C. Cuvelier, P. Thunis and A. Gzella

1
The CityDelta project CD1,2,3 Methodology,
Results, Conclusions
C. Cuvelier, P. Thunis and A. Gzella JRC,
European Commission, Italy
CityDelta community
Cost-effective control of urban air
pollution 16-17 November, 2006 IIASA, Laxenburg,
Austria
2
Presentation Overview

• How did it start driving force
• CD1
• CD1 CD2
• O3 validation, PM10 validation
• Conclusions CD1,2
• CD2 CD3
• CD3 and FRs (Tomorrow)

3
How did it start
October 2001 CityDelta_1 kick-off workshop at
JRC CAFÉ programme launched by the Commission in
2001 and concluded in 2005 by the adoption of
Thematic Strategy on Air Pollution under the 6th
Environmental Action Programme Question Which
measures will lead to a cost-effective reduction
of air-pollution health-related problems in
European Cities? In particular problems related
to O3 and PM CityDelta Objective How to
include sub-grid effects into a Europe-wide
health impact assessment for O3/PM? First idea
was to study episodic situations peak values O3
and PM However
4
WHO Systematic review of health aspects of air
pollution in Europe
5
CityDelta1

• A model inter-comparison exercise for
  urban-regional dispersion models focusing on 8
  European cities to identify
• the systematic differences (deltas) between
  rural and urban background AQ (Scale),
• how these deltas depend on emissions
  (Emissions),
• how these deltas vary across cities
  (Cities),
• how these deltas vary across models (Models)
• how these deltas vary for O3 and PM
  (Pollutants).

6
CityDelta Model validation

• What are the skills of state-of-art models in
  simulating
• air quality within European cities and in
  their vicinity?
• How do model skills depend on the city and how
  large is
• their variability amongst models?
• How do the skills of regional-scale models, with
  a typical
• resolution of 50 km, compare to city-scale
  models, with a
• typical resolution of 5 km
• Do models have sufficient skills to be used for
  the evaluation
• of emission-regulation policies at the scale
  of the continent
• and its main cities?

7
8 Cities for CityDelta1
London Paris Prague Berlin
Copenhagen Katowice Milan Marseille
8
8 Emission Scenarios for CD1
0 --- 1999 1 --- 2010 CLE Current
Legislation 2 --- 2010 NOx MFR Maximum
Feasible Reduction 3 --- 2010 NOx (CLEMFR)/2
4 --- 2010 VOC MFR 5 --- 2010 NOX and
VOC MFR 6 --- 2010 PMcoarse MFR 7 ---
2010 PM2.5 MFR

• Meteo 1999 provided by Meteo-France (Aladin 10
  km) or calculated.
• Boundary conditions provided by EMEP or
  calculated.
• Long term simulations full year for PM, 6
  months for O3
• Outputs delivered with resolution of 5-10 or 50
  km

9
CityDelta1 18 Model configurations 7
regional-scale, 11 urban-scale
10
CityDelta1 CityDelta2

• CityDelta1 was focused mainly on O3
• Inconsistencies in the emission inventories
• Not enough PM10 observational stations for
  validation
• Not all models were able to perform the PM
  calculations

• For CityDelta 2
• Only 5 models participated in the 2nd phase
• CHIMERE, LOTOS, REM-CALGRID, OFIS, CAMx
• 4 cities were selected Berlin, Milan, Paris,
  Prague
• Base year 2000, related to the availability of
  Observations
• Scaling of the city-emission inventory per EMEP
  grid-cell,
• per SNAP sector, per pollutant
• Focus on PM10, and PM2.5

11
Emission Inventories Local vs Regional
Milan
Emission densities (All Domain) Milan
20
18
High source
Low source
Traffic
16
14
2
12
10
Tons/km
8
6
4
2
0
City
City
City
City
City
EMEP
EMEP
EMEP
EMEP
EMEP
NOx
COx/2
PM105
VOC
SOx

• NOx, CO, SOx, VOC estimates seem quite robust
• PM estimates show 40-50 differences.

• CityDelta has also contributed to a
  considerable revision of
• the regional-scale emission data

12
CityDelta2 OBS stations
Berlin German Federal Environmental
Agency Milan Regione Lombardia
Terraria Paris AirParif Prague Czech
Hydrometeorological Institute
13
4 Cities for CityDelta2
London Paris Prague Berlin
Copenhagen Katowice Milan Marseille
14
7 Emission Scenarios for CD2
0 --- 2000 2000 1 --- 2010
CLE 2010 CLE 2 --- 2010 MFR 2010
MFR 3 --- 2010 MFR1 (NOx, SO2, PPM) 2010
MFR1 4 --- 2010 MFR2 (NMVOC, NH3, CO) 2010
MFR2 5 --- 2010 MFR1 (City domain
only) 2010 CLE 6 --- 2010 MFR2 (City domain
only) 2010 CLE
15
CityDelta1,2 O3 validation
OBS grey Left LS Right FS
Annual Mean O3 ppb
16
CityDelta1,2 O3 validation (Taylor)
Standard deviation, Correlation Coeff, and CRMSE
in one diagram
17
CityDelta1,2 O3 validation (Taylor)
Annual Mean O3 ppb
18
CityDelta1,2 O3 validation
OBS grey Left LS Right FS
Sum of max daily 8-hour mean O3 concentrations
over 35 ppb, calculated over the entire year
19
CityDelta1,2 O3 LS-FS variability
LS FS
O3 ensemble
20
CityDelta1,2 Conclusions(1)

• O3
• All models have reasonable skills to simulate
  the concentrations in both
• city centres and rural areas located nearby
• Successful simulation of daily-max, 95
  percentile and SOMO35
• differences between Milan and the N-European
  cities
• Better simulation in rural than in urban areas
  (effect of titration, mixing
• near sources, meteorology)
• Most models overestimate mean concentrations in
  city centres,
• no bias is observed for Ox O3 NO2
• Models capture relative variations of daily
  averages and maxima, with
• time correlation coefficients up to 0.8
  (day-time)
• Overestimation of daily averages and maxima is
  due to the difficulty in
• simulating the lower ozone values

21
CityDelta1,2 PM10 validation
OBS grey Left LS Right FS
Annual Mean PM10 µg/m3
22
CityDelta1,2 PM10 validation
Annual Mean PM10 µg/m3
23
CityDelta1,2 Winter PM10 LS-FS
LS FS
Annual Mean PM10 µg/m3
24
CityDelta1,2 Conclusions(2)

• PM10
• Simulation of PM10 concentrations by the models
  is rather poor
• Large-Scale models largely underestimate the
  total mass, this bias is
• reduced for Fine-Scale models
• The difference between PM10 concentrations in
  Milan and North-
• European cities is not very clear
• Correlation coeff are of the order of 0.6-0.7
  (smaller than for O3)
• No PM2.5 validation lack of monitoring data

25
CityDelta1,2 Conclusions(3)

• General
• Models are skillful for simulation of ozone, but
  fail to provide an accurate
• prediction of PM10 (uncertainty in knowledge,
  dynamical/chemical
• processes, emission processes, obs,
  susp./resusp. of particles, )
• Fine-Scale models show better performance for
  PM10 in the cities, and have
• better treatment of the titration effects
• Probably more improvement when using small-scale
  meteorology, and
• increased vertical resolution
• A further increase in horizontal resolution, to
  1x1 km, may be necessary to
• increase the model skills
• No model is superior to the others (the best one
  is ENS!)
• Hypothesis The ensemble of models gives a
  representation of the
• uncertainty of our knowledge

26
CityDelta1,2 Emission reduction
LS FS
Impact of emission-reductions on ensemble LS and
FS
27
CityDelta2 CityDelta3

• The preliminary FRs based on CityDelta2 need
  improvement
• Other city characteristics than only wind speed
  and emission
• densities need to be taken into account
• The models have undergone further development
  (PM)
• Monitoring stations for PM2.5 are available

• For CityDelta 3
• Focus on PM2.5 for FR (? PPM2.5 ? Emiss)
• 3 models participated in the 3rd phase
• CHIMERE, REM-CALGRID, CAMx (and TCHAM)
• 7 cities were selected Berlin, Lisbon, London,
  Krakow, Milan,
• Paris, Prague

28
7 Cities for CityDelta3
London Paris Prague Berlin
Copenhagen Katowice Milan Marseille
29
8 Emission Scenarios for CD3

• Meteo 2004 provided by INERIS (MM5 run)
• Boundary conditions provided by INERIS

30
CityDelta3 Monitoring Stations
BERLIN
LONDON
KRAKOW
PARIS
PRAGUE
MILAN
LISBON
31
CityDelta3 PM2.5 annual vs OBS
01CAMx 02REM 03CHIMERE
Prague-(SUB)
London-(UB)
Prague-(UT)
Prague-(UT)
Prague-(UT)
London-(UT)
32
http//aqm.ei.jrc.it/citydelta

• Main
  Publications
• Overview
• C. Cuvelier et al., CityDelta A model
  intercomparison study to explore the impact
• of emission reductions in European cities
  in 2010. To appear Atm. Env, 2006
• Validation
• R. Vautard et al., Evaluation and
  intercomparison of Ozone and PM10 simulations
• by serveral chemistry transport models over
  four European cities within the
• CityDelta project. To appear Atm. Env, 2006
• Delta
• P. Thunis et al., Analysis of model responses to
  emission-reduction scenarios within
• the CityDelta project. To appear Atm. Env,
  2006
• CD-rom containing all CityDelta 1 and 2 results
• Many other contributions to Workshops,
  Conferences (oral and poster presentations)

33
(No Transcript)