Sub-keV Ring Current Ions: Source, Transport, and O+/H+ difference M. Yamauchi, R. Lundin, H. Nilsson, S. Arvelius (IRF-Kiruna),

Y. Ebihara (NIPR), and Cluster-CIS team

westward drift

eastward drift

eastward drift

"Wedge-like dispersion" Sub-keV trapped ions are seen almost all satellites at around L=4-6. They are wedge-like energy-latitude dispersed as shown in both Viking data (mid-altitude) and Cluster data (equatorial plane). They are trapped ions drifting eastward, i.e., the ExB drift (including corotation) is stronger than the magnetic drift (B and curvature) at this energy range.

We show (1) Viking statistics (2) Cluster event studies.

H+

Previous Workssub-keV ion precipitation @ subauroral region):

•Aureol 1 (400~2500km): Sauvaud et al., 1980

00-06 MLT: increases after substorms.

* DMSP F6/F7 (800 km):Newell & Meng, 1986 0830 MLT: correlated with Kp with some hours delay, and event may last a day.

* Viking (2~3 RE): Yamauchi et al., 1996a,b"Wedge-like dispersed structures"modulation by pc-5 pulsation.

* Simulation: Ebihara et al., 2001drift model (ExB, |B|, and curvature)many hours after nightside injectiondispersion patterns + MLT dependence.

* Freja/Viking/Cluster: Yamauchi et al., 2005morning peakO+ at low-altitude / H+ at high-altitude

* others: Shelley et al., 1972; Chappel et al., 1982

Reversed dispersion Both

dispersions

Ebihara et al., 2001

extend >1keVlimited < 1keV

Cluster(2001-2003)

(1) Vikingbackward superposed

epoch analyses

Probabilities of observing the wedge-like structure after the start of AE activity. Probability is calculated from numbers of traversals with/without the structure for each 3-hr bin (3-hr running sum) for each 3-hr MLT bin.

The peak probability moves eastward, while the peak value of the probability decrease as the peak moves eastward.

6 MLT

9 MLT

12 MLT

15 MLT

18 MLT

Time-lag (hours)

Evacuation is seen (the probability is even lower than asymptotic one)

Viking Summary MLT Minimum

Quiet case

After end of 300 nT activity

Asymptotic

Quiet case

After end of 300 nT activity

Maximum

Clear case

After start of 400 nT activity

5~7 1~3h (0%) 8~9h (30%) 0~3h (85%)

8~10 2~3h (5%) 9~10h (50%) 2~4h (75%)

11~13 3~5h (10%) 10~11h (70%) 4~6h (70%)

14~16 4~6h (35%) 12~13h (80%) 6~7h (50%)

17~19 6~8h (50%) 14~16h (100%) 10h (25%)

The wedge-like structure drifts eastward, and is a fossil of substorm activity (model is right!). Decay time is several hours (charge exchange model is right!).

However, it appears much earlier than prediction, suggesting that a substantial amount of "wedge" might be formed in the morning sector during substorms. We need to identify the source location from event study.

(2) ClusterWe have several possibilities

(A) Strong electric field push ions quickly.(B) Scattering of <10 keV ions(C) Energetic ions precipitate and sputter

ionospheric ions into the space.(D) Unknown local energization process.

Need to find events when the wedge is formed during a substorm.

We found one case. Wedge is seen only at outbound.

Case study!

Ion source dispersion scenario

night night No !

night morning (A)

morning morning (B) (C) (D) H+

O+

+1° 0° -1° -2°

No

Yes

Yes

?

No

No

2001-10-2123:40-24:00 UT

Yes

No

No

S/C-4

S/C-1

S/C-3

Relative S/C position: all at 9.0±0.1 MLT

H+

O+

H+

O+

H+

O+

Observation summaryS/C-1 (23:45 UT), S/C-4 (23:50 UT), and S/C-3 (23:40 UT) passed through the same magnetic flux tube at 9 MLT (L≈4).

Wedge-like dispersion at 23:50 UT.No low-energy signature at 23:40 UT.

Butterfly-trapped distribution Bounce inside the geomagnetic bottle. Difference between 23:40 UT and 23:50 UT in the same flux tube means an temporal variation although observation is made in the opposite hemisphere.

VE: eastward ExB drift speed = energy independent, MLT dependentVB: westward magnetic (|B|+curvature) drift velocity = energy dependent VE >> VB at low energy (<100 eV) and VE ~ VB at high energy (value depends on E-field strength). From dispersion curve, the last-coming ions are 10-20 keV. Therefore, VE ~ VB at 20 keV in the present case.

V1*t = V2*(t+∆t) or

(t+∆t)/∆t = V1/(V1-V2)

~ VE/VB (note : VB@10 keV)

= (E/B)*(q*R*B/3*W*g)

~ E [mV/m]/g or

t = ∆t*E [mV/m]/g - ∆tfor observation near equatorial plane, where E and B are the field strengths, q is the charge, R = 4 RE is the geocentric distance, W = 10 keV is the ion energy, and g ~ 1, 0.9 &0.7 for 90°, 40° & 0° pitch angles

V1 = VE-VB ~ VE @ 0.1 keV V2 = VE-VB << VE @ 10 keV

V1 ~ VE

V2 ~ VE-VB

t

t+∆t

0.1 keV

10 keV

time-of-flight principal

Dispersion analysisPitch angle of the "wedge" is about 40~90° (g=0.9~1.0) t ≤ (1.1*E[mV/s] - 1) * ∆t

Electric field is 1~3 mV/m for half an hour t = 0.1~2.3*∆t & VE = 3~10 km/s

• 0.1 keV @ 23:50 UT, S/C-1 Nothing @ 23:40 UT, S/C-3 : temporal chance

(b) 10 keV @ 23:53 UT, S/C-1 Nothing @ 23:40 UT, S/C-3 : temporal change

(c) 0.1 keV @ 23:50 UT, S/C-1 10 keV @ 23:53 UT, S/C-1 : temporal or spatial

Combination of (a)+(b) : it is temporal change ∆t < 13 min t < 30 min before 23:40 UT drift distance = VE * t < 20000 km dispersion started at 7~9 MLT.Combination of (b)+(c) : if temporal ∆t~3 min t = 0.5~8 min before 23:50 UT drift distance = VE * t = 100~5000 km dispersion started at 8~9 MLT.

1~3 mV/m

On the other hand, we observed O+ "wedge" at 0.05-0.3 keV (20 km/s ~ 50 km/s). The 0.05 keV O+ takes 20~30 min to travel from the ionosphere to the Cluster location along B in best case. From this:

(1) Source timing is about 20~30 min before, i.e., at 23:20~23:30 UT.

(2) The combination (b)+(c) cannot be true, i.e., the observed dispersion is mostly the spatial structure.

(3) O+ pitch angle is uni-direction, i.e., should not have been mirror-bounced, endorsing point (1).

O+ source ≠ H+ sourceO+ motion ≠ H+ motionThe 2001-10-21 event showed a clear O+/H+ difference inside the wedge, with H+ bounce-averaged feature (with butterfly pitch-angle distribution), whereas O+ is not bounce-averaged.

Statistically the wedge-like structure is O+ rich at low-altitudes (Freja) whereas it is H+ rich as high-altitudes (Cluster).

These fact suggests that O+ source could be different from H+ source. We found couple of good Cluster examples that endorse this idea

correlated & anti-correlated

H+

O+

H+/O+ differences

Correlation part means that H+ and O+ has the same bounce-average drift motion. Then, how can we understand the anti-correlation part just 15 minutes later?

Summary and conclusions(1) The dispersion might start in the morning for a substantial

numbers of the wedge-like structure. This is suggested by the local time distribution, superposed epoch analyses, and a case study on the 2001-10-21 event (source <30 min, <3 RE distance).

(2) Pitch-angle distribution, particularly for O+, suggest

ionospheric source (consistent with morning source).

(3) In addition to the altitude dependence of the O+/H+ ratio, O+ are sometimes behaving in a different way from H+.* non-bounce-average feature (2001-10-21 event).* correlation and anti-correlation in a single traversal. ???

Future task : understand the source of the wedge-like structure for both O+ and H+. This final target is still far away.

END

What is "wedge-like dispersion" / What is the problem?Sub-keV trapped ions seen almost all satellites at around L=4-6. They are wedge-like energy-latitude dispersed as shown in both Viking data (mid-altitude) and Cluster data (equatorial plane). See Poster XY0868 by Yamauchi and Lundin in this session.

Data analyses and simulation confirmed that they are drifting trapped ions (I.e., Cluster should see the same phenomena as Viking). Past analyses raised two main questions.

(1) Past identification of "wedge" observed by Cluster was in-appropriate because the resultant distribution does not agree with the simulation or other satellites. We need to refine the criterion.

(2) Past statistics suggests that the source can be in the morning sector during substorms. We need to identify the source location from event study.

westward drift

eastward drift

New criterion and statistics

Past criterion: completely isolated from > 5 keV ring current. Only (a) is identified as "wedge" but not (b) or (c)New criterion: isolated from > 5 keV component at the wedge location, as long as wedge is extended from sub-keV. All of (a), (b), and (c) are identified as "wedge"

New statistics shows morning peak, which is consistent with the other statistics and simulation.

(a)(b)+(c)

Present Work

(1)case study

(2) statistics

Simulation indicates:

* Drift slowly eastward

* Originated from past substorm-related injections into the ring current region 5~20 hours before.

However,

No solid data analyses has been done to confirm the dynamic part of the model.

Are they drifting? If so, velocity?

Are they nightside origin?

Are they related to substorms?

If so, time lag?

(1) Case study : It requires isolated substorm activity + consecutive traversals, but even best case can be interpreted in many way.

Result : 1

Probabilities of observing the wedge-like structure after the end of AE activity. (cf. (2) in explanation)

Quiet probability corresponds to the last injection Lowest quiet probability start increase (=last wedge passing through) start to increase at later time-lag at larger MLT, it moves eastward, while the value itself increase eastward.

Time-lag (hours)

6 MLT

9 MLT

12 MLT

15 MLT

18 MLT

Conclusions : wedge-like structures1. The structure is related to the past AE activity but not directly to Dst

2. After hourly AE>400 nT, the majority of the structure reaches the noon, and nearly half of them reaches the early afternoon sector.

3. The structures in the evening sector most likely have traveled by eastward drift rather than directly from the nightside by westward drift.

4. The response at 6 MLT is nearly immediate after high AE activities. Source of wedge shifts or extends to the early morning, e.g., 4-5 MLT.

5. The drift speed for hourly AE>400 nT is somewhat faster than model prediction even taking into account of the morning-shift of source.

6. The decay time of several hours at all MLT is consist with the charge exchange life time.

7. Sub-keV ions are sometimes evacuated right after the onset of substorm or storm.

More statistics (∑3h)

Peak probabilities vs AE threshold

Optimum time-lag vs AE threshold

For different AE threshold values

Best threshold value values are 400 nT for start of activity and 300 nT for end of activity

Result of superposed epoch analyses (∑3h)From end of activity (cf (2) in explanation) From start of activity (cf (1) in explanation)

Time-lag (hours) Time-lag (hours)

6 MLT

9 MLT

12 MLT

15 MLT

18 MLT

Previous Works

Reversed(Type 2)

Ordinary (Type 1)

Both (island)

sub-keV ion precipitation @ subauroral region):

* Aureol 1 (400~2500 km): Sauvaud et al., 1981 00-06 MLT: increases some hours after substorms.

* DMSP F6/F7 (800 km):Newell and Meng, 1986 0830 MLT: correlated with Kp with some hours delay, and event may last a day.

* Viking (2~3 RE): Yamauchi et al., 1996a,b"Wedge-like dispersed structures"modulation by pc-5 pulsation

* Simulation: Ebihara et al., 2001drift model (ExB, grad|B|, and curvature)many hours after nightside injectiondispersion patterns + MLT dependence

* Freja/Viking/Cluster: Yamauchi et al., 200505-19 MLT: morning peakaltitude comparison: O+/H+ ratio change

* Others: Shelley et al., 1972; Chappel et al., 1982Reversed (Type 2) Both

(island)

Viking Observation

Ebihara et al., 2001

Present Work

(1) simple statistics(2) case study(3) advanced statistics

Simulation indicates:

* Drift slowly eastward

* Originated from past substorm-related injections into the ring current region 5~20 hours before.

However, no solid data analyses has been done to confirm the dynamic part of the model.

Are they drifting? If so, velocity? from where?

Related to substorms? If so, time lag?

Simple AE correlation? misleadingAfternoon sector show negative correlation. Thus, one may not take direct correlation.

Simple Dst correlation? misleadingBefore all, only 58 out of 700 are during Dst < -30 nT. Magnetic storm activity is not the direct cause.

159 209 280 58 # traversals

80 198 192 159 77 # traversals

(2) Case study

It requires isolated substorm activity + consecutive traversals = rare.

Even the best case (860912) show superficial anti-correlation

(1) Backward Superposed Epoch Analyses

* Probabilities with/without “wedge” signature at various MLT in dayside is obtained for different time-lags from latest AE increase.

(a) Ideal AE profile gives three characteristic times (1)(2)(3).

(b) & (c) We must fight against realityHope statistics helps.

Total only 700 traversals, sorted by * 3-hour MLT bins * 3-hour windows (running

summation) for the time-lag * add all types of dispersions

Case studyWe have several possibilities

(A) Strong electric field push ions quickly.(B) Scattering of <10 keV ions(C) Energetic ions precipitate and sputter

ionospheric ions into the space.(D) Unknown local energization process.

Need to find events when the wedge is formed during substorms. We found one case. Wedge is seen only at outbound.

Ion source dispersion scenario

night night No !

night morning (A)

morning morning (B) (C) (D)

General context (deduced from ENA image)ENA image indicates (1) strong E (2) No energetic H+ in the late morning sector(3) qualitative difference between O+ and H+

Strong E scenario (A) in the Table: Ions < 10 keV could have convected to the morning sector quickly without forming the dispersion.

No energetic H+ difficult for scenario (B) in the Table unless electron is important

H+ - O+ difference At least O+ wedge can be formed local

post-midnight preference= strong E (could be ~10 mV/s)