17 July 2002

1
17 July 2002
2
July 17 2002 flare TRACE observations
3
130 separate tracks
Fletcher, Pollock Potts 2004
4
Flare footpoints on simultaneous magnetogram
Can calculate F vBLOS
.
5
.
UV footpoint source intensity variations
correlated with F
Typical examples
I1600
I1600
Peaks in v BLOS for individual footpoints show
significant correlation in time with peaks in the
UV brightness, during impulsive phase
Observations Monte-Carlo
simulations Peaks within ? 2s 25 ? 5 8 ?
2 Peaks within ? 8s 45 ? 5 25
? 5
6
Footpoints and HXRs
7
.
HXR sources tend to be where F is highest
RHESSI 50-100 keV
RHESSI 25-50 keV
070131 070201
070201 070230
07023- 070300
8
Search for Correlated UV footpoint pairs
My quick-and-dirty previous work identified
highly-correlated pairs from linear and rank
correlation coefficients. Problem this tended to
find correlations in overall trends in data,
rather than correlated bursty structure on 10s
of seconds timescales Solution separate out
slowly and rapidly varying components (c.f.
Aschwanden work with Fourier filtering) and
correlate rapidly-varying parts
Technique used wavelet à trous Iteratively
decomposes signal into aperiodic variations on
different timescales seems to be best wavelet
for the job
9
Pairs of correlated footpoints
pairs of footpoints for which UV time profiles
highly correlated (lines join pairs with linear
correlation coefficient gt 0.8)
10
image
32 px
16 px
8 px
4 px
2 px
11
Example of à trous decomposition 2 sources
(black green)
Shortest timescale component removed signal
cleaned of high frequency noise
2 longest timescale components removed
detrended data
Only the component with characteristic timescale
10s
12
Et cetera
13
Examples of reasonably well-correlated pairs
Neighbours
Neighbours
Well-separated pairs
14
(No Transcript)
15
Systematic time delays or just coincidence??
If footpoint intensity variations due to particle
heating, how does a 5s time delay arise?
16
(No Transcript)
17
Sample UV source pairs at HXR footpoint regions
Green eastern source, blackwestern source
18
Conclusions
HXR sources occur where flux reconnection rate
highest Just starting looking again at
correlations. - Highly correlated UV footpoint
pairs hard to find - Evidence for time lags of
4-5 seconds between footpoint brightenings? -
Have to look particularly for evidence of
correlation at HXR source locations