Template Template Meteorological Modeling Protocol for the Three States Air Quality Study (3SAQS) Ralph Morris and Bart Brashers ENVIRON International Corporation Zac Adelman and Aijun Xiu University of North Carolina at Chapel Hill 3SAQS Workshop -- CSU Fort Collins, CO May 28, 2013
19
Embed
Template Meteorological Modeling Protocol for the Three States Air Quality Study (3SAQS) Ralph Morris and Bart Brashers ENVIRON International Corporation.
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
TemplateTemplate
Meteorological Modeling Protocol for the Three States Air Quality Study (3SAQS)
Ralph Morris and Bart BrashersENVIRON International Corporation
Zac Adelman and Aijun XiuUniversity of North Carolina at Chapel Hill
3SAQS Workshop -- CSU Fort Collins, COMay 28, 2013
2
Purpose• To perform prognostic meteorological modeling to
support the development of Photochemical Grid Model (PGM) inputs– PGM modeling will analyze the air quality and air quality related
values (AQRVs, which include visibility and sulfur and nitrogen deposition) at key receptor areas across the western U.S.
• Goal is for 3SAQS PGM modeling to start in fall 2013• To achieve this schedule requires the meteorological
modeling process to start now
3
3SAQS Meteorological Modeling Leverages off of Several Recent Studies
• Denver RAQC and CDPHE MM5 and WRF sensitivity modeling (ENVIRON & Alpine, Jun 2011)– Develop improved PGM modeling procedures for Denver area– Bridge between MM5 and WRF
• WDEQ WRF modeling (Alpine & ENVIRON, Sep 2011)– 2008 annual 36/12 km and winter episode fine-scale
• WestJumpAQMS 2008 36/12/4 km WRF modeling (ENVIRON & Alpine, Feb 2012)– Use as template for 3SAQS meteorological modeling– Update to more recent model options and procedures
4
3SAQS Meteorological Modeling Steps
• Prepare Draft Modeling Protocol – Done– (ENVIRON and UNC, May 2013)
• 3SAQS Cooperators Comment on Protocol• Finalize Modeling Protocol (Jun 2013)• Meteorological Modeling (Jun-Jul/Aug 2013)• Meteorological Model Application/Evaluation
3SAQS Meteorological Modeling Protocol• Model Selection: WRF ARW
Performing better than MM5 that is no longer supported
• Episode Selection: 2011– Corresponds to NEI year and not atypical AQ/Met
• Domain Definition: 36/12/4 km– 36 km CONUS Domain (RPO, WestJumpAQMS, etc.)– 12 km WESTUS Domain (WestJumpAQMS)– 4 km Domain – Need Feedback from 3SAQS:
Option 1: WestJumpAQMS 12 km Inter-Mountain West Processing Domain (IMWD)
Option 2: Smaller 4 km domain tailored for 3SAQS
6
WestJumpAQMS 36/12/4 WRF Domain Modeling Domains
7
Alternative 3SAQS 36/12/4 km WRF Domain
8
36/12/4 km WRF Domain Definitions
• RPO Lambert Conformal Projection– Center @ (40°, -97°) & true latitudes of 33° and 45°
• 36 km CONUS Domain: ~165 x ~130• 12 km WESTUS Domain: ~255 x ~255• 4 km Domain Options:
– WestJumpAQMS IMWD: ~325 x ~525– Alternative: ~250 x ~290 (~half the size of the IMWD)
9
3SAQS 4 km WRF Domain Definition?• WJ WRF modeling w/ 4 km IMWD took 25,000 processing
hours each 5-day segment using 8 cores– Elapsed time of ~100 days using 80 cores– Can 4 km domain focus on CO, UT and WY?
10
WRF Vertical Domain• 37 vertical layers• Surface to 50 mb (~19 km
above msl)• Layers 1 & 2 = 12 m thick each• Max layer thickness = 2,000 m• For PGM layer collapsing to 25
vertical layers?– Used in WJ, Denver, Allegheny Cty– Collapse 2 WRF layers in lowest 3
layers = 24, 32 and 40– Max layer thickness = 3,900 m
Microphysics Thompson scheme New with WRF 3.1. Longwave Radiation RRTMG Rapid Radiative Transfer
Model for GCMs includes random cloud overlap and improved efficiency over RRTM.
Shortwave Radiation RRTMG Same as above, but for shortwave radiation.
Land Surface Model (LSM) NOAH Four-layer scheme with vegetation and sub-grid tiling.
Planetary Boundary Layer (PBL) scheme YSU Yonsie University (Korea) Asymmetric Convective Model with non-local upward mixing and local downward mixing.
Cumulus parameterization Kain-Fritsch in the 36 km and 12 km domains, with KF trigger option 2 or 3. None in the 4 km domain.
4 km can explicitly simulate cumulus convection so parameterization not needed.
Analysis nudging Nudging applied to winds, temperature and moisture in the 36 km and 12 km domains
Temperature and moisture nudged above PBL only.
Observation Nudging Nudging applied to surface wind only in the 4 km domain
Surface temperature and moisture observation nudging can introduce instabilities.
Initialization Dataset ERA-Interim Daily reanalysis product or 12 km NAM
WestJumpAQMS used 12 km North American Model (NAM)
14
WRF Model Performance Evaluation• Quantitative Evaluation
– Surface Temperature, Mixing Ratio and Winds– Use METSTAT and AMET evaluation tools– MADIS surface observation database– 36, 12 and 4 km domains, subregions, individual states
WestJumpAQMS went down to individual monitor
• Qualitative Evaluation– Comparison of spatial patterns of precipitation with
analysis fields based on observations (PRISM/CPC)– Comparison of spatial distribution of clouds with visual
satellite observations
15
Model Performance Benchmarks
• Met Model Performance Benchmarks originally derived after analysis of “good” MM5 performance to support air quality modeling (Emery et al., 2001)– Primarily ozone studies under simple (flat) terrain and simple
meteorological conditions (e.g., stationary high pressure) Sometimes sea breezes involved (e.g., Houston and Los Angeles)
• As part of WRAP, Kemball-Cook (2005) proposed alternative benchmarks to account for more complex conditions in the inter-mountain west and Alaska– McNally (2009) also proposed alternative benchmarks for
complex terrain conditions
16
Model Performance Benchmarks
Parameter Simple ComplexTemperature Bias
≤ ±0.5 K ≤ ±2.0 K
Temperature Error
≤ 2.0 K ≤ 3.5 K
Mixing Ratio Bias
≤ ±1.0 g/kg NA
Mixing Ratio Error
≤ 2.0 g/kg NA
Wind Speed Bias
≤ ±0.5 m/s ≤ ±1.5 m/s
Wind Speed RMSE
≤ 2.0 m/s ≤ 2.5 m/s
Wind Direction Bias
≤ ±10 degrees
NA
Wind Direction Error
≤ 30 degrees
≤ 55 degrees
Simple (Emery et al., 2001) and Complex (Kemball-Cook et al., 2005) Meteorological Model Performance Benchmarks
17
Soccer Plot Example Quantitative Evaluation
• Temperature Bias vs. Error– 36 km CONUS, 12 km
WESTUS & 4 km IMWD– Compared against Simple
and Complex Benchmarks– From WestJumpAQMS
18
Example Qualitative Precipitation EvaluationJan (top) and Jul (bottom ) CPC analysis fields (left) vs. WRF (right)
19
3SAQS WRF Modeling Next Steps
• Decide on 4 km domain• Comments on Draft Modeling Protocol
– By Monday June 10, 2013• Start setting up WRF for 2011 modeling• Limited sensitivity tests