Measurement Needs for AQ Models

Slides:



Advertisements
Similar presentations
Some recent studies using Models-3 Ian Rodgers Presentation to APRIL meeting London 4 th March 2003.
Advertisements

EPA PM2.5 Modeling Guidance for Attainment Demonstrations Brian Timin EPA/OAQPS February 20, 2007.
Experiments with Assimilation of Fine Aerosols using GSI and EnKF with WRF-Chem (on the need of assimilating satellite observations) Mariusz Pagowski Georg.
Section highlights Organic Aerosol and Field Studies.
Incorporation of the Model of Aerosol Dynamics, Reaction, Ionization and Dissolution (MADRID) into CMAQ Yang Zhang, Betty K. Pun, Krish Vijayaraghavan,
The semi-volatile nature of secondary organic aerosol (SOA) in the Mexico City Metropolitan Area November 2, 2007 EAS Graduate Student Symposium Christopher.
Title EMEP Unified model Importance of observations for model evaluation Svetlana Tsyro MSC-W / EMEP TFMM workshop, Lillestrøm, 19 October 2010.
Christian Seigneur AER San Ramon, CA
The Framework of Modeling SOA Formation from Toluene Oxidation Di Hu and Richard Kamens Department of Environmental Sciences and Engineering, University.
A Comparative Dynamic Evaluation of the AURAMS and CMAQ Air Quality Modeling Systems Steven Smyth a,b, Michael Moran c, Weimin Jiang a, Fuquan Yang a,
The robustness of the source receptor relationships used in GAINS Hilde Fagerli, EMEP/MSC-W EMEP/MSC-W.
Evaluation of the AIRPACT2 modeling system for the Pacific Northwest Abdullah Mahmud MS Student, CEE Washington State University.
Impact of chemistry scheme complexity on UK air quality modelling Met Office FitzRoy Road, Exeter, Devon, EX1 3PB United Kingdom Tel: Fax:
Air Quality Impact Analysis 1.Establish a relationship between emissions and air quality. AQ past = a EM past + b 2.A change in emissions results in an.
Effects of Pollution on Visibility and the Earth’s Radiation Balance John G. Watson Judith C. Chow Desert Research Institute Reno,
GEM-MACH Global The Canadian Global Air Quality Modeling/Forecasting System Dr. Sunling Gong Science and Technology Branch January 16-17, 2012.
EVALUATION AND IMPROVEMENT OF GAS/PARTICLE MASS TRANSFER TREATMENTS FOR 3-D AEROSOL SIMULATION AND FORECAST Xiaoming Hu and Yang Zhang North Carolina State.
PM2.5 Model Performance Evaluation- Purpose and Goals PM Model Evaluation Workshop February 10, 2004 Chapel Hill, NC Brian Timin EPA/OAQPS.
Summer Institute in Earth Sciences 2009 Comparison of GEOS-5 Model to MPLNET Aerosol Data Bryon J. Baumstarck Departments of Physics, Computer Science,
Ozone MPE, TAF Meeting, July 30, 2008 Review of Ozone Performance in WRAP Modeling and Relevance to Future Regional Ozone Planning Gail Tonnesen, Zion.
Presentation by: Dan Goldberg Co-authors: Tim Vinciguerra, Linda Hembeck, Sam Carpenter, Tim Canty, Ross Salawitch & Russ Dickerson 13 th Annual CMAS Conference.
A comparison of PM 2.5 simulations over the Eastern United States using CB-IV and RADM2 chemical mechanisms Michael Ku, Kevin Civerolo, and Gopal Sistla.
PM Model Performance in Southern California Using UAMAERO-LT Joseph Cassmassi Senior Meteorologist SCAQMD February 11, 2004.
4. Atmospheric chemical transport models 4.1 Introduction 4.2 Box model 4.3 Three dimensional atmospheric chemical transport model.
GEM/AQ Simulations on Intercontinental Transports Science and Technology Branch Environment Canada.
Photochemical Modeling For Regulatory Applications Jim Boylan – GA EPD Praveen Amar – NESCAUM CMAS Users Forum Meeting October 14, 2010.
Model Evaluation Comparing Model Output to Ambient Data Christian Seigneur AER San Ramon, California.
Operational Evaluation and Comparison of CMAQ and REMSAD- An Annual Simulation Brian Timin, Carey Jang, Pat Dolwick, Norm Possiel, Tom Braverman USEPA/OAQPS.
1 Modeling the Atmospheric Transport and Deposition of Mercury Dr. Mark Cohen NOAA Air Resources Laboratory Silver Spring, Maryland Mercury Workshop, Great.
Evaluation of the VISTAS 2002 CMAQ/CAMx Annual Simulations T. W. Tesche & Dennis McNally -- Alpine Geophysics, LLC Ralph Morris -- ENVIRON Gail Tonnesen.
11 September 2015 On the role of measurements and modelling in Dutch air quality policies Guus Velders The Netherlands (RIVM)
SEAC4RS Payload Payload Synergies Synergies. Complementarity between aircraft can be considered to fall into three categories. Each has considerations.
Diagnostic Study on Fine Particulate Matter Predictions of CMAQ in the Southeastern U.S. Ping Liu and Yang Zhang North Carolina State University, Raleigh,
A ir Quality Research Branch Meteorological Service of Canada Environment Environnement Canada Performance Evaluation of AURAMS for Multiple Cases Michael.
How accurately we can infer isoprene emissions from HCHO column measurements made from space depends mainly on the retrieval errors and uncertainties in.
Georgia Institute of Technology SUPPORTING INTEX THROUGH INTEGRATED ANALYSIS OF SATELLITE AND SUB-ORBITAL MEASUREMENTS WITH GLOBAL AND REGIONAL 3-D MODELS:
AoH/MF Meeting, San Diego, CA, Jan 25, 2006 WRAP 2002 Visibility Modeling: Summary of 2005 Modeling Results Gail Tonnesen, Zion Wang, Mohammad Omary, Chao-Jung.
WORKSHOP ON CLIMATE CHANGE AND AIR QUALITY : part I: Intercontinental transport and climatic effects of pollutants OBJECTIVE: Define a near-term (-2003)
Breakout Session 1 Air Quality Jack Fishman, Randy Kawa August 18.
AoH Work Group Weight of Evidence Framework WRAP Meeting – Tucson, AZ January 10/11, 2006 Joe Adlhoch - Air Resource Specialists, Inc.
The application of Models-3 in national policy Samantha Baker Air and Environment Quality Division, Defra.
7. Air Quality Modeling Laboratory: individual processes Field: system observations Numerical Models: Enable description of complex, interacting, often.
Meteorological drivers of surface ozone biases in the Southeast US
Report on “what are the emerging research needs for CAMS”
Emerging Science EPA’s ORD Supports Regional Haze Program
Development of a Multipollutant Version of the Community Multiscale Air Quality (CMAQ) Modeling System Shawn Roselle, Deborah Luecken, William Hutzell,
The science of urban air quality
Model Future: Nesting with Regional Models
Sunil Kumar TAC, COG July 9, 2007
A New Method for Evaluating Regional Air Quality Forecasts
Nitrogen Deposition: Measurement Techniques and Field Studies
AQMEII3: the EU and NA regional scale program of the Hemispheric Transport of Air Pollution Task Force The AQMEII 3 modelling team S. Galmarini, C. Hogrefe,
Impact on Recent North American Air Quality Forecasts of Replacing a Retrospective U.S. Emissions Inventory with a Projected Inventory Michael Moran1,
Yongtao Hu, Jaemeen Baek, M. Talat Odman and Armistead G. Russell
M. Samaali, M. Sassi, V. Bouchet
Steve Griffiths, Rob Lennard and Paul Sutton* (*RWE npower)
Joanna Struzewska Warsaw University of Technology
Alison Redington* and Derrick Ryall* Dick Derwent**
Update on 2016 AQ Modeling by EPA
CMAQ model as a tool for generating input data for HM and POP modeling
About comparability of measured and modeled metrics
PM modelling assessment in Northern Italy
Time-Integrated Particle Measurements : Status in Canada
Michael Moran Air Quality Research Branch
Potential applications of small sensor technology at the nexus of land-atmosphere-society Kirsti Ashworth, Royal Society Dorothy Hodgkin Research Fellow,
UNCERTAINTY ANALYSIS IN OZONE MODELS
Summary: TFMM trends analysis
Regional Modeling for Stationary Source Control Strategy Evaluation
EMEP/MSC-W How can EMEP Intensive measurement periods help to improve modelling of acidification, eutrophication, O3 and PM? Views from MSC-W H. Fagerli.
Joe Adlhoch - Air Resource Specialists, Inc.
Presentation transcript:

Measurement Needs for AQ Models Paul Makar Air Quality Research Branch, Meteorological Service of Canada, Environment Canada Environment Environnement Canada Canada

What kind of data have been used in the past (by MSC) Points to be covered: What kind of data have been used in the past (by MSC) for model evaluation? How have these data been used? What are their limitations / benefits? Some ideas on what would be useful for future work. Environment Environnement Canada Canada

AQ Model/Measurement Comparisons: Three types of data for comparisons: Laboratory / Process Studies Monitoring Networks Measurement Intensives Environment Environnement Canada Canada

Laboratory / Process Studies Represent transition between “pure” and “applied” science. Fundamental basis for every component of an AQ model. Some examples: Smog chamber data for gas-phase mechanism evaluation Environment Environnement Canada Canada

Laboratory / Process Studies Represent transition between “pure” and “applied” science. Fundamental basis for every component of an AQ model. Slope = 0.46 R2 = 0.98 Slope = 0.64 R2 = 0.90 Slope = 0.85 R2 = 0.72 Some examples: Comparison of laboratory water uptake on organics to absorptive partitioning model results. Environment Environnement Canada Canada

Laboratory / Process Studies Last example: 1-D column model of Isoprene emissions and chemistry. Sensitivity of isoprene emissions measurements to chemical loss: 30% underestimate in emission possible. Environment Environnement Canada Canada

Laboratory and Process Studies are necessary for improvements to science in regional AQ models! Test the model components as stand-alone processes, if possible. (2) Test them in the complete AQ model and compare to monitoring and measurement intensive data. Environment Environnement Canada Canada

Monitoring Networks: Multiple locations across a wide spatial domain.  Long time periods available.  Limited chemical speciation.  Limited time intervals (24 hr averages, 1 day in 3 or 1 day in 6)  Evaluation of model over a wide range of locations, hence over a wide range of conditions. The “depth” of the evaluation is limited by the lack of detail (speciation and time interval). Environment Environnement Canada Canada

Example: An AURAMS model grid showing measurement intensive (Pacific2001) and Monitoring sites: Intensive Monitoring Networks Very detailed data, but only at a few sites, Better representation over domain, but less detailed data. Environment Environnement Canada Canada

Useful for different things: Monitoring networks can be used to get scatterplots. e.g. this shows that the model is over-predicting particle nitrate – it doesn’t tell you why. Environment Environnement Canada Canada

Measurement Intensives: Highly speciated High time-resolution Data only available at small number of sites. Measurements are for a limited total time. e.g. measurement intensive data showed precursor to particle nitrate was being over-predicted. The point: if you want to determine the cause of an inconsistency between model and measurements, you need the highest amount of speciation and time resolution possible. Environment Environnement Canada Canada

AQ Measurement - Model Interactions: Monitoring Networks: New Directions for AQ Measurement - Model Interactions: Monitoring Networks: Provide real-time data for AQ forecast evaluation and data assimilation. Continuous monitoring of some species at high time resolution starting. Adjoint models under construction for data assimilation of air quality variables (to improve AQ model forecast skill). Environment Environnement Canada Canada

AQ Measurement - Model Interactions: Measurement Intensives: New Directions for AQ Measurement - Model Interactions: Measurement Intensives: On-site model-measurement links: AQ models are being used to drive measurement intensive programs. E.g. MSC contribution to ICARTT MSC contribution: study on cloud processes and cloud chemistry. Resulted from the research staff being asked, “What measurements could be made that would provide the most useful data for model improvement.” AQ Model simulations used to help determine site location. AQ Model forecasts to be run to determine aircraft flight plans. Environment Environnement Canada Canada

Some thoughts on “user needs”: What suite of measurements are needed to support model evaluation? How much money do you have to spend? I want measurements of everything, on 5 minute time intervals, every 3 km, over the entire planet. Failing that… The monitoring networks currently in place across North America are sufficient to evaluate an AQ model’s performance. (Operational Evaluation) The monitoring networks currently in place across North America are NOT sufficient to evaluate the causes behind an AQ model’s performance. (Diagnostic Evaluation) Environment Environnement Canada Canada

The monitoring networks currently in place across North America are sufficient to evaluate an AQ model’s performance. (Operational Evaluation) Why? Already measuring O3, PM2.5, PM10 mass at high time resolution, PM speciation at low time resolution. Sufficient to “evaluate” a forecast of these species. Not sufficient to explain the reason for a poor forecast. The monitoring networks currently in place across North America are NOT sufficient to evaluate the causes behind an AQ model’s performance. Why? High time resolution, simultaneous measurements of PM, O3 and precursors are not available in the networks. Need: speciated VOCs, HNO3, NH3, NOx, SO2 at high time resolution, ditto for PM speciation. Environment Environnement Canada Canada

(1) Is the measurement location representative of the grid-square? How do we treat point-measurement versus grid-prediction incommensurability? (1) Is the measurement location representative of the grid-square? This is a model grid-square. 0.85 Its concentration is valid at the point in the centre. Assumed to represent the entire square. 0.82 0.78 0.35 0.85 0.75 This is a measurement site. 0.61 0.65 Is its concentration representative of the entire grid-square? 0.99 Make ten simultaneous measurements. Calculate the slope and correlation between the proposed site and the others. If the correlation is high, the slope is 1, and the intercept is small: the within-grid-square variability is low, and one site will be sufficient to characterize the grid square. Some grid squares may be better than others! Environment Environnement Canada Canada

(2) Is the measurement location representative of the region? Use the same strategy as above; how do the sites correlate across the region? (3) Is the model grid-square representative of the larger region? Use a similar strategy: are the model results at the grid-square well correlated with others in the surrounding region? The point: some statistics on measurements and model results across the domain will quantify “incommensurability”. Environment Environnement Canada Canada

But, if you want to know why the control programs are working What suite of measurements are needed to establish that we have met the requirements of the control programs in terms of achieving the target reductions in air concentrations? Depends: If the target reductions are for total PM, you measure total PM. O3 O3 PM2.5 PM2.5 But, if you want to know why the control programs are working or not, or if they’re working for the reasons you think they are, you need to look at other species, more detailed time resolution (Diagnostic level – wish list coming). Environment Environnement Canada Canada

How does measurement error affect model evaluation? Depends: Measurement limitations or bias, if unreported, will result in a lot of (possibly) wasted effort on model “improvements”. In the absence of information to the contrary, the measurements will be assumed to have no errors, no bias, and to be an exact representation of reality. What are model evaluation goals? Operational (forecast): Is it giving the right result? Diagnostic (Control strategy): Is it giving the right result for the right reason? Environment Environnement Canada Canada

Wish list for Control Strategy / Diagnostic Model Development: Process-Specific Studies at multiple sites with: Size-resolved PM: SO4, NO3, NH4, Na, Cl, H, Ca, Si, OC, EC, other elemental cations, anions, metals, at high time resolution. VOCs: (1) High time resolution on reactive O3 precursor classes. (2) SOA precursor classes speciated (detailed speciation for members of Aromatics, Alkenes, Monoterpenes, etc.). (3) Oxygenated first and second level products for the above. OC PM speciation: PM measurements analyzed using New Directions methods (see yesterday’s talks) to get Humic-Like Substances, polymer chains, and their precursors identified. Transport: Aircraft as well as surface-based measurements, Lidar for determining aerosol layering structure. How well does the met. model simulate local as well as synoptic meteorology? Meteorology: Environment Environnement Canada Canada

Wish list for Process Studies: Three studies that would help reduce uncertainties in AQ models: Multi-species (VOC and organic) gas deposition flux and deposition velocities. (2) Hulis and polymer formation – lab and ambient, with New Directions instrumentation. (3) Detailed emissions speciation for the top 80% of mass of primary PM, VOC, inorganic gas emissions. Environment Environnement Canada Canada

Fundamental process measurements are necessary for Summary: Fundamental process measurements are necessary for model construction and evaluation of components. Monitoring networks: good for forecasting/operational use and identifying broad-scale problems, but still limited in terms of time and speciation, hence limited for diagnostic evaluation. Measurement Intensives / diagnostic measurements: necessary to work perform diagnostic evaluations for control strategy runs. Environment Environnement Canada Canada