UrbanEmissions (UEinfo) was founded in 2007 with the vision to be a

repository of information, research, and analysis related to air

pollution.

This primer is the 2nd edition on source apportionment.

All the primers are available @www.urbanemissions.info/publications

Send your questions and comments to [email protected]

Most of the illustrations are made using the open access “excalidraw”
portal.
 1 2 3

.. it is also important to
In addition to knowing
know
(1) how much is the
(4) the contributions of
pollution (2) where is the
various sources to the
pollution and (3) when is
air pollution problem.
the pollution…
This includes sources
within the city, as well
as those outside the
4 city’s administrative
boundary and the
designated airshed.
1
 How can we
determine how much
various sources are
contributing to the
air pollution problem?

We can determine this

through a source
apportionment study.

I already know
what are the main
sources, then why
do I need a study?

A study will scientifically

determine these contributions
and help prioritize the sources
that can lead to cleaner air more
quickly and cost-effectively.
2
 Are these studies
easy to conduct?

These studies are

scientifically robust but
never easy to conduct.
They require a lot of skilled
personnel for monitoring,
data modeling, laboratory
facilities, and financial
support. There are multiple
institutions with resources
that can help in conducting
How often do
these studies.
we have to
conduct these
studies?

Depending on the availability of resources and technical

facilities, these studies can be repeated every year. This
will allow us to study the trends in ambient pollution
levels and the impacts of various control options
implemented in the city airshed.
3
 What are the
different ways to
conduct a source
apportionment
study?

There are 2 ways

(a) Top-down (sampling way)
(b) Bottom-up (emissions way)

Are the results

from these two
methods same?
Ideally, yes.

The top-down approach provides results specific to

sampling locations, which are then averaged to
create a city-level profile.

The bottom-up approach provides results for the

city’s designated airshed, with spatial granularity
based on the available emissions inventory.
4
 Top-
Top-down approach (sampling way)

Ambient Lab-
Lab-Chemical Receptor
Sampling Analysis Modelling

This approach measures pollution using

filters, which are analysed in the
laboratory for chemical signatures.
Then the sample profiles are
statistically matched with source
profiles to assess source contributions.

These steps are

straightforward, but
each component is a
project in itself..

5
 Are there any
rules for how many
samples must be
collected?

Yes, there are some thumb rules.

e.g., India defined the following based
on total population (TP)
For TP under 100,000 -- 4 units
For TP under 1 million - 4 + 0.6 per 100,000
For TP under 5 million – 7.5 + 0.25 per 100,000
For TP above 5 million - 12 + 0.16 per 100,000
Table source: CPCB, India
For example:
City of Bengaluru requires 41 locations

Finally, the city's

financial, personnel, and
operational capacity
decides how many
locations are selected.

6
 How do you select
the sampling
locations?

If locations are not selected

representatively, the sample
results will be biased.

To minimize these biases, the

protocol suggests selecting multiple
locations representing all areas and
activities in the study airshed.

7
 MASS -- gravimetric analysis
ELEMENTS -- Na, Mg, Al, Si, P, S, Cl, K, Ca, Ti, V, Cr,
Mn, Fe, Co, Ni, Cu, Zn, Ga, As, Se, Br, Rb, Sr, Y, Zr,
Mo, Pd, Ag, Cd, In, Sn, Sb, Ba, La, Au, Hg, Ti, Pb and U
IONS -- SO42-, NO3-, NH4+, Cl-, Na+, K+
CARBON -- elemental, organic, and total

The sampled filters are stored in a

controlled environment and sent to the lab
for chemical analysis.

We require information on the total

sample weight, fractions of elements, ions,
and carbon. Each of these elements is
analyzed using different machines.

How expensive is
this step?

The lab equipment is not cheap. If the city

does not operate its own lab, samples can be
sent to a certified lab.

With the city's own lab, the study can be

replicated at many locations and over more
years.
8
 The most common
receptor models used
are CMB and PMF.

Local source profiles

are preferred. If not
available, take Here the
representative profiles term source
from the library. is often
referred to
the fuel.

9
 These are the common chemical markers of
various sources for PM2.5. The source profiles
are developed by collecting filter samples
immediately closer to the sources.

There are overlapping signatures between

sources using the same fuels. Extra scrutiny is
required in dissecting these results. For
example: road dust versus construction dust;
coal in a cooking stove versus an industrial
boiler; and diesel in a bus versus a truck versus a
generator set.
10
 A pollution map with concentrations and
source contributions at various locations can
be averaged to represent the city airshed.

11
 Bottom-
Bottom-up approach (emissions way)

Gridded-
Gridded-Emission Meteorology Chemical-
Chemical-Transport
Inventories Modelling

This approach utilizes grid-level

emission inventories of all known
sources, along with meteorology and
airshed boundary conditions, as input
to chemical-transport models to
assess source contributions.

These steps are

straightforward, but
each component is a
project in itself..

12
 Depending on the size of the airshed
and the chosen model, this approach
can be computationally demanding.
It requires personnel with training to
simulate meteorological and
atmospheric models, as well as skills in
GIS databases, mapping, and data
analytics.

13
 What are the basic
layers of information
we must gather for
this approach?

Define your area of interest. Outside your city, include as

much as possible all the areas with sources that could
influence local air quality in the short term. This area is often
referred to as the city’s airshed.

Collect information from archives on geography, land use-land

cover, population, urbanization maps, industrial maps,
residential maps, commercial activities, energy consumption
rates, vehicle registration and usage numbers, meteorological
statistics, and ambient monitoring data.

While this may seem like a lot, most cities already have this
information, which needs to be archived for final use.
14
 For urban-scale studies, typical grid-size is 1-km and typical
airshed size is 30x30 for small cities to 80x80 grids for big
cities, in north-south and east-west directions.

For larger airsheds with more grids, the grid-size increases as

a compromise to support computational needs.
15
 What is the
fundamental equation
to start building an
emissions inventory?

Depending on the level of detail in the

activity data for each sector, the
equation can become more complicated.
If detailed information is not available,
begin with total city activity data.

If local emission factors are

unavailable, use averages from a library.

However, it is crucial to start!

16
 The first step is to utilize available emission
inventories—either from a local resource or
extract from a global inventory.

Then, expand the data collection exercises

to collate activity data through primary
surveys.

Spatial allocation of the emissions to

airshed grids requires multiple layers of
information in GIS formats.
17
 3-dimensional meteorological information is
must for modelling studies. This includes
information on pressure, temperature, wind
speed, wind direction, precipitation, relative
humidity, inversion heights, and others.

Weather Research Forecasts (WRF) model is

state-of-the-art for urban, regional, and
global applications.
18
 Select an appropriate model depending on
your computational capacity and analysis
needs. Consult experts.

19
 How can we
validate the
modelling results?

A large pool of ground monitoring data means a

better understanding of the spatial and temporal
trends of air pollution in the airshed.

This information can be used to validate the

models and increase their confidence levels.

New modeling systems can also utilize satellite

observations to validate the modeling results.

20
 Excellent.

These are two distinct approaches to

reach the same conclusions.

Depending on our financial and technical

resources, we can use either of the
approaches (or both). Neither of them
is easy, but we must start the study.

Exactly.

Both methods are robust and well-tested in

cities across the world. Ideally, both
approaches should be applied to consolidate the
results.

The most important step is making the decision

to conduct the source apportionment study.
Then, based on the resources and information
available, we can decide which approach is more
suitable.
21