PROJECTIONS OF AIR QUALITY IN EUROPE: A TWO-WAY APPROACH 1 Institute for Environmental Research and Sustainable, Development, National Observatory of Athens, Greece 2 Division of Environmental Physics and Meteorology, National and Kapodistrian University of Athens, Greece K.V. VAROTSOS 1, 2 , C. GIANNAKOPOULOS 1 , M. TOMBROU 2
25
Embed
PROJECTIONS OF AIR QUALITY IN EUROPE: A TWO-WAY …acmg.seas.harvard.edu/geos/.../2011/pdf/Day2/...pc.pdf · PROJECTIONS OF AIR QUALITY IN EUROPE: A TWO-WAY APPROACH 1 Institute for
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
PROJECTIONS OF AIR QUALITY IN EUROPE:A TWO-WAY APPROACH
1 Institute for Environmental Research and Sustainable, Development, National Observatory
of Athens, Greece2 Division of Environmental Physics and Meteorology, National and Kapodistrian University
of Athens, Greece
K.V. VAROTSOS1, 2, C. GIANNAKOPOULOS1, M. TOMBROU2
2 approaches
based on empirical relations between ozone and atmospheric historical measurementscombined with future projections
based on climate-chemical models for various greenhouse gas emission scenarios
CLIMATE CHANGE IMPACT ON AIR-QUALITY
(Jacob and Winner, 2009)
MOTIVE:
CLIMATE CHANGEMETEOROLOGICAL
CONDITIONSAIR-QUALITY
(Ο3)
Question : Potential increase of Tmax (Climate Change) Ο3 ?
1st Approach
Recent Studies:
Ο3 sensitivity to meteorology
• temperature, (Jacob et al., 1993;Silmann et al., 1995)• solar radiation, (Ordonez et al., 2005)• number of days since the last frontal passage,
(Wise and Comrie, 2005)• humidity , (Camalier et al., 2007)• frequency of summertime mid latitude cyclones
(Leibensperger et al., 2008)
Data– Study domain
Field measurementsOzone:daily maximum of 8-h running average concentrations from 47 non-urban stations in Europe (EMEP)
Temperature: 2 periods of daily maximum surface temperatures from the E-OBS gridded dataset (Gridded dataset derived through interpolation of station data), for each ozone station closest grid point.
• 1961-1990, for validation purposes with the RACMO2 model• same period to each ozone site year range observations
Regional Climate Model -RACMO2 (KNMI, ENSEMBLES)Simulation periods for each ozone station closest grid point.
• 1961-1990 period for evaluating the model performance compared to thegridded maximum temperatures
• 2021-2050 and 2071-2100, based on the IPCC SRES A1B scenario .
E-OBS gridded dataset is on the same grid with the Regional Climate Model (horizontal resolution 22km x 22km
1st Approach
Table 1.Stations and the year range used in the analysis. See location of the
stations in Fig. 1
Figure. Locations of the ozone stations (red) and their closest grid points (black) for
temperature used in the analysis. See the correspondence between numbers and names of
the stations in Table 1. For stations where only black squares are visible the station location
coincides with the closest grid point of the gridded maximum temperature
Station Code Station Name Altitude (m a.s.l) Year Range Station Type
1.AT02 Illmitz 117 1995-2004 rural
2.AT04 St. Koloman 851 1995-2004 -
3.AT05 Achenkirch 960 1995-2004 mountaineous
4.AT30 Pillersdorf 315 1995-2004 rural
5.AT32 Sulzberg 1020 1995-2004 -
6.AT33 Stolzalpe 1302 1995-2004 -
7.AT45 Dunkelsteinerwald 320 1995-2004 -
8.AT46 Gaenserndorf 146 1995-2004 -
9.BE01 Offagne 420 1991-2002 -
10.BE32 Eupen 295 1991-2002 -
11.BE35 Vezin 160 1991-2002 -
12.CH02 Payerne 489 1993-2003 rural
13.CH03 Tänikon 539 1993-2003 rural
14.DE02 Langenbrügge 74 1992-2001 forestry
15.DE03 Schauinsland 1205 1992-2001 forestry
16.DE04 Deuselbach 480 1992-2001 rural
17.DE05 Brotjacklriegel 1016 1992-2001 forestry
18.DE07 Neuglobsow 62 1992-2001 forestry
19.DE08 Schmücke 937 1992-2001 forestry
20.DE12 Bassum 52 1992-2001 rural
21.DE26 Ueckermünde 1 1992-2001 rural
22.DE35 Lückendorf 490 1992-2001 rural
23.ES01 Toledo 917 1993-2000 rural
24.ES04 Logrono 370 1993-2000 rural
25.FR08 Donon 775 1998-2005 forestry
26.FR09 Revin 390 1998-2005 forestry
27.FR13 Peyrusse Vieille 236 1998-2005 rural
28.BG02 Eskdalemuir 243 1991-2002 rural
29.BG06 Lough Navar 126 1991-2002 rural
30.BG13 Yarner Wood 119 1991-2002 rural
31.BG14 High Muffles 267 1991-2002 rural
32.BG15 Strathvaich Dam 270 1991-2002 rural
33.BG31 Aston Hill 370 1991-2002 rural
34.BG32 Bottesford 32 1991-2002 rural
35.BG33 Bush 180 1991-2002 forestry
36.BG36 Harwell 137 1991-2002 rural
37.BG37 Ladybower 420 1991-2002 rural
38.BG39 Sibton 46 1991-2002 rural
39.GR01 Aliartos 110 1996-2005 rural
40.IT01 Montelibretti 48 1995-2004 -
41.IT04 Ispra 209 1995-2004 rural
42.NL09 Kollumerwaard 1 1991-2002 -
43.NL10 Vreedepeel 5 1991-2002 -
44.PL02 Jarczew 180 1995-2004 rural
45.PL03 Sniezka 1604 1995-2004 -
46.PL04 Leba 2 1995-2004 rural
47.PL05 Diabla Gora 157 1995-2004 rural
METHODOLOGY
• Identify the ‘representative stations’ using rotated PCA and correlation analysis betweenthe daily maximum 8-h average ozone concentrations and the observed daily maximumtemperature
• Derive threshold temperature for ozone exceedances.
• Construct an empirical-statistical model, based on the probability distribution of dailymaximum 8-h average ozone concentration with daily maximum temperature.
Bins of 1oC above the pre-calculated threshold for temperature and bins of 5 ppb for ozoneconcentration, are calculated.
• Evaluate the spatial temperature behavior of RACMO2 vs EOBS for the 1961-1990
• Apply the empirical statistical model with the future modeled temperatures.
1st Approach
exceedance days are days with daily maximum 8-h average >= 60 ppb
OZONE EXCEEDANCE SEASON – PCA – ‘REPRESENATIVE STATIONS’
• PCA implemented on the ozone exceedance season (1 April- 30 September)
• 5 PCs (5 sub regions) explaining ~71% of the variability in the daily maximum 8-h average ozoneconcentrations
PC
Number
Sub
region
Variance
explained
(%)
Mean
NOx
emissions
NO2
Equivale
nts (Mg)
Mean
NMVOC
emissions
(Mg)
Mean
ozone
exceedan
ce days
(%)
1 South
east
37.21 3241 3975 14
2 North
west
12.45 5041 10268 2
3 South
west
11.13 3431 4963 13
4 Central
north
6.51 3831 3656 8
5 North
east
3.53 2465 1922 8
Table 2. List of the sub regions identified by PCA, variance
explained by each principal component, mean NOx, mean
NMVOC and mean percentage ozone exceedance days for the
stations constituting each sub region.
• 2 ‘representative stations’ from each sub region selection based on communality(red X) and correlation coefficient between ozone and temperature (black X)
• meteorology plays a dominant role regarding the stations grouping
X
X
X
X
XX
X
X
X
X
ESTIMATED FUTURE O3 PROBABILITY DISTRIBUTIONFUTURE EXCEEDANCE DAYS
Southeast Subregion
Northwest Subregion
Southwest Subregion
Centralnorth Subregion
ESTIMATED FUTURE O3 PROBABILITY DISTRIBUTIONFUTURE EXCEEDANCE DAYS
• Future ozone probability distributions follow the shape pattern of the observed.
• A temperature increase of about 26% in the future could lead to an increase of about 24 extra ozone exceedance days/year at polluted stations.
• Although this relation can been seen as a starting point for how the future atmosphere will behave it contain a caveat: it assumes constant emissions of ozone precursors
2ndAPPROACH Giss/Geos-Chem global GCM-CTM modelling system
• High correlation coefficients was found between the modelling system and observations for both O3 and Tmax
• model mainly underestimates Tmax compared to the observations
• model overestimates O3 compared to the observations especially in the warm months
• results from the future scenarios simulations revealed the most profound impact of climate change will take place in Southern Europe
Empirical statistical model based on the O3 – T relation by means of probability distributions
FUTURE WORK
• Examine GISS meteorology• update emissions inventories• simulations with a finer resolution and/or under different future emissions
scenarios
NKUA-NOA Group Poster by Anna Protonotariou
Title: European CO budget : regional sources and its links with synoptic circulation and long range transport
Extra slides
Relationship of ozone exceedance days with daily maximum temperature.
Figure. Probability that the daily maximum 8-h average
ozone will exceed 60 ppb for a given daily maximum
temperature for the stations selected from each sub
region.
Tthresh P(ozone exceedance days) > 4%
Evaluation of the Regional Climate Model using gridded temperature
observationsSpatial Evaluation
Figure Difference between the RCM and the EOBS for the 1961-1990 period.
•small underestimation on the west
•small overestimation on the east
RACMO2 - EOBS
Regional Climate Model Future Simulations
Figure. Mean differences between the 2021-2050 future simulation (left), the 2071-2100 future simulation (right) and the 1961-1990period for the Average Annual maximum temperature.
Both periods are warmer compared to the 1961-1990 period with the 2071-2100 period
exhibiting higher daily maximum temperatures than the 2021-2050 period (~2 oC)
2021-2050 – reference period 2071-2100 – reference period