The Environmental Benefits Mapping and Analysis Program (BenMAP)

1
The Environmental Benefits Mapping and
Analysis Program (BenMAP)
Dr. Ajay Ojha Manager Air Quality Management
Cell of PMC
2
Key Features

• BenMAP is a PC-based GIS program developed by
  USEPA
• Estimates the Health Benefits associated with air
  quality changes by creating population-level
  exposure surfaces
• Estimates Changes in Incidences of a wide range
  of health outcomes associated with ambient air
  pollution
• Places an Economic Value on these reduced
  incidences
• BenMAP produces both Point Estimates and
  Distributions of Incidence and Value Estimates

3
Steps in Analysis
Air quality changes
Air monitoring data Air modeling
data Effect estimates (epidemiology) Popu
lation data Health surveillance
data Valuation functions
Baseline conditions (air quality WITHOUT
regulation/policy scenario in place)
Control conditions (air quality WITH
regulation/policy scenario in place)
Health effects incidence estimation
Reduction in population-level exposure to air
pollution
Reduction in health effects incidence (deaths
and disease cases)
Valuation
Monetary value (benefits) of health effects
incidence reductions
4
STEP I in Analysis Air Quality Grid Creation

• Model direct use model-generated air
  concentration coverage (no interpolation)
• Monitor direct interpolate air concentration
  surface based on monitor (empirical) data
• Monitor and model relative use model data to
  support interpolation of monitor data
• Monitor rollback incremental or roll back to
  standard reduction in monitor values followed by
  interpolation

5
Air Quality Grid Creation Monitor Direct
Monitor direct spread (interpolate) monitor data
using different approaches
VNA
Population unit from Census
Interpolation
Closest Neighbor
PM2.5 monitors
6
Loading of monitor Monitoring data gathered at
20 locations from 1995 to 2004 Limitations in
missing database for regular PM10 for few
monitors 3 monitors having regular PM10 database
utilized Fixed fraction of SPM assumed to be PM10
as average from the monitored data
PM10 SPM-Estimated PM10 Monitor in 2000)
PM10 SPM-Estimated PM10 Monitor 2004


7
Generation of Control Scenario
Percentage reduce all values same percentage
required for exceedance point to achieve
standard Incremental reduce all values same
increment required for exceedance point to
achieve standard Standard - (1) associated with
the Attainment Test, which determines whether
monitor is in attainment, (2) used to bring out
of attainment monitors into attainment.
ug/m3
33 (percentage) all monitor values reduced set
percentage (e.g., 25)
DAY
10ug/m3 (incremental) all monitor value rolled
back a set increment (e.g., 10 ug/m3)
60 ug/m3 daily standard all monitors rolled back
so they achieve specified standard (e.g., 60
ug/m3 24hr daily average)
8
Estimating Exposure Creating air quality grids

• User specified area via selected grid definition
• Ward grid definition selected for Mumbai case
  study
• Assuming estimations at centre of the ward is
  good exposure estimate can be changed
• Population data available for ward level and
  wards are most common geographical features
• Monitors estimate only at one point and most
  wards do not have monitors
• Use of Voronoi Neighbor Average (VNA)
  identifies monitors near centre of each ward and
  uses inverse distance weighting
• Limiting the distance of interpolation for more
  accuracy
• Both Uniform grid and Ward grid were used for
  sensitivity analysis

9
Air Quality Grids All monitors Rolled Back to
100 µg/m3
Without Action (Basline)
With Action (Control)
10
Step II in Analysis
Baseline conditions (air quality WITHOUT
regulation/policy scenario in place)
Control conditions (air quality WITH
regulation/policy scenario in place)
Change in air quality (difference between
baseline and control air quality conditions)
Reduction in population-level exposure to air
pollution
Reduction in health effects incidence (deaths
and disease cases)
Monetary value (benefits) of health effects
incidence reductions
11
Exposure Estimates for Mumbai

2000
2004
12
Health effects incidence and benefits calculation
(OVERVIEW)
Project to future simulation year
Multiple studies for the same endpoint meta
analysis
Health impact function (lung cancer PM2.5)
Demographic (population count for tract)
REDUCTION in incidence rate of lung cancer
related deaths in population
Number of lung cancer deaths avoided in tract
PM (annual mean PM2.5)
Subtract tract-level annual mean PM2.5
concentrations
Project to future simulation year
Control air quality grid
Baseline incidence rate for lung cancer-related
mortality in study population
Baseline air quality grid
13
Health Impacts and Studies Used in Analysis for
Mumbai

• Only 3 time series studies considered several
  uncertainties may exists
• Cohort studies Pope (2002) shown different
  results and may be used

Health Effect Source Description
Premature Mortality Health Effects Institute (2004) Summary of studies in Asia
Cropper et al (1997) Analysis conducted in New Delhi
Ostro (2004, p. 9) Based on a variety of developing and developed countries
Cardiovascular Hospital Admissions Environment Protection Training Research Institute (2005, p. 341) Estimate appears to be based on U.S. studies
Cough Kumar (undated) Punjab, India
Lower Upper Respiratory Symptoms Vichit-Vadakan (2001, Table 3) Bangkok, Thailand
Minor Restricted Activity Days Ostro Rothschild (1989) United States
14
Health Impact Function

• Baseline Mortality Rate
• Studies conducted by Kumar et. Al (2002)
• Gathered mortality statistics from the Bombay
  Municipal Corporation
• Ward level population data was used gathered in
  2001 through National Census
• Three age groups
• Under 18,
• Between 18 64,
  and
• Over 64 years of
  age.

15
Generation of Health Incidence (risk) Estimates
(Based on figure developed by Abt Associates)
Mortality reduction
Air quality change
Effect estimate
Incidence rate

Population



Change in annual pollution concentration (baseline control ) (ug/m3) Percent change in incidence per ug/m3 Annual mortality rate (deaths/pop per year) People within specified age group
Mortality reduction
8 0.00402
0.01 3700
1.5 (rounded to 2) deaths avoided
16
Contd

• Hospital Admissions
• No studies available in India
• One carried by the Hyderabad Environment
  Protection Training Research Institute (2005)
• However, study based on US estimates Estimation
  to be used with CAUTION
• Incidence Rate
• No good estimates for hospital admissions
• Environment Protection Training Research
  Institute (2005) study
• The study calculates incidence rate for Hyderabad
  may have some uncertainty

17
Contd

• For Cough, Only one unpublished study by Kumar
  was available and used
• For Lower Upper Respiratory Symptoms (asthma,
  itchy eyes, cough, phlegm, and sinusitis) - study
  by Vichit-Vadakan (2001) for Bangkok, Thailand
  considered
• Minor Restricted Activity Days - Ostro and
  Rothschild (1989)
• Health data expected to introduce uncertainty
• Preferred for the case study to provide point of
  comparison between different sets of results

18
Health Impact Estimation Function Choice
DRAG and DROP
19
Step III in Analysis
Baseline conditions (air quality WITHOUT
regulation/policy scenario in place)
Control conditions (air quality WITH
regulation/policy scenario in place)
Change in air quality (difference between
baseline and control air quality conditions)
Reduction in population-level exposure to air
pollution
Reduction in health effects incidence (deaths
and disease cases)
Monetary value (benefits) of health effects
incidence reductions
20
Generation of Monetized Benefits Estimates
Unit (monetary) value
Monetized benefit estimate
Mortality reduction


Reduced number of premature deaths due to PM exposure Rupee value assigned to statistical life (based on relatively small mortality risk increments)
Benefits (rupees)
2 avoided deaths 5.5 million
11 million Example data
21
Valuation Mumbai Case Study Whole Metropolitan
Area
22
Valuation Considering Maximum Distance 5 Km
23
Running Configurations

• Reusable file that stores the choices involved in
  estimating health impacts
• Air quality grids, selected health impact
  function, population data, uncertainty
  characterization, etc
• Once created can be used many times to generate
  health impact estimates that are directly
  comparable
• Eg. 5 km maximum interpolation and no
  interpolation
• Also can be done across analyses

24
Aggregation

• AGGREGATION Convert from one spatial scale
  (grid) to another grid.
• Summarize results at spatial scale of interest
  to policy makers
• Based on area-weighted apportionment (where
  required) which is conducted within BenMAP
• Can save run speed and reduce data volume for
  results datasets.

25
Pooling

• POOLING Combining estimates together (e.g.,
  summing across age groups, or pooling results for
  the same endpoint derived using different
  studies)
• Meta Analysis (inverse weighted pooling)
  random/fixed effects pooling
• User-specified (subjective) weights
• Addition and subtraction

STUDY ID LOCATION HEALTH ENDPOINT AGE GROUP Estimated Number of Hospital Admissions
A3 Hirani ST Mortality 18-60 120
A3 Hirani ST Mortality 60 and up 200
B2 Kumar ST Mortality 18 and up 390
Different Age Groups (ADDITION)
Different Studies (same endpoints) AVERAGING
variance weighted
26
Aggregation and Pooling - Mumbai
27
Pooling Incidence Results
28
Preliminary Results of some scenarios used for
understanding BenMAP tool
Health Effect Wards Wards, 5km Max Grid Grid, 5km Max
Premature Mortality 4540E6 3120E6 4500E6 2820E6
Cardiovascular Hospital Admissions, All Ages 16.5E6 11.2E6 16.3E6 10.1E6
Lower Respiratory Symptoms, All Ages 6070E6 4070E6 6030E6 3680E6
Cough 352E6 252E6 349E6 219E6
Upper Respiratory Symptoms, All Ages 6680E6 4410E6 6640E6 3980E6
Minor Restricted Activity Days 217E6 145E6 216E6 131E6
29
Applications

• National-scale regulatory development (United
  States)
• - Cost-benefits analysis (Regulatory Impact
  Analysis) for PM and Ozone regulations focusing
  on mobile sources and power plants
• Regional/urban analysis (in support or
  local-scale policy analysis)
• Estimate benefits for PM control options (diesel
  bus retrofit)
• Include consideration for all diesel sources
  (on-road and off-road)
• Revealed harbor (shipping) as significant source
  
• International
• Number of training/collaborative initiatives
  (Mexico, China, Taiwan, Philippines, S. Korea)
• Week-long training session followed by
  policy-relevant case-study assessments
• ONGOING initiatives

30
Potential Applications in India

• Policy Research Analysis
• Actions Plans in India for 17 cities
• Evaluation of actions already taken and that of
  proposed action plans
• Risk Assessment, Cost Effectiveness Analysis,
  Cost Benefit Analysis
• Integration with the present efforts of Pune Air
  Quality Management work already in pipeline

31
BenMap Web Page Search Navigational Capacity
Mumbai Air Quality Database
Year 1987 2000 2005
Parameters SO2 SPM NOx
Air Quality
Agencies MCGB
Mobile Data Industrial Location Dumping
Site Traffic Junctions
Parameters NH3, SO2, H2S, NO2, SPM, RSPM, CO,
O3, Benzene
Health Data
Report Available
Parameters SPM RSPM SO2 NO2 H2S NH3
Year 1987 2000 2005
Agencies NEERI
Literature /Papers
Study Summaries
Additional Source
Agencies KEM Municipal Hospital
Respiratory Disease
URBAIR PMRAP ASIAN Benchmark Study
Health Report
Publication Paper Title List, Author, Subject
Area
List of References of Different Studies
Trans Thane Creek Area Industrial Data
From The project carried out by NEERI funded by
MPCB
32
Monthly Averages of Polynuclear Aromatic
Hydrocarbons in RSPM Fraction at Bombay
Source NEERI
33
Monthly Averages of Trace Metals Concentrations
in RSPM at Bombay (24 Hr. Avg. in ug/m3)
Source NEERI
34
South and Central Bombay City
Eastern Western Suburbs
Petrochemical Area
Benzo (?) Pyrene in Mumbai Area (1978 -1986)
Source Municipal Corporation of Greater Mumbai

35
South and Central Bombay City
Eastern Western Suburbs
Petrochemical Area
SPM (?g/m3) in Mumbai Area (1978 -1986) Source
Municipal Corporation of Greater Mumbai
36
BenMAP Strengths and Limitations

• STRENGTHS
• Flexibility can use BenMAP anywhere in the
  world, using own data and map files
• Data management users can easily manage their
  own data and share with others
• Speed quickly perform analyses, and generate
  maps reports
• Pooling several methods for combining health
  impact monetary value estimate
• Transparency easy to check assumptions made in
  analysis
• LIMITATIONS
• No air quality model. Air quality modeling not
  available in BenMAP.
• Loading data. Identifying and loading data
  requires careful attention to detail.
• Map files. The ability to load maps should be
  improved - takes lot of time.

37
Thank You
Available at the URL http//www.epa.gov/ttnecas1/b
enmodels.html http//www.epa.gov/ttnecas1/benmap2
download.html
Sincere Acknowledgements Ted Mac Donald,
USEPA Zachary Pekar, USEPA Donald Mc Cubbin, Abt
Associates, USA Dr. Rakesh Kumar, NEERI, Mumbai,
India Dr. Dilip Boralkar, MPCB, Mumbai, India