International Reference Ionosphere

Introduction and Open Problems

Dieter Bilitza

George Mason University

Department of Physics and Astronomy, Fairfax, Virginia, USA

NASA, Goddard Space Flight Center

Heliospheric Laboratory

Greenbelt, Maryland, USA

1

IRI – Terms of Reference

� IRI is an international project jointly sponsored by the Committee on Space Research(COSPAR) and the International Union of Radioscience (URSI) to develop and improvea reference model for the most important plasma parameters i n the Earth ionosphere

� COSPAR’s prime interest is in a general description of the io nosphere as part of theterrestrial environment for the evaluation of environment al effects on spacecraft andexperiments in space.

� URSI’s prime interest is in a reference model for defining th e background ionospherefor radiowave propagation studies and applications.

� By charter the model should be primarily based on experimental evidence using allavailable ground and space data sources and should not depen d on the evolvingtheoretical understanding of ionospheric processes.

� As new data become available and as older data sources are ful ly evaluated andexploited, the model should be revised in accordance with these new results.

� Where discrepancies exist between different data sources t he IRI team should facilitatecritical review and discussions to determine the reliability of the different data sets andto establish guidelines on which data should be used for IRI m odeling.

irimodel.org2

� - are based on ground and/or space data.

� - use appropriate mathematical functions to represent the characteristic variation patterns seen in long data records.

� - do not depend on our evolving understanding of the processes that shape the ionosphere environment.

� - are, generally, available to users in the form of a computer program that runs faster than the more complex, theoretical models and that can be easily adapted a specific problem.

� - consider the production, loss, and transport of ionospheric particles (electrons and ions) and energy, and determine the electron and ion densities and temperatures from the balance of these processes

� - use a numerical (iterative) scheme to solve the Boltzmann equations for the ionospheric gas.

� - are primarily used in an investigative mode, to elucidate specific processes and their effects on the ionosphere

� - require considerable computer time, even on fast machines, and their complex computer code and logic make it difficult for people other than the model developers themselves to apply the model to a specific problem.

In actual fact, most ionospheric models are hybrids, using empirical as well as theoretical elements. An empiricalmodel, for example, may use theoretical values to fill data gaps, whereas a theoretical model may use a data-based mapping for specific parameters as input (e.g., to represent neutral densities and temperature), or as anormalization tool for areas of model-data discrepancies.

Empirical Models Theoretical Model s

3

4

M. Abdu (Brazil)

R. Ezquer, M. Mosert-Gonzalez (Argentina)

K. OyamaK. Igarashi S. Watanabe (Japan)

P. Wilkinson B. Ward (Australia)

O. Obrou (Ivory Coast)

M.-L. ZhangJiankui Shi, W. Wan (China)

J.-Y. Liu , S.-Y. Su (Taiwan)

Kyoung Min(South-Korea)

M. Rycroft, L. Cander (U.K.), A. Alcayde, R. Hanbaba (France), W. Singer, C. Stolle (Germany), B. Zolesi, S. Radicella (Italy), M. Friedrich (Austria),D. Altadill (Spain), F. Arikan (Turkey)

D. Anderson, E Araujo-Pradere, D. Bilitza, M. Codrescu, T. Fuller-Rowell, I. Galkin, C. Mertens, B. Reinisch, L. ScherliessJ. Sojka, V.WickwarS-R Zhang (USA)

J.B. Habarulema (Uganda, S.-Africa)4

K. Mahajan S.P. Gupta P.K. Bhuyan(India)

P. Sibanda(Zambia)

A. Poole, L-A. McKinnell(South-Africa)

J. AdeniyiE. Oyeyemi (Nigeria)

A. Danilov, V. Depuev, T. Gulyaeva, A. Mikhailov, I. Zakharenkova, S. Pulinets, K. Ratovsky (Russia)

L. Triskova, V. Truhlik, D. Buresova (Czech Rep) I. Kutiev (Bulgaria) A. Krankowski, H. Rothkaehl I. Stanislawska (Poland), I. Cherniak (Ukraine)

P. Supniti, P. Kenpankho(Thailand)

IRI Working Group Members (65/28)IRI Working Group Members (65/28)IRI Working Group Members (65/28)IRI Working Group Members (65/28)

D. Themens (Canada)

2010 Representation of the Auroral and Polar Ionosphere in IRIAdvances in

Space ResearchVolume 51, Number 4,

IRI Workshops and Publications

2017Real-time Ionospheric Predictions with IRI

and COSMIC and other GNSS Data

2018 COSPARGeneral Assembly Pasadena, California, USAJuly 14-22, 2018Description of the Ionosphere through Data Assimilation

irimodel.org

2019 IRI CCBWFrederick University,

Nicosia, CyprusSeptember 2-13, 2019

Real-time ionospheric modeling in the European and African sector

IRI CCBWNCU, Taiwan

Advances in Space Research

Volume 63Issue 6

2020COSPAR

General Assembly

Sydney, AustraliaAugust 15-23, 2020

Improving the Description of Hemispheric Differences in Ionospheric Models

5

URSIGeneral Assembly

Rome, ItalyAug 29- Sep 5, 2020

IRI: Progress in Retrospective and Real-Time Ionospheric Predictions

2015IRI at equatorial latitudes and

progress towards Real-Time IRI

2016

IRI CCBWBangkok, Thailand

COSPAR Istanbul, Turkey

Advances in Space Research

Volume 60Number 2

Improved Description of the Iono-sphere through Data Assimilation

Which Parameters in What Range?

IRI describes monthly averages of• electron density• electron temperature• ion temperature• ion composition (O +, H+, He+, N+, NO+, O2+, Cluster ions)

IRI represents variations with• altitude (50km – 2000 km)• latitude, longitude (geographic or geomagnetic)• date and time of day

External drivers:• solar indices: F10.7 (daily, 81-day, 365-day), R (13 month)• ionospheric index: IG (13 month)• magnetic indices: ap and kp (3-hour, daily)

Additional output parameters:• vertical ionospheric total electron content (vITEC)• ion drift at equator• occurrence probability for spread-F and for F1 layer

irimodel.org 6

Data Sources

Instrument Platform Parameter CommentsIonosonde Worldwide Ne from E From fifties

~170 stations to F peak to now

Incoherent Jicamarca, Ne whole Few radarsScatter Arecibo, profile incl. worldwideRadar St. Santin E-Valley Comparisons:

Millst. Hill, Te, Ti EISCAT,Malvern Ni, vi Kharkiv

Topside Alouette 1,2, Ne topside Data mostly Sounder ISIS 1,2, profile from 70s 80s

ISS-b, IK-19

In situ AE, Aeros,DE Ne Te Ti Coverage inDMSP, IK Ni vi LT, season, h CNOFS,Grace lat. dependsChamp … on s/c orbit

Rocket Rocket data D-region Sparse compilations parameters data set

GNSS GPS etc. TEC Incl. pTECAltimeter Topex, Jason EC < s/c alt. Oceans only

Radio LEO-GNSS: Ne profile SphericalOccultation COSMIC I,II symmetry

Champ Grace of Ne 7

Data Base of Satellite Insitu Measurements

Solar cycle variations are largest for

M3000F2 model, which is closest to

AMTB at high and Shubin at low solar

activity.

hmF2Comparison of the three hmF2 model

options now available in IRI-2016:

- CCIR-M3000F2 Ionosonde data

- AMTB (default) Digisonde data

- Shubin-2015 Radio occultation data

Discrepancies between different data sources

Te (830 km)Coincident measurements between DMSP 12,

13, 14, 15 and Millstone Hill incoherent scatter

radar over the time period 1996 – 2006 (Bilitza

et al. , 2007; Truhlik et al., 2009)

Similar discrepancies for daytime and

nighttime. MH shows much larger variability

than the DMSP probe data. The low electron

densities at 830 km make data analysis

difficult for both techniques. 9

SP

10

Build-up of IRI electron density profile

Mathematical functions:

Global Variations: • Special spherical harmonic functions

using modified magnetic coordinates (modip, invdip)

• Interpolation/transition functions(Epstein, others)

Time Variations: • Harmonics of different order• Smooth transitions between day and

night values (Epstein)

Height Variations:• Epstein skeleton function approach • Chapman

function

Global models for anchor points: foF2/NmF2 foF1/NmF1, foE/NmE, foD/NmD,hmF2/M(3000)F2, hmF1 , hmE, hmD

Normalized to E and F peaks

Built-up of the IRI electron density profile

F2 bottomside: B0, B1

F1 ledge: C1

Topside parameters

E-Valley: HBR, HABR, Depth,Slope at HEF

D-Region

E Bottomside

11

Longitude

Latit

ude

ISS-b data

12

NmF2/m-3 = 1.24E10 (foF2/MHz)2

Global Models for NmF2

Location of ionosonde stations used for the global foF2 model.

SP

CCIR (1967) global model for foF2

Jones and Gallet 1962 Jones et al., 1969

Ionosonde 1954–1958plus screen points in ocean areas

ComitéConsultatifInternationalRadio-communication

of the

International Telecommunication Union

..using a special coordinate: modified dip latitude

Order 6 for foF2Order 4 for M(3000)F2

13

… and a special set of geographic coordinate functions

76 coefficients for foF249 coefficinets for M(3000)F2

14

CCIR 1967Jones and Gallet, 1962Jones et al., 1969

Ionosondes 1954-1958

Screen points, especially over the oceans and southern hemisphere, through extrapolation along lines of constant modified magnetic dip angle

6th order harmonics in UT and special functions for global specification

Recommended over continents

URSI 1988Fox and McNamara,1988 Rush et al.,1989

45,000 station months of ionosondedata

Theoretical model was adjusted such that it agreed with measured foF2 values over land, and then was used to compute screen points over oceans and in data sparse regions

Same functions as CCIR

Recommended over ocean areas

Room for improvement: Large volume of more recent d ata, better models for screen point analysis, better functions, better computers

15

Digisonde GIRO Network

Global Ionosphere Radio Observatory

16

LGDC

Longitude

Latit

ude

ISS-b data IRI URSI-88

Damboldt and Suessmann (2012) validation with large volume of ionosondedata over 7 solar cycles finds average foF2 error of just a few %. 17

• Developed by Fuller-Rowell et al., 1998, 2000 and Aruajo-Pradere et al., 2003, 2004.• Based on data from 75 ionosonde stations for 43 storms prior to 2001.• Tested meanwhile with many more storms and stations. • Uses an index based on the weighted integral of the ap index over the previous 33 hours. • Describes quite well the mostly negative ionospheric storm effects at mid-latitude summer.• An evaluation with ionosonde data for 2001 storm periods showed an overall improvement of IRI by 50%.

foF2storm(t)/foF2quiet(t) = a0 + a1 X(t) + a2 X2(t) + a3 X

3(t)

X(t) = ∫ F(τ) ap(t- τ ) dτ over previous 33 hours, foF2quiet(t) is monthly mean

foF2 STORM MODEL

18

- 4

- 3 .5

- 3

- 2 .5

- 2

- 1 .5

- 1

- 0 .5

0

0 . 5

1

0 1 0 2 0 3 0 4 0 5 0

P r e v i o u s H o u r s

ampl

itude

o u t p u tf i l t e r

The optimum shape and length of the filter was obtained by the singular value decomposition method, minimizing the mean square difference between the ap index and the ionospheric ratios

foF2 Storm Model

SP

foF

2

UT/consecutive hours

ObservationsIRI-without stormIRI with storm

19

20

Epsteinfunctions

dT/dh = m1+Σi (mi+1 – mi)/(1+exp[-(h-hi)/di] Epstein step function:� Transition from day to night

value� Combining regions of

constant gradient (Booker function)

Epstein layer function:� Topside electron density

profile

International Reference Ionosphere - IRI

--------- Plasma Temperature ---------------

21

Te

Ti

Te

Equator

Mid-latitudes

daynight

daynightHei

ght /

km

Geomagnetic Latitude

IRI Plasma Temperatures

3000km1400km

night

day

Ele

ctro

n Te

mpe

ratu

re /

K

SP

International Reference Ionosphere - IRI

------------ Ion Composition ------------------

N+ Ne

O+

H+

He+

O+

NO+

O2+Ne

O+ O2+NO+

NeCluster +

22

IRI Ion Composition

NeO+

H+

He+

N+

Cluster Ions : Ion formed by the combination via noncovalent forces of two or more atoms or molecules of one or more chemical species with an ion, for example:[(H20)nH]+ or [(NaCl)nNa]+.

SP

Red lines = Auroral boundaries at 0.25 ergs/(cm2s).

Model representation of mean energy and energy flux of precipitating

electrons in terms of MLat, MLT, and Kp (0-10) by Zhang and Paxton (2008)

based on TIMED/GUVI measurements from 2002 to 2005.

� Auroral boundaries for IRI varying with Ap

� Real-time updating of IRI using GUVI or SSUSI data (Zhang et al., 2010).

International Reference Ionosphere - IRI

------------ Ion Composition ------------------ O+He+IRI Auroral Boundaries

Being an empirical model IRI has the advantage that it already includes some of the newly discovered phenomena that theoretical modelers still need to include in their modeling framework.

International Reference Ionosphere - IRI

------------ Ion Composition ------------------ O+He+Advantage of Empirical Models

Weddell Sea and Yakutsk Anomalies : foF2 in summer larger during nighttime than during daytime.

foF2midnight – foF2noon

IK-19

GSM TIP

IRI-2012

Causes for the anomalies:

(1) The meridional component of the neutral wind reaches maximum values at the anomaly sites (during summer nights) because of the combined effect of solar EUV heating (sunlit magnetic pole in summer) and heating from precipitating auroral electrons.

(2) The effect of the neutral wind on the vertical plasma transport is proportional to sinΨ · cosΨ with Ψ the magnetic inclination.

Adewale et al., Advances in Space Research 49 316–326, 2012 (GPS Receiver in Tukuyu, Tanzania and 12 other African stations)

de La Beaujardière et al. , C/NOFS observations of deep plasma depletions at dawn, Geophys. Res. Lett., 36, L00C06, 2009.

SP

International Reference Ionosphere - IRI

------------ Ion Composition ------------------ O+He+Comparing New Measurements against IRI

IMPROVEMENTS� New data and better data averaging/fitting methods and functions

� More accurate description of the driver->parameter relationship • IRI uses 13-month running means of R and IG for F-peak and other parameters• The sunspot number index R was recently significantly revised; a factor 1.4

increase (Clette et al., SW, 2015).• The ionosonde-based IG is better than EUV indices because it includes

dynamical effects; EUV indices are better than R and F10.7 because the ionospheric plasma is created by EUV radiation.

• Time delays of 1-3 days need to be considered when using solar indices.

� Updating IRI with real-time or retrospective data• IRI allows input of NmF2/foF2, hmF2, NmF1/foF1, NmE/foE, hmE, B0, B1 if

appropriate measurements are available.• Equivalent R and/or IG are obtained by adjusting IRI output with F-peak

measurements and/or TEC.

� Assimilating real-time or retrospective data into I RI• IRTAM: Real-time global digisonde data for foF2, hmF2, B0, and B1 • Various techniques for assimilating foF2 and TEC data into IRI.

27

o The global, continuous coverage provided by GNSS data makes them an important resource for improving IRI and IRI Real-Time.

o GNSS parameters:� GPS: TEC (slant, vertical); need additional analysis to get profile parameters

(correlation; tomography; need to subtract plasmaspheric contribution)

� Altimetry (TOPEX, Jason): vTEC; difficult to deduce profile information.

� Radio Occultation: whole profile; limitations in regions of strong spatial or

temporal gradients.

o Need to consider the error sources for the different techniques� GPS: receiver and transmitter biases, multi-path corrections, slant-to-vertical

transformation

� Altimetry: calibration

� Radio Occultation: spherical symmetry assumption

Build-up of IRI electron density profileGNSS Data and IRI

irimodel.orgirimodel.org

Build-up of IRI electron density profileIRI Specifics

Combining the global picture recorded by satellites for different LTs and levels over solar activity with the 24/7 356 analysis provided by ground stations.

Modular approach, e.g., global models for profile anchor points: Te at different heights from ISIS, AE, IK connected by Epstein skeleton function.

Options to switch on a different model: 3 model options for hmF2 using different data sources (Ionosonde M(3000)F2 or hmF2, COSMIC/RO); 3 model options for topside Ne profile using different formalism.

Avoid introducing interdependence between parameters because replacing one parameter model will affect the related parameter.

New, better models are easily phased in with validation help from the users

IRI drivers: F10.7 (daily, 81-day, 356-day), R (13 month), IG (13 month), Ap, Kp (3-hour, daily)

User input of measured parameters: NmF2/foF2, hmF2/M(3000)F2, NmF1/foF1, NmE/foE, hmE, NmD/foD, hmD, B0

irimodel.org30

Build-up of IRI electron density profileIRI +/-

+ Synthesis of almost all available and reliable ionospheric data.+ Widely used and validated; IRI is often the reference against which a

satellite/rocket team compare their new data. + Recognized standard by COSPAR, URSI, ITU, and ISO+ Improvements continue as new data become available.+ Includes effects not yet discovered/explained by theory (4-wave pattern,

Weddell Sea and Yakutsk Anomalies)+ IRI team’s global distribution guarantees access to the global data base+ Fared very well in model validation studies including the CEDAR

Challenge

− Only as good as the data foundation on which it was build. − Bias towards Northern mid-latitudes.− Data-sparse regions/times require inter/extra-polation− Recent very low and broad solar minimum brought conditions not

covered by previous solar minimum data sets.− Funding

irimodel.org31

Recent Publications

32

STATUS AND NEW VERSIONS:Bilitza D., D. Altadill, Y. Zhang, C. Mertens, V. Truhlik, P. Richards, L.-A. McKinnell, and B.

Reinisch: The International Reference Ionosphere 2012 – a model of international collabora-tion, J. Space Weather Space Climate, 4, A07, 1-12, doi:10.1051/swsc/2014004, 2014.

Bilitza D., The International Reference Ionosphere – Status 2013, Adv. Space Res., 55(8), 1914-1927, doi: 10.1016/j.asr.2014.07.032, 2015.

Bilitza, D., D. Altadill, V. Truhlik, V. Shubin, I. Galkin, B. Reinisch, and X. Huang, International Reference Ionosphere 2016: From ionospheric climate to real-time weather predictions, Space Weather, 15, 418–429, doi:10.1002/2016SW001593, 2017.

VALIDATION:Shim J. S., et al., CEDAR Electrodynamics Thermosphere Ionosphere (ETI) Challenge for

systematic assessment of ionosphere/thermosphere models: Electron density, neutral density, NmF2, and hmF2 using space based observations, Space Weather, 10, S10004, doi:10.1029/2012SW000851, 2012.

Shim, J. S., et al., (2014) Systematic Evaluation of Ionosphere/Thermosphere (IT) Models, in Modeling the Ionosphere-Thermosphere System (eds J. Huba, R. Schunk and G. Khazanov), John Wiley & Sons, Ltd, Chichester, UK. doi: 10.1002/9781118704417.ch13, 14 March 2014.

Shim J. S, et al., CEDAR-GEM Challenge for Systematic Assessment of Ionosphere/Thermos-phere Models in Predicting TEC during the 2006 December Storm Event, submitted to Space Weather, 2017.

THANK YOU

33

Build-up of IRI electron density profileNew Inputs for IRI-2018

� FIRI-2017: An updated model for the D-region electron density based on a simple ion-chemical model that is corrected with rocket radio propagation data (M. Friedrich, Graz University of Technology, Austria)

� IRI-Plas: Extension to the plasmasphere of electron density profile based on TEC-deduced adjustments and indices (T. Gulyaeva, IZMIRAN, Russia)

� hmF2 storm model: Representation of the effects of ionospheric storms on F2 peak height based on ionosonde data (D. Altadill and E Blanch, Ebro Observatory, Spain)

� Scintillation model: Occurrence probability for scintillations based on COSMIC data (Shihping Chen, National Central University, Taiwan).

� IRI Real-Time: Special IRI subroutines for reading IRTAM coefficients and calculating NmF2, hmF2, B0 and B1 (Dieter Bilitza, George Mason University)

irimodel.org34

3-hour daily 81-day 13-month

solar F10.7 F10.7 R, F10.7

ionospheric IG

magnetic ap, kp ap, kp

IRI drivers:

F10.7 solar photon flux at 10.7cm wavelength; can be measured from ground

R sunspot number

IG ionosphere-effective solar index, based on CCIR foF2 model and ionosondedata from globally distributed ionosonde stations. First 13 stations now 5.

ap, kp are derived from magnetometer measurements of the variation of the magnetic field due to currents flowing in the earth's ionosphere

Ma et al., 2009

IRI Drivers

35

F10.7 daily – blackF10.7 81-day - greenF10.7 13-month - blue

D-Region Bilitza, 1981 R12 Friedrich, 2001 F10.7D Danilov et al. 1995 F10.7D

E-Peak Kouris & Muggleton 73 COV F1-Region Ducharme et al. 1973 R12

Scotto et al. 1997 R12Bottom- Bilitza et al., 2000 R12

-side Altadill et al. 2009 R12F-Peak Jones&Gallet, 1965 IG12

Rush et al., 1989 IG12hmF2 Bilitza et al. 1979 R12

Altadill et al. 2013 R12Shubin 2015 F10.7_81

Topside Rawer et al. 1978 COVSATRadicella et al. 2006 R12

Spread-F Abdu et al. 2003 F10.7D

Te Truhlik et al. 2011 PF10.7

Tn CIRA Picone et al. 2002 F10.7_81o, F10.7DPo

Ni(h300km) Triskova et al. 2003 PF10.7

Ni(h300km) Danilov&Yaichnikov 85 F10.7_12

vi Scherliess&Fejer 1999 F10.7D

Electron

Density (Ne)

Temperatures

(Te, Ti, Tn)

Ion

composition

and velocity

(Ni, vi)

R12 = 12-month running mean of sunspot number

F10.7D = daily F10.7 index (adjusted to AU)

F10.7DP = daily F10.7 index of previous day

F10.7DPo = F10.7DP observed

F10.7_81 = 81-day running mean of F10.7

F10.7_12 = 12-month running mean of F10.7

PF10.7 = (F10.7_81 + F10.7D)/2

COV= 63.75+RSSN*(0.728+RSSN*0.00089)

COVSAT= COV if(COV.gt.188.) COVSAT=188

IG = Global Ionospheric index

IG12 = 12-month running mean of IG

36

Problems with R and F10.7 :

R is related to the number spots observed in sun’s photosphere. F10.7 is related to the solar irradiance at 10.7cm wavelength while the irradiance responsible for ionospheric ionization is in the range 1-120nm (X-ray to EUV).

□ F10.7+ R

37

Problems with solar indices R and F10.7

R was recently significantly revised; a factor ~1.5 increase (Clette et al., SW, 2015). Since the IRI models were developed with the ‘old’ R12 we use a scale factor of 0.7 with the new R12.

Correlation between F10.7 (R) andEUV fluxes measured by the EUVspectrometer on the AEROS-Asatellite at different wavelengths.

Solar irradiance driven indices

Using measured fluxes at specific wavelength bands:

MgII core-to-wing ratio Core lines at 279.5-284 nm Lyman α or β, He I or II 121.5nm, 102.6nm, 58.4nm, 30.4nmE10.7 SOLAR2000 (Tobiska)EUVAC Fluxes in 36 wavelength bands (Richards)

Full EUV spectrum models (0-120nm in steps of 1nm):

FISM (Chamberlin), NRLEUV2 (Lean): They use different methods to inter-calibrate the measurements from different instruments and to fill data gaps by assuming close correlation with proxies.

Ionizing solar irradiances for the different layers:

D-layer X-rays, Cosmic rays N2, O2Lyman α 121.5 nm NO

E-layer soft X-rays 1-10 nm O2Lyman β 102.6 nm O2He II 30.4 nm O

F-layer EUV 10-105 nm O

38

Solar EUV Indices

Problems with EUV indices :

Only available since satellite measurements became feasible. Long time gap with no EUV measurements.

39

Problems with EUV indices I

Inter-calibration issues between different satellites are still being debated.

Solar cycle amplitude varies with wavelength and is still uncertain because of the calibration uncertainties.

40

Correlation between the Lyman-βflux and the EUV fluxes at other wave-lengths measured by the AEROS-A satellite weakens as we move to smaller wavelengths.

Solar indices describe only solar influence (ionization) not dynamics of F region. Good in solar-controlled E and F1 region.

Problems with EUV indices :Problems with EUV Indices II

A time delay of 2-5 days has been observed between foF2 and TEC measurements and F10.7 as well as EUV indices. Most likely caused by the very long lifetime of O in the F-region.

Tai, Rawer, Bilitza and v. d. Mühle, KB 1983

41

Based on work of Liu et al., Telecomm. J., 1988:• Using noontime measurements of foF2 from 13 globally

distributed ionosondes. • Adjust R12 in CCIR foF2 model until agreement is achieved

between data and model.• Take median of the 13 adjusted R12 values.

PROBLEMS :• Stations have varied over time and are now down to 4

(Chilton, Port Stanley, Kokubunji, and Canberra)• IG-CCIR but not an equivalent IG-URSI• Only noon time data are being used.• Ionosonde indices work best locally or regionally; averaging

over too wide a latitude range greatly diminishes the usefulness of these indices.

41

Ionosonde Index IG

42

Forecasting Indices

IG-12 Ne(400km)

Feb 2011

Feb 2010

Nov 2008

Nov 2007

During the 2007-2010 minimum period:

Using indices forecasted one year ahead resulted in an overestimation of Ne(400 km) by 20%.

Using indices forecasted two years ahead results in an overestimation by 70%.

For TEC the corresponding numbers are 15% and 50%.

Effect of Prediction Uncertainties on IRI

(Neyyyy(400)-Ne

2011(400))/Ne2011(400)

Build-up of IRI electron density profileOngoing IRI Projects

Sporadic E: Model for the occurrence probability of sporadic E based on COSMIC radio occultation data (Christina Arras, GeoResearchCentre, GFZ, Potsdam, Germany).

Vary-Chap: New representation for the topside electron density profile using Alouette and ISIS topside sounder data and the TOPIST s/w (Osman Mohamed, University of Massachusetts Lowell, USA).

New IG indices: Using hemispheric indices and different IG indices for CCIR and URSI models (Steven Brown, George Mason University, USA)

Ne Tops Cor: Introducing solar cycle dependence in the IRI-2007cor option for the topside electron density profile based on CHAMP, GRACE, Swarm, Alouette, and ISIS data with the main goal to accurately represent the conditions at very low solar activity (Chao Xiong, GeoResearch Center, Germany)

irimodel.org44

• Direct modeling of hmF2 instead of using relationship with M(3000)F2.

• foF2 storm model includes only negative effects at mid-latitudes.

• Need to consider storm effects on hmF2 and topside and bottomside profile

• Representation at very low solar activity, recent very low and extended minimum (2008/2009)

• Replace older R-depending models with models depending on F10.7 and/or IG and use of monthly

indices instead of now 13-month averages. 45

Open Problems - 1

irimodel.orgirimodel.org

Done

Done for Ni working on Ne

Done

• Extending IRI to plasmaspheric altitudes • Better description of Equatorial Ionization

Anomaly characteristics

• Including models of the ion transition heights (molecular/O+/light ions) as anchor points for

the ion composition model

• Improved global models for occurrence probability of spread-F and of scintillations

• Assimilation of GNSS data for real-time IRI• IRI at high-latitudes (auroral oval, subauroral

trough, Te enhancement, polar cap)

46

Open Problems - 2

►►►►STANDARD FOR ENGINEERING APPLICATIONS

SOME APPLICATIONS

►►►► VISUALIZATION AND ONLINE TOOLS FOR SPACE ENVIRONMEN T

►►►► BACKGROUND IONOSPHERE FOR EVALUATING DATA RETRIEVAL TECHNIQUES (TOMOGRAPHY, RADIO OCCULATATION)

►►►► IONOSPHERIC CORRECTIONS FOR EARTH OBSERVATIONS FROM SPACE

►►►► IONOSPHERIC PARAMETERS FOR THEORETICAL MODELS

►►►► HF COMMUNICATIONS (FREQUENCY MANAGEMENT, HAM RADIO)

Density profile functions for the topside electron density for a scale height of 100 km

Difference between AE-C average profile and best-fitted approximation with the five most used topside functions.

Topside Modeling Functions - Stankov et al., 2003, Bilitza et al., 2005

Nighttime

Epstein N(h)/NmF2 = 4 exp(z)/(1+exp(z))2

often called sech-squared function

Chapman N(h)/NmF2 = exp{c[1-z-exp(-z)]}c=0.5, 1 Chapman α, β function

Exponential N(h)/NmF2 = exp(-z)

z=(h-hmF2)/H H=scale height

49

Components of foF2 Variability

Contribution to total variability in %

10-15 %30-40 %

5-7 %

1-5 %

LT: 10-14

1983 Stara Zagora, Bulgaria Towards an Improved IRI ASR 4( 1) 19841985 Louvain, Belgium IRI - Status 1985/86 ASR 5(10) 19851987 Novgorod, Russia Ionospheric Informatics ASR 8( 4) 19881989 Abingdon, UK Development of IRI-90 ASR 10(11) 19901991 Athens, Greece Adv. in Global/Regional Descript. Ionospheric Parameter ASR 12( 7) 19921993 Trieste, Italy Off Median Phenomena and IRI ASR 14(12) 19941995 New Dehli, India Low and Equatorial Latitudes in IRI ASR 18( 6) 19961997 Kühlungsborn, Germany New Develops. in Ionospheric Modeling and Prediction ASR 22( 6) 19981999 Lowell, MA, USA IRI- Workshop 1999 ASR 27( 1) 20012001 Sao Jose dos Campos, Brazil Description of the Low Latitude Ionosphere in the IRI ASR 31( 3) 2003 2003 Grahamstown, South Africa Quantifying ionospheric variability ASR 34( 9) 20042005 Roquetes, Spain New Data and Reg. Models ASR 39( 5) 20072006 Buenos Aires, Argentina Improved IRI TEC Repres. ASR 42( 4) 20082007 Prague, Czech Rep. IRI/COST Workshop ASR 43(11) 2009 2009 ColoradoSprings, USA Task Force Activity: Real-Time IRI2009 Kagoshima, Japan IRI'2009 Workshop EPS 63(4) 2011 and 64(4) 2012 2011 Hermanus, South Africa IRI over Africa 2012 Prague, Czech Republic COST & Real-Time IRI 2013 Olztyn, Poland IRI and GNSS2015 Bangkok, Thailand Rep. scintillations in IRI (MO: Pornchai Supnithi)

Fig. 2. Group photo of participants of 1993 IRI meeting in Trieste, Italy.

Fig. 2 . Group pho to of partici pant s of 1993 IRI meet ing in Trieste, It aly.

foF2 UT: 0 - 24 foF2 LT: 0 - 24

hmF2 UT: 0 - 24 log(Ne) UT: 0 - 24

52