Comparison of UAM Results with Ambient

Definitions
Introduction
Urban Airshed Model
Evaluating UAM Performance using Surface Air Quality Data
Example Graphical Displays
Statistical Parameters for Model Performance Evaluation
Comparison of UAM Results with Aloft Air Quality Data
Characterization of Ozone Concentrations Aloft During SCAQS
Comparing UAM Predictions and Aloft Air Quality
Result of Comparison of Aloft Ozone and NOx Measurements with Model Predictions in SCAQS
Air Quality Regional Boundaries
Summary
References
Lumped Carbon-Bond Species
Relationship Between Individual Chemical Species and Lumped Carbon-Bond Species

CALMET = Diagnostic meteorological model
CALRAMS = Prognostic meteorological model
CBIV = Carbon bond IV chemical reaction mechanism scheme
NO_x = NO + NO₂ + poorly defined fraction of other NO_x species (given conventional analyzers)
NO_y = NO_x+ HNO₃ + organic nitrates + inorganic nitrates = NO_x + NO_z
NO_z = Oxidation products of NO_x = NO_y * (1 - NO_x/NO_y)
RHC = Reactive hydrocarbons
ROM = Regional oxidant model
UAM = Urban Airshed Model
IV = EPA Regulations version using CBIV; V = with variable grid
VOC = Volatile organic compounds

•  Important to compare Urban Airshed Model (UAM) output with ambient air quality data to assess model performance.
•  Three broad types of ozone and precursor data useful for comparisons to UAM output:
•  Surface air quality
• Aloft air quality
• Boundary conditions

• The UAM is a 3-D grid model designed to calculate the concentrations of both inert and chemically reactive pollutants by simulating physical and chemical processes that take place in the atmosphere.
• The UAM uses a mass balance in which relevant emissions, transport, chemical reaction, and removal processes are expressed in mathematical terms.
• Simulations are usually 24- to 72-hour periods during which episodic meteorological conditions persist.
• Typical UAM application:
•  Select episode (usually widespread exceedance of ozone NAAQS, typical meteorological conditions).
• Select modeling domain to encompass ozone monitors that reported exceedances and all major source regions.
• Prepare model inputs using observed meteorological, emission, and air quality data for an episode.
• Evaluate model performance.
• The UAM is used for analysis of spatially and/or temporally differentiated future emission control strategies and their effect on air quality in various parts of the modeling region.

EVALUATING UAM PERFORMANCE USING SURFACE AIR QUALITY DATA

 Analytical Tools

•  Graphical displays
• Statistics

Species

• Ozone
• NO, NO₂, NO_x, NO_y
• VOC in CBIV groups
• VOC/NO_y or VOC/NO_x ratios

Supplemental Data

•  Meteorological data
• Emissions estimates
• Geophysical data
• Data quality and completeness information

•  Time series plots
• Compare observed and simulated pollutant concentrations for ozone, NO, NO₂ (or NO_y) and selected VOC for all monitoring sites within model domain.
• Compare observed ozone concentrations with the minimum and maximum simulated concentrations within nine surrounding grid cells of a monitoring site.
• Contour plots
• Show simulated pollutant concentrations and observed concentrations for ozone, NO, NO₂ (or NO_y), and selected VOC for each hour.
• Frequency distributions
• Of residuals (differences between hourly observed and predicted concentrations) for ozone.
• Scatter plots
• Observed versus predicted hourly concentrations.

 Introduction 

Analysis Approach 

• Obtain UAM air quality output (all layers) for grid cells surrounding air quality measurement location.
• Obtain gridded mixing heights for these cells.
• Bin aloft data to match UAM output layers.
• Plot comparisons.
• Investigate spatial and temporal variability of the UAM output.
• Review results with respect to meteorological conditions leading to the ozone episode.

Vertical profiles of NO_x concentrations measured by aircraft spiral compared to the UAM average for El Monte, June 25, 1987. The 15-m vertical average aircraft NO_x concentrations, 5-layer model predictions, and model layer-averaged aircraft data are shown. The model predictions were significantly lower aloft (particularly at about 500 to 100 m msl) and significantly higher in the surface layer all day. (Roberts et al., 1993b)

• Model predictions of ozone below 200 m agl in the morning were ³ measured concentrations.
• Model predictions of ozone above about 200 m agl were typically about 50 to over 100 ppb lower than measured concentrations.
• Model predictions of NO_x below about 400 m agl in the morning were both higher and lower than measured concentrations.
• Model predictions of NO_x above 400 m agl in the midday and afternoon were typically lower than measured concentrations.
• Modeling recommendations included the need to properly represent significant ozone concentrations aloft (esp. 500-1000 m), mixing of ozone layers to the surface, horizontal pollutant structure aloft, and mixed-layer ozone greater than surface ozone concentrations.

•  Surface air quality data not necessarily sufficient
•  Regional models often underpredict ozone at upwind boundaries
•  Important to look at precursors
•  Definition of boundary

Average aloft (above 600 m msl) and surface NO_x concentrations along the 1991 LMOS southern boundary in the morning, midday, and evening on July 18 (Roberts et al., 1994).

Aloft and surface NMOC concentrations along the 1991 LMOS southern boundary in the morning, midday, and evening on July 18 (Roberts et al., 1994).

• Comparisons should also be made with ozone and ozone precursors, including NO, NO_y, and speciated VOC.
• Comparisons at domain boundary are important.

Adamski W. (1997) An analysis of measured and predicted concentrations aloft of ozone and total reactive oxides of nitrogen in the eastern U.S. during July 1995. Draft report prepared by Wisconsin Department of Natural Resources, Racine, WI, January.

California Air Resources Board (1995) Sacramento area modeling analysis for the 1994 state implementation plan. Report prepared by Technical Support Division, California Air Resources Board, Sacramento, CA, April.

Cassmassi J., Mitsutomi S., Bassett M., Lester J.C., and Zhang X. (1994) Ozone modeling - performance evaluation. Draft technical report V-B prepared by South Coast Air Quality Management District, Diamond Bar, CA, June.

Eisinger, D.S., Deakin E.A., Mahoney L.A., Morris R.E., and Ireson R.G. (1990) Transportation control measures: state implementation plan guidance. Revised final report prepared by Systems Applications International, San Rafael, CA, SYSAPP-90/084, September.

Hanna S.R., Moore G.E., and Fernau M.E. (1996) Evaluation of photochemical grid models (UAM-IV, UAM-V, and the ROM/UAM-IV couple) using data from the Lake Michigan Ozone Study (LMOS). Atmos. Environ. 30, 3265-3279.

McNair L.A., Harley R.A., and Armistead G.R. (1996) Spatial inhomogeneity in pollutant concentrations, and their implications for air quality model evaluation. Atmos. Environ. 30, 4291-4301.

Roberts P.T. and Main H.H. (1992a) Characterization of three-dimensional air quality during the SCAQS. In Southern California Air Quality Study Data Analysis. Proceedings from SCAQS Data Analysis Conference, University of California, Los Angeles, CA, July 21-23, Air & Waste Management Association, Pittsburgh, PA, (STI-1223), VIP-26.

Roberts P.T., Main H.H., Lindsey C.G., and Korc M.E. (1993a) Ozone and particulate matter case study analysis for the Southern California Air Quality Study. Final report prepared for the California Air Resources Board, Sacramento, CA by Sonoma Technology, Inc., Santa Rosa, CA, STI-90020-1222-FR, May.

Roberts P.T., Main H.H., and Korc M.E. (1993b) Comparison of 3-D air quality data with model sensitivity runs for the South Coast Air Basin. Paper No. 93-WP-69B.05 presented at the Air & Waste Management Association Regional Photochemical Measurement and Modeling Studies Conference, San Diego, CA, November 8-12, (STI-1244).

Roberts P.T., Dye T.S., Korc M.E., and Main H.H. (1994) Air quality data analysis for the 1991 Lake Michigan Ozone Study. Final report prepared for Lake Michigan Air Directors Consortium, Des Plaines, IL by Sonoma Technology, Inc., Santa Rosa, CA, STI-92022-1410-FR.

Stoeckenius T.E., Ligocki M.P., Shepard S.B., and Iwamiya R.K. (1994a) Analysis of PAMS data: application to summer 1993 Houston and Baton Rouge data. Draft report prepared by Systems Applications International, San Rafael, CA, SYSAPP94-94/115d, November.

Systems Applications International, Sonoma Technology Inc., Earth Tech, Alpine Geophysics, and A.T. Kearney (1995) Gulf of Mexico Air Quality Study. Vol. I: summary of data analysis and modeling. Final report prepared for U.S. Department of the Interior, Minerals Management Service, Gulf of Mexico OCS Region, New Orleans, LA, OCS Study MMS-95-0038.

Tesche T.W., Georgopoulos P., Seinfeld J.H., Cass G., Lurmann F.W., and Roth P.M. (1990) Improvement of procedures for evaluating photochemical models. Draft final report prepared for Research Division, California Air Resources Board, Sacramento, CA by Radian Corporation, Sacramento, CA, Contract No. A832-103, March.

U.S. Environmental Protection Agency (1991) Guideline for regulatory application of the Urban Airshed Model (UAM). Report prepared by U.S. Environmental Protection Agency, Research Triangle Park, NC, EPA 450/4-91-013.

 a The relationships between FORM and ALD2 and individual chemical species are not indicated because ambient data for important components of those lumped species are not available.

Source: Stoeckenius et al., 1994a