Better modeling through MATLAB:
Connecting the auroral oval and the convection reversal boundary
Liam Kilcommons
Mentors:
Dr. Barbara Emery
Dr. Delores Knipp
Dr. Liying Qian
Dr. Wenbin Wang
Background source: http://blog.thetechnonaut.com/?p=91
Motive, Means, and Opportunity
Granted, the aurora is
beautiful, but why bother
studying it?
You can't really get to the
aurora, so where do you
get your data?
Actually, it's pretty important: In general:
Auroral particle precipitation
associated with Joule heating,
biggest energy input (besides
solar radiation) into upper
atmosphere

Ionospheric disturbances and
accompanying electric and
magnetic fields cause:
 Power grid damage
 GPS inaccuracy
 Air travel re-route
 Pigeon loss

ground based remote sensing

direct satellite measurements

The aurora is pretty well
studied, what else is there
to know?
Lots!
Need better representation of
atmosphere response to solar
storms

Models' representation of
aurora needs to be improved
(2nd biggest energy input!)
Defense Meteorology Satellite
Program satellites – aurora data
In this case:
OMNI – satellite solar wind and
ground based data from various
sources


What's this Joule thing?
Most familiar incarnation:
Wires getting hotter due to current flow (a.k.a ohmic heating)
Qj µ vi * Flux
In upper atmosphere:
2
Neutral atoms getting hotter due to
collisions with energetic electrons and
ions
e
ve
ion
Drives 'neutral winds'

vi
Neutrals
Momentum transfer

Strong solar storms (lots of
precipitation and fast ion convection)
cause worldwide changes in neutral
winds

Auroral
precipitation: e- and
p+
Ion convection: ions
of O, O2, N2 etc

vp
p
Duskward, anti-sunward, and other mysterious directional adjectives
Graphing as if one were looking down on the magnetic north pole
Dawn
6
Noon
12
N Pole
(90º)
0/24
70º Midnight
80º
Not to scale
http://www.atigun.com/photography/aurora-photographs/47-aurora-photographs
18
Dusk
Angular coordinate = magnetic local time (MLT)
Radial coordinate = magnetic latitude (MLAT)
Ion Convection (or ion drift; the physics are the same)
6
View from the top
N
The
Earth
E
12
B
0/24
S
Dawn-Dusk E-field!
18
Looks familiar...
6
6
DMSP Satellite in a
dawn-dusk orbit
12
E x B gives direction
of vi
18
Vi?
The Ion Drift Velocity
0/24
12
0/24
vi
18
Convection
Reversal
Boundary (CRB)
Our Data Sources: DMSP Satellites
+X indicates Along-Track Direction
Y indicates Cross-Track Direction
-Z is Earthward Direction
SSJ4 – Electron Flux
IDM – Ion Drift Meter
The beginning of this presentation, finally
6
The (scientific) problem:
TIE-GCM needs to calculate Joule heating
Joule heating calculation requires relative
locations (difference in radii) of CRB and auroral 12
oval

N.P.
Positions change constantly and drastically with
changes in solar wind

TIE-GCM uses a constant (CRB 2 degrees
duskward of auroral max flux)

My goal:
Gather statistics on real life relative CRB and
aurora positions
Revise a MATLAB data display program to
display the CRB and aurora positions for a
particular pass of a particular DMSP satellite on
a particular day.

18
0/24
How I Spent My Summer Vacation
The (Programming) Problem:
Need a flexible graphical display environment to visualize many kinds of data with
different formats
Why?
'Big picture' view of sun interacting with high latitude upper atmosphere
The Specifics:
Incorporate solar data, auroral flux, and ion drift velocities
Visualize many different variables, whatever user wants
Data Name
Data Source
Date Range
(DMSP F-13)
(on site)
Frequency of
Measurements
Format
Headache
Rating
Particle Flux
DMSP SSJ4
1987-2005
Every second
Binary
High
Ion Drift Velocity DMSP IDM
1987-2002
Every 4
seconds
ASCII
Text
Low
OMNI Solar
Various:
Wind Composite ACE, Ground
Based, etc.
1978 to
Present
Every Hour
ASCII
Text
Low
No Need to Reinvent the Wheel, Just Figure Out How It Rolls
The foundation for our visualization tool was Data_Plotter:
Matlab GUI ~ 6000 lines of code
Originally designed to study neutral density response for specific solar events

Feature
Original Program: Data
Plotter
Desired Features for New
Program
Input Data Type
Heavily Preprocessed
Raw Satellite Data
Available Times
Only Selected Events,
2004,2005
Any date for which data for
desired plots exists
Available Datasets
DMSP PF, IDM, SSJ
CHAMP Neutral Density
DMSP IDM, SSJ
OMNI Solar Data
Types of Plots
Polar, Cartesian
Polar, Cartesian
Calculations
None
Convection Reversal
Boundary, Auroral Max
Flux Points
No Need to Reinvent the Wheel, Just Figure Out How It Rolls
The foundation for our visualization tool was Data_Plotter:
Matlab GUI ~ 6000 lines of code
Originally designed to study neutral density response for specific solar events

Biggest Challenge
Feature
Original Program: Data
Plotter
Desired Features for New
Program
Input Data Type
Heavily Preprocessed
Raw Satellite Data
Available Times
Only Selected Events,
2004,2005
Any date for which data for
desired plots exists
Available Datasets
DMSP PF, IDM, SSJ
CHAMP Neutral Density
DMSP IDM, SSJ
OMNI Solar Data
Types of Plots
Polar, Cartesian
Polar, Cartesian
Calculations
None
Convection Reversal
Boundary, Auroral Max
Flux Points
My Solution: DMSP_Plot.m
CRB
CRB
Max
Flux
Max
Flux
....and then, last night, around 11 pm there were: (Preliminary) Results!
Max
Flux
CRB
CRB
Max
Flux
Progress Report
Feature
Original Program: Data
Plotter
Desired Features for New Goal Met?
Program
Input Data Type
Heavily Preprocessed
Raw Satellite Data
Available Times
Only Selected Events,
2004,2005
Any date for which data for Yes
desired plots exists
Available Datasets
DMSP PF, IDM, SSJ
CHAMP Neutral Density
DMSP IDM, SSJ
OMNI Solar Data
Yes
Types of Plots
Polar, Cartesian
Polar, Cartesian
Partially;
Yes
Code still a
little buggy
Calculations
None
Convection Reversal
Boundary, Auroral Max
Flux Points
Partially;
Partial
functionality in
seperate code
Future Work
Perfect CRB and auroral maximum-flux-point finding algorithms

Gather statistics on position of aurora center and CRB center and their relative
radii as a function of solar wind parameters

Compare results to modeled results

Use results to devise empirical model for estimating center position and radii of
auroral oval with respect to CRB

Integrate model into TIEGCM for better Joule heating calculation

THANK YOU!
QUESTIONS?