Transcript A Brief Intoduction to Air Quality Data Analysis (ppt file)

A Brief Introduction to
Air Quality Data Analysis
Susan S. G. Wierman
Executive Director
Mid-Atlantic Regional
Air Management Association (MARAMA)
[email protected]
About me
Outline & Key Points
• Motivation
o The purpose is to understand the
world & answer questions
o Analytical skills can help you land a
job
• Conceptual Model
o Understand the data –put your
study in context
• Examples
• Available data, tools,
and resources
o Basic/intermediate/advanced
o Air quality measurements and
estimates of emissions
o Models, techniques, more
examples
Why air quality data analysis?
• The analysis of air quality data can provide
important insights into the formation and
transport of air pollution
• Air quality
–
–
–
–
Affects public health
Can damage plants and ecosystems (e.g., acid rain)
Causes deterioration of buildings, monuments, etc.
Contributes to climate change
Sample of air pollution health effect study
Tracking CO2 and Climate Change
The Economist March 6 ,
2010 p. 99
Prof. Ellis B.
Cowling,
University
Distinguished
Professor AtLarge,
NC State U.
"The purpose of research
is to create simple
declarative sentences that
tell the truth."
• "Truth is a perception of reality that is consistent
with all relevant evidence and not contradicted by
any important evidence.“
• One of the traps for measurement campaigns is
insufficient analysis and interpretation compared
EPA Center on
to data collection efforts. The solution includes
Airborne
Organics: 1996
involving analysts in the measurement phase of
Summer
the program, realizing the need for "customers"
Symposium
who derive value from analysis and interpretation
Report
of data, and the formulation of specific science
and policy questions that are to be addressed by
specifically designed measurement systems.
What is the role of analysis?
Decisions
(interpretations, conclusions, guidelines, regulations
Raw data
Science
(knowledge, examples,
expertise, theory)
(observations, methods)
Analysis
(Integration, reanalysis, modeling,
inputs from other systems)
Processed data
(calculated,
normalized, value-added)
Getting a job
• For yourself:
– Does the purpose and function of the employer
interest you?
– Can you communicate with and understand the
people you’ll be working with?
– What are they working on & how can you help?
• Be prepared to demonstrate:
– What analysis tools and techniques have you
used?
– How well can you write and speak?
Outline & Key Points
• Motivation
o The purpose is to understand the
world & answer questions
o Analytical skills can help you land a
job
• Conceptual Model
o Understand the data –put your
study in context
• Examples
• Available data, tools,
and resources
o Basic/intermediate/advanced
o Air quality measurements and
estimates of emissions
o Models, techniques, more
examples
Conceptual Model is First Step
1. Develop a conceptual description of the
problem to be addressed
– You must understand something about the data
you are analyzing
– Formulate relevant science and policy questions
Conceptual Model
• A conceptual model is a summary or what we
know and don’t know about the system we
want to study
• A conceptual model includes a statement of
relevant questions we want to answer
Who/what/where/when/why/how?
•
•
•
•
•
•
Who or what causes air pollution?
Where and when is air pollution a problem?
For whom is it a problem?
Why does air quality get better or worse?
What is good air quality? What’s our goal?
How do we measure progress in solving air
pollution problems?
• What don’t we know about air pollution?
Activities cause emissions of air pollutants,
Pollutants affect air quality, Air quality affects health & welfare
Human
Activity
Pollutant
Emissions
Health and
Welfare
Air Quality
Air quality results from complex interactions of many
processes, both natural and human-made
Human
Activity
Pollutant
Emissions
Natural Systems,
Weather,
Climate
Health and
Welfare
Air Quality
One Air Pollutant—Ozone
• Ozone is formed from a mix of pollutants
via sunlight-driven reactions
volatile
organic
compounds
(VOC)
+
nitrogen
oxides
(NOx)
Sunlight
(time
passes)
Ozone
(O3)
– Summer is sometimes called the Ozone Season
– Chemical reactions take time, so ozone forms
downwind of sources of VOC and NOx
17
Ozone is a “secondary” pollutant
Primary pollutants come directly from sources
Secondary pollutants are formed in the atmosphere
18
Ozone: “Good up high, bad nearby”
Ozone layer: we want ozone here for UV protection
Smog: we don’t want ozone, as it’s bad to breath
-- ozone layer
-- smog
19
NOx Emissions Sources (NC 2007)
Area (2%)
Non-road (23%)
Point (44%)
Mobile (31%)
20
VOC Sources (NC, 2007)
Area (7%) Non-road (2%)
Mobile (3%)
Point (3%)
Biogenic (85%)
21
Another Air Pollutant—Particles
• Particulate matter (PM)
– Particles in the air include solids and liquid
droplets.
– Particulate matter is a complex mixture of
many substances!
• Aerosol particles are both primary (resulting
from direct emissions to the atmosphere) and
secondary (formed in the atmosphere).
22
Primary Particle Sources
• Power plants, factories, diesel exhaust,
construction sites, unpaved roads, wood
burning, agriculture sites, forest fires, etc.
23
Secondary Particle Sources
• Gases from burning fuels, ammonia from agricultural and
other sources, and other organics react with sunlight and
water vapor and are chemically transformed into particles
24
E.g.: Formation of Sulfate Particles
• Sulfate particles are an important component
of particle pollution in the Mid-Atlantic region
– Sulfates are formed in the air through reactions
ammonia
+
sulfur
dioxide
Ammonium
Sulfate
particles
25
Particle Size Matters
• Smaller particles seem to have more serious health and
environmental impacts.
Human hair cross section (70 µm)
Coarse Particles (PM10)
• Size: 2.5 to 10 μm
• Smaller than a human hair
Fine Particles (PM2.5)
• Size: < 2.5 μm diameter
• Greater health concern
PM10
(10 µm)
PM2.5
(2.5 µm)
M. Lipsett, California Office of
Environmental Health Hazard
Assessment
26
Regional SO2 Sources by State, 2002
1.2
Millions tons/y
1
Point
Mobile
Non-Road
Area
0.8
0.6
0.4
0.2
0
Delaware
District of
Columbia
Maryland
New Jersey
Pennsylvania
North
Carolina**
Virginia**
West Virginia**
27
Weather influences Pollution Transport
– Wind
Mixes and disperses air pollution
Moves clean or polluted air
from “upwind” regions
into the local area
– Temperature
Influences vertical movement
and mixing of pollutant
– Precipitation
Cleans the atmosphere of pollutants
28
Stagnation Increases Pollution
• Stagnant air allows air pollution to
accumulate.
• These conditions exist for many
reasons, but a couple common ones
include:
– High pressure system can produce very
stagnant air, often for several days in a
row
– Clear evenings can produce stagnant air
that usually dissipates after sunrise
Source: NPS (2006), http://69.41.173.145/ww/www2.nature.nps.gov/air/webcams/parks/nacccam/washcam.cfm
August 1, 2006, 2 pm
August 5, 2006, 2 pm
29
Three Transport Patterns
• Westerly: Prevailing winds bring
pollution from the Midwest.
Winds turn northward after
crossing the Appalatians.
• Inter-city transport: The heat
island effect of cities causes
pollution to rise. As it moves
downwind it cools and returns
to the surface.
• Recirculation: Winds can
reverse and bring pollution back
over a source region, adding
current and previous pollution
30
More info.
• MARAMA’s report, A Guide to Mid-Atlantic
Regional Air Quality, at:
http://www.marama.org/reports/GuideMidAtlantic_RegAQ_Final.pdf
• EPA’s AirNow “Air Quality Basics” links at
http://www.airnow.gov/
Outline & Key Points
• Motivation
o The purpose is to understand the
world & answer questions
o Analytical skills can help you land a
job
• Conceptual Model
o Understand the data –put your
study in context
• Examples
• Available data, tools,
and resources
o Basic/intermediate/advanced
o Air quality measurements and
estimates of emissions
o Models, techniques, more
examples
Simple – Check data against goals
• What air pollutant concentrations are of
concern?
– How does air quality in my area compare to those
concentrations?
• How do air pollutant concentrations change
spatially and by time of day, day of week,
season, and year?
What concentrations are significant?
AQI
Air Quality
0-50
Good
51-100
Moderate
101-150
Unhealthy for
sensitive people
151-200
Unhealthy
201-300
Very unhealthy
>300
Hazardous
• The Air Quality Index (AQI) is
like a yardstick for measuring
many kinds of air pollution
on one scale.
• A higher AQI means worse
air quality
– If the AQI is over 100, air
quality can cause health
problems for sensitive
individuals
34
Ozone concentration & the AQI
Air Quality Index
(no units)
• Good (up to 50)
• Moderate (51-100)
• Unhealthy for Sensitive
Groups (101-150)
• Unhealthy (151-200)
• Very Unhealthy (201-300)
• Hazardous (301-500)
Ozone (ppm, 8 hr average)
 Good (0 – 0.064)
 Moderate (0.065 – 0.084)
 Unhealthy for Sensitive
Groups (0.035 – 0.104)
 Unhealthy (0.105 – 0.124)
 Very Unhealthy {0.125 (8 hr)
– 0.404 (1 hr)}
 Hazardous {0.405 – 0.60 (1
hr)}
Daily fine particles & the AQI
Air Quality Index
(no units)
• Good (up to 50)
• Moderate (51-100)
• Unhealthy for Sensitive
Groups (101-150)
• Unhealthy (151-200)
• Very Unhealthy (201-300)
• Hazardous (301-500)
PM2.5 (24-hour – µg/m3)
 Good (0 – 15)
 Moderate (>15 - 40)
 Unhealthy for Sensitive
Groups (>40 - 65)
 Unhealthy (>65 - 150)
 Very Unhealthy (>150 –
250)
 Hazardous (>250 – 500)
Data analysts developed the AQI to make it easier for people to understand air quality.
Where and when is air quality of concern?
www.airnow.gov
EPA’s AIRNOW website is an easy place to find
the AQI for major cities
Current conditions and Forecasts
37
Examples of Data Analysis
• Does ozone exceed (i.e.,
violate) the new 8-hour
ozone standard more
often than the old 1-hr
standard?
• Are the same areas
affected?
• Do the episodes of bad
air quality last longer
during the day and last
more days?
Examples of Data Analysis
Frequency
chart - days
exceeding 8hr and 1-hr
ozone
standards by
month in the
Mid-Atlantic
region, 1997
More Questions to consider
• What’s causing the air pollution?
• What reasons might explain trends or changes
in the data?
Examples of Data Analysis
• What can we learn about
the sources of fine
particles in rural areas
and urban areas by
analyzing the chemical
constituents found on
filters from monitoring
stations?
• Are the sources similar
throughout the Midwest,
Mid-Atlantic, and
Northeast regions?
Examples of Data Analysis
• The probability of a high
source day (top 10%)
given the hourly wind
direction for PM source
8, the incinerator
source, at Acadia
National Park.
• The + indicates the
wind direction with
probability significantly
greater than 10%.
More Questions to consider
• How do I ensure the data I plan to analyze are
of good quality?
• How are monitoring or analysis methods
affecting the data?
Examples of Data Analysis
• How do two different
methods of measuring
concentrations of PM2.5
compare to each other?
• What are the reasons for
any differences?
• Can the cheaper method
that gives continuous
data (TEOM) be used to
estimate the more
expensive Federal
Reference Method (FRM)
required for compliance?
Examples of Data Analysis
Outline & Key Points
• Motivation
o The purpose is to understand the
world & answer questions
o Analytical skills can help you land a
job
• Conceptual Model
o Understand the data –put your
study in context
• Examples
• Available data, tools,
and resources
o Basic/intermediate/advanced
o Air quality measurements and
estimates of emissions
o Models, techniques, more
examples
Where to find air quality data
• AirNow, the EPA’s website for current and
forecasted air quality index (AQI),
http://www.airnow.gov
• AirData, the EPA’s website to research emissions
in your community,
http://www.epa.gov/air/data/
• Air Quality System Datamart
http://www.epa.gov/ttn/airs/aqsdatamart/
• Air Quality System (AQS) Data
http://www.epa.gov/ttn/airs/airsaqs/detaildata/
downloadaqsdata.htm
Where to find air quality data
• VIEWS – The Visibility Information Exchange Web
System – http://views.cira.colostate.edu/web/
EPA Clean Air Markets Division
Emissions Monitoring Data
Outline & Key Points
• Motivation
o The purpose is to understand the
world & answer questions
o Analytical skills can help you land a
job
• Conceptual Model
o Understand the data –put your
study in context
• Examples
• Available data, tools,
and resources
o Basic/intermediate/advanced
o Air quality measurements and
estimates of emissions
o Models, techniques, more
examples
Useful guidance & tools
• EPA Modeling Guidance (www.epa.gov/scram001)
– Guidance on the Use of models and Other Analyses for
Demonstrating Attainment of Air Quality Goals for Ozone,
PM2.5, and Regional Haze, EPA-454/B-07-002, April 2007
• Tools for using R to analyze air quality data
– www.openair-project.org (Kings College London and the
University of Leeds)
– Possible tools and analyses for the upcoming network
assessments (pdf) (tools being tested, not yet public)
(http://www.epa.gov/ttn/amtic/files/ambient/2009confer
ence/natlmonconf.pdf
Useful guidance & tools
• Field Guide to Air Quality data
– Explains the fields used in the AQS database
– Posted on the website for the AQS Datamart under
“Interpretation of Data”
• VIEWS - http://views.cira.colostate.edu/web/
–
–
–
–
–
Contour maps & Air Quality Index maps
Trends analysis
Composition analysis
US Power Grid maps
Webcam links
EPA Guidance for Regional Air Quality
Modeling (10 steps)
1. Develop a conceptual description of the
problem to be addressed
2. Develop a modeling/analysis protocol (a plan)
3. Select an appropriate model
4. Select appropriate meteorological time periods
to model
5. Choose an appropriate geographic area to
model, with appropriate horizontal/vertical
resolution, and establish the initial and
boundary conditions that are suitable for the
application
EPA Guidance for Regional Air quality
Modeling (10 steps, cont.)
6. Generate meteorological inputs to the air quality
model
7. Generate emissions inputs to the air quality model
8. Run the air quality model with initial year (base case)
emissions and evaluate model performance. Perform
diagnostic tests to improve the model, as needed.
9. Perform future year modeling, including additional
control strategies if needed to meet air quality
standards
10. Perform supplemental analysis – consider the weight
of evidence to confirm the modeled results
Put your results in context
10.Perform supplemental analysis – consider the
weight of evidence to confirm the modeled
results
– This means use more than one method – maybe
another model, maybe data analysis
– You must put your results in context – are they
reasonable? – are there unexpected results?
– Sometimes the analysis results in a change in the
conceptual model or the statistical/numerical
model
Interesting Links
• The Smog Blog http://alg.umbc.edu/usaq/ , a daily diary of
air quality in the U.S. prepared using information from
satellites, ground-based measurements, and models.
Interpretation and analysis are provided by the staff of the
University of Maryland, Baltimore County
Atmospheric Lidar Group.
• American Lung Association “Air Quality Facts”
http://www.stateoftheair.org/2009/facts/
• Calculating your Carbon Footprint (instructional outline)
http://www.ie.unc.edu/erp/resources/Calculating_Your_Car
bon_Footprint.pdf
57
Technical Reports
• Report on analysis of carbonaceous particulate matter
for New York
http://www.nescaum.org/documents/assessment-ofcarbonaceous-pm-2-5-for-new-york-and-the-region/
• Modeling mercury in the Northeast US
http://www.nescaum.org/documents/mercurymodeling-report_2007-1005b_final.pdf/
• Contributions to regional haze in the Northeast and
Mid-Atlantic States
http://www.nescaum.org/documents/contributionsto-regional-haze-in-the-northeast-and-mid-atlantic-united-states/
Air Quality Trends Analysis
• Midwest ozone trends report
http://www.ladco.org/reports/ozone/pre2008/oz
one_trends_report_january_2002.pdf
• EPA air quality trends report
http://www.epa.gov/airtrends/
MARAMA Reports
• www.marama.org/reports
Getting started: Sample questions
• How has air quality in the Baltimore area changed since
1990? (Look at one monitor for one pollutant, or look at
the AQI for the region and consider several pollutants.)
How often is the air unhealthy in each year? Unhealthy
for sensitive groups?
• What were the worst air quality days in 2007 in the
Baltimore area? (Create a timeline of days and
concentrations or AQI levels for several monitors for
ozone and for PM2.5) Are there seasonal differences in
which pollutant is highest?
• What are the characteristics of the worst air quality days
in 2007? (Look for records of temperature, wind speed
and direction, clouds, rain, month, day of week.)
More sample questions
• What were the trends in emissions from large SO2 sources
within 200 km of the Baltimore region during 2007? Was
there a correlation between daily emissions of SO2 and
concentrations of PM2.5? Was it seasonal?
• What were the daily high temperatures in Baltimore for
2007? Was there a correlation between ozone and
temperature? Compare to data for 1988 or 1995.
Illustrate with archived maps from the AirNow website.
• Map the locations of ozone monitoring sites from
Washington, DC to Philadelphia. In 2007 did all the
monitors experience poor air quality on the same days?
During the same hours? Was there a pattern that might
indicate pollutant transport?
Consider: What’s your purpose?
• Exploratory data analysis
– What patterns can I find in the data?
– What interesting questions are suggested?
– What other data would be helpful?
• Analysis for a specific reason
–
–
–
–
What questions have I been asked to answer?
Does my data allow me to answer them?
What are the limitations of the data and the analysis?
How can those limitations be addressed?
Thank you.
Any questions?