.9
.0
MM-DO-YY B HR
(1st)
- zO estimated(cm), zd(cm)
- ws measurement ht. (m), temp
. meas. ht
. (m)
- LAI (estimated), vegetation state
- no. of diameters, density (gm/cm**3)
E HR
(1st)
US
(m/s)
-999.9
- diameter (microns)
- mass fraction
TEMP SU RAO USTAR MONIN
(C) (W/m**2) (m/s) (m)
20. -999.9 0.35 9.0e9
HEAT FLUX
(W/m**2)
-999.9
RA
(s/cm)
-999.9
RD
(s/cm)
-999.9
RC
(s/em)
-999.9
VD
(cm/s)
1.40
ENODATA
-------
Pb 1 3
Garland (1982), Conference Proceedings, 849-858
GRASS - WIND TUNNEL
2.0.
-999.9,
-999.8,
1
-999.9
-999.9
-999.9
.0
1
7.5
1.
MM-OD-YY B HR
E HR
US
(1st) (1st) (m/S)
-999.9
- zO estimated(cin), zd(cm)
- ws measurement ht. (m), temp. meas. ht. (m)
- LAI(estimated), vegetation state
- no. of diameters, density (gm/cm**3)
- diameter (microns)
- mass fraction
TEMP SU RAO USTAR MONIN HEAT FLUX
(C) (U/m**2) (m/s) (m)
20. -999.9 0.35 9.0e9
RA RD RC VD
(V/m**2) (s/cm) (s/cm) (s/cm) (cm/s)
999.9 -999.9 -999.9 -999.9 1.00
ENOOATA
Pb 1 3
Garland (1982), Conference Proceedings, 849-858
GRASS - WIND TUNNEL
- zO estimated(cm), zd(cm)
- ws measurement ht. (m), temp. meas. ht. (m)
- LAI(estimated), vegetation state
- no. of diameters, density (gm/cm**3)
- diameter (microns)
- mass fraction
WS TEMP SU RAD USTAR MONIN
(m/s) (C) (U/m**2) (m/s) (m)
-999.9 20. -999.9 0.35 9.0e9
2.0, -999.9
-999.9. -999.9
-999.8, -999.9
1 1.0
5.0
1.
MM-OD-YY B HR E HR
(1st) (1st)
HEAT FLUX
RA RD RC VD
(U/m**2) (s/cm) (s/cm) (s/cm) (cm/s)
-999.9 -999.9 -999.9 -999.9 0.45
ENDDATA
Pb 1 3
Garland (1982), Conference Proceedings, 849-858
GRASS - WIND TUNNEL
2.0,
-999.9,
-999.8,
1 1
3.2
1.
MM-DD-YY
ENDDATA
-999.9
-999.9
-999.9
.0
B HR
(1st)
- zO estimated(cm), zd(cm)
- ws measurement ht. (m), temp. meas. ht. (m)
- LAI (estimated), vegetation state
- no. of diameters, density (gm/cm**3)
- diameter (microns)
- mass fraction
E HR US TEMP SU RAD USTAR MONIN HEAT FLUX RA
(1st) (m/s) (C) (W/m**2) (m/s) (m) (U/m**2) (s/cm)
-999.9 20. -999.9 0.35 9.0e9 -999.9 -999.9
RD RC VD
(s/cm) (s/cm) (cm/s)
-999.9 -999.9 0.15
Pb 1 3
Garland (1982), Conference Proceedings, 849-858
GRASS - WIND TUNNEL
2.0, -999.9
-999.9, -999.9
-999.8, -999.9
1 1.0
1.6
1.
MM-DD-YY B HR E HR WS
(1st) (1st) (m/s)
-999.9
ENDDATA
- zO estimated(cm), zd(cm)
- ws measurement ht. (m), temp. meas. ht.
- LAI(estimated), vegetation state
- no. of diameters, density (gm/cm**3)
- diameter (microns)
- mass fraction
USTAR MONIN
(m/s) (m)
0.35 9.0e9
(m)
TEMP SU RAD
(C) (W/m**2)
20. -999.9
HEAT FLUX RA RD RC VD
(W/m**2) (s/cm) (s/cm) (s/cm) (cm/s)
-999.9 -999.9 -999.9 -999.9 0.04
-------
Pb 1 3
Garland (1982), Conference Proceedings, 849-858
GRASS - WIND TUNNEL
2.0,
-999.9,
-999.8,
1 1
0.7-5
1.
MM-DD-YY
ENODATA
-999.9
-999.9
-999.9
.0
B HR E HR
(1st) (1st)
- zO estimated(cm), zd(cm)
- ws measurement ht. (m), temp. meas. ht. (m)
- LAI (estimated), vegetation state
- no. of diameters, density (gm/cm**3)
- diameter (microns)
- mass fraction
WS TEMP SW RAD USTAR MOW IN HEAT FLUX RA RD
(m/s) (C) (W/m**2) (m/s) (m) (W/m**2) (s/cm) (s/cm)
-999.9 20. -999.9 0.35 9.0e9 -999.9 -999.9 -999.9
RC VD
(s/cm) (cm/s)
-999.9 0.02
Pb 1 3
Garland (1982), Conference Proceedings, 849-858
GRASS - WIND TUNNEL
2.0, -999.9
-999.9, -999.9
-999.8, -999.9
1 1.0
0.4
1.
- zO estimated(cm), zd(cm)
- ws measurement ht. (m), temp. meas. ht. (m)
- LAI(estimated), vegetation state
- no. of diameters, density (gm/cm**3)
- diameter (microns)
- mass fraction
MM-DD-YY
ENDDATA
B HR E HR - WS
(1st) (1st) (m/s)
-999.9
TEMP
(C)
20.
SW RAD
(W/m**2)
-999.9
USTAR
(m/s)
0.35
MONIN
(m)
9.0e9
HEAT FLUX
(W/m**2)
-999.9
RA
(s/cm)
-999.9
RD
(s/cm)
-999.9
RC
(s/cm)
-999.9
VD
(cm/s)
0.02
Pb 1 3
Garland (1982), Conference Proceedings, 849-858
GRASS - WIND TUNNEL
2.0,
-999.9,
-999.8,
1
0.04
1
-999.9
-999.9
-999.9
.0
1.
- zO estimated(cm), zd(cm)
- ws measurement ht. (m), temp. meas. ht.
- LAI(estimated), vegetation state
- no. of diameters, density (gm/cm**3)
- diameter (microns)
- mass fraction
(m)
MM-DD-YY
ENDDATA
FEOOH
Garland
GRASS
2-0,
10.0,
-999.9,
1 1
2.8
1.
MM-DD-YY
05-21-81
09-23-81
ENDDATA
B HR
(1st)
2 3
(1982),
-999.9
-999.9
-999.9
.0
B HR
(1st)
E HR WS
(1st) (m/s)
-999.9
TEMP SW RAD USTAR MONIN
(C) (W/m**2) (m/s) (m)
20. -999.9 0.35 9.0e9
HEAT FLUX
(W/m**2)
-999.9
RA RD
(s/cm) (s/cm)
-999.9 -999.9
RC VD
(s/cm) (cm/s)
-999.9 0.09
Conference Proceedings, 849-858
E HR WS
(1st) (m/s)
2.5
3.0
- zO estimated(cm), zd(cm)
- estimated ws meas. ht. (m),
- LAI (estimated), vegetation
- no. of diameters, density
- diameter (microns)
- mass fraction
TEMP SW RAD USTAR MONIN
(C) (W/m**2) (m/s) (m)
20.0 -999.9 -999.9 9.0e9
20.0 -999.9 -999.9 9.0e9
temp. meas.
state
HEAT FLUX
(W/m**2)
0.0
0.0
ht. (m)
RA RD
(s/cm) (s/cm)
-999.9 -999.9
-999.9 -999.9
RC VD
(s/cm) (cm/s)
-999.9 0.05
-999.9 0.12
-------
FEOOH 2 3
Garland <1982), Conference Proceedings, 849-858
GRASS - WIND TUNNEL
2.0,
10.0,
-999.9
-999.9
-999.8, -999.9
1 1.0
2.8
1.
MM-DD-YY
06-01-81
08-17-81
ENDDATA
8 HR E HR
(1st) (1st)
US
(m/s)
3.5
3.5
- zO estimated(cm), zd(cm)
- estimated us meas. ht. (m), temp. meas. ht. (m)
- LAI(estimated), vegetation state
- no. of diameters, density
- diameter (microns)
- mass fraction
TEMP SU RAO USTAR NONIN
(C) (W/m**2) (m/s) (m)
20.0 -999.9 -999.9 9.0e9
20.0 -999.9 -999.9 9.0e9
HEAT FLUX RA RD RC VD
(U/m**2) (s/cm> (s/cm) (s/cm) (cm/s)
0.0 -999.9 -999.9 -999.9 0.18
0.0 -999.9 -999.9 -999.9 0.12
FEOOH 1 3
Garland (1982), Conference Proceedings, 849-858
GRASS
- zO estimated(cro), zd(cm)
- estimated ws meas. ht. (m).
2.0, -999.9
10.0, -999.9
-999.9, -999.9
1 1.0
3.8
1.
MM-DD-YY B HR E HR
(1st) (1st)
09-30-82
ENDDATA
temp. meas. ht. (m)
- LA!(estimated), vegetation state
- no. of diameters, density
- diameter (microns)
- mass fraction
TEMP SW RAD USTAR MONIN HEAT FLUX RA RD RC VD
(m) (W/m**2) (s/cm) (s/cm) (s/cm) (cm/s)
9.0e9 0.0 -999.9 -999.9 -999.9 0.42
WS TEMP SW RAD USTAR
(m/s) (C) (W/m**2) (m/s)
2.4 20.0 -999.9 -999.9
(m)
0.17 0.16 0.17 0.17 0.16 0.17
Fine Prt 12 4
Wesely et al (1983), BLM, 27, 237-255.
LEAFLESS DECIDUOUS FOREST IN WINTER (North Carolina)
100., -999.9 - zO estimated(cnr). zd(cm)
39.0, 42.0 - ws measurement ht. (m), temp. meas. ht.
-999.9, 3.0 - LAI(estimated), vegetation state
6 1.0 - no. of diameters, density (gm/cm**3)
0.05 0.06 0.07 0.08 0.09 0.1 - particle diameters
- mass fraction
TEMP
(C)
9.2
15.4
14.8
13.2
15.0
15.2
7.5
8.4
8.5
9.2
9.7
9.7
MM-OD-YY
01-26-81
01-26-81
01-26-81
01-27-81
01-27-81
01-27-81
01-28-81
01-28-81
01-28-81
01-28-81
01-28-81
01-28-81
ENDDATA
B HR E HR
(1st) (1st)
10:00 10:30
17:30 18:00
18:00 18:30
10:30 11:00
11:30 12:00
12:30 13:00
10:30 11:00
11:30 12:00
12:00 12:30
12:30 13:00
13:00 13:30
15:00 15:30
WS
(m/s)
4.0
2.1
2.5
2.5
2.6
1.8
3.6
2.7
2.9
3.3
2.8
1.5
SW RAD
USTAR
l/m**2) (m/s)
-999
-999
-999
-999
-999
-999
-999
-999
-999
-999
-999
-999
.9
.9
.9
.9
.9
.9
.9
.9
.9
.9
.9
.9
0
0
0
0
0
0
0
0
0
0
0
.64
.26
.28
.38
.38
.25
.56
.57
.48
.56
.57
0.35
MONIN HEAT FLUX
(m)
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
-999e9
(W/m**2)
49.
-16.
-20.
85.
56.
26.
43.
209.
89.
231.
237.
55.
0
0
0
0
0
0
0
0
0
0
0
0
RA
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RD
(s/cm)
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
9
9
9
9
9
9
9
9
9
-999.9
-999.9
-999.9
RC
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
VD
(cm/s)
0.010
0.020
0.120
0.110
0.070
0.140
0.680
0.320
0.800
0.550
0.200
0.090
-------
18 3
esely et al (1982), Conference Proceedings, 943-952
lORT GRASS, TEXAS
1.0, -999.9 - zO estimated(cm), zd(cm)
39.0, 42.0 - ws measurement ht. (m), temp. mess. ht. (m)
-999.9, -999.9 - LAI(estimated), vegetation state
12 1.0 - no. of diameters, density (gm/cm**3)
0.10 0.13 0.16 0.19 0.23 0.29 0.36 0.44 0.54 0.66 0.81 1.0 - particle diameters
0.05 0.10 0.14 0.19 0.16 0.11 0.05 0.04 0.05 0.06 0.04 0.0 - fractional mass
MM-DO-YY
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
08-25-78
ENDDATA
B HR
(1st)
11:18
11:48
12:18
12:48
13:18
13:48
14:18
14:48
15:18
15:48
16:18
16:48
17:18
17:48
18:18
18:48
19:18
19:48
E HR
(1st)
11:48
12:18
12:48
13:18
13:48
14:18
14:48
15:18
15:48
16:18
16:48
17:18
17:48
18:18
18:48
19:18
19:48
20:18
US
(m/s)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9'
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
TEMP
(C)
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
30.
SU RAD
(U/m**2)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
USTAR MOW IN
(m/s) (m)
0.24 -999e9
0.25 -999e9
0.26 -999e9
0.32 -999e9
0.20 -999e9
0.21 -999e9
0.22 -999e9
0.33 -999e9
0.24 -999e9
0.27 -99969
0.33 -999e9
0.34 -999e9
0.32 -999e9
0.25 -999e9
0.27 -999e9
0.28 -999e9
0.20 -999e9
0.18 -999e9
HEAT FLUX
(W/m**2)
144.
135.
156.
170.
173.
112.
45.
105.
102.
66.
52.
15.
-14.
-29.
-41.
-36.
-35.
-42.
RA
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RD
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RC
(s/cm)
-999.9
-999.9
-999.9
.-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
VD
(cm/s)
0.220
0.140
0.110
0.150
0.230
0.240
0.150
0.150
0.190
0.150
0.160
0.290
0.200
0.200
0.250
0.150
0.070
0.140
PART 61 5
Lorenz & Murphy(1989)
PINE PLANTATION
28.0, 790.0
9.8, -999.9
9.0 -999.9
6 1.
0.5 0.6 0.7 0.8
0.17 0.16 0.17 0.17 I
MM-OD-YY B HR E HR
(1st) (1st)
BLM, 46, 355-366.
- zO (cm), zd(cm)
- ws measurement ht. (m), temp. meas. ht.
- LAI, vegetation state
- no. of diameters, density (gm/cm**3)
0.9 1.0 - particle diameters (microns)
1.16 0.17 - mass fraction
US TEMP SU RAO USTAR MOW IN HEAT FLUX
(m/s)
1.84
2.22
3.22
2.07
2.16
2.26
2.10
2.16
2.17
2.28
2.39
2.38
(C) (U/m**2)
17.8
17.8
17.8
17.8
17.8
17.8
17.3
17.8
17.8
17.8
17.8
17.8
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
(m/s)
-999.
-999.
-999.
-999.
-999.
-999.
-999.
'-999.
-999.
-999.
-999.
-999.
9
9
9
9
9
9
9
9
9
9
9
9
9.
9.
9.
9.
9.
9.
9.
9.
9.
(m) (U/m**2)
e09
e09
e09
e09
e09
e09
e09
e09
e09
9.e09
9.
9.
e09
e09
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
. (m)
RA
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RD
(s/cm)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
RC
VD
(s/cm) (cm/s)
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
0.15
0.02
0.04
0.13
0.13
0.11
0.17
0.22
0.19
0.17
0.14
0.15
-------
ENDDATA
2.27
2.31
2.33
2.43
2.48
2.53
2.54
2.64
2.87
2.59
2.64
2.66
2.81
3.04
3.54
2.32
2.36
2.41
2.46
2.62
2.63
2.70
2.72
2.80
2.30
2.94
3.12
3.13
3.20
3.28
3.23
3.22
3.10
3.25
3.34
3.50
3.64
3.73
3.84
3.88
4.02
3.90
3.89
4.08
4.39
3.85
4.41
2.81
3.39
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.8
17.3
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.
9
-999.9
-999.9
-999.
9
-999.9
-999.9
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.c09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
9.e09
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.
9
-999.9
-999.9
-999.
.-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
9
9
9
9
9
9
9
9
9
-999.9
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
-999.
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
-999.9
0.24
0.23
0.30
0.34
0.28
0.32
0.25
0.14
0.14
0.35
0.31
0.26
0.23
0.25
0.20
0.45
0.44
0.52
0.49
0.52
0.41
0.39
0.39
0.43
1.17
0.49
0.56
0.61
0.63
0.75
0.90
0.97
0.34
0.39
0.48
0.32
0.38
0.44
0.42
0.36
0.46
0.59
0.91
0.72
0.69
1.32
0.90
0.00
0.00
-------
APPENDIX E
Implementation of the Modified Source
Depletion Method in ISC2
-------
E.I Overview of Method
Horst (1983) [reproduced as Appendix F] describes a method for incorporating the
effects of deposition on the vertical distribution of material in a plume, without resorting to a
full surface depletion treatment (see the discussion of plume depletion techniques in Section 3).
The method includes a vertical profile adjustment factor as well as a source depletion factor, so
that concentrations in the lower portion of the plume approximate those produced by the
surface depletion model. This adjustment factor is important because deposition is proportional
to near-surface concentrations. The source depletion method by itself overestimates near-
surface concentrations, thereby overpredicting deposition rates which hastens removal of
material from the plume.
Let V(x,z,h) denote the vertical distribution of plume material in the absence of
deposition. In terms of the notation used by Horst (1983),
= u
(E-l)
where D is the crosswind-integrated concentration distribution for a plume at a height h above
the ground, released from a source of unit source strength. Then the vertical distribution factor
that is modified to account for the depletion resulting from deposition, Vd (x,z,h), is defined to
be:
FQ(x) PtoQ (E-2)
FQ(x) is the fraction of material that remains in the plume'at the downwind distance x (i.e., the
mass that has not yet been deposited on the surface). This factor may be thought of as a source
depletion factor, a ratio of the "current" mass emission rate to the original mass emission rate.
P(x,z) is the vertical profile adjustment factor.
The effect of Equation (E-2) is illustrated in Figures E-l and E-2. Figure (E-l) displays
a depletion factor FQ, and the corresponding profile correction factor P(x,z) for a distance at
which oz is 1.5 times the plume height. This assures that the plume has been in contact with the
ground for a long enough time that significant deposition has occurred. The depletion factor is
constant with height, whereas the profile correction shows that most of the material is lost from
the lower portion of the plume. Figure (E-2) compares the vertical profile of concentration both
with and without deposition and the corresponding depletion of material from the plume. The
depleted plume profile is computed using Equation (E-2).
E-l
-------
2.0 -i
N 1.5 -
O
£
1.0 -
N
-------
2.0 n
N 1.5 -
D
en
N
1.0 -
CD
0.5 -
0.0
Original Profile
Depleted Profile
0.0
0.5
1.0
1.5
2.0
Concentration
FIGURE E-2. VERTICAL PROFILE OF CONCENTRATION BEFORE AND AFTER
APPLYING FQ AND P(x,z) SHOWN IN FIGURE E-l.
E-3
-------
FQ(x) is a function of the total deposition velocity (vd), V(x^d,h), and
FQ(x) = EXP
(E-3)
where zd is a height near the surface at which the deposition flux is calculated. This equation
reflects the fact that the material removed from the plume by deposition is just the integral of
the deposition flux over the distance that the plume has traveled. For general forms of V(x,z,h)
and P(x^), Equation (E-3) is evaluated numerically.
The deposition velocity for particles generally contains a component related to the
settling velocity, vs. A "tilted plume" is used to simulate the effect of gravitational settling on the
plume as a whole. This approximation entails replacing the plume height, h, in Equations (E-l)
and (E-2) with
(E-4)
For large travel-times, hs can become less than zero. However, the tilted plume approximation
is not a valid approach in this region. Therefore, a minimum value of zero must be imposed on
hj. In effect, this limits the settling of the plume, although the deposition velocity continues to
account for gravitational settling near the surface.
The profile correction factor P(x,z) is developed by Horst (1983) for the case in which
reflection of material from the mixing lid is not important. He finds
-'(1 -EXP[-vs
-1
S o
(E-5)
where R(z,zd) is an atmospheric resistance to vertical transport. When the product vsR(z,zd) is
of order 0.1 or less, the exponential function is approximated (for small argument) to simplify
P(x,z):
-i
(E-6)
E-4
-------
This simplification is important, since the integral in Equation (E-6) can be computed using
analytical approximations for many forms of R(z,zd) that are consistent with the Briggs' formulas
for oz (Gifford, 1976). Typically, only the largest particles may have a settling velocity v, large
enough to require the numerical integration of Equation (E-5).
The atmospheric resistance is defined as
where K(z) is the vertical eddy diffusivity. Because we will be using empirical expressions for o2,
K(z) should be consistent with these. Horst (1983) points out that
K = u ^ (E-8)
and that
(E-9)
for a Gaussian plume from a ground-level source, where z is the mean height of the
distribution of mass in the plume. Using Equation (E-9) to map z to oz, Equation (E-7) is
represented by
1 dx
This allows R(z,zd) to be evaluated for particular forms of oz. Horst provides solutions for
o = ax
oz = ox(l + foe)-1/2 (E-H)
or = ox(l + fee)'1
which encompass the forms used for Briggs' rural dispersion curves, as well as Briggs' urban
curves for stability classes C, D, E, and F. The solution for urban classes A and B,
o = «(1 * fee)1/2 (E-12)
E-5
-------
is derived in Section E-2. (Solutions to Equations (E-7) and (E-6) for each form of oz are listed
in Section E-2).
E.2 Extension of Solutions for Urban Classes A and B
The Briggs' curves for oz for urban locations during stability classes A and B have the
form of Equation (E-12). Therefore, Equation (E-10) becomes
ax
fl
\ Tt
(E-13)
_L m
l + bx
The limits are implicit functions obtained from Equation (E-9):
(E-14)
That is, x(z) is the distance at which oz equals z / v/2~7rc . If both sides are squared, x(z) can be
expressed as the root of a cubic equation. In developing the FORTRAN code to implement
Horst's method, we use an iterative method to solve for the root of Equation (E-14).
Adding Equation (E-13) to the solutions given by Horst for the other forms of oz, we
have:
= ax:
1 -L In (zjz4)
n au v '
(E-15)
= axl(\
(E-16)
E-6
-------
oz = axl(l
(E-17)
oz = ax(l + fee)1
fec(z) - 1
bx(zd)
bx(zd) -
(E-18)
The profile correction factor, P(x,z), requires the integral of the product of R(z,zd) and
the vertical distribution factor V(x,z,h), which is dominated by a series of exponential functions
of height. An analytic solution to this integral is possible for the R(z,zd) terms involving ln(z/zd),
(z-zd) and (z^z/); but the complex form of Equation (E-18), coupled with the supplementary
relation in Equation (E-14), precludes such a solution. After trying several approximation
techniques, the solution for P(x,zd) with R(z,zd) given by Equation (E-18), was approximated as
follows.
First, a program was developed that solves the integral in Equation (E-6) numerically.
This not only allows us to test various approximate results, it can also serve as a numerical
solver if no analytical approximations are found to be adequate. Then, we developed an
approximate expression for R(z,zd) for small z, which is facilitated by the fact that the constant
b = 0.001 for urban classes A and B:
or
(E-19)
where k =
= 2b nr
a N| 2 '
E-7
-------
This allows Equation (E-18) to be written as
1-Lfc
n ait
(1 + kz)V4 - 1 (1 + fa,)* + 1
fe)
V4
(E-20)
Further expanding (1 + kz)1/4 as 1 + kz/4, the natural log expression in Equation (E-18)
becomes
(E-21)
This gives a leading term that is the same as Equation (E-15), for oz = ax, and the approximate
result for P is
-1
(E-22a)
where
= ln
*
-------
bx):
- v.
In
- 1 +
(E-25)
ua
In
'v/2or
,-1
n 8 )
(E-26)
For the last form, k
2b fT
a N 2 '
and
°t(l ~ -0006
0.6724 o
300m
300m
and
1000 a.
^ 1000m
> 1000m
The approximation to the integral in P(x,zd) for oz = ax(l + bx)1/2 matches the numerical
solution to within 1% for zd = 0.03 m and oz <; 4000 m, and it matches to within about 2% for
zd = 1 m over the same height range.
E.3 Mixing-Lid Treatment
The results presented above do not include the presence of a mixing lid. With such a lid,
the profile correction factor must operate only within the mixed layer, so that the upper limit in
the integral for P(x,zd) is zt, not infinity, in Equation (E-6). Furthermore, the standard
formulation of V(x,z,h) is also a function of z-v since the distribution of material in the plume is
E-9
-------
"reflected" from z = z,. In principle, the additional reflections in the V(x,z,h,Zj) just add to the
number of terms in P(x,zd), since the form of the integral in Equation (E-6) remains the same.
However, most of the emphasis in obtaining P(x^ Zj), and compare the results with Equations (E-23) through
(E-26).
In the well-mixed limit,
_!
*,-
(E-27)
so that P(x,zd) involves the integral of terms involving just ln(z/zd), (z-zd), and (z2-zd2), since the
exponentials are not present in Equation (E-27). Performing the integrations yields functions
that are equivalent to Equations (E-23) through (E-25), except oz is replaced by a constant times
(E-28)
(E-29)
(E-30)
Therefore, the effect of the lid is to limit the size of oz in evaluating P(x,zd). Hence, we retain
Equations (E-23) through (E-25), but restrict the use of oz to values no larger than those that
satisfy Equations (E-28) through (E-30). In other words, caps are placed on the value of oz used
when a mixed layer exists.
Equation (E-26), for oz = ax(l + bx)1/2, is a special case. We were able to solve the
integral for the case of complete mixing (Equation E-27), but the result involves many terms.
Since the approximate form used in Equation (E-26) performed well in the absence of a mixing
lid, we also tested a version in which oz was replaced by z4 as in Equation (E-28). This result
E-10
-------
also performed well when compared to the numerical integration, so this more compact result
was adopted.
E.4 Numerical Integration For P(x,zd)
Because Equation (E-3) involves the numerical evaluation of an integral over the
distance from the source to each receptor, an analytic representation of P(x,zd) is preferred in
order to streamline the computations. As discussed in Section E.I, P(x,zd) can be represented
by simple analytic functions so long as vs R(z,zd) z 0.1. However, larger and denser particles
(greater than 10 \im in diameter) have settling velocities great enough to violate this condition
at times. For these situations, the full expression for P(x,zd) (Equation E-5) must be solved
numerically as well. This means that each point evaluated in Equation (E-3) involves a
numerical evaluation of Equation (E-5). This can be time-consuming.
The subroutines developed make use of a general integration routine. It subdivides the
interval into more and more equally-spaced segments until the value of the integral converges to
within an imposed tolerance. Further study is recommended to optimize these integration
procedures. For example, the tolerance level might be too restrictive. Or, an integration
technique might be specifically designed for the integrand. We know the form of V(x,z,h) and
R(z,zd), and may be able to increase the efficiency of the integration by designing an algorithm
that "knows" where the integrand changes most rapidly, and least rapidly. Fewer points are
needed to integrate across regions in which the variation of the integrand is nearly linear.
There is also a possibility that the solution to Equation (E-5) can be approximated
simply enough to avoid its numerical solution. We have been able to recast the integral in
Equation (E-5) to one of the form:
(zlzd}v'A e -W * */° dz
for R(z,zd) of the form
R(Vd) = A ]n(zlzd) + B(z - zd] + c(z2 - zd2) (E-32)
With a suitable definition of a new variable of integration, it appears that Equation (E-5) could
be solved to yield a representation made up of the product of a Gamma function, an exponential
function, and a parabolic cylinder function. Such a solution has not been completely worked
out, since it is not clear that such a representation would lead to a more efficient evaluation of
the integrals.
E-ll
-------
References
Gifford, FA., Jr., 1976: Turbulent diffusion - "typing schemes: A review. NucL Saf., 17, 68-86.
Horst, T.W., 1983: A correction to the Gaussian source-depletion model. In Precipitation
Scavenging, Dry Deposition and Resuspension, H.R. Pruppacher, R.G. Semonin, W.G.N.
Slinn, eds., Elsevier, NY.
E-12
-------
TECHNICAL REPORT DATA
(Please read Instructions on reverse before completing)
1. REPORT NO.
EPA-454/R-92-017
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Development and Testing of Dry Deposition Algorithms
5. REPORT DATE
May 1993
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Joseph A. Scire, Gary E. Moore and David Strimaitis
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Sigma Research Corporation
196 Baker Avenue
Concord, MA 01742
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68D9007, Work Assignment 3-1
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Air Quality Planning And Standards
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
EPA Work Assignment Manager: Jawad S. Touma
16. ABSTRACT
The primary objective of this study is to implement a generalized dry deposition algorithm for
paniculate matter, suitable for regulatory use, into the Industrial Source Complex (ISC2) model.
Reviews were conducted of methods for computing dry deposition velocities, plume depletion, and
certain micrometeorological parameters from routinely-available observations. Several observational data
bases were identified from the literature and used in the testing and evaluation of ten particle deposition
models. Recommendations for computing particle deposition velocities, plume depletion, and
meteorological variables were made. These techniques have been incorporated into a revised version of
the ISC2 model and related processor programs.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Atmospheric Dispersion Modeling
Dry Deposition
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (Report)
Unclassified
21. NO. OF PAGES
128
20. SECURITY CLASS (Page)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION IS OBSOLETE
------- |