1 / 26

Advanced CMAQ Concepts

Advanced CMAQ Concepts. Plume in Grid Process Analysis Model Performance Evaluation and QA Procedures. Plume in Grid (PinG) (1). Subgrid scale treatment of major emitting point sources (MEPSE) For more realistic treatment of dynamic and chemical processes impacting elevated point sources

kristian-vidar
Download Presentation

Advanced CMAQ Concepts

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Advanced CMAQ Concepts • Plume in Grid • Process Analysis • Model Performance Evaluation and QA Procedures

  2. Plume in Grid (PinG)(1) • Subgrid scale treatment of major emitting point sources (MEPSE) • For more realistic treatment of dynamic and chemical processes impacting elevated point sources • CMAQ currently has one implementation of a PinG treatment • CMAQ PinG consists of a Plume Dynamics Model (PDM) and a Lagrangian reactive plume model • Capable of both gas-phase and aerosol treatment

  3. Plume in Grid (PinG)(2) • The PDM is a stand-alone preprocessor that simulates plume rise, horizontal and vertical growth, dispersion, and transport at sub-grid scales • The PDM controls the interaction between the plumes and the parent grid • The Lagrangian plume model is internal to the CCTM and simulates the chemistry within the plumes themselves • Intended for grid resolutions of 20-40 km • Both physical and chemical criteria for plume handover to parent grid

  4. CCTM PinG Plume in Grid (PinG)(3) Emissions PDM Meteorology Adapted from: Gillani and Godowitch (1999), Science Algorithms of the EPA Models-3 CMAQ Modeling System , EPA/600/R-99/030, pp. 9.1 9.31

  5. Plume in Grid (PinG)(4) • CMAQ implementation requires compiling the CCTM with the PinG option invoked and running the PDM preprocessor to prepare special emissions inputs • Two CCTM compiler options for PinG • ping_noop: No PinG treatment • ping_smvgear: PinG with internal Gear chemistry solver • Emissions requirements: 2-d MEPSE file that defines which sources to receive PinG treatment • SMOKE instrumented to create CMAQ-ready MEPSE files

  6. Plume in Grid (PinG)(5) • The PDM uses a MEPSE file and meteorology inputs to create a CCTM input file • Build and execute the PDM similar to the other CMAQ preprocessors (e.g. ICON, BCON) • CCTM compiled with the PinG option will look for the additional PDM and MEPSE input files during execution • Additional CCTM PinG output includes an unmerged/active plume netCDF file • Post-processing utility to overlay the active plumes onto the parent grid without chemical coupling

  7. Process Analysis (PA)(1) • Eulerian grid models are based on partial differential equations that define the time-rate of change in species concentrations due to chemical and physical processes • PA is a configuration system within Eulerian models that provides quantitative information about the impacts of individual processes on the cumulative chemical concentrations • PA is an optional feature of CMAQ that provides insight into the reasons for a model’s predictions

  8. Process Analysis (PA)(2) • Two classes of PA • Integrated reaction rates (IRR) • Integrated process rates (IPR) • PA is useful for • Identifying sources of error • Interpreting model results • Determining the important characteristics of chemical mechanisms (IRR) • Determining the important characteristics of different implementations of physical processes (IPR) • IPR quantifies the contribution of each source and sink process for a particular species at the end of each time step

  9. Process Analysis (PA)(3) • CMAQ implementation requires compiling the CCTM with PA include files generated by the PROCAN preprocessor • PA include files specify • IRR or IPR • Chemical species or groups to collect PA information about • A PROCAN configuration file defines the contents of the include files • A PROCAN run script uses information in the configuration file and calls the executable to create the include files

  10. CCTM PA_CMN PA PA_CTL PA_DAT Process Analysis (PA)(4) pa.inp PROCAN Configuration File Include Files

  11. Process Analysis (PA)(7) • IRR quantifies the mass throughput of a particular reaction within a chemical mechanism • IRR can diagnose mechanistic and kinetic problems within the chemistry model • IRR can reveal NOx vs. VOC sensitivity regimes • IRR generally more difficult to interpret than IPR

  12. Model Performance Evaluation (MPE) • Question why a model is doing what it is doing • What are the inherent uncertainties and how do they impact the model results • Qualitative and quantitative evaluation • Diagnostic versus operational evaluation • Comparisons against observations • Evaluate at different temporal and spatial scales • Categorical model evaluation (used for Forecasting) • Contingency Table, False Alarm Rate, Skill Scores, CSI, etc.

  13. Quantitative vs Qualitative • Qualitative model evaluation targets intuitive features in results • Effects of urban areas • Boundary layer effects • Effects of large point sources and highways • Diurnal phenomena • Quantitative evaluation provides statistical evidence for model performance • Daily, seasonal, annual comparisons with observed data • At coarse grids, compare observations with the concentrations in the model grid cell in which the monitor is located • At fine grids, compare observations with the concentrations in a matrix of cells surrounding the cell in which the monitor is located

  14. Problems/Issues • Modeling scales have grown tremendously both spatially and temporally • Datasets becoming larger • Need to process and digest voluminous amount of information • Heterogeneous nature of observational datasets • Vary by network, by quality, by format, by frequency • Measurement or model artifacts • What is modeled is not always measured • Need adjustments before comparisons • Problem of incommensurability • Comparing point measurement with volume average

  15. Observational Databases • AIRS (hourly) (~4000) • IMPROVE (every 3rd day) (~160) • CASTNET (hourly, weekly) (123) • NADP (weekly) (over 200) • EPA Supersites (sub-hourly) (8) • EPA STN (hourly) (215) • PAMS (hourly) (~130) • AERONET • Special field campaigns • e.g. AIRMAP, ASACA, BRAVO, CCOS, CRPAQS, NARSTO, SEARCH, SOS, TXAQS, etc. • Aircraft Data • Remote Sensing Data (AURA, MODIS, etc.)

  16. Operational Evaluation(mostly quantitative) • Compute suite of statistical measures of performance • Peak Prediction Accuracy, Bias metrics (MB, MNB, NMB, FB), Error metrics (RMSE, FE, GE, MGE, NMGE), etc. • “Goodness-of-fit” measures (based on correlation coefficients and their variations) • Various temporal scales • Time-series analyses • Hourly, weekly, monthly • Grid (tile) plots • Scatter plots • Pie-charts

  17. Diagnostic Evaluation(qualitative and quantitative) • Compute various ratios • Metrics different for each problem being diagnosed / studied • O3/NOz,,H2O2/HNO3 for NOx versus VOC limitation • NOz/NOy for chemical aging • PM species ratios such as NH3/NHx, NO3/(total nitrate) for gas-particle partitioning, NH4/SO4, NH4/NO3, etc. • Others? • Innovative Techniques • Empirical Orthogonal Functions • Principal Component Analyses • Process Analyses • Source Apportionment (available for Carbon and Sulfur) • Decoupled-direct method (DDM) • Others?

  18. Analyses Tools for MPE • sitecmp to prepare obs-model pairs • Part of CMAQ Distribution • PAVE • http://www.cmascenter.org • I/O API Utilities • http://www.baronams.com/products/ioapi • netCDF Operators • http://nco.sourceforge.net • NCAR Command-line Language • http://www.ncl.ucar.edu • Python I/O API Tools • http://www-pcmdi.llnl.gov/software-portal/Members/azubrow/ioapiTools/index_html • Atmospheric Model Evaluation Tool (AMET) • Under development at EPA

  19. MPE Example 1 Grid Resolution Variability 36-km 4-km 12-km

  20. MPE Example 2Spatial Variability of Peak Predictions

  21. MPE Example 3Wind and Obs Overlay

  22. MPE Example 4Scatter Plot Analyses • Regression analyses present model results across multiple observation points or time periods SO4 O3

  23. MPE Example 5Time Series Analyses

  24. MPE Example 6 Attainment Demonstration for O3

  25. MPE Example 7Forecast Model Evaluation

More Related