thermal structure of the topside ionosphere at low ... · at low latitudes: new observational...

Thermal Structure of the Topside Ionosphere

at Low Latitudes:

New Observational Opportunities

Pei Chen Lai and William J. Burke

Boston College/Institute for Scientific Research

19 March 2014

Knowledge about the range of states that the topside ionosphere can assume and the conditions

that give rise to them is essential for improving existing models used in a host of practical and

research applications. Still, the topside is at best a partially explored region. Electron density

profiles (EDPs) acquired during COSMIC-GPS radio occultation events offer promise for fuller

views of the topside. However, their reliance on Abel inversions has given users pause.

This presentation proceeds in three stages: First, we review the physics and mathematical

techniques underlying EDP extractions during occultation intervals as well as the measurement

capabilities of other relevant in situ sensors. Second, recognizing that reliance on Abel inversions

constitutes the technique’s Achilles heel, we performed a limited comparison of COSMIC EDPs

with electron densities measured by sensors on the Communications/Navigation Outage Forecast

System (C/NOFS) satellite during conjunction intervals. Results show that COSMIC EDPs were

in closer agreement with ion densities measured by the C/NOFS than were those predicted by

widely used models. Third, we outline a new technique that combines the capabilities of sensors

on the COSMIC, C/NOFS and DMSP satellites to infer altitude profiles of electron and ion

temperatures (Te, Ti), mean ion masses <mi> and ionospheric scale heights H at altitudes between

the F-layer peak and 850 km. Data acquired during eight COSMIC-C/NOFS-DMSP conjunctions

on 24 – 25 October 2011 are used to demonstrate the proposed method’s feasibility then outline

our plan to apply it to large databases. Our ultimate goal is to specify topside EDP taxonomies

that occur at low latitudes as well as the temperature and mass distributions required to support

them.

Abstract

Outline and Objective

• This presentation addresses five questions:

(1) Where is GPS and how does it operate?

(2) What is COSMIC and what does it measure?

(3) How do COSMIC receptions of GPS signals get turned into EDPs?

(4) Can we trust COSMIC EDPs and learn something new from them?

(5) What value added do data from RPAs on C/NOFs and DMSP bring?

• Our objective is to introduce this audience to GPS-COSMIC capabilities

for providing useful information about the thermal structure and dynamics

of the low-latitude ionosphere in the altitude range 200 to 800 km.

Global Position System

• GPS consists of 24 operational satellites flying

in 55o inclined, circular orbits at an approximate

altitude of 22,000 km.

- 6 orbital planes with 4 satellites per plane

- Dual frequency transmitters

f1 = 1.57542 GHz

f2 = 1.2276 GHz

• GPS receivers identify incoming code with a

precision of better than 1 ns (30 cm) by receiver

clock from the carrier waves “precise ranging

code” (10.23 Mb/s).

• The accuracies of clocks between COSMIC and

GPS are 10 to 15 m and < 1μs.

http://upload.wikimedia.org/wikipedia/commons/9/9c/ConstellationGPS.gif

Constellation Observing System for

Meteorology Ionosphere and Climate

• COSMIC is a joint Taiwan - US mission

that was launched into 72o inclined orbits

from Vandenberg AFB on 14 April 2006.

• It consists of 6 identical 3-axis stabilized

satellites

• Initial altitude: 500 km.

• Sequentially raised to ~ 800 km.

• Orbital nodes separated by ~ 2 hours

in local time.

COSMIC Payload

GOX Occultation Antenna

GOX Occultation Antenna

COSMIC-GPS Radio Occultations

Alt

itu

de

Temperature

EDP

Schematics of COSMIC-GPS

radio occultation events

Cosmic may rise above

or

sink below GPS horizon

Schreiner et al. (1999), Radio Sci.,

34, 949–966.

GPS-to-COSMIC Propagation:

• The well-known dispersion relation for EM waves propagating in the

ionosphere with >> pe >> ce >> en is: ,

where and

• The phase v and group vg speeds of the waves are:

• Group delays and phase advances depend only on electron

densities ne encountered along propagation path lengths s:

2 2 2 2( ) /pek c

2

22

2

2 2

2 2

(1 )2

1

1 (1 )2

pe

pe

pe pe

g

cv c

v c c

2

2 2 2

0

40.3

2 (2 )

COSMIC

e

e GPS

qt n ds STEC

c m f cf

2

0/ 2pe e e peq n m f 3( ) 8.98 ( )pe ef kHz n cm

Define the excess phase parameter S as the difference between signals

propagating in vacuum over distance |rCOSMIC – rGPS| and along S0 the

actual ray path through the ionosphere.

S = S0 - |rCOSMIC – rGPS|

where 1- fpe

2 / 2 f 2 is the index of refraction.

0

COSMIC

GPS

S ds

GPS-to-COSMIC Propagation: Applications


Schreiner et al. (1999) showed that

bending angles for L1 and L2 signals

are very small:

• 1.118 10-4 for ne = 1010 m-3

• 2.07 10-3 for ne = 1013 m-3

Hence, we assume that GPS signals

propagate along very similar paths.

Thus STEC can be calculated with

either f 1 or f 2 or both frequencies.

Bottom Line: Approximate GPS-COSMIC propagation paths as straight lines.


1 1 2 2 1 2 1 2

2 2

1 2

( )

40.3 40.3 40.3( )

S f S f S S f fSTEC

f f

• Using both frequencies any clock-based errors exactly cancel.

• Occultation intervals last about 12 minutes

• Since STEC measurements are recorded at rate of ~1 per second,

about 700 are accumulated to form each EDP.

• Inverse Abel transformations are then used to calculate ne along

vertical tangent lines.

0

0

0 0

2 2

0

( ) /1( )

COSMICr r

e

r r

dSTEC r drn r dr

r r


• r0 = distance from center of Earth to a specific altitude

• r = distance from center of Earth to height of tangent point

• ne profiles can then be integrated to obtain vertical TEC

Critical assumptions for valid Abel inversions:

(1) Propagation paths are nearly straight lines

(2) GPS and COSMIC orbits are nearly circular

(3) Electron density profiles are spherically symmetric,

i.e. horizontal gradients along ray paths are weak.

800

200

( )

km

e

km

VTEC n h dh

Cosmic Data Products and Availability

• COSMIC is administered by the

National Space Organization (NSPO)

and

the University Corporation for

Atmospheric Research (UCAR)

• EDP data are available in tabular

format via the internet from the

\Taiwan Analysis Center for COSMIC

(TACC): http://tacc.cwb.gov.tw/

and

UCAR: http://www.cosmic.ucar.edu/.

• The example to the right shows an EDP

and useful ephemeris information

derived from downloaded data files.

http://tacc.cwb.gov.tw/

http://www.cosmic.ucar.edu/

Complementary Data Sources: SSIES on DMSP & CINDI on C/NOFS

Current – Voltage Sweeps

e-

negative bias

swept voltage

negative bias

collector plate

i+ i+

to electronics

photo & secondary

electrons

Vsat

RPA Schematic RPA Measurements:

(1) 3 components of ion drift

velocities (Vi)

(2) Ni tot, Ni O+, and Ni light

(3) Ion temperatures Ti

(4) Mean ion mass < mi >

• SSIES has a boom-mounted

spherical Langmuir probe to

measure electron densities Ne

and temperatures Te.

• Infer topside scale heights

H = kB(Te +Ti) / < mi> g

Log I

-6 -4 -2 0 2 4 6 8 10

Slope -1/Ti

Intercept Ni

Applied Voltage

Two COSMIC Case Studies: Solar Min and Max Storms

• To help understand COSMIC measurements we undertook two case studies

• The first case focused on VTEC measurements acquired during an 80-day period

in late 2007 in which ejecta from a coronal hole swept by Earth three times.

• During the first and third encounters the corotating interaction region (CIR)

at the leading edge of the high-speed stream evoked weak responses in

the dayside ionosphere.

• Strong responses seen during the second encounter demonstrated effects of

penetrating electric fields generated by complex interplanetary sources that

included the near simultaneous arrival of an ICME.

• The second case study sought to determine whether COSMIC EDPs are

trustworthy. Their differences from predictions of the NeQuick model led

previous investigators to conclude that horizontal electron density gradients

degrade COSMIC EDPs to unacceptable levels. We present empirical

tests of this dire conclusion.

Case Study 1: November 2007 Storm

EIT image from SOHO on

18 November 2007 showing

a large coronal hole near the

Sun’s central meridian

Schematic representation of a corotating

region at the leading edge of a high speed

stream in the solar wind approaching Earth.

Within streams nSW is very low.

Magnetic flux emanating from coronal holes are unipolar, with radial

components that point either toward or away from the Sun.


80-day period centered on

the November 2007 storm:

(A) F10.7: daily and 81-day running

averages

(B) NSW (red) and VSW (blue)

(C) IMF BX (blue) and BY (red)

(D) Magnetospheric electric field

(E) Dst index

Vertical dash, marking the arrivals of high speed streams in the vicinity

Earth, are separated by 27-day solar rotation periods.


Days 322 – 327, 2007

(A) Solar wind density (red) and speed

(blue)

(B) IMF BX (blue) and BY (red):

Note: crossing of heliospheric current

sheet (HCS).

(C) IMF BY (red) and BZ (blue)

(D) Magnetospheric E field (~ 1 mV/m)

(E) Sym-H index minimum (~ -70 nT)

Note: relative UTs of ICME, CIR,

HSS and HCS

UT

• VTEC Distribution sampled by

COSMIC plotted as functions

of local time in 9 latitude bins

in northern (left) and southern

(right) hemispheres.

• From this perspective VTEC

increased during the storm’s

main phase and soon relaxed.

• Apparently, no surprises!


COSMIC VTEC versus Local Time

Case Study 1: November 2007 Storm COSMIC VTEC versus Universal Time

• Distribution of VTEC sampled by

COSMIC: days 322 – 327, 2007

plotted as functions of universal time

in 9 latitude bins in northern (left)

and southern (right) hemispheres.

• Sym-H index and VS in bottom plots

• Viewed from this perspective we

see that VTEC increased on during

day of the storm’s main phase then

relaxed.

• Decreased during first half of say 325!

• Why? Penetration electric field

and deviation from photo-

chemical equilibrium.

① ②

④ ⑥

COSMIC EDPs: Study 2 COSMIC – C/NOFS – DMSP

Conjunctions

CV

24 and 25 October 2011

Interplanetary Drivers and

Geomagnetic Responses

0

10

20

30

40

0

200

400

600

800

NS

W (

cm-3

)

PS

W (

nP

a)

VS

W (k

m/s)

-30

-20

-10

0

10

20

30

BX

BY

BZ

(n

T)

-200

-150

-100

-50

0

50

100

297:00 297:12 298:00 298:12 299:00

Sy

m H

(n

T)

Main

Phase

Recovery

Phase SSC

Top Panel Traces:

• NSW density (red)

• VSW speed (black)

• PSW dynamic pressure (blue)

Middle Panel Traces

• IMF BX (black)

• IMF BY (red) 1-minute averages

• GSM coordinates • IMF BZ (blue)

Bottom Panel Trace:

• Sym H index (black)

• Red dots indicate UT of EDP acquisitions.

• Vertical dashed lines mark beginnings

of main & recovery phases

COSMIC EDP: Study 2 Interplanetary and Storm Dynamics

COSMIC EDP: Study 2 C/NOFS and Model Comparisons

Z

Pre

-Sto

rm

Main

Ph

ase

Reco

ver

y

Midnight Dawn Noon Dusk (A-P)

(A-M)

(A-R)

(B-M)

(B-P)

(B-R)

(C-P)

(C-M)

(C-R) (D-R)

(D-P)

(D-M)

COSMIC EDP Study Statistical Comparisons with C/NOFS

1000

104

105

106

107

1000 104

105

106

107

PB

Mod

:

Ne [

cm-3

]

C/NOFS: Ni [cm

-3]

Ne = 6.31 * N

i 0.848

R = 0.774

1000

104

105

106

107

1000 104

105

106

107

NeQ

uic

k:

N

e [cm

-3]

C/NOFS: Ni [cm

-3]

Ne = 37.88 * N

i 0.698

R = 0.799

1000

104

105

106

107

1000 104

105

106

107

CO

SM

IC:

N

e [

cm-3

]

C/NOFS: Ni [cm

-3]

Ne = 0.612 * N

i 1.03

R = 0.858

Electron densities from EDPs:

• COSMIC (left) Log – Log Plots

• NeQuick (middle) Power - Law Regression Analyses

• PBMod (right)

Plotted as functions of Ni measured at C/NOF altitudes on days 297 ( ) and 298 ( ).

Dotted lines are guides indicating results if inferred Ne from EDPs = Ni from CNOFS

650 700 750 800 850 900 950

Altitude (km)

650 700 750 800 850 900 950

10

11

12

13

14

Altitude (km)

Ln

Ne (

cm-3

)

( ) ( )1/

( )

e B e i

i

dLn N k T TH

dh m g h

Top:

• Orange, black and purple

lines mark EDPs from

PBMod, COSMIC & NeQuick

• Red/blue dots show Ni

from C/NOFS / DMSP

Bottom:

• Linear regressions Ln (Ne)

versus altitude for h > 700 km

COSMIC Electron Density Profiles

Topside Scale Heights

COSMIC-CNOFS-DMSP Conjunctions

Event Te (K) Ti (K) HDMSP (km) HCOS (km) HNeQ (km) HPBM (km)

1 1697 1801 236.7 233.5 231.4 236.0

2 2997 2512 372.6 346.7

204.4 351.5


Topside Scale Heights

At DMSP altitudes RPA data showed that O+ was the dominant ion

High degree of agreement achieved between scale heights calculated with mi, Te, Ti

from DMSP and those from COSMIC and PBMod EDP slopes.

Recently we developed a low-pass, Fourier fitting procedure that

is piecewise continuous at the altitude of the F layer peak h = hp

12

12.5

13

13.5

14

14.5

15

12 12.5 13 13.5 14 14.5 15L

n (

Ne*)

Ln (Ne COSMIC

)

Ln (Ne*) = 0.00016 + 0.9999 Ln (N

e COSMIC )

R = 1

12

12.5

13

13.5

14

14.5

15

200 300 400 500 600 700 800 900

Ln

(N

e C

OS

MIC

)

Ln

(N

e*)

h (km)

UT: 297:10:00

GLat: 6.8o

GLong: 104o

LT: 17:00

Ne max

: 1.47 106 cm

-3

hp 433 km

4

*

0

ln ( ) ( ) ( )e p k k

k

N h h a Cos k b Sin k

| |

800

p

p

h h

h

4

*

0

ln ( ) ( ) ( )e p k k

k

N h h c Cos k d Sin k

| |

200

p

p

h h

h

1*( ) ( )

( )

B e i e

i

k T T dLn NH

m g h dh

( ) ( ) ( )( ) ( )i

e i

B

m h g h H hT h T h

k


Fitting Procedure

Case 1

In all examples R > 0.999

CINDI on CNOFS measures: <mi>, Ti and Vi

Since

COSMIC & PBMod EDPs

with 1st and 2nd Derivatives

during conjunctions with CNOFS & DMSP

10

11

12

13

14

15

200 300 400 500 600 700 800 900

CASE 1

Ln Ne PBMLn(Ne cosmic)Ln Ni CNOFSLn Ni DMSP

Ln

Ne L

n N

i C

osm

ic

PB

M

CN

OF

S

DM

SP

h (km)

UT: 297:10:00

GLat: 6.8o

GLong: 104o

LT: 17:00

Ne max

: 1.47 106 cm

-3

hp 433 km

-0.01

-0.005

0

0.005

0.01

200 300 400 500 600 700 800 900

d L

n N

e /

dh

(k

m-1

): C

OS

MIC

PB

M

h (km)

-0.005

0

0.005

200 300 400 500 600 700 800 900

d2 L

n N

e/d

h2 (

km

-2):

C

OS

MIC

P

BM

h (km)

10

11

12

13

14

15

200 300 400 500 600 700 800 900

Ln (Ne*)Ln Ne PBMLn (Ne CNOFS)

Ln (Ne DMSP)Ln

Ne:

C

OS

MIC

P

BM

C

NO

FS

D

MS

P

h (km)

Case 2 UT: 297:04:39

LT: 08:00

GLat: 4.8o

GLong: 50.2o

Ne max: 1.49 106 cm

-3

hp: 279 km

-0.01

-0.005

0

0.005

0.01

200 300 400 500 600 700 800 900

d L

n N

e /d

h (

km

-1):

C

OS

MIC

P

BM

h (km)

-0.002

-0.001

0

0.001

0.002

200 300 400 500 600 700 800 900

d2 L

n N

e /

dh

2 (

km

-2)

CO

SM

IC

PB

M

h (km)

COSMIC & PBMod EDPs



-0.01

0

0.01

100 200 300 400 500 600 700 800 900

d2 L

n N

e / d

h2 (

km

-2):

C

OS

MIC

P

BM

h (km)

-0.02

-0.01

0

0.01

0.02

200 300 400 500 600 700 800 900

d L

n N

e / d

h (

km

-1):

C

OS

MIC

P

BM

h (km)

10

11

12

13

14

15

200 300 400 500 600 700 800 900

CASE 3

Ln (Ne*)

Ln Ne PBMLn (Ne CNOFS)Ln (Ne DMSP)

Ln

Ne:

CO

SM

IC P

BM

C

NO

FS

D

MS

P

h (km)

UT: 297:19:00

LT: 08:00

GLat: -5.8o

GLomg: 154o

Ne max

: 3.77 10U cm-3

hp: 300 km

-0.02

-0.01

0

0.01

0.02

200 300 400 500 600 700 800 900

d L

n N

e /d

h (

km

-1):

C

OS

MIC

P

BM

h (km)

-0.02

-0.01

0

0.01

0.02

200 300 400 500 600 700 800 900

d L

n N

e /d

h (

km

-1):

C

OS

MIC

P

BM

h (km)

9

10

11

12

13

14

200 300 400 500 600 700 800 900

Ln (Ne*)

Ln Ne PBMLn (Ne CNOFS)Ln (Ne DMSP)

Ln

Ne:

C

OS

MIC

P

BM

C

NO

FS

D

MS

P

h (km)

Case 4 UT: 297:19:15

LT: 04:55

GLat: -4.8o

GLong: 145o

Ne max

: 2.7 105 cm

-3

hp: 270 km

COSMIC & PBMod EDPs



-0.01

-0.005

0

0.005

0.01

200 300 400 500 600 700 800 900

d2 L

n N

e / d

h2 (

km

-2):

C

OS

MIC

P

BM

h (km)

-0.01

-0.005

0

0.005

0.01

200 300 400 500 600 700 800 900

d L

n N

e /d

h (

km

-1):

C

OS

MIC

P

BM

h (km)

10

11

12

13

14

15

200 300 400 500 600 700 800 900

CASE 5

Ln (Ne*)Ln Ne PBMLn (Ne CNOFS)Ln (Ne DMSP)

Ln

Ne:

C

OS

MIC

P

BM

C

NO

FS

D

MS

P

h (km)

UT: 297:04:39

LT: 18:37

GLat: 7.1o

GLong: 59.4o

Ne max

: 1.42 106 cm

-3

hp: 342 km

-0.005

0

0.005

200 300 400 500 600 700 800 900

d2

Ln

Ne /

dh

2 (

km

-2):

C

OS

MIC

P

BM

h (km)

-0.02

-0.01

0

0.01

0.02

200 300 400 500 600 700 800 900

d L

n N

e/dh

(k

m-1

) C

OS

MIC

P

BM

h (km)

10

11

12

13

14

15

200 300 400 500 600 700 800 900

CASE 6 Ln (Ne*)Ln Ne PBMLn (Ne CNOFS)

Ln (Ne DMSP)

Ln

Ne:

C

OS

MIC

P

BM

C

NO

FS

D

MS

P

h (km)

UT: 298:12:55

LT: 18:00

GLat 4.5o

GLong: 76o

Ne max

: 1.66 106 cm

-3

hp: 472 km

COSMIC & PBMod EDPs



-0.001

0

0.001

200 300 400 500 600 700 800 900

d2 L

n N

e /

dh

2 (

km

-2):

C

OS

MIC

P

BM

h (km)

-0.02

-0.01

0

0.01

0.02

200 300 400 500 600 700 800 900

d L

n N

e /

dh

C

OS

MIC

P

BM

h (km)

10

11

12

13

14

15

200 300 400 500 600 700 800 900

Case 8

Ln (Ne*)Ln Ne PBMLn (Ne CNOFS)

Ln (Ne DMSP)

Ln

Ne:

C

OS

MIC

P

BM

C

NO

FS

D

MS

P

h (km)

UT: 297:03:39

LT: 05:45

GLat: -9.5o

GLong: 34.1o

Ne max

: 7.3 105 cm

-3

hp: 238 km

10

11

12

13

14

15

200 300 400 500 600 700 800 900

CASE 7

Ln (Ne*)Ln Ne PBMLn (Ne CNOFS)Ln (Ne DMSP)

Ln

Ne:

C

OS

MIC

P

BM

C

NO

FS

D

MS

P

h (km)

UT: 298:03:15

LT: 05:12

GLat: -6.3o

GLong: 26.7o

Ne max

: 4.4 105 cm

-3

hp: 277 km

-0.005

0

0.005

200 300 400 500 600 700 800 900

d2 L

n N

e / d

h2 (k

m-2

): C

OS

MIC

P

BM

h (km)

-0.01

-0.005

0

0.005

0.01

200 300 400 500 600 700 800 900

d L

n N

e / d

h (k

m-1

): C

OS

MIC

P

BM

h (km)


Summary and Conclusions:

• Presented case studies were undertaken to establish a feasible methodology

for testing the reliability of COSMIC-based topside EDPs and VTEC estimates

• The November 2007 and October 2011 storms provided a variety of external

driving conditions.

• The PLP on CNOFS provided high resolution ion densities for comparison with

EDPs estimated from COSMIC STEC measurements and model predictions.

• EDPs from COSMIC acquired within 15 were in better agreement with CNOFS

Ni measurements than model predictions.

• Early comparisons indicate that further study and analysis is worthwhile, using

Ni , Ti and Te from DMSP and C/NOFS to estimate topside thermal distributions.

• We have taken first steps towards developing an AI approach to EDP evaluations.

thermal structure of the topside ionosphere at low ... · at low latitudes: new observational...

Documents