Air Quality Modeling for Teachers

PRESENTER:

Robert R. Gotwals, Jr.

The Shodor Education Foundation, Inc.

Ozone

O3 creation

normal decay

depletion

O3 per cl radical

impact of cl

Post 1994 message

O Molecules

Ozone Guess

About the Shodor Education Foundation

a non-profit Durham-based science and mathematics education group

goal: the appropriate and authentic use of technologies for science and mathematics education
- Appropriate: the technology fits the problem, rather than technology looking for a problem
- Authentic: students using real tools like real scientists

small staff, mostly former supercomputing center scientists and educators

About Shodor

operating philosophy- being “Shodorific”
- what is the incremental cost of extending materials and resources that are funded through other sources?
  - EPA is a good example!
- Mentoring and apprenticeships
  - Middle school, high school, undergraduate and graduate students

Overview of Workshop

Introduction to the technologies, techniques, and tools of computational science

Specific overview of the science of air quality and air pollution
- Photochemical smog

Use of advanced scientific tools -- air quality models (AQM) as learning and instructional tools

Lecture, “seat work”, and guided exploration on the Web

Why this workshop?

YOU
- Learn about a very important area of environmental science - air pollution and air quality modeling
- develop an understanding of how scientists understand the problem of air pollution
- Learn about technology tools - AQMs - used to study this science

ME
- Help you to learn about an important area of environmental science
- Reality check: are these tools appropriate, useful, authentic?

Overview of Workshop

Day One: Morning
- Computational Science
- Air Pollution and Atmospheric Science
- Using the Online Air Quality Modeling Tool

Day One: Afternoon
- Guided Experimentation with EKMA/OZIP

Day Two: Morning
- AQM Scenario

Day Two: Afternoon
- Presentations
- Discussion: reality check

Part One: Introduction to Computational Science

SCIENCE: the study of how nature behaves

Observational Science

Experimental

Science

Theoretical

Science

Computational

Science

Computational Science: A Tripartite Approach

What is Computational Science?

the joining of application, algorithm and architecture to solve a complex scientific problem

The "Tripartite" of Computational Science
- Applications
  - understanding the science of the problem
- Algorithm
  - creating a mathematical representation of the problem -- the "mathematical model"
  - Choosing the right numerical method to solve the mathematical model
- Architecture
  - Choosing the right platform (hardware) to solve the problem

Applications

Chemistry: electronic structure determinations

Physics: astrophysics (galaxy simulations)

Biology: population dynamics

Mathematics: fractals

Environmental Science: acid rain deposition models

Linguistics: analysis of language transference

Economics: Adam Smith models

Political Science/History: causative factors in Bosnian War conflict

Medicine: epidemiology, pharmacokinetics models

Algorithms

creating a mathematical representation of the problem -- the "mathematical model"

choosing the appropriate numerical "recipe" to solve the problem
- Examples
  - Linear Least Squares: for fitting data to a line
  - Newton's Method: for finding roots of an equation
  - Euler's Method: for solving integrals
  - Runge-Kutta Methods: for solving integrals
  - Cramer's Rule: for solving systems of equations

Architecture

choosing the appropriate "platform" to solve the problem
- single-user personal computer (IBM PC, Macintosh)
- scientific workstation (SGI Indy 2)
- workstation clusters
- supercomputer (Cray T3D MPP system)
  - scalar/serial processors
  - vector processors
  - parallel processors
  - vector/parallel machines

Simple Example: Tying your shoe

Application: what is the science of typing a shoe?

Algorithm: determining the mathematics

Assumptions:
- there are two strings
- we ignore “roundoff error” in the knot

Architecture: pencil/paper, calculator

TYS =
- X / Y
- X / Y + 0.5X / 0.5Y

Computational Science Tools

Types of Tools for solving computational problems
- Programming: Fortran, BASIC, C, Pascal
- Spreadsheets
- Equation-Solvers: Mathematica, TKSolver, Maple, MathCAD
- Dynamic Modelers: STELLA II, VenSim
- Scientific Visualization Programs: NCSA Scientific Visualization Tools, Spyglass, AVS, Wavefront
- Discipline-Specific Software: EKMA/OZIP, Models-3, UAM, PAVE, etc.

A Sample Problem: Behavior of Gases (Chemistry)

Ideal gas law: PV = nRT
- pressure times volume = amount of gas times constant times temperature

this mathematical model makes two assumptions about gases:
- gas molecules in a closed container will never bump into each other
- we can compress the molecules down to a volume of zero

need a better mathematical model for understanding the behavior of gases
- van der Waals
- Beattie-Bridgeman
- Redlich-Kwong

Van der Waals equation

takes into account the two assumptions by adding two new constants, a and b

generates a mathematical equation that is very difficult to solve analytically (using algebra)

Van der Waals equation:

the problem is to choose a way to solve this computationally
- we know the application we wish to solve
- we know the algorithms (we have a mathematical model)
- we need to select the architecture and the tool
  - architecture: PCs are powerful enough!
  - a variety of tools can be used!

Spreadsheet Implementation

Fortran Implementation

STELLA Implementation

MathCAD Implementation

BUT!! Why do we need this?

there are many interesting problems that can be solved using this technology that cannot be easily solved using traditional methods
- too tedious to solve problems using calculators
- too dangerous to try to solve problems in the laboratory
- too expensive to try to solve problems in the laboratory
- problems are only solvable using mathematical techniques or models

establish a true marriage between mathematics, computing and science

What Computational Science is NOT!

putting numbers into a spreadsheet

analyzing data gathered in the field or experimentally

fitting data to an equation

visualizing data collected experimentally or in the field

writing a computer program

using a computer to do databases, word processing or presentations

Who Cares?

21st Century Science: The Grand Challenges
- Molecular and structural biology
- Cosmology
- Environmental Hydrology
- Warfare and Survivability
- Chemical Engineering and electronic structure
- Weather prediction
- Nanomaterials

Solve any PART of one of these problems, and ….

PPT Slide

Part Two: The Atmosphere and Tropospheric Chemistry

The Atmosphere

Ozone in the Earth’s Atmosphere

Ozone Production

Precursor chemicals + hv = Ozone (O3)

NOx + VOCs + (CO) -------> O3 + PAN, etc.

NOx: a family of chemicals known as oxides of nitrogen

VOCs: volatile organic compounds that include carbon (C), hydrogen (H), and oxygen (O)

PAN: peroxyacetyl nitrates (strong irritants, toxics)

Photochemical Smog

Chemistry of Photochemical Smog

this is a partial list of the chemical reactions that describe the production of ozone and photochemical smog

numbers refer to the kinetics (rates) of each reaction

species with dots are “radicals”, highly reactive, short-lived molecules

Tropospheric Ozone Production

NO2 + hv -----> O + NO

O + O2 ----> O3 + M

OH + CO ----> H + M

OH + CO ----> H + CO2

H + O2 + M ----> HO2 + M

HO2 + NO + hv ----> OH + NO2

Net equation: CO + 2 O2 ----> O3 + CO2

NOx Emissions Sources

VOC Emissions Sources

CO Emissions Sources

Lagrangian Transport Schematic

The Air Pollution System

Generalized AQM

Box Model

EKMA/OZIP

EKMA: Empirical Kinetics Modeling Approach
- Empirical: using experimental data from the field
- Kinetics: based on rates of chemical reactions in the atmosphere

OZIP: Ozone Isopleth Plotting Program
- Isopleth: a chart showing equal (“iso”) concentrations (“pleths”) of ozone

OZIP: Ozone Isopleth Plotting Program

OZIP is an early-generation (late 70s) ozone concentration modeling program
- 25 FORTRAN programs, working in series
- only calculates ozone concentrations for a single day (vs.. multiple-day of newer models)
  - UAM: Urban Airshed Model
  - MODELS-3
- requires understanding of
  - chemical reactions of the troposphere
  - emissions inventories
  - meteorology
  - various simulation scenarios

Shodor’s Approach to EKMA/OZIP

in support of EPA, Shodor has:
- installed OZIP on its high-performance workstation (SGI)
- constructed a Web-interface with user support
- provided full documentation to the user

Sample Isopleth Chart