This paper describes a model based assessment of air quality impacts of
a large industrial complex development in Porto do Acu, Brazil.
Over an area of more than 100 km2, more than 350 distinct emission sources
including coal and gas fired thermal power pants with a total of 5.4 GW generating capacity,
iron and steel industry, cement and building material, automotive and miscellaneous industries,
a large harbor and a major traffic corridor are estimated to
generate total emissions exceeding 4,000 g/s NOx.
The impact assessment using "standard" tools
(USEPA AERMOD regulatory Gaussian model according to Brazilian standards)
indicates possible exceedances of air quality standards inside
(hourly NOx and daily and annual PM10 and TSP values)
but also outside the industrial district and immediate transportation corridor
area (hourly NOx values).
The results, however, are strongly influenced by the choice of models
and model configuration options, source geometry and emission heights for area sources,
available input data (primarily the meteorological observation).
There also is a mismatch between the standards (NO2) and the regulatory
model result (NOx) that leads to very conservative estimates.
Extensive sensitivity analysis including the use of alternative models
was performed to better understand and quantify the uncertainty introduced
by standard choices and data limitations, but also to identify the data
requirements to reduce this uncertainty, suggesting a more iterative approach
to impact assessment.
In a final step, we demonstrate the potential of adaptive emission control
based on real-time simulation, monitoring data assimilation,
and operational forecast as a powerful and cost-effective mitigation instrument,
to obtain "robust compliance" despite considerable uncertainty in the
original environmental impact assessment.