Observation of radiation belt losses into the ionosphere

1
Observation of radiation belt losses into the
ionosphere
Craig J. Rodger Department of Physics University
of Otago Dunedin NEW ZEALAND
REPW 2007 245pm Wednesday, 8 August 2007
2
Energetic electrons, the Radiation Belts, and
Geomagnetic Storms
Losses overall response of the RB to geomagnetic
storms are a "delicate and complicated balance
between the effects of particle acceleration and
loss". Reeves et al., GRL, 2003. Essentially
all geomagnetic storms substantially alter the
electron radiation belt populations, and losses
play a major role. A significant fraction of the
particles are lost into the atmosphere, although
storm-time non-adiabatic magnetic field changes
also led to losses through magnetopause shadowing
(particularly during storm-times). In today's
talk I want to discuss some studies we have done
on particle precipitation from the RB during
storms (FAST and SLOW events) after storms
3
Energetic Precipitation from the RB affects the
lower ionosphere
For electrons gt100keV, the bulk of the
precipitated energy is deposited into the middle
and upper atmosphere (30-100km), and can be
detected through changes in subionospheric VLF
propagation.
REP
Subionospheric VLF signals reflect from the
ionospheric D-region. ? received radio waves
are largely determined by propagation between the
boundaries ? Very-long range remote sensing is
possible (1000's of km )!
4
For frequencies gt10 kHz man-made transmissions
from communication and navigation transmitters
can be observed in almost every part of the
world, as seen in the spectra below. Very-long
range remote sensing is possible as the
transmitter signals can be received 1000's of km
from their sources!
South Africa
New Zealand
Israel
Antarctica
5
Our AARDDVARK
An aarmory of AARDDVARKs. This map shows our
existing network of sub-ionospheric energetic
precipitation monitors. MORE INFORMATION
www.physics.otago.ac.nz\space\AARDDVARK_homepage.h
tm
6
Part I FASTREP

• Part I is based on the following publication
• Rodger, C. J., M. A. Clilverd, D. Nunn, P. T.
  Verronen, J. Bortnik, and E. Turunen, Storm-time
  short-lived bursts of relativistic electron
  precipitation detected by subionospheric radio
  wave propagation, J. Geophys. Res., 112, A07301,
  doi10.1029/2007JA012347, 2007.

7
Microbursts
These pulses here!

• RELATIVISTIC MICROBUSRTS
• gt1 MeV microbursts lasting ltlt1 s (L4-6)
• typically observed at the outer edge of the
  radiation belt
• observed at all local times, but occur
  predominantly in the morning sector
• Thought to be associated with VLF chorus waves
• Lorentzen et al., 2001

SAMPEX satellite observed fluxes show that
microburst precipitation losses can essentially
"flush out" the entire relativistic electron
population during the main phase of the storm.
However, the global loss estimates are based upon
the assumption that the microburst flux is
isotropic and constant over selected L and MLT
ranges. Ground based observations would
complement the space-based point measurements!
FAST Process!
8
REP Microbursts
The distinction between relativistic and
non-relativistic microbursts is an important one.
Non-relativistic microbursts have been known of
for some time Non-relativistic (gt150 keV)
microbursts are also of short duration (0.2-0.3
s), and occur on the dayside. Rocket and balloon
measurements indicate that the precipitating
electrons are primarily in the tens to hundreds
of keV range, with negligible fluxes above 1
MeV. SAMPEX observations have also shown that
REP microbursts are often accompanied by
non-relativistic (gt150 keV) microbursts, but the
bursts in the two energy channels do not exhibit
a one-to-one correspondence.
So REP microbursts and non-REP microbursts can
occur in the same time periods and rough
locations, but are very unlikely to be created
simultaneously through the same process.
9
REP Microbursts Chorus
REP microbursts are correlated with satellite
observed VLF chorus wave activity short
duration of microbursts similar chorus elements
similarity in LT distributions have lead
to the widely held assumption that REP
microbursts are produced by wave-particle
interactions with chorus waves. However, this
has yet to be confirmed, and a one-to-one
correlation of REP microbursts and chorus
elements has yet to be demonstrated. Here we
examine ground-based ionospheric data during REP
activity to better understand the energy spectra
and hence the source mechanism
Chorus observed in Dunedin on 7 Feb 2005
10
We start by focussing on the late January 2005
period. This was a fairly disturbed time, with
numerous solar flares (X-rays striking the upper
atmosphere). A fairly large solar flare on 20
January 2005 produced an unusually hard SPE
(protons striking the upper atmosphere). An
associated CME triggered a geomagnetic storm,
leading to a relativistic electron drop-out on 21
Jan 2005.
11
Observation setup
During 21 January 2005 (and also 4 April 2005)
our subionospheric VLF receivers were operating
at several different locations, observing
transmitters that span a wide range of longitude
sectors and L-shells.
12
REP observed at SGO on 21 Jan 2005
Summary Plot. Amplitude subionospheric VLF
observations at 17-18UT from SGO.
FAST losses on top of the SLOW losses.
NOTE There appears to be two different time
scales seen in this data changes on 5min
timescales and also short lived spikes, which is
what I want to focus on first.
13
REP observed at SGO on 21 Jan 2005
Different Summary Plot. Period which emphasizes
the FAST subionospheric amplitude perturbations
observed on multiple transmitters at SGO.
14
Examples of FASTREP in SGO data
The residual amplitudes that remain after
removing the smoothed background with a 5s
filter, drawing out the rapid changes. Shows
many large short-lived spike events of 1-5 dB,
both increases and decreases. Similar short lived
large negative and positive perturbations are
also observed in the transmitter phase. Almost
certainly the first ground-based observations of
relativistic electron microbursts.
15
Examples of FASTREP in SGO data
Over the 4 hour period 17-21 UT, 221 FAST
perturbations were observed on transmissions from
NAA and 109 on NDK, i.e. a rough rate of 1 per
minute on each GCP. The majority of the FAST
perturbations are not simultaneous with
perturbations on other paths even though they
occur during the same periods. However, a very
small fraction, on the order of 1-2 appear to
be simultaneous on multiple paths. This behaviour
is consistent with ionospheric changes with small
spatial size, occurring near to the Sodankylä
receiver.
A useful analogy is a rainstorm, with many small
raindrops, each representing a burst of
precipitating electrons, produced by the same
physical process spanning a much larger spatial
region.
16
Timescales of FASTREP
60s worth of Sodankylä-received subionospheric
amplitude from VLF transmitters on 21 January
2005, plotted at the maximum 0.1s time
resolution. Several well defined FAST features
are evident, each lasting roughly 1 s, as well
as a feature on NDK lasting 6 s which appears to
be made up of multiple discrete pulses.
The well defined FAST features are typical for
the short-lived pulses seen on 21 January 2005,
made up of a 0.2 s rise followed by a 0.5 s decay
to background amplitude levels.
17
Timescales of FASTREP
Another example, this time on NDK. Generally a
0.2s risetime, following by a 0.5s decay, but
with some events appearing to be made up of
multiple events. The rapid decay time is highly
suggestive of a short-lived highly energetic
precipitation burst producing significant
additional ionization at 60km altitude (peak
ionization altitude for a 1.5MeV electron). We
therefore conclude that the FAST subionospheric
VLF perturbations are most consistent with the
ionization signature caused by REP microbursts.
Time resolution0.1 s points
18
REP precipitation on 3 April 2005
Fundamentally similar to the perturbations
observed on 21 Jan 2005, the decay time of these
events are a factor of 3 longer (1.4 s c.f.
0.5 s), consistent with masking of higher
altitude ionization changes from subionospheric
detection due to the proton precipitation
continuing on from the 20 Jan 2005 SPE. This
provides strong evidence that that the FASTREP is
produced by short-lived bursts of REP, with the
"real" perturbation-decay time being 1.5 s when
unmasked by ionization produced from other
precipitation.
19
Modelling of Chorus produced VLF perturbations
It is widely assumed in the community that REP
microbursts are due to chorus waves. In order to
test the REP microburst observations against the
expected signature for chorus produced
precipitation, we make use of calculated
precipitation fluxes produced by a single chorus
element at L5 at 100km altitude. The chorus rays
are launched from the geomagnetic equator . As
these rays propagate, they resonate with
counter-streaming energetic electrons through the
m1 resonance mode.
20
Modelling of Chorus produced VLF perturbations
N. Hemisphere
S. Hemisphere
However, because these chorus waves are oblique,
higher order resonances are also possible,
including both counter- and co-streaming
orientations. The calculation is undertaken using
a test-particle simulation assuming AE-8 trapped
electron flux levels, and an isotropic
distribution that is sharply cutoff at the loss
cone. N. hemisphere 2 energy groupings,
4-30keV 150 keV-4MeV S. Hemisphere single
high energy grouping, spanning 50keV-3MeV
21
Effect of precipitation on atmosphere
Assume northern hemisphere chorus precipitation
pattern (but it doesnt matter for D-region)
To describe the atmospheric impact of the
chorus-produced precipitation, we make use of the
Sodankylä Ion Chemistry model. The SIC model runs
were undertaken for polar night conditions
appropriate for 17 UT on 21 Jan 2005 at a
location of (69ºN, 22.5ºE), i.e., close to our
receiver in Finland. This figure shows the time
varying electron number density profiles
determined from the SIC model.
22
Modelled VLF perturbation from chorus REP
213654.80 UT, 5 April 2005
100s keV precipitating electrons make
long-lasting ionospheric modification gt70 km and
hence negative amplitude perturbation.
Apply the SIC-determined electron densities to
the VLF propagation code Modefinder and see what
the perturbation looks like. It does not do a
very good job of reproducing the FAST VLF
signature too much low energy electrons present
which impact at high altitudes.
23
Modelled VLF perturbation from mono-energetic REP
beams
213654.80 UT, 5 April 2005
Try 0.1s monoenergetic beams with flux of
100 el.cm-2s-1sr-1 (similar to SAMPEX) and
contrast with the same microburst decay time. We
need a highly energetic beam with no significant
electron flux less than 1.5MeV to reproduce the
decay time of our VLF perturbations. This is not
consistent with chorus production, but is
consistent with the satellite observations of
relativistic REP microbursts.
24
Summary - Part I

• On 21 January 2005 a relativistic electron
  drop-out at GEO was associated with several
  hundred short-lived VLF perturbations observed by
  the AARDDVARK receiver at SGO over 4 hours.
• The vast majority of these perturbations were not
  simultaneous with perturbations on other nearby
  paths ? precipitation "rainstorm" producing
  ionospheric changes of small spatial sizes around
  the Sodankylä receiver.
• 1.5s VLF perturbation decay time is consistent
  with short-lived bursts of REP, observed by
  satellites as relativistic microbursts. To the
  best of the authors knowledge, this is the first
  ground based observation of microburst REP.
• Contrasts between predicted VLF perturbations for
  chorus driven precipitation, and those due to
  more highly energetic beams indicate that the
  recovery time is most consistent with
  precipitation containing only highly relativistic
  populations, principally from energies greater
  than about 2 MeV.
• The chorus-driven VLF perturbations were very
  unlike the experimentally observed VLF
  perturbations, and unlike the reported energy
  spectra of satellite observed relativistic
  microbursts.

25
Part II REPJGR

• Part II is based on the following publication
• Clilverd, M. A., C. J. Rodger, R. M. Millan, J.
  G. Sample, M. Kokorowski, M. P. McCarthy, Th.
  Ulich, T. Raita, A. J. Kavanagh, and E.
  Spanswick, Energetic particle precipitation into
  the middle atmosphere triggered by a coronal mass
  ejection, J. Geophys. Res., (in press), 2007.

26
Minutes-to-hours MeV precipitation

• MeV Precipitation
• gt0.5 MeV precipitation lasting minutes-hours
• Located between L3.8-6.7
• observed only in the late afternoon/dusk sector
  (1430-0000 MLT)
• May be produced by EMIC waves
• Millan et al., 2002

MAXIS balloon observed fluxes show that this
precipitation mechanism is a primary loss process
for outer zone relativistic electrons.
Plot taken from Millan, R. M., R. P. Lin, D. M.
Smith, K. R. Lorentzen, and M. P. McCarthy, X-ray
observations of MeV electron precipitation with a
balloon-borne germanium spectrometer, Geophys.
Res. Lett., 29(24), 2194, doi10.1029/2002GL015922
, 2002.
SLOW Process!
27
Initial conditions - CME impact
ACE
ACE data is delayed to reflect expected CME
arrival time.
17.2 1710 UT
ACE
GOES
28
In Jan 2005 MINIS balloons were launched through
the AARDDVARK network
29
Scandinavia
Dusk Sector 20 MLT
L 5-6 Riometer VLF (Iceland SGO)
L 2.5-4 Riometer VLF (Germany SGO)
Pulsation mags
30
Dusk Sector
Initial pulse
L 5-6 Riometer VLF (Iceland SGO)
Initial pulse
L 2.5-4 Riometer VLF (Germany SGO)
Extended precipitation
31
Prolonged precipitation for 1 hour
4 November 2003
Thomsons Great Flare (X45)
Peak flux 17.3-17.35 UT
32
Nurmijarvi Pulsation Magnetometer
EMIC band
L3.4
33
IMAGE EUV observation of plasmasphere
Compressed plasmasphere 5 hours after CME, with
Kp8
SUN
The position of the plasmapause is consistent
with EMIC-induced precipitation of electrons at
L2-3
Lpp 2-3
34
17-18 UT, 21 Jan 2005Global Observation sites
35
Dayside10-15 MLTNorthern H
gt30 keV
Using Walt 1994 to calculate an azimuthal drift
period at L4.0 14 mins 1.5 MeV 17 mins 0.8
MeV
14 mins
?
?
gt100 keV
?
?
17 mins
36
Dayside10-15 MLTSouthern H
Halley-
gt100 keV
Some bursty precipitation events, although the
timing is not clearly linked to the northern
hemisphere events, apart from Flt3S L4.1 Some
interesting comparisons between MINIS balloon
X-ray counts and riometer data
gt30 keV
gt20 keV -10 MeV
37
Map of precipitation
Initial arrival of the CME causes precipitation
over L from 3-8, and probably over all MLT.
(initial burst also seen by Macquarie Island
riometer. Nothing in Dunedin-Australian VLF data,
hence lower cutoff). The nightside
precipitation is more localised, and the dayside
rather bursty.
38
Summary Part II

• Detectable levels of energetic precipitation into
  the atmosphere following the 21 January 2005
  CME.
• Initial precipitation lasting 5 minutes,
  observed over a large proportion of the Earth.
• Dayside precipitation bursty and suggests
  drifting electrons 1 MeV interacting with a loss
  region at L4.
• Duskside precipitation appears to be associated
  with the location of the plasmapause and produced
  by EMIC waves, which were simultaneously recorded
  in southern Finland.
• We are now working to increase our database of
  EMIC waves with similar events, and work on the
  precipitation fluxes and energy spectrum.

39
Part III CRRESREP

• Part III is based on the following publication
• Rodger, C. J., M. A. Clilverd, N. R. Thomson, R.
  J. Gamble, A. Seppälä, E. Turunen, N. P.
  Meredith, M. Parrot, J. A. Sauvaud, and J.-J.
  Berthelier, Radiation belt electron precipitation
  into the atmosphere recovery from a geomagnetic
  storm, J. Geophys. Res., (in press), 2007.

40
Proton fluxes
DEMETER measured drift-loss cone electron fluxes
at L3.2
Dst
Geomagnetic storm of 11 Sept 2005 led to an
increase in the energetic electron population in
the inner edge of the outer radiation belt 1000
times ambient 100 above pre-storm levels decays
over 14 days to pre-storm levels (and 5 times
above ambient), after which there is a DEMETER
data-gap
Kp
41
DEMETER 128 energy channels (70keV-2.34
MeV) CRRES 17 energy channels (153keV -1.582
MeV)
NOTE CRRES shifted to fit DEMETER fluxes
Hardening in electron spectra seen by the
low-Earth orbiting DEMETER in drift-loss cone is
consistent with that seen by CRRES near the
geomagnetic equator in June 1991. We represent
the spectra by the fitted power law shown.
42
Subionospheric Observations
NYA, Ny Ålesund
SGO, Sodankylä
CAM, Cambridge
NAA, 24kHz 600kW
From 1 September 2005 we undertook subionospheric
observations of NAA from our 3 European AARDDVARK
receiver sites.
43
Dominated by Solar Proton Event line SPE
Mid L12
Solar Proton Event and something else?
Mid L7.5
Undisturbed level
Must be precipitation from RB (too low L for
protons)
Mid L3.2
Change in NAA at CAM due to energetic electron
precipitation 141 dB decrease at
Midnight 2.40.3 dB increase at Midday
44
Model the ionospheric changes
gt150keV electron flux 200 el.
cm-2 s-1

• Sweep through range of precip. fluxes
• Use a simplified ionospheric chemistry scheme to
  determine the D-region electron density after
  precipitation (checked by the Sodankylä Ion
  Chemistry Model)
• Use this as an input for the subionospheric
  propagation code LWPC

HENCE predict expected change in amplitude for a
given precipitation along our NAA-CAM path.
45
Subionospheric measurements of precip.
Precipitation fluxes required to reproduce the
changes in subionospheric propagation observed
(NAA -gt CAM).
Peak Fluxes 3500300 el. cm-2s-1 at midday
18515 el. cm-2s-1 at midnight. NOTE after
storm, day fluxes are 20 times higher than night
fluxes (for 6 days).
This is consistent with plasmaspheric hiss
observations. During geomagnetically disturbed
conditions (AEgt150 nT) CRRES found 0.2-0.5kHz
hiss was 10 times stronger on dayside than
nightside. With DEMETER wave-data we can test
this!
46
DEMETER observations of plasmaspheric hiss

• At L3.2 resonance 0.5 kHz waves with 160keV
  electrons40 Hz waves with 1 MeV electrons
• Use DEMETER to look at this wave range and
  L-range above our transmitter-receiver Great
  Circle Path.

• Both wave and particles show a factor 200
  increase during 9-11 Sept.
• Daytime wave powers 10 times nighttime in
  post-storm period, much like seen in the
  precipitating particle measurements.

Dawnside equatorial chorus does not reproduce the
day-night differences seen in our precipitation
fluxes, as it has peak intensities on the morning
and evening sides. Off-equatorial chorus is 100
times stronger on day than night.
47
Summary - Part III
After 11 September 2005 geomagnetic storm the
gt150 keV electron fluxes in the drift loss cone
at L3.2 increased by a factor of 1000 above
ambient conditions. The fluxes decayed to within
a factor of 5 of the ambient levels over the
following 14 days. Measured energy spectrum is
consistent with previous observations of
storm-time flux increases. Subionospheric VLF
measurements allow us to quantify the
precipitation losses, effectively "resolving" the
bounce loss cone. The peak precipitated fluxes
of gt150 keV electrons into the atmosphere were
3500300 el. cm-2s-1 at midday and
18515 el. cm-2s-1 at midnight. Plasmaspheric
hiss (PH) observations show reasonable agreement
with precipitating particle behaviour, suggesting
PH with freqs lt500 Hz primary driver of energetic
electron losses outside storm periods.
48
Conclusions
The precipitation of energetic radiation belt
particles can be detected from the ionosphere,
providing a complementary way of looking at these
events. We have advanced our tools a great deal,
and are now able to provide indications of
precipitating flux in many situations. The
AARDDVARK network is well positioned to work
alongside upcoming missions, like DSX, RBSP and
ORBITALS.
49
1st international HEPPA Workshop
2008"High-Energy Particle Precipitation in the
Atmosphere"

• 28-31 May 2008
• Finnish Meteorological Institute, Helsinki,
  Finland
• Focus Observations and modelling of atmospheric
  and ionospheric changes caused by energetic
  particle precipitation, e.g. solar proton events,
  relativistic electron precipitation, and auroral
  electron precipitation. Topics ranging from
  short-term ionospheric changes to long-term
  atmospheric changes are welcome, including
  defining spectra of precipitating particles and
  the effects on atmospheric dynamics.
• http//heppa2008.fmi.fi
• Registration begins in October 2007

50
Craig visiting the Artillery Museum. St.
Petersburg, October 2006
Thankyou! Are there any questions?