Title: How well can we model air pollution meteorology in the Houston area?
1How well can we model air pollution meteorology
in the Houston area?
Wayne Angevine CIRES / NOAA ESRL Mark Zagar Met.
Office of Slovenia Jerome Brioude, Robert Banta,
Christoph Senff, HyunCheol Kim, Daewon Byun
2Orientation
- Surface sites to be used for temperature and wind
comparisons - LaPorte wind profiler in green
50km
Galveston Bay
55km
Gulf of Mexico
3Orientation
- Satellite image on 1 September 2006 1137 LST
- Coasts low and sandy, little elevation change or
terrain
4Measurements and simulations
- Texas Air Quality Study II (August-October 2006)
- Surface meteorological and pollution monitoring
sites - Mixing heights and winds from a radar wind
profiler at LaPorte (on land) - WRF simulations
5How can we tell if one model run is better than
another?
- Need metrics that clearly show improved
performance - Several approaches
- Traditional bulk statistics
- Case studies
- Sea breeze and stagnation frequency
- Plume locations
6WRF simulations
- 75 days, 1 August 14 October 2006
- 5 km inner grid spacing
- Three styles
- FDDA of 3 wind profilers, reduced soil moisture,
and hourly SST - FDDA of 3 wind profilers and reduced soil
moisture - Reduced soil moisture only
- All with ECMWF initialization every 24 hours (at
0000 UTC) - Retrospective runs, not forecasts
7Impact of FDDA on wind profile
- Full run, all hours
- FDDA reduces random error in direction
- Note this is not an independent comparison (this
data was assimilated) - Red is FDDA run
- Blue has FDDA, 1-h SST, and reduced soil moisture
- Green has reduced soil moisture only
8Impact of FDDA on surface winds
- Full run, all hours
- FDDA reduces random error in direction (more
clearly seen if only daytime hours are
considered) - ECMWF has less speed bias at C35 and C45 and less
random error in speed at all sites - ECMWF has similar direction bias and random error
to WRF runs over all hours, but WRF w/FDDA is
better in daytime - Red is FDDA run
- Blue has FDDA, 1-h SST, and reduced soil moisture
- Green has reduced soil moisture only
- Black is ECMWF
9Impact of FDDA and soil moisture on surface winds
- Episode days (17) only
- Site C45, southeast of Houston very near
Galveston Bay - FDDA improves random error in both speed and
direction - 1-h SST improves random error in the afternoon,
but makes it worse at night - ECMWF has different but comparable errors, but
WRF w/FDDA is better at hours 18 and 21 (and
worse at hour 3) - Red is FDDA run
- Blue has FDDA, 1-h SST, and reduced soil moisture
- Green has reduced soil moisture only
- Black is ECMWF
10Impact of FDDA and soil moisture on surface
temperatures
- When are the errors worst?
- 10 days have at least one hour with temperature
difference gt 5K at site C35 (28 hours total) in
FDDA run - All differences gt 5K have model gt measurement
(model too warm) - All 10 days have convection or a cold front in
reality - Model also has clouds and fronts but different
amount, timing, or location
11New metricsSea breeze frequency
- How often does a sea breeze occur in the
simulation AND measurement? - Definition Northerly component gt1 m/s between
0600 and 1200 UTC and southerly gt1 m/s after 1200
UTC - FDDA or FDDA1hSST run closer to measurement at
all 7 sites (at least a little) - Results not sensitive to threshold
Red is FDDA run Blue has FDDA, 1-h SST, and
reduced soil moisture Green has reduced soil
moisture only
12New metricsNet trajectory distance
- Trajectories starting midway along the Ship
Channel at 1400 UTC each day, extending for 10
hours at 190 m AGL - WRF run w/FDDA
- Comparing total distance to net distance
- A rough measure of recirculation
- The lower left portion of the diagram is of most
interest
13New metricsNet trajectory distance
- Net distance was found by Banta et al. to
correlate well with maximum ozone - Also holds for trajectories from WRF simulated
winds, shown here - r -0.85, r2 0.72
- Run with FDDA
- Run with 1-h SST about the same
- Total distance correlation much worse (r -0.57)
14New metricsVector average wind
- Averaging u and v vs. averaging speed
- Over 10 hours 1400-2400 UTC
- Interesting points are those below the 11 line
since they have significant curvature - Run with FDDA and 1-h SST
- Correlates well with measured wind (r gt 0.9) in
either run with FDDA - Non-FDDA run not as good (r lt 0.85)
15New metricsVector average wind
- Good correlation with max ozone from airborne
measurements - r -0.91, r2 0.83
- Run with FDDA and 1-h SST
- Runs without 1-h SST about the same
- Without FDDA results are much worse
- Scalar speed correlation slightly worse(?) (r
-0.88) but still better than net trajectory
distance
16Lagrangian plume comparisons
- FLEXPART dispersion model with real emissions
- Met fields from WRF (red) and ECMWF (blue)
- SO2 measurements from NOAA aircraft (black)
- WRF result has much better resolution and plume
locations, even if averaged to same grid
17Conclusions
- ECMWF model used for initialization is already
quite good, making it difficult to demonstrate
improvement with high-resolution simulations - Traditional statistics (bias and std. dev.) dont
crisply display differences between runs,
although they generally indicate improvement with
FDDA - Different sites show different results
- Looking at distribution of errors is useful
- Large errors in temperature (gt5K) occur when
moist convection is present - New metric of sea breeze correspondence shows
improvement at all 7 surface sites with FDDA - Net trajectory distance correlates better with
ozone than total distance - Vector average wind correlates still better with
ozone, scalar average wind speed almost as good - Average wind (vector or scalar) shows clearly
that FDDA makes an important improvement under
high-ozone conditions - Improvement above the surface is easy to
demonstrate (eg. by comparison with wind profiler
data) - Lagrangian plume model provides clear information
about directly relevant performance of the model,
but how to encapsulate? - Uncertainty analysis is needed
- How good is good enough?
- What if we know we have improved the model, but
cant show that we have improved the results?
18Thanks to
- Bryan Lambeth, Texas Commission on Environmental
Quality - NOAA P3 scientists
- Richard Pyle and Vaisala, Inc. for funding
- and many others
19New metricsSea breeze frequency
- How often does a sea breeze occur in the
simulation or measurement? - Definition Northerly component gt1 m/s between
0600 and 1200 UTC and southerly gt1 m/s after 1200
UTC - FDDA or FDDA1hSST run closer to measurement at 4
of 7 sites
Red is FDDA run Blue has FDDA, 1-h SST, and
reduced soil moisture Green has reduced soil
moisture only Black is surface site measurement
20New metricsStagnation frequency
- How often does stagnation occur in the simulation
or measurement? - Definition Wind speed lt 1 m/s at any
hour between 1500 and 2300 UTC - FDDA or FDDA1hSST run closer to measurement at 3
of 7 sites
Red is FDDA run Blue has FDDA, 1-h SST, and
reduced soil moisture Green has reduced soil
moisture only Black is surface site measurement
21New metricsStagnation frequency
- How often does stagnation occur in the simulation
AND measurement? - Definition Wind speed lt 1 m/s at any
hour between 1500 and 2300 UTC - No clear improvement with FDDA or FDDA1hSST
- Results not sensitive to threshold
Red is FDDA run Blue has FDDA, 1-h SST, and
reduced soil moisture Green has reduced soil
moisture only
22New metricsSea breeze and stagnation
- Other things we can learn from these metrics
- Sea breeze correspondence is good at C45, closest
to Bay and Gulf, with high frequency - Even better sea breeze correspondence at C81 with
lowest frequency - C45 has the lowest stagnation frequency
Red is FDDA run Blue has FDDA, 1-h SST, and
reduced soil moisture Green has reduced soil
moisture only