Transcript COHA Overview

Causes of Haze Assessment
Mark Green
Desert Research Institute
Marc Pitchford, Chair
Ambient Monitoring & Reporting Forum
Causes of Haze Assessment
Goals & Objectives
 Assess causes of haze for all WRAP Federal Class I
Areas on a periodic basis – every five years
 Encourage broad-based stakeholder participation
throughout the assessment process
 Enhance the utility and accessibility of the results for
SIP & TIP development,
Regional air quality model evaluation & interpretation,
Identification of monitoring gaps,
Improved methodology for setting natural haze levels, &
Tracking effectiveness of emission control programs
Causes of Haze Assessment
Approach
Data analysis methods are selected to respond to a series
of questions concerning the causes of haze
 Will require numerous methods applied to ambient
monitoring data, but not regional air quality models
 As they become available, AMRF reviews draft responses
to each question & posts final responses to a web site
 Results are designed for computer searches, with internal
links and directories for an easily navigated virtual report

Causes of Haze Assessment Process
WRAP/AMRF
Causes of Haze Questions
Contractor
Data Analyses
WRAP/AMRF
Review Draft Results
Each Analysis Method
Addresses 1 or More
of the Questions
Separate Review &
Posting for Each
Analysis & Question
WRAP/AMRF
Post Final Results on Web
Each Question is Addressed at Each Class I Areas
Grand
Canyon
Question 1a
Question 1b
Question 1c
Question 1d
Question 2a
etc.
Mount
Rainier
Lost
Wood
etc.
Each Analysis Method Addresses One or More Questions
Method 1 Method 2 Method 3
Question 1a
Question 1b
Question 1c
Question 1d
Question 2a
etc.
etc.
Causes of Haze Assessment
Questions
 What aerosol components are responsible for haze?
What are the major components for best, worst &
average days & how do they compare?
How variable are they episodically, seasonally,
interannually?
What site characteristics best group sites with similar
patterns of major components?
How do the relative concentration of the major
components compare with the relative emission rates
nearby & regionally?
Causes of Haze Assessment
Questions - continued

What is meteorology’s role in the causes of haze?
How do meteorological conditions differ for best, worst and typical
haze conditions?
What empirical relationships are their between meteorological
conditions and haziness?
How well can haze conditions be predicted solely using
meteorological factors?
What site characteristics best group sites with similar relationships
between meteorological conditions and haze?
How well can interannual variations in haze be accounted for by
variations in meteorological conditions?
Causes of Haze Assessment
Questions - continued

What are the emission sources responsible for haze?
What geographic areas are associated with transported air that
arrives at sites on best, typical & worst haze days?
Are the emission characteristics of the transport areas consistent
with the aerosol components responsible for haze?
What do the aerosol characteristics on best, typical and worst
days indicate about the sources?
What does the spatial & temporal pattern analysis indicate about
the locations and time periods associated with sources
responsible for haze?
Causes of Haze Assessment
Questions - continued

What are the emission sources responsible for haze?
- continued What evidence is there for urban impacts on haze & what is the
magnitude & frequency when evident?
What connections can be made between sample periods with
unusual species concentrations & activity of highly sporadic
sources (e.g. major fires & dust storms)?
What can be inferred about impacts from sources in other states,
other RPOs & other countries?
What refinements to default natural haze levels can be made
using ambient monitoring and emission data?
Causes of Haze Assessment
Questions - continued
 Are there detectable &/or statistically significant
multi-year trends in the causes of haze?
Are the aerosol components responsible for haze
changing?
Where changes are seen, are they the result of
meteorological or emissions changes?
Where emissions are known to have changed, are
there corresponding changes in haze levels?
Assessment Approach
 Start with basics, sequentially increase complexity
 Most thorough effort for 35 WRAP sites with 7 or
more years data and 4 long-term CENRAP sites
 Reduced set of analyses for remaining 44 WRAP
sites and 20 CENRAP sites with <3 years of data
 Descriptive analyses, trajectory analyses, episode
analysis, cluster analysis, factor analysis, receptor
modeling, statistical tests
2 Meteorology
& Haze
2a.
Meteorology
for best,
middle, worst
days
2b. Emp irical
relationships
between
meteorology
and haze
-2c. Exp lain
best & worst
days by
meteorological
factors
2d. Site
characteristics
& relationship
between
meteorology
& haze
2e. Interannual
variation in
haze &
meteorology
Descripti ve
Trajectory
Episode
Climatology of
haze –
mesoscale,
synoptic scale
factors
Wind
speed/direction,
RH,
precipitation
and haze
Extinction rose
Residence
time,
conditional
probability
Meteorology
Cluster
Factor
Similar Days
(wind fields,
trajectories)
Wind
fields
Frequency of
clusters
Site
meteorology
Meteorological
based site
clustering
Year-toyear
Residence
time
variation
Year-to-year
variation in
met. cluster
frequency
Receptor
Modeling
Statistical
Tests
Period of record for IMPROVE
/protocol sites in WRAP region
119 of 156 visibility protected Class I areas in WRAP region, 10
in CENRAP region
 78 in WRAP and all 10 in CENRAP have IMPROVE sampler in
or nearby Class I area
 Additional 14 CENRAP sites not at Class I area
 3 Class I areas (Grand Canyon, Saguaro, and Yellowstone) have
2 IMPROVE monitoring sites
 37 WRAP and 4 CENRAP sites with relatively long-term data,
starting between 1988 and 1994
 28 sites >10 years data, 9 sites 7-9 years data (WRAP)
 Remaining sites started between 1999 and 2002 ,0-3 years data

Prepare emissions density maps
Help in interpreting the aerosol component data;
 Determine relationship of sources to the Class I areas;
 Interpreting results of backtrajectory analysis;
 To examine relationships between mesoscale
meteorological transport and efforts of the sources upon
Class I areas
 For CENRAP, need to include emissions east of
CENRAP (Midwest RPO, VISTAS)

Describe monitoring sites
 Their representation of the Class I area and
nearby Class I areas;
 Relationship to terrain features, bodies of water,
etc.;
 Proximity to major point sources, cities, etc.
Information from the emissions compilation
described above will be quite useful.
Assess meteorological setting of sites
 Expected mesoscale flow patterns of interest
(sea/land breeze, mountain/valley winds,
convergence zones, nocturnal jets, etc.);
 Orographic precipitation patterns (i.e. favored for
precipitation, or in rain-shadow);
 Inversion layers;
 Potential for transport from cities and other
significant sources/source areas.
Aerosol data analysis
Descriptive statistics and interpretation for aerosol dataindividual components and reconstructed extinction
 Document, interpret component spatial and seasonal
patterns- Best 20%, middle 60%, worst 20% reconstructed
extinction days and seasonal patterns by site
 Compile, describe spatial and seasonal patterns of
aerosol components frequency distributions.
 Interpret aerosol component data in light of emissions
sources, monitoring site settings, backtrajectories
 Cluster analysis to group sites with similar patterns in
aerosol component contributions to haze

Backtrajectory analysis
Gather backtrajectory endpoint data
Compute and map backtrajectory summary statistics
residence time by season, best 20% and worst 20%
reconstructed extinction and aerosol components for all
sites with 5 years or more of data.
 Prepare conditional probability maps for high and low
extinction and aerosol components.
 Interpret maps using emissions density, location
information, site setting information
 Mesoscale meteorological analysis needed for
many sites –backtrajectories will be misleading


Phase 1 conceptual model and
virtual report
Develop preliminary conceptual models regarding the
sources of haze at every Class I area in the WRAP and
CENRAP regions + 14 additional sites in CENRAP;
 Note uncertainties and limitations of the conceptual
models;
 Suggest methodologies to refine conceptual models in
next phase of study
 Make information available over Internet as “virtual
report”

Subsequent phases
 Compile additional meteorological, gaseous,
aerosol, emissions, and source profile data as
needed to complete remaining tasks
 Episode analysis -Use combination of
backtrajectory, synoptic, mesoscale meteorological
analysis, aerosol and emissions data to
conceptually understand regional or sub-regional
episodes of high aerosol component concentrations
In-depth meteorological analysis
 Mesoscale flow patterns affecting sites
 Cluster analysis to group days with similar
patterns and examine aerosol components for
each cluster
 Interannual variability of meteorological patterns
 Diurnal variations in flow patterns, comparison
with diurnal variation in optical data.
Emissions changes and receptor
modeling
Evaluation of changes in emissions since 1988 and
relationship to aerosol component concentration changes
 Source profile analysis- compile source profiles- note
changes over time since 1988
 Establish chemical abundances against which enrichment
factors can be evaluated
 Use carbon fractions from TOR analysis –can
contributions of different carbon sources be
distinguished?

Emissions changes and receptor
modeling -continued
 Apply Chemical MassBalance (CMB) model
 Apply Positive Matrix Factorization (PMF) at
sites with sufficient periods of record of aerosol
data
 Apply UnMix model to aerosol data for each site
with sufficient data
Trends and comprehensive
assessment
 Statistical significance tests to determine
significance of trends in component concentrations
 Interpret trends in light of trends in emissions and
interannual variability of meteorological patternsTrend due to emissions or meteorological changes?
 Comprehensive assessment of causes of haze- all
Class I areas + 14 additional CENRAP areasformulation of refined conceptual models applicable
to all areas
QUESTION Descripti ve
1 Aerosol
Components of
Haze
1a.
Co mponents
for best,
middle, worst
days
1b. Temporal
variations of
components
1c. Site
characteristics
& co mponents
1d. Co mponent
contributions
vs. emissions
near & reg ional
Trajectory
Episode
Cluster
Aerosol
Temporal
patterns
Sites,
spatial
patterns
Emissions
Spatial
emissions &
aerosol
Sites
Factor
Receptor
Modeling
Statistical
Tests
2 Meteorology
& Haze
2a.
Meteorology
for best,
middle, worst
days
2b. Emp irical
relationships
between
meteorology
and haze
-2c. Exp lain
best & worst
days by
meteorological
factors
2d. Site
characteristics
& relationship
between
meteorology
& haze
2e. Interannual
variation in
haze &
meteorology
Descripti ve
Trajectory
Episode
Climatology of
haze –
mesoscale,
synoptic scale
factors
Wind
speed/direction,
RH,
precipitation
and haze
Extinction rose
Residence
time,
conditional
probability
Meteorology
Cluster
Factor
Similar Days
(wind fields,
trajectories)
Wind
fields
Frequency of
clusters
Site
meteorology
Meteorological
based site
clustering
Year-toyear
Residence
time
variation
Year-to-year
variation in
met. cluster
frequency
Receptor
Modeling
Statistical
Tests
3 Emissions
sources
responsible
for haze
3a. Transport
patterns best,
middle, worst
days
3b. Emissions
characteristics
for hazy
transport
patterns
3c. Aerosol
characteristics
and emissions
best, middle,
worst days
Descripti ve
Trajectory
Trajectory
Spatial
analysis
Residence
time,
conditional
probability
Episode
Cluster
Factor
Receptor
Modeling
Emissions,
Source
profiles,
aerosols
Aerosol
Chemical
abundances
Extreme
events
CM B
Statistical
Tests
3d. Spatial &
temporal
pattern
analysis re:
sources
responsible
for haze
3e. Urban
source effects
3f. Sporadic
emissions
sources &
haze
3g. Emissions
outside US
and haze
3h.
Refinement of
natural
visibility
conditions
EOF
(e.g.
PCA)
Aerosol
composition,
optical
Mesoscale
transport
Emissions
Residence
time,
conditional
probability
Temporal &
spatial
patterns
Chemical
abundances,
transport
Emissions,
Transport,
aerosol
composition
CM B,
UnMix,
PMF
Time series
analysis
CM B,
UnMix,
PMF
Time series
analysisaerosol
components
4 Trends in
causes of
haze
4a. Aerosol
components
changing?
4b. Changes
fro m
meteorology
or emissions?
4c.
Emissions
changes lead
to changes in
haze?
Descripti ve
Trajectory
Episode
Cluster
Aerosol
Receptor
Modeling
Statistical
Tests
Significance
tests
Frequency of
meteorological
clusters
Emissions,
aerosol
components
Factor
Time series
analysisaerosol,
emissions