Louisville 2006 PM2.5 Season Forecasting Case Study

1
TITLE Louisville 2006 PM2.5 Season Forecasting
Case Study The Forecast Predictor Variables
Analysis of the Relationship between the Highest
PM2.5 Episode (July 18-20) and A Mid-Atlantic
Coastal Tropical Storm (Beryl)   AUTHORS   Arthur
F. Chang and Cynthia H. Lee Louisville Metro
Air Pollution Control District (APCD),
Louisville, Kentucky, 40204   Steve Sherman
Indiana Department for Environmental Management
(IDEM), Indianapolis, Indiana, 46204   ASSIGNED
NUMBER 16
2
ABSTRACT   In this case study,
we continue to analyze our PM2.5 two new forecast
predictor variables for our highest PM2.5
concentrations in 2006 (July 18-20). These
highest PM2.5 pollution predictor variables are
surface synoptic systems interaction between 1) a
pending frontal passage and 2) a Mid-Atlantic
coastal tropical storm system. They
seem to follow 2005s predictor variables for the
highest PM2.5 episode in 2005. Base on the
coordinates from the ten (10) highest PM2.5 data
points of two storm systems (Beryl and Ophelia),
the latitude (N) ranges from 28 to 40 and
longitude (W) spans from 70 to 80. By
adding these predictor variables to consider the
forecast of very high concentration episodes, it
provided our agency better accuracy to timely
call Air Quality Alert (AQA) to reach the public
and sensitive groups to avoid the exposure when
the forecasted Air Quality Index reaching
Unhealthy to Sensitive Groups (USG when PM2.5 gt
40 ug/m3). We studied and compared
the information among synoptic surface conditions
and maps of East U.S. surface meteorological
hourly observations by Louisvilles NWS (National
Weather Services) station HYSPLIT model
forecast backward trajectory by NOAA GOES
Aerosol/Smoke products and the coordinates of
Mid-Atlantic coastal tropical storm systems
(Beryl-2006 and Ophelia-2005) from Tropical
Cyclone reports summarized by NHC (Nation
Hurricane Center). We identified that
the relative location of a Mid-Atlantic Carolina
Coastal Tropical System to a pending frontal
passage is contributing to the severe PM2.5
stagnation (single city, an area, or/and
regional). Plus upwind transport by
recirculation can be demonstrated by HYSPLIT
backward trajectory plots and contributed to
these PM2.5 pollution built-ups.
3
ABSTRACT (contd)Our highest 2006 PM2.5
episode had three days PM2.5 FRM maximum
concentrations (ug/m3) of 39.6, 39.3 and 48.2
associating with TS Beryl latitude/longitude
coordinates of (-32.3,73.3), (-34.5, 73.7) and
(-37.4, 73.2), respectively. The 48.2 ug/m3 was
the 1st maximum in 2006 Louisville PM2.5 FRM data
set to date.From an earlier analysis of the
2005 hurricane season (the most active in record)
with PM2.5 concentrations, several tropical
storms systems (Franklin, Harvey, Irene and
Ophelia) were identified to associate with high
Louisville PM2.5 concentrations. Ophelia had the
most pronounce impact on PM2.5 by triggering the
highest PM 2.5 episode that included the 2005 1st
maximum concentrations of 48.8 ug/m3 on September
9, 2005. The six days from September 8th to13th
had concentration (ug/m3) of 43.5, 48.8, 45.9,
47.8, 40.1 and 42.9, respectively.The
similarities between these two tropical systems
are 1. A pending surface cold front passage
that the movement of the cold front was
stalled by a Mid-Atlantic coastal tropical system
(Ophelia-05 Beryl-06) . 2. Very high PM2.5
concentrations (including the 1st PM2.5 annual
maximum) associate with both TS Beryl and
H. Ophelia tropical systems with similar
coordinates that were off Mid-Atlantic Carolina
coast lines. 3. HYSPLIT backward trajectory
from the east and southeast showed vertical
mixing and subsidence.The similarities
derived from this case study had been used for
reviewing the1st- 4th maxima of Louisville PM2.5
FRM data sets for 2001-2004 (See Table 1 2 for
more correlations) and will be used for future
PM2.5 forecast model development.
4
METHODOLOGY

• 1. Louisville Metro FRM PM2.5 Monthly Average
  Trends (1999-2006) Figures 1 and 2
• Louisville Metro Federal Reference Method (FRM)
  PM2.5 network has four PM2.5 sites along with one
  collocated site. FRM PM2.5 network consists of
  five RP 2025 FRM samplers. Each monthly average
  was derived from all runs of five samplers from
  that month in Louisville Metro. In Figure 1, it
  displays seven (7) years of monthly averages from
  1999 to 2006. In Figure 2, the month-to-month
  average comparisons were performed between 2005
  and previous six year averages of respective
  monthly averages (1999-2004).
• Daily FRM and TEOM PM2.5 Time Series Plots
  (Summer 2005) Figures 3 4
• FRM PM2.5 daily maximum concentrations were
  derived from five FRM samplers. In this study,
  the Maximum Values Reports (AMP440) from EPA-AQS
  database were used for Louisville Metro PM2.5
  maximum values. TEOM (Tapered Element
  Oscillating Microbalance) PM2.5 daily maximum
  concentrations were derived from four samplers.
  Both daily FRM and TEOM maximum concentrations
  were plotted for the summer of 2005.
• The tropical storms and hurricanes information
  were downloaded from National Hurricane Center
  (NHC). The locations (A-Atlantic Ocean, G- Gulf
  of Mexico, and E- end of storm or hurricane) and
  the remnant of each storm or hurricane after
  landing (L-Landing) were extracted from the
  NOAAs storm tracking tables. The Figure 3
  displays storm names and types along with time
  series plots of both FRM and TEOM for the summer
  of 2005.
• Surface Weather Maps and PM2.5 AIRNow Maps
  Figures 6, 8, 10,11 and 7
• With permission from NOAA Unisys Corp., the
  surface weather maps were downloaded for studying
  the scope of interaction between each cyclone and
  stalled frontal movement.
• Backward Trajectories Maps - Figures 12 13
• The NOAA HYSPLIT model was used for these high
  PM2.5 episode days. Three level of trajectories
  were generated by NAM forecast initiations e.g.
  500, 1000 and 1500. Either 24 hours or 36 hours
  backward time duration were used for generating
  backward trajectories.
• 5. Tropical Storms/Hurricane Track Information
  Figures 5, 7, and 9
• This study uses tropical storm and hurricane
  tracking data (see Figure 5 7) from Unisys
  Corp. (with permission) or National Hurricane
  Center (NHC). The force winds profile product
  (Figure 9) of NHC delineates the wind impact by
  storms or hurricanes. Figure 9 shows the outer
  limit of force winds by TS Beryl on 20th of July
  2006. This experimental product from NHC can
  potentially be used to identify a stagnated area
  that is outside the outer boundary of force winds
  of each storm.

5
Figure 1. Louisville Metro Monthly PM 2.5
Averages Historical Trend 1999-2006
6
Figure 2. Unique FRM PM2.5 Monthly Averages Peak
for 2005 - September
2005 TROPICAL STORMS SEASON ALTERED
NORMAL LOUISVILLE METRO FRM PM2.5
MONTHLY AVERAGES TRENDS (1999 2005)
30.0
DECREASED PM 2.5
INCREASED PM 2.5
Hurricane DENNIS
7/5-10/05
Hurricane OPHELIA
9/6-18/05
25.0
Remnant over
KY-IL border
with rain for
3
stalled on 11th (
48.9 ug/m
) over
Cape
more than a week.
Hatteras N. Carolina
caused long
--See surface map and PM2.5 daily
stagnation for Louisville.
tracking chart for details (see Fig. 3 4)
--See surface map and PM2.5 daily
tracking chart for details (see Fig. 3 4).
20.0
6 yr Avg
CONC. (ug/m3)
15.0
05 M Avg
10.0
5.0
0.0
1
2
3
4
5
6
7
8
9
10
11
12
6 yr Avg
14.8
14.3
13.1
12.9
16.0
19.9
24.1
22.2
17.0
15.3
15.8
15.9
05 M Avg
12
17.1
14.3
12.9
14.9
20.5
19.7
21.1
23.4
16.4
12.7
14.8
MONTH
7
Figure 3. 2005 Summer Time Series Plots of
Louisville Metro Daily PM 2.5 FRM and TEOM
Fall (9/22/2005)
Summer (6/22/2005)
H1 OPHELIA9/6A-9/14Carolina-17E/05
Louisville Metro 2005 Tropic Systems Impacted on
PM2.5 FRM and TEOM Data
TS IRENE
H Nate
Unhealthy for Sensitive Groups (USG)
8/9- 8/15/05
9/05 -9/10
NO
6/21/05-
influence by TS
TS Franklin
TS Harvey
H3 Dennis
Moderate
H5 KATRINA
7/21 -7//29
8/02 -8/08
7/05 -7//10
8/20A-8/25G-28L-31E/05
H Arlene
Remnant (light
6/10 -6/12/05
color arrow)
over KY-lL
7/10 -
PM2.5 Concentration, ug/m3
7//17
raining daily
Good
USG
FRM_Max
Good
Obs. Midnight-t-M TEOM
31-
2-
4-
6-
8-
10-
12-
14-
16-
18-
20-
22-
24-
26-
28-
30-
1-
3-
5-
7-
9-
11-
13-
15-
17-
19-
21-
23-
Jul
Aug
Aug
Aug
Aug
Aug
Aug
Aug
Aug
Aug
Aug
Aug
Aug
Aug
Aug
Aug
Sep
Sep
Sep
Sep
Sep
Sep
Sep
Sep
Sep
Sep
Sep
Sep
Day of Month
8
Figure 4. 2006 Summer Time Series Plots of
Louisville Metro Daily PM 2.5 FRM and TEOM
9
Figure 5. 2006 TS BERYL Best Track (18-21,
July) Blocking and Rolling Blocking
10
Figure 6. Louisville 1st Max of 2006 (7/20) -
48.2 ug/m3 TS Beryl (-39,72.6)
11
Figure 7. 2006 TS BERYL Best Track- Blocking near
Carolina Coast
12
Figure 8. Louisville 1st Max of 2006 (7/20)-
48.2 ug/m3 TS Beryl (-37.4,73.2)
13
Figure 9. Louisville 1st Max of 2006 (7/20)-
48.2 ug/m3 TS Beryl (-37.4,73.2)
14
Figure 10. Louisville 2nd Max of 2006 (7/18)
39.6 ug/m3 TS Beryl (-32.8,73.4)
15
Figure 11. Louisville 2nd Max of 2006 (7/18) -
39.6 ug/m3 TS Beryl (-32.8,73.3)
16
Figure 12. Louisville 2nd Max of 2006 (7/18) -
39.6 ug/m3 Backward Trajectory
17
Figure 13. Louisville 4th Max of 2006 (8/25) - 38
ug/m3 Subsidence and E/SE Vertical Mixing
18
Table 1. 2001-2006 Correlations Between Tropical
Systems and Louisville PM2.5 FRM Annual 1st to
4th Maximum Readings
19
Table 2.Tropical Cyclone Location (Lat/Long)
Statistics of 2001- 2006 Relating to
LouisvillePM2.5 1st to 4th Annual Maximum
20
GOES Aerosol/Smoke Product (GASP)The GASP
product is a retrieval of the Aerosol Optical
Depth (AOD) made from the current GOES East
visible imagery

• July 17 Louisville Pre-July 18 -20 PM2.5
  Episode (high AOD in TN, AL, MS, GA, SC)
• July 19 Louisville had 3rd annual PM 2.5
  maximum of 39.3 ug/m3.

21
7/17/2006 GOES Aerosol/Smoke Product (GASP)
22
7/19/2006 GOES Aerosol/Smoke Product (GASP)The
GASP product is a retrieval of the Aerosol
Optical Depth (AOD) made from the current GOES
East visible imagery
23
CONCLUSIONS This 2006 Louisville PM2.5
case study has illustrated the impact of Tropical
Storm Beryl (TS Beryl) on our annual maximal
(e.g. 1st , 2nd and 3rd max.) concentrations.
They were 1st Max (48.2 ug/m3) -7/20/2006, 2nd
Max (39.6 ug/m3)-7/18/2006, and 3rd Max (39.3
ug/m3)- 7/19/2006. The NOAA Hurricane Prediction
Center tracked and summarized the activities of
this tropical system from July 18 to July 20,
2006. By studying both Beryl in 2006 and Ophelia
in 2005, the finding of the longer resident time
of Carolinas coastal tropical systems associated
with the highest PM 2.5 readings in Louisville.
By extending this findings to applying
to the maximal data set of 2001 2004, it again
shows a strong association ( 20/24 83 with 2
maximum from winters in 2001 and 2003 See Table
12 for details) with Louisville annual maximal
values. The statistical analysis of the latitude
and longitude data for these tropical systems is
listed in Table 2. It shows significant
geographical affinity to the Carolinas
Mid-Atlantic areas, namely the Mean is ( 32.3,
71.9) and the Median is (-32, 75).
Further studies the relationship of annual
maximal values with surface synoptic conditions
and type of NOAA HYSPLIT trajectories, this case
study learned that three strong predictor
variables can be identified and utilized for
future 1st- 4th maximum PM2.5 episode forecasts.

24
CONCLUSIONS (Contd) The three strong predictor
variables are 1) a pending frontal passage,
2) blocking or rolling blocking by
Carolinas coastal tropical systems or even a
stalled low pressure system (e.g.
8/20/03-1st max.,09/09/03-2nd max.,and
8/25/06-4th max., and 3) backward
trajectories have a) vertical mixing, b)
east, southeast and south re-circulation of
polluted air mass, and c) subsidence to
associate with a Louisville and regional
stagnation.
25
Credits

• UNISYS Corp.
• NOAA/National Hurricane Center
• NOAA/National Weather Services - Louisville
• AIRNow/EPA
• NOAA Environmental Satellites, Data and
  Information Service (NESDIS)
• Dr. W. Geoffrey Cobourn, Mechanical Engineering
  Department, University of Louisville

26
References  1. Arthur F. Chang, Cynthia H.
Lee, and Steve Sherman. US EPA 2006 National Air
Quality Conference, Observations of 2005
Tropical Storms and Hurricanes Impact on
Louisville Metro PM 2.5 Pollution Forecasting,
2006, Poster 14   2.      Hubbard, M.C.
Cobourn, W.C. Development of a Regression
Model to Forecast Ground-Level Ozone
Concentration in Louisville, KY Atmos.
Environ. 1998, 32, 2637-2647.   3.      Cobourn,
W.G. Hubbard, M.C. An Enhanced Ozone
Forecasting Model Using Air Mass Trajectory
Analysis Atmos. Environ. 1999, 33,
4663-4674.   4.      Cox, W.M. Chu, S.
Meteorologically Adjusted Ozone Trends in Urban
Areas A Probabilistic Approach Atmos.
Environ. Part B. Urban Atmos. 1993, 27,
425-434.   5.      Bloomfiedl, P. Royle, J.A.
Steinberg, L.J. Yang, Q. Accounting for
Meteorological Effects in Measuring Urban Ozone
Levels and Trends Atmos. Environ. 1996, 30,
3067-3077.   6.      Tropical Storms and
Hurricanes Data Available at National Hurricane
Center http//www.nhc.noaa.gov/archive/2005/inde
x.shtml (accessed 2005 and 2006)   7.      PM 2.5
2003 2005 Mapping Data Available at EPA
AIRNow http//www.airnow.gov/ (accessed 2005 and
2006)   8.      U.S. Surface Weather Mapping
Data Available at UNISYS Corp
http//weather.unisys.com/surface (accessed 2005
and 2006)   9.      Louisville Local Weather
Data Available at Louisville National Weather
Service (NWS) http//www.srh.noaa.gov/data/fore
casts/OHZ032.php?warncountyOHC151cityLouisville
(accessed 2005 and 2006) 10. NOAA
Environmental Satellites, Data and Information
Service (NESDIS), GOES Aerosol/Smoke Product
(GASP) http//www.orbit.nesdis.noaa.gov/star/AQ_i
ndex.php