The PANDA Project Guy P. Brasseur Jan. 2015 Objective of the PANDA - - PowerPoint PPT Presentation

SLIDE 1

The PANDA Project

Guy P. Brasseur

• Jan. 2015

SLIDE 2

Objective of the PANDA Project

• To establish a team of European and Chinese

scientists who will jointly use space

• bservations and in-situ data as well as

advanced numerical models to monitor, analyse and forecast global and regional air quality.

• PANDA will disseminate methodologies, tools

and data to a variety of users.

SLIDE 3

Specific Goals (1)

• Identify space remote sensing data that are available

and could be used to better monitoring air quality (ozone and its precursors, aerosols) at the global and regional scales;

• Improve remote sensing data where needed, taking

into account specific conditions in Asia;

• Identify and collect surface in-situ observations that

are available and will complement space observations;

• Improve and evaluate the current knowledge on

anthropogenic and natural emissions in Asia and at the global scale;

• Use remote-sensing and in-situ data to analyse specific

air quality situations, using state-of-the-art global and regional chemical transport models.

SLIDE 4

Specific Goals (2)

• Conduct specific air quality analyses in urbanized areas of the

Asian continent and assess the importance of intense emissions in the populated areas of Beijing, Shanghai and Guangzhou;

• Develop a methodology to downscale existing global scale

predictions of air quality (specifically those provided by the EU MACC-II Project) for the East Asian region, including the populated areas of China;

• Quantify the impact of regional air pollution emissions in Asia on

the rest of the world in response to long-range transport;

• Develop user-friendly toolboxes providing easily accessible data

related to air quality;

• Disseminate information and products to users and stakeholders,

including members of national, regional and local environmental agencies and other decision makers

SLIDE 5

The PANDA Project

Coordinator: Guy Brasseur Deputy Coordinator: Prof. Xuemei Wang Period: Jan 2014 - Dec. 2016 Budget: 2 Millions Euros

SLIDE 6

The Analysis and Prediction System

SLIDE 7

SLIDE 8

Satellite Ground based Assimilation Consistent fields Model

Integration and Prediction

Toolbox for communicating air pollution predictions

SLIDE 9

Modeling Strategy

Regional

Target ~15km

• EMEP (MET NORWAY)
• WRF-CHEM (MPI)
• CHIMERE-ECMWF (KNMI)
• LOTOS-EUROS-ECMWF (TNO)
• IFS/C-IFS not all species (ECMWF)

Global

Target ~50km

• EMEP (MET NORWAY)
• IFS/C-IFS (ECMWF)
• MOZART (NUIST)
• MOZART (CNRS)

Sub-regional

Beijing / Tianjin Shanghai / YRD Guangzhou / PRD Target ~5km

• WRF-CHEM E China (SCUEM)
• WRF-CHEM (SESE-SYSU)
• Enviro-HIRLAM (DMI)
• AURORA (VITO)

Boundary Condition Boundary Conditions

Urban

Shanghai? Target ~2km

• enviro-HIRLAM (DMI)
• WRF-CHEM (MPI-M)

Boundary Conditions

SLIDE 10

Carbon Monoxide Nitrogen Dioxide

Global Forecast of Air Quality for Monday 26 January 2015

EU MACC Project ECMWF

Aerosol

SLIDE 11

Study periods

Year 2010: Global

• ~2 months January 2010 and July 2010 : Regional
• ~2 weeks within the selected months : Local

Year 2013: As far as possible (emissions… observations...):

• ~2 months January + December 2013 : Global+Regional
• ~2 weeks within the selected months : Local

Forecasts:

• MACC (global) already available and can be used for

e.g. demo of AQI on toolbox

• besides, main strategy will develop later based on the

experience with coupling/nudging/assimilation

SLIDE 12

Topics: engineering/applied and science

• Sensitivity to emissions: high vs low anthropogenic

(uncertainty ), top-down / inverse emissions from MarcoPolo, policy oriented / mitigation scenarios and their impacts?

• (lateral) boundary conditions, coupling of chemical and

aerosol schemes between global / regional / local configurations

• Assimilation and nudging approaches
• NO2 in winter, CO underestimation
• Link aerosol/haze/meteorology, link with WMO/WGNE

and COST.

• Field campaign (June 2014)

SLIDE 13

Assimilation in MACC

WRF-Chem prediction 20x20km MACC forecast/reanalysis as IC & BC WRF-Chem 60x60km 20x20km

Downscaling to Regional Scale in Asia

7 x 7 km Satellite data

Air Quality Index (AQI) WRF-Chem prediction AQI

SLIDE 14

Choice of Initial Conditions for the Regional Downscaling

MACC surface CO analysis MOZART surface CO MOZART surface O3 MACC surface O3 analysis

SLIDE 15

CO with HTAPv2 emissions CO with MACCity emissions

20x20k m WRF-Chem CO 60x60km

Observations WRF 60x60km WRF 20x20km WRF 20x20+htap

Downscaling of CO with different surface emissions: HTAPv2 (left) MACCity (right)

SLIDE 16

The Workpackages

SLIDE 17

SLIDE 18

WP-1 Remote Sensing Data

SLIDE 19

Satellites used in the PANDA Project

SLIDE 20

WP1: Remote Sensing Observations

T1.2 - First version of historical satellite data set (Lead: IUP-UB; ISM, CNRS, ULB) [Months: 1-12 D1.2]

• Full dataset of O3 and CO vertical profiles (1-km vertical gridding) from the existing FORLI processing chain

(Sept. 2007 - Dec. 2013) (including averaging kernels and error covariance)

Monitoring pollution in Asia with IASI - Achievements so far CO SO2

Boynard et al, GRL 2014 FORLI-CO: Available on French database Ether - Cfr. D1.1 Report

• Extreme concentrations in

several pollutants (CO, SO2, NH3)

• Temperature inversion 

increased sensitivity of IASI at surface level

• NH3 total columns with suitable error estimates
• Ensure that the dataset is suitable for validation activities

Pollution event as seen by IASI (January 12, 2013)

SLIDE 21

IASI tropospheric O3

Tropospheric ozone in urban and rural regions in China

Courtesy S. Saffedine

• Seasonality in tropospheric
• zone in several urban centres

and surrounding rural areas, including in East Asia

Monitoring pollution in Asia with IASI - Achievements so far

FORLI-O3: Available upon request (ULB/LATMOS) - Cfr. D1.1 Report

Population density

• Trop. O3 col. (DU)

SLIDE 22

Courtesy S. Saffedine

Monitoring pollution in Asia with IASI - Achievements so far Comparisons between ground-based measurements and WRF-Chem in China (7 stations in North China plain, 13 stations in the Pearl River delta)

Courtesy S. Saffedine

SLIDE 23

Comparisons with ground-based measurements in China

IASI NH3 total columns

NH3 emission in 1 km grid cell (kg/yr) (Huang et al., GBC 2012) 0.05ºx0.05º grid (mg/m²) (Van Damme et al., ACP 2014) Shangzhuang site in the North China Plain

IASI Ground-based meas.

IASI NH3: Available upon request (ULB/LATMOS) - Cfr. D1.1. Report Quick look available on Ether (Van Damme et al., AMTD 2014)

Monitoring pollution in Asia with IASI - Achievements so far

SLIDE 24

IUP Bremen PANDA contribution

Work on deliverables:

• Report on satellite data availability
• Preparation of new UV/vis satellite data
• Planning and preparation of first summer school
• Creation of NO2 validation data set

NO2 above EC China

• Clear decrease in 2014
• Technological

improvements?

• Less coal burnt?
• Less economic

growth?

SLIDE 25

Trends in NO2 tropospheric columns from CAM-Chem and different satellite observations.

SLIDE 26

Deliverables WP 1

SLIDE 27

WP-2 In-situ Observations

SLIDE 28

Ozone weekend effects in the Beijing–Tianjin–Hebei metropolitan area, China

• The ozone weekend effect (OWE) was first investigated in the metropolitan area of Beijing–Tianjin–Hebei

(BTH), China, using in situ measurements from the Atmospheric Environment Monitoring Network from July 2009 to August 2011.

• An obvious weekly periodical variation in the surface ozone concentration, with a lower ozone concentration

from Wednesday to Friday (weekday) and a higher concentration from Saturday to Monday (weekend);

• Y. H. Wang, B. Hu, D. S. Ji, Z. R. Liu, G. Q. Tang, J. Y. Xin, H. X. Zhang, T. Song, L. L. Wang, W. K. Gao, X. K. Wang, and
• Y. S. Wang（2014） Ozone weekend effects in the Beijing–Tianjin–Hebei metropolitan area, China。 Atmos. Chem.

Phys., 14, 2419-2429, 2014

UV radiation Ratio of VOCs/NOx A clear weekly cycle in the fine aerosol concentration was observed, Higher concentrations of aerosol on weekdays can reduce the UV radiation flux by scattering or absorbing, which leads to a decrease in the ozone production efficiency O3 PM Concentrations A smaller decrease in volatile organic compounds (VOCs; using CO as a proxy) and much lower NOx concentrations

• n the weekend may lead to higher

VOC/NOx ratio, which can enhance the

• zone production efficiency in VOC-limited

regime areas.

SLIDE 29

Deliverables WP 2

SLIDE 30

WP-3 Anthropogenic and Natural Emissions

SLIDE 31

CO emissions in China from 1980 to 2010

SLIDE 32

Collection of bottom-up emission inventories over China

Inter-comparisons of emission estimates in 2008

• Prefer MEIC and REAS, too high estimates of EDGAR

Available emission inventories:

• MEIC v.1.0
• EDGAR v4.2
• REAS version 2
• NH3 emissions from PKU (Peking Univ.)

SLIDE 33

Comparisons of bottom-up emission inventories over China –power plants

Comparisons of MEIC, REAS and EDGAR on power plants

• Prefer MEIC, which

compiled based on unit- based methodology Ratio of SO2 to CO2: Decreasing along with CO2 emissions in MEIC, reflecting the high FGD implementation on large units

SLIDE 34

Deliverables WP 3

SLIDE 35

WP-4 Integration of Observations and Models

SLIDE 36

Implement 2 Ozone-Tagging Methods into MOZART-4

Hou & Zhu, 2014, Atmos. Environ.

Wang(1998a, 1998b), Sudo (2007):The

Chemical tendency of ozone produced in the region i is given as:

       

, , , , , , , ,

i i i

dQ x y z P x y z x y z Q x y z dt   

   

, , , , P x y z Pi x y z      

: inside region i : outside region i

Qi : mixing ratio of O3 tagged by the region I β: the chemical loss rate constant Pi the gross chemical production within the region i.

Tagging ozone by regions Tagging Ozone by Species

Emmons (2012) : quantify contributions of

• zone by “tagging” emissions of NO and its

resulting products and following them to the production of ozone.

SLIDE 37

Reasons of spring O3 peak in mid-latitude east Asia Pacific rim

mostly due to regional photochemical productions — 2 tagging agree well O3 contribution rate of region-tagging

8ppb

Regions in MOZART

O3 contribution of NO-tagging ,China

8ppb

NO from China

SLIDE 38

Downscaling in the Pearl River Delta

• Observations of meteorological elements and

air pollutants over PRD are compared with WRF‐Chem simulations for cases in January and July 2010.

• The resolutions of three-nesting domains for

WRF-Chem are 45 km, 15 km and 3 km,

• respectively. The temporal variations for O3

and PM10 and the spatial distributions for SO2, NO2, PM10 and O3 are presented.

SLIDE 39

Physical options Chemical options Microphysics Lin et al. Chemical mechanism RACM Long wave RRTM Aerosol mechanism MADE/VBS Short wave Goddard Photolysis option F-TUV Urban physics BEP Land-surface Noah LSM Boundary -layer MYJ

Grid resolution 45 km, 15 km, 3 km Center point Guangzhou (23.5°N,113.7°E) Top of the model 50 hPa Vertical layers 24

WRF/Chem Model Settings

SLIDE 40

Time serious of daily average PM10 and O3 (2010.1.12~2010.1.21) January 2010

Sites OBS. SIM. MB MAGE RMSE r LH 156 109

• 47

54 58 0.85 WQS 129 62

• 67

67 79 0.56 TH 74 21

• 54

54 61

• 0.24

LY 72 118 46 48 62

• 0.22

TJ 86 32

• 54

54 62

• 0.28

JJJ 150 88

• 61

61 70 0.80 HJC 183 178

• 5

36 48 0.74 DH 118 165 46 54 67 0.71 CZ 196 94

• 102

102 131 0.86 XP 114 60

• 54

55 64 0.32 JGW 98 20

• 77

77 80 0.24 HG 129 302 173 173 188 0.81 ZML 102 84

• 18

51 61 0.08 Average 124 103

• 21

68 79 0.40 Sites OBS. SIM. MB MAGE RMSE r LH 125 78

• 47

47 55 0.48 WQS 165 96

• 69

85 101

• 0.47

TH 177 74

• 103

103 115 0.15 LY 104 51

• 53

61 65

• 0.07

TJ 146 100

• 46

59 78 0.61 JJJ 117 100

• 17

42 51 0.05 HJC 110 83

• 27

29 36 0.69 DH 129 95

• 34

47 54 0.24 CZ 94 99 5 28 34 0.48 XP 119 64

• 55

62 65

• 0.36

JGW 147 78

• 69

74 79

• 0.30

HG 129 79

• 50

52 54 0.33 ZML 108 95

• 12

44 66

• 0.42

Average 128 84

• 44

56 66 0.11

PM10 O3

SLIDE 41

The average wind field, SO2, NO2, PM10 and O3 horizontal distribution during day time in Pearl River Delta region 2010.1.12~2010.1.21（07:00~18:00LST）

January 2010

SLIDE 42

Deliverables WP 4

SLIDE 43

WP-5 Toolbox Development

SLIDE 44

The PANDA Toolbox

• ‘Delivered’ in December 2014 as D5.1
• To be further developed during the

lifetime of PANDA!

• Currently located at

http://www.marcopolo-panda.eu/

under menu item RESULTS

• Possible collaboration between

MarcoPolo and PANDA ( the Panda-MarcoPolo toolbox?)

SLIDE 45

Deliverables WP 5

SLIDE 46

WP-6 Cooperation Dissemination and Capacity Building

SLIDE 47

SLIDE 48

SLIDE 49

PANDA Summer School: 23-29 August 2015

Open to all PANDA, MarcoPolo and MACC-III partners

Remote Sensing of the Atmosphere, Emissions and Modeling

Institute of Environmental Physics (IUP), University of Bremen Applications: http://www.marcopolo-panda.eu/summer-school/european- summer-school/ Programme available on the website Deadline: February 15th, 2015

SLIDE 50

Deliverables WP 6

SLIDE 51

Management

SLIDE 52

Management Board

• A Management Board of 5 members will be

responsible for the daily management of the Project, monitor progress on a frequent basis, conduct periodic teleconferences with the partners, prepare the scientific and financial reports for the Commission and for the General Project Assembly.

SLIDE 53

Advisory Board

• A small Project Advisory Board will guide the

Project and advise the Management bodies including the project General Assembly and the Management Board.

• The Board will meet once a year. It elects its
• Chairperson. Members of the Management

Board will attend meetings of the Advisory Board and if requested, may provide input to the discussions.

SLIDE 54

Advisory Board

SLIDE 55

Educational Board

The project will establish a strong educational program that will support targeted schools for the potential users in China, and also in

• Europe. Exchanges of students and junior scientists will contribute

to this dissemination effort. A committee of 5 persons (2 from China and 3 from Europe) will be established to organize the meetings related to educational and dissemination activities.

SLIDE 56

User Interface Panel

• A User Interface Panel will develop working links between

the PANDA partners and potential users.

• Users are representatives of the scientific community,

European and Chinese space agencies, international

• rganizations specialized in data management and

dissemination, air quality forecasting groups, national and local environmental agencies, members of municipalities and provincial administrations as well as climate services.

• Representatives of other sectors including the health

sector will also be invited.

• The Panel will meet two times during the project in

conjunction with the summer schools organized by PANDA. It will be co-chaired by the Coordinator and the Deputy Project Leader.

SLIDE 57

Thank You

SLIDE 58

SLIDE 59

O3 with HTAPv2 emissions O3 with MACCity emissions

WRF-Chem O3 60x60km 20x20km

O3 diurnal cycle (Pudong)

Observations WRF 60x60km WRF 20x20km WRF 20x20+htap

Downscaling of ozone with different surface emissions: HTAPv2 (left) MACCity (right)

SLIDE 60

8h max

R=0.69

Pudong, Shanghai

SLIDE 61

• ---- Observations
• ---- 60km + MACCity
• ---- 20km + MACCity
• ---- 20km + HTAPv2

O3 in Beijing PM2.5 in Beijing PM10 in Beijing

Evaluation of the Prediction in Beijing

PM10 PM2.5 Ozone

Beijing January 2010

SLIDE 62

July 2010 Modeling period：2010.6.28 00:00~2010.7.5 00:00（ UTC） Evaluation period：2010.6.30~2010.7.4（daily mean） Evaluation meteorological sites：17

Variables OBS. SIM. MB MAGE RMSE r T at 2 m（℃） 29.9 28.6

• 1.3

1.3 1.4 0.57 WSP at 10 m（m/s） 2.3 2.6 0.3 0.9 1.1 0.09 Vapor pressure （hPa） 30.3 30.3 0.0 0.8 0.9 0.51 Relation humidity at 2 m （%） 73.5 80.4 7.0 7.2 7.7 0.47 Variables OBS. SIM. MB MAGE RMSE r T at 2 m（℃） 14.8 14.7

• 0.1

1.6 1.9 0.91 WSP at 10 m（m/s） 1.9 3.2 1.2 1.5 1.7 0.70 Vapor pressure（hPa） 12.0 10.6

• 1.5

2.0 2.3 0.96 Relation humidity at 2 m （%） 69.0 58.5

• 10.5

12.6 14.8 0.87

January 2010 Modeling period：2010.1.10 00:00~2010.1.22 00:00（ UTC） Evaluation period：2010.1.12~2010.1.21（daily mean）

• OBS. is the average of observation, SIM. is the average of simulation, MB is

mean bias, MAGE is mean absolute gross error, RMSE is root mean square error, r is correlation coefficient.

SLIDE 63

Sites OBS. SIM. MB MAGE RMSE r LH 58 153 95 95 101

• 0.65

WQS 31 118 87 87 89 0.37 TH 58 60 2 8 8 0.23 LY 20 165 144 144 145

• 0.25

TJ 12 12 7 7

• 1.00

JJJ 38 108 70 70 71 0.09 HJC 87 104 16 16 22 1.00 DH 20 244 224 224 227 0.74 CZ 49 94 45 45 50

• 0.12

XP 44 100 56 56 56 0.74 JGW 44 23

• 21

21 25

• 0.33

HG 31 300 269 269 271 0.23 ZML 30 68 37 37 39

• 0.05

Average 40 119 79 83 86 0.08

Time serious of daily average PM10 and O3 (2010.7.12~2010.7.21) July 2010

Sites OBS. SIM. MB MAGE RMSE r LH 152 150

• 1

19 25 0.73 WQS 132 87

• 45

48 56 0.06 TH 227 223

• 4

38 42 0.23 LY 68 69 1 22 24

• 0.03

TJ 94 69

• 24

24 26 0.78 JJJ 138 112

• 26

29 48 0.03 HJC 144 104

• 40

40 42 0.51 DH 49 87 39 39 45

• 0.71

CZ 69 86 17 22 26

• 0.59

XP 211 121

• 90

90 100 0.47 JGW 190 116

• 74

74 83 0.39 HG 159 126

• 33

45 52 0.41 ZML 92 85

• 7

22 25 0.06 Average 133 110

• 22

39 46 0.18

PM10 O3

SLIDE 64

The average wind field, SO2, NO2, PM10 and O3 horizontal distribution during day time in Pearl River Delta region 2010.6.30~2010.7.4（07:00~18:00LST）

July 2010