March 17 2015 Wen LI - SCIENCE GATEWAY: OVERVIEW

Radiation Belt Electron Dynamics During the March 17 2015

Geomagnetic Storm: Observations and Simulations

Wen Li, Qianli Ma, Richard Thorne, Jinxing Li, Jacob Bortnik, Van Allen Probes ECT and EMFISIS team, POES, THEMIS

team, and other potential coauthors

Acceleration and loss mechanisms of outer radiation belt electrons

• Acceleration mechanisms– Electron injection– Inward radial diffusion (ULF)– Local acceleration by wave-

particle interactions• Chorus wave• Magnetosonic wave

– Time domain structure– …

• Loss mechanisms– Loss to magnetopause– Outward radial diffusion (ULF)– Precipitation loss by pitch

angle scattering• EMIC wave• Chorus wave• Plasmaspheric hiss

– …

Key science questionsØ When and where is each mechanism dominant?Ø What is their quantitative role in radiation belt electron dynamics?

2

ObjectiveQuantify the role of various important physical processes during the March 17 2015 geomagnetic storm

Ø Radial diffusion Ø Chorus waveØ Plasmaspheric hissØ Magnetosonic wave

Methodology§ Electron and wave measurements from Van Allen Probes§ A technique to infer event-specific chorus wave intensity on a

global scale using POES electron measurements [Li, W. et al., 2013; Ni et al., 2014]

§ 3D diffusion code to simulate radiation belt electron dynamics [Ma et al., 2015]

3

Overview of March 17 2015 storm

4

[Shue et al. 1998]

§ Early March 17: Ø a significant increase in P

pushed MP location to ~ 5 REØ MeV electrons dropout

§ After 0 UT on March 18: Ø P becomes much weaker, MP

location moved out to 8-10 REØ a significant fluctuation in Bz

with averaged values in south Ø MeV electrons accelerated at

lower energy first followed by multi-MeV with a time delay up to ~1.5 day

PP moved to ~2 RE in the main phase and gradually moved out in the recovery phase

◇:Plasmapause location near premidnight and postmidnightinferred from EMFISIS data

Evolution of electron PSD

L* (TS04D)

5

Clear rising peaks in electron PSD are observed for > 1000 MeV/G, indicating local heating process is operating.

Chor

us B

w(P

OES

)

(00-04 MLT)

(04-08 MLT)

(08-12 MLT)

(00-24 MLT)

4h reso.

(00-24 MLT)

1h reso.

THEMIS

RBSPChorus Bw(THEMIS)

Chorus Bw(RBSP)

SYM-H

AL

Chorus wave evolution

6

POES technique to infer chorus intensity [Li, W. et al., 2013; Ni et al., 2014]

0012

06

18

Electron pitch angle distribution

(L = 3)

SYM-HChorus Bw

(MLT ave.POES)

MS Bw(RBSP)

1.8 MeV

3.4 MeV

6.3 MeV

7

1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8

Electron PAD:Ø |MLAT| < 5ºØ ◇: Local pitch angle

line: fitted pitch angle after mapping into equator

Butterfly distribution

Electron pitch angle distribution

(L = 4.5)

8

1 2 3 4 5 6 7 8 1 3 4 6 7 82 5

SYM-HChorus Bw

(MLT ave.POES)

MS Bw(RBSP)

1.8 MeV

3.4 MeV

6.3 MeV

Ø MeV electrons experienced fairly strong acceleration up to ~ 7 MeV over 18-20 March, and show flat-top distribution

Ø Chorus wave intensity is modestly strong

3D radiation belt modeling

Pitch-angle Scattering<Daa(α, E)>

Mixed diffusion<Dαp(α, E)>

Energy Diffusion

<Dpp(α, E)>

Radial Diffusion

<Dpp(α, E)>

3D Fokker-Planck equation:Dαα, Dpp, DLL, mixed terms[e.g., Schultz and Lanzerotti, 1974; Fok et al., 2008; Albert et al., 2009; Shprits et al., 2009; Xiao et al., 2009; Glauert et al., 2014; Tu et al., 2014; Ma et al., 2015]

L:Inner boundary: f (L = 2) = 0Outer boundary: f (L = 6) = Obs.

Energy:f (µmax= 22330 MeV/G) = low valuef (µmin=80 MeV/G) Varying based on observations

Pitch angle:∂f/∂α(α=0º)=0, ∂f/∂α(α=90º)=0α < αLC: lost within ¼ τb

Boundary conditions

9

Radial diffusion coefficients (DLLE+DLL

M)

[Brautigam and Albert, 2000]

Electric field fluctuation: Magnetic field fluctuation:

Energy and pitch angle diffusion by chorus and hissCalculate Dαα, Dpp, and Dαp using Full Diffusion Code [Ni et al., 2008] § Chorus: event-specific wave evolution using POES

technique [e.g., Li, W. et al., 2013; Ni et al., 2014] § Hiss: Statistical wave intensity distribution dependent on AL

using EMFISIS wave data [Li, W. et al., 2015]10

Comparison between observed and simulated electron flux

Observation DLL 0.3 DLL

11

Chorus

Hiss

0.59 MeV

0.75 MeV

2.6 MeV

6.3 MeV

◇:Averaged plasmapause location over 00-12 MLT from Jerry Goldstein’s model

Observation DLL 0.3 DLL

0.59 MeV

0.75 MeV

2.6 MeV

6.3 MeV

Chorus Chorus + Hiss Chorus + Hiss + 0.3 DLL

0.59 MeV

0.75 MeV

2.6 MeV

6.3 MeV12

13

Evolution of electron PAD

at L = 4.5(Obs. Vs. Sim.)

Observation

Formation of the flat-top electron PAD, a typical signature of chorus-drivenelectron acceleration

Observation Simulation

Chorus 0.3 DLLChorus + 0.3 DLL

03-20/00 At the electron peak location, chorus plays an important role in accelerating electrons to MeV range, and radial diffusion helps lead to even stronger electron acceleration > 6 MeV.

[Li, J. et al., in preparation]

Butterfly distribution caused by MS waves (L = 3)

§ Use test particle simulation to calculate diffusion coefficients for MS waves

§ Use 2D diffusion code to simulate electron PSD evolution caused by MS wave and hiss

§ Simulated 100s keV electrons develop a peak at 60º in a few hours due to parallel acceleration by Landau resonance, but the butterfly formation is slower for MeV electrons.

§ Including hiss scattering helps smooth sharp butterfly profiles.

14

MS wave

102 keV

350 keV

743 keV

1.8 MeV

2.6 MeV

4.2 MeV

15

SUMMARYWe performed a 3D diffusion simulation by including the effects of radial diffusion, chorus, hiss, and MS waves during the March 17 2015 storm§ Near the peak location, chorus plays an important role in

accelerating seed electrons to MeV in the early recovery phase (from 03/18 to 03/20) and radial diffusion helps accelerate electrons to even higher energies.

§ Away from the PSD peak, radial diffusion, hiss scattering loss, and other processes are required to fully understand electron dynamics.

§ At L = 3, MS waves are primarily responsible for causing butterfly distribution in the mid recovery phase.

Future Work§ Evaluate the potential effect of EMIC-driven precipitation loss§ Incorporate more realistic DLL into our diffusion model