PM3 Deliverable: A05.1
ESS GmbH
PM3 - Particulates monitoring, modelling and
management
LIFE09 ENV/CY/000252
Project Deliverable A05.1
Deliverable Number
A05.1
Deliverable Name:
Integrated Model System: Reference and User Manuals
DDr. Kurt Fedra
Environmental Software & Services GmbH
A-2352 Gumpoldskirchen, Austria
Due Date: PM 09, September 30 2011
Dissemination Level
PU
PP
CO
Public
Restricted to other program participants (including the Commission
Services)
Confidential, only for members of the Consortium (including
Commission Services)
Page 1 of 95
X
PM3 Deliverable: A05.1
ESS GmbH
Introduction
This Deliverable relates to Actions/tasks A05/5.2, Aerosol modelling, long range
transport (using the 3D nested grid model CAMx) with inputs from tasks 1.3 (wind
statistics as the major driving force) , 4.2 (emission estimates), and 4.3, (the
DUST entrainment model), and A05/5.3Nested Grid Modelling (MM5, DUST,
CAMx), which provide input to the emission control optimization (Actions A07,
A08).
The Deliverable describes the integrated model system, embedding the DUST
emission model (see also: Deliverable DA04/4.2) into the AirWare (EUREKA
E!3266 model framework. (for more detail, see the parallel deliverable DA05/5.2
that describes the operational implementation of the model system), and provides
an introduction to the embedded hypermedia on-line manuals, embedded with the
on-line web implementation of the models system.
Technical aspects of the deliverable
All results and associated interactive
model and analysis tools are
accessible from the PM3 model home
page: http://www.ess.co.at/LIFE. The
manual pages are implemented a
hyperlinked HTML pages, that can also
include links to animations (mpeg
files), and in principle, sound files.
Every manual page can also include
any number of further links to related
Manual pages or the GLOSSARY, and,
in principle, any other external web
page through the appropriate URL.
The integrated model system is based on AirWare (http://www.ess.co.at/AIRWARE),
mainly developed in the EUREKA projects E!1388 AIDAIR, and its follow-up, E!3266
WEBAIR (http://www.ess.co.at/WEBAIR).
While not all components of the AirWare system are relevant (and used) for the PM3
LIFE project, the main extension for PM3 is the integration of the DUST model (see
Deliverable DA04/4.2) in the central MM5-CAMx nested grid model system, and its
inclusion in the optimization model component.
The basic philosophy of the embedded hypertext manual page is a combination of
 rapid prototyping (as part of the OO development philosophy)
 concurrent requirements engineering.
The manual pages and their continuous updates and additions are understood as a
service (part of system support and maintenance) more than a product. A link to an
on-line problem reporting system will be provided to help users submit questions,
Page 2 of 95
PM3 Deliverable: A05.1
ESS GmbH
comments and problem reports to the system, including the manual pages, which will
trigger a specific update or extension of the respective component.
On-line Manuals: Implementation and Table of Contents
While the individual screens of the web interface provide
links to the corresponding specific manual pages,
 Each individual page include in its header a link to the
Table of Contents start page;
 The start page can also be directly addressed, leading
to all primary manual pages (which in turn may contain further links in the
hypertext system.
Manual pages are centrally managed
(for a number of parallel application
servers), and distributed to the
individual servers with incremental
changes. A daily test compares the
latest changes (modification dates)
with the individual server installations
and updates whenever necessary.
The update system compares the
central repository with the manual
implementation on any (local or
remote) application server, and
generates a list of updates due,
missing or unexpected files. The
Subsequent script manages the files
transfers and any deletions or redundant or unexpected files.
Each individual manual page
includes a header that shows
the last release the manual
page was updated for, and
the date of the last update.
The embedded hypertext
manual is structured into the
following sections:
1.
2.
3.
4.
5.
6.
7.
8.
Third-party documentation and background documents
Expert System and DSS
Data Management and GIS
Monitoring data
Emission inventories and emission modeling
Scenarios and Scenario Analysis
Simulation Models
Model output, impact analysis
Page 3 of 95
PM3 Deliverable: A05.1
ESS GmbH
User Manual/Tutorial
The Reference manual is designed as a hierarchically structured hyperlinked
document that covers both the reference manual (including information for system
administrators) and the average, infrequent user. While the first page on every topic
provide a basic description of background and functionality/use, optional links on
each page lead to further, detailed material for experienced users and system
administrators.
CourseWare eLearning environment
In addition to the on-line embedded
Reference manual, an interactive
Tutorial implemented in a web-based
eLearning system is under
development. This adapt course
material from AirWare training
sessions and a seminar “Urban
Ecology” at the University of Innsbruck
(LV: 825073), offered every winter
semester.
The lectures consist of a conditional
sequence of dynamically HTML page
and integrated questions and tests.
The tutorial/course content is
dynamically structures in response to
learner behavior, using a real-time rulebased expert system to drive the
“Socratic Dialog” engine. In addition to
individual questions of embedded test
sequences, the web based
(hypermedia) course content can
include video, animation, but also links
to specific model components and
tools, offered as on-line exercises as
part of the tutorial.
The course content also links to the
Model Glossary (also linked from http://www.ess.co.at/LIFE) for the explanation of
any technical terms and concepts introduced.
Page 4 of 95
PM3 Deliverable: A05.1
ESS GmbH
Reference Model structure and content
The Reference manual is structured into the following eight main sections:
1. Third-party documentation and background documents
2. Introduction, installation, configuration
3. Expert System and DSS
4. Data Management and GIS
5. Monitoring data
6. Emission inventories and emission modeling
7. Scenarios and Scenario Analysis
8. Simulation Models
9. Model output, analysis, impacts
Please note that the structure is based on a simple HTML page, and thus easy to
modify to adapt to any specific application/implementation, based on user feedback.
1. Third-party documentation and background documents
This section refers to the various
on-line manuals for the embedded
public domain models (MM5, WRF,
AERMOD, CAMx).
The manual also links to
related documents, compiled in the
on-line PDF library, currently
holding 654 titles with the primary
classification: air.
The documents in the PDF library
include basic bibliographic
reference, keyword classification,
an abstract, and a copy of or link to
the on-line PDF document for online reading or optional download.
The library is open for users
(excluding guest) to upload new
documents or suggest modification
to the classification and keywords.
Page 5 of 95
PM3 Deliverable: A05.1
ESS GmbH
2. Introduction, installation, configuration
 Introduction


AirWare is an interactive information and decision support system for
urban air quality assessment and management. It is developed in a series
of parallel and synergetic projects, including:
 the EUREKA Projects EU1388 AIDAIR and E! 3266 WEBAIR
 the Telematics Applications Project ECOSIM
 the EESD City-of-Tomorrow project SUTRA
 a number of related research and development projects, see:
http://www.ess.co.at/docs/gallery.html
The current AirWare system is implemented as a distributed client-server
system that supports web access from any industry standard browser,
networked computer or mobile device (SmartPhone, tablet).
 Servers: Linux (Open Source) or Unix platforms
 Clients: any PC or workstation running a standard web browser,
mobile devices (SmartPhones, tablet, NetBook) with Internet access
and a standard browser.
Installation notes
AirWare is installed as a client-server web application, using
 an Apache web server (2.2.9 or);
 several data bases (implemented with MySQL (5.0); Alternatively:
PostgreSQL, ORACLE);
 PHP (5.2.6)
 a number of cgi's (with the models and analytical tools;
 various data and configuration files:
o /gis: map data
o /var/www/html//MANUALS/AIRWARE: manual pages and
imagery
o /var/www/html/templates: HTML templates for OBJECT display.
 configuration of the scheduled tasks in several crontab files.
The system [APPLICATION] is located under the Apache httpd server
root directory:
/var/www/html/[APPLICATION]
and the model cgi's in:
/var/www/cgi-bin
which includes the configuration files for the scheduled model runs.
Configuration files
PHP include files
file: /var/www/html/phpincludes/config.inc.php /* user, host, password used
by php scripts */
Most of the entries are referring to localhost and thus should work in any
installation, unchanged. The following entries need location specific
adaptation:
 $java_host=[JAVAHOST]; << localhost does not work here!!>>
 $db_user['[DBSERVER]']='root';
 $db_password['[DBSERVER]']='';
Page 6 of 95
PM3 Deliverable: A05.1
ESS GmbH
Data Base (MySQL) configuration
file: /var/www/cgi-bin/db_config /* default cgi database connection */
 host
 user
 password
 port-number 3306 default
Color configuration
file: /var/www/cgi-bin/mapcolors /* substance specific color ranges for
model results display*/
 default color scheme with equidistant color ranges, dynamic range
definition
 pre-defined, static, non-linear substance and aggregation-time specific
color schemes.
file: /var/www/cgi-bin/colorranges /* default lower and upper bounds for
dynamic color ranges */
This uses the following format:
ANY ANY 1 100
POLLUTANT_NAME PERIOD MIN MAX
where the first record contains the definition of a global default, used when the
dynamic scheme is selected, but no default definition corresponding to the
current combination of substance/aggregation period can be found.
Exposure thresholds
A similar file defines the exposure thresholds for the computation of population
exposure: /var/www/cgi-bin/exposurelimits which is again a simple ascii file
with the format:
ANY
ANY
100
POLLUTANT_NAMEPERIODTHRESHOLD
which defines:
 the pollutant or index
 the aggregation period
 the default threshold
where the first record with ANY, ANY defines a global default if no specific
value for a given combination of substance and aggregation period can be
found. The values are defined in the unit that is associated with the respective
pollutant or index in the Knowledge Base.
CAMx model configuration
Parameter files needed by CAMx:
/var/www/html/templates/camx/data/
 CAMx5.4.chemparam.4_CF
 CAMx5.4.chemparam.NOx
 CAMx5.4.chemparam.SO2
 CAMx5.4.chemparam.CB4
 CAMx5.4.chemparam.PM10
 rate.2002164.do.mech3
 topconcCB5.dat
 topconcPM10.dat
 topconcSO2.dat
 topconcNOx.dat
Page 7 of 95
PM3 Deliverable: A05.1
ESS GmbH

topconcPM2_5.dat
Shell scripts
Shell scripts for running interactive, Nowcast and Forecast scenarios:
/var/www/cgi-bin/camx/
 runcamx.sh
 runNowcast.sh
 runNOxForecast.sh
 runO3Forecast.sh
/var/www/cgi-bin/aermod/
 runaermod.sh
 runNowcast.sh
 Model CGIs
cgis are located in /var/www/cgi-bin/MODEL in sub-directories under the
corresponding model of function group name. For tabular summary and
description, see: http://www.ess.co.at/MANUALS/AIRWARE/modelcgis.html
DUST specific cgis are located in /var/www/cgi-bin/dustent and include:
 main.cgi
main emission calculations
 cal.cgi
calibration options
 new.cgi
new scenario constructor
 pick.cgi
readback and results for a specific location
 Shell scripts
Shell scripts for running interactive, Nowcast and Forecast model scenarios
are located in the corresponding model or function specific sub-directories
under cgi-bin; please note that the DUST model is triggered by the CAMx
script as part of the emission pre-processing;
/var/www/cgi-bin/camx/
 runcamx.sh
 runNowcast.sh
 runNOxForecast.sh
 runO3Forecast.sh
/var/www/cgi-bin/aermod/
 runaermod.sh
 runNowcast.sh
For the application specific scripts, refer to
http://80.120.147.34/MANUALS/AIRWARE/scripts.html
 Data bases
Several data bases (implemented with MySQL (5.0); Alternatively:
PostgreSQL, ORACLE) are used for data management and the
communication between the different models.
They include the following data bases:
mysql
information_schema
AERMOD
AIR
mysql system data base
mysql system data base
AERMOD model scenarios, for sequences of 24 hourly runs
emission sources: plants, boilers, small stacks, line and area sources
Page 8 of 95
PM3 Deliverable: A05.1
ESS GmbH
AIRDESCRIPTORS Knowledge Base: descriptors
AIRMATRIX
model output and emission matrices(by model/domain, parameter, time
step, and layer)
CAMx
CAMx model scenarios
CONTACT
address DB for source contacts
COUNTRIES
countries for address DB (source contacts and institutions)
DOMAIN
model and emisison inventory domains
GLOBAL
general configuration
GLOBAL_LOG
log of model runs
MM5
MM5 model scenarios
MODEL_TS
timeseries descriptors (parameters used in the monitoring time series)
MONITORING_TS timeseries data
RBO
object class information, monitoring stations, timeseries
TEHRAN
census/population data for census tracks (polygon/areas)
USER_MAN
User/access management
gis
individual (user configured) maps
mapc
GIS meta data, map set, overlays
for a list of the data base TABLES in each of these data bases, please refer to
http://www.ess.co.at/MANUALS/AIRWARE/databases.html
 Data requirements
The basic requirements to set up an AirWare system and configure and run
the models are as follows:
1. Domain definition:
o Definition of the geographical scope/area, background maps
including the definition of the individual city/industrial areas of
interest, land use, topography (DEM);
2. Monitoring data:
o Meteorological data (minimally for one reference year) at an
hourly resolution:
 wind speed and direction
 air temperature
 precipitation
 early-morning soudings, (optional) hourly mixing height,
please note that these can be estimated from (optional)
cloud cover data alternatively.
Please note that alternatively, meteorological data can be
generated by a prognostic meteorological model (MM5, WRF)
driven with global GFS data from NOAA on a daily basis. These
data can either be used as dynamic data fields (multi-layered) or
by extracting meteorological parameters into existing or
simulated meteorological monitoring stations.
o Air quality observation data (matching the emission and meteo
data sets !) including station locations.
3. Meteo forecasts: to run the model(s) in forecast mode, dynamic
boundary conditions (met data for selected grid points within the
domain) from an appropriate regional meteorological forecasting model
are required. The meteorological forecasts can optionally be suplied by
the prognostic meteorolgical model MM5 Output from MM5 is
Page 9 of 95
PM3 Deliverable: A05.1
ESS GmbH
generated to a 3 km, hourly resolution, automatically interpolated
further with MM5toCAMxto a 1 km resolution by use with the CAMx
nested grid air quality model.
4. Emission data:
o Point (stacks) and area source emissions, locations, source
(stack) characteristics, emission rates for any or all pollutants
covered. Additional (alternative) data include fuels and fuels
consumption, combustion technologies, industrial classifications,
etc.
o Transportation graph/network with link specific data
(frequencies) as well as global properties such as fleet
composition;
o Fugitive emission from the fuel system including parked cars,
gas stations.
Alternative: traffic emission grid;
o Emission inventories for (industrial) point sources and fugitive
emissions, and area sources (domestic heating);
o For the ozone modelling (PBM photochemical box model; for
the 3D (optional) dynamic CAMx model see below):
Hydrocarbons, divided into the following eight groups: Non
reactives, Ethilene, Olefines, Alkanes, Formaldehyde, Aldehyde,
Aromatics and Toluene; and CO, NOx (fixed ratio between NO
and NO2), data for biogenic VOC emissions.
o DUST model:
 wind field (dynamic MM5 or WRF output)
 soil moisture (MM5/WRF generated)
 DEM (30 meter resolution
 land cover, vegetation (several alternative data sources)

5. Population data:
o Population distribution (for exposure analysis) by census tracts
or larger administrative entities, or gridded (1ha to 1 km2);
6. Receptor points: for the monitoring of compliance with standards,
model performance (validation), and impact assessment a number of
receptors should be defined. These are:
o Air quality monitoring stations
o Building points (user positioned arbitrary receptor points)
o Receptor areas (user defined arbitrary polygons).
7. Economic data:
o Cost functions for emission reduction strategies, by emission
class of major (point source) object or traffic related (fleet
composition).
8. Logo(s) and start-page image.
For the model specific data requirements and default data sets used for the
individual models, please refer to the model specific descriptions.
Page 10 of 95
PM3 Deliverable: A05.1
ESS GmbH
 User Management
ASP systems provide their own user management and authentication system in addition to the Linux operating system level - that can be configured to
provide selective access for local or external users to different levels of
functionality in the system.
User access is based on the concepts of GROUPS and SERVICES: a user
belongs to a group which has read-write access to set of services provided by
the system.
User administration is only accessible to admin group users. The link to the
user administration tools is on the <b<="" b="" style="font-family: Arial, sansserif; ">of the respective application.</b
The primary access to the tool is a list of current users.
The listing provides user name, first and last name, creation date, status
(activated or not) and the last access date for the user.
The auxiliary buttons are GROUPS and NEW;
NEW offers and empty template to add a new user;
GROUPS edits groups and the associated service access rights.
For details, see: http://www.ess.co.at/MANUALS/AIRWARE/useradmin.html
 Model Accuracy
Model accuracy is usually determined by comparing model results with air
quality observations; the basic problems encountered are primarily related to
1. the intrinsic uncertainty of the observations themselves and their
sample nature;
2. the vastly different scales involved when observing and modeling a
turbulent process.
While monitoring data describe concentrations over time at a point location
based on small volumes (several liter) of air samples, the model estimates
average concentration within a time step (e.g., an hour) for a comparatively
very large volume of air (usually several million liters), under assumptions of
complete mixing within the model grid cell/volume or as resulting from a
(tenous at best) steady state assumption.
Annex I of 2008/50/EC defines Data Quality Objectives, but none are defined
for modeling in the context of PM10 or PM2.5 other than for the annual
average: 50%. In general, the Directive states (Annex I):
“The uncertainty for modelling is defined as the maximum deviation of the
measured and calculated concentration levels for 90 % of individual
monitoring points, over the period considered, by the limit value (or target
value in the case of ozone), without taking into account the timing of the
events. The uncertainty for modelling shall be interpreted as being
applicable in the region of the appropriate limit value (or target value in the
case of ozone). The fixed measurements that have to be selected for
comparison with modelling results shall be representative of the scale covered
by the model.”
Page 11 of 95
PM3 Deliverable: A05.1
ESS GmbH
AirWare offers two mechanisms to
compare monitoring data and model
results for the grid cells containing the
monitoring station:
 Model scenario pages: the
button Compliance at the model
scenario levels leads to a pop-up
window that provide a tabular
listing of all monitoring stations in
the domain with their measured
values (if available for the current
model time step) and the model
generated values as well as the respective air quality standard for the
current substance and temporal aggregation; the top of the window
shows the comparison of measured versus simulated for the 24 hourly
values of the scenario.

Time series data analysis: in the time series analysis (display and
analysis of a single time series) the time series of observed data can be
compared with model generated values for that location if the model
was run for the period selected. The two time series are shown in a
common graph for direct comparison.
The tabular summary on the left side includes a selection of
(applicable) error statistics.
 Pollutants
AirWare can represent any number of pollutants, as well as air quality indices
such as PSI, as well as the PSI contributions of each of the five pollutants
used as the basis for the PSI calculations (PSIP):
the set of pollutants covered is completely open for monitoring data and
emissions.
AirWare uses an open list of pollutants. These can be grouped into two distinct
classes:
1. Pollutants emitted (which includes CO2 and VOC, but excludes ozone)
2. Pollutants modeled (which includes O3, ozone, but excludes CO2 and
VOC)
The list of pollutants considered is, in principle, open and defined in a
Descriptor in the systems Knowledge Base and can be changed by the
system administrator. However, numerous and complex dependences make
changing that list nontrivial.
The basic list of pollutants considered in
 the emission inventories
 the emission factors for stationary combustion and road traffic
include:
 CO, CO2, SO2, NO, NO2, NOx, PM10, PM2.5, VOC
For the individual models, there are restrictions based on the specific
substance behavior and the corresponding model processes.
Page 12 of 95
PM3 Deliverable: A05.1
ESS GmbH
PSI contributions:
In addition to the PSI index value (based on data over the last 24 hours, the
following individual subtance data are available as "individual pollutants" on an
hourly, moving average, basis:
 CO: maximum of the 8-hour-averages of the last 24 hours (in reference
to the timestep)
 O3: maximum of the 1-hour-averages of the last 24 hours (in reference
to the timestep)
 NO2: maximum of the 1-hour-averages of the last 24 hours (in
reference to the timestep)
 SO2: 24-hour-average (ending at the timestep)
 PM10: 24-hour-average (ending at the timestep)
The basic set of pollutants can be grouped in terms of their behavior:
1. Physical properties:
o bouyant, dense/heavier than air (heavy gases, particles);
2. Chemical properties:
o conservative
o first order decay, transformation (e.g., NO=>NO2, SO2=>SO4)
o complex photochemistry (NOx+VOC=>O3)
Emission inventories for point, area, and line sources should include:
1. CO2 (for statistical purposes only)
2. SO2
3. NOx (or NO/NO2)
4. CO (for traffic)
5. PM-10, PM-2.5
6. Volatile Organic Compounds (VOC), fugitive from area and line
sources.
The latter group is broken into a number of sub-groups or species, based
on a speciation table that can be associated with any source class, type,
or individual source. The global default speciation table includes:
TOG (Total Organic Gases) is calculated from VOC:
TOG = factor*VOC (default: factor=1); values above one represent the
Methane fraction.
The mass speciation factors for TOG are defined in the
default profile (#0) which can be overloaded by emission source class,
type, and individual source. (Source:
http://www.epa.gov/ttn/chief/emch/speciation/)
ALD2 Higher Aldehyde, based on Acetaldehyde 0.02
ETH Ethene
0.05
FORM Formaldehyde
0.02
OLE Olefin carbon bond C=C
0.05
PAR Paraffin carbon bond C-C
0.49
TOL Toluene and other monoalkyl aromatics 0.05
XYL Xylene and other polyalkyl aromatics
0.04
NR
non reactives
0.28
Page 13 of 95
PM3 Deliverable: A05.1
ESS GmbH
 PSI air quality index
In addition to the basic set of air pollutants, AirWare can also compute and
display air quality indices such as PSI.
PSI requires values from several pollutants and for different aggregation
periods that are generated by the
photochemical model CAMx, both a daily forecasts and as hourly nowcasts
with the corresponding moving averages over the most recent 24 hours,
including the current hour.
1. All CAMx result matrices show averages over a time period. The period
can be 1 hour, 8 hours or 24 hours. All matrices are assigned to their
ending time. For example, the 8-hour-average-matrix of 10:00 - 18:00 is
assigned to 18:00. The 24-hour-average-matrix of September 29th,
7:00 - September 30th, 7:00 is assigned to September 30th, 7:00.
2. For calculation of PSI matrices the following matrices are generated for
each timestep:
o CO: maximum of the 8-hour-averages of the last 24 hours (in
reference to the timestep) is taken for each grid cell.
o O3: maximum of the 1-hour-averages of the last 24 hours (in
reference to the timestep) is taken for each grid cell.
o NO2: maximum of the 1-hour-averages of the last 24 hours (in
reference to the timestep) is taken for each grid cell.
o SO2: 24-hour-average-matrix (ending at the timestep) is taken.
o PM10: 24-hour-average-matrix (ending at the timestep) is taken
3. The concentration values are transformed into PSI values using the
following formula:
psip = ( psi(hi) - psi(lo) ) * (concp - bp(lo) ) / ( bp(hi) - bp(low) ) +
psi(low)
where:
psip
calculated psi index for the pollutant
concp the concentration of the pollutant
bp(lo) lower breakpoint of the interval of pollutant concentration
bp(hi) higher breakpoint of the interval of pollutant concentration
psi(lo) PSI values corresponding to bp(lo)
psi(hi) PSI values corresponding to bp(hi)
bp(lo), bp(hi), psi(lo) and psi(hi) are taken from the PSI color definition.
If the concentration is above the upper limit of the table, PSI is not
defined. In this case the maximum PSI value + 1
is assigned (psip=301 )
4. PSI matrix is generated by taking the maximum of the psip value of the
5 pollutants for each grid cell.
 Particulates
For the calculation of particulates (PM10) AERMOD uses
 the fraction of particle mass emitted in the fine class, less than 2.5
microns,
 the representative mass mean particle diameter in microns.
Page 14 of 95
PM3 Deliverable: A05.1
ESS GmbH
The following default values are applied:
1. If no PM2.5 emissions are specified for a source, the default value of
m(PM2.5)/m(PM10) = 0.4 is used.
2. Default value for mass mean particle diameter for PM2.5 (fine fraction):
Dmm(PM2.5) = 2.0
3. Default value for mass mean particle diameter for PM10 (coarse
fraction) without PM2.5: Dmm(PM10-PM2.5) = 7.5
4. Using the default value of m(PM2.5)/m(PM10) (0.4)leads to a default
average particle diameter Dmm(PM10) = 5.3
CAMx: Dust is split into 'Fine Crustal'/'Coarse Crustal 50/50 for CAMx input.
PM10 is split into 'Fine Other Primary'/'Coarse Other Primary' as
before. Next forecast run will use the new input.
For the DUST model, the following (primary, initial) default sizes are
associated with the soil fractions derived from the FAO data:
Sand (coarse soil fraction)
2.00 – 0.060 mm
Silt (medium soil fraction)
0.06 – 0.002 mm
Clay (heavy soil fraction)
< 0.002 mm
Emission Scenario:
Column header from in the summary now reads: "Matrix".
 "PM10" - pyrogenic PM10 emissions from emission sources,
 "dust" - dust emission from dust entrainment.
In case complex sources exist for the project, an additional column
showing complex sources emissions is added.
For the details of different representation mechanisms for particulates in
CAMx, please refer to the model specific description or the on-line CAMx User
Manual.
 Air Quality Standards
AirWare supports any set of user defined air quality standards (defined in the
application specific configuration files, see above). The standard are used for
the evaluation of model results (at simulated monitoring stations) or monitoring
data time series.
For the implementation for Cyprus and the PM3 project, based on
2008/50/EC, please refer to: http://www.ess.co.at/LIFE/standards.html
 Pasquill Stability Classes
Only applicable to AERMOD, not used in PM3
 ISIC Rev 4.0 codes
For a description of the ISIC code (used for emission source classification),
please refer to: http://unstats.un.org/unsd/cr/registry/regcst.asp?Cl=27
Page 15 of 95
PM3 Deliverable: A05.1
ESS GmbH
3. Expert System and DSS
 XPS: Expert System
The embedded expert system can server a number of functions:
 estimation of parameters, e.g., for simulation models;
 checking completeness, consistency, and plausibility of user
specifications;
 interpretation of complex (model generated) data sets.
 real-time control of scheduled and event-driven complex and
conditional task sequences.
The expert system uses Descriptors as its basic variables, which are
linked in processed in Rules.
Main System Components are defined in separate manual pages, linked
from http://www.ess.co.at/MANUALS/AIRWARE/xps.html
 DESCRIPTORS: the variables of the expert system
 RULES: first order nested production rules
 KB editor: interactive Knowledge Base editor
 Knowledge Base Editor
A dedicated editor is available to manage the Knowledge Base
(Descriptors and associated RULES) for the embedded XPS backward
chaining expert system.
The primary selector shows the Descriptors available, which can be sorted
by name, author, or modification date. With every Descriptor name, the
basic meta data attributes such as name, owner, and modification dates
are shown.
For the descriptor, an interactive editor is available. The editor compiles
the meta data name, description, author, creation and modification dates,
and descriptive information such type and format, and unit where
applicable.
The descriptor definition includes:
 the legal values (symbols and numerical ranges);
 the question text that is used in the editor tool (Java applet);
 the list of associated rules.
The editor also offer s preview function for the definition.
The preview shows how the definition is interpreted by the interactive
editing tool (a Java applet) that uses the Descriptor definition to configure
the tool used to set a specific value.
The rules associated with a Descriptor fall into two groups:
 Rules that set that Descriptor
 Rules that use the Descriptor.
These are first order production rules, implementing first order logic
principles or modus ponens.
The list of rules leads directly to the editing tool for RULES. A dedicated
editor assists in formulating syntactically correct RULES.
Page 16 of 95
PM3 Deliverable: A05.1
ESS GmbH
Within a given template, the user can select the possible values for
operators and symbolic variables (Descriptors) that can be used in the
IF-THEN Rules, including nested AND and OR clauses.
 KB: Descriptors
The facts (data) of expert systems XPS and RTXPS are stored in
DESCRIPTORs.
A value is assigned to a DESCRIPTOR either by direct editing or by
starting the rule-based inference, that will recursively compile information
from the data base including asking the user where necessary, to assign a
value to the target (Descriptor) of the inference.
The system then uses a set of alternative methods enumerated in the
DESCRIPTOR definition to obtain or update the DESCRIPTOR value in
the current context. The inference engine compiles all necessary
information for the appropriate Backward Chaining Rules' input conditions
recursively, evaluates the Backward Chaining Rules, and eventually
updates the target DESCRIPTOR.
Descriptor syntax and descriptions
Descriptors are defined in terms of:
 an internal name used by the code, not to be changed
 a DISPLAY NAME freely changeable by the user;
 a UNIT (text string)
 A TYPE. Currently supported types:
o V (variable, text only values),
o H (hybrid): numerical ranges and associated symbolic range;
o S (symbolics: list of symbols only.
 META DATA: user, modification date, explanatory text.
 TABLE of legal value ranges (type dependent); The legal range is
defined from the lower bound of the first to the upper bound of the
last range.
The range definitions include triplest of numbers
minimim - median (display) - maximum values, followed by a
symbolic description of the range.
Range definitions must be contiguous; the upper bound of the lower
rangte must be identical to the lower bound of the following range.
Median (display default) values are optional, but must be between
(including) the upper and lower bounds for the range If no median
value is defined the arithmetic mean is used.
 Question: HTML style text that is shown in the Descriptor editor.
 KB: Rules
Backward Chaining Rules define how values for DESCRIPTORs are
derived from values of other DESCRIPTORs, user inputs, simulation model
results or other information available (data bases, GIS, etc.).
Rule Syntax
RULE <rule#>
IF <condition>
Page 17 of 95
PM3 Deliverable: A05.1
ESS GmbH
THEN <action>
ENDRULE
<condition> := <condition <logical operator> <condition>
:= descriptor <comparative operator> <operand>
:= descriptor <singular operator>
:= TRUE
<action> := descriptor = <operand>
:= descriptor <meta operator> <constant>
<operand> := <operand> <arithmetic operator> <operand>
:= descriptor
:= <constant>
<logical operator> := AND | OR
<comparative operator> := < | > | <= | >= | == | !=
<singular operator> := EXISTS | NOT_EXISTS
<arithmetic operator> := [ | ] | * | / | + | <meta operator>
:= INCREASES_BY | DECREASES_BY
:= BECOMES
<constant> := string
:= number
Rules can result in the absolute assignment of DESCRIPTOR values, their
relative, incremental modification, or they can be used to control the
inference strategy depending on context. Rules can also include simple
formulas; more complex functions can be used through the generic model
interface.
An example for a rule used in the reservoir expert system is:
RULE 1020231
IF
average_reservoir_depth == small
AND retention_time
< 30
THEN reservoir_stratification = unlikely
ENDRULE
Also, the user can call up a knowledge base browser, that allows to
navigate in the tree structure of the knowledge base within the context of
individual problems. The browser can descend the inference tree,
displaying sets of rules referring to a list of DESCRIPTORs and allow to
inspect individual DESCRIPTOR definitions.
The possibility to integrate models in place of rules in an expert system and
at the same time use embedded rule-based components in models
provides a very rich repertoire of building blocks for interactive software
systems.
 RTXPS Real-time System
RTXPS is a forward-chaining expert system operating in real-time (time
aware): the forward chaining loop is executed continuously, and several
function obtain, compare and use variables based on Absolute (systems)
and Elapsed Time, including user defined TIMERS.
This makes it possible to execute sequences of conditional and scheduled
tasks with complex interdependencies, including the execution and
Page 18 of 95
PM3 Deliverable: A05.1
ESS GmbH
interpretation of dynamic simulation (forecasting or operational control)
models, or ACTIONS based on the values of remote (real-time( sensors.
RTXPS can be run in two distinct modes:
1. automatic, autonomous, continuously; here the forward chaining
loops runs continuously, and is controlled only bythe RULES and
ACTIONS executed;
2. interactively, which include functions that trigger display, interactive
editors, and user dialof in general. Here, the forward chaining loop
can be controlled by the user with interactive input: whenever input
from the user is requested, the system will wait for that input, unless
a TIMEOUT is defined for the dialog with a default continuation
without interactive user input.
A special case or application of RTXPS is CourseWare, an interactive webbased distance learning tool. On the basis of available rules, RTXPS
decides which action is to be conducted next. Actions consist of up to 3
function out of which one should create a html page which is sent to the
client. Pages can contain in-text functions - tools to display or edit
descriptors or graphical elements which are created dynamically.
Files and data bases
The basic files for RTXPS are located in the (default) directory:
/var/www/RTXPS
RTXPS contains the rtxps.cgi executanle, and the following subdirectories,
KB
knowledge base
bin
script and functions as part of ACTIONS
defaults configuration files
logs
dynamic log files
RTXPS uses for its operation:
 the RTXPS executable, rtxps.cgi located in the base directory
/var/www/RTXPS; the RTXPS executable is started manually, but
can be included in the rc.* scripts for automatic start-up at boot time;
please note that RTXPS needs to be retsarted whenever the
knowledge base changes.
 A dynamic Knowledge-Base that stores the current state of the
system in the data base;
 The Knowledge Base configuration files in /var/www/RTXPS/KB:
these include:
 KBconfig: a knowledge base configuration fil, the sets the
maximum dimensions, and enumerates the different files to be
use, multiple entries will be concatenated in the sequence of
their listing;
MaxNr-Of-Descriptors:100
MaxNr-Of-Rules:
MaxNr-Of-RTRules: 100
MaxNr-Of-Actions: 100
100
Page 19 of 95
PM3 Deliverable: A05.1
ESS GmbH
DESCRIPTOR-FILES
Descriptors
END-DESCRIPTOR-FILES
RTRULE-FILES
Rules
END-RTRULE-FILES
ACTION-FILES
Actions
END-ACTION-FILES
 RTXPS Rules
The forward chaingin RULES are very similar to the first order production
rules used in the backward chaining version of the expert system, XPS.
The main difference is the ACTION or consequence part: while in the
backward chaining branch, this is always an ASSIGNMENTsetting the
value of a DESCRIPTOR, the forward chaining system uses a repertoir
of ACTIONS so that the general rule syntax looks like:
IF [CONDITION]
AND/OR [CONDITION] (evaluates to TRUE)
THEN (execute) ACTION
where ACTION can be
 an ASSINGMENT of the form: DESCRIPTOR == VALUE,
 any one of the built in FUNCTIONS;
 a user defined, named ACTION from the RTXPS ACTION
Configuration file.
 Functions
The consequence, THEN or ACTION part of an RTXPS RULE triggers
an ACTIONtriggers an ACTION AN ACTION is either one of the built-in
FUNCTIONS for a user defined, named ACTION.
The built-in generic FUNCTIONs include triggers for simulation models, the
backward chaining expert system, or external communication tasks such
as data acquisition from monitoring systems or tasks such as automatic
dialing for phone connections, or sending automatically generated eMail,
fax or SMS messages.
Other ACTION functions trigger special interactive editors to obtain
information on more complex risk objects (such as trains, plants, etc.)
which require specific dialogue windows for consistent editing of the
attributes of the the risk objects and provide additional functionality such as
links to on-line databases.
From an implementation perspective there are two groups of functions:
 those defined in the library (and thus available in all applications)
 and the application specific FUNCTIONs only implemented for a
specific application such as CourseWare.
FUNCTIONs implemented in the expert system library:
Page 20 of 95
PM3 Deliverable: A05.1







ESS GmbH
set_descriptor_value(<descriptor name>,<value>)
Initializes the DESCRIPTOR <descriptor name> with the
specified value.
Should the DESCRIPTOR already have a value set it will be
overwritten with the new one.
clear_descriptor_value(<descriptor name>)
The value of the DESCRIPTOR <descriptor name> is cleared.
The previous value (if present) is saved in the default value of
the DESCRIPTOR.
execute_rule(<rule#>[,<rule#>,...])
The backward chaining rules with the specified rule numbers
(<rule#>) are fired in the given sequence until one of them has
been sucessfully completed.
At least one rule number must be specified.
execute_rt_rule(<rule#>[,<rule#>,...])
All forward chaining rules with the specified rule numbers
(<rule#>) are fired in the given sequence.
Whenever an ACTION is set to ready the ACTION is
automatically started and afterwards set to done.
At least one rule number must be specified.
export(<descriptor name>)
The value of the DESCRIPTOR <descriptor name> is written to
a file. The file is put into the subdirectory defined by the default
"rtxps.fax.dir" in the directory specified by the log-files-path
(defined in the CONFIG file of the application). The name of the
file is the name of the DESCRIPTOR cut to its first 14
characters.
If the DESCRIPTOR does not have a value the
string *unknown* is written to the file.
system(<descriptor name>)
The value of the DESCRIPTOR <descriptor name> is run as a
UNIX shell command in the directory specified by the default
"rtxps.bin.dir".
If the DESCRIPTOR value is emplty no command is executed.
shell(<cmd>)
The string <cmd> is run as a UNIX shell command in the
directory specified by the default "rtxps.bin.dir".

Interactive RTXPS sessions
 update_descriptor(<descriptor name>[,ask])
The value for the DESCRIPTOR <descriptor name> is set most
commonly triggering the backward chaining inference engine, but
more generally using the resources specified in its DESCRIPTOR
definition (eg., backward chaining rules, data base access,...).
Please note: depending on the parameters and current status of the
dynamic knowledge base, the function update_descriptor() may
also be applicable in an automatic, autonomous RTXPS sessions,
see above.
Page 21 of 95
PM3 Deliverable: A05.1


ESS GmbH
The user is prompted only in case of missing information.
If the value of the DESCRIPTOR has already been set before the
function is called the value is not modified unless the update
parameter is specified. Then the DESCRIPTOR AskBox is
displayed showing the current value and the user can decide if he
wants to keep the value, modify it, or re-run the deduction.
display_html(<filename>)
The content of the file <filename> is displayed in HTML format over
the map display of the RTXPS screen.
Position and extent are taken from the defaults also defining the
position of the map display, ie., "rtxps.map.x", "rtxps.map.y",
"rtxps.map.w", and "rtxps.map.h".
close_html()
The HTML display previously opened with the
FUNCTION display_html(<filename>) is closed.<
 Timers (RTXPS)
There are 3 predefined descriptors and several buil-in functions relating
to time:
1. START_TIME: start time of the current RTXPS cycle, mainly used in
interactive applications such as CourseWare or theRiskWare realtime emergency management applications.
2. CURRENT_TIME: system (local) time for the system defined time
zone, including day saving time shifts, HH:MM:SS
3. CURRENT_UTC: global standard time (Zulu, UTC, GMT),
HH:MM:SS
4. ELAPSED_TIME, difference between CURRENT_TIME and
START_TIME.
5. MINUTE(number): : TRUE if the current minute from HH:MM:SS
matches;
6. HOUR(number): TRUE if the current hour from HH:MM:SS
matches;
7. DAY(number), DAYN(string):
o TRUE if DAY(number) argument matches the number of the
day (1-31) within the current monthy.
o TRUE if the string argument of DAY matches the threecharacter abbreviation for the current days name (system
date function) as in SUN, MON, TUE, WED, THU, FRI, SAT,
case insensitive;
o MONTH(), MONTHM();
o YEAR() TRUE if the argumenrt matches the currenr year
from date (YYYY).
The Descriptor definitions (in the respective application's KnowledgeBase)
are shown below, with the TIMERS being of the TYPE: variable.
Page 22 of 95
PM3 Deliverable: A05.1
ESS GmbH
RTXPS first creates dynamic knowledge base, then sets predefined
descriptors to their current values.
Implementation: CourseWare
START_TIME and ELAPSED_TIME rely on the fact that the first action of
every course calls function init(). These descriptors can be used for
computation as well as display. The behaviour of value(ELAPSED_TIME) is
equivalent to header in-text function\elapsed_time and behaviour
of value(CURRENT_TIME) is equivalent to header in-text function \time.
Besides user can define application specific time descriptors and set their
value using rtxps function set_time(descriptor_name).
All time descriptors can be displayed with rtxps in-text function
 \value whose format is hh:mm:ss or with rtxps in-text function
 \valuef which displays an integer number in seconds.
There are two operations defined for time descriptors:
 addition: add_time(result,operand1,operand2)
 substraction: subst_time(result,operand1,operand2)
see also: RTXPS functions
 Emission control technologies
A data base of Emission control technologies, that are designed to be used
in the optimization scenarios. Access to the data base is from within the
AirWare ONBECT class selector, or directly (linked) from
http://www.ess.co.at/LIFE
 Emission control MCA
The Emission control optimization is currently based on PBM a simple
numerical air quality model that simulates photochemical smog at an urban
scale.
For a given PBM scenario, the buttons EMISSIONS and in the emission
scenario, optimize lead to the optimization interface.
The optimization scenario is a standard OBJECT with header and META
data.
The scenario will list a set of sources, for which control options (emission
reduction at at a cost) can be defined. The set of sources includes:
1. all boilers (major point sources from industrial plants)
2. small stacks (as a class)
3. area sources (as a class)
4. line sources (as a class).
As a basic scenario control parameter, the user can define the number of
runs to be made in the Monte Carlo framework.
Emission control strategies
for each of the emission sources or source classes, the user can edit:
Page 23 of 95
PM3 Deliverable: A05.1
ESS GmbH
F
oINV
rOP
Ref
annualized investment cost of the control policy or technology
annual operating costs (OMR)
reference unit for the OMR costs (time step)
eReduction%the relative emission reduction achievable with a 100%
(complete) application of the policy of technology
a
a minimum application rate
cMin
Max
a maximum application rate
h
a weight factor that defines the probability that the instrument will
Weight
be applied
Once a scenario is completely defined, it can be run with the run button. A
message Running ..... will indicate the status, an optional eMail message
will indicate completion for registered user with their eMail stored in the
user data base.
Optimization results
When the set of runs specified by the user is completed, the results page
(also accessible from the corresponding button) will be shown.
The results list the summary over the set of model runs with the average
application rate and associated investment and operating costs for each
emission source or source class. The columns %min and &max indicate
the minimum and maximum remaining emissions (as % of the original
uncontrolled emissions) from the corresponding source.
The button details generates a ranked listing of the model runs
(alternatives) with a set of four (user selected) criteria displayed, the last
one associated with a graphical representation (color coded horizontal bar)
to symbolize the criteria value. The list can be sorted (ascending or
descending) for any one of the criteria displayed.
Selecting any individual alternative from the list will display its specific
results, i.e., the various technology application rates, cost, and associated
emission reduction.
Post-optimal analysis
To analyse the set of alternatives generated, the data can be exported
to DMC a discrete multi-criteria DSS tool.
The buttonexport will export a data set if results have been generated;
The button DMC will directly lead to the corresponding DMC scenario.
 DMC: discrete multi-criteria
The DMC tool is an interactive discrete, multi criteria optimization system,
using a reference point approach and normalization of the decision space
between nadir and uotpia.
The sets of alternatives to analyze can either be:
1. imported as a set from WRM optimization scenarios (use the button:
DMC" after a successful optimization run at the optimization
scenario overview level to export the feasible alternatives
generated);
2. imported individually from WRM model runs;
3. imported from a CSV file (exported from a spreadsheet)
Page 24 of 95
PM3 Deliverable: A05.1
ESS GmbH
4. created (and edited) manually.
Basic Concepts
The tool operates on named sets of alternatives. Each alternative in an
object that has a set of attributes or criteria that define it. The set of critiera
or dimensions in decision space is the same for all alternatives within a set.
The user can select or deselect any one of the criteria (with a minimum of
two) for consideration or use in the analysis.
Every one of the criteria used is define in the Knowledge Base (KB)
associated with the tool.
For each of the criteria, there is
 a (scalar) value defined
 an optimization strategy or direction (minimize, maximize, meet)
defined; the latter case is a shortcut for minimizing the deviation in
either direction from the reference, in this case, by default
the median of the distribution for that criterion in the set.
 an (optional) constraint (upper or lower bound, or normalized
distance depending on the strategy or direction of optimization)
defined.
The main concepts used include:
 feasible subset: the set of alternatives that meets all of the
constraints;
 UTOPIA: a hypothetical point defined by the "best" value for each
the criteria in the set.
 NADIR: a hypothetical point defined by the "worst" value for each of
the criteria in the set;
 dominated subset: an alternative is considered dominated, if we
can find another in the set that is better in at least one, and not
worse in any other (i.e., at least equal), of the criteria. The
dominated subset is identified and excluded from consideration.
 PARETO set: this consists of all non-dominated alternatives;
 Reference Point: a hypothetical point defined by the user who
select his preferred value for each of the criteria. The default
Reference Point is UTOPIA.
 Efficient Point: this is the feasible and non-dominated alternative
"nearest" to the Reference Point. The distance is determined as the
N (number of criteria) dimension Euclidean distance in the spce
define by all (active) critiera normalized between NADIR and
UTOPIA. The distance is described as a (normalized in %) level of
achievement or (100% - distance): 100% would be UTOPIA itself,
50% halfway between UTOPIA and NADIR, and 0% would
represent the position of NADIR.
In addition to the achievement level of the efficient point, the tools
also shows the (arithmetic) average achievement level of the
PARETO set.
Basic functionality
1. define the set, add and delete alternatives; with every new
alternative, NADIR and UTOPIA may change;
Page 25 of 95
PM3 Deliverable: A05.1
ESS GmbH
2. define the criteria, select or deselect from consideration; a good
strategy is to start with two criteria only, and add critiera one by one
towards a more complex problem.
3. introduce constraints on any of the criteria to limit the subset of
feasible alternatives to be considered;
4. introduce a reference point different from its default (UTOPIA).
The results is the efficient point, which is the current solution or the best
non-dominated alternative, nearest to the reference point.
Auxiliary analysis
In addition to the graphical display of the set of alternatives with different
colors for the different types of alternatives (NADIR (BLACK); infeasible
(small dark GREY); dominated (larger light GREY), PARETO (GREENE);
current (BLUE); efficient (YELLOW); and REFERENCE (RED), the tools
shows either
 a summary of the optimization scenario with the import and
configuration interface;
 a tabular summary and editor interface for an individual alternative
selected from the main list of alternatives with the associated
scatter-gram of all alternatives, the current alternative marked;
 a tabular summary of the set with the associated scatter-gram;
 an enlarge scatter-gram;
 the statistics for an individual criterion with a histogram, with the
possibility to select a second criterion for a display of a scatter-gram
and correlation (covariance) estimate.
Criteria
For WaterWare: The set of criteria can, in principle, be defined by the user
interactively: criteria can be deleted from the set, and new ones can be
defined in the Knowledge Base with the KB Editor and added with the
RECONFIGURE button. These will be set to a value of undefined for all
existing alternatives, which the user then can edit.
For AirWare/PBM: the current set of criteria is pre-defined depending on
the air quality model used.
4. Data Management and GIS
 Data Types
AirWare includes a set of basic and composite data types and OBJECTS
for which dedicated import, editing, analysis and display functions are
provided.
The data types include:
1. Descriptors are scalar, symbolic, numeric (with associated
symbolic range definitions and defaults), or text variables that are
managed by the embedded expert system and a dedicated
interactive editing tool. Descriptors are used to define the
parameters for more complex types such as:
2. Time Series which in addition to their numeric value have an
associated explicit time stamp;
Page 26 of 95
PM3 Deliverable: A05.1
ESS GmbH
3. Matrices, which in addition to the numeric value of each cell have a
location (geo-reference) and geometry, including possible vertical
layering, and again may have explicit time stamps.
4. GIS data (geo-referenced) including points, lines, topological
networks, polygons, and rasters as individual map overlays and
composite maps.
5. Hypertext including imagery, used to describe OBJECTS (see
below);
These basic data types are grouped into various composite structures or
OBJECTS: OBJECTS are complex composites of basic and derived data
types; examples for OBJECT classes in AirWare include:
 Monitoring Stations that contain time series data;
 Point Sources that hold emission data;
 Industrial plants, that combine several point and possibly area
sources;
 Area sources of emissions;
 Line sources representing road segments;
 Road networks that provide a topological framework for line
sources;
 Matrix objects, that manage the 2D data sets for the distributed
models;
 Model domains that define areas of interest for simulation;
 Meteorological scenarios, that hold consistent and complete data
sets for the simulation models, bound to a model domain and
derived from monitoring stations.
 Emission scenarios, dynamically derived for a model domain;
 Model scenarios, that compile the above sub-scenarios and can
trigger model and display and analyze their results;
 Object data base
AirWare includes a general purpose OBJECT DATA BASE that manages
geo-referenced or generic objects. A multi-attribute selector/navigator
provides access to classes and instances, and support a range of filtering
and sorting operations.
The objects in AirWare are grouped by CLASS. The list of CLASSES is
open and can be configured and extended easily.
OBJECT CLASSES currently supported are:
 Monitoring stations
 Time series data sets;
 Meteorological scenarios (complete, consistent and synchronized
data sets)
 Emission sources, including
o point sources (boilers/stacks),
o area sources,
o line sources (road segments)
 Emission patterns
 Fleet compositions
Page 27 of 95
PM3 Deliverable: A05.1
ESS GmbH

Road networks (topological structure as used in transportation
models)
 Industrial plants or installations (containers for point sources,
possibly also parallel area sources for fugitive emissions);
 Model domains (geographically defined);
 Overlays and Maps (set of overlays in specific stacking order).
 Matrices (model input related)
Each of these CLASSES may have any number of elements.
Each class has a set of specific attributes, organized in a set of data
structures and associated METHODS for instantiation and display.
Selecting an OBJECT from the listing within a CLASS starts a display and
editing page for the object.
Data structures include
1. HEADER with name and id (mandatory), location, and meta-data
information, common to ALL object classes; The LOCATION of the
object is used, together with a specifications of EXTENT (a radius
around it) to define a VIEPORT to generate a MAP of the
surrounding of the object with the object diaplyed as a SYMBOL in
the center of the MAP.
2. ATTRIBUTE SET that is class specific; they may include:
o Hypertext including imagery for a detailed description;
o Tables of Descriptors, using XPS library functions for editing
and inference;
o Time series of scalars, vectors, or lists;
o Matrices
o References to related objects (children or sibblings);
 Object header
All OBJECTS have a HEADER, that includes their
1. Basic identification and short textual description,
2. META data (user, modification date).
3. geo-reference (and parallel map display of the OBJECT position where
applicable in the main OBJECT display page driven by the OBJECT
TEMPLATE, see also:OBJECT LOCATION);
4. (optional) symbolic geo-reference.
Elements of the HEADER, common to all OBJECT classes, include:
 Object name, class, and a unique ID.
 Reference year (validity of the data).
 Description (free text)
 Owner (user with RW access), creation data, last modification date.
 Help file link (URL), class specific;
 Hypertext link (URL); (embedded display for the description page)
 Geo-reference: lat, long, elevation;
 Symbolic geo-reference: up to four tuples of administrative unit - unit
name, usually: location (case study domain name used for
geographical data filtering), province, community/city.
From this set, any subset may be displayed depending on the layout for
the object class: see also: OBJECT TEMPLATE).
Page 28 of 95
PM3 Deliverable: A05.1

ESS GmbH
Hidden attributes:
o Mapset ID, leads to the selection of the background map shown.
o Reference point, x/y in map coordinates, used for the positioning
of the symbol shown.
o Map extent around the reference point.
o Display symbol or style (color, line style and width) reference
o Type specific GIS coupling: area (polygon) or line segment for
area and line sources, respectively.
 Object attributs
Attribute sets consist of
1. Descriptors (which can be: variable, symbolic, or hybrid); their layout,
format and editor interface is controlled by the class-specific
TEMPLATE file and the variable (Descriptor) definition in the
Knowledge Base.
2. Structures: these include more complex data structures with specific
constraints (like summing up to 100% ...) such as time series, tables of
different dimensionality, and matrices. Their layout is defined by a set
of WIDGETS in or available to the PARSER cgi that is triggered by the
respective TEMPLATE tags. or escape sequences. The TEMPLATE is
responsible for the overall positioning of the widget.
o Time series (references);
o Temporal scaling factors (monthly, daily, hourly);
o Contact addresses, contact persons;
o Fleet composition;
o VOC speciation (ISIC industry class specific);
o Link lists to other, related objects.
3. Methods for the attribute instantiation in a dynamic context (scenario,
absolute or simulation/scenario time)
 Object TEMPLATES
Individual objects are displayed using HTML-style TEMPLATES that
conreol layout and compile content (the attributes).
The TEMPLATE to use for a given OBJECT class is defined in the data
base RBO, table: CLASS.
The TEMPLATE files themselves are located in the directory:
/var/www/html/templates
and service/application specific sub-directories, including:
aermod
boiler
emission_pattern mapserver
monitoring
receptor_area
windf
airmodels
building_points domain
emission_report metamatrix
small_stack
timeseries
animations
camx
emission
line_source
meteo
pop
traffic
area_source camx_result
emission_factor
map
mm5
powerplant
The TEMPLATE is a standard HTML file; it is parsed before display, and
specific escape sequences or tags to add dynamic content into the HTML
file include:
Page 29 of 95
PM3 Deliverable: A05.1
ESS GmbH
1. \include(template/standard_head.html) including any common files
(in this case, for the common object header);
2. \name(DESCRIPTOR), which displays the display name: of the
Descriptor DESCRIPTOR;
3. \ask(DESCRIPTOR), which will display the current value (including
undefined) for the Descriptor as an active link in blue, which will
trigger the editing dialog.
4. \value(DESCRIPTOR), which will display the current
value (including undefined) for the Descriptor as a read-only string;
5. \unit(DESCRIPTOR), will display the unit of the Descriptor
DESCRIPTOR.
 Object location
Most OBJECTS are located in space can be geo-referenced
In AirWare, the embedded GIS and MapServer provide background maps
for the display of object location. The default background map is
associated with the LOCATION of the application or project.
For the display of a each object in space, the following data are used:
1. a reference point that defines the position;
2. a MAP ID that defines the map to be used (default associated with
the project/application through the concept of LOCATION);
3. an EXTENT that defines a radius around the object that should be
shown on the MAP; together with the reference point (x and y
coordinates) this defines a VIEWPORT to use for the object/map
display;
4. A SYMBOL ID that defines HOW the object should be shown on the
map (default symbol: a red circle).
Changing Position/Viewport and MAP ID
To locate an OBJECT on the map, change the map extent to be shown
around it, and compose/choose the background maps to be shown, and
EDIT button at the OBJECT DATA BASE level map display leads to the
COMPOSER.
There the current (default) map with the current (OBJECT specific) location
and extent are shown.
The user can:
1. move the OBJECT position by clicking on any point on the map; the
map will be redrawn with the new position in the center, the same
background and extent subject to the constraint that the background
map will cover the entire map window.
2. change the extend (FOCUS) around the OBJECT (subject to the
constraint above; if the extent chosen exceeds the size of the
underlying map given the current location/viewport, the object
location will be shifted out of the center).
3. select an alternative background map by name from the pull-down
menu; if no appropriate map is available,
4. compose a new background by selecting an alternative set of
overlays at the MapComposer level and press the GENERATE
Page 30 of 95
PM3 Deliverable: A05.1
ESS GmbH
BUTTON; the new map will then be available in the MapCatalog and
can now be selected from the above positioning pull-down menu.
Upon RETURNING to the original OBJECT DISPLAY page, the
(new) MAP ID, position, and extent around the OBJECT will be
stored with the OBJECT and will affect the OBJECT map display.
 Model Domains
All simulation model runs or scenarios are implemented for a certain area
of interest, location and extent, or model domain. Also, from the master
emission inventory, domain specific inventories can be generated.
Domains are OBJECTS that can be selected from the OBJECT DATA
BASE hierarchical selection/navigator tools.
A domain usually will cover a specific area of interest such as a city,
industrial district, part of the transportation system, the area around a
major emission source like a thermal power station or major industry, or
the entire region of an application. The domain definition provides a shared
data structure to ensure the consistency between
1. all the geo-referenced data such as the emission sources,
2. a meteorological scenario,
3. model grid and sub-grids and thus geo-referenced output,
4. results post-processing such as exposure analysis.
A Domain is a named object with the standard object MEATA
DATA header.
Due its multiple use by different models, scenarios, and also as embedded
grid in CAMx, domains are READ ONLY and can only be modified by
ADMIN group user. The attributes of a domain include:
1. Object META DATA including name and description;
2. Symbolic georeference (data driven set of administrative or
geographical concepts such as region, province, community)
3. Location and (rectangular) extent, defined by:
o Reference point X and Y (in the center of the domain) used
to locate the domain; this is shown as the usual default
OBJECT location indicator as a circle on the map, it can
edited by ADMIN users, also used to locate the object on the
map interactively;
o Origin X and Y: map coordinates of the origin (lower left
corner) of the domain, computed from reference and extent,
read only;
o Aspect ratio (symbolic selection between square 1:1 and
landscape 4:3);
o East-West extent or width in km
o North-South extent in km (computed, read only)
o area covered (in km², computed, read only)
o the upper right corner is computed and stored/accessible in
the data base, but not displayed.
o Map margin (in % of the domain extent, generates a margin
around the domain rectangle).
Page 31 of 95
PM3 Deliverable: A05.1
ESS GmbH
4. Background map: MAP ID for the background map (hidden), the
map can be selected interactively through the MapCatalog.
The Domain is described by these attributes in tabular format together with
the background map indicating the extent with the rectangular domain
boundaries drawn on the background map.
For a new domain, the ADMIN user can specify:
 name and description;
 select the aspect ratio and horizontal extent;
 select the map window size in % of the domain width;
 define the location of the reference point i(center of the domain);
 select an alternative background map from the MapCatalog.;
Nested grid constraints
CAMx uses a nested grid scheme to calculate results for sub-domain at a
higher resolution. To make that possible, a number of constraints apply for
domain that are to be used as sub-domains with CAMx. See
also: Configuration of Scheduled Scenarios.
1. The domain must be square;
2. The corner point must be positioned on a 3km master grid node;
3. The domain size (extent) must be modulo 3.
 Matrix Objects
Spatially distributed models require spatially distributed input data, and
produce spatially distributed results.
In AirWare, all model related spatial data use the same, common data
format and are described as matrix OBJECT.
Matrix objects are read-only, they are generated by mode pre-processors
or are model generated output. Matrix objects (which can contain one or
mode 2D matrix) are selected from a standard selector/navigator, selection
properties rea:
 matrix name
 model domain
 parameter (Descriptor)
 generation method
 generation date
 N of layers
 N of time steps
A Matrix is a named object with the standard object MEATA DATA header.
The attributes of a matrix include:
 Object META DATA including name and description;
 Class specific attributes include:
o generation method (symbolic name of the pre-processor or
model)
o the model domain that provides location references; these
include:
 map coordinates of the origin (lower left corner);
 aspect ratio (symbolic selection between square 1:1
and landscape 4:3);
 East-West extent in km
Page 32 of 95
PM3 Deliverable: A05.1
ESS GmbH


North-South extent in km (computed, read only)
area covered (in km², computed, read only).
o Matrix geometry:
 cell size (only square cells are supported;
 number of cells in X and Y directions.
o Parameter (based on a Descriptor, the parameter is either
scalar or a 2D vector in case of wind fields)
o number of layers
o time stamp of first (set of) matrices
o number of time steps
 Statistics:
average, minimum, maximum, non-zero average, histogram (10
classes, color coded with automatically generated rainbow colors
also used for 2D display).
Please note: a complex system of matrix analysis/comparison similar to the
time series analysis is in preparation for future releases.
Matrix display
When a matrix is selected from the selector/navigator, its default display
shows the first instance( lowest layer, first time step) color coded
corresponding to the histogram. Selector/buttons (tape deck) support
switching between layers and time steps. The matrix can also be shown in
a full-size pop-up window for more detail.
Matrix import
An interactive import function supports the upload of matrices prepared by
external tools such as a spreadsheet.
Matrix export
Model results matrices can be exported in CSV format with a 14
record header for META DATA.
 Matrix Export
All spatially distributed models in AirWare produce their output as MatrixObjects.
These matrices can be exported in CSV format using the export button at
the details model results display page.
The exported CSV file starts with 16 lines of header information with the META data
describing the exported matrix:
Source
Resolution
Substance
Unit
Start time
Duration
Interval
Time steps
Layers
Components
Rows
Columns
CAMx
1000 m
NOx
ug/m3
2008-08-25 19:00
24 hours
1 hour
24
1
1
300
300
Page 33 of 95
PM3 Deliverable: A05.1
ESS GmbH
Domain name
NW corner of domain (UTM coordinates)
Domain extent in x direction
Domain extent in y direction
TEHRAN
x=380000 m - y=4105000 m
300000 m
300000 m
This is followed by the data (comma separated floats) for each grid cell.
For each time step the matrix is given in the following format (time steps are
separated by an empty line):
 Data start at the north-west corner of the domain.
 Each line corresponds to a row of the matrix containing values ordered from
west to east:
o The first value corresponds to the north-west corner of the domain,
o The last value of the first line corresponds to the north-east corner.
o The first value of the last line corresponds to south-west corner of the
domain,
o The last value of the last line corresponds to the south-east corner.
 Population data
Population data are an OBJECT CLASS, associated with polygons similar
to area sources. These data (usually from a census with associated census
tracks as the spatial units) form the basis for the pollution data raster used
for the calculation of population exposure.
The primary entry from the object class selection leads to a ranking list of
population objects, i.e., populated areas or census tracts.
The listing shows the object names, an associated symbolics address
(e.g., district) and the absolute and total population numbers.
The associated map can be zoomed, and queried: to start the info mode,
click the info button under the map, the click on any one of the color coded
areas on the map. The name of the object will be shown. To return to the
zooming mode, click the zoom button.
 Land use data
Land use surface characteristics and associated roughness (length) as well
as UV albedo are important inputs for CAMx.
CAMx landuse categories and the default surface roughness values (m)
assigned to each category by season. Winter is defined for conditions
where there is snow present; winter months with no snow are assigned to
the Fall category. Roughness for water is taken as the maximum of the
baseline value given in the table, and the function Z0 = 0.000002 w**2.5 ,
where w is surface wind speed (m/s).
The listed UV albedo values can be used to assign UV albedo from
landuse data in preparing the albedo/haze/ozone (AHO) input file.
Surface Roughness (in meters) and UV Albedo by land cover and
season
N
Land cover
1 Urban
2 Agricultural
Spring Summer
1.00
1.00
0.03
0.20
Fall
Winter UV Albedo
1.00
1.00
0.08
0.05
0.01
0.05
Page 34 of 95
PM3 Deliverable: A05.1
3 Rangeland
4 Deciduous forest
5 Coniferous forest, wetland
6 Mixed forest
7 Water
8 Barren land
9 Non-forested wetlands
10 Mixed agricultural/range
11 Rocky (with low shrubs)
ESS GmbH
0.05
1.00
1.30
1.15
0.0001
0.002
0.20
0.04
0.30
0.10
0.01
0.001
1.30
0.80
0.50
1.30
1.30
1.30
1.30
1.05
0.90
0.0001 0.0001 0.0001
0.002
0.002
0.002
0.20
0.20
0.05
0.15
0.03
0.006
0.30
0.30
0.15
0.05
0.05
0.05
0.05
0.04
0.08
0.05
0.05
0.05
 Embedded GIS
AirWare uses an embedded MapServer that manages a MapSet together
with a
 MapComposer configuration tool to combine several individual
topical overlays, over an optional raster background map.
 MapCatalog overview and map selection to associate named maps
with OBJECTS and OBJECT CLASSES including model domains,
o MapImport and overlay description tool,
to provide geographical background and orientation/reference for its data
(OBJECTS and spatial model inputs and results.
In addition, a range of geo-referenced data that can be viewed as topical
maps are used as model inputs.
For the user, the MapServer is transparently embedded with all georeferenced OBJECT CLASSES, model scenarios and their underlying
domains, model inputs data sets, and model results.
Within a given application, a MapSet of available overlays is defined.
A MapImport function is associated with the display and definition of the
raw map layers or overlays, accessible from the central MapComposer.
GIS formats supported for import include
 ArcView Shape files (topical maps, one topic at a time)
 Raster formats (TIFF).
User interaction
GIS related operations the user can perform include:
1. Map Composition from the set of individual topical overlays,
including the definition of optional ViewPorts for individual named
maps;
2. Map Import, interactive import of maps (raster format) or (single
topic) topical map overlays from client PCs (please note that
composite maps with several topics can be constructed with the
Map Composition tool)
3. Map Catalog Browsing and selection of maps as the background for
individual OBJECTS and OBJECT CLASSES including model
domains;
4. Background Map Selection: interactive selection of background
maps for
1. Model domains
2. all georeferenced OBJECT classes
3. all georeferenced OBJECTS (instances).
Page 35 of 95
PM3 Deliverable: A05.1
ESS GmbH
5. Object Geometry: Import for Area Sources and Receptor Area
polygon boundaries
(single feature of a single-topic ArcView Shape file);
6. Object Geometry: Import for Line Sources polylines or line
segments (single topic ArcView Shape file).
7. Object Location, definition of the location of point sources (or object
with a reference point) directly on the map;
 MAP Set
Maps are organized in an application specific MapSet.
The MapSet defines:
1. the common projection for all maps
2. the common coordinate system for all maps
3. the bounding box for the application, which should coincide with
the main
model domain (AirWare) that covers the area of interest;
4. the location in the file system of the raw map data
5. the default map shown in the MapCompser and MapSet
interface.
The latter (the default map) can be changed by the user; all other
features must be consistently configured by the administrator in the
data base:
Data bases, tables, and files
DATA BASES:
mapc, gis
DATA BASE TABLES:mapset_data, overlay_data
DATA PATH:
user configured in the MapSet main page, default: /gis
 MAP Composition
AirWare uses an embedded GIS or MapServer together with a Map
Catalog and configuration tool (Map Composer) to provide geographical
background for its data (OBJECTS and spatial model inputs and
results.
The Map Composer offers a choice of
 one background overlay (raster or polygon feature, opaque, full
coverage, from a list of all applicable background overlays: only
one can be chosen at a time) and
 all applicable foreground overlays line features, transparent
polygon or raster, partial coverages, any number can be
combined) together with
 a central point of interest (X and Y coordinates) and
 extent (in meters, interpreted as a radius around the central point
of interest)
from which a named MAP (configuration) can be generated for
1. preview (temporary) or
2. permanent storage
Page 36 of 95
PM3 Deliverable: A05.1
ESS GmbH
and retrieval/display from the Map Catalog for selection as default
background for all geo-referenced OBJECTS.
The list of overlays leads to a description and individual preview of any
of the overlays listed by the Map Composer.
Navigation BUTTONS
1. The BUTTON MAPSET leads to a description of the
applications MapSett and its parameters (projection, coordinate
system, bounding box) that are needed for Map Import. Also
supports the definition of a default map for the MapSet, and the
directory of all map data.
2. The BUTTON OVERLAYS leads to a listing of all available map
layers or overlays, and offers the possibility to ADD/IMPORT a
new overlay with the Map Import dialog.
3. The BUTTON MAPS leads to the Map Catalog.
4. The BUTTON COMPOSITION leads back to the main Map
Composer level from any its secondary pages.
If a MAP (configuration) is to be stored permanently in the Map
Catalog, the user can specify a name before generating a permanent
configuration. Other descriptive META data (creation date, user/owner,
overlays used, projection, extent) are stored automatically with the new
map. At the level of the Map Catalog, a more detailed (free) textual
description of the map selection can also be entered.
 MAP Catalog
The MapCatalog is accessible from the Map Composer level.
The MapCatalog shows all MAP (configurations) that have been defined
and stored as permanent as a thumbnail images together with their user
defined short name and ID.
Clicking on any one of the small MAP images lead to a pop-up page that
shows that MAP (in its original viewport) in a larger window together with
its META DATA (name, description, owner, creation and modification
dates) as well as the list of overlays it was generated from, the parameter
of its (native) viewport (center coordinates and extent defined as radius in
meters) and the coordinates of the bounding box of the map.
The maps shown can be selected as background map for
1. any geo-referenced OBJECT CALSS
2. any geo-referenced OBJECT
3. model domains and thus modeling results
A SELECT button allows to select the MAP as the (new) background for an
OBJECT that has called the MapComposer from the map display EDIT
button at the OBJECT CLASS selection or OBJECT display level.
From the set of maps shown in the MapCatalog, any individual maps can
be selected and shown individually, in a larger window together with its
legend,
The SELECT operation changes the default (class specific) background
map for the for an OBJECT or OBJECT CLASS, but retains its latest
viewport (center point and extent).
Page 37 of 95
PM3 Deliverable: A05.1
ESS GmbH
 MAP Import
The import of maps or map layers/overlays is available as an option for
 NEW overlays, triggered from the MapComposer tool through the
button: OVERLAYS.
 EXISTING overlays to replace their current data;
 specific OBJECT geometry representation:
1. line sources (road segments)
2. area sources
3. receptor areas.
For an existing overlay from the list, an alternative map/overlay data set
can be imported; they must conform to the parameters defined for the
current MapSet in terms of projection, coordinate system, and extent (at
least partly within the overall application bounding box).
For a NEW map, map layer or overlay, the IMPORT button will trigger a
dialog pop-up window where the users specifies (or select from browsing
the local file system) a file name for the map data to be imported from the
local client PC, browsing the local file system.
GIS formats supported
Vector formats
 ArcView shape files: import supported for lines and polygons one
topic, one color, one line width at a time attribute table is ignored
display color and line width can be selected
 Arc/Info export files: supported are single-precision,
uncompressed files containing lines and/or polygons one topic, one
color, one line width at a time attribute table is ignored display color
and line width can be selected
Raster formats
 tiff: import supports pseudo and true color uncompressed and
compressed images. Supported compression algorithms, are:
CCITT 1D Huffman compression , CCITT Group 3 Facsimile
compression , CCITT Group 4 Facsimile compression , Lempel-Ziv
& Welch compression , baseline JPEG compression , word-aligned
1D Huffman compression , and PackBits compression . In addition,
several nonstandard compression algorithms are supported: the 4bit compression algorithm used by the ThunderScan program ,
NeXT 2-bit compression algorithm, an experimental LZ- style
algorithm known as Deflate , and an experimental CIE LogLuv
compression scheme designed for images with high dynamic range.
Import Dialog
1. To import a map data set from the Overlay Description page, the
user can first select the type: vector (the default) or raster. If a
reaster (TIFF) is to be imported, the origin and extent have to be
defined first, else and error message will be generated. In case of a
vector coverage, the origin and extent will be determined
automatically from the data imported.
For a vector coverage, the type (shapefile or e00) will be determined
automatically. The user can specify in the dialog box:
Page 38 of 95
PM3 Deliverable: A05.1
ESS GmbH
line/border color (from an interactive color selector), default
border color is BLACK;
o line thickness in pixels;
o Fill attributes: the default for the filling of polygons is NONE;
 fill color (from the the color selector)
 fill style (opaque or transparent)
To generate a filled polygon without visible border, select the same
color for border and opaque filling.
The file to be imported is identified by its local (client) path and file
name, or through a local interactive file browser and selection.
A successful import is indicated by the imported data being
displayed on the Overlay Description page.
2. To import a map data set for an area sources or receptor area
boundary, a similar dialog is used. For these polygons, only vector
data are applicable, and the user can specify a feature ID to extract
a single feature (element) from a larger file. The display style is
defined by the system to make all elements of this class appear in
the same style.
Error handling
The following error conditions are checked and reported; Warnings are
only reported in the system log file:
 For rasters:
1. bounding box partially outside the MapSet bounding box
(warning);
2. bounding box completely outside the MapSet bounding box
(error, abort, file rejected);
3. pixels are not square (error, file rejected).
 For vectors:
1. coordinates partially outside MapSet bounding box (warning);
2. coordinates completely outside MapSet bounding box (error,
abort, file rejected).
3. wrong (not recognized) format and file structure (error, abort,
file rejected);
The Map Data Import and any error encountered is also logged
automatically in the System log.
o
5. Monitoring data
 Monitoring stations
Monitoring Stations describe the physical monitoring equipment, its
location, and lead to the actual monitoring time series. This includes
meteorological variables, emissions, and ambient concentrations. A special
case are traffic observations in real time.
A Monitoring Station Object is defined by its HEADER including a hypertext
page describing it, and the list selector for all the time series collected
there.
Other attributes are grouped into topics that are shown in pop-up windows
triggered by (optional) BUTTONS. They include:
Page 39 of 95
PM3 Deliverable: A05.1
ESS GmbH
1. Links to other objects
2. Administrative information
3. Technical information
Time series (parameter) selection
Under the Object HEADER, the page displays a selector of available
time series, organized in terms of their parameters. The selector (full page
width) shows:
1. Parameter name
2. Last measurement (end date)
3. First measurement (start date)
4. Time step (symbolic, i.e., a STRING that can also indicate irregular
sampling or bi-monthly);
5. unit
6. Number of observations
7. Coverage in % (100-% of missing values).
The MONITORING TIME SERIES themselves consist of a HEADER, and
the actual time-stamped values. TIME SERIES META DATA compiled
during the import process, include:
1. USER: (determined automatically)
2. NAME: a descriptive (short) name for the time series that will be
used in the selector list;
3. TYPE: one of historical or real-time.
4. DESCRIPTION: a free text field that defines, e.g., data source (such
as the original monitoring station/sensor), method for patching,
assumptions (for synthetic time series);
5. LOCATION: defines the geographical scope (on import, this
selected by the user from a list of pre-defined options, or generated
automatically from the application's CONFIGURATION information if
the data base covers a single location);
6. VARIABLE: (or parameter selected from a pre-defined list) with its
associated UNIT;
7. START DATE: will be determined automatically from the time
stamps of the imported data.
8. TIME STEP: symbolic label for the time stamp (e.g., hourly, daily,
monthly, irregular); internally, we can derive the time tep in minutes
from the delta between the obervations;
9. GENERATED: date of original imnport;
10. LAST MODIFIED: date of last update, mainly for real-time data;
Attribute set
Other descriptive information is organized in two linked pages:
1. Administrative information, describing the owner or operator,
contact addresses etc.
2. Technical information, describing the actual monitoring equipment
used, telemetry, etc.
Data Items
Administrative information
 Address block for the institution and a contact person with an
optional second (local, on site) address, derived from the
Page 40 of 95
PM3 Deliverable: A05.1
ESS GmbH
INSTITUTIONS data base, relational contact/address data base
TABLE
List of sensors/methods used for the parameters monitored; for each
sensor, the following information can be stored and displayed in a
separate pop-up window:
 Sensor name
 Hypertext link (URL) for descriptive text, optional imagery.
 Manufacturer
 Model name
 Sensor technology
 Year of installation
 Date of last maintenance/calibration
 Monitoring sensors
Monitoring Stations have one or more sensors or instruments to collect
their data. For every station, a list of sensors or instruments is available,
leading to a description of the respective objects;
New sensor/instruments can be added (selected) from a generic shared
set, which is accessible from the main Object Data Base level and can be
edited there.
 Locating stations
AirWare provides an option to identify appropriate locations for monitoring
stations in the vicinity of major emission sources. It identifies location
where the maximum annual average concentration over populated or
specifically designated receptor areas can be expected, given annual
meteorological data.
This is based on identifying the points a maximum annual average
concentrations using AERMOD for either buoyany gases or particles,
based on a complete set of 8640 hourly runs for a calender year. Please
note that the annual model run requires a COMPLETE data set iof wind
and temperature data. Small gaps of up to four hours can be patched
automatically. For larger gaps, the model run will be aborted and a
corresponding warning issued.
The concentration field from a successful run is analysed for populated
areas (a thematic map layer) and the set of user defined receptor areas in
a user defined domain around the source. Up to three location for
monitoring stations (with a user defined exclusion radius) are defined with
decreasing annual average concentration.
The user defined parameters include:
1. the emission source (point sources only);
2. the pollutant to be considered (SO2, NOx, PM10);
3. year and meteo station: please note that the annual run requires a
complete data set; gaps of up to four hours can be patched
automatically, larger gaps will be identified by a pre-processor
(checker) that will abort the scenario.
Page 41 of 95
PM3 Deliverable: A05.1
ESS GmbH
4. the model domain and grid resolution; please note that extreme
values will lead to a model run of several hours !
5. exclusion radium around any solution/location found: the model
identifies the first location (maximum annual average concentration
over populated area or receptor area); around this location, an area
defined by the exclusion radius will be ignored for the location of the
second and third locations.
The results are shown as color coded crosses at the proposed locations,
together with a tabular summary of the annual average concentrations at
these locations.
Details of the results can be viewed with the options to create isolines of
concentration and zoom into the domain.
 Time series analysis
AirWare uses two a common data base representation for OBSERVATION
and monitoring time series data in a simple format:
TIMESTAMP VALUE;
1. for meteorological observations;
2. for air quality observations
3. for emission data (large stacks//boilers).
These data can be historical or real-time with continuous, regular updates.
Time series data are associated either with:
1. A monitoring station object;
2. A stack/boiler object in case of emission data.
The available time series can be sorted by variable as well as starting date.
Buttons are available (view, select) to select a given time series, or to first
view it, which shows the time series graph together with basic descriptive
statistics.
For monitoring stations that include WIND DIRECTION A special analysis
of the wind rose is provided by the ANALYSIS button: the wind rose is
generated for 22.5 degree slots.
Patching missing values
Gaps of up to three missing values can be patched with
the PATCH button: the values inserted are either based on linear
interpolation, or, in the case of wind direction, simplex extrapolation of the
last legal value. A small dialog (pop-up)wind reports on the number of
successful or failed patch attempts.
Patched values are marked and shown in GREY in the time series display.
To drop all patched value, a reset button can be used. All patched values
will again be replaced by a missing value code (-999).
 Adding a time series
Time series data can be imported
 manually, bulk, for larger data sets;
 automatically, from any local or remote data base, in real-time;
Page 42 of 95
PM3 Deliverable: A05.1
ESS GmbH
respective node input data. An interactive tool to IMPORT/UPLOAD additional
time series by remote users to this data base and selector list is available with
the IMPORT button at the TS selector page.
Bulk import, manual:
The import dialog for bulk import consists of the following steps:
 an appropriate time series file is prepared on the local client PC, e.g.,
using a spreadsheet, then exported in CSV format.
The following file format is supported:

TIMESTAMP,VALUE

e.g:
2004-01-01,111.0

2004-01-02 10:00,112.0

2004-01-03 11:00,111.5

....
TIMESTAMP is in the format YYYYMMDD hh:mm
 in the web dialog, the descriptive data (meta data) are entered;
They include:
o USER: (determined automatically)
o NAME: a descriptive (short) name for the time series that will be
used in the selector list;
o DESCRIPTION: a free text field that defines, e.g., data source
(such as the original monitoring station), method for patching,
assumptions (for synthetic time series),
o LOCATION: (selected by the user from a list of pre-defined
options);
o VARIABLE: (or parameter selected from a pre-defined list) with
its associated UNIT;
o SOURCE FILE: local (on the client PC) file name and path from
where the data set will be taken from; a browser option supports
an interactive definition of the source file;
o START DATE: will be determined automatically from the time
stamps of the imported data.
o TIME STEP: will be determined automatically from the time
stamps of the imported data where possible.
 after the IMPORT button was pressed, the file will be uploaded to the
WaterWare data base location on the WaterWare server;
Preview: after import, an analysis program will check it, and provide the
user with a graphical display, plus descriptive statistics and an echo of
all data and information provided.
Incremental manual data import
As an additional function, data sets can be added manually to an existing time
series, e.g., for completing or correcting existing data sets. The user can
decide whether existing values should be overwritten (replaced with the new
import), or only missing values patched.
Real-time automatic data import
For a given monitoring station, a real-time connection to a (remote or local)
data base that holds real-time monitoring data can be defined.
Page 43 of 95
PM3 Deliverable: A05.1
ESS GmbH
For the real-time data acquisition, parameters include:
 User name and password for data base access;
 IP address of the data base server;
 Data base type (ORACLE, MySQL, PostgreSQL);
 Data base and Table names; these will be used for all parameters that
do not have specific, individual data base and table names defined;
 Date and time column names;
 Time shift: delay in minutes after the full hour, when the data acquisition
will be attempted;
 Status to switch the import on or off.
followed with a list of parameters (time series) defined for this
monitoring stations, where for each parameter the data base, table,
column name as well as a condition (SQL WHERE clause) can be
defined.
 Manual TS data Import
AirWare uses a common data base representation for OBSERVATION and
monitoring time series data in a simple format:
TIMESTAMP VALUE
1. for meteorological observations, including simulated observations
populated from model results;
2. for air quality observations, including simulated observation
populated from model results;
3. for emission data (large stacks//boilers).
Data import
To import data three different mechanisms are supported:
1. automatic import from a data base at scheduled intervalls;
2. manual import of a complete time series (historical data)
(at the level of the monitoring station object)
3. manual updated of an existing time series
(at the level of the individual time series)
In both cases a button in the header of the page is offered:
 import TS at the level of the monitoring station, leading a
pop-up window for the specific import dialog to add a completely
new time-series to a station.
 import at the level of the time series, again leading to a
pop-up window with the specific import dialog to update existing
time series.
 Automatic TS data Import
The default resolution of Time Series in AirWare is hourly, all time steps
are defined in minutes.
Import Dialog
Time series data can be imported
 manually, bulk, for larger data sets or to create a new time series;
Page 44 of 95
PM3 Deliverable: A05.1


ESS GmbH
manually, for small additions/corrections to add or replace selected
data;
automatically, from any local or remote data base, in real-time.
Real-time data import:
For a given monitoring station, a real-time connection to a (remote or local)
data base that holds real-time monitoring data can be defined.
For the real-time data acquisition, parameters include:
 User name and password for data base access;
 IP address of the data base server;
 Data base type (ORACLE, MySQL, PostgreSQL);
 Data base and Table names; these will be used for all parameters
that do not have specific, individual data base and table names
defined;
 Date and time column names;
 Time shift: delay in minutes after the full hour, when the data
acquisition will be attempted;
 Status to switch the import on or off.
followed with a list of parameters (time series) defined for this
monitoring stations, where for each parameter the data base, table,
column name as well as a condition (SQL WHERE clause) can be
defined
 MM5 data import
The following scripts and functions are used to control the download of
meteorological input data and boundary conditions for the MM5 prognostic
meteorological model:
/app/data/gfs/get_gfs.php
/app/data/gfs/ungrib_gfs.sh
/app/data/gfs/wget_fnl.sh
/app/data/gfs/ungrib_fnl.sh
/app/data/pregid/pregrid_gfs.sh
/app/data/pregid/pregrid_fnl.sh
downloads gfs data, scheduled by cron
ungribs gfs data calling pregrid_gfs.sh
downloads (historical) fnl data, on demand
ungribs fnl data calling pregrid_fnl.sh
reads gfs grib data for forecast
reads fnl grib data for historical scenarios
For MM5 model runs, the main path is /app/mm5/mm5_location/:
run_mm5_forecast.sh
run_mm5_fnl.sh
log_import.sh
TERRAIN/TERRAIN_DOMAINx
regridder/create_namelist.sh
regridder/regridder
INTERPF/create_namelist.sh
INTERPF/interpf
MM5/configure.user
MM5/mm5.deck
MM5/Run/cp_bc.sh
MM5/Run/mm5.exe
runs daily forecast
runs historical scenario
imports log message to GLOBAL_LOG.LOG
mm5 grid files
creates namelist.input file for regridder
regrids input data to mm5 horizontal grid
creates namelist.input file for interpf
interpolates input data to mm5 vertical grid
mm5 model configuration file
mm5 setup file for model run, creates namelist file
mmlif
copies input files to model run directory
mm5 model executable
Page 45 of 95
PM3 Deliverable: A05.1
ESS GmbH
MM5/Run/GENPARM.TBL
MM5/Run/LANDUSE.TBL
MM5/Run/SOILPARM.TBL
MM5/Run/VEGPARM.TBL
MM5toCAMx/create_namelist.sh
MM5toCAMx/run_mm5camx
table with general parameters
table with land use categories
table with soil parameters
table with vegetation parameters)
creates mm5camx.job input file for mm5camx.linux
runs mm5camx.linux and saves data in
MM5_OUTPUT
MM5_IMPORT/scenario_import.sh imports general mm5 scenario info
6. Emission inventories and emission modeling
 Emission Inventories
AirWare maintains a set of emission inventories as OBJECT CLASSES in
the OBJECT DATA BASE with a class specific selector/navigator for Area
SOurces, Boilers, Industrial Plants, LIne SOurces, and Small Stacks
(default alphabetical order in the OBJECT ClassLIst;
Alternatively, a direct entry point Emission DB leads to a list of <="" b=""
style="font-family: Arial, sans-serif; ">organized by domains. The selector
of emission inventories shows their name, as well as the number of boilers,
small stacks, area sources, and line sources within each of the domains.
The listing can be sorted by any of these attributes.
Detailed data structures and data base representations are described in :
 Point Source data (everything with an explicit stack);
 Area Source data (emission without explicit stack data)
 Line Source data (traffic generated).
The basic groups of emission sources are:
1. Point sources (raised (industrial) stacks;
o Industrial plants are groups of stacks and associated
boilers, representing industrial installations that also hold
administrative data such as street address, contact person,
permits, etc.
o Boilers and stacks which are larger boilers and associated
stacks which are :
 normally grouped into plants possibly together with
one or more area sources;
 usually have time-series of emission and related data
such as flue gas temperature and velocity or fuel
consumption.
o Small stacks and their associated boilers, representing
smaller sources for which no time series data (but optional
description of a generic emission pattern) exist. The may be
associated with a primary source other than an industrial
boiler (including private central heating plants), e.g., from a
restaurant or smaller commercial entity. They can have an
(optional) street/contact address of an enterprise attached.
2. Area sources (commercial and domestic areas or associated with
industrial plants);
Page 46 of 95
PM3 Deliverable: A05.1
ESS GmbH
3. Line sources (road segments with traffic data, optionally grouped
into traffic networks that provide an interface to transportation
models).
4. Complex sources are composite area sources that can also
include explicit individual point, line and area sources, that are
automatically deducted from the complex source emission totals. An
application example are the European level EMEP emission data.
Access, display, editing, data import
The various emission classes are accessed from:
1. Main Menu: directly through the OBJECT DATA BASE navigation;
2. Main Menu: through dedicated summary pages for the three main
classes, points, areas, and lines;
3. Model Menus: through the model specific, dynamically generated
emission scenarios which are part of the various model scenarios.
Each of these three CLASSES has its own start page (directly accessible
from the main application menu) with summary information describing the
current status of the emission inventory. This overview and statistical
summary is provided for the default domain of the application, and has a
domain selector to provide any local sub-set analysis, e.g., on a province
or city level.
The following emission inventory summary pages are available:
1. Inventory Summary (summary over all source classes);
2. Point Source Interface (everything with an explicit stack);
3. Area Source Interface (emission without explicit stack data)
4. Line Source Interface (traffic generated).
5. Complex Source Interface (composite area sources).
Data structures and attributes
All emission sources are OBJECTs and have the standard OBJECT META
DATA associated that include:
 name, short description, owner, creation and last modification date;
 geo-reference: symbolic and (type specific, reference point,
polygon, line segment);
 Hypertext description (with import function);
 Type specific attributes (Descriptors);
 source specific temporal scaling factors (for month, day, hour)
 a TABLE of substance specific average emission values;
 optional one or more substance (or parameter) specific temporal
scaling factors;
 optional time series of emission values (include emission
temperature and speed for industrial stacks), and related
parameters, with import function.
The emission inventory covers the territory of the entire domain or location,
sources for specific case study (model) areas are extracted dynamically.
Modification of emission parameters for scenario analysis are achieved
with an additive and multiplicative term that can be specified in the model
related scenarios for each average emission value.
Page 47 of 95
PM3 Deliverable: A05.1
ESS GmbH
 Point sources
Point sources represent raised (industrial) stacks and the associated
boilers. There are two closely related OBJECT CLASSES:
1. Boilers and Stacks that are always linked to an industrial plant, and
can have (optional) time-series of emission data;
2. Small Stacks, that are "singular", have no plant grouping but may have
an optional contact address of a plant, and have a single, average
emission value for each pollutant substance.
The detailed data structure and corresponding Data Base
Representation depends on the set of Descriptors that the OBJECT specific
HTML TEMPLATE includes.
Point sources or stacks/boilers can optionally be grouped into
 plants representing industrial installation that also hold administrative
data such as street address, contact person, ISIC classification, start of
operations, etc.
All point sources are OBJECTs and have the standard OBJECT META DATA
associated that include:
 Name, short description, owner, creation and last modification date;
 Geo-reference: symbolic and (type specific, reference point at the base
of the stack);
 Hypertext description (with import function);
 Type specific attributes (Descriptors):
o Stack height, diameter
o Industry classification (ISIC)
o Boiler type, power rating, primary fuel, average fuel
consumption.
o Other optional attributes (open list of Descriptor-Value pairs that
can be configured with the KB editor and the OBJECT
TEMPLATEs). Examples are operating hours, start time and
stop time of operations.
 Source specific temporal scaling factors (for month, day, hour)
 a TABLE of substance specific average (default) emission values;
 Industrial plants
Industrial plants or installations are primarily administrative objects that can
group one or more point and area sources together and provide
administrative information such as contact addresses of owner/operator,
ISIC classification, permits, etc. and also provide direct access to any
monitoring stations and data associated with the installation.
A plant OBJECT is described by
1. a standard OBJECT HEADER
2. A hypertext description including optional imagery;
3. A map (supplied by the MapServer) with the location of the object
indicated; please note that the map include an EDIT button that
leads the the MapComposer which can be used to configure the
map background for the object.
Page 48 of 95
PM3 Deliverable: A05.1
ESS GmbH
4. Type specific attributes that can include (PLEASE NOTE: the list of
descriptive attributes is open !) the
o type of operation,
o technology describing the primary combustion process,
o pollution control technology
o ISIC classification code,
o year of construction or start of operations,
o number of local employees,
o power rating (for TPS),
o primary product
o annual production volume
o primary fuel,
o average annual fuel consumption;
5. A CONTACT ADDRESS link;
6. A TABLE (selector) of
o point sources (boilers/stacks) and
o area sources
with their aggregated emission (average), by substance;
7. An emission pattern for the plant; please note that each point or are
source may have its own, individual patter down to the substance
level. In the absence of any more detailed information, the plant
level pattern is inherited by all sources attributed to that plant;
8. Links to any local monitoring stations.
 Boilers/stacks
1. Boilers or Industrial Stacks are always linked to an industrial plant,
and can have (optional) time-series of emission data;
2. Small Stacks, that are "singular", have no plant grouping but may have
an optional contact address of a plant, and have a single, average
emission value for each pollutant substance.
Point sources or stacks/boilers can optionally be grouped into
 plants representing industrial installation that also hold
administrative data such as street address, contact person, ISIC
classification, start of operations, etc.
All point sources are OBJECTs and have the standard OBJECT META
DATA associated that include:
 Name, short description, owner, creation and last modification date;
 Georeference: symbolic and (type specific, reference point at the
base of the stack);
 Hypertext description (with import function);
 Type specific attributes (Descriptors):
o Stack height, diameter
o Industry classification (ISIC)
o Boiler type, power rating, primary fuel, average fuel
consumption.
o Other optional attributes (open list of Descriptor-Value pairs that
can be configured with the KB editor and the OBJECT
Page 49 of 95
PM3 Deliverable: A05.1




ESS GmbH
TEMPLATEs). Examples are operating hours, start time and
stop time of operations.
Source specific temporal scaling factors (for month, day, hour)
a TABLE of substance specific average (default) emission values;
optional one or more Substance (or parameter) specific temporal
scaling factors;
optional time series of Emission values (for boilers and stacks),
includes flue gas temperature and velocity), and optinal related
parameters such as flue gas volume or fuel consumption, with a
generic time series import function.

Boilers and Industrial Stacks:
Name
Stack height [m]
Stack diameter [m]
Exit temperature [Degree Celsius]
Exit velocity [m/s]
Turbine type [steam, gas, n/a, ...]
Year of construction
Fuel type [gas, heavy oil, ...]
Fuel consumption [kg/year]
Emission Data [g/s]
NO/NO2 ratio
GIS coordinates
Link button: Emission temporal
pattern
Link button: Profile
Link button: documents
average consumption
NOx, NO,NO2, SO2, PM10, PM2.5, CO,VOC
current MapSet projection, UTM or Lat/Long
select or define and link a pattern
pop-up window with hypertext and images
selector listing and search for text documents
(PDF)
Optional: emission time series
 Mobile sources
This simple case is linked to the Mobile Source emission object; a Mobile
Source OBJECT is similar to a point source (small stack), but has three
additional button links in the header:
1. MultiPUFF, the basic scenario interface and results display;
2. Trajectory, that describes the movement of the object in the domain
together with the hips "activity level" (a multiplier for the average
substance-specific emissions) at each location and time step;
3. Run the model scenarios interface and dialog.
Object properties
The Mobile Source is a standard OBJECT with header, META data, and a
set of attributes the user can edit. This includes hypertext and imagery,
a contact button for the administrative data (owner/oprtator address and
contact information), basic description and properties including source
categotry, type, model, size, engine type, rating/horespower, fuel (the list of
attributed is open and controlled by the corresponding Object TEMPLATE),
and the substance specific emission rates for "normal operation".
1. MultiPUFF the model scenario is defined by:
o a model domain, that also defines the meteorological data
source, and spatial resolution;
Page 50 of 95
PM3 Deliverable: A05.1
ESS GmbH
A START TIME (YYYYMMDDHH[mm])
the duration of the model scenario and run (default: 24 hours;
this value is read only, defined with the source trajectory and
the model scenario interface)
o The substance that will be displayed;
o The model results display options, including the basic map
overaly, tape deck, color coding editor diaslog, details page,
animation, and scenario comparison..
2. Trajectory leads to an editor page that supports:
o zooming the domain background map;
o editing the mobile source trajectory and associated activity
levels:
selecting a position on the map (left mouse button) will copy
its coordinated into the list of coordinates left of the map; the
user then defines the TIME STAMP (in minutes of elapsed
time after the STARTTIME, and the associated activity level,
symbolic, or as a numerical multiplier for the average
emission rates. To change emission while the source is at a
fixed location, the coordinates for this location need to be
repeated with a different TIMESTAMP and emission factor.
3. Model Dialog offers several model specific parameters (including
output time step) and starts the model run; depending on the
expected duration of the run (for a single source the model takes
approximately 30 seconds per hour simulated) the model offer
notification of the completion of the run by eMail or optional SMS.
o
o
 Area sources
Area sources represent distributed sources of small stacks and chimneys
and fugitive emissions.
The default Data Structure and data base representation of area sources is
defined by the object header and the open list of attributes defined in the
HTML OBJECT TEMPLATE.
They can represent
 industrial areas or individual enterprises (possible in combination
with one or more point sources and associated with a plant
OBJECT);
 commercial (e.g., gas stations) or residental areas;
 city areas representing diffuse contributions from the transportation
system including both moving and stationary vehicles;
 intersections as "hot spots" of the transportation network if not
covered by appropriate line sources.
All area sources are OBJECTs and have the standard OBJECT META
DATA associated that include:
 name, short description, owner, creation and last modification date
of the data set;
 geo-reference: symbolic, by a reference point (center of the
rectangular bounding box of the polygon, see below);
 OPTIONAL Hypertext description (with import function);
Page 51 of 95
PM3 Deliverable: A05.1

ESS GmbH
Type specific attributes (Descriptors)
o ISIC classification;
o Type (describing a sectoral classification, e.g., commercial,
industrial, residential, public, traffic)
o average height;
o total area;
o primary fuel;
o fuel consumption (kg/day)
o an optional link to a contact address for larger
objects/buildings.
o
DATA SUMMARY:
Area sources
Name
Location: Province
Location: Community
Category (symbolic)
ISIC Code
Year of construction
Source size (symbolic)
Fuel type (symbolic)
Fuel consumption average
[kg/year]
Total Area [km2]
Average Height [m]
Emission Data, average [g/s]
Residential Areas, Airports, Industrial
Park...
begin of operations
very small, small, medium, large, very
large
list of legal values from the knowledge
base
(consistency with GIS polygon
representation ?
NOx, NO,NO2, SO2, PM10, PM2.5, CO,
VOC
GIS Data
 Type specific geometry: georeference (beyond the symbolic
association with a province/county/community include:
o the reference point;
o a circle (defined by the reference point and the radius
derived from the total area;
o an arbitrary polygon.
Polygon: the AREA OBJECT has an attribute that defines the
NUMBER OF POINTS in the polygon (one single closed area,
no crossing lines, no islands); In a separate table
(AREA_SOURCE_POLYGONS) the coordinates are stored
together with the OBJECT ID in a number indicating the drawing
sequence:
OBJ_ID n X Y
the drawing routine finds N pairs of coordinates X,Y from n=1 to
N and draws the polygon, closing it by connectin N with 1 again.
Page 52 of 95
PM3 Deliverable: A05.1




ESS GmbH
Data import: The area source geometry (polygon) can be
imported as a CSV text file in a separate import dialog.
Source specific temporal scaling factors (for month, day, hour)
a TABLE of substance specific average (default) emission
values (g/s);
optional one or more substance (or emission related parameter
such as production volume, energy consumption, etc) specific
temporal scaling factors;
 Area source import
Bulk Data Import:
To import a set of area sources together, the data from a spreadsheet
can be imported directly to the data base, with parallel import of the
corresponding polygon (geometry) data.
Polygon import:
An import function supports the import of polygon definition as shape
files from ArcView.
The interactive polygon import dialog includes a browser option on the
local client, and requires the specification of the features ID of the
polygon boundaries in the shape file imported.
 Line Sources
Line sources represent street or road segments and hold their emission
data. They are, in principle independent entities ([set of] poly lines, i.e.,
possibly with unconnected segments) but can part of a topological road
network as used by any traffic model such as EMME/s or VISUM.
A Line Source is a named OBJECT. If it is associated with a network, its
position and geometry is supplied by that network; else, the coordinates of
its start and end node are specified with the line source directly.
Line Source descriptions consists of the standard OBJECT header and
META DATA. It is characterized by its position and length, and it is shown
on the associated MAP.
The set of data/attributes that define the line source data structure
Descriptor (defined in the systems knowledge base) in the HTML
TEMPLATE that defines the object display and editor.
The attributes of a line source include:
 name, administrative designation (number or ID), road type,
length, number of lanes [optional: surface, slope, road network
link, fleet composition link].
 symbolic georeference to a model domain;
 vehicle frequency total (with optional vehicle class specific
values)
 average speed;
 Substance specific emissions in g/s/km.
 optional scaling factor and additive term for each pollutant.
Page 53 of 95
PM3 Deliverable: A05.1
ESS GmbH
 Road network
Road networks describe topological networks of line sources that represent
street or road segments and hold their emission data.
A Road network is a named OBJECT. Its description consists of the
standard OBJECT header and META DATA, including name and
description. It is characterized by its component segments, and it is shown
on the associated MAP. A large-scale display and zooming function
support the viewing of detailed networks.
The specific attributes of a road network beyond its standard META DATA
include:
1. associated model domain;
2. number of nodes and segments, total length;
3. method of generation (reference to external transportation models).
The network consists of a list of NODES with their X and Y coordinates,
and a list of segments, that connect to at most two nodes.
The segment link to the corresponding line source object and its attribute.
Every node must be connected to at least one SEGMENT. The number of
nodes in a well connected network must not exceed the number of
segments by more than one.
Please Note: a network is an optional and auxiliary construct. Line sources
are, in principle, potentially independent objects that require no
connectivity for their use as emission sources.
 Fleet Composition
Fleet compositions are named OBJECTS in AirWare. They are specified
for all line sources in the overall domain:
To estimate emission from traffic, several components are required:
1. traffic frequencies (number of vehicles passing a segment per unit
of time)
2. average speed of traffic (km/hour)
3. fleet composition.
4. cold start fraction and road type dependent emission factors for
each pollutant considered.
A fleet composition OBJECTS consists of the relative share (number of
vehicles) for any user defined set of classes, for example:
1. Gasoline Passenger cars
2. Diesel Passenger cars
3. LPG Passenger cars
4. Gasoline Light Duty vehicles
5. Diesel Light Duty vehicles
6. Heavy Duty vehicles
7. Urban Busses
8. Coaches
9. Motorcycles
10. Other (new technologies).
Page 54 of 95
PM3 Deliverable: A05.1
ESS GmbH
The set of vehicle classes is user defined: a Descriptor vehicle_classes in
the Knowledge Base contains the list of classes to be considered in the
fleet composition and in the emission factor table.
The fleet composition object consists of the standard object header and
META DATA, and a table of relative shares (in % of vehicles) that should
sum to 100%
The fleet composition is global the same for all elements of a
network); alternative fleet compositions can be selected at the level
of emission scenarios.
 Complex sources
Complex sources are composite area sources that can also include
explicit individual point, line and area sources, that are automatically
deducted from the complex source emission totals. An application example
are the European level EMEP emission data.
 Scenarios
All simulation model runs or scenarios use a combination of different sets
of assumptions or pre-processed input data, that must be kept consistent
to generate meaningful results.
An emission scenario is a dynamically instantiated OBJECT, similar to
the emission inventories, but model specific, which includes a START
TIME and a DURATION, as well as a choice of pollutant for the emission
data sets. The elements of a model scenario include:
1. Spatial extent or model domain due to computational requirements,
this implies the possible model (grid) resolutions and, indirectly, the
feasible range of duration of a model run. Time step (for model
output, internal time steps for dynamic models are derived from
computational stability requirements of the respective numerical
solvers) is closely related, but is largely determined by the
temporaral variability of the processes simulated.
The choice of location or area of interest constrains the set of
applicable
o Meteorological scenarios that requires a set of consistent
parameters (time series of fields) the latter being preprocessed dynamically based on the applicable model grid
resolution and for the model domain selected.
Depending on the model implied by the choice of domain,
this implies meteorological data either as (time series) of
scalar values, or dynamic fields (time series of matrices) in
one or more vertical layers.
2. Time and duration: together with the location and extent, a
scenario is defined by the start date/time and duration for the
simulations. This, in turn is constrained by the model chosen which
will either represent
o a sequence of hourly steady-state solutions (AERMOD),
Page 55 of 95
PM3 Deliverable: A05.1
ESS GmbH
an individual meteorological event with a one hour averaging
time (Convolution/traffic)
o a short-term (2-6 hours) dynamic and highly transient event
(MULTI-PUFF)
o a 48 hour scenario at hourly resolution (CAMx).
The choice of model and its pollutant emission requirements as well
as the implied duration and temporal resolution defines the:
o Emission scenario: this is dynamically derived from all
sources in the shared emission inventory within the spatial
(domain) and temporal scope of the model scenario.
Dynamic Emission Scenarios
Emission data for a model scenario are generated dynamically for a given
scenario. The specification for the pre-processing defined below are kept
with the actual model scenario, together with the complete emission data
derived from the original data in the emission inventories.
The emission pre-processing includes:
1. Pollutants: definition of a model specific list of pollutants for data
are required;
2. Sources identification of all sources located within the model
domain.
3. Temporal patterns: extraction of the emission data for the
substances required and the period of interest (model start
date/time and duration);
The data used can either be:
o Derived from a time series of emission observations if it
overlaps with the model period;
o Derived from the average emission value for each pollutant of
concern and adapted to the model period by means of the
temporal correction factors for month, day of the week, and
hour of the day;
the values extracted for point sources from the emission inventory
are then adjusted with the multipliers and additive term optionally
specified by the user for each source/pollutant combination.
o
Emission scenarios cover two basic formats:
1. Stack data (time series, direct or through temporal patterns)
2. Gridded data (small stacks, area sources, traffic for CAMx), also as
time series.
3. Line sources (used explicitly in the convolution model, gridded for
CAMx).
Please note: to keep the models consistent and ensure reasonable
operation, the emission scenario generated will always include a
COMPLETE set of source/pollutant combinations, even if there are no
values for any one of the required source/pollutant combinations: thisa will
be represented by a value of 0.00 that the user can correct with the
additive correction (and possibly a multiplier). By the same token, any
given source in the set can be "switched off" with a 0 multiplier.
Page 56 of 95
PM3 Deliverable: A05.1
ESS GmbH
Scenario display
The display page for a dynamically generated emission scenario
corresponds to the display page of the overall emission inventory summary
page.
Hypothetical sources
A hypothetical (e.g., future) source can be generated by copying an
existing source in the corresponding emission scenario, selecting
the type "hypothetical" and modifying its location and source and
emission characteristics as required.
 Matrices and Export
Emission data are also available as gridded, topical maps can can be
viewed as map overlays or topical maps, and exported in numerical format
for further processing with external tools (statistics, GIS).
The emission matrices are either calculated
1. automatically by the pre-processors for CAMx and AERMOD
scheduled (hourly and daily) runs;
2. interactively, on demand, from the emission inventories;
3. interactively, as part of a model scenario (CAMx or AERMOD).
The listing of the currently available emission matrices is triggered by the
text link Emission matrices at the bottom of the main menu page.
Emission Matrix Data
Emission Matrices are OBJECTS with the standard OBJECT data and
meta data.
The attributes of an emission matrix include:
 Matrix name
 Descriptor and unit
 Domain reference
 Size and resolution
 Dimensionality (layers and time steps)
 Basic statistics
Data Export
From the main emission matrix page, a button labeled export triggers an
export dialog.
This will start a small po-up window with a summary of the emission data,
and a text link export to the actual gridded emission data. The data are
organized (for all non-zero values) in a simple text/ascii format:
X-coordinate Y-coordinate emission value
A left mouse button click will display the data;
A right mouse button click will start a download dialog on the local
client PC.
This includes the option Save Link as ... which will download the
emission data into a local user defined text file.
Page 57 of 95
PM3 Deliverable: A05.1
ESS GmbH
 Emission Patterns
Emissions vary in time.
To reflect that, AirWare supports three mechanisms:
o explicit editing of emission time series (e.g., for PBM, PUFF);
o use of time-series data of emissions (only for boilers/major stacks
associated with industrial plants);
o use of temporal scaling factors.
The temporal scaling factor can be specified for:
o monthly variations, 12 values for seasonal variations;
o day of the week, 7 values, for differences between workdays and
weekends;
o hourly variations, 24 values, for daily patterns such as working
hours or rush hour.
The values (multipliers) can be normalized (BUTTON: NORMALIZE) so
that their total effect is neutral or mass conservative over a full year, when
applied to the annual average value at an hourly basis.
The temporal patterns can be assigned to
o individual emission sources
classes of emission sources (by sector) at the level of model
specific emission scenarios.
 Exceptions: holidays
A list of statutory holidays with their associated emission reduction (or
more general: correction) factors is available as a special object class.
Unless otherwise specified, a holiday uses the scaling factor for
Sunday.
The list (read only) is linked from the temporal emission pattern editor
page, and can be extended/edited from the Object Data Base level.
A holiday is defined by:
1. the DATE (YYYY-MM-DD, can be selected with in interactive
date editor (calendar widget);
2. a name (short text string, the Object Name, see also: Object
data structure
3. an optional explanatory text field;
4. the emission reduction factor to be applied for that day.
 Emission Estimates
The emission inventories in AirWare include, for every source and the list
of pollutants defined and configured in the system Knowledge Base on
emission value in g/s. This represents the long-term average emission rate
for that source, which is adapted for any emission scenario by temporal
scaling factors to estimate values for any specific hour.
The basic emission values for each substance can be edited interactively.
If no value was defined the string undefined is shown to indicate the
missing data value.
Page 58 of 95
PM3 Deliverable: A05.1
ESS GmbH
The emission inventories offer two ways of estimating emission values
where no primary data are available:
 from fuel type and consumption;
 from a qualitative size classification.
 Estimating emission from fuel type and consumption
For every emission source of the types: boilers, small stacks, area sources
a fuel type (one of a user defined list configured in the systems knowledge
base) and a an annual average fuel consumption can be specified.
Whenever the emission inventory is regenerated (button RECALCULATE)
an estimate based on fuel type, fuel consumption, and a fuel type and
pollutant specific emission factors.
Estimated emission values are marked with a red asterisk. *
Estimating emissions from qualitative size classification
If neither a directly edited value nor a complete data set for estimation from
fuel type and consumption (and corresponding emission factor) is
available, a rule-based semi-quantitative approach can be used. This
requires a qualitative classification of a source in terms such as small,
medium, large (the list of classes and designating symbols is user defined
in the KnowledgeBase). The estimation method first established for all
source classes the distribution of available emission data. This is then split
in linear intervals according to the number of classes defined by the user.
The median of the class corresponding to the qualitative source
characterization is then assigned to the source.
Emission values that have been estimates are marked with a red asterisk. *
Update strategy
The emission inventory primarily shows the direct emission estimates; IF
and only IF no direct emission estimate is in the data base (for any or all of
the pollutants), the estimation is attempted IF and only IF fuel type and
annual fuel consumption are both defined. Then, whenever the emission
inventory ir recalculated (the button: RECALCULATE) is pressed) missing,
undefined emission values will be replaced with the fuel factor derived
estimates wherever possible.
In a second pass, all remaining undefined values will be replaced with the
rule based estimate where the relative source size has been defined.
User define values always replace estimated values;
Estimates based on fuel replace the rule-based estimates.
 Emission Factors (combustion)
Emissions for combustion related sources are taken by the models from
the emission data base; these values are either directly entered with the
respective source, or can be estimates from source specific data on
 fuel type
 annual fuel consumption
Page 59 of 95
PM3 Deliverable: A05.1
ESS GmbH
together with the temporal emission patterns
The basic emission estimates are implemented as follows: the emission
factor defines the amount of pollutant generated (in g) from one unit of
“activity”, e.g., burning one kilo of fuel. This is then converted to the
required emission unit of g/s by correcting for the number of seconds in
a year and the annual fuel consumption data values.
The emission inventory primarily uses direct emission estimates; If and
only IF no direct emission estimate is in the data base (for any or all of
the pollutants), the estimation is attempted IF and only IF fuel type and
annual fuel consumption are both defined. Then, whenever the
emission inventory ir recalculated (the button: RECALCULATE) is
pressed) missing, undefined emission values will be replaced with the
value estimated from fuel type, consumption, and the emission factor if
defined in the data base. derived estimates wherever possible.
 Emission Factors (traffic)
Emissions for traffic sources are taken by the models from the emission
data base, associated with each line source for any or all pollutants
considered.
The emission values are either directly entered with the respective soucre,
or can be estimates from source specififc data on
 vehicle class (defined in the fleet composition object)
 average speed over a segment
 traffic frequency over all classes
 road type or category and road type correction factor
 cold start fraction and correction factor
 data base tables of vehicle class, speed, and pollutant specific
emission factors
 correction factors for cold starts and road type.
together with the temporal emission patterns.
 Greenhouse Gas Emissions
AirWare maintains a set of emission inventories as OBJECT CLASSES in
the OBJECT DATA BASE with a class specific selector/navigator for Area
Sources, Boilers, Industrial Plants, Line Sources, and Small Stacks (default
alphabetical order in the OBJECT ClassList;
For Greenhouse gases (GHG) the emission inventories for "classical" air
pollutants are extended to
1. include additional source classes, based a more process oriented
classification;
2. include additional substances, primarily CO2 and Methane.
 GHG emission estimates
AirWare maintains a set of emission inventories as OBJECT CLASSES in
the OBJECT DATA BASE with a class specific selector/navigator for user
defined classes of sources including
Page 60 of 95
PM3 Deliverable: A05.1
ESS GmbH

combustion related (point, are and line sources), based on fuel type
and consumption as well as technological parameters;
 process related fugitive emissions (area sources), based on process
specific emission factors and control technologies;
 biogenic emissions (are sources) based on land-use and climatic
parameters;
Conceptually, the estimates are based on:
Monitoring and direct measurement may involve continuous emission
monitoring systems (CEMS) (emissions recorded over an extended and
uninterrupted period), predictive emission monitoring (correlations
developed between measured emission rates and process parameters) or
source testing (e.g. stack sampling).
Mass balance involves the application of the law of conservation of mass
to a facility, process or piece of equipment. Emissions are determined from
the difference in the input and output of a unit operation where the
accumulation and depletion of a substance are included in the calculations
Emission factors uses emission factors (EF) to estimate the rate at which
a pollutant is released into the atmosphere (or captured) as a result of
some process activity or unit throughput. The EFs used may be average or
general EFs, or technology-specific EFs.
Engineering estimates may involve estimating emissions from
engineering principles and judgement, using knowledge of the chemical
and physical processes involved, the design features of the source, and an
understanding of the applicable physical and chemical laws.
In AirWare, this is implemented as an backward chaining Expert System
with first order production RULES.
 Volatile Organic Compounds
Volatile organic compounds (VOCs) together with Oxides of Nitrogen
(NOx) are the precursors of ozone formation.
Emissions of VOCs is usually from distributed, area sources rather than
stacks, and due to evaporation rather than combustion processes. Fugitive
VOC emissions originate from industrial and commercial areas and
households (e.g., from chemical production processes, use of points and
solvents), and the transportation (fuel) system, including
 gas stations
 mobile vehicles
 stationary vehicles.
Models that go beyond the city areas may also include biogenic VOC
sources.
Basic VOC groups required by the photochemical (ozone) model CAMx
include:
 Source emission rate of CO from area and line sources (kg/hour)
 Source emission rate of CO from point sources (kg/hour)
 Source emission rate of NOx from area and line sources (kg/hour)
 Source emission rate of NOx from point sources (kg/hour)
 Source emission rate of total hydrocarbons (THC) from area and
line sources (kg/hour)
Page 61 of 95
PM3 Deliverable: A05.1







ESS GmbH
Source emission rate of total hydrocarbons (THC) from point
sources (kg/hour)
Source emission rate of hydrocarbon classes from point sources
(moles/hour)
Source emission rate of hydrocarbon classes from area and line
sources (moles/hour)
Ratio of NO2/NOx in area and line source NOx emissions
Ratio fo NO2/NOx in point source NOx emissions
Ratio of CH4/THC in area and line source THC emissions
Ratio fo CH4/THC in point source THC emissions
 VOC speciation
Emissions of VOCs is usually from distributed, area sources rather than
stacks, and due to evaporation rather than combustion processes as well
as from biological processes.
Fugitive VOC emissions originate from industrial and commercial areas
and households (e.g., from chemical production processes, use of points
and solvents), and the transportation (fuel) system.
Since the simulation models require more detailed emission data (in terms
of chemical substances or "species") than are realistically measured,
assumptions on the distribution of bulk data (VOC) are made referred to
as speciation (default speciation profile “0”
For the 3D dynamic photochemical model CAMx,
Species TOG fraction
AirWare uses the EPA speciation list
ALD2
0.02
(http://www.epa.gov/ttn/chief/emch/speciation/cbivETH
0.05
profiles_mar_4_2002.xls ) to create CB5 lumped
FORM
0.02
OLE
0.05
species from TOG (total organic gases) taken from:
PAR
0.49
http://www.epa.gov/ttn/chief/emch/speciation/
TOL
0.05
The system does not differentiate by emission class. It
XYL
0.04
assumes that the VOC values contained in the data
NR
0.28
base refer to non-methane organic compounds (the
conversion factor for profile "0" is 1).
AirWare uses the default profile "0" for TOG .
The system also uses the default profile "0" for NOx conversion to NO and
NO2:
NO: 0.9; NO2: 0.10
 DUST entrainment model
The dust entrainment model estimates non-pyrogenic dust emission from
natural surfaces as a function of wind speed, land cover/vegetation, soil
characteristics, and soil moisture, primarily.
The total Dust PM10 emission [g/s/ha] is calculated as the product of
1. WindFactor,
2. ErosionFactor (erodibility)
3. CalibrationFactor.
Page 62 of 95
PM3 Deliverable: A05.1
ESS GmbH
The Calibration factor
is a user defined multiplier, that should be made part of a data assimilation
scheme to minimize overall bias.
The Wind factor
is computed from average
hourly (monitored ar generated
by MM5) ground level wind
speed (m/s) using a Weibull
function to generate a
distribution of wind speeds v
and their relative frequency
around that mean, as follows:
for v>TR: f(v) = (v - TR) **EXP
for v<TR: f(v) = 0
where TR is the (user-defined)
Wind Threshold,
and EXP is the (user-defined)
Exponent
windFactor = sum over all frequency classes of ( f(v)*frequency(v) )
frequency(v) = (k/c) * pow( (v/c), (k-1) ) * exp( -pow( (v/c), k) )
where v is the wind speed, k is the shape parameter, c is the scale parameter
of the Weibull distribution.
A user defined multiplier can be
used to scale the original, hourly
average wind speed.
The Erosion factor depends on:
 Vegetation index, derived
e.g., from land use (e.g.,
CORINE classification)
satellite imagery
(MODIS), scaled NDVI (010)
 FAO soil type (that


defines the fractions of
coarse (sand)a medium
(silt) and fine (loam) soil
components,
slope, orientation
(aspect) and gradient,
soil moisture, and in
conjunction with
drying/recently dried
soils, saltation as a
mechanical process to
break up the crust and
thus increase erodibility.
Page 63 of 95
PM3 Deliverable: A05.1
ESS GmbH
Soil moisture
the effect of soil moisture is described as a simple linear treshold function:
soil moisture > THRESHOLD (e.g., 80 % of field capacity) => sm_factor
=0
soil moisture < THRESHOLD => sm_factor = (THRESHOLD (in %) soil_moisture) * 2
A user defined soil moisture multipier can be used to adjust any observed soil
moisture data.
Drying soils, crusts and the saltation factor
Depending on composition (loam component) drying soils will form a
crust that limits erodibility. Depending on the coarse sand fraction, and
the time/wind exposure since drying (to or below the soil moisture
threshold) this crust will be broken through the process of saltation,
which will increase erodibility asymptotically to its default value
corresponding to the actual soil moisture.
The corresponding correction factor is a linear function of "wind-hours"
above a saltation threshold.
Deposition, loss and re-entrainment
Particulates are deposited on all surfaces, and depending on a surface
specific loss term, available for re-entrainment. The loss factor is
applied once, at the hour of deposition. The remaining mass
accumulates until the wind exceeds the threshold in any one of the
speed classes and the wind factor > 0.0.
The loss coefficient is a function of land cover, 100% for open water.
The loss factor is one of the properties defined in the land cover
parameter table.
Re-entrainment from deposited dust is limited to the accumulated
deposition, corrected by a land cover (surface) specific loss factor
(under development).
 DUST sensitivity analysis
The main dust scenario page leads to a sensitivity analysis tool.
This displays the basic emission matrix from the calling scenario page,
together with the basic model
parameters and result
statistics;
The page offers the
possibility to edit
(temporarily) any of the
model parameters and rerun
the model, select an
alternative model domain,
and zoom into the resulting
emissions matrix.
Page 64 of 95
PM3 Deliverable: A05.1
ESS GmbH
Sensitivity analysis functions also include the possibility to disable (or set
constant over the entire domain) selected inputs and associated
processes; these include:
o the three soil fractions coarse, medium, heavy (fine);
o soil moisture
o vegetation index
o slope/gradient
o wind speed
The results are visualized as an hourly matrix at a time, that can be switched
by a tapedeck; as a set of 24 hourly thumbnails, and as a histrogram.
The tools also supports visualizations of input matrices:
 MM5 scenario (wind field, soil moisture
 the distribution of the three soil fractions
 vegetation index
 the slope/gradient
Finally, the user can select any arbitrary position on the map to display the
Weibull distribution and corresponding (potential) erosion for that location.
 DUST calibration
The calibration of the DUST model uses any or all of a set of model
parameters, accessible from the DUST sensitivity analysis page.
The default values for the forecast scenarios are defined (and read for
every new run) in the location.setup data base, accessible from the
administrative menu.
Parameters available for calibration include:
o wind speed threshold for the three soil fractions
o sold moisture threshold (global)
o soil moisture linear dependency (slope)
o wind multiplier
o wind exponent
o Weibull shape parameter
o global scale factor.
For the comparison of
simulated and observed
values, two different
mechanisms are supported:
o comparison of
individual station data
(hourly or daily
aggregate)
o comparison of
matrices (e.g.,
synoptic AOT data)
In both cases, basic
statistical moments and
nonparametric regression
Page 65 of 95
PM3 Deliverable: A05.1
ESS GmbH
parameters are computed. optional tools include also a coarse
“coincidence analysis) for an arbitrary number of classes/bin of the
underlying continuous variable.
Please note that calibration
(and validation) of the DUST
model against
monitoring data is indirect, as
the monitoring station will
also be affected by any fate
and (long-range) transport
from a wide range of source
areas. Direct calibration of
the emission would require
specific near field
experimental setup, with the
monitoring very close
downwind of a suspected
source area.
The DUST emission model and the subsequent fate and transport model
CAMx operates at an hourly resolution. For the direct comparison of the
hourly model results and the daily gravimetric monitoring data, an
aggregation function to generate average 24 hour values has been
implemented in support of the calibration and validation tools.
 DUST input data
The dust model uses three
types of input:
o model parameters
(static, unless
adjusted with data
assimilation)
o dynamic driving forces
from MM5 (wind field,
soil moisture)
o spatial data (static
with the exception of a
possible dynamic
vegetation index with
seasonal adjustments.
These data cab be visualized in the sensitivity analysis page, that
provides a pull-down menu for the display of visual input data,
including:
o DEM and slopes,
o soil fraction (coarse, medium, fine (heavy)
o vegetation index,
o MM5 outputs (wind field, soil moisture)
Page 66 of 95
PM3 Deliverable: A05.1
ESS GmbH
7. Scenarios and Scenario Analysis
 Real-time scenarios
For CAMx, (and the implicit pre-processors MM5 and DUST), AERMOD
and PBM (not applicable in PM3) , real-time scenarios (daily forecast and
hourly nowcasts) real-time scenarios can be constructed.
There are two possible real-time operations:
1. Now-casting based on the most recent meteorological observations
and air quality data for possible data assimilation;
2. Fore-casting based on the most recent meteorolgical forecast,
usually over 24 hours (MM5 prognostic model run with NCEP/GFS
input data)
3.
For the real-time scenario, the input data for meteorology and emission are
generated automatically, scheduled or event driven (availability of the
necessry in put data), and the models are executed automatically, again
schedules or even based.
The results of a ral-time run are offered either as default result on
dedicated results pages, or within the normal scenario selection.
Configuration of scheduled runs
The configuration of schedules runs consists of the following step:
1. Select the corresponding scenario from the scenario list, or create a
NEW scenario with the NEW button at that level; For CAMx, the
main scenario for the master domain is mandatory, but the nested
sub-domain scenario can be configured interactively;
2. Edit the scenario (existing or new) to define
o model domain (or nested sub-domain for CAMx);
o pollutant;
o emission scenario;
o meteorological scenario.
3. Editing the corresponding shell script (see below, requires ADMIN
privileges) to add the new scenario to the list of scheduled runs;
4. optionally edit the crontab entry to modify the frequency of runs.
Please note: all scenarios of the same type must run with the same
temporal resolution, like every N hours.
5. The number of CAMx subdomains for one scenario must not excced
9.
6. The subd-omains have to be square, and FIT to the main domain
master grid (origin on a master grid node, i.e., the coordinates in
meter must end with 000 (see the CYPRUS domain definition in the
MapSet), size is arbitrary.
Editing the scenarios
The model scenarios consist of the following elements that can be
interactively configured by the user:
Page 67 of 95
PM3 Deliverable: A05.1




ESS GmbH
Selection of the model domain; the button DOMAIN leads to a listing
of all domains available, including the option to generate a new one;
viewing a domain also shows a button: SELECT to select the
domain currently displayed;
Selection of a pollutant (for CAMx, and the associated chemical
mechanism in CAMx, the list of applicable pollutants is defined in
the systems knowledge base);
Selection of a meteorological scenario; depending on the model and
local configuration, this primarily involves selection of one
(AERMOD) or several meteorological stations to be used for the
meteorological input data time series; for CAMx, this can also
include the option to use the MM5 meteorological fields rather than
individual monitoring stations and the diagnostic wind model DWM.
Editing of the emission scenario, including switching on or off the
DUST emission pre-processor. This is primarily derived from all the
sources in the current model domain. However, the user can:
o edit the multipliers for the classes: Boilers; Small stacks; Area
Sources; and Line Sources subject to:
 ADMIN user access rights;
 the model not currently running.
o switch on or off individual sources;
o select or edit the temporal patterns associated with the
individual source classes;
o go to the emission inventory and edit the sources
themselves.
Please note: while the changes to a scenario are local to the
scenario, any changes to the inventory are SHARED by all
models that do refer to the inventory: the on/off switch or a
sectoral multiplier are scenario specific, but the basic emission
rate is global and potentially shared !
Editing the real-time shell scripts
The scripts to drive the scheduled model runs are located on /var/www/cgibin in the respective model specific sub-directories. Below, the example
shows a typical script for CAMx NOx nowcasts:
#!/bin/sh
# This file is used to run the camx forecast scenario for NOx
# A new run is only started when the preceeding run is already
finished
# move error log from time to time
if ( test `find /var/log/camx_NOxforecast.error -size +50k` )
then
mv /var/log/camx_NOxforecast.error
/var/log/camx_NOxforecast.error.old
fi
date >> /var/log/camx_NOxforecast.error
if ! test -e /var/www/cgi-bin/camx/runForecast.lck
Page 68 of 95
PM3 Deliverable: A05.1
ESS GmbH
then
touch /var/www/cgi-bin/camx/runForecast.lck
/var/www/cgi-bin/camx/camxmain.cgi -s 18534 -p
/var/cyprusair/camx/ -f /var/www/cgi-bin/db_config --forecast 1>
/var/log/camx_NOxforecast.out 2>>
/var/log/camx_NOxforecast.error
rm -f /var/www/cgi-bin/camx/runForecast.lck
else
echo "Forecast locked" >> /var/log/camx_NOxforecast.error
fi
exit 0
editing crontab entries
To trigger or execute the scripts at a given schedule, they are entered in
the systems crontab; this can be done as user root with the command:
crontab -e;
#Scheduled Model Execution:
1 * * * * sh /var/www/cgi-bin/aermod/runNowcast.sh
1 0 * * * sh /var/www/cgi-bin/pbm/pbmForecast.sh
11 * * * * sh /var/www/cgi-bin/camx/runNOxNowcast.sh
21 * * * * sh /var/www/cgi-bin/camx/runPM10Nowcast.sh
31 * * * * sh /var/www/cgi-bin/camx/runO3Nowcast.sh
5 2 * * * sh /var/www/cgi-bin/camx/runNOxForecast.sh
15 2 * * * sh /var/www/cgi-bin/camx/runPM10Forecast.sh
45 4 * * * sh /var/www/cgi-bin/camx/runO3Forecast.sh
For a description of the crontab syntax, please consult the serves
on-line system manual with: man crontab.
As an optional alternative for more complex scheduling, the sequence and
timing of the model execution can be event driven rather than following a
fixed schedule, and based on the RTXPS real-time rule based expert
system.
 Configuration: HOW TO
The setup and configuration of
1. CAMx Nowcast model scenarios (automatically uses previous
results as initial conditions)
2. CAMx Forecast model scenarios (automatically uses previous
results as initial conditions)
consists basically of three steps:
1. Generate the nowcast or forecast scenario
2. Generate or edit the shell script in /var/www/cgi-bin/camx or
/var/www/cgi-bin/aermod, resp.
3. Make an entry into crontab
Plese note:
o to edit an existing scenario, it has to be reset (discarding any
results) if available to avoid inconsistencies;
Page 69 of 95
PM3 Deliverable: A05.1
ESS GmbH
o for nested scenarios, the configuration must be set for all domains,
starting with the top level.
o for multi-day forecasts scenarios, please note that consecutive
scenarios must have the same configuration (e.g., resolution)
generate a CAMx nowcast/forecast scenario
This currently needs to be done directly in the data base; for specific
configuration requests, please contact: info@ess.co.at
Scheduled run configuration
The scheduled runs are configured in the following files: The file names are
arbitrary, they only need to be the consistent with the filenames in the
crontab.
CAMx Nowcast:
/var/www/cgi-bin/camx/run[APPLICATION]Nowcast.sh
To add a new nowcast run, only the corresponding scenarioID needs to be
added to the list of scenarioID's (line26)
CAMx Forecast:
/var/www/cgi-bin/camx/run[APPLICATION]NOxForecast.sh
To add a new forecast run, make a copy, and substitute the scenarioID
(line 4)
In addition to the scenario ID the "run*.sh" files contain the following
information (nothing needs to be changed here!):
# database configuration file
config="/var/www/cgi-bin/[APPLICATION]_db"
# base directory of cgi's
cgiDir="/var/www/cgi-bin/"
# where to write log and error file
errorFile="$path/camx/log/forecastNOx.error"
logFile="$path/camx/log/forecastNOx.log"
# base directory of output files
outDir="$path/camx/forecast/"
# where to find camx input files generated by mm5 (CAMx only)
mm5Dir="/var/airdata/mm5/"
# the simulation date (forecast only)
tmw=`date -d tomorrow +'%Y-%m-%d'`
Scheduling (crontab entries
The runs are started via the following (root) crontab entries: (edit with:
crontab -e)
#camx
mn * * * * /var/www/cgi-bin/camx/run[APPLICATION]Nowcast.sh mf hf * * *
/var/www/cgi-bin/camx/run[APLICATION]NOxForecast.sh
Any new CAMx forecast runs need to be added here.
mn is the minute of nowcast start - depending on server time relative to
location time (mn=1 or mn=31). If server time is equal to location time
mn=1.
Page 70 of 95
PM3 Deliverable: A05.1
ESS GmbH
mf is the minute, hf is the hour of forecast start, depending on the
availability of MM5 results. It can be started as soon as MM5 results are
available.
Make sure that each of the Nowcast runs will be finished within an hour (for
all scenarios included in the shell script). Otherwise the nowcast of the next
hour will be aborted.
Make sure that only ONE CAMx forecast is running at a time. Otherwise
the run which starts later will be aborted.
 Dynamic scenarios
Scenarios for the air quality models are named model-specific objects,
selectd from a model specific scenario selector/navigator, and generated
or modified with a separate scenario generator/editor. The primary
attributes of a scenario on the listing of available scenario include:
 Scenario name
 Short description (first few characters)
 Model domain or area of interest
 Start date/time
 Owner/user of the scenario
 Date of last modification
There are two distinct modes of operation:
1. Selection and modifications of an existing scenario;
2. Generation of a new scenario
Modifying and running an existing scenario
Within an existing scenario, the user can change:
 Start date/time within the constraints of the meteorological scenario
chosen;
 The pollutant, depending on the choices offered by the respective
model;
 Emissions by modifying the sources defined by the domain and the
emission values defined by the start date and time;
PLEASE NOTE: existing scenario with available results must be rest
(discarding the results) before they can be changed to avoid
inconsistencies between scenario and results.
Creating a new model scenario
Creating a new model scenario involves the choice of:
1. the model domain
2. the meteorological scenario
After that, the procedure is the same as with any pre-existing
scenario.
 Steady-state scenarios
not applicable in PM3
Page 71 of 95
PM3 Deliverable: A05.1
ESS GmbH
8. Simulation Models
 AERMOD description
The AERMOD model system (regulatory Gaussian model) is not being
used in PM3.
 Dynamic Multi PUFF
The Lagrangian/Gaussian Puff Model in AirWare is derived from the
USEPA model INPUFF 2.4 (Petersen and Lavadas, 1986).
INPUFF is a Gaussian INtegrated PUFF model. It is designed to simulate
dispersion from semi-instantaneous or continuous and optionally mobile
point sources over a spatially and temporally variable wind field (provided
by prognostic meteorological models such as MM5 or WRF, or
diagnostically interpolated wind fields). The algorithm is based on a
Lagrangian transport framework for a sequence of emission events which
disperse with Gaussian puff assumptions including a vertically uniform
wind direction field and no chemical reactions.
The analytical solutions for
atmospheric concentration of
a gaseous or suspended
particulate pollutant,
incorporating dry deposition
and gravitational settling were
given by Rao (1982).
PUFF is implemented for:
mobile sources, including 3D
(airplane TOL)
 Basic scenario
analysis, for a given
domain, start date and
duration with the wind
field and
meteorological data
provided either
o by the data base of historical (re-analysis, FNL) prognostic
model results;
o daily (continuous, GFS) prognostic meteorological forecasts;
 Daily forecasts and now-casts;
 Scenario analysis for individual moving sources linked to the mobile
source object data base.
 Point Source screening
A simple JAVA applet, implementing the basic Gaussian equations for
elevated point sources, is linked to individual point sources in the emission
data base. It is intended for fast screening level analysis of maximum
concentration and the distance of the maximum fro the sources, as well as
the direct comparison of alternative scenarios including stack parameters.
Page 72 of 95
PM3 Deliverable: A05.1
ESS GmbH
 Mobile Sources (PUFF)
Implementation of Multi-puff
for mobile sources, including
a 2D or 3D trajectory editor,
intended for road
transportation (cars, buses,
trucks, motorcycles), railways
(diesel/diesel electric
locomotives), shipping, and
airplanes.
Depending on the type of
vehicle, the model is run with
an internal computational time
step of milliseconds. Several
vehicles can be run in
parallel, and scaled for the
estimation of emission and dispersion over longer periods.
 PBM Ozone Model
Support for PBM (photochemical box model) has been discontinued as of
release 5.8; it is now fully replaced by CAMx, see below.
 CAMx 3D Dynamic O3
The Comprehensive Air quality Model with extensions (CAMx) is an
Eulerian photochemical dispersion model that allows for an integrated oneatmosphere assessment of gaseous and particulate air pollution (ozone,
PM-2.5, PM-10, air toxics, mercury) over many scales ranging from suburban to continental.
See also: CAMx User Guide V5.40, PDF
It is designed to unify all of the technical features required of state-of-thescience air quality models into a single system that is computationally
efficient, easy to use, and publicly available. The model code has a highly
modular and well documented structure which eases the insertion of new
or alternate algorithms and features.
CAMx simulates the emission, dispersion, chemical reaction, and removal
of pollutants in the troposphere by solving the pollutant continuity equation
for each chemical species (l) on a system of nested three-dimensional
grids. The Eulerian continuity equation describes the time dependency of
the average species concentration (cl) within each grid cell volume as a
sum of all of the physical and chemical processes operating on that
volume.
Page 73 of 95
PM3 Deliverable: A05.1
ESS GmbH
The governing equations are expressed mathematically in terrain-following
height (z) coordinates. It considers a horizontal wind vector, net vertical
entrainment rate, multiple vertical layers, atmospheric density, and
turbulent exchange (or diffusion). The terms on the right-hand side
represents horizontal advection, net resolved vertical transport across an
arbitrary space- and time-varying height grid, and sub-grid scale turbulent
diffusion. Chemistry is treated by simultaneously solving a set of reaction
equations defined from specific chemical mechanisms. Pollutant removal
includes both dry surface uptake (deposition) and wet scavenging by liquid
precipitation (rain).
 CAMx basic data
Data requirements
DATA CLASS
DATA TYPES
Meteorology
Supplied by MM5 or
WRF)

3-Dimensional Gridded Fields:
o Vertical Grid Structure
o Horizontal Wind Components
o Temperature
o Pressure
o Water Vapor
o Vertical Diffusivity
o Clouds/Rainfall
Air Quality
Obtained from
Measured Ambient Data
Emissions
Supplied by DUST,
EMEP, emission
inventories and patterns
(EMEP, national)





Gridded Initial Concentrations
Gridded Boundary Concentrations
Time/space Constant Top Concentrations
Elevated Point Sources
Combined Gridded Sources
o Low-level Point
o Mobile
o Area/Non-road Mobile
Biogenic
Gridded DUST (wind erosion)
Gridded Surface Characteristics
o Land Use/Vegetative Cover
o UV Albedo
o Snow Cover
o Land/Water Mask
o Roughness Length
Drought Stress
Atmospheric Radiative Properties
o Gridded Haze Opacity Codes
o Gridded Ozone Column Codes
Photolysis Rates Lookup Table
Geographic
Developed from
Landuse/Landcover
Maps (CORINE, USGS)
Drought Index Maps,
Modeled or Satellite
derived Snow Cover
Photolysis
Derived from Satellite
Measurements and
Radiative Transfer
Models






Page 74 of 95
PM3 Deliverable: A05.1
ESS GmbH
Roughness Length
Land use surface characteristics and associated roughness (length) as well as UV
albedo are important inputs for CAMx.
CAMx landuse categories and the default surface roughness values (m) assigned to
each category by season. Winter is defined for conditions where there is snow
present; winter months with no snow are assigned to the Fall category. Roughness
for water is taken as the maximum of the baseline value given in the table, and the
function Z0 = 0.000002 w**2.5 , where w is surface wind speed (m/s).
The listed UV albedo values can be used to assign UV albedo from landuse data in
preparing the albedo/haze/ozone (AHO) input file.
Surface Roughness (in meters) and UV Albedo by land cover and season
N.
1
2
3
4
5
6
7
8
9
10
11
Land cover
urban
agriculture
range land
deciduous forest
coniferous forest, wetland
mixed forest
water
barren land
non-forested wetlands
mixed agricultural/range
rocky, low shrubs
spring
1.00
0.03
0.05
1.00
1.30
1.15
0.0001
0.002
0.20
0.04
0.30
summer
1.00
0.20
0.10
1.30
1.30
1.30
0.0001
0.002
0.20
0.15
0.30
fall
1.00
0.05
0.01
0.80
1.30
1.05
0.0001
0.002
0.20
0.03
0.30
winter
1.00
0.01
0.001
0.50
1.30
0.90
0.0001
0.002
0.05
0.006
0.15
UV albedo
0.08
0.05
0.05
0.05
0.05
0.05
0.04
0.08
0.05
0.05
0.05
CAMx Basic data requirements
1. Landuse [time invariant 2-dimensional matrix]- one of the following categories,
derived from USGS or CORINE land cover data sets:
o Urban
o Agricultural
o Rangeland
o Deciduous forest
o Coniferous forest, wetland
o Mixed forest
o Water
o Barren land
o Non-forested wetlands
o Mixed agricultural/range
o Rocky (with low shrubs)
2. Ozone Column density (Dobson Units) [time variant 2-dimensional matrix]
3. Haze turbidity value (unitless) [time variant 2-dim matrix]
4. Drought stress (Palmer Drought Index) [time invariant 2-dim matrix]
5. Layer Interface Height (m ABG) [time variant 3-dim matrix]
6. Pressure (mb) [time variant 3-dim matrix]
7. Layer average horizontal wind components (m/s) [time variant 3-dim matrix]
8. Temperature (Degree Celsius) [time variant 3-dim matrix]
Page 75 of 95
PM3 Deliverable: A05.1
ESS GmbH
9. Surface Temperature (Degree Celsius) [time variant 3-dim matrix]
10. Water vapor concentration (ppm) [time variant 3-dim matrix]
11. cloud water content (g/m3) [time variant 3-dim matrix]
12. precipitation water content (g/m3) [time variant 3-dim matrix]
13. total integrated column cloud optical depth (dimensionless) [time variant 3-dim
matrix]
14. vertical diffusivity (m2/s) [time variant 3-dim matrix]
15. Initial concentrations (ppm for gases, microgram/m3 for aerosols) [time
invariant 3-dim matrix] for all or for a subset of the modeled species
16. Upper boundary condition (ppm for gases, microgram/m3 for aerosols) [time
invariant scalar] for all or for a subset of the modeled species
17. Lateral boundary concentrations (ppm for gases, microgram/m3 for aerosols)
[time invariant scalar] for all or for a subset of the modeled species for each of
the 4 lateral boundaries
18. Emissions for each Elevated Point Source:
o Location (x,y coordinates)
o Stack height (m) [time invariant scalar]
o Stack diameter (m) [time invariant scalar]
o Stack exit temperature (degree Celsius) [time invariant scalar]
o Stack exit velocity (m/hr)[time invariant scalar]
o Stack flow rate (m3/hr) [time variant scalar]
o Effective plume height override (m) [time variant scalar]
o Emission rate (mol/hr for gasses, g/hr for aerosols) [time variant scalar]
for each species
19. Gridded emissions (mol/hr for gasses, g/hr for aerosols) [time variant 2-dim
matrix] for each species
Chemistry data requirements:
For the CB5/6 chemical mechanism option in CAMx, data on the following chemicals
(species and groups) are required: Species marked with (-) are reaction products of
the primary emitted species and are therefore not contained in the Emission Input
files, but are required for the initial concentrations, upper and lateral boundary
concentrations file. Where these data are not available, model defaults and standard
VOC speciation table are used.
Input chemical species for the CB4 mechanism (not all are required, a minimal data
set includes NO/NO2, CO, and VOV with a default speciation list):
1. Inorganic species represented individually:
o NO Nitric oxide
o NO2 Nitrogen dioxide
o CO Carbon monoxide
o -N2O5 Dinitrogen pentoxide
o -HNO3 Nitric acid
o -HONO Nitrous acid
o -O3 Ozone
o -PNA Peroxynitric acid
Page 76 of 95
PM3 Deliverable: A05.1
ESS GmbH
-H2O2 Hydrogen peroxide
2. Organic species represented individually:
o ETH Ethene
o FORM Formaldehyde
o ISOP Isoprene
o MEOH Methanol
o ETOH Ethanol
3. Organics species grouped together:
o PAR Paraffin carbon bond (C-C)
o OLE Olefin carbon bond (C=C)
o TOL Toluene and other monoalkyl aromatics
o XYL Xylene and other polyalkyl aromatics
o CRES Cresol and higher molecular weight phenols
o ALD2 Higher aldehyde (based on acetaldehyde)
o -PAN Peroxyacyl nitrate (based on peroxyacetyl nitrate)
o -NTR Organic nitrate (RNO3)
o -MGLY Methylglyoxal and other aromatic products
o -ISPD Isoprene product (lumped methacrolein, methyl vinyl ketone,
etc.)
o -OPEN Aromatic ring opening product
4. Additional chemical species for (optional) Aerosol Chemistry
o NH3 Ammonium
o OLE2 Biogenic olefin to represent terpenes
o PNO3 Particulate nitrate
o PSO4 Sulfate
o PNH4 Particulate ammonium
o POA Primary organic aerosol
o PEC Primary Elemental Carbon
o FPRM Fine other primary (<2.5 micrometer)
o FCRS Fine Crustal (<2.5 micrometer)
o CPRM Coarse other primary
o CCRS Coarse crustal
o NA Sodium
o PCL Particulate chloride
o PH2O Aerosol water content
o -CG1-CG5 Condensable organic gasses (products of PAR, TOL,
XYL,CRES, OLE2)
o -SOA1-SOA5 Secondary organic aerosols (PRODUCTS of CG1-CG5)
o -HCl Hydrogen Chloride (as a product of acidified sea salt aerosol)
o
Page 77 of 95
PM3 Deliverable: A05.1
ESS GmbH
 CAMx scenarios
A CAMx scenario is a named OBJECT that compiles, displays, and suports
editing of model scenarios.
Model scenarios can be:
 Scenarios proper, based on user defined set of input parameters;
 Now-casting scenarios, automatically run every hour, driven by realtime meteo data feeds, using the previous hourly result as initial
condition,
 Fore-casting scenarios, automatically run every 24 hours, driven by a
24 hour meteorological forecast, using the applicable now-casting result
for inital conditions.
Model implementation
CAMx is implemented for a pre-defined master domain covering an area
large enough to account for the air flow during the 48 hour simulation period at
a relatively coarse resolution e.g., 2 km; and a number of sub domains
covering areas of specific interest like major settlements or industrial areas, at
a finer resolution, e.g., 500 m.
Computations for the sub-domains can be switched on or off by the user.
Scenario Parameters
The following parameters define a CAMx scenario:
Scenario
parameter
Comment
Start date/time
user selectable, subject to the availability of input data in the
meteo scenario
Duration of
simulation
fixed at 1 hour for nowcasting, 48 hours for forecasting and
scenario runs
Master Domain
Fixed to the applicvation domain, e.g., Cyprus
Main Grid
resolution
scalar [m]; fixed to maintain consistency with all matrix preprocessors
Subdomains
user selectable from a list of (possibly user defined) domains
Subdomain Grid
resolution
fixed for performance reasons, scalar [m], e.g., 500
Vertical grid
structure
number of layers, layer interface heights [m]; pre-defined at the
application level, read only
Chemistry
mechanisms
user selectable:






conservative gas (NOx or SO2)
conservative particulate matter (PM < 2.5 micrometer
conservative particulate matter (PM > 2.5 micrometer
simple chemistry SO2/H2SO4
simple chemistry NO/NO2
CB4 (gas phase chemistry, 96 reactions and 37 species)
Page 78 of 95
PM3 Deliverable: A05.1
ESS GmbH

CB5/CF (CB5 including Aerosol chemistry, 117 reactions
and up to 67 species)
Output layers
only bottom layer or 3 dimensional output, default: bottom layer
only
Output time step
scalar [min], default is 60 minutes
Output pollutants defined through the chemistry mechanism, fixed for
conservative and simple chemistry; variable only for chemistry
mechanisms CB4 and CB4/CF; there a list of chemical species
is given; default setting
Meteo scenario
user selectable, constrained by domain, model specific
Emission
scenario
dynamically generated for the domain and period, model
resolution (for gridded data) and chemistry mechanism that
defines the set of pollutant emissions required)
Initial conditions
3D matrix for at least one chemical species, selected from the
set of matrix objects
Lateral boundary Four 2D matrices for at least one chemical species
conditions
Top
concentrations
Upper boundary condition: scalar for at least one chemical
species
Restart
true or false, if 'true' CAMx uses instanteous concentration files
of preceding run instead of initial concentration file; default in
true for nowcasting and forecasting)
Error Check
true/false 'true'=will stop after 1st timestep; default set of false
Maximum
timestep
scalar [min], default set to 15 minutes
Dry deposition
true/false, default set to true
Wet deposition
true/false, default set to true
PiG Submodel
possible values: None, GREASD, IRON, default set to None
Map projection
Possible values: LAMBERT, POLAR, UTM, LATLON, default set
to UTM
UTM Zone
defined by the main domain
Advection Solver Possible values: PPM, BOTT, default set to PPM
Chemistry Solver Possible values: CMC, IEH, LSODE, default set to CMC
Probing Tool
possible values: None, OSAT, PSAT, GOAt, APCA, DDM, PA,
RTRAC, default set to None
Staggered winds true/false, default set to false
Gridded
true/false, default is set to true
Page 79 of 95
PM3 Deliverable: A05.1
ESS GmbH
Emissions
Point Emissions
true/false, default set to true
Ignore Emission
Dates
true/false, default set to true
 CAMx sub-domains
1. The largest domain is the "main domain", smaller domains inside the main
domain are called "subdomains".
2. Each subdomain has to cover at least on
grid cell of the master domain. Therefore
the size as well as the position of each
subdomain depends on the grid definition of
the master domain. For example, if master
domain has a resolution of 30km, the
smallest subdomain possible has a size of
30km and it has to be positioned exactly on
the grid of the master domain, compare
figure.
3. Subdomains may not overlap.
4. A subdomain may be located within another
subdomain. The current implementation
allows 4 levels of domains, but this could be extended in principle.
5. When subdomains are nested, the larger subdomain has to be selected first.
6. The maximum number of subdomains is 9.
7. There are further restrictions to resolutions (compare below), but Airware's
CAMx domain selection offers only 'allowed' resolutions.
The following error messages will displayed on selecting a subdomain which violates
one of the above rules:
Violation of
point 2:
Subdomain 'name' is not compatible to grid of main domain
Violation of
point 3:
Subdomain 'name1' overlaps with 'name2'
Violation of
point 5:
Subdomain 'name1' is inside domain 'name2'. Larger domain needs
to be selected first
Violation of
point 6:
Too many subdomains. Maximum number of subdomains is 9
Violation of
point 7:
in case there no 'allowed' resolution for the subdomain: "No
compatible resolution for subdomain 'name".
Excerpt of CAMx User Guide:
Each grid nest is defined over a subset of master (coarsest) grid cells. The range of
master grid row and column indices that define the coverage of each nested grid
Page 80 of 95
PM3 Deliverable: A05.1
ESS GmbH
must be specified in the run control file. An integer number of nested grid cells must
span one master grid cell; this number is referred to as a “meshing factor”.
1. The ratio of master grid cell size to nested grid cell size must be an integer
(e.g., a “meshing factor” of 3 means that 3 nested cells span the
distance of 1 master cell, resulting in an area of 9 nested cells per master
cell);
2. For telescoping grids (a nested grid containing an even finer grid), the cell size
of the finest grid must be a common denominator for all parent grids above it
(e.g., a 36-12-4 km or 36-12-2 km arrangement is allowed, but a 36-12-9 km is
not);
3. The restriction in (2) above does not apply to parallel nested grids of the same
generation (e.g., 4 km and 5 km grids can be located in different areas of a
master grid provided that the master cell size is some multiple of 20 km);
4. Nested grids cannot overlap, although they may share a common boundary or
edge;
5. Nested grids cannot extend into a boundary, or non-modeled, area of the
master grid;
6. CAMx is currently configured to allow four "generations" of nests (e.g., four
levels of telescoping grids); this can be extended in the code easily if more
than four levels of nests are required;
 CAMx meteo data
CAMx reads and writes several 3-D time resolved files that will require
significant disk space. Therefore, these files are written as Fortran
unformatted binary to minimize storage requirements.
In contrast to the usual CAMx input files written in the IEEE "big_endian",
for AIRWARE the IEEE "little_endian" has to be used for compatibility
reasons. A detailed description of the file formats required can be found in
the CAMx User's Guide, Version 4.30, Environ International Corporation,
Novato, CA, 2006, pp.5.30 - 5.34. CAMxUsersGuide_v4.30.pdf
Main input files
1. Height/Pressure File: The Fortran binary height/pressure input file
contains three-dimensional gridded fields of layer interface heights
and layer-average pressure.
2. Wind File: The Fortran binary wind file contains three-dimensional
gridded fields of layer average horizontal (u and v) wind
components.
3. Temperature File: The Fortran binary temperature file contains
three-dimensional gridded fields of layer average temperature and
two-dimensional gridded fields of surface temperature.
4. Water Vapor File: The Fortran binary water vapor file contains
three-dimensional gridded fields of layer average water vapor
concentration.
5. Cloud/Rain File: The Fortran binary cloud/rain file contains threedimensional gridded fields of cloud parameters to be used for
chemistry and wet/dry deposition calculations. Note that
precipitation rate is not explicitly provided to the model; instead, it is
Page 81 of 95
PM3 Deliverable: A05.1
ESS GmbH
internally calculated from the three precipitation water content forms
provided on the cloud/rain file.
6. Vertical Diffusivity File: The Fortran binary vertical diffusivity file
contains three-dimensional gridded fields of layer interface vertical
diffusivity (i.e., turbulent exchange or diffusion coefficients).
7.
CAMx can use the (optional)MM5 meteorological pre-processor to
generate all necessary meteorological input files. If no MM5
meteorological pre-processor output is available, meteorological data
are derived in the following way:
1. Height/Pressure File:
Layer interface heights are fixed at
25 m, 50 m, 100 m, 200 m, 500 m, 1000 m, 2000 m and 4000 m.
Pressure (mb) [time variant 3-dim matrix]
calculated from ground atmosheric pressure and temperature and
elevation
M = 0.028964; [kgmol-1] molweight of air
g = 9.81; [ms-2] acceleration due to gravity
R = 8.314; [JK-1mol-1] Gas Constant
h: height
t[i][j] temperature of grid cell i,j, h=0
pres[i][j] pressure of grid cell i,j, h=0
p[(i,j,h) pressure of grid cell i,j, h
p[(i,j,h) = pres[i][j] * exp((-M*g*h/(R*t[i][j])));
Since most monitoring stations do not contain values for pressure,
pres[i][j] = 1001 mb at h = 0 m is assumed for all i, j of the domain
2. Wind File
The 3D wind dynamic field is calculated using the Diagnostic Wind
Field Model DWM and a synthetic geostrophic wind (assumed to be
homogeneous over the domain) estimated from the (vector average)
of anemometric wind data, using an exponential vertical profile for
speed and a Coriolis induced rotation.
This is based on the DEM ( 1 km resolution) for the domain, and the
anemometric wind speed and wind direction of one or more
monitoring stations the user can select interactively for the meteo
scenarios for scenario analysis (AERMOD and CAMx). The station
selection for nowcast and forecast runs are configurable but not
interactively to better maintain consistency.
3. Temperature File
Temperature [time variant 3-dim matrix] is
calculated from ground temperature [2-dim matrix] and height,
assuming adiabatic lapse rate alr =0.003 K/m
h: height (elevation + layer height)
T(x,y,h)* = T(x,y,h=0) - alr*h
Page 82 of 95
PM3 Deliverable: A05.1
ESS GmbH
The temperature of the selected monitoring station is the basis of the
calculation.
Surface Temperature [time variant 2-dim matrix]
The temperature of the selected monitoring station is the basis of the
calculation.
Water Vapor File
Water vapor concentration (ppm) [time variant 3-dim matrix]
calculated from temperature, pressure and relative air moisture
(Clausius-Clapeyeron, assuming temperature independent enthalpy)
R = 8.314 JK-1mol-1 gas constant
TR = 273.15 K reference temperature
H = 45050 Jmol-1 molar enthalpy of evaporation at TR
PR = 611 Pa saturated partial pressure of water vapor at TR
M_H2O = 18 gmol-1 molecular weight of water
M_AIR = 29 gmol-1 molecular weight of air
f = 0.4 assumed relative humidity
ps: saturated partial pressure of water vapor
ph2o: partial pressure of water vapor
t: temperature of grid cell i,j
pg: (total) pressure of grid cell i,j
w: absolute humidity of grid cell i,j
ps = PR*exp( -(H/R)*((1/t) - (1/TR)) );
ph2o = f*ps;
w = (M_H2O/M_AIR)*(ph2o/(pg-ph2o));
no data a re available: relative air moisture
is set to 40% for the whole domain.
Cloud/Rain File
no data available: set to zero.
Vertical Diffusivity File
Vertical diffusivity (m2/s) [time variant 3-dim matrix] is
calculated from surface roughness, the temporal temperature gradient
and layer height:
vd(i,j) = 10*sqrt(rough(i,j))*(t(i,j) - t_min(i,j) + 1)*(exp(cc*h))
vd(i,j): vertical diffusivity (m2/s) of grid cell i,j
rough(i,j): surface roughness of landuse categorie of grid cell i,j
t(i,j): temperature of grid cell i,j
t_min(i,j): minimal temperature of the day of grid cell i,j
h: layer height
cc = -0.004 exponential factor
Page 83 of 95
PM3 Deliverable: A05.1
ESS GmbH
 CAMx chemistry data
For the CB5 chemical mechanism option in CAMx, data on the following
chemicals (species and groups) are required: Species marked with (-) are
reaction products of the primary emitted species and are therefore not
contained in the Emission Input files, but are required for the initial
concentrations, upper and lateral boundary concentrations file.
Input chemical species for the CB4 mechanism (not all are required):
1. Inorganic species represented individually:
o NO Nitric oxide
o NO2 Nitrogen dioxide
o CO Carbon monoxide
o -N2O5 Dinitrogen pentoxide
o -HNO3 Nitric acid
o -HONO Nitrous acid
o -O3 Ozone
o -PNA Peroxynitric acid
o -H2O2 Hydrogen peroxide
2. Organic species represented individually:
o ETH Ethene
o FORM Formaldehyde
o ISOP Isoprene
o MEOH Methanol
o ETOH Ethanol
3. Organics species grouped together:
o PAR Paraffin carbon bond (C-C)
o OLE Olefin carbon bond (C=C)
o TOL Toluene and other monoalkyl aromatics
o XYL Xylene and other polyalkyl aromatics
o CRES Cresol and higher molecular weight phenols
o ALD2 Higher aldehyde (based on acetaldehyde)
o -PAN Peroxyacyl nitrate (based on peroxyacetyl nitrate)
o -NTR Organic nitrate (RNO3)
o -MGLY Methylglyoxal and other aromatic products
o -ISPD Isoprene product (lumped methacrolein, methyl vinyl ketone,
etc.)
o -OPEN Aromatic ring opening product
4. Additional chemical species for (optional) Aerosol Chemistry
o NH3 Ammonium
o OLE2 Biogenic olefin to represent terpenes
o PNO3 Particulate nitrate
o PSO4 Sulfate
o PNH4 Particulate ammonium
o POA Primary organic aerosol
Page 84 of 95
PM3 Deliverable: A05.1
o
o
o
o
o
o
o
o
o
o
o
ESS GmbH
PEC Primary Elemental Carbon
FPRM Fine other primary (<2.5 micrometer)
FCRS Fine Crustal (<2.5 micrometer)
CPRM Coarse other primary
CCRS Coarse crustal
NA Sodium
PCL Particulate chloride
PH2O Aerosol water content
-CG1-CG5 Condensable organic gasses (products of PAR, TOL,
XYL,CRES, OLE2)
-SOA1-SOA5 Secondary organic aerosols (PRODUCTS of CG1CG5)
-HCl Hydrogen Chloride (as a product of acidified sea salt aerosol)
9. Model output, analysis, impacts
 Model output display
All model interface (GUI) in
AirWare is designed to be
viewed by any industry
standard web browser,
without the need for specific
plugins, cookies, etc. Only
Javascript and pop-up
windows should be enabled.
All spatially distributed
models in AirWare produce
their output as MatrixObject,
in the form of color coded
overalys over a user selected background map.
These matrices are organised in sets along time and (vertical) space
dimensions. After a successful model run, the corresponding model result
is shown together with a summary of the scenario (meteorological and
emission data) a a color coded overlay over the default background map of
the domain for which the scenario was run.
Nowcast results
For the start page, the results of the current nowcast run for a user defined
(configured) substance, e.g., NOx is displayed. This uses opaque color,
and a color scheme defined in /var/www/cgi-bin/mapcolors which defines
21 RGB values. The minimum concentration (values below are ignored),
the maximum concentration (values above are shown in color 21), and the
color style can be defined interactively (clicking on the color ramp will start
an editor dialog). The selection of
 the model (must be configured to produce nowcast results)
Page 85 of 95
PM3 Deliverable: A05.1
ESS GmbH


the model domain
the pollutant
are configured in the data base: GLOBAL/LOCATIONS
Scenario results
In this primary output (transparent colors over the background map), the
user can select time steps and vertical layer; for a more interactive control
over the output display, the button DETAILS leads to a new page that
shows the map and model results in a bigger window. In addition, the
following controls area avilable:
 selection of time step (next and previous buttons);
 color coding (upper and lower bounds)
 isolines (4 user defined values)
 additiopnal map features (toggle on or off)
 zooming (arbitrary multi-leyer)
 data export (ascii, x,y,values for all non-zero values).

Color coding
here the user can define the upper and lower bounds of the color ramp to
be used. Values below the minimum set are transparent, not shown;
values above the maximum set are shown in white. In between, there are
20 classes (linear equidistance spacing) allocated to the colors of the
rainbow from blue (lowest values) to red (highest).
Isolines
Four user selected isolines can be displayed with pre-defined colors; the
user sets the value for which the isoline should be drawn, and the
REDRAW button will redisplay the model results with the isolines defined.
.
Map features
In addition and parallel to the isolines, selected topics (emission source
classes) can also be displayed, toggled on or off with the corresponding
buttons.
Zooming
The map display and model output overlay support arbitrary zooming. To
zoom in, the user select a point of interest, that will be in the center of the
area shown. This point is selected with a left mouse button click. Now
moving the mouse away from the point in any direction will show a
dynamically sized window (the size depends on the distance of the mouse
from the initial point of interest); another mouse click will now select the
area showen by the red box over the map, the map will be redrawn for this
area. To zoom in again, repeat the procedure; to zoom out, (one step at a
time) select the corresponding button under the map.
 Color coding
The display of model results over user selected background maps is based
on a set of color scheme definitions.
Page 86 of 95
PM3 Deliverable: A05.1
ESS GmbH
Two different systems are being used:
1. pre-defined, static arbitrary (non-linear) color ranges, defined for any
or all combination of
o pollutant;
o aggregation period (1,8,24, 8760 (annual) hours).
2. a (default) dynamic color definition.
The color are defined in a simple ASCII configuration file, /var/www/cgibin/mapcolor.
Dynamic default color range
0 0 255
The first entry in the file consists of 21 decimal,
integer RGB color definitions:
0 100 255
The interpretation is as follows: the range
0 150 255
between a user-defined minimum and
... ... ...
maximum concentration is divided into 20 equal
intervalls, each being assigned one of the first 20
255 100 0
colors; this range (in terms of minimum and
255 75 0
maximum) can be dynamically re-defined. Values
255 50 0
below the minimum are invisible (transparent);
values above the maximum (open ended) are
255 0 0
displayed in color 21.
150 0 0
Named, static color ranges
Separated by a blank record, any number of named color definitions follow.
Their first record consists of a variable (pollutant ) name, and an
aggregation period in hours:
O3_ppb 1 3
would indicate ozone in ppb, hourly values,
0 0 255 12
and a MIN (starting) concentration for the
0 100 255 24
first concentration range and color of 3. This
0 150 255 36
header is followed by any number of color
range definitions, defined by their upper limit
0 255 150 48
value:
... ... ... ...
The file consists of the three decimal, integer
255 50 0 360
RGB definition, values and the upper limit of
the corresponding color range
255 0 0 400
(concentrations lower or equal that value,
150 0 0
and larger than the previous bound, are
shown in the corresponding color. The first range start at the MIN value
defined in the header (concentrations below MIN are invisibble); The last
range is open ended.
Implementation strategy
Any model output display always shows one well defined combination of
pollutant and aggregation period: the model scenario has either a single
substance and aggregation period, or selectors for several pollutants and
periods.
Whenever there is color scheme defined for this combination of pollutant
and aggregation period, this is being used automatically to render the
results.
Page 87 of 95
PM3 Deliverable: A05.1
ESS GmbH

Color definitions starting at a lower limit of 0 (and thus covering the
entire model domain and map window) are rendered transparent as
the default;
 Color definitions starting with a lower limit >0 are rendered opaque
as the default.
If no color scheme has been defined for a given combination, the default
(dynamic) scheme is used with an initial default of MIN and MAX values
taken a second configuration file, /var/www/cgi-bin/colorranges This uses
the following format:
where the first record contains the definition of a global default, used when
the dynamic scheme is selected, but no default definition corresponding to
the current combination of substance/aggregation period can be found.
Please note that POLLUTANT_NAME is the (internal) name of the
corresponding Descriptor defined in the systems knowledge base.
Interactive selection and configuration
At the level of the scenario
display, the color bar under
the map is used as a trigger
for a color scheme selection
dialog: A pop-up window
shows the current selection:
o predefined or dynamic
o lower and upper
bounds,
o transparent/opaque
switch.
Interactive changes:
 the user can toggle the
transparent to opaque
selection;
 IF the default color scheme is pre-defined (regulatory interpretation),
the user can switch to the dynamic scheme and change the display
bounds;
 IF the default scheme is already dynamic, the display bounds can be
changed.
 Any such modification will only be valid for the current session. No
interactive changes will affect the rendering of the automatically
generated forecast and nowcast imagery at the first page, multi-time
step displays or animation.

Any long-term and
permanent re-definition
must be done by editing the
color configuration.
ANY
ANY
1
100
POLLUTANT_NAME PERIOD MIN MAX

Page 88 of 95
PM3 Deliverable: A05.1
ESS GmbH
 Output matrix export (CSV)
All spatially distributed models in AirWare produce their output as
MatrixObjects.
These matrices can be exported in CSV format using the export
button at the details model results display page.
The exported CSV file starts with 16 lines of header information with the
META data describing the exported matrix:
Source
CAMx
Resolution
1000 m
Substance
NOx
Unit
ug/m3
Start time
2008-08-25 19:00
Duration
24 hours
Interval
1 hour
Time steps
24
Layers
1
Components
1
Rows
300
Columns
300
Domain name
TEHRAN
NW corner of domain (UTM
coordinates)
x=380000 m - y=4105000
m
Domain extent in x direction
300000 m
Domain extent in y direction
300000 m
This is followed by the data ( comma separated floats ) for each grid cell.
For each time step the matrix is given in the following format (time steps
are separated by an empty line):
 Data start at the north-west corner of the domain.
 Each line corresponds to a row of the matrix containing values
ordered from west to east:
o The first value corresponds to the north-west corner of the
domain,
o The last value of the first line corresponds to the north-east
corner.
o The first value of the last line corresponds to south-west
corner of the domain,
o The last value of the last line corresponds to the south-east
corner.
Page 89 of 95
PM3 Deliverable: A05.1
ESS GmbH
 Model output animation
Wherever models produce a time series of results (24, 48, 72, or 24*365
hours) these results can be converted to and viewed as an MPG
animation, using a Java applet as player.
Creating PNG images and MPG movies from
model results matrix data
uses: mpeg2 MPEG-2
Encoder / Decoder, Version
1.2, July 19, 1996, Copyright
(c) 1996 MPEG Software
Simulation Group,
mssg@mpeg.org (author
contact),
http://www.mpeg.org/MSSG/
The utilities mpeg2encode /
mpeg2decode are used to
convert PNG into MPG.
Contains implementation of
an ISO/IEC DIS 13818-2
codec and converts
uncompressed video frames
into MPEG-1 and MPEG-2
video coded bitstream
sequences and vice versa. The codec is used by 'convert' program for
encoding/decoding. mpeg2encode and mpeg2decode binaries are located
in /usr/bin/
Databases:
 AIRMATRIX.MATRIX_DATA
 AIRMATRIX.MATRIX_METADATA

Function:
 /var/www/cgi-bin/matrixmpg.cgi
Usage: matrixmpg.cgi [-?] [-f|--db-config STRING] [-i|--matrix-id INT] [-?|-help] [--usage]
 input options: matrixID and db config file
 reads matrix MetaData for given matrixID
 sets image options (WIDTH, HEIGHT, BACKGROUND MAP,...)
 uses MatrixImageCreator() functions to create png images (source
directory mp2:/d0/acatk/libsrc/libmetamatrix/)
 creates png images for all matrix layers and timesteps
 uses background map saved as PNG images named
mapserver$IMAGEMAP in
 /var/www/html/templates/metamatrix/data
 converts png images into mpg animation using 'convert' program
 saves png files into PNG_OUT_DIR defined in MM5.MM5_scenario
(/var/www/html/templates/mm5/img)
Page 90 of 95
PM3 Deliverable: A05.1

ESS GmbH
saves mpg files into MPG_OUT_DIR defined in MM5.MM5_scenario
(/var/www/html/templates/animations/mpg)

Utility program:
 /usr/bin/convert
 converts list of png images sorted by time step in ascending order
into a mpg animation
 program belongs to package ImageMagick 5.4.7
web:http://www.imagemagick.org

ANIMATIONS DISPLAY:
 html directory: /var/www/html/templates/animations

java applet:
 /var/www/html/templates/animations/mpeg_java-3.6
 Java class files for the MPEG decoder and player:
/var/www/html/templates/mm5/php
The Java applet is installed in a php file for animations display
(var/www/html/templates/animations/php/animation_matrix.php)
The parameters that can be controlled within the applet are:
 WIDTH of animation,
 HEIGHT of animation,
(default: 50).
Animation control:
 oneleft() - one step back
 stopanim() - stop
 oneright() - one step forward
 moreright() - play animation
The movie files are in /var/www/html/templates/animations/mpg/
naming convention: movie.$matrixID.$layer.mpg
The class "MPEG_Play" is the main applet. Its work is divided into 2
phases:
1. scanning: A "ScanThread" and an "AnimatorThread" work
concurrently The "ScanThread" produces a list of frames (images)
and informes the "AnimatorThread" if a frame is decoded. The
number of the frame (timestep) is displayed under the loaded frame.
2. display: By means of the method "Element.close_chain()" the list of
frames is closed to a ring of frames. After that the frames are
displayed. The "ScanThread" dies and the "AnimatorThread" begins
to display the frames.
Page 91 of 95
PM3 Deliverable: A05.1
ESS GmbH
 Temporal aggregation
To facilitate the direct
comparison of hourly model
results with daily air quality
standards (PM10), a
temporal aggregation tool
summarizes 24 hour of
simulation in a single matrix
that can display:
o daily average
o maxima during the
day
o average above a user
defined threshold
o a non-linear exposure
function (user defined
exponent)
a secondary page (details) support the creation of user defined isolines on top
of the any of the previous pages concepts.
 Population exposure
In addition to the exposure function described above, AIrWare offers a
basic population exposure tools that is based on overlay analysis of
concentrations fields and population data.
Model results scenarios can also be interpreted in terms of population
exposure, if an appropriate data set of population data (polygon or raster
based) is available for the model domain.
The population exposure is available from every model results scenario
from the button: popexp, upper edge of the map window.
The button leads to a separate page that inherits the model scenario
including domain and color configuration. On top of the concentration (or
air quality index) display, the populated area where a substance and
aggregation time specific standard is exceeded, is shown in RED. In
addition, a numerical summary (area and number of people exposed) is
also shown under the concentration distribution histogram.
Data and files
The population distribution (density) is defined in a regular raster (default:
100 m resolution, 1 ha) generated from an appropriate shapefile, which is
also available in the embedded GIS as an overlay.
The default path for the file is:
/gis/APPLICATION/data/population.density.
The default thresholds for the computation of exposure are defined in a
configuration file, namely: /var/www/cgi-bin/exposurelimits which is a
simple ascii file with the format:
ANY
ANY
100
POLLUTANT_NAME PERIOD THRESHOLD
Page 92 of 95
PM3 Deliverable: A05.1
ESS GmbH
which defines:
 the pollutant or index
 the aggregation period
 the default threshold
where the first record with ANY, ANY defines a global default if no
specific value for a given combination of substance and aggregation
period can be found. The values are defined in the unit that is
associated with the respective pollutant or index in the Knowledge
Base.
 Receptor points and areas
Receptor areas describe arbitrary areas of special interest that can be
used for impact assessment or the identification of monitoring station
locations.
All receptor areas are OBJECTs and have the standard OBJECT META
DATA associated that include:
 name, short description, owner, creation and last modification date
of the object description data or geometry;
 Hypertext description (with import function) including optional
imagery; HTML files and images (GIF, JPEG) can be imported from
the client with a local file browser.
 Type specific attributes (Descriptors):
o Total area (used defined)
o Population (user defined)
Please note that the set of descriptive attributes is open and data driven;
the set can be extended by:
 defining an appropriate Descriptor;
 adding a reference to the object class specific TEMPLATE file
the list can be extended, fully data driven.
 Type specific geometry:
a receptor area is define by a single, arbitrary polygon that can be
imported as an ArcView Shapefile with a selected feature ID, see
below.

Geometry/Polygon import:
An import function supports the import of polygon definition as shape files
from ArcView.
The interactive polygon import dialog includes a browser option on the
local client, and requires the specification of the features ID of the polygon
boundaries in the shapefile imported.
Receptor or Building Points
Building Points describe arbitrary points of special interest in a model
domain that can be used to represent simulated monitoring stations or
receptor points for the analysis of compliance with air quality standards
but based on spatial model results rather than point observations.
Air quality models export computed concentration values for Receptor
points as time series for all model scenarios.
Page 93 of 95
PM3 Deliverable: A05.1
ESS GmbH
All Building Points are OBJECTs and have the standard OBJECT META
DATA associated that include:
 name, short descriptiopn, owner, creation and last modification date
of the object description data or geometry;
 Hypertext description (with import function) including optional
imagery; HTML files and images (GIF, JPEG) can be imported from
the client with a local file browser.
 Type specific attributes (Descriptors):
o A TYPE designation (user defined in the Descriptor:
building_type)
o
The building point is represented by an point location that the user can
define interactively on the (default) background map by selecting the
desired location with the mouse or alternatively entering the coordinated
directly.
Additional configuration options include choice of the background map
from the MapCatalog and the definition of the extent of the area around the
point to be shown.
Page 94 of 95
PM3 Deliverable: A05.1
10.
ESS GmbH
Proof of deliverable
The Reference and User Manual pages (hypertext) are embedded with the web
interface of the model system. A green button (help) on every page provides a link to
the corresponding help (manual) page. The start page (Table of Content) of the
AirWare on-line manual pages can also be reached directly at:
http://80.120.147.34/MANUALS/AIRWARE/TOC.html.
Page 95 of 95