A statistical study of plasma line enhancements
during IPY 2007-2008
May 19, 2016
Michael Hammarsten, Nicola Schlatter, and
Nickolay Ivchenko
EISCAT 3d user meeting 2016
Plasma line enhancements
Method of detection and classification
I Method of detection
I
Frequency power profile distribution based thresholding
I
I
Mean power for each frequency bin
Threshold on the mean value of the resulting normalised vector.
I Method of classification
I
Range power profile distribution based thresholding
I
I
Normalised range distribution of the peak frequency
Thresholds based on variance of that vector
I Three types of enhancements
I
I
I
Single
Profile
Diffuse
I Results
I NOTE: Centre frequency changed from ±4MHz to ∼ ±3.2MHz during mid of
December 2007
Michael Hammarsten, Nicola Schlatter, and Nickolay Ivchenko | A statistical study of plasma line enhancements during IPY 2007-2008
1
Enhancement detection - Cut 1
Frequency power profile distribution based thresholding



33
P
I
Pr 1
..
.
P11
···
.
..
···
2

Pr ν
.. 
. 
P1ν
Pr ν
r =1
33
(Mean power for each frequency bin)
Pmν =
Pm1
···
Pm768
I Normalising vector Pmν → pm
ˆ ν (Retrieving the relation between min and max)
I Threshold using the mean value of pm
ˆ ν
I
I
ˆ ν ∼ 0.5
During noise: mean of pm
ˆ ν ≤ 0.2
When enhanced: mean of pm
I
The frequency bin with the highest power is used during classification
Michael Hammarsten, Nicola Schlatter, and Nickolay Ivchenko | A statistical study of plasma line enhancements during IPY 2007-2008
Enhancement classification - Cut 2
Range power profile distribution based thresholding

Pr νp

P33νp


..

=


.
P1νp
I Normalising vector Pr νp → p̂r νp
I Classification using the variance of p̂r νp
I
I
I
Single: variance of p̂r νp ≤ T 1
Profile: variance of p̂r νp ≥ T 2
Diffuse: variance of T1 < p̂r νp < T2
I
Diffuse enhancements consists mostly of
wrong classifications and artefacts
Michael Hammarsten, Nicola Schlatter, and Nickolay Ivchenko | A statistical study of plasma line enhancements during IPY 2007-2008
3
Single Enhancement Characteristics
A small and distinct frequency and range distribution
Cut 2 - Variance of p̂r νp : Small
Michael Hammarsten, Nicola Schlatter, and Nickolay Ivchenko | A statistical study of plasma line enhancements during IPY 2007-2008
4
Profile Enhancement Characteristics
A broad, but focused, range distribution
Cut 2 - Variance of p̂r νp : High
Michael Hammarsten, Nicola Schlatter, and Nickolay Ivchenko | A statistical study of plasma line enhancements during IPY 2007-2008
5
Diffuse Enhancement Characteristics
Diffuse enhancements consists mostly of wrong classifications and artefacts
Cut 2 - Variance of p̂r νp : Medium
Michael Hammarsten, Nicola Schlatter, and Nickolay Ivchenko | A statistical study of plasma line enhancements during IPY 2007-2008
6
Downshifted Plasma Line Enhancements
Frequency and range distributions for each type
I Dumber of dumps analysed: 1,876,165 (∼173 days)
I DWS: 30,797 → 1.64% || UPS: 88,058 → 4.96%
I DWS: one peak at ∼190km || Single: ∼170km
I Main distributions at -3.4MHz to -4MHz || Single: bandwidth edges and -4MHz
Michael Hammarsten, Nicola Schlatter, and Nickolay Ivchenko | A statistical study of plasma line enhancements during IPY 2007-2008
7
Upshifted Plasma Line Enhancements
Frequency and range distributions for each type
I Dumber of dumps analysed: 1,876,165 (∼173 days)
I DWS: 30,797 → 1.64% || UPS: 88,058 → 4.96%
I UPS: two peaks at ∼190km and ∼ 230km || Single: ∼170km
I Main distributions at -3.4MHz to -4.4MHz || Single: bandwidth edges and -4MHz
Michael Hammarsten, Nicola Schlatter, and Nickolay Ivchenko | A statistical study of plasma line enhancements during IPY 2007-2008
8
Plasma Line Enhancements
Time distribution for each type of enhancement
I The main distribution of enhancements occurs at noon MLT.
I
Single enhancements occurs a few hours earlier
Michael Hammarsten, Nicola Schlatter, and Nickolay Ivchenko | A statistical study of plasma line enhancements during IPY 2007-2008
9
Plasma Line Enhancements
Seasonal distribution of peak frequency during IPY
I Seasonal dependency
I
I
I
Summer - Broader spectral distribution
Fall - Transition - Bandwidth edges
Winter - Centre frequency changed
Michael Hammarsten, Nicola Schlatter, and Nickolay Ivchenko | A statistical study of plasma line enhancements during IPY 2007-2008
10
Plasma Line Enhancements
Number of enhancements as a function of range vs frequency
I Spectral distribution
I
I
I
Main distributions at ±3.4MHz to ±4MHz
Secondary distributions at bandwidth edges
UPS has a third distribution at 3.7MHz to 4.4MHz at 230km height.
Michael Hammarsten, Nicola Schlatter, and Nickolay Ivchenko | A statistical study of plasma line enhancements during IPY 2007-2008
11
Conclusion of study
Thank you for listening! – [email protected]
I Range distribution
I
I
DWS has only one peak at ∼190km
UPS has two peaks at ∼190km and ∼ 230km
I
The main distribution of Single enhancements occurs at the ∼170km
I Spectral distribution
I
I
I
Main distributions at ±3.4MHz to ±4MHz
Secondary distributions at bandwidth edges
UPS has a third distribution at 3.7MHz to 4.4MHz at 230km height.
I Time distribution
I
The main distribution of enhancements occurs during noon MLT.
I
Single enhancements occurs a few hours earlier
I Artefacts of system, interference, and filter aliasing
I
At ∼ 4.1MHz, ∼ 4.4MHz, and ∼ -4.4MHz
Michael Hammarsten, Nicola Schlatter, and Nickolay Ivchenko | A statistical study of plasma line enhancements during IPY 2007-2008
12