NORPAC emission factor estimation and model exercises

NORPAC - database, emission factor estimation and model exercises



Background


At the Copenhagen meeting (March 2004), it was decided that the Nordic
model exercise during 2004-2005 should focus the street and urban
environment. Good data sets from street canyons and open roads from the
Nordic countries should be commonly used by different model
implementations. The final purpose of the model exercise is to establish
high quality emission factors for PM10, PM2.5 and UFP, taking into account
their dependence on meteorological and street related parameters. These
factors may later be used for regional modelling purposes.

Matthias K and Lars G met in Lund (August 17, 2004) to discuss further work
on the database interchange and model work. For both DMU and SMHI particle
mass (PM10, PM2.5) stand out as being more important than UFP, consequently
we will start to work on PM analysis and the determination of exhaust and
non-exhaust emission factors. Gunnar O at SMHI and Christer J at ITM have
worked on Stockholm data for this purpose and they are willing to
collaborate with their data and their experiences as part of NORPAC
activities.

The tasks to perform are thus:
. Create a database with checked and complete PM data sets from street
canyons and open roads in Stockholm and in Copenhagen (hopefully also
Oslo and Helsinki will join, Question to FMI, UH and NILU, METNO).
Complete means that except for PM10/PM2.5 there will be also NOx data, in
order to use the tracer method. Urban background and rural background
should also be given, so that the local contribution can be analysed.
Traffic counts and meteorological data are also required.


. Determine emission factors for vehicle generated PM, separating exhaust
and non-exhaust emissions.


. Using the emission factors for a model exercise.


Database overview (format, login, get started)


We suggest the use of the "NORDIC" database that has been created in Sweden
for model validation purposes. The database includes an Internet tool for
time series analysis, which means that all data can be compared and
analysed through the web (downloading of the raw data or elaborated
variables is of course possible). The data will be loaded by SMHI (mail to
gunnar.omstedt@smhi.se). Data will be loaded in local time (including the
summer time shift) with 1 hour time resolution and the time stamp
indicating the end of the averaged hour (e.g. "1400" means average between
1300 and 1400). PM10/2.5 measurements by TEOM should be provided as "direct
readings" without any factor or correction for losses due to evaporation of
volatile compounds. The units for all compounds in the database are µg/m3.
The link to the NORDIC database can be found via www.luftkvalitet.se
directly to: http://slb05.slb.nu/iairviro/ or through the NORPAC web page
link under SMHI / ITM. The username /password for each group will be
distributed by Email.
As platform for to access the database you need a newer version (e.g. 6.0,
older versions might work) of Internet Explorer and Java VM (version 1.4xx,
standard in newer IE versions, can be installed from free down load under
http://www.java.com). If you can not login to the database, the missing or
outdated Java might be the probable reason.

After login: click 'nordic' on the left panel, than click 'Indico
Presentation' on the top.
Now use left menu points from top to bottom to select stations, parameters
time period etc. you wish to plot and analyse. The output is possible as
graphic or text.
All settings and selection can be saved under 'macro' in one personal
directory per user.
All users can see and use macros (= plots, analysis etc.) of other users.
A more detailed technical description is on the way.

Detailed description of the content and some plots of data of the complete
NORDIC database can be found in a report by Lars G et al., SMHI: NORDIC -
A database for evaluation of dispersion models on the local, urban and
regional scale. (Download:
http://simair.smhi.se/luftkvalitet/nordic_rep04.pdf)
A user reference of the Indico Presentation tool for the database can be
found under: http://simair.smhi.se/luftkvalitet/indico_doc_volume12p.pdf)


Emission factors

The emission factor determination will be made on by the different groups
according to preferred methods. Recent work on separating exhaust and non-
exhaust emissions have been done at SMHI (Gunnar Omstedt), DMU (Peter
Wåhlin) and has also been extensively documented by the Lohmeyer group in
Germany (Düring et al.).

The immediate goal is to present and discuss results from this data
analysis and emission factor determination on the NORPAC general meeting
Nov 10 2004 in connection with the NOSA workshop (Stockholm Nov 11-12,
2004). In good time prior to this (ca. Middle of October), we should
exchange documents or reports on the different ways of estimating emission
factors e.g. English summaries of SMHI PM10 model (Gunnar O) and the DMU
study (Peter Wåhlin et al.). Question to FMI /NILU, please give a brief
overview of emission factors used by your groups.
At the meeting in Stockholm we will also discuss the modelling exercise.


Overview of data for the PM analysis


The basic idea is that for each country well defined data sets are included
with the following criteria (absolutely needed in black, desired in red):

. PM10, PM2.5, NOx from one or various street canyons
. PM10, PM2.5, NOx from one or various open roads
. PM10, PM2.5, NOx from one urban background station
. PM10, PM2.5, NOx from one rural background station
. Traffic volume data and vehicle speed (time series or averaged for a
typical week)
. Traffic composition (LDV_gasoline, LDV_diesel, HDV)
. Wind speed and direction, temperature, precipitation, humidity, std for
wind direction, global radiation, cloudiness, vertical temperature
gradient.

It is of interest that the data set is as "up to date" as possible,
preferably not older than 2000. The data sets will be divided into station
groups, easily separated in the internet interface, e.g. STO_street,
STO_open, COP_street etc.

For the Stockholm case ('STO_street'), we included into the database:

Hornsgatan: periods 2000 (SEC exercise) and October 2002 to June 2004.
Hourly data of:

. Traffic counts and vehicle speed (total number of vehicles per lane, in 4
lanes)
. Meteorology (wind speed, direction, std of wind direction, temperature,
humidity, global radiation, temperature gradient, precipitation)
. PM10 at Hornsgatan (northern side), Rosenlundsgatan (urban background)
and Aspvreten (rural background)
. PM2.5 at Hornsgatan (northern side), Rosenlundsgatan (urban background)
and Aspvreten (rural background)
. NOx at Horngatan (northern and southern side) and Torkel Knutsson (urban
background).

One open road/highway location : Period March - May 2003. Hourly data of:

. Traffic counts and vehicle speed (LDV and HDV per lane, 4 lanes)
. Meteorology (wind speed, direction, std of wind direction, air
temperature, humidity, temperature gradient, road temperature,
precipitation)
. PM10 from W1 (west of road), E1 and E2 (east of road), Rosenlundsgatan
(urban background) and Aspvreten (rural background)
. NOx from E1 and E2 (east of road), Torkel Knutsson (urban background) and
Norra Malma (rural background).

For the Copenhagen street case ('COP_street') we provide:

Jagtvej and H.C. Andersens Boulevard (street) from Jan 2002 to June 2004
hourly data. Continuous total traffic counts and traffic speed are also
available for Jagtvej. We will provide average traffic profiles for both
streets.
Station HCOE (urban background, roof) measures meteorology and background
concentrations.
NOx , CO, PM10 and PM2.5 are provided for the database (no PM2.5 at
Jagtvej).
Rain was measured at Jagtvej 2001-2004 and starting March 2004 data from a
road sensor at HCAB are available (street wetness, surface temperature,
salt content, ...)

Two additional Stations Lille Valby (NOx) and/or Vavihill (PM10/PM2.5)
might complement the dataset with respect to near-city /rural background
levels.