DREAM: The Dynamic Radiation Environment Assimilation Model

1
DREAM The Dynamic Radiation Environment
Assimilation Model

• Developed by LANL to quantify risks from natural
  and nuclear belts
• Uses Data Assimilation with GEO, GPS and other
  observations
• Couples ring current, magnetic field, and
  radiation belt models
• Goals Specification, Prediction, and
  Understanding

• Radiation belts produced by high altitude
  nuclear explosions are not included in this talk

2
Building Blocks of the DREAM Model
Plasma Sheet Boundary _at_ Geo
Ring Current Model (RAM)
Global B Model
Radiation Belt Electron Data
H
Physics Model
Phase Space Density vs. µ, K, L
DREAM Data Assimilation
3
Model ValidationPredict HEO fluxes using GPS,
GEO, Polar
1 MeV electron flux at HEO-3
?
HEO is a different orbit and is not used as input
to DREAM
Time (14 months 1990-1991)
4
Different Radiation Belt Specification Models
compared to HEO data
DREAM Data Assimilation
CRRES Model
AE8 Model
HEO Observations
Geomagnetic Activity
5
The ratio of Model Flux/Observed Flux gives
quantitative assessment of accuracy
Model 20x too low
Model 100x too high
6
Average HEO flux(L) compared to DREAM,
CRRES-EL, and AE8

• Three radiation belt models compared to HEO
• Averages and Standard Deviations (dotted)

7
Average HEO flux(L) compared to DREAM,
CRRES-EL, and AE8

• AE-8 model
• Wrong average profile
• Standard deviation 0

8
Average HEO flux(L) compared to DREAM,
CRRES-EL, and AE8

• CRRES Electron model based on 1990-1991 data
• Wrong average profile for 1 MeV
• Small standard deviation due to limited model
  variation with activity

9
Average HEO flux(L) compared to DREAM,
CRRES-EL, and AE8

• DREAM with GEO and GPS data
• Model has losses but no acceleration or pitch
  angle scattering
• HEO data not used in the assimilation

10
Distribution of HEO Fluxes at L 6Relative to
average flux
Log10 (Flux / ltFluxgt)
10 x
11
Average Ratio of Fluxes Model Flux / Observed
Flux

• Flux ratio is one metric for comparison
• Log scale 0 perfect match
• Inclined profiles can only be made to match at
  one L

12
Prediction Efficiency tests ability to model
variations around average

• Prediction Efficiency 1 perfect match
• PE 1 -

S (model - measurement)
S ( measure - ltmeasuregt)
13
Radiation Belt Climatology Available Data Sets
(so far)
14
Phase Space Density Peaksidentify local
acceleration

• Radial Diffusion (the classical theory) produces
  smoothly increasing profiles
• Wave-Particle Acceleration accelerates particles
  locally and should produce peaks
• A time-variable outer boundary of the belts can
  also make peaks
• PSD Peaks investigated by Selesnick and Blake
  1997, Green and Kivelson 2004, Iles et al.
  2006, Shprits et al. 2007,Chen et al. 2007

15
Simultaneous, 5-satellite measurements show
formation of equatorial PSD peak
Chen et al., Nature Physics, 2007
16
PSD Radial Profiles 2001 2002 LANL Geo, GPS,
Polar
Mu2083MeV/G 1MeV at GEO K0.03G1/2RE
17
Average PSD Radial Profile 2001 2002 LANL
Geo, GPS, Polar
Mu2083MeV/G 1MeV at GEO K0.03G1/2RE
18
Radiation Belt Climatology Data Intervals
Analyzed for Radial PSD Profiles
19
PSD Radial Profile 1984 1985 LANL Geo, GPS,
SCATHA
Mu2083MeV/G 1MeV at GEO K0.03G1/2RE
20
Average PSD Radial Profile 1984 1985 LANL
Geo, GPS, SCATHA
Mu2083MeV/G 1MeV at GEO K0.03G1/2RE
21
PSD Radial Profile 2005 GPS, Polar
Mu2083MeV/G 1MeV at GEO K0.03G1/2RE
22
Average PSD Radial Profile 2005 GPS, Polar
Mu2083MeV/G 1MeV at GEO K0.03G1/2RE
23
Initial Results Solar Phase does not (yet) show
statistically significant differences in PSD
profiles
24
Precipitation Distribution Solar-cycle
Dependence 1992-2008
SAMPEX
NOAA POES
25
Precipitation Distribution IV Local-Time
Dependence
26
Precipitation Distribution Kp Dependence