EPA 600/R-10/007 | August 2009 | www.epa.gov/ord
OtrA
United States
Environmental Protection
Agency
Development of Gridded Fields of
Urban Canopy Parameters for
Advanced Urban Meteorological and
Air Quality Models
Office of Research arid Development
National Exposure Research Laboratory, Atmospheric Modeling and Analyses Division

-------
EPA/600/R-10/007 August 2009 www.epa.gov/ord
Development of Gridded Fields of
Urban Canopy Parameters for
Advanced Urban Meteorological and
Air Quality Models
Steven J. Burian
Department of Civil and Environmental Engineering
University of Utah
Jason Ching
Atmospheric Modeling and Analyses Division
National Exposure Research Laboratory
Office of Research and Development
United States Environmental Protection Agency
Research Triangle Park, NC
August 2009

-------
Notice
The information in this document has been funded wholly, or in part, by the U.S. Environmental
Protection Agency (EPA) Office of Research and Development under Contract No. PO-2D-
6217-NTEX issued to the University of Arkansas. Mention of trade names or commercial
products does not constitute endorsement or recommendation for use.

-------
Acknowledgments
This report is based on research completed by Steven J. Burian, University of Arkansas (now at
the University of Utah), for the U.S. Environmental Protection Agency National Exposure
Research Laboratory (NERL), Atmospheric Modeling Division. Jason Ching, NERL's
Atmospheric Modeling Division (now with its Atmospheric Modeling and Analyses Division),
served as the technical project manager and the driving force behind the project. Special
acknowledgment is given to Sylvain Dupont for guidance on parameter estimation procedures.
Woo Suk Han, Srinivas Pradeep Velugubantla, and Sri Ram Kumar Maddula contributed
significant time and effort to manage and process the data described in this report.

-------

-------
Abstract
Urban dispersion and air quality simulation models applied at various horizontal scales require
different levels of fidelity for specifying the characteristics of the underlying surfaces. As the
modeling scales approach the neighborhood level (~1-km horizontal grid spacing), the
representation of urban morphological structures requires much greater detail. To provide the
most accurate surface characterization possible for an air quality modeling study of Houston,
TX, airborne Light Detection and Ranging (LIDAR) data were obtained from TerraPoint LLC at
1-m horizontal grid cell spacing for Harris County, TX, an area of approximately 5800 km2. The
data were managed in the ESRI ArcView 3.3 and ArcGIS 8.2 GIS software packages. Scripts
and computer codes were written in Avenue, Visual Basic for Applications, and Fortran to
compute 20 urban canopy parameters (UCPs), including building height statistics and
histograms, height-to-width ratio, plan area density function, frontal area density function,
roughness length, displacement height, mean orientation of streets, and sky view factor. In
addition, procedures were developed to approximate several UCPs that could not be
determined from the elevation data, including surface cover type, building material fraction, and
percent directly connected impervious area.
The modeling phase of the study involved applying an urban energy budget model, DA-SM2-U,
the Penn State/National Center for Atmospheric Research Mesoscale Model (MM5), and the
U.S. Environmental Protection Agency (EPA) Models-3 Community Multi-scale Air Quality
(CMAQ) modeling system. To accommodate the modeler needs, area-weighted UCPs were
defined for 82,368 1-km2 grid cells corresponding to the DA-SM2-U, MM5, and CMAQ modeling
domains. Phase I of the UCP computation project focused on a 1653-km2 section centered on
the downtown of the City of Houston. A building footprint data layer was developed and
validated using 0.5-m horizontal resolution True Color orthophotos. Using the building footprint
data layer refined by the aerial photos and the LIDAR data products, UCPs were calculated for
the buildings only, the vegetation only, and the full canopy. The second phase of the project
involved computing only full-canopy UCPs for the remaining 3589 km2 of Harris County not
included in Phase I. The third, and final, phase of the UCP computation project involved
developing an accurate land use data layer for the 1653-km2 Phase I study area, correlating the
UCPs in the study area to the underlying land use, extrapolating building and vegetation UCPs
to the 3589 grid cells in the Phase II study area, and extrapolating all UCPs to the 77,126 grid
cells outside of the Phases I and II study areas. In total, the results spreadsheets accompanying
this report contain approximately 84 million UCP values.
v

-------

-------
Table of Contents
List of Tables		ix
List of Figures		x
Nomenclature		xiii
1.	Project Overview		1
1.1	Project Objectives		1
1.2	Houston Study Area		1
2.	Houston Urban Database		4
2.1	TerraPoint Elevation Data		4
2.2	Building Footprints		4
2.3	Land Use/Land Cover		6
3.	Urban Canopy Parameters		9
3.1	Building Height Characteristics		9
3.2	Vegetation Height Characteristics		11
3.3	Canopy Height Characteristics		11
3.4	Plan Area Density Function		11
3.5	Top Area Density Function		12
3.6	Frontal Area Density Function		12
3.7	Sky View Factor		13
3.8	Displacement Height and Roughness Length		13
3.9	Mean Orientation of Streets		15
3.10	Building Material		15
3.11	Percent Directly Connected Impervious Area		15
3.12	Fraction Land Cover		16
4.	Urban Canopy Parameter Extrapolation Based on Underlying Land Use		17
4.1	Average Urban Canopy Parameter Derivation		19
4.1.1	Building Height Characteristics		19
4.1.2	Vegetation Height Characteristics		20
4.1.3	Canopy Height Characteristics		22
4.1.4	Sky View Factor		24
4.1.5	Displacement Height and Roughness Length		24
4.1.6	Building Material		24
4.1.7	Percent Directly Connected Impervious Area		25
4.1.8	Fraction Land Cover		25
4.2	UCP Assessment and Validation		26
4.2.1	Building Height Characteristics		27
4.2.2	Canopy Height Characteristics		30
5.	Houston Urban Canopy Parameters		50
5.1	Building Height Characteristics		50
5.2	Vegetation Height Characteristics		50
5.3	Canopy Height Characteristics		60
5.4	Displacement Height and Roughness Length		60
5.5	Sky View Factor		60
5.6	Percent Directly Connected Impervious Area		60
vii

-------
Table of Contents (cont'd.)
6.	Summary	 72
7.	References	 73
viii

-------
List of Tables
No.
1	Digital Datasets Obtained or Created for Urban Canopy Parameter (UCP)
Computations	 5
2	Anderson Classification System Used in U.S. Geological Survey (USGS) Land
Use/Land Cover Datasets	 7
3	Anderson Level 2 Land Use Types in Modeling Domain	 8
4	Urban Canopy Parameters Computed for This Project	 9
5	Numeric Angles Corresponding to Street Orientations	 15
6	Comparison of Land Use Distribution in the 1653-km2 Phase I Study Area for
Original USGS Land Use Dataset and Revised Dataset	 18
7	Comparison of Areas of Each Land Use Type in the UCP Derivation and
Validation Areas	 20
8	Sample Site Descriptions	 21
9	Calculation and Extrapolation Extent for Each Urban Canopy Parameter	 22
10	Average Building Height Characteristics per Land Use in the 1235-km2 UCP
Derivation Area	 23
11	Mean Vegetation Height per Land Use Type	 28
12	Mean Canopy Height per Land Use Type	 32
13	Mean Sky View Factor per Land Use Type	 36
14	Mean Displacement Height per Land Use Type	 37
15	Mean Roughness Length per Land Use Type	 38
16	Assumed Building Material Fraction per Land Use Type	 39
17	Mean Percent Directly Connected Impervious Area per Land Use Type	 40
18	Fraction Land Cover per Land Use Type	 41
19	Comparison Statistics for the Calculated and Extrapolated Building Height
Characteristics	 41
20	Comparison of Mean Building Height and Standard Deviation for Selected
USGS Level 2 Land Use Types	 42
21	Comparison Statistics for the Calculated and Extrapolated Building Height
Histograms, Plan Area Density, Top Area Density, and Frontal Area Density	 46
22	Comparison Statistics for the Calculated and Extrapolated Canopy Height
Characteristics	 46
23	Comparison Statistics for the Calculated and Extrapolated Canopy Plan Area
Density and Top Area Density	 49
ix

-------
List of Figures
No.
1	Houston vicinity map and modeling domain	
2	Land use/land cover of the modeling domain inner grid	
3	Three phases of UCP computation project	
4	Land use of the Phase I building study area	
5	UCP derivation and validation zones overlaying land use	
6	Building height histograms	
^ Building plan area density computed for the 1235-km2 UCP derivation section in
Houston	
g Building top area density computed for the 1235-km2 UCP derivation section in
Houston	
9 Building frontal area density computed for the 1235-km2 UCP derivation section in
Houston for a wind from the north direction	
1f) Vegetation plan area density computed for the 1235-km2 UCP derivation section in
Houston	
^ ^ Vegetation top area density computed for the 1235-km2 UCP derivation section in
Houston	
12 Vegetation frontal area density computed for the 1235-km2 UCP derivation section in
Houston for a wind from the north direction	
,|2 Canopy plan area densities computed for the 1235-km2 UCP derivation section in
Houston	
^ Canopy top area densities computed for the 1235-km2 UCP derivation section in
Houston	
15	Canopy frontal area densities computed for the 1235-km2 UCP derivation section in
Houston for a wind from the north direction	
16	Scatter plot of extrapolated versus calculated mean building height for the
418 grid cells in the validation area	
17	Scatter plot of extrapolated versus calculated standard deviations of building
height for the 418 grid cells in the validation area	
18	Scatter plot of extrapolated versus calculated plan-area-weighted mean building
height for the 418 grid cells in the validation area	
19	Scatter plot of extrapolated versus calculated wall-to-plan area ratio for the
418 grid cells in the validation area	
20	Scatter plot of extrapolated versus calculated height-to-width ratio for the
418 grid cells in the validation area	
21	Scatter plot of extrapolated versus calculated mean canopy height for the
418 grid cells in the validation area	
22	Scatter plot of extrapolated versus calculated roughness lengths for the
418 grid cells in the validation area using the Macdonald et al. (1998) equations
and a north wind azimuth	
23	Scatter plot of extrapolated versus calculated displacement heights for the
418 grid cells in the validation area using the Macdonald et al. (1998) equations
and a north wind azimuth	
24	Scatter plot of extrapolated versus calculated roughness lengths for the
418 grid cells in the validation area using the Raupach (1994) equations and a
north wind	
25	Scatter plot of extrapolated versus calculated displacement heights for the
418 grid cells in the validation area using the Raupach (1994) equations and a
north wind	
2
3
3
17
19
24
25
26
27
29
30
31
33
34
35
43
43
44
45
45
46
47
47
48
48
x

-------
List of Figures (cont'd.)
No.
26	Spatial distribution of mean building height in Harris County	
27	Spatial distribution of standard deviation of building height in Harris County	
28	Spatial distribution of number of buildings in Harris County	
29	Spatial distribution of number of buildings in the modeling domain	
30	Spatial distribution of average wall-to-plan area ratio in Harris County	
31	Spatial distribution of average wall-to-plan area ratio in the modeling domain	
32	Spatial distribution of average building height-to-width ratio in Harris County	
33	Spatial distribution of average building height-to-width ratio in the modeling domain..
34	Spatial distribution of building plan area fraction in Harris County	
35	Spatial distribution of building plan area fraction in the modeling domain	
36	Spatial distribution of building frontal area index in Harris County	
37	Spatial distribution of building frontal area index in the modeling domain	
38	Spatial distribution of mean vegetation height in Harris County	
39	Spatial distribution of mean vegetation height in the modeling domain	
40	Spatial distribution of vegetation plan area fraction in Harris County	
41	Spatial distribution of vegetation plan area fraction in the modeling domain	
42	Spatial distribution of vegetation frontal area index in Harris County	
43	Spatial distribution of vegetation frontal area index in the modeling domain	
44	Spatial distribution of mean canopy height in Harris County	
45	Spatial distribution of mean canopy height in the modeling domain	
46	Spatial distribution of canopy plan area fraction in Harris County	
47	Spatial distribution of canopy plan area fraction in the modeling domain	
48	Spatial distribution of canopy frontal area index in Harris County	
49	Spatial distribution of canopy frontal area index in the modeling domain	
50	Spatial distribution of roughness length in Harris County calculated using the
Macdonald et al. (1998) set of equations for a north wind azimuth	
51	Spatial distribution of roughness length in the modeling domain calculated using the
Macdonald et al. (1998) set of equations for a north wind azimuth	
52	Spatial distribution of displacement height in Harris County calculated using the
Macdonald et al. (1998) set of equations for a north wind azimuth	
^2 Spatial distribution of displacement height in the modeling domain calculated using
the Macdonald et al. (1998) set of equations for a north wind azimuth	
54	Spatial distribution of roughness length in Harris County calculated using the
Raupach (1994) set of equations for a north wind azimuth	
55	Spatial distribution of roughness length in the modeling domain calculated using the
Raupach (1994) set of equations for a north wind azimuth	
56	Spatial distribution of displacement height in Harris County calculated using the
Raupach (1994) set of equations for a north wind azimuth	
57	Spatial distribution of displacement height in the modeling domain calculated using
the Raupach (1994) set of equations for a north wind azimuth	
58	Spatial distribution of roughness length in Harris County calculated using the
Bottema (1997) set of equations for a north wind azimuth	
59	Spatial distribution of roughness length in the modeling domain calculated using the
Bottema (1997) set of equations for a north wind azimuth	
60	Spatial distribution of displacement height in Harris County calculated using the
Bottema (1997) set of equations for a north wind azimuth	
61	Spatial distribution of displacement height in the modeling domain calculated using
the Bottema (1997) set of equations for a north wind azimuth	
51
51
52
52
53
53
54
54
55
55
56
56
57
57
58
58
59
59
61
61
62
62
63
63
64
64
65
65
66
66
67
67
68
68
69
69
xi

-------
List of Figures (cont'd.)
No.
62	Spatial distribution of sky view factor in Harris County	 70
63	Spatial distribution of sky view factor in the modeling domain	 70
64	Fraction of each grid cell in Harris County that is directly connected impervious
area	 71
65	Fraction of each grid cell in the modeling domain that is directly connected
impervious area	 71
xii

-------
Nomenclature
Az
A*
Xs
X|/k
H'sky
0
a
empirical coefficient for Macdonald morphometric roughness equations
correction factor for the drag coefficient
height increment
frontal area index
building height-to-width ratio
building wall-to-plan area ratio
roughness sublayer influence function
sky view factor
wind direction
length-weighted mean orientation of streets
two dimensional
three dimensional
Afb(z)	building frontal area density
Afc(z)	canopy frontal area density
Afv(z)	vegetation frontal area density
Apb(z)	building plan area density
ApC(z)	canopy plan area density
Aproj	frontal area of roughness elements projected into plane perpendicular to wind
ApV(z)	vegetation plan area density
At	area of grid cell
Atb(z)	building top area density
Atc(z)	canopy top area density
Atv(z)	vegetation top area density
CD	drag coefficient
CD	compact disc
cd|	parameter in Raupach morphometric roughness equations
CMAQ	Community Multi-scale Air Quality
COH	City of Houston
cR	drag coefficient of an isolated roughness element mounted on the surface
CRE	cumulative relative error
cs	drag coefficient for the substrate surface at height h
CTG	composite theme grid
DCIA	directly connected impervious area
DEM	digital elevation model
DTM	digital terrain model
EPA	U.S. Environmental Protection Agency
ESRI	Environmental Systems Research Institute
GIRAS	Geographic Information Retrieval Analysis System
GIS	geographic information system
7?	mean height
hAW	area-weighted mean height
hc	mean canopy height
HGAC	Houston-Galveston Area Council
k	von Karman
LIDAR	light detection and ranging
LULC	land use/land cover
L(z)	area index
xiii

-------
MM5	Penn State/National Center for Atmospheric Research Mesoscale Model
MrSID	multiresolution seamless image database
NCAR	National Center for Atmospheric Research
NLCD	National Land Cover Dataset
RMSE	root mean square error
sh	standard deviation of height
TIGER	Topographically Integrated Geographic Encoding and Referencing System
TSARP	Tropical Storm Allison Recovery Project
u*	friction velocity
U	large-scale wind speed
UCP	urban canopy parameter
USGS	U.S. Geological Survey
zd	displacement height
zdpi	in-plane sheltering displacement
z0	roughness length
xiv

-------
1. Project Overview
Urban morphological characteristics are required to accurately run many mesoscale
meteorological, surface energy budget, and air quality models. Traditionally, best guess
estimates of urban morphological parameters were made based on literature values and an
underlying base dataset (e.g., land use/land cover) that typically had coarse horizontal
resolution. However, currently the state-of-the-practice is to analyze three-dimensional (3D)
digital datasets of buildings and trees integrated with two-dimensional (2D) digital datasets of
land use/land cover, infrastructure, aerial photographs, and satellite imagery to derive the
necessary urban canopy parameters. The tools used in the analyses include geographic
information system (GIS), image analysis, database management software, and computer
programming of the numerical algorithms.
1.1	Project Objectives
The objectives of this project were as follows.
(1)	Compile and manage the digital datasets necessary to compute urban canopy parameters
for a selected Community Multi-scale Air Quality (CMAQ)/Penn State/National Center for
Atmospheric Research Mesoscale Model (MM5)/DA-SM2-U modeling domain, including the
City of Houston (COH), TX.
(2)	Write computer codes in Avenue, Visual Basic for Applications, and Fortran to process
digital data to compute a set of requested urban canopy parameters (UCPs).
(3)	Execute the computer codes to calculate the following urban canopy parameters.
•	Building-specific parameters: mean height, standard deviation of height, mean height
weighted by plan area, wall-to-plan area ratio, height histograms, plan area density, top
area density, frontal area density, and height-to-width ratio
•	Vegetation-specific parameters: mean height weighted by plan area, plan area density,
top area density, and frontal area density
•	Canopy-specific parameters: mean height weighted by plan area, plan area density, top
area density, and frontal area density
•	General morphological parameters: roughness length, displacement height, sky view
factor, and mean orientation of streets
•	Surface cover parameters: building material fraction, land cover fraction, land use fraction,
and percent directly connected impervious area
(4)	Compile the results into spreadsheets and summarize in a final report.
The project required the following data to be supplied: (1) a database consisting of the bald
earth elevation, building rooftop elevation, and vegetation (trees and large shrubs) elevation
data collected using airborne light detection and ranging (LIDAR) scanning system (in raster
form) at both 1- and 5-m horizontal resolutions; (2) a set of derived building polygons (in vector
form) with elevation as an attribute based on the data for Harris County, TX; and (3) the grid
mesh for Harris County of the nested computational domains in the MM5. As described below,
conditions 1 and 2 were not met satisfactorily, and adjustments had to be made to complete the
project with an acceptable level of quality.
1.2	Houston Study Area
The CMAQ/MM5/DA-SM2-U modeling domain for this project is centered on the Houston
Metropolitan Area in Southeast Texas (see Figure 1). The modeling domain covers an 82,368-
km2 area, including approximately two-thirds land surface and one-third water surface (primarily
the Gulf of Mexico). This project involved the computation of requested UCPs for the entire
1

-------
^7

IS	KB
Houston
1J3? !4sikM*l
vTTx^-vn
X LJir 1
tL
-H
1

M
to



yj
fZ?

Gulf
of
Mexico
Figure 1. Houston vicinity map and modeling domain. The location of Houston within the State of
Texas is shown on the left and the innermost grid of the modeling domain is shown as a red box
surrounding the Houston Metropolitan Area on the right.
82,368-km2 modeling domain. The modeling domain is subdivided into a modeling grid mesh
with 1-km horizontal spatial resolution. Each 1-km2 grid cell (82,368 total) must have all UCPs
defined. The land use/land cover (LULC) for the modeling domain is shown in Figure 2. Overall,
the land surfaces of the modeling domain are predominantly rural, consisting of significant
fractions of Cropland and Pasture and Forest Land. The highest concentration of urban land use
is in the Houston Metropolitan Area located at the left center of Figure 2.
The data processing task was divided into three phases, as illustrated in Figure 3. Phase I of the
UCP computations focused on a 1653-krrr section of the COH and surrounding areas. A
building footprint data layer was developed and validated using 0.5-m horizontal resolution
digital orthophotos. Using the building footprint data layer and refined by the digital orthophotos,
UCPs were calculated for the buildings only, the vegetation only, and the full canopy. Phase II of
the project involved computing only full canopy UCPs for the remaining 3589-krrr of Harris
County not included in Phase I. Phase III of the project involved developing an accurate land
use data layer for the 1653-krrr Phase I study area, correlating the UCPs in the study area to
the underlying land use, extrapolating building and vegetation UCPs to the 3589 grid cells in the
Phase II study area, and extrapolating all UCPS to the 77,126 grid cells outside of the Phase I
and II study areas (labeled as Phase III in Figure 3).
2

-------
Land Use Class
KE31DCNTTAX
¦	GfiHHBIBM*. AND SEAVCCl
¦	UDOtmML
¦	T*AMS*iQ*TATTC*. COM»JN»CAT»e*, UTVJT*
3 NBU5T ANO COftlMEKC CUFLX3
¦	WJCU UUM m fiUtLf-Uf4
om* uiwan o* puilt vp
CRQPWAND AMP BMTVIIt
CONFINED rCEOMO OPS
^ I CKMMH, MNtTAAO* MCi NUttftEfhE*
OIMCT AOWCULTURAL LAN©
MCPUQUS fdWIT LANO
¦	rvt wokjin ram st land
HUrtO	LAND
fO«fSTEO WETLAND
MOHTO«e*T«) WT\AHO
¦IttClillACCOUS HAH Qf: LAND
5m« utt A NO BAV9H KANMLAN&
WOOED RAN fit LA HO
IUKC FKPOKD ROC*
BtACWS
&ANDT AREA INCN-UEACHI
¦	•TRVkUMl
T*AN*ntONAL AREA*
¦	BAYS AND aTUAJCTS
¦	LAKE*
¦	OEUfffOAt
¦	smTAUt AMD CAHAi .1
4kU
0	25 50	100 Kitometers
	1	I	I I	
Figure 2. Land use/land cover of the modeling domain inner grid. Land use data based on the
USGS dataset, classified according to the Anderson level 2 scheme.
Figure 3. Three phases of UCP computation project.
3

-------
2. Houston Urban Database
The Houston GIS Urban Database includes multiple surface topography and surface cover
digital datasets that were purchased for the project and several derivative data products created
during the project. Table 1 lists all datasets that were obtained or created. The data name,
format, and source are given, and a brief description of the dataset is included.
2.1	TerraPoint Elevation Data
TerraPoint LLC provided the base layer of elevation data for this project. The elevation data
products were derived from data collected using LIDAR technology. LIDAR technology
produces x, y, z representation of topography via airborne lasers. Data products are created as
an even distribution of data points in evenly spaced grids. The TerraPoint data products were
spaced at intervals of 1 and 5 m, with a horizontal accuracy of 15 to 20 cm and a vertical
accuracy of 5 to 10 cm.
The following seven data products were included in the delivery from TerraPoint.
(1)	Digital Elevation Model (DEM)—full feature (raster)
(2)	Digital Terrain Model (DTM)—bare earth (raster)
(3)	Ground—ground feature (raster)
(4)	Nonground—nonground feature (raster)
(5)	Building—building feature (raster)
(6)	Vegetation—vegetation feature (raster)
(7)	Building Polygons—building footprint shapefile (vector)
The Building, Vegetation, and Building Polygons were found to be inadequate or unnecessary
for the present project. The data layers did not cover enough area, and the data itself was found
to contain significant errors when cross-referenced with aerial photos taken at approximately the
same time as the LIDAR fly over. The TerraPoint algorithm for separating buildings from
vegetation did not provide a building polygon coverage sufficiently accurate or extensive for the
present project. The most significant deficiency was the incorrect placement of building
polygons or missing polygons. Thus, the Building and Vegetation raster datasets and the
Building Polygon datasets obtained from TerraPoint were not used in the calculations described
below. Instead, other datasets were obtained, derived, and created to meet the project needs.
The UCP calculation algorithms operate with absolute heights of the canopy and objects (as
opposed to top elevations relative to mean sea level). The required canopy height data product
was derived by subtracting the DTM data layer from the Nonground data layer. The Ground and
DEM data products were not needed for this project.
2.2	Building Footprints
Because the TerraPoint building polygon dataset was inadequate to establish accurate building
footprints, other options had to be explored. A building footprint dataset was obtained from the
City of Houston (COH), but the dataset dated to 1983, with small updates in the mid 1990s.
Comparison of the COH building dataset with a recent aerial photo indicated that significant
parts of the dataset were outdated. Buildings were not shown in areas of recent development
and buildings were shown in areas that had been redeveloped. Approximately 2 person-months
were invested in checking and correcting the COH building footprint dataset for a 1653-km2
area. The 1653-km2 area was chosen to include the downtown core area of Houston, the ship
channel industrial district, and extensive coverage of the level 2 U.S. Geological Survey (USGS)
4

-------
Table 1. Digital Datasets Obtained or Created for UCP Computations
Dataset
Format
Source
Description
Digital Elevation Model
(DEM)
Arclnfo Export
TerraPoint
Full feature; LIDAR gridded data product with
1- and 5-m horizontal spatial resolution
Digital Terrain Model
(DTM)
Arclnfo Export
TerraPoint
Bare earth; LIDAR gridded data product with
1- and 5-m horizontal spatial resolution
Ground
Arclnfo Export
TerraPoint
Ground feature; LIDAR gridded data product
with 1- and 5-m horizontal spatial resolution
Nonground
Arclnfo Export
TerraPoint
Nonground feature; LIDAR gridded data
product with 1- and 5-m horizontal spatial
resolution
Buildings
Arclnfo Export
TerraPoint
Building feature; LIDAR gridded data product
with 1- and 5-m horizontal spatial resolution
Vegetation
Arclnfo Export
TerraPoint
Vegetation feature; LIDAR gridded data
product with 1- and 5-m horizontal spatial
resolution
Building Polygons
ESRI shapefile
TerraPoint
Polygons of building footprint derived by
TerraPoint for selected areas where buildings
are distinct from vegetation
Building Polygons
ESRI map
library
City of Houston
Building footprints in the City of Houston dating
to 1983
Building Polygons
ESRI shapefile
Created
Building footprint coverage for 1653-km^ area
of the City of Houston. Dataset based on City
of Houston building polygon dataset; dataset
corrected and improved using 0.5-m digital
orthophotos dating to 2000.
Street Centerlines
ESRI shapefile
U.S. Census
TIGER files
Downloaded from www.esri.com
Aerial Photos
MrSID
Houston-Galveston
Area Council
(HGAC)
0.5-m True Color Geo-referenced Digital
Orthophotos
Land Use/Land Cover
ESRI shapefile
USGS
Standard USGS land use digital dataset with
Level 2 land use classification
Land Use/Land Cover
ESRI shapefile
Created
The USGS dataset was modified using aerial
photos as base map for Phase I study area
(1653 km2). Major improvement was the
addition of newly developed areas—affected
more than 50% of the Phase I and II study
areas.
Modeling Grid Cells
ESRI shapefile
Created
1-km2 horizontal resolution grid cell coverage
of modeling domain; coordinates and
projection of grid cell coverage specified by
EPA
Parcel
ESRI map
library
City of Houston
Multiple CD set of data from City of Houston
containing a great amount of parcel level
information
Elevation Contours
ESRI shapefile
Tropical Storm
Allison Recovery
Project (TSARP)
and Harris County
Flood Control
District
2-ft elevation contours
land use types. The original COH building dataset within the 1653-km2 Phase I study area
included 523,920 building footprints. The modified building dataset contained 664,861 building
footprints.
5

-------
2.3 Land Use/Land Cover
After comparing several LULC datasets for possible use in this project, the LULC dataset made
available by USGS eventually was chosen. The USGS dataset contained a more detailed
classification system for urban areas than the National Land Cover Dataset (NLCD), although it
was more outdated (1970s versus 1990s). The COH land use parcel data is also available, and
it is the most accurate and current compared with the USGS and the NLCD. But, the coverage
is limited to the COH and would not be available for extrapolation of UCPs to the entire
modeling domain.
The USGS is the Federal agency primarily responsible for development, maintenance, and
distribution of a nationwide LULC dataset. The land use and cover is classified according to the
level 2 classification scheme described by Anderson et al. (1976) (see Table 2). The level 2
classification scheme has relatively coarse detail but permits finer classification to level 3 or 4.
The basic sources of land use compilation data are National Aeronautics and Space
Administration (NASA) high-altitude aerial photographs and National High-Altitude Photography
(NHAP) program photographs, usually at scales smaller than 1:60,000 (USGS, 1990).
USGS LULC datasets can be downloaded in two formats. The first format was developed as
part of the Geographic Information Retrieval Analysis System (GIRAS) and is polygon based.
Each polygon represents a contiguous area of homogeneous land use/cover. The minimum size
of polygons depicting all "urban or built-up land" (categories 11-17); "water" (51-54); "confined
feeding operations" (23); "other agricultural land" (24); "strip mines, quarries, and gravel pits"
(75); and urban "transitional areas" (76) is 4 ha. All other categories have a minimum polygon
size of 16 ha (USGS, 1990). In the urban or built-up land and water categories, the minimum
width of a feature to be shown is 200 m. Although the minimum-width consideration precludes
the delineation of very narrow and very long 4 ha polygons, triangles or other polygons are
acceptable if the base of the triangle or minimum width of the polygon is 200 m in length, and if
the area of the polygon is 4 ha. Exceptions to this specification are limited access highways (14)
and all double-line rivers (51) on the 1:250,000-scale base, which have a minimum width of 92
m. For categories other than urban or built-up land and water, the 16-ha minimum size for
delineation requires a minimum-width polygon of 400 m. The second data format is termed
Composite Theme Grid (CTG). The CTG data are grid cell based. The grid cells are actually
based on a regular point sampling, where the center point of each cell is 200 m apart from other
center points in adjacent cells (USGS, 1990). The GIRAS-polygon-based dataset was chosen
for use in this project.
Table 3 summarizes the amounts of each level 2 land use type within the MM5/CMAQ/DA-SM2-
U modeling domain. The majority of the domain is comprised of water (Gulf of Mexico),
Cropland and Pasture, Forestland, and Wetlands. Less than 5% of the modeling domain is
classified as urban, but that area is, for the most part, concentrated within and adjacent to the
COH. UCPs will be computed for each land use type primarily based on an analysis of the
1653-km2 Phase I study area. Following UCP derivation, the land-use-specific UCPs will be
extrapolated to the entire innermost (i.e., finest spatial resolution) grid of the modeling domain
using an area-weighted averaging approach based on the underlying USGS land use dataset.
More details of the UCP extrapolation are provided in Section 3 of this report.
6

-------
Table 2. Anderson Classification System Used in USGS LULC Datasets
Level 1
Level 2
1. Urban or Built-Up Land
11.	Residential
12.	Commercial and Services
13.	Industrial
14.	Transportation, Communication, and Utility
15.	Industrial and Commercial Complexes
16.	Mixed Urban or Built-Up Land
17.	Other Urban or Built-Up Land
2. Agricultural Land
21.	Cropland and Pasture
22.	Orchards, Groves, Vineyards, Nurseries, and Ornamental
Horticultural Areas
23.	Confined Feeding Operations
24.	Other Agricultural Land
3. Rangeland
31.	Herbaceous Rangeland
32.	Shrub and Brush Rangeland
33.	Mixed Rangeland
4. Forest Land
41.	Deciduous Forest Land
42.	Evergreen Forest Land
43.	Mixed Forest Land
5. Water
51.	Streams and Canals
52.	Lakes
53.	Reservoirs
54.	Bays and Estuaries
6. Wetland
61.	Forested Wetland
62.	Nonforested Wetland
7. Barren Land
71.	Dry Salt Flats
72.	Beaches
73.	Sandy Areas Other Than Beaches
74.	Bare Exposed Rock
75.	Strip Mines, Quarries, and Gravel Pits
76.	Transitional Areas
77.	Mixed Barren Land
8. Tundra
81.	Shrub and Brush Tundra
82.	Herbaceous Tundra
83.	Bare Ground
84.	Wet Tundra
85.	Mixed Tundra
9. Perennial Snow or Ice
91.	Perennial Snowfields
92.	Glaciers
7

-------
Table 3. Anderson Level 2 Land Use Types in Modeling Domain
USGS Level 2 Land Use Name
Area (km2)
Percent of Area
Residential
1867
2.3
Commercial and Services
309
0.4
Industrial
543
0.7
Transportation, Communications, and Utility
236
0.3
Mixed Industrial and Commercial
2
0.0
Mixed Urban or Built-Up Land
110
0.1
Other Urban or Built-Up Land
157
0.2
Cropland and Pasture
19,113
23.2
Orchards, Groves, Vineyards, and Nurseries
13
0.0
Confined Feeding Operations
3
0.0
Other Agricultural Land
28
0.0
Herbaceous Rangeland
466
0.6
Shrub and Brush Rangeland
132
0.2
Mixed Rangeland
446
0.5
Deciduous Forest Land
2951
3.6
Evergreen Forest Land
14,737
17.9
Mixed Forest Land
5025
6.1
Streams and Canals
322
0.4
Lakes
1019
1.2
Reservoirs
671
0.8
Bays and Estuaries
1624
2.0
Gulf of Mexico
27,091
32.9
Forested Wetlands
791
1.0
Nonforested Wetlands
4142
5.0
Beaches
16
0.0
Sandy Areas Other Than Beaches
36
0.0
Bare Exposed Rock
1
0.0
Strip Mines, Quarries, and Gravel Pits
32
0.0
Transitional Areas
494
0.6
8

-------
3. Urban Canopy Parameters
The Models-3/CMAQ/MM5/DA-SM2-U modeling framework requires many UCPs to be defined
to adequately represent the urban effect in the simulations. Table 4 lists the UCPs computed for
this project and the name of the Excel spreadsheet that contains the results from the
computations (all spreadsheets are available on the accompanying compact disc [CD] set). In
addition, a gridded coverage of land use is provided on the CD set. Following the table, a brief
review of the parameter calculation procedures is given for each parameter.
Table 4. Urban Canopy Parameters Computed for This Project
Urban Canopy Parameter
Results
Mean Building Height
GridCellUCPs.xls
Standard Deviation of Building Height
GridCellUCPs.xls
Mean Building Height Weighted by Footprint Plan Area
GridCellUCPs.xls
Wall-to-Plan Area Ratio
GridCellUCPs.xls
Building Height-to-Width Ratio (ks)
GridCellUCPs.xls
Building Height Histograms
BuildingHeightHistograms.xls
Mean Vegetation Height Weighted by Plan Area
GridCellUCPs.xls
Mean Canopy Height Weighted by Plan Area
GridCellUCPs.xls
Mean Orientation of Street
GridCellUCPs.xls
Building Plan Area Density Function [ADb(z)]
BuildingPlanAreaDensity.xls
Vegetation Plan Area Density Function [ADV(z)]
VegetationPlanAreaDensity.xls
Canopy Plan Area Density Function [ADC(z)]
CanopyPlanAreaDensity.xls
Building Rooftop Area Density Function [Atb(z)]
BuildingRooftopAreaDensity.xls
Vegetation Top Area Density Function [Atv(z)]
VegetationTopAreaDensity.xls
Canopy Top Area Density Function [Atc(z)]
CanopyTopAreaDensity.xls
Building Frontal Area Density Function [Afb(z)]
BuildingFrontalAreaDensity.xls
Vegetation Frontal Area Density Function [Afv(z)]
VegetationFrontalAreaDensity.xls
Canopy Frontal Area Density Function [Afc(z)]
CanopyFrontalAreaDensity.xls
Roughness Length and Displacement Height (Raupach, 1994)
GridCellUCPs.xls
Roughness Length and Displacement Height (Macdonald, et al.
1998)
GridCellUCPs.xls
Roughness Length and Displacement Height (Bottema, 1997)
GridCellUCPs.xls
Sky View Factor
GridCellUCPs.xls
Plan Area Fraction of Buildings, Roadways/Pavement, Vegetation,
Open Water, and Other Cover
LandCoverFraction.xls
Building Material Fraction
GridCellUCPs.xls
Percent Directly Connected Impervious Area (DCIA)
GridCellUCPs.xls
3.1 Building Height Characteristics
For the Phase I study area, all building height parameters were computed using either the
building footprint shapefile dataset digitized during this project or based on the extrapolation of
the land use sample values (described later).
The mean and standard deviation of building height were calculated using the following
equations:
N
h = -M—
9

-------
S„ =
N i -\2
/=1
/V-1
(2)
where h is the mean building height, Sh is the standard deviation of building height, h, is the
height of building /, and N is the total number of buildings in the area. The mean building height
weighted by building plan area was calculated using the following equation:
ZA",
hAW = 	
IA
/=1
(3)
where hAw is the mean building height weighted by building plan area, and A; is the plan area
at ground level of building /.
The wall-to-plan area ratio is defined as the summed surface area of building walls divided by
the grid cell plan area,
A
o _ w
(4)
where Aw is the combined surface area of the building walls, and At is the plan area of the grid
cell.
The building height-to-width ratio (As) is calculated for two buildings by dividing the average
height by the distance between the two buildings,
where Hi is the height of the upwind building, H2 is the height of the downwind building, and S12
is the horizontal distance between the two buildings (i.e., the canyon width). The calculation of
As can be performed for each pair of adjacent elements in a building array, which can be very
tedious for the complex building shapes and patterns in a city. Instead, As is computed along
linear traverses across the city at different angles using equation 5. The calculation strategy
involved converting the urban building database into a raster DEM (a matrix of numbers
representing building height). Then, traversing along each row or column of grid cells, the
height-to-width ratio was calculated between each pair of buildings. Because this approach
yields As values in nonpreferred directions (e.g., running along a street, not across a street), the
matrices of traverses done at different angles are superimposed and the largest height-to-width
ratio within each grid cell is selected.
Building height histograms for each grid cell were derived at 5-m height bin sizes.
10

-------
3.2 Vegetation Height Characteristics
The mean vegetation height weighted by plan area was calculated using the following equation:
LA",
hAw = ' 1
N
IA
" .	(6)
where haw is the mean vegetation height weighted by plan area, /?, is the vegetation height in
raster cell /, and A is the plan area of raster cell /.
3.3 Canopy Height Characteristics
The mean canopy height weighted by plan area was calculated using the following equation:
N
Iiaw = 	
IA
»' .	(7)
where hAw is the mean canopy height weighted by plan area, /?, is the canopy height in raster
cell /, and A is the plan area of raster cell /.
3.4 Plan Area Density Function
The plan area density function [Ap(z)] is defined as the average roughness element plan area
within a height increment divided by the volume of the height increment,
z + —Az
2
iz lAp(z>'
Ad(z).
z —A z
2
ATAz
(8)
where Ap(z') is the plan area of roughness elements at height z\ At is the plan area of the site,
and Az is the height increment for the calculation. Because At is not a function of height, we
can bring it into the integral in the numerator and obtain,
1.
z + — Az
1 I
At J
2 (z L,'
kiz
i —Az
Ap(z) =	*
Az ^ At
Az	(9)
Knowing Xp(z') = Ap(z')/At and assuming that the roughness element plan area does not
change appreciably within a height increment Az, equation 9 can be approximated by,
11

-------
A„{ z) = ^l
Az .	(10)
The plan area density function is computed for buildings only, vegetation only, and canopy.
3.5 Top Area Density Function
The roughness element top plan area within a height increment Az can be approximated by the
difference between the top plan areas at two heights,
A(*) = A
z -
2
Az^i „ f Az^i

z + ¦
2
(11)
where Ap(z) is the top plan area of roughness elements at the specified height (flat-roofed and
canopy top assumption has been made). The top area density function [At(z}\ can then be
defined as the top plan area of roughness elements per height increment Az divided by the
volume of the height increment,
A{Z.^\.A(Z+^
Mz)" A'iz) " 1 2j 1 2-
AT Az	AT ¦ Az
(12)
where At is the total grid cell area. Analogous to the leaf area index used in the plant canopy
community, the integration of At(z) from a specified elevation above ground (z) to the height of
the canopy (hc) is equal to the area index [L(z)],
L(z) = \ar(z')dz'.	(13)
z
The integration of At(z) from ground elevation to the canopy height (hc) is equal to Ap,
Z_(o) = Ap =jAt{z')dz'
0	(14)
The top area density function is computed for buildings only, vegetation only, and canopy.
3.6 Frontal Area Density Function
The frontal area density function [Af(z)] is defined as,
. ( s)=^Ui£1
(15)
where A(0)prOj(Az) is the area of roughness surfaces projected into the plane normal to the
approaching wind direction for a specified height increment (Az), 0is the wind direction angle,
12

-------
and A7- is the total plan area of the grid cell. For a specified wind direction, the integral of Af(z)
over the canopy height equates to the frontal area index (If).
The frontal area index is defined as the total area of roughness elements projected into the
plane normal to the approaching wind direction (Aproj) divided by the plan area of the study site
where 6is the wind direction.
The frontal area density function is computed for buildings only, vegetation only, and canopy.
For this project, the frontal area density was determined for each grid cell for an approach wind
from four directions: north, northeast, east, and southeast. Roughness elements were assumed
to be rectangular solids in all cases (i.e., their width does not change with height for a single
polygon or grid cell). This, of course is not true for trees and other objects that have a changing
width with height, but the data used to find the UCPs did not represent the variable width of
objects.
3.7	Sky View Factor
The ratio of radiation received (or emitted) from a planar surface to the radiation emitted (or
received) by the entire hemispheric environment is called the sky view factor (\j/Sky)¦ Vsky varies
from 0 to 1, where \|/Sky = 0 means the sky is completely obstructed by obstacles and \|/Sky = 1
means that there are no obstructions. To expedite the processing, the raster data product was
resampled to 3-m horizontal resolution, and the sky view factor was determined for each ground
level 9-m2 grid cell by finding the fraction of visible sky for rays directed in eight azimuths. As an
approximation, the average visible sky fraction is computed and set as the sky view for that grid
cell. Sky view factor is computed using the canopy dataset only (building- or vegetation-specific
sky view factors are not calculated).
3.8	Displacement Height and Roughness Length
The displacement height (Zd) and roughness length (z0) are computed using three sets of
equations reviewed by Grimmond and Oke (1999). The first set of equations was developed by
Raupach (1994),
(^t),
(16)
(17)
and
z.
O
H zd	, u
^-= exp -k— + y/k
J
(18)
where
13

-------
y.
— = min
U
(cs + cRXf) ,
u*_
U
(19)
and y/k is the roughness sublayer influence function, U and u* are the large-scale wind speed
and the friction velocity, respectively; Cs and Cr are drag coefficients for the substrate surface at
height Zh in the absence of roughness elements and of an isolated roughness element mounted
on the surface, respectively; and Cdi is a free parameter. Raupach (1994) suggested y/j< =
0.193, (u*IU)max = 0.3, Cs = 0.003, Cr = 0.3, and Cdi = 7.5.
The second set of equations was derived by Macdonald et al. (1998). These equations
incorporate the drag coefficient and displacement height into the expression for roughness
length (z0),
^k = 1 + a~Ap (JL -1)
(20)
and
1-=^
. ZHJ
exp
f
0.5^
f
1-i

-0.5
-H J
(21)
where oris an empirical coefficient, Co is a drag coefficient, k is the von Karman constant, and j5
is a correction factor for the drag coefficient (the net correction for several variables, including
velocity profile shape, incident turbulence intensity, turbulence length scale, and incident wind
angle, and for rounded corners). Macdonald et al. (1998) recommended for staggered arrays of
cubes that a~ 4.43 and /?- 1.0. These values were used by Grimmond and Oke (1999) and are
also used here. The drag coefficient Co was set to 1.2, and the von Karman constant (k) is
equal to 0.4.
The third set of equations was developed by Bottema (1997),
ZH Zdpl
= = —	 exp
k
(0.5 Z,CJ°"
(22)
where Zdpi is an in-plane sheltering displacement height calculated based on the density and
pattern of the building arrangement (normal or staggered), and Cdh =1.2 max[1 - 0.15(Lx/zh),
0.82] min[0.65 + 0.06{Lylzn), 1.0], Bottema (1997) provides additional equations to compute Zd
and Zdpi that are dependent on the arrangement and density of buildings. For the study area, a
repeatable calculation method that can be automated must be used. Therefore, the building
arrangement is assumed to be staggered arrays of buildings with high densities for all areas.
This is a reasonable assumption for most of the land uses, except for the low-density residential
areas on the outskirts of the Phase I study area. All roughness lengths and displacement
heights are determined for four wind directions: north, northeast, east, and southeast.
14

-------
3.9 Mean Orientation of Streets
The mean orientation of all streets in a grid cell is computed by finding the length-weighted
average,
N
IM
" ,	(23)
where 6lw is the mean orientation weighted by street length, 6} is the angle of street segment /,
and L, is the length of street segment /. The numeric angles in the range of 0 to 179.99 are
defined in Table 5.
Table 5. Numeric Angles Corresponding to Street Orientations
Street Orientation
Angle (degrees)
North-South
0 (179.99 is nearly north-south)
Northeast-Southwest
45
East-West
90
Northwest-Southeast
135
3.10	Building Material
Information on building material of commercial and residential structures is available through the
COH Planning and Development Department Tax Roll Records. These parcel level datasets
include information on commercial structure type, residential exterior type, and residential
exterior masonry description. This tax roll information was used as guidance to develop a matrix
to define the fraction of building material used in each Level 2 land use type. The five building
materials possible were
(1)	concrete,
(2)	wood,
(3)	steel,
(4)	brick, and
(5)	combination.
The assumed building material fraction for each land use type is included in the spreadsheet
UCP Land Use Averages and shown later in this report. Values for all land uses except
Residential and Commercial and Services were estimated from assessment of the digital
orthophotos.
3.11	Percent Directly Connected Impervious Area
The percent directly connected impervious area (DCIA) in each grid cell is estimated using
roadway pavement, building footprint, and uncovered parking data to estimate the total
impervious area. Roadways and parking areas are assumed directly connected to the drainage
system. This is a very reasonable assumption for most urban areas given that roadway
drainage is most likely mandated by local ordinances. Single-family residential rooftops are
assumed disconnected from the drainage system because most downspouts from houses are
probably directed to pervious areas (e.g., yards). Multifamily rooftops are assumed connected.
Commercial and industrial building rooftops are assumed directly connected to the drainage
system. A summary of the assumed DCIA values for each land use are included in the
spreadsheet UCP Land Use Averages and shown later in this report.
15

-------
3.12 Fraction Land Cover
The fraction ground cover type was determined by integrating several datasets in GIS. The
vegetation, the building footprint, roadway pavement, and open water datasets were used to
find the fraction of each grid cell covered by the following land covers.
•	Building
•	Roadway/Pavement (includes sidewalks and driveways)
•	Vegetation
•	Open water
•	Other (includes gravel, bare soil, rock, beach, railroad lines, and other built structures that are
not buildings or roadways)
The values used in the UCP extrapolations are contained in the spreadsheet UCP Land Use
Averages and shown later in this report.
16

-------
4. Urban Canopy Parameter Extrapolation Based on
Underlying Land Use
The building dataset covers only 1653 km2 of the 82,368 km2 modeling domain (primarily the
COH and adjacent areas). The LIDAR data covers only 5242 km2 of the 82,368 km2-modeling
domain. Clearly, the elevation, building, and surface cover data extent is not enough to provide
coverage for UCP computation in each grid cell of the entire modeling domain. Therefore,
average UCPs were calculated for each USGS level 2 land use type, and area-weighted
averages then were determined for each grid cell that did not have data coverage (i.e., UCPs
were extrapolated from the area of data to the area lacking data using the land use). A random
spot check of the USGS land use dataset within the 1653-km2 Phase I study area against the
0.5-m color ortho images from the Houston-Galveston Area Council indicated that the dataset
was inaccurate and not acceptable for UCP derivation and validation because of recent
unrepresented urbanization. Therefore, the land use dataset was modified using the aerial
photos as a base map. Figure 4 shows the final land use dataset used for UCP derivation and
validation. Table 6 lists the amount of each level 2 land use type found within the Phase I study
area boundary according to the original USGS dataset and according to the modified dataset.
The amount of urban land use increases substantially, although the amount of Cropland and
Pasture decreases about the same amount. This is expected because the area has urbanized
significantly during the 20+ years since the USGS dataset originally was derived. There are also
several areas of potential classification differences (e.g., Reservoirs, Strip Mines, and
Transitional Areas) between the original dataset and the modified dataset. All land use types
with potential large percent changes caused by classification differences do not have a
significant effect on the project. For example, many smaller reservoirs would have been too
small to be noted in the original dataset but were included in the revised dataset.
Ship
Channel
... , '•
y jA J,
V
*
Downtown
Houston
Land Use Ciass
*£&3EWTlAi_
HUMPOmftTKM. COUWyWCArO* UTIUTT
CMPU3
lam UHMM q« iwfgp
OtHtR UHBftt* Qfl ttUR-t-Ur
iSOHAKS AMO M&fUflfi
eOWHU FtElfcnti 0*5
| QftCHJWDS VIHE	AMD PTUftB&RlES
D€C«KKXJ9 FOREST LMD
| rVIRMEIN FOWE S1 LAW
MxiDrcmcsi unto
rOW*TI0 WITLAKO
WOWORMnfl' WWTVJUJD
| ttrfbFUMRS
rWAMSimQWAL AAEA5
| LAKES
| fcESEUVCMftt
| 4nt£AU£AN0GAMALS
0	2 5 5 10 Kilometers
1	I B I 1 I I I I
Figure 4. Land use of the Phase I building study area. Land use classified according to the
Anderson level 2 categories.
For most parameters, a 1235-km2 section of the 1653-km2 Phase I study area was used to
derive the average UCP for each land use type, whereas the remaining 418 km2 was used for
validation (see Figure 5). Table 7 contains the land use areas within the derivation area and the
17

-------
Table 6. Comparison of Land Use Distribution in the 1653-km2 Phase I Study Area for
	Original USGS Land Use Dataset and Revised Dataset	

Original Area
Revised Area
Percent Change
USGS Level 2 Land Use Name
(km2)
(km2)
(%)
Residential
533.7
633.1
+19
Commercial and Services
85.9
115.2
+34
Industrial
79.9
156.6
+96
Transportation, Communications, and Utility
53.3
61.9
+16
Mixed Industrial and Commercial
2.0
1.0
-50
Mixed Urban or Built-Up Land
16.7
19.4
+16
Other Urban or Built-Up Land
67.1
88.5
+32
Cropland and Pasture
532.9
297.5
-44
Orchards, Groves, Vineyards, and Nurseries
0.7
0.0
—
Confined Feeding Operations
0.1
0.0
—
Other Agricultural Land
0.0
1.1
—
Herbaceous Rangeland
0.0
0.0
—
Shrub and Brush Rangeland
0.0
0.0
—
Mixed Rangeland
0.0
0.0
—
Deciduous Forest Land
46.7
68.6
+47
Evergreen Forest Land
22.3
12.7
-43
Mixed Forest Land
172.6
147.5
-15
Streams and Canals
14.3
21.8
+53
Lakes
1.0
0.6
-40
Reservoirs
2.8
7.7
+175
Bays and Estuaries
0.0
0.0
—
Gulf of Mexico
0.0
0.0
—
Forested Wetlands
0.8
0.3
-63
Nonforested Wetlands
3.5
3.0
-14
Beaches
0.0
0.0
—
Sandy Areas Other Than Beaches
0.0
0.0
—
Bare Exposed Rock
0.0
0.0
—
Strip Mines, Quarries, and Gravel Pits
2.0
11.0
+450
Transitional Areas
14.7
5.5
-63
validation area for comparison. For some parameters, calculation of values for large areas was
not possible because of data, time, or computational limitations. For these situations, five
samples of each land use type located within the 1235-km2 section of the Phase I study area
were arbitrarily selected (the same five sample sites were used for all UCPs that had to be
calculated in this fashion). Table 8 contains a description and size of each sample site. When
necessary, the UCPs were calculated for the five samples, and the average value was used in
the extrapolation calculation. Time constraints prevented validation for some of these UCPs.
As a summary of the UCP-land use derivation process, Table 9 lists each UCP and indicates
the areas (Phases I, II, III; see Figure 3) that have the parameter value based on actual
calculation and the areas that have the value based on the area-weighted averaging of land use
specific parameters (i.e., extrapolation). Only eight parameters were based on the samples to
compute the average value per land use, and two parameters (mean orientation of streets and
land use fraction) were computed for the entire modeling domain. The remaining parameter
extrapolations are based on Phase I study area averages. The average UCP values per land
18

-------
Validation
Land Use Class
REStOEKTUU.
¦	COMMERCIU. AW MBVICtS
| INDtiaTOUl.
¦	m*.»SPW(Ano» i;»««uuN!«iu;j« ™h
INIWST AND GOMWEjRC CMPLKS
1 Hid UMW CXI uuii.T ;ir
OTHER URBAN OK HUH.T-UP
CBOPUUiC *tttt KMWRE
CONFWEB FEEDIWi OPS
CKWHjuicia, vtNewuisa uiBNmtscniEi
DECIDUOUS f WEST LAND
| CVCReRECU tMEST LAW
ItXEO FOREST LAND
FORESTED WCTAND
NONFOftESTCD WETLAND
I STWP W(S
TRAHSTTTOHAL «HEU
| LAKI»
| RESERVOIRS
| STKEMQ AltS CANALS
0	2,5 S to Kilometers
1	I I ¦ 11 I 11
Downtown
Houston
Ship
Channel
Figure 5. UCP derivation and validation zones overlaying land use.
use and per sample site are included in the UCP Land Use Averages Excel spreadsheet. Only a
summary and brief description of the values is given here.
4.1 Average Urban Canopy Parameter Derivation
4.1.1 Building Height Characteristics
Average building height characteristics were determined for the 1235-krrr UCP derivation
section of the Phase I study area. The average values were computed and are listed in Table
10. In some cases, the average values were discarded and other values were assumed (see the
UCP Land Use Averages Excel spreadsheet for a full accounting of the assumptions).
Building height histograms were found for each land use based on the buildings contained in the
1235-krrr UCP derivation section of the Phase I study area. Figure 6 shows the building height
histograms for the seven urban land use types. The height histograms are given in terms of
number of buildings per square kilometer, which can be used for area-weighted extrapolation.
To extrapolate, the number of buildings in each height increment is multiplied by the area of
each land use type within each grid cell boundary. The building height histograms for the
nonurban land use types are not shown because those land uses contained very few buildings.
The building plan area density, top area density, and frontal area density were also found for the
1235-krrr UCP derivation section of the Phase I study area. These functions of height are
shown in Figures 7, 8, and 9. Figure 7 indicates that the Commercial and Services and
Industrial land use types have the highest ground level building plan area fraction on average
(0.22 for Commercial and 0.18 for Industrial) and have the most gradual decrease in plan area
fraction as height increases.
The building roof area density function shown in Figure 8 clearly shows the concentration of
rooftops between 5 and 13 m (predominantly one- to three-story structures). The Commercial
and Services, Industrial, and Mixed Urban or Built-Up land use types have the largest roof area
19

-------
Table 7. Comparison of Areas of Each Land Use Type in the
UCP Derivation and Validation Areas
USGS Level 2 Land Use Name
Derivation (km2)
Validation (km2)
Residential
481.4
151.5
Commercial and Services
95.6
19.7
Industrial
133.8
22.8
Transportation, Communications, and Utility
46.4
15.5
Mixed Industrial and Commercial
1.0
—
Mixed Urban or Built-Up Land
17.4
1.9
Other Urban or Built-Up Land
71.1
17.4
Cropland and Pasture
148.7
148.7
Orchards, Groves, Vineyards, and Nurseries
—
—
Confined Feeding Operations
—
—
Other Agricultural Land
—
1.1
Herbaceous Rangeland
—
—
Shrub and Brush Rangeland
—
—
Mixed Rangeland
—
—
Deciduous Forest Land
36.7
31.9
Evergreen Forest Land
8.7
—
Mixed Forest Land
147.5
—
Streams and Canals
21.8
—
Lakes
0.6
—
Reservoirs
4.8
2.9
Bays and Estuaries
—
—
Gulf of Mexico
—
—
Forested Wetlands
0.3
—
Nonforested Wetlands
2.7
0.3
Beaches
—
—
Sandy Areas Other Than Beaches
—
—
Bare Exposed Rock
—
—
Strip Mines, Quarries, and Gravel Pits
9.3
1.7
Transitional Areas
4.9
0.6
density values (within the 5- to 9-m increment). Figure 9 indicates that the Commercial and
Services, Industrial, and Mixed Urban or Built-Up land use types also have the highest building
frontal area densities.
4.1.2 Vegetation Height Characteristics
The mean vegetation height was determined for the five land use samples, and the average is
shown in Table 11. As expected, the forest land use types have the highest mean vegetation
height. Of the urban land use types, the Residential land use has the highest.
The vegetation plan area density, top area density, and frontal area density were also found for
the five land use samples. The average functions for each land use type are shown in Figures
10, 11, and 12. Figure 10 indicates that the two forest land use types have the greatest plan
area density at all heights. One other interesting note is the higher plan area density for
Residential land use compared to Cropland and Pasture.
20

-------
Table 8. Sample Site Descriptions (size of area shown in parentheses)
Land Use
Sample 1
Sample 2 |
Sample 3 |
Sample 4
| Sample 5
Residential
Older high-density
single-family
(13 ha)
Older high-density
single-family (5 ha)
Newer multifamily
apartments (10 ha)
Newer high-
density single-
family (15 ha)
Older low-density
single-family
(16 ha)
Commercial and
Services
Business and
distribution (28 ha)
Office buildings
(24 ha)
Shopping mall
(58 ha)
Strip commercial
(8 ha)
Downtown
(28 ha)
Industrial
Mixed
warehousing and
manufacturing
(18 ha)
Manufacturing
(25 ha)
Refinery—ship
channel (62 ha)
Heavy industrial—
ship channel
(51 ha)
Warehousing and
distribution—ship
channel (53 ha)
Mixed Industrial
Not enough samples
—averaged Commercial and Services and Industrial

and Commercial





Mixed Urban or
Mixed industrial,
Mixed industrial,
Industrial,
Industrial,
Light industrial,
Built-Up
commercial, and
commercial, and
commercial, and
warehousing, and
commercial, and

residential (16 ha)
residential
(16 ha)
multifamily
residential (21 ha)
mobile and single-
family homes
(17 ha)
mobile and single-
family homes
(14 ha)
Other Urban or
Educational (high
Golf course
New buildings with
Parking area with
Industrial-like
Built-Up
school) (21 ha.)
(43 ha)
large tracts of
landscaping
(19 ha)
adjacent
landscaping
(17 ha)
activity with open
pit and open space
(18 ha)
Cropland and
Pasture
Cropland (56 ha)
Cropland
(85 ha)
Mixed cropland
and pasture
(33 ha)
Mixed cropland
and pasture with
trees (47 ha)
Mixed cropland
and pasture with
trees (51 ha)
Orchards,
Groves,
No samples—used Cropland and Pasture



Vineyards, and
Nurseries





Confined Feeding
Operations
No samples—used Cropland and Pasture
Other Agricultural
Land
Not enough land area for sufficient samples—used Cropland and Pasture
Herbaceous
Rangeland
No samples—used Cropland and Pasture
Shrub and Brush
No samples—used Cropland and Pasture



Rangeland





Mixed Rangeland
No samples—used Cropland and Pasture
Deciduous Forest
Dense forest cover
Dense forest cover
Dense forest cover
Dense forest cover
Dense forest cover
Land
(75 ha)
(56 ha)
(96 ha)
(77 ha)
(34 ha)
Evergreen Forest
Dense forest cover
Dense forest cover
Dense forest cover
Dense forest cover
Dense forest cover
Land
(75 ha)
(62 ha)
(21 ha)
(15 ha)
(47 ha)
Mixed Forest
Dense forest cover
Dense forest cover
Dense forest cover
Dense forest cover
Dense forest cover
Land
(74 ha)
(32 ha)
(58 ha)
(231 ha)
(88 ha)
Streams and
Canals
No samples - parameters easily defined
Lakes
No samples—parameters easily defined
Reservoirs
No samples—parameters easily defined
Bays and
Estuaries
No samples—parameters easily defined
Gulf
No samples—parameters easily defined
Forested Wetland
Not enough land area for sufficient samples—used Mixed Forest Land
Nonforested
Wetland
Not enough land area for sufficient samples—used Cropland and Pasture
Beaches
No samples—parameters easily defined
Sandy Area
Non-Beach
No samples—parameters easily defined
Bare Exposed
Rock
No samples—parameters easily defined
Strip Mines,
Quarries, and
Gravel Pits
Open pit (75 ha)
Gravel pit
(11 ha)
Open pit (185 ha)
Gravel mine
(41 ha)
Open pit (51 ha)
Transitional
Areas
Not enough land area for sufficient samples—used Strip Mines
21

-------
Table 9. Calculation and Extrapolation Extent for Each Urban Canopy Parameter
Urban Canopy Parameter
Calculation3
Extrapolation3
Mean Building Height
Phase I
Phase II and III
Standard Deviation of Building Height
Phase I
Phase II and III
Mean Building Height Weighted by Footprint Plan Area
Phase I
Phase II and III
Wall-to-Plan Area Ratio
Phase I
Phase II and III
Building Height-to-Width Ratio (Xs)
Phase I
Phase II and III
Building Height Histograms
Phase I
Phase II and III
Mean Vegetation Height Weighted by Plan Area
Samples
Phase I, II, and III
Mean Canopy Height Weighted by Plan Area
Phase I and
II
Phase III
Mean Orientation of Streets
Phase I, II,
and III
None
Building Plan Area Density Function [ADb(z)]
Phase I
Phase II and III
Vegetation Plan Area Density Function [ADV(z)]
Samples
Phase I, II, and III
Canopy Plan Area Density Function [ADC(z)]
Phase I
Phase II and III
Building Rooftop Area Density Function [Atb(z)]
Phase I
Phase II and III
Vegetation Top Area Density Function [Atv(z)]
Samples
Phase I, II, and III
Canopy Top Area Density Function [Atc(z)]
Phase I
Phase II and III
Building Frontal Area Density Function [Afb(z)]
Phase I
Phase II and III
Vegetation Frontal Area Density Function [Afv(z)]
Samples
Phase I, II, and III
Canopy Frontal Area Density Function [Afc(z)]
Phase I
Phase II and III
Roughness Length and Displacement Height (Raupach, 1994)
Phase I
Phase II and III
Roughness Length and Displacement Height (Macdonald et al.,
1998)
Phase I
Phase II and III
Roughness Length and Displacement Height (Bottema, 1997)
Phase I
Phase II and III
Sky View Factor
Samples
Phase II and III
Plan Area Fraction of Buildings, Roadways/Pavement, Vegetation,
Open Water, and Other Cover
Samples
Phase I, II, and III
Land Use Fraction
Phase I, II,
and III
None
Building Material Fraction
Samples
Phase I, II, and III
Directly Connected Impervious Area (DCIA)
Samples
Phase I, II, and III
Of the 1653-km Phase I study area, 1235 km were used to derive the average UCP for each land use type,
whereas 418 km2 were used to validate the averaging and extrapolation. The Phase II study area is the 3589-km2
area of Harris County that has LIDAR data for canopy but no buildings and vegetation. Phase III is the remainder of
the modeling domain outside the 5242-km2 Phases I and II areas of Harris County (see Figure 3).
The vegetation top area density function in Figure 11 clearly shows the concentration of
vegetation tops in the 5- to 15-m range. The forest land use types have the largest top area
density. Figure 11 indicates that the forest land use types also have the highest vegetation
frontal area densities.
4.1.3 Canopy Height Characteristics
The mean canopy height per land use was determined for the 1235 km2 section of the Phase I
study area. Values are listed in Table 12. Note the high canopy height of the forest land,
whereas the Residential and Commercial and Services land use types have the highest mean
canopy height of the urban land use types.
The canopy plan area density, top area density, and frontal area density also were found for the
five land use samples. These functions of height are shown in Figures 13, 14, and 15. Figure 13
22

-------
Table 10. Average Building Height Characteristics per Land Use in the
1235-km2 UCP Derivation Area
USGS Level 2
Land Use Name
Number of
Buildings
Mean
Building
Height
(m)
Standard
Deviation
of
Building
Height (m)
Plan-Area-
Weighted
Mean
Building
Height (m)
Wall-
to-Plan
Area
Ratio
Height-to-
Width
Ratio
Residential
433,811
5.70
3.04
5.49
0.251
0.087
Commercial and
Services
28,372
6.05
7.07
8.97
0.214
0.089
Industrial
25,789
6.09
3.73
8.09
0.136
0.054
Transportation,
Communications, and
Utility
823
4.81
3.54
10.04
0.001
0.013
Mixed Industrial and
Commercial
331
4.97
3.42
6.21
0.151
0.060
Mixed Urban or Built-Up
Land
11,918
5.72
3.20
6.04
0.246
0.091
Other Urban or Built-Up
Land
5246
4.95
2.73
6.00
0.037
0.019
Cropland and Pasture
2187
5.02
3.20
5.50
0.004
0.005
Orchards, Groves,
Vineyards, and
Nurseries
0
5.02
3.20
5.50
0.004
0.005
Confined Feeding
Operations
0
5.02
3.20
5.50
0.004
0.005
Other Agricultural Land
3
5.02
3.20
5.50
0.004
0.005
Herbaceous Rangeland
0
5.02
3.20
5.50
0.004
0.005
Shrub/Brush Rangeland
0
5.02
3.20
5.50
0.004
0.005
Mixed Rangeland
0
5.02
3.20
5.50
0.004
0.005
Deciduous Forest Land
677
7.32
4.42
7.64
0.007
0.006
Evergreen Forest Land
61
5.51
3.45
5.27
0.002
0.003
Mixed Forest Land
1525
6.74
4.13
5.89
0.003
0.004
Streams and Canals
0
—
—
—
—
—
Lakes
0
—
—
—
—
—
Reservoirs
0
—
—
—
—
—
Bays and Estuaries
0
—
—
—
—
—
Gulf of Mexico
0
—
—
—
—
—
Forested Wetlands
0
6 74
4 13
5 39
0 003
0 004
Nonforested Wetlands
0
5.02
3.20
5.50
0.004
0.005
Beaches
0
—
—
—
—
—
Sandy Areas
0
—
—
—
—
—
Bare Exposed Rock
0
—
—
—
—
—
Strip Mines, Quarries,
and Gravel Pits
100
4 10
1 33
4 72
0 003
0 004
Transitional Areas
16
—
—
—
—
—
indicates that the two forest land use types have the greatest plan area density near the ground
surface (<25 m), whereas, above 25 m, the Commercial and Services land use has the highest
plan area density. One other interesting observation is the higher plan area density for
Residential land use near the ground level.
23

-------
400
350
300
E
d" 250

at
a.
O) 200
c
|5
'5
OQ
_ 150
O
100
50
0
Figure 6. Building height histograms. Number of buildings on average in 5-m height increments
for each urban land use type.
The canopy top area density functions in Figure 14 clearly show the concentration of canopy top
in the 5- to 10-m range for most land uses, except forested types. The forest land use types
have the largest top area density. Figure 15 indicates that the forest land use types also have
the highest vegetation frontal area densities near the ground, whereas the Commercial and
Services has the highest above 28 m.
4.1.4	Sky View Factor
The mean sky view factor was determined for the five sample sites for each land use type. The
average of the five is shown in Table 13.
4.1.5	Displacement Height and Roughness Length
The displacement height and roughness length were determined according to three sets of
morphometric equations as described above. The average displacement heights and roughness
lengths based on the five land use samples are listed in Tables 14 and 15, respectively.
Interestingly, there is significant variability in roughness length and displacement height for the
different equation methods used. This discrepancy is being explored further, and the findings
will be reported later.
4.1.6	Building Material
Using guidance developed from the COH Planning and Development Department Tax Roll
Records, a matrix of assumed fractions of building materials per land use type was developed
(see Table 16). These values are by far the most uncertain of the UCPs presented in this report.
< 5 m
5 -10 m
~	Residential
¦	Commercial and Services
¦	Industrial
¦	Transportation, Communication, Utility
¦	Mixed Industrial and Commercial
¦	Mixed Urban or Built-up
~	Other Urban or Built-up
10 -15 m	15 - 20 m
Building Height (m)
20 - 25 m
25 - 30 m
24

-------
60
50
40
% 30
"35
20
10
0
0.00	0.05	0.10	0.15	0.20	0.25
Apb (Z)
Figure 7. Building plan area density computed for the 1235-km2 UCP derivation section in Houston
(shown are the values for the USGS level 2 urban land use types).
If buildings are riot a significant fraction of the land use surface area (i.e., low plan area
fractions), then the building material fractions are listed as zero for all materials.
4.1.7	Percent Directly Connected Impervious Area
The percent DCIA was determined for the five sample sites for each land use type. The mean
value for each land use type is listed in Table 17. Because of insufficient samples the following
assumptions were made.
•	The DCIA for Mixed Industrial and Commercial is assumed to be the average of Industrial and
Commercial and Services.
•	Cropland and Pasture DCIA was used for the following land use types: Orchards, Groves,
Vineyards, and Nurseries; Confined Feeding Operations; Other Agricultural Land;
Herbaceous Rangeland; Shrub and Brush Rangeland; Mixed Rangeland; Nonforested
Wetland. Mixed Forest Land DCIA was used for Forested Wetland.
•	All water surfaces were assumed to have a DCIA of zero (this might be modified if
downstream water flow is of interest).
•	Beaches, Sandy Non-Beach, and Bare Exposed Rock all were assumed to have a zero DCIA;
Transitional Areas were assumed to have the same DCIA of Strip Mines.
4.1.8	Fraction Land Cover
The average fraction land cover for each land use is determined from the five sample sites.
Table 18 lists the average fractions computed.
-•-Residential
— Commercial and Services
Industrial
-^Transportation, Communication, Utility
Mixed Industrial and Commercial
Mixed Urban or Built-up
Other Urban or Quilt-up
25

-------
40
35
30
|> 20
"33
X
15
10
5
0
0.00	0.02	0,04	0.06	0.08	0.10	0.12
Atb (z) (1/m)
Figure 8. Building top area density computed for the 1235-km2 UCP derivation section in Houston
(shown are the values for the USGS level 2 urban land use types).
4.2 UCP Assessment and Validation
An assessment of the accuracy of the UCP extrapolation procedure was performed for most of
the parameters computed. The assessment was conducted by comparing the actual calculated
values for each grid cell in the 418-km2 UCP validation area with the extrapolated values. The
validation area is south of the UCP derivation area (see Figure 5) and contains a fairly
representative distribution of land use types. In addition, because the validation area is
adjacent, much of the styles of the land use development will be similar (e.g., residential
subdivisions likely will be fairly similar in form). This assessment will indicate the relative level of
confidence for each UCP value.
-•-Residential
— Commercial and Services
Industrial
-^Transportation, Communication, Utility
Mixed Industrial and Commercial
Mixed Urban or Built-up
Other Urban or Built-up
The relative similarity of the 418 pairs of calculated and extrapolated UCPs is determined
through the computation of several summary statistics:
±-£(ucp,-ucp;)
Bias = n '=1	,	(24)
(ucp,
Root Mean Square Error (RMSE) = I 1=1	, and	(25)
26

-------
u>
'55
40
35
30
25
20
15
10
0
0.000 0.002 0.004 0.006 0.008 0.010 0.012 0.014
Aft>(Z)
Figure 9. Building frontal area density computed for the 1235-km2 UCP derivation section in
Houston for a wind from the north direction (shown are the values for the USGS level 2 urban land
use types).
z
100 • —
UCP, - UCP;
Cumulative Relative Error (CRE) =
T.UCP,
;=1
(26)
where UCP, is the extrapolated UCP for the ith grid cell, UCPi¦ is the calculated UCP for the ith
grid cell, and n is the number of grid cells (418). In addition to the calculation of the statistics,
scatter plots also were produced to visualize the relative match between the pairs.
4.2.1 Building Height Characteristics
The bias, root mean square error (RMSE), and cumulative relative error (CRE) for the building
height characteristics extrapolation are shown in Table 19. The bias column includes a percent
change indicating the relative increase or decrease for the extrapolated values compared with
the calculated values. For example, the results indicate that the extrapolated mean building
height per grid cell is 23% higher on average than the calculated mean building height. The
comparison statistics clearly show that, on average, the extrapolation will result in higher
building height UCPs compared with calculation from actual data. One possible explanation is
the fact that the UCP derivation area included downtown Houston, which would have elevated
the average building height characteristics This was anticipated at the outset of the project but,
-¦-Residential
—Commercial and Services
Industrial
-•-Transportation, Communication, Utility
Mixed Urban or Built-up
Other Urban or Built-up
27

-------
Table 11. Mean Vegetation Height per Land Use Type
USGS Level 2 Land Use Name
Mean Vegetation
Height (m)
Notes
Residential
4.03

Commercial and Services
2.90

Industrial
1.16

Transportation, Communications, and Utility
0.27

Mixed Industrial and Commercial
2.03

Mixed Urban or Built-Up Land
2.77

Other Urban or Built-Up Land
2.43

Cropland and Pasture
1.13

Orchards, Groves, Vineyards, and Nurseries
1.13
Used Cropland and Pasture
Confined Feeding Operations
1.13
Used Cropland and Pasture
Other Agricultural Land
1.13
Used Cropland and Pasture
Herbaceous Rangeland
1.13
Used Cropland and Pasture
Shrub and Brush Rangeland
1.13
Used Cropland and Pasture
Mixed Rangeland
1.13
Used Cropland and Pasture
Deciduous Forest Land
11.83

Evergreen Forest Land
9.09

Mixed Forest Land
9.76

Streams and Canals
—
No vegetation
Lakes
—
No vegetation
Reservoirs
—
No vegetation
Bays and Estuaries
—
No vegetation
Gulf of Mexico
—
No vegetation
Forested Wetlands
9 76
Used Mixed Forest Land
Nonforested Wetlands
1.13
Used Cropland and Pasture
Beaches
—
No vegetation
Sandy Areas Other Than Beaches
—
No vegetation
Bare Exposed Rock
—
No vegetation
Strip Mines, Quarries, and Gravel Pits
3 16

Transitional Areas
—
Assumed no vegetation
given the small area of the downtown core compared with the entire UCP derivation area, the
effect was determined to be small.
For the building height characteristics, it was also feasible to compare parameters as a function
of land use type. Table 20 shows the mean building height and standard deviation of building
height as calculated in the UCP derivation and validation areas. The Residential and Industrial
land uses have the greatest difference in mean building height between the derivation and
validation areas. This was unexpected because it was anticipated that the Commercial and
Services land use, which included downtown Houston, would have the greatest difference
because of the presence of skyscrapers in the derivation area and not in the validation area.
The large difference in the Industrial area may be caused by the ship channel industries making
up most of the Industrial land use in the derivation area, whereas standard industrial parks
made up the Industrial land use in the validation area.
28

-------
40
35
30
25
% 20
"33
15
-•-Residential
— Commercial and Services
Industrial
-^Transportation, Communication, Utility
Cropland and Pasture
Deciduous Forest
-•-Mixed Forest
Apv (z)
Figure 10. Vegetation plan area density computed for the 1235-km2 UCP derivation section in
Houston (shown are the values for selected USGS level 2 land use types).
Scatter plots were produced to visualize the relative comparison between calculated and
extrapolated values in the validation area. Figures 16 and 17 show the scatter plots for the
mean and standard deviation of building height, respectively. The figures visualize the increased
frequency of overprediction by the extrapolation process, as was shown in the summary
statistics above. Another observation that must be explained is the lower threshold for the
extrapolated values visible in the plots. This may be occurring because of several reasons. First,
the average values for building height characteristics used in the extrapolation assume that
buildings will be present in all land uses, even if it is a very small number of buildings. But, some
grid cells will not contain buildings. The grid cells with zero calculated height did not contain
buildings, and these significantly affect the appearance of the scatter plot (i.e., no points found
near 0,0). Another possible reason for the appearance of a lower threshold on extrapolated
values is the potential for errors in building heights used to determine the calculated values.
Errors in height will be more pronounced in grid cells with a small number of buildings.
The scatter plot for the plan-area-weighted mean building height is shown in Figure 18, which is
similar to the mean building height and standard deviation scatter plots. Figures 19 and 20
display the scatter plots of the wall-to-plan area ratio and the height-to-width ratio, respectively.
The match between extrapolated and calculated values is much better for these two ratios,
except for the extremely large calculated values. But the lower threshold is not noted because
the ratio values take into account the density of the buildings (i.e., how many per area and how
far apart they are). Therefore, the average values used for extrapolation will include rather small
values that accurately can represent the small values calculated in the grid cell. This was not
possible for the mean height, standard deviation, and plan-area-weighted height because the
29

-------
40
35
30
|> 20
"33
X
15
10
5
0
0.00	0.05	0.10	0.15	0.20	0.25
A,v (2)
Figure 11. Vegetation top area density computed for the 1235-km2 UCP derivation section in
Houston (shown are the values for selected USGS level 2 land use types).
average values used in the extrapolation did not account for the number of buildings or density.
Clearly, the extrapolation will not be able to capture the extremes (small or large calculated
values) because average values are used, but, for most parameters, the match will be fair
except for the extremely large values. Thus, the average values may be sufficient to provide an
acceptable estimation for most of the UCPs.
The accuracy of the extrapolation of the building height histograms, building plan area density,
top area density, and frontal area density was assessed by computing the bias, RMSE, and
CRE for 25 randomly selected grid cells of the 418 validation cells. The mean, minimum, and
maximum bias, RMSE, and CRE are shown in Table 21. The CRE suggests that the building
frontal area density and building plan area density are predicted across all elevations with the
most accuracy. The building height histograms and building top area density also are predicted
fairly well. In most cases, the mean values are affected significantly by the maximum value, and
the maximum value is an outlier that is especially poorly predicted.
4.2.2 Canopy Height Characteristics
The Bias, RMSE, and CRE for the canopy height characteristics extrapolation are shown in
Table 22. The bias column includes a percent change indicating the relative increase or
decrease for the extrapolated values compared with the calculated values. For example, the
results indicate that the extrapolated mean canopy height per grid cell is 36% higher, on
average, than the calculated mean canopy height. The comparison statistics clearly show that,
on average, the extrapolation will result in elevated canopy height UCPs compared with
-•-Residential
— Commercial and Services
Industrial
-^Transportation, Communication, Utility
Cropland and Pasture
Deciduous Forest
Mixed Forest
30

-------
40
35
30
| 20
"33
X
15
10
5
0
0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.045
Afv (Z)
Figure 12. Vegetation frontal area density computed for the 1235-km2 UCP derivation section in
Houston for a wind from the north direction (shown are the values for selected USGS level 2 land
use types).
calculation from actual data. The anticipated overprediction of roughness length and
displacement height is expected to be approximately 30% to 40%.
Scatter plots were produced to visualize the relative comparison between calculated and
extrapolated values in the validation area. Figure 21 shows the scatter plot for the mean canopy
height. The scatter of the data pairs for the mean canopy height is greater than the scatter for
the mean building height Thus, it appears that the mean canopy height extrapolation will
produce, on average, less accurate values than the mean building height.
The scatter plots for the roughness length and displacement extrapolations also were prepared.
Figures 22 and 23 display the scatter plots for roughness length and displacement height
computed using the Macdonald et al. (1998) set of equations, whereas Figures 24 and 25 show
z0 and zd for the Raupach (1994) set of equations. It was not possible to perform the validation
for the Bottema (1997) equations because the extrapolation was based on samples. Figures 22
and 23 indicate that the roughness length extrapolation is less accurate than the displacement
height extrapolation, in fact, the displacement height extrapolation appears to be very accurate.
Figures 24 and 25 indicate that the displacement height calculated by the Raupach (1994)
equations is more accurately extrapolated than the roughness length calculated by the
Macdonald et al. (1998) equations.
-•-Residential
— Commercial and Services
Industrial
-^Transportation, Communication, Utility
Cropland and Pasture
Deciduous Forest
Mixed Forest
31

-------
Table 12. Mean Canopy Height per Land Use Type
USGS Level 2 Land Use Name
Mean Canopy
Height (m)
Notes
Residential
7.39

Commercial and Services
7.11

Industrial
5.59

Transportation, Communications, and Utility
5.96

Mixed Industrial and Commercial
4.81

Mixed Urban or Built-Up Land
6.31

Other Urban or Built-Up Land
6.15

Cropland and Pasture
6.02

Orchards, Groves, Vineyards, and Nurseries
6.02
Used Cropland and Pasture
Confined Feeding Operations
6.02
Used Cropland and Pasture
Other Agricultural Land
6.02
Used Cropland and Pasture
Herbaceous Rangeland
6.02
Used Cropland and Pasture
Shrub and Brush Rangeland
6.02
Used Cropland and Pasture
Mixed Rangeland
6.02
Used Cropland and Pasture
Deciduous Forest Land
11.06

Evergreen Forest Land
11.79

Mixed Forest Land
10.51

Streams and Canals
0.00

Lakes
0.00

Reservoirs
0.00

Bays and Estuaries
0.00

Gulf of Mexico
0.00

Forested Wetlands
10.51
Used Mixed Forest Land
Nonforested Wetlands
6.02
Used Cropland and Pasture
Beaches
0.00

Sandy Areas Other Than Beaches
0.00

Bare Exposed Rock
0.00

Strip Mines, Quarries, and Gravel Pits
2.56

Transitional Areas
0.00

The accuracy of the extrapolation of the canopy plan are density and top area density was
assessed by computing the bias, RMSE, and CRE for 25 randomly selected grid cells of the 418
validation cells (the same 25 that were used for the building height histogram, plan area density,
top area density, and frontal area density validation assessments). The mean, minimum, and
maximum bias, RMSE, and CRE are shown in Table 23. The results suggest that the
extrapolation of the canopy plan and top area densities are fairly accurate. The minimum CRE
values indicate that, for some grid cells, the extrapolation produces a density function very close
to that calculated, and the mean CRE indicates that only 50% cumulative relative error is to be
expected, on average.
Overall, the assessment indicates that, although there may be significant error associated with
the extrapolation for individual grid cells, on average, the errors will be moderate and have a
minimal impact on model results.
32

-------
40
35
30
25
20
"33
X
15
10
0
0.000	0.200	0.400	0.600	0.800	1.000
Ape (Z)
Figure 13. Canopy plan area densities computed for the 1235-km2 UCP derivation section in
Houston (shown are the values for selected USGS level 2 land use types).
-•-Residential
— Commercial and Services
Industrial
-^Transportation, Communication, Utility
Cropland and Pasture
Deciduous Forest
-•-Mixed Forest
33

-------
40
35
30
25
20
"33
X
15
10
5
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
A,c (Z)
Figure 14. Canopy top area densities computed for the 1235-km2 UCP derivation section in
Houston (shown are the values for selected USGS level 2 land use types).
-•-Residential
— Commercial and Services
Industrial
-^Transportation, Communication, Utility
Cropland and Pasture
Deciduous Forest
-•-Mixed Forest
34

-------
60
Residential
Commercial & Services
50
Industrial
40
Transportation, Communication, Utility
Cropland & Pasture
Deciduous Forest Land
Mixed Forest Land
20
10
0 —
0.000
0.010
0.020
0.030
0.040
0.050
0.060
af(z)
40
g>
35
30
25
20
15
10
0
0.000 0.010 0.020 0.030 0.040 0.050 0.060
Afc (z)
Figure 15. Canopy frontal area densities computed for the 1235-km2 UCP derivation section in
Houston for a wind from the north direction (shown are the values for selected USGS level 2 land
use types).
Residential
Commercial and Services
Industrial
Transportation, Communication, Utility
Cropland and Pasture
Deciduous Forest
Mixed Forest
35

-------
Table 13. Mean Sky View Factor per Land Use Type
USGS Level 2 Land Use Name
Mean Sky View Factor
Residential
0.70
Commercial and Services
0.81
Industrial
0.82
Transportation, Communications, and Utility
0.91
Mixed Industrial and Commercial
0.81
Mixed Urban or Built-Up Land
0.78
Other Urban or Built-Up Land
0.84
Cropland and Pasture
0.94
Orchards, Groves, Vineyards, and Nurseries
0.94
Confined Feeding Operations
0.94
Other Agricultural Land
0.94
Herbaceous Rangeland
0.94
Shrub and Brush Rangeland
0.94
Mixed Rangeland
0.94
Deciduous Forest Land
0.61
Evergreen Forest Land
0.65
Mixed Forest Land
0.78
Streams and Canals
1.00
Lakes
1.00
Reservoirs
1.00
Bays and Estuaries
1.00
Gulf of Mexico
1.00
Forested Wetlands
0.78
Nonforested Wetlands
0.94
Beaches
1.00
Sandy Areas Other Than Beaches
1.00
Bare Exposed Rock
1.00
Strip Mines, Quarries, and Gravel Pits
0.96
Transitional Areas
1.00
36

-------
Table 14. Mean Displacement Height per Land Use Type (for north wind)

Mean
Mean
Mean
USGS Level 2 Land Use Name
Displacement
Height—
Raupach (m)
Displacement
Height—
Macdonald (m)
Displacement
Height—
Bottema (m)
Residential
4.69
5.94
2.89
Commercial and Services
3.42
4.74
3.65
Industrial
2.81
4.21
2.63
Transportation, Communications, and Utility
1.68
2.04
0.17
Mixed Industrial and Commercial
3.13
4.52
3.14
Mixed Urban or Built-Up Land
4.21
4.82
3.02
Other Urban or Built-Up Land
3.90
3.42
2.22
Cropland and Pasture
3.06
1.39
0.18
Orchards, Groves, Vineyards, and Nurseries
3.06
1.39
0.18
Confined Feeding Operations
3.06
1.39
0.18
Other Agricultural Land
3.06
1.39
0.18
Herbaceous Rangeland
3.06
1.39
0.18
Shrub and Brush Rangeland
3.06
1.39
0.18
Mixed Rangeland
3.06
1.39
0.18
Deciduous Forest Land
8.15
10.87
13.23
Evergreen Forest Land
8.02
10.94
11.26
Mixed Forest Land
7.69
10.16
10.36
Streams and Canals
0.00
0.00
0.00
Lakes
0.00
0.00
0.00
Reservoirs
0.00
0.00
0.00
Bays and Estuaries
0.00
0.00
0.00
Gulf of Mexico
0.00
0.00
0.00
Forested Wetlands
7.69
10.16
10.36
Nonforested Wetlands
3.06
1.39
0.18
Beaches
0.00
0.00
0.00
Sandy Areas Other Than Beaches
0.00
0.00
0.00
Bare Exposed Rock
0.00
0.00
0.00
Strip Mines, Quarries, and Gravel Pits
1.08
0.82
0.02
Transitional Areas
0.00
0.00
0.00
37

-------
Table 15. Mean Roughness Length per Land Use Type (for north wind)

Mean
Mean
Mean
USGS Level 2 Land Use Name
Roughness
Length—
Raupach (m)
Roughness
Length—
Macdonald (m)
Roughness
Length—
Bottema (m)
Residential
0.86
0.24
1.05
Commercial and Services
0.72
0.24
0.98
Industrial
0.61
0.11
0.74
Transportation, Communications, and Utility
0.15
0.12
0.17
Mixed Industrial and Commercial
0.67
0.17
0.86
Mixed Urban or Built-Up Land
0.67
0.35
1.10
Other Urban or Built-Up Land
0.72
0.83
1.84
Cropland and Pasture
0.67
1.15
1.01
Orchards, Groves, Vineyards, and Nurseries
0.67
1.15
1.01
Confined Feeding Operations
0.67
1.15
1.01
Other Agricultural Land
0.67
1.15
1.01
Herbaceous Rangeland
0.67
1.15
1.01
Shrub and Brush Rangeland
0.67
1.15
1.01
Mixed Rangeland
0.67
1.15
1.01
Deciduous Forest Land
0.93
0.00
0.48
Evergreen Forest Land
1.21
0.07
0.54
Mixed Forest Land
0.90
0.02
0.29
Streams and Canals
0.00
0.00
0.00
Lakes
0.00
0.00
0.00
Reservoirs
0.00
0.00
0.00
Bays and Estuaries
0.00
0.00
0.00
Gulf of Mexico
0.00
0.00
0.00
Forested Wetlands
0.90
0.02
0.29
Nonforested Wetlands
0.67
1.15
1.01
Beaches
0.00
0.00
0.00
Sandy Areas Other Than Beaches
0.00
0.00
0.00
Bare Exposed Rock
0.00
0.00
0.00
Strip Mines, Quarries, and Gravel Pits
0.20
0.24
0.48
Transitional Areas
0.00
0.00
0.00
38

-------
Table 16. Assumed Building Material Fraction per Land Use Type
USGS Level 2 Land Use Name
Concrete
Wood
Steel
Brick
Mixed
Residential
0.10
0.30
0.00
0.30
0.30
Commercial and Services
0.30
0.10
0.30
0.10
0.20
Industrial
0.10
0.05
0.60
0.05
0.20
Transportation, Communications, and Utility
—
—
—
—
—
Mixed Industrial and Commercial
0.20
0.08
0.45
0.07
0.20
Mixed Urban or Built-Up Land
0.20
0.20
0.20
0.20
0.20
Other Urban or Built-Up Land
0.50
0.10
0.00
0.30
0.10
Cropland and Pasture
—
—
—
—
—
Orchards, Groves, Vineyards, and Nurseries
—
—
—
—
—
Confined Feeding Operations
—
—
—
—
—
Other Agricultural Land
—
—
—
—
—
Herbaceous Rangeland
—
—
—
—
—
Shrub and Brush Rangeland
—
—
—
—
—
Mixed Rangeland
—
—
—
—
—
Deciduous Forest Land
—
—
—
—
—
Evergreen Forest Land
—
—
—
—
—
Mixed Forest Land
—
—
—
—
—
Streams and Canals
—
—
—
—
—
Lakes
—
—
—
—
—
Reservoirs
—
—
—
—
—
Bays and Estuaries
—
—
—
—
—
Gulf of Mexico
—
—
—
—
—
Forested Wetlands
—
—
—
—
—
Nonforested Wetlands
—
—
—
—
—
Beaches
—
—
—
—
—
Sandy Areas Other Than Beaches
—
—
—
—
—
Bare Exposed Rock
—
—
—
—
—
Strip Mines, Quarries, and Gravel Pits
—
—
—
—
—
Transitional Areas
—
—
—
—
—
39

-------
Table 17. Mean Percent DCIA per Land Use Type
USGS Level 2 Land Use Name
Mean DCIA (%)
Residential
26.5
Commercial and Services
86.4
Industrial
74.1
Transportation, Communications, and Utility
70.3
Mixed Industrial and Commercial
80.3
Mixed Urban or Built-Up Land
49.7
Other Urban or Built-Up Land
24.5
Cropland and Pasture
1.2
Orchards, Groves, Vineyards, and Nurseries
1.2
Confined Feeding Operations
1.2
Other Agricultural Land
1.2
Herbaceous Rangeland
1.2
Shrub and Brush Rangeland
1.2
Mixed Rangeland
1.2
Deciduous Forest Land
2.3
Evergreen Forest Land
0.0
Mixed Forest Land
1.2
Streams and Canals
0.0
Lakes
0.0
Reservoirs
0.0
Bays and Estuaries
0.0
Gulf of Mexico
0.0
Forested Wetlands
1.2
Nonforested Wetlands
1.2
Beaches
0.0
Sandy Areas Other Than Beaches
0.0
Bare Exposed Rock
0.0
Strip Mines, Quarries, and Gravel Pits
3.3
Transitional Areas
0.0
40

-------
Table 18. Fraction Land Cover per Land Use Type
USGS Level 2 Land Use Name
Roadway
/Parking
Building
Vegetation
Open
Water
Other
Residential
0.210
0.219
0.570
0.000
0.001
Commercial and Services
0.610
0.255
0.135
0.000
0.000
Industrial
0.471
0.240
0.230
0.002
0.057
Transportation, Communications, and Utility
0.507
0.007
0.267
0.000
0.219
Mixed Industrial and Commercial
0.541
0.248
0.182
0.001
0.028
Mixed Urban or Built-Up Land
0.356
0.219
0.424
0.000
0.000
Other Urban or Built-Up Land
0.184
0.070
0.725
0.005
0.017
Cropland and Pasture
0.022
0.002
0.975
0.001
0.000
Orchards, Groves, Vineyards, and Nurseries
0.022
0.002
0.975
0.001
0.000
Confined Feeding Operations
0.022
0.002
0.975
0.001
0.000
Other Agricultural Land
0.022
0.002
0.975
0.001
0.000
Herbaceous Rangeland
0.022
0.002
0.975
0.001
0.000
Shrub and Brush Rangeland
0.022
0.002
0.975
0.001
0.000
Mixed Rangeland
0.022
0.002
0.975
0.001
0.000
Deciduous Forest Land
0.023
0.000
0.977
0.000
0.000
Evergreen Forest Land
0.000
0.000
1.000
0.000
0.000
Mixed Forest Land
0.012
0.000
0.988
0.000
0.000
Streams and Canals
0.000
0.000
0.000
1.000
0.000
Lakes
0.000
0.000
0.000
1.000
0.000
Reservoirs
0.000
0.000
0.000
1.000
0.000
Bays and Estuaries
0.000
0.000
0.000
1.000
0.000
Gulf of Mexico
0.000
0.000
0.000
1.000
0.000
Forested Wetlands
0.012
0.000
0.988
0.000
0.000
Nonforested Wetlands
0.022
0.002
0.975
0.001
0.000
Beaches
0.000
0.000
0.000
0.000
1.000
Sandy Areas Other Than Beaches
0.000
0.000
0.000
0.000
1.000
Bare Exposed Rock
0.000
0.000
0.000
0.000
1.000
Strip Mines, Quarries, and Gravel Pits
0.034
0.008
0.079
0.027
0.851
Transitional Areas
0.000
0.000
0.000
0.000
1.000
Table 19. Comparison Statistics for the Calculated and
Extrapolated Building Height Characteristics

Bias
RMSE
CRE
Mean Building Height
1.024 (+23%)
1.643
28%
Standard Deviation of Building Height
1.471 (+75%)
1.849
83%
Footprint Area-Weighted Mean Height
1.153 (+23%)
2.392
34%
Wall-to-Plan Area Ratio
0.009 (+9%)
0.052
31%
Height-to-Wdth Ratio
0.004 (+11%)
0.022
37%
41

-------
Table 20. Comparison of Mean Building Height and Standard Deviation for Selected
USGS Level 2 Land Use Types (Note: The derivation area is 1235 km2 and the
validation area is 418 km2.)
USGS Level 2 Land Use
Name
Mean Building
Height (m)-
Derivation
Mean Building
Height (re-
validation
Standard
Deviation (m)-
Derivation
Standard
Deviation (re-
validation
Residential
5.70
4.74
3.04
2.06
Commercial and Services
6.05
5.85
7.07
5.81
Industrial
6.09
4.95
3.73
2.24
Transportation,
Communications, and Utility
4.81
4.17
3.54
2.26
Other Urban or Built-Up Land
4.95
4.68
2.73
2.80
Cropland and Pasture
5.02
4.94
3.20
2.78
Deciduous Forest Land
7.32
5.67
4.42
3.32
42

-------
15
Mean Building
Height
¦a 10

5
0
0
5
10
15
Calculated
Figure 16. Scatter plot of extrapolated versus calculated mean building height (m) for the 418 grid
cells in the validation area.
15
Standard Deviation of
Building Height
10
5
rl
0
0
5
10
15
Calculated
Figure 17. Scatter plot of extrapolated versus calculated standard deviations of building height
(m) for the 418 grid cells in the validation area.
43

-------
15
Plan-Area-Weighted
Mean Building Height
¦a 10
~V
5
0
0
5
10
15
Calculated
Figure 18. Scatter plot of extrapolated versus calculated plan-area-weighted mean building height
(m) for the 418 grid cells in the validation area.
44

-------
0.75
¦a 0.50
a»
4-1
o
Q.
2
X
111 0.25
0.00
Wall-to-Plan
Area Ratio
*~ ~*#. --
0.00
0.25	0.50
Calculated
0.75
Figure 19. Scatter plot of extrapolated versus calculated wall-to-plan area ratio for the 418 grid
cells in the validation area.
0.25
0.20
~o
-------
Table 21. Comparison Statistics for the Calculated and Extrapolated Building Height
Histograms, Plan Area Density, Top Area Density, and Frontal Area Density

Mean
Bias
Min.
Bias
Max.
Bias
Mean
RMSE
Min.
RMSE
Max.
RMSE
Mean
CRE
Min.
CRE
Max.
CRE
Building Height
Histograms
-6.02
-41.52
22.42
39.35
1.55
130.44
80%
26%
576%
Building Plan
Area Density
0.0035
-0.0129
0.0406
0.0143
0.0003
0.0669
65%
23%
193%
Building Top
Area Density
0.0002
-0.0031
0.0063
0.0046
0.0003
0.0119
88%
45%
178%
Building Frontal
Area Density
0.0002
-0.0010
0.0012
0.0008
0.0001
0.0019
55%
25%
185%
Table 22. Comparison Statistics for the Calculated and
Extrapolated Canopy Height Characteristics

Bias
RMSE
CRE
Mean Canopy Height
1.81 (+36%)
2.48
42%
z0, North Wind, Macdonald et al., 1998
0.29 (+107%)
0.44
125%
zd, North Wind, Macdonald et al., 1998
0.61 (+17%)
1.79
38%
z0, North Wnd, Raupach, 1994
0.14 (+23%)
0.24
33%
zd, North Wnd, Raupach, 1994
1.00 (+33%)
1.43
40%
15
Mean Canopy
Height
10

~~~ ~
5
0
0
5
10
15
Calculated
Figure 21. Scatter plot of extrapolated versus calculated mean canopy height (m) for the 418 grid
cells in the validation area.
46
-------
2.0
zo, North Wind, Macdonald
et al. (1998) Equations
T3
a)
re
Q. 1.0
(0
X
HI
•	* >.«V J *
~»»v* ~ ~
~ t / ~, ~ «1 *
~ „ %*~ * ~~
%:* *
~ \ ~~ .~ ~
v
~	*>*
0.0
0.5
1.0
Calculated
1.5
2.0
Figure 22. Scatter plot of extrapolated versus calculated roughness lengths (m) for the 418 grid
cells in the validation area using the Macdonald et al. (1998) equations and a north wind azimuth.
15.0
Zd, North Wind, Macdonald
et al. (1998) Equations
-a 10.0
o
4->
re
o
Q.
re
4->
X
lu 5.0
. ~ t	:

0.0
0.0
5.0	10.0
Calculated
15.0
Figure 23. Scatter plot of extrapolated versus calculated displacement heights (m) for the 418 grid
cells in the validation area using the Macdonald et al. (1998) equations and a north wind azimuth.
47
-------
2.0
Zo, North Wind, Raupach
(1994) Equations
O
Q.
(0
X
LU
~ ~~~
~~
0.5
0.0
0.0	0.5	1.0	1.5	2.0
Calculated
Figure 24. Scatter plot of extrapolated versus calculated roughness lengths (m) for the 418 grid
cells in the validation area using the Raupach (1994) equations and a north wind.
15.0
Zd, North Wind, Raupach
(1994) Equations
¦a 10.0
Q>
O
Q.
(0
t-
X
LU
5.0
~
0.0
0.0	5.0	10.0	15.0
Calculated
Figure 25. Scatter plot of extrapolated versus calculated displacement heights (m) for the 418 grid
cells in the validation area using the Raupach (1994) equations and a north wind.
48
-------
Table 23. Comparison Statistics for the Calculated and Extrapolated
Canopy Plan Area Density and Top Area Density

Mean
Bias
Min.
Bias
Max.
Bias
Mean
RMSE
Min.
RMSE
Max.
RMSE
Mean
CRE
Min.
CRE
Max.
CRE
Canopy Plan
Area Density
0.0194
-0.0492
0.2529
0.0708
0.0107
0.3132
57%
8%
171%
Canopy Top
Area Density
-0.0013
-0.0117
0.0101
0.0125
0.0040
0.0356
53%
15%
124%
49
-------
5. Houston Urban Canopy Parameters
This section contains a brief overview and summary of the Houston UCPs. The figures shown
were created to display the final gridded dataset and begin the process of validating the dataset
and interpreting the spatial distribution patterns. The complete UCP dataset is included with the
accompanying spreadsheets.
5.1	Building Height Characteristics
For the discussion of the building height characteristics, the focus primarily will be the Harris
County grid cells because of the concentration of buildings in this region of the modeling
domain. Figures 26 and 27 display the spatial distribution of the mean and standard deviation of
building height for Harris County. For reference, Figures 28 and 29 show the building count per
grid cell for Harris County and the modeling domain, respectively. The building count
information probably somehow should be factored into, or paired with, the other building height
data when trying to determine drag effects. Some building height characteristics might not
accurately represent the entire grid cell if the number of buildings in the cell is small. For
example, assume the mean building height in one grid cell is 7.5, but that grid cell only contains
three buildings. The overall effect on the wind flow because of those three buildings is small, but
may be exaggerated by simply using the mean height without a count of buildings. Note in
Figures 26 and 27 the distinct separation between the data contained in Harris A (actual
building data) and the other parts of Harris County.
Figures 30 and 31 illustrate the building wall-to-plan area ratio for Harris County and the
modeling domain, respectively. Clearly, a trend is noticeable with the building parameters.
Houston contains the highest values and a concentration of the highest values of the
parameters, with pockets of elevated values spread among the other urbanized areas in Harris
County and the modeling domain. Figure 2 contains a figure of the land use for the modeling
domain, which follows closely with the building parameter results (for the domain, this is
expected because of the UCP-land use correlation). Figures 32 and 33 show the building
height-to-width ratio parameter for Harris County and the modeling domain, respectively. The
overall spatial pattern is again similar to the other building parameters. One overall observation
from the building parameters is the aggregation of elevated values within the Downtown Core
Area. High values also are observed in several Commercial and Services and Industrial centers
within and near Houston. Some high values are noticed in Residential (multifamily) districts.
Figures 34 and 35 display the building plan area fraction parameter in Harris County and the
modeling domain, respectively. One interesting observation is the location of the highest values
in the Commercial and Services and Residential districts in the city, in addition to the expected
high values within the downtown core area. Industrial areas near the ship channel, however, did
not have high building plan area fractions. Similar patterns are noted in Figures 36 and 37,
which show the spatial distribution of the building frontal area index parameter for Harris County
and the modeling domain, respectively.
5.2	Vegetation Height Characteristics
Figures 38 and 39 illustrate the spatial distribution of mean vegetation height per grid cell in
Harris County and the modeling domain. The elevated vegetation heights in the northern part of
Harris County correspond to a heavily forested region. This correlation is more noticeable in
Figure 39 where the forested regions of the northern part of the modeling domain are delineated
clearly. Similar spatial distributions to the vegetation height are noted for the vegetation plan
area fraction and frontal area index (see Figures 40-43).
50
-------
Mean Building Height (m)
No Buildings
2-4
4-5
5-6
6-7
7-15
15-45
Downtown
Core Area
Ship Channel
Industry
o	20 Kilometers
Figure 26. Spatial distribution of mean building height in Harris County ,
Standard Deviation {m)
No Buildings
0~ 1
1 -2
2-3
3-4
4 - 5
5- 15
15-55
20 Kilometers
Figure 27. Spatial distribution of standard deviation of building height in Harris County.
51
-------
Number of Buildings
no
H 0-100
¦I 100 -300
| 300 - 500
500 -750
750 -1250
1250 - 2500
20 Kilometers
Figure 28. Spatial distribution of number of buildings in Harris County.
1


Number of Buildings
" Gulf of Mexico
No Buildings
1 I 0 -100
g 100 -300
j 300 - 500
500 - 750
750 -1250
1250 -2500
N
~

100 Kilometers
Figure 29. Spatial distribution of number of buildings in the modeling domain.
52
-------
Wall-to-Plan Area Ratio

o

0-0.02

0,02-0.05

0.05-0.1

0.1 - 0.2

0.2 - 0.3

0.3 - 0,4
i ]
0,4-0.9
N

20 Kilometers
Figure 30. Spatial distribution of average wall-to-plan area ratio in Harris County.

Wall-to-Plan Area Ratio
HGUtf of Mexico
No Buildings
0 - 0.02
0.02 - 0.05
H 0.05-0.1
gj 0.1 - 0,2
0.2 - 0.3
J 0.3 - 0.4
0.4 - 0.9
N

100 Kilometers
Figure 31. Spatial distribution of average wall-to-plan area ratio in the modeling domain.
53
-------
.• -li I :,| _ .igiU;
Height-to-Width Ratio
No Buildings
gg 0-0.01
H 0 01 -0 025
| 0.025-0.05
" 0.05-0.1
0.1-0.15
J 0.15-0.3
H03 - 06
; -V:: 'nffllh •

N
A
' *L\

A


20 Kilometers

Figure 32. Spatial distribution of average building height-to-width ratio in Harris County.
*
'
*	V"	i
* '	• • r ....	} '	,	i«
' • '	.	s!. -£ <	-	f	' '	"
1 I i -	,	t **¦ ,
' i 't *
H-r" 4 v*• v '
r «
JJ'r
Q'r- .&
;• mm
Height-to-Width Ratio
j Gulf of Mexico
no Buildings
ao - 0.01
O.Ot - 0.025
0.025 - 0.05
0.05-0.1
0.1 *0.15
~ 0.15-0.3
0.3 - 0.6
N
\
100 Kilometers
Figure 33. Spatial distribution of average building height-to-width ratio in the modeling domain.
54
-------
Building Plan Area Fraction
20 Kilometers
Figure 34. Spatial distribution of building plan area fraction in Harris County.
bs*h
- •;

- ^
¦'*. Ha
Kt * '
. %
• '	• >, / V . .
' i. r'A rSt 1
Hill I

u
£
•w
(5 Si
Building Plan Area Fraction
Gulf of Mexico
Q
0-0.025
0.025 -0.05
0.05-0.1
0.1 -0.15
0.15-0.2
0.2 - 0.4
N
\
100 Kilometers
Figure 35. Spatial distribution of building plan area fraction in the modeling domain.
55
-------
Building Frontal Area Index

0

0 - 0.01

0.01 - 0.025

0.025 -0,05

0.05 - 0.075

0.075 -0.1

0.1 - 0.3
N

20 Kilometers
Figure 36. Spatial distribution of building frontal area index in Harris County.

S3 < I •:
I	\ *1 • I "* J

xi /
is -< ¦
. f'
. b .
«¦ •
-4
Building Frontal Area Index

Gulf of Mexico

No Buildings

0 -0.01

0.01 -0.025

0.025 - 0.05

0.05 - 0.075
zz
0.075 -0.1

0.1 -0,3
N
A
100 Kilometers
Figure 37. Spatial distribution of building frontal area index in the modeling domain.
56
-------
Mean Vegetation Height (m)
| No Vegetation
0-1
731-2
~~2-4
~ 4-6
IB6'8
8-12
20 Kilometers
Figure 38. Spatial distribution of mean vegetation height in Harris County.
Mean Vegetation Height (m)

Gulf of Mexico
~
No Vegetation
I—I
0-1

1 -2
Hi
2-4

4 -6
I
6-8
!	I
B -12
N
\
100 Kilometers
Figure 39. Spatial distribution of mean vegetation height in the modeling domain.
57
-------
.'..-'..WI#
mm
Will
Vegetation Plan Area Fraction
0
0-0.1
0.1-0.2
0.2 - 0.3
0.3-0,4
0.4 - 0.5
0.5 - 0.7
0.7 - 0.95
20 Kilometers
Figure 40. Spatial distribution of vegetation pian area fraction in Harris County.
y\ ff
ff -v
Vegetation Plan Area Fraction
Gulf of Mexico
No Vegetation
0-0,1
0.1 - 0.2
0.2 - 0.3
0.3 - 0.4
0.4 - 0.5
0.5 - 0.7
0.7-0.95
N
\
100 Kilometers
Figure 41. Spatial distribution of vegetation plan area fraction in the modeling domain.
58
-------

Vegetation Frontal Area Index
H
m
1
No Vegetation
0-0.05
0.05-0.1
0.1 - 0.2
0.2 - 0,3
0.3 - 0 4
0.4 - 0.5
0.5 - 0.7


N
A
20 Kilometers
Figure 42. Spatial distribution of vegetation frontal area index in Harris County.
i
^ \
Vegetation Frontal Area Index
Gulf of Mexico
No Vegetation
0 - 0.05
0.05-0.1
0.1 - 0.2
0.2-0.3
0.3 - 0.4
0.4 - 0.5
0.5 - 0.96
N
\
100 Kilometers
Figure 43. Spatial distribution of vegetation frontal area index in the modeling domain.
59
-------
5.3	Canopy Height Characteristics
Figures 44 to 49 contain graphics of the spatial distribution of the mean canopy height, canopy
plan area fraction, and canopy frontal area index parameters for Harris County and the
modeling domain. An important observation from the canopy data distribution is the spatial
heterogeneity in Harris County. This can be attributed, first of all, to the spatial heterogeneity of
the urban terrain but also likely is partially because of the fact that the canopy parameters in
Harris County were calculated based on actual data, whereas the parameters outside Harris
County are based on the area-weighted average extrapolation. As noted earlier, the
extrapolation will tend to reduce the heterogeneity of the parameter across the modeling
domain.
5.4	Displacement Height and Roughness Length
The displacement height (zd) and roughness length (z0) were calculated using three sets of
equations: (1) Macdonald et al. (1998), (2) Raupach (1994), and (3) Bottema (1997). Figures
50-61 display the roughness length and displacement height for the three calculation methods in
both Harris County and the modeling domain. Comparison of the results of the three techniques
indicates that there is considerable variability from one method to another. The investigation of
the variability between the three morphometric techniques will be part of a separate study and
will be reported in the future.
5.5	Sky View Factor
Figure 62 shows the spatial distribution of the sky view factor for Harris County. The grid cells
containing mostly open space, including grid cells within the ship channel and near Galveston
Bay, have the highest sky view factors. The forested areas at the north end of Harris County
have the lowest sky view factors. Sky view factors are similar in urban areas and forest areas;
therefore, the urban area is not distinguished in the figure as it is for other parameters. Figure
63 shows the distribution of the sky view factor for the modeling domain. Areas near the coast
have sky view factors near 1.0, whereas the forested areas to the north are mostly in the range
of 0.6 to 0.7.
5.6	Percent Directly Connected Impervious Area
Figure 64 shows the spatial distribution of the fraction of DCIAs for Harris County. The
downtown core area has the highest values (approaching 90% impervious area directly
connected to the stormwater drainage system), with slightly lower values in adjacent
Residential, Commercial and Services, and Industrial areas. Outside of Houston, the fraction
DCIA values are very close to zero (shown in gray in the figure) for most areas. Figure 65
shows the distribution of the fraction DCIA for the modeling domain. The Houston metropolitan
area is clearly delineated, as are the major interstate highways (1-10 is the east-west line and I-
45 runs north out of Houston). Other smaller cities also are shown. Water bodies other than the
Gulf of Mexico are shown as contiguous white areas. White areas to the north of the modeling
domain include lakes, forested land, and wetlands.
60
-------
Canopy Height (mj
0
0-2
2-4
4-6
6-6
8-12
12-30
N
A
r i
20 Kilometers
Figure 44. Spatial distribution of mean canopy height in Harris County.
00 Kilometers
Figure 45. Spatial distribution of mean canopy height in the modeling domain.
Mear
61
-------
ttfiWr fttH
jScJMjL^L£i?%BL ffnWPfV% Canopy Plan Area Fraction
f

0
0-0.2

0.2 - 0.4
0.4 - 0.5


0.5 - 0.6
0.6 - 0.8

¦i
0.8 - 1
N
A
20 Kilometers

Figure 46. Spatial distribution of canopy plan area fraction in Harris County.
SsSwV'jifW'	'
mm.- v
- 1 W

>i <¦ W~ '
i*
K I
Canopy Plan Area Fraction
Guff of Mexico
0
0 - 0.2
0.2 - 0.4
0.4 - 0.5
0,5-0.6
0.6-0.8
0.8-1
N

r*="
100 Kilometers
Figure 47. Spatial distribution of canopy plan area fraction in the modeling domain.
62
-------
0	20 Kilometers
Canopy Frontal Area Index
0
0-0.1
0.1 -0.2
| 0.2 - 0.3
0.3 - 0.5
I 0.5 - 0.7
0.7 - 1
Figure 48. Spatial distribution of canopy frontal area index in Harris County.
,.V
r>-
Canopy Frontal Area Density
J Gulf of Mexico
0
J 0-0.1
0.1 - 0.2
I 0.2 - 0.3
0.3 - 0.5
j 0,5 - 0.7
N

100 Kilometers
Figure 49. Spatial distribution of canopy frontal area index in the modeling domain.
63
-------



¦
0
0-0.1
0.1 - 0.2
0.2 - 0.4
0.4 - 0.6
0.6 -1
1-4
la * * g sz * xj "i

1 1 
-------

Displacement Height (m)
LJ
0	O
1

1 -2

2-4

4-6
6-10
¦1
10-25
20 Kilometers
Figure 52. Spatial distribution of displacement height in Harris County calculated using the
Macdonald et al. (1998) set of equations for a north wind azimuth.
Displacement Height (m)
Guff of Mexico
100 Kilometers
Figure 53. Spatial distribution of displacement height in the modeling domain calculated using the
Macdonald et al. (1998) set of equations for a north wind azimuth.
65
-------
Roughness Length (m)
0
0-0,1
0.1 -0.2
0.2 - 0.4
0.4 - 0.6
0.6-1
1	-4
20 Kilometers
Figure 54. Spatial distribution of roughness length in Harris County calculated using the Raupach
(1994) set of equations for a north wind azimuth.
/-v
IW J
w We
1 .tfi -

Roughness Length (m)
Golf of Mexico
0
0-0.1
0.1-0.2
0.2 - 0.4
0.4 - 0.6
0.6-1
1	-4

•• ' J
.%« '* V" "
'T*^4
i £
¦ ij? ¦
N

100 Kllornelers
Figure 55. Spatial distribution of roughness length in the modeling domain calculated using the
Raupach (1994) set of equations for a north wind azimuth.
66
-------
Displacement Height (m}
20 Kilometers
Figure 56. Spatial distribution of displacement height in Harris County calculated using the
Raupach (1994) set of equations for a north wind azimuth.
Y Displacement Height (m)
Gulf of Mexico
too Kilometers
Figure 57. Spatial distribution of displacement height in the modeling domain calculated using the
Raupach (1994) set of equations for a north wind azimuth.
67
-------
fw^sSnin?*'
20 Kilometers
Roughness Length (m)
0
0-0.1
0.1 - 0.2
0.2 - 0.4
0.4 - 0.6
0.6-1
1	-4
Figure 58. Spatial distribution of roughness length in Harris County calculated using the Bottema
(1997) set of equations for a north wind azimuth.
¦t" 4 V -	v, «*•
* & - •	• - - v ¦ *
Roughness Length (m)
Gulf of Mexico
0
0-0.1
0.1-0.2
0.2 - 0.4
0.4 - 0.6
0.6-1
1	-4
N

100 Kilometers
Figure 59. Spatial distribution of roughness length in the modeling domain calculated using the
Bottema (1997) set of equations for a north wind azimuth.
68
-------
20 Kilometers
Displacement Height (m)
0
0-1
1-2
2-4
4-6
6-10
10-25
Figure 60. Spatial distribution of displacement height in Harris County calculated using the
Bottema (1997) set of equations for a north wind azimuth.
Displacement Height (m)
Gulf of Mexico
100 Kilometers
Figure 61. Spatial distribution of displacement height in the modeling domain calculated using the
Bottema (1997) set of equations for a north wind azimuth.
69
-------
• i*5f» i jltk. Sky View Factor


0.6 - 0.7
¦ jJtit:it: ¦ if ^ *wyyr

0.7 - 0.8


0.8 - 0.9


0.9-0.95
* Wf§ f ¦ -*¦ ¦ r/^TC:: -iMFy |»

0.95 -1
nl1! 1 f rili ,rrt~" ^ il J jr^TTTTiiiiijfjTp
M ' -TfcW* *** < i! ' to- •«•••' •* M "^3 "C.
JL '"'' U mti- *J|:T i t * Kit:1 •+*'*¦*Biii'i PJff • »: jjf lefffHfeg: • ' jfk




N
A

1
l\
IF. .L ift —.
0
10 Kilometers
* "¦V'v " »>#C jHifffc i "


Figure 62. Spatial distribution of sky view factor in Harris County.


5 Sky View Factor




Gulf of Mexico

y ^aPuJj\ '¦" ]^V9-Y • H y K. \:^pi R *y\


0.6-0.7

• ¦&flSSm '0|tJe$b S 0 p '$£?&'
jL- * 1 Jk.ilt-A,r'.T**' 'atflmTJ V> ¦* *5 « - . ¦» -J . ' - • * s i \T*' ?


0.7-0.8

/— t.t "* d J ' -J «4 '*
^•V'T Wm^1 * ' **-"
~ * ©''Sfc.-t * " jSScfiK' i£\f® ¦*


0.8-0.9
0.9-0.95
0.95-1
1
& VvSF 7^'*
fliw

N
A
1

0
190 Kilometer*
-
U i


A
-


Figure 63. Spatial distribution of sky view factor in the modeling domain.
70
-------
Fraction DCiA
0
"| 0 - 0.05
0 05 ~ 01
| 0.1 - 0.2
] 0.2 - 0.4
| | 0.4 - 0.6
ri, o.6 - 0.9
N
20 Kilometers
Figure 64. Fraction of each grid cell in Harris County that is directly connected impervious area
(DCIA).
Fraction DCIA
i Guff of Mexico
Figure 65. Fraction of each grid cell in the modeling domain that is directly connected impervious
area (DCIA).
71
-------
6. Summary
The project described in this report involved the processing of a high-spatial-resolution digital
terrain dataset using GIS and other computational tools. The objective was to compute a
gridded set of urban canopy parameters for use in the CMAQ/MM5/DA-SM2-U modeling
system. The set of derived UCPs for each land use, the UCPs used for extrapolation, and the
final gridded dataset are included in the spreadsheets accompanying this report. The final
dataset contains the following number of UCPs:
•	16 UCPs required only one value per grid cell; 82,368 grid cells [1,317,888 total values]]
•	9 UCPs (Plan Area Densities, Top Area Densities, and Frontal Area Densities) are given as a
function of height (one value per meter for a range of 33 to 297 m) for each grid cell
[-74,000,000 total values/;
•	2 UCPs (Land Cover Fraction and Building Material Fraction) have five values per grid cell
[823,680 total values]]
•	1 UCP (Building Height Histograms) has 62 values per grid cell (62 height increments)
[5,106,816 total values/; and
•	and the land use fraction has 29 values per grid cell [2,388,672 total values].
In total, the results spreadsheets contain approximately 84 million UCP values! Because of the
voluminous amount of UCPs, a comprehensive summary and presentation is not possible.
Instead several figures displaying the data spatially were prepared. The Excel spreadsheets
contain the entire set of UCPs.
72
-------
7. References
Anderson, J. R., Hardy, E. E., Roach, J. T., and Witmer, R. E. (1976). A land use and land cover
classification system for use with remote sensor data. USGS Professional Paper 964, U.S.
Geological Survey.
Bottema, M. (1997). "Urban roughness modelling in relation to pollutant dispersion." Atmospheric
Environment, 31: 3059-3075.
Grimmond, S., and Oke, T. (1999). "Aerodynamic properties of urban areas derived from analysis of
surface form." Journal of Applied Meteorology, 38: 1262-1292.
Macdonald, R. W., Griffiths, R. F., and Hall, D. J. (1998). "An improved method for estimation of surface
roughness of obstacle arrays." Atmospheric Environment, 32: 1857-1864.
Raupach, M. R. (1994). "Simplified expressions for vegetation roughness length and zero-plane
displacement as functions of canopy height and area index." Boundary-Layer Meteorology, 71:
211-216.
USGS (1990). USGeoData 1:250,000 and 1:100,000 Scale Land Use and Land Cover and Associated
Maps Digital Data. U.S. Geological Survey, Reston, VA.
73
-------
¦ fV ^
SEPA
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGE & FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
Recycled/Recyclable Printed on paper that contains a minimum of
50% postconsumer fiber content processed chlorine free
-------