1. MODELING AT NEIGHBORHOOD SCALE Sylvain Dupont and Jason Ching
Introduction
MODELING AT NEIGHBORHOOD SCALE
Sylvain Dupont and Jason Ching
Collaborators: Tanya Otte and Avraham Lacser
University Corporation for Atmospheric Research
U.S. Environmental Protection Agency
Research Triangle Park, NC

2. Objective: modeling air-quality for estimating human exposure to air pollution in urban area.
Modeling at neighborhood scale: development of an Urban Canopy Parameterization (UCP) inside MM5 for CMAQ.
Estimating the sub-grid-scale variability in the pollutant concentration fields.

3. Outline Definition of neighborhood scale
Urban Canopy Parameterization (UCP)
# General scheme
# Dynamic, thermal, humidity and TKE components
Preliminary results: Philadelphia case
# MM5 results
# CMAQ results
Conclusions and Perspectives

4. Neighborhood scale Interaction between meso and local scales.
Meso scale
Neighborhood scale
1 km.
Roughness Sub-Layer
Local scale
Rural
Rural
Urban

5. The details of the whole urban canopy can not be represented:
Neighborhood scale
The details of the whole urban canopy can not be represented:
Parameterization of the urban surface effects.
Meso scale
Neighborhood scale
1 km.
Roughness Sub-Layer
Local scale
Rural
Rural
Urban

6. Majority of pollutants are emitted inside the roughness sub-layer:
Neighborhood scale
Majority of pollutants are emitted inside the roughness sub-layer:
Necessity to have a good representation of meteorological fields.
Meso scale
Neighborhood scale
1 km.
Roughness Sub-Layer
Local scale
Rural
Rural
Urban

7. Neighborhood scale
The ground conditions used by mesoscale models are not satisfactory at neighborhood scale:  Drag-force approach.
Meso scale
Neighborhood scale
1 km.
Roughness Sub-Layer
Local scale
Rural
Rural
Urban

8. Drag-Force approach Meso scale Neighborhood scale Rural Rural Urban
I Modèle de sol urbain SM2-U
Drag-Force approach
Meso scale
Neighborhood scale
1 km.
Roughness Sub-Layer
Rural
Rural
Urban

9. Urban Canopy parameterization
The UCP is introduced inside the Gayno-Seaman PBL model.
Complete the drag-force approach introduced by Lacser & Otte in MM5 following the work of Martilli (2002).
Extend the drag-force approach to all roughness elements inside the canopy: buildings and vegetation.
Introduce the detail soil model SM2-U considering both rural and urban surfaces.

10. Urban canopy parameterization
SM2-U

11. Urban canopy parameterization
New version of the UCP

12. Urban canopy parameterization
Urban morphology
The knowledge of the vertical and horizontal distribution of the different surface types is necessary.
Roof area density
Vegetation area density
Building plan area density
Vegetation plan area density
Building frontal area density

13. Momentum equation = forcing terms
Urban canopy parameterization
Dynamic component
Momentum equation = forcing terms
(modification of vertical turbulent transport term)
+ momentum sources due to horizontal and vertical building surface
+ momentum sources due to vegetation

14. Heat equation = forcing terms
Urban canopy parameterization
Sensible heat flux
Latent heat flux
Net radiation: solar, atmospheric, and earth radiations
Storage heat flux
Anthropogenic heat flux
Hsens i LE
Gs i
Qanth i
Rn i
Thermal components
Heat equation = forcing terms
(modification of vertical turbulent transport term)
+ heat sources from surfaces + anthropogenic heat sources

15. Effects of the canopy thickness
Urban canopy parameterization
Effects of the canopy thickness
Modification of paved surface temperature equation
# Heat capacity of the wall
# Heat exchange between through the buildings
# Radiative trapping: introduction of an effective albedo parameterization deduced from Masson (2000).
Extinction of the radiation through the canopy

16. Humidity equation= forcing terms
Urban canopy parameterization
Humidity components
Precipitations
Infiltration
Draining network
Return towards equilibrium
Draining
Evapotranspiration
Water draining outside the system
Humidity equation= forcing terms
(modification of vertical turbulent transport term)
+ humidity sources from surfaces + anthropogenic humidity sources

17. TKE equation= forcing terms
Urban canopy parameterization
TKE components
TKE equation= forcing terms
(modification of vertical turbulent transport and dissipation terms)
+ TKE sources due to horizontal and vertical building surface
+ TKE sources due to vegetation
+ TKE sources due to sensible heat fluxes

18. Summary of MM5 versions
Roughness approach
Drag approach

19. Preliminary results: Philadelphia case
14 July 1995 (sunny day).
MM5 has been run in a one-way nested configuration: 108, 36, 12, 4 and 1.33 km horizontal grid spacing.
UCP uses only for the 1.33 km domain.
Turbulent scheme model: Gayno-Seaman PBL with the turbulent length scale of Bougeault and Lacarrere (1989).

20. 1.33 km domain 112x112x40 grid points 4 km domain 85x88x30 grid points
Philadelphia case
1.33 km domain
112x112x40 grid points
4 km domain
85x88x30 grid points

21. 7 urban categories have been defined following Ellefsen (1990-91).
Philadelphia case
For the 1.33 km domain:
7 urban categories have been defined following Ellefsen ( ).
23-category (USGS) vegetation categories.

22. Mixing height and wind vectors at 50 m AGL
Philadelphia case
Mixing height and wind vectors at 50 m AGL
a) the standard version of MM5 using GS PBL
b) GS PBL including TLSP (B-L,89)
Without UCP (nocan)

23. Vertical profiles in central Philadelphia,
Philadelphia case
Vertical profiles in central Philadelphia,
Ratios: a) local u*, and b) TKE to local u* max at 2 p.m.
c) potential temperature at 6 a.m.
Solid line (can), dash line (nocan); Roof percentage bottom right

24. Meteorological fields
Philadelphia case
Meteorological fields
Can simulations
Left: mixing height
Right: air temperature and wind vectors at 50 m

25. (can-nocan) simulations
Philadelphia case
Difference fields
(can-nocan) simulations
Left: mixing height
Right: air temperature and wind vectors at 50 m

26. CMAQ Results MM5 v 3.5 (w/UCP). CB-IV mechanism.
Introduction
Philadelphia case
CMAQ Results
MM5 v 3.5 (w/UCP).
CB-IV mechanism.
Turbulent scheme from the G-S PBL scheme.
CMAQ computational domain and grid structure based on MM5 domains:
# 21 layer gridding for 36, 12, and 4 km simulations
# 31 layer gridding for 1.33 km runs with UCP
Emission processing using SMOKE
# Near surface emissions distributed into lowest 10 vertical layers for 1.33 km grid simulations

27. Philadelphia case
Normalized Difference with and without BL89 (nocan) ( 2pm EDT) (a) CO; (b) HCHO; © NOx; (d) O3 a c d b

28. Parameter Sensitivity Case Study
Philadelphia case
Parameter Sensitivity Case Study
July 14, 1995
Grid size 1.33 km
(Pcan – P nocan) /P can

29. Normalized Difference for CO (6 a.m. local)
Philadelphia case
Normalized Difference for CO (6 a.m. local)

30. Normalized Difference (6 p.m. local)
Philadelphia case
Normalized Difference (6 p.m. local)
NOx Ozone

31. Normalized Difference
Philadelphia case
Normalized Difference
for Fine Particle Number (Left: 6 a.m. Right 6 p.m.)

32. Multi-scale Simulations
Philadelphia case
Multi-scale Simulations
km grid sizes
July 14, 1995
(6 p.m. local)

33. Philadelphia case CO

34. Philadelphia case
NOx

35. Philadelphia case
Ozone

36. Fine Particle Number (x10 9)
Philadelphia case
Fine Particle Number (x10 9)

37. Philadelphia case
Sulfate (mg/m3)

38. Philadelphia case
Ammonium (mg/m3)

39. Elemental Carbon (mg/m3)
Philadelphia case
Elemental Carbon (mg/m3)

40. Aldehydes (with UCP) HCHO CH3CHO
Philadelphia case
Aldehydes (with UCP) HCHO CH3CHO

41. Neighborhood-Scale Modeling
Summary Points
UCP introduced into MM5
# Modified turbulence length scale parameterization in GS-PBL model: Suppresses undesired undulations
#Improved Dispersion parameters: Mixing heights, U*, stability, …
Air quality fields
# Sensitivity to introduction of UCP
# Spatial pattern details resolved at N-S
# Resolution requirements differ for different pollutants

42. Project Status, Future plans
Testing and refining UCPs in MM5 and CMAQ
Develop PDFs for sub- grid variability for different parent grid resolutions
Work-in-Progress: Prototype study
Preliminary results for Philadelphia
Advanced N-S modeling for Houston, Texas
Detailed urban morphology data base