Mathematical modelling of environmental impacts has often been an important part of Impact Assessment. Modelling of air emissions, noise, thermal discharges, wastewater releases on the shore and through outfalls have been some of the common applications.
Mathematical models are used to simulate emissions, meteorology, and hydrology/hydraulics and arrive at ambient concentrations to compare with applicable standards under various scenarios. These results are then used to make decisions on stack heights, outfall length, noise barriers and design of green belts, level of control or pre-treatment of emissions needed, sulphur content in the coal etc. There are significant implications of model-based decisions on the costs.
Results of simulation depend on the model being used. In air quality modelling, for instance, results of Industrial Source Complex (ISC) model and that of AERMOD can differ by an order of magnitude! Hence, which model is used for impact assessment becomes a very important question. There have been instances in India wherein a Flue Gas Desulphurization (FGD) unit has been slapped on a Coastal Thermal Power Plant based on results of a basic Gaussian Plume Model (GPM) that did not factor-in coastal conditions (i.e. sea breeze and fumigation). Underneath the façade of science – there can be the elegance of ignorance!
In countries like the United States, Australia and EU select models are recognized as “regulatory” models and these models are recommended for impact assessment. Importantly, this list of models is regularly updated. See http://www.epa.gov/airquality/modeling.html for recommended air quality models by US EPA. The Ministry of Environment and Forests (MoEF) in India has listed a number of models for impact predictions. See http://envfor.nic.in/divisions/iass/eia/Annex5.htm. This list of models is however dated (e.g. it lists ISC-2 as one of the recommended air quality models, when today AERMOD is recommended in the United States replacing even the ISC-3). Further, no guidance is provided on the use of these models. Last publication on Guidance on Air Quality Models was in 1997-1998 by the Central Pollution Control Board (CPCB).
No research studies have been commissioned by the MoEF to test the accuracy of these mathematical models under Indian conditions and using Indian data. Have we ever conducted verification studies on model predictions?
Modelling is hardly taught at universities, so most Indian ‘modelers’ use modelling software as a black box. Plug data in and spew the results!
Against this background, how accurate the results of modelling are, is another important question. Indeed, the results of models in most cases are quite different from the observations. Reasons are several, including data inaccuracy in quantifying emissions (especially nonpoints and fugitives), inability to mimic complex meteorology/ hydraulic conditions and reactions. The differences can lead to incorrect decisions and inappropriate
I would like to draw your attention here to a very interesting observation made in Hong Kong (see http://www.legco.gov.hk/yr12-13/english/panels/ea/ea_anlp/papers/ea_anlp0628cb1-1393-2-e.pdf). This observation pertains to the importance of model accuracy or confidence and communication of modelling results regarding air quality. I am reproducing two interesting paragraphs from this note that are worth pondering over.
| We welcome the Subcommittee’s review of “Air Pollution Modelling in Hong Kong”. Hong Kong’s air quality is a topic of great public concern. The lack of progress in improving air quality has led many to question the effectiveness of our air quality management system. Most people understand that air quality modelling is important for our air quality management system because of its role in air quality impact assessment.
However, not many people understand the details of air quality modelling as modelling is relatively technical and complicated. Hence, most people prefer to treat it as a black box. People don’t like black boxes, particularly when our air quality management system does not seem to be working. Therefore, there have been growing concerns about of the accuracy and application of air quality models in Hong Kong.…….So, we have a peculiar situation: for the past 15 years, the air quality models used in the EIAs are projecting compliance with the AQOs, while observations by EPD showed continual noncompliance for Hong Kong. This apparent contradiction is an important reason why many are sceptical about the EIA process in general, and lost trust in the air quality models in particular.
In India, we include modelling more like a ritual in the EIA reports. We use out-dated models and our model application quality is poor. Results of modelling, especially model limitations, are not well communicated to the stakeholders and to the public. Model verification and validation studies are seldom carried out. If used on a mature basis, modelling indeed has a value – but we haven’t learnt the art and science of modelling in this perspective.
We need to ask MoEF/CPCB to update the list of recommended math models, provide guidance on their use and interpretation, and sponsor model verification studies. Let us also work together to offer hands-on training programmes on application of some commonly used models. Let us not make modelling a ritual.
For Students – Research on how a model reaches a state of a ‘regulatory model’ for impact assessment. Take a case study of ISC-2, ISC-3 and AERMOD in US EPA. Understand how models are field-verified and what is involved in Prediction Audits of a model.
For more serious research in this arena, I would recommend you to see an excellent article by Matthew Cashmore ‘The role of science in environmental impact assessment: process and procedure versus purpose in the development of theory’ (see http://faculty.mu.edu.sa/public/uploads/1338109315.1011EIA-8.pdf ) as a start.The list of references in this article is very comprehensive and very helpful for a beginner.