Abstract: Evan Couzo

Atmospheric models: Why you should never ever ever trust them,
why some people do, and why you will, too.

Dr. Evan Couzo (Williams Class of 2005)
Postdoctoral Associate, MIT
http://globalchange.mit.edu/about/our-people/personnel/staff_id/391

Since time immemorial, the modeling community has debated the nature and purpose of numerical models in a way that would make Plato smile. Are models real or just shadows on the wall? Must models maintain strict fidelity to natural laws or can they make pragmatic assumptions? Air quality managers and regulatory agencies don’t have the luxury of debating modeling philosophy, virtuous though it is. They need decision-making tools today and cannot wait for the body of science to be completed tomorrow. Hence, regulatory photochemical modeling takes the pragmatic approach. These models incorporate the atmospheric community’s current understanding of environmental principles and best guesses at input parameters (e.g. emissions, meteorological fields, boundary conditions). The process of model formulation necessarily introduces error and uncertainty because models cannot be identical to an open environmental system. Remaining cognizant of this fact is paramount to a responsible and justifiable use of model predictions. In the right hands, regulatory photochemical models are powerful predictive tools that assist in the development of pollution control strategies.

The first part of this talk will be an introduction to how air quality models work, ways in which they fail to represent the natural environment, and why, despite their shortcomings, they are invaluable tools. The second part will describe a recent attempt to add physical and chemical accuracy to the model vis-à-vis nitrous acid formation. Nitrous acid is an important atmospheric radical precursor (radicals control pollutant formation), and laboratory experiments and field measurements have identified heterogeneous sources of nitrous acid that are not included in modeled chemistry. We have updated an air quality model to include heterogeneous nitrous acid reactions, and demonstrate that, as a result, model performance improves and pollutant formation is enhanced.