Low (web) Quality - BALTEX

288
Impact of megacities on the regional air quality: A South American case
study
Claas Teichmann and Daniela Jacob
Max-Planck-Institute for Meteorology, Hamburg, claas.teichmann@zmaw.de
1. Introduction
Natural as well as anthropogenic emissions determine the
aerosol and chemical composition of the atmosphere.
This has a major impact on cloud formation, on the
hydrological cycle and on air quality. In many South
American regions the effects of megacities - such as São
Paulo, Buenos Aires, etc. - are crucial and have an impact
on a regional scale. In other regions the emissions are
dominated by natural sources as well as by land-use
change and biomass burning. The goal of the study is to
estimate the impacts of different emission sources on the
regional air quality.
In South America the Andes have a significant influence
on the atmospheric circulation and on the transport of
chemical species, because of the pronounced orographic
features. An adequate representation of the Andes within
a climate model is only possible with the relatively high
horizontal resolution of a regional climate model. The
high resolution is also needed to resolve the megacities
with their highly concentrated emissions and the
corresponding chemical reactions.
2. The regional model REMO including online
chemistry and tracer transport
In this study the newest operational version of the
regional climate model REMO (Jacob et al., 2001, 2007)
is used including on-line chemistry and tracer transport.
The chemistry module is based on the second generation
Regional Acid Deposition Model (RADM2) (Stockwell et
al., 1990).
The model calculates the meteorological processes
directly together with photochemistry and tracer
transport. The advantage over off-line chemistrytransport
models – which are driven by the, e.g., hourly
output from a meteorological model – is the direct
coupling of meteorological and chemical fields, which
both are available for each model timestep.
In the current model setup, the atmospheric mechanisms
influence the chemical mechanism and the tracer
transport, while there is no feedback from chemistry and
tracer concentrations onto the atmospheric properties and
processes (see Fig 1).
Fig 1. REMO with chemistry and tracer transport
The advantage of this setup is that different case studies,
e.g., with modified emission inventories, are subject to
exactly the same meteorological conditions. Differences
in resulting trace gas concentrations can be attributed to
the chemical mechanism and to the modifications made,
e.g., to the emission inventory.
3. Boundary data
Boundary and initial data for the chemical species
concentrations is provided by global model output from
the Model for Ozone and Related Chemical Tracers
(MOZART) (Kinnison et al., 2007). Anthropogenic
emission data and fire emission data is taken from the
REanalysis of the TROpospheric chemical composition
over the past 40 years (RETRO) emission database (e.g.,
Schultz et al., 2008).
Meteorological boundary conditions are provided by
ERA-40 re-analysis data (Uppala et al.,2005).
4. Model simulations
In order to assess the impact of the different emission
sources, several model runs are performed in a case study
for the year 2000. They include the full chemistry and
tracer transport and are embedded in a hindcast which
comprises the whole ERA-40 period for meteorology
only.
A so called reference run includes the full emission
inventory, natural as well as anthropogenic emissions. In
two sensitivity runs the emissions from different sources
are modified.
Emissions from eight megacities (cities with more than
five million inhabitants) are reduced by 90% in the first
sensitivity run (denoted as reduced megacity emissions
run). In the second sensitivity run, denoted as no-fires
run, fire emissions are removed from the emission
inventory. In a comparison, the impact of the different
emission sources on the regional air quality is obtained.
5. Results
As an example, results from two sensitivity runs are
shown for the simulated April 2000. As April is not part
of the main fire season, the impact of the fire-emissions is
relatively low, compared to the maximum fire impact of
the year.
In Figure 2 the maximum weighted difference between
the sensitivity runs and the reference run of near surface
CO concentrations (lowest model layer) is shown for the
reduced megacity emissions run. Figure 3 shows the
maximum weighted difference of CO concentrations for
the no-fires run. This gives an impression of the extent of
air pollution episodes originating from megacities
compared to fire emissions and shows which regions are
affected.
As South American megacities are located mainly in
coastal regions, pollution can be transported long
distances over the ocean. This is shown for Buenos Aires
where the CO pollution plume reaches far over the
southern Atlantic with a relative impact (i.e., the increase
from the reduced megacity emissions to the reference
case) of more than 10%.