Combining HYSPLIT and CMAQ to resolve urban scale features: an example of application in Houston, TX Ariel F. Stein (1), Vlad Isakov (2), James Godowitch (2), and Roland R. Draxler (3) 1 Earth Resources & Technology (ERT) on assignment to NOAA Air Resources Laboratory, Silver Spring, MD 2 NOAA ARL Atmospheric Sciences Modeling Division**, RTP, NC 3 NOAA Air Resources Laboratory, Silver Spring, MD **In partnership with the U.S. EPA
19
Embed
Ariel F. Stein (1), Vlad Isakov (2), James Godowitch (2), and Roland R. Draxler (3)
Combining HYSPLIT and CMAQ to resolve urban scale features: an example of application in Houston, TX. Ariel F. Stein (1), Vlad Isakov (2), James Godowitch (2), and Roland R. Draxler (3) 1 Earth Resources & Technology (ERT) on assignment to NOAA Air Resources Laboratory, Silver Spring, MD - PowerPoint PPT Presentation
Welcome message from author
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
Combining HYSPLIT and CMAQ to resolve urban scale features:
an example of application in Houston, TX
Ariel F. Stein (1), Vlad Isakov (2), James Godowitch (2), and Roland R. Draxler (3)1 Earth Resources & Technology (ERT) on assignment to NOAA Air Resources
3 NOAA Air Resources Laboratory, Silver Spring, MD**In partnership with the U.S. EPA
Motivation• Resolving urban scale features is critical for air toxics modeling and exposure
assessments.
• 3D Eulerian models are not designed for fine scale applications. Even at 1x1 km resolution we are pushing the limits of this approach. Very resource intensive.
• Objective: to test the feasibility of developing an urban hybrid simulation system
• In this example, CMAQ provides the regional background concentrations and urban-scale photochemistry, HYSPLIT provides the spatially resolved concentrations due to stationary emission sources, and AERMOD – from mobile sources.
• In this first application, the HYSPLIT, AERMOD and CMAQ models are used in combination to calculate high resolution Benzene concentrations in the Houston area. The study period is from August 22th to September 3rd of 2000.
• Furthermore, multiple HYSPLIT simulations with varying model inputs and physical parameters are used to create a concentration ensemble to estimate the concentration variability. Health assessment and exposure models.
Scale interactions
Choosing models
• Looking for the appropriate tools to handle different scales.
• Regional scale: 3D Eulerian models with full chemistry CMAQ
• Local scale: – Lagrangian models (simple or no chemistry)
• Exploring two sources: – Dispersion: created by different particle trajectory
pathways in the turbulent atmosphere. 27 members, in which the particle trajectory variability in HYSPLIT is calculated by using a different seed number to estimate the random component of the particle diffusion.
– Transport: different flow regimes that might be introduced when using gridded data to represent meteorological data fields. 18 members created by shifting the position of the meteorological field variables.
Advection and Dispersion• X(t+t)=Xmean(t+t)+U'(t+t) t,
• U'(t+t)=R(t) U'(t)+U"( 1-R(t)2 )0.5 ,
• R(t)=exp(-t/TLx),
• U"=u , where U’ is the random velocity component, Xmean is the original position due to only advection by the
mean winds, R is the turbulent velocity autocorrelation, u is the standard deviation of the turbulent velocity, and is a computer generated random number with 0 mean and of 1. Additional terms to account for gradients in the turbulent velocity near the ground are required for vertical particle dispersion.
The growth of the particle distribution, or the “puff” mode, is represented by a much simpler formulation, where the growth rate of the horizontal standard deviation of the particles is given by
• dh/dt = (2 u)0.5
The dispersive growth rate for particles or puffs is controlled by the standard deviation of the turbulent velocity.
• u = (Kx / TL)0.5
where K represents the turbulent diffusivity and TL a constant Lagrangian time scale. For vertical turbulence and within the boundary layer K is a function of height and surface stability. Above the boundary layer it depends upon the local stability, a ratio of the wind shear and thermal stratification. Horizontal turbulence is computed from the deformation of the wind field.
Modeling domain: Houston, TX
Monitors
Census tract centroids
Stationary sources
Major roads
36 km x 36 km areaBenzene emissionsHYSPLIT: Top six point sources ~60% emissions in Harris CountyAERMOD: 330 road links in 36x36 km area ~25% of all on-road benzene emissions in Harris County
Models settings• The study period is from August 22th to September 3rd of 2000 • All models driven by MM5 meteorology• Emissions based on 1999 National Emission Inventory
• CMAQ:– For the CMAQ domain with 36-km grid cell spacing:
• 45 columns x 46 rows with 24 vertical layers• Layer 1 ~ 38 m thickness• Central Houston/Harris County is covered by grid cell: column16, row 14• Boundary conditions: tropospheric “clean” background• Benzene emissions from “local” emission sources being treated by the dispersion
models were removed from the Houston 36-km cell to avoid a “double-counting” situation.
– For the CMAQ domain with 1-km grid cell spacing:• 108 x 108 with 24 vertical layers• Layer 1 ~ 38 m thickness• Boundary conditions from 12, 4 km nests
– General Model Configuration:– SAPRC99 photochemical mechanism + toxics chemistry, includes
• HYSPLIT:– 5000 3D lagrangian particles released– No chemistry– Meteorological fields calculated with MCIP2ARL preprocessor – Horizontal resolution : 0.01x 0.01 degree (1x1 km2 horizontal resolution)
• AERMOD:– No chemistry– Meteorology: Hourly surface observations from the nearest NWS
meteorological station at George Bush Intercontinental Airport (IAH), and vertical profiles from the Lake Charles upper air station, processed by AERMET
– Emissions: 330 road links modeled as area sources in 36x36 km2 area ~25% of all on-road benzene emissions in Harris County
– Receptors: 38 x 31 receptor grid, 1178 receptors total (consistent with HYSPLIT: 0.01 degree spacing, 29.65 – 29.95 deg. Lon, 95.14 – 95.51 Lat).