D. Bilitza

The 2002 IRI Task Force Activity was held from August 5 to 9 at the Aeronomy and Radiopropagation Laboratory (ARPL) of the Abdus Salam International Centre for Theoretical Physics in Trieste, Italy. The meeting was hosted by the director of the ARPL, S. Radicella and was attended by J. Adeniyi from Nigeria, G. Miro and M. Cueto from Spain, K Alazo and B. Lazo from Cuba, B. Reinisch and D. Bilitza from USA, L. Ciraolo, P. Coisson, B. Nava, B. Forte, E. Canessa, M. Zennaro, and C. Fonda from Italy, S. Pulinets from Russia/Mexico, R. Leitinger from Austria, M. Mosert and M. Gende from Argentina. The main goal of this years’ activity was the establishment of a table of variability values for inclusion in the next version of IRI. This will provide IRI users with an estimate of the deviation from the monthly mean that could be expect for specified conditions. Another important topic was the development and testing of new models for the IRI topside electron density profile.

(1) Measure of Variability

Different options were discussed regarding the selection of the best parameter to describe ionospheric variability. The obvious choice is the monthly standard deviation in conjunction with the monthly mean. But ionospheric conditions often include a few disturbed days (storm days) during a month with values significantly different from the rest of the values during the month. A good method to exclude these outlying parameter values is the use of median and quartiles; median is the value at the midpoint of the monthly data set and upper and lower quartiles are defined as the values that are above 1/3^rd and 2/3^rd, respectively, of the data values. It was decided to choose the monthly median as the representative value for the month and the parameter

v = | upper quartile – lower quartile | / median * 100 [%]

as the measure of monthly variability. Monthly mean and the variability parameter

s = standard deviation / mean * 100 [%]

are a secondary choice. But when computing mean and s values periods of disturbance should be first excluded.

(2) Parameters Considered

During the 2001 IRI Task Force Activity it had been decided to establish a database to study the following parameters:

foF2, hmF2, B0, B1, D1, and Ne (hmF2+h) with h = -150, -50, 50, 100,

200, 300 km

for the following conditions:

LT: 10:00-14:00, 18:00-20:00, 22:00-2:00, 5:00-7:00

Month: representative month for Winter, Spring, Summer and Fall

Solar Activity: Low and High (e.g., best R12 = 0, 100)

Discussions during this TFA focused again on the subject of how best to describe the variability for the whole profile. Will it be sufficient to study and determine the variability of the peak and profile parameters (e.g., foF2, hmF2, B0, B1, D1) and then use variability models for these parameters to describe the variability at any height? But this could be a quite involved process since taking the highest/lowest peak and profile parameter values will not necessarily generate the highest/lowest density value at a specific height. Contributions from v(foF2) and v(hmF2) or v(B0) may cancel each other out. Alternatively one could consider studying the variability at fixed heights, e.g. at 110, 170, 250, 350, and 500 km. It was decide to first focus on variability models for the peak and profile parameters and then use these models to investigate how best to represent the variability at various heights.

(3) Stations Included

A list of ionosonde stations that are candidates for being involved in this variability study was included in the last report and an updated copy is included below

*Station*	*Location*	*Iono-sonde*	*GPS*	*lead*	*Used at TFA*
Ushuaia, Argentina	high-lat	yes	?	M. Mosert	yes
Belgrano, Argentina	high-lat	yes	?	M. Mosert	yes
Rom, Italy	mid-lat	yes	yes	G. Miro
El Arensillo, Spain	mid-lat	yes	yes	G. Miro	yes
Ebro, Spain	mid-lat	yes	no	G. Miro
Pruhonice, Czech Rep	mid-lat	yes	no	D. Buresova
San Juan, Argentina	mid-lat	yes	no	M. Mosert	yes
Buenos Aires, Argentina	mid-lat	yes	?	M. Mosert	yes
Toluca, Mexico	mid-lat	yes	?	S. Pulinets
Santiago, Chile	mid-lat	in past	yes	M. Mosert
Havana, Cuba	mid-lat	yes	no	A.Calzadilla	yes
Millstone Hill, USA	mid-lat	yes	no	?
Conception, Chile	mid-lat	yes	no	Foppiano
Grahamstown, South Africa	mid-lat	yes	?	L. McKinell
Tucuman, Argentina	crest	in past	yes	M. Mosert	yes
Shung-Li, Taiwan	crest	yes	yes	TBD
Cacheira Paulista, Brazil	crest	yes	?	I. Batista
Salta, Argentina	crest	no	yes	TBD
Acension Island	crest	?	?	TBD
Ougadougou, Burkina Faso	equator	yes	no	J. Adeniyi	yes
Ibadan, Nigeria	equator	yes	no	J. Adeniyi
Korhogo, Ivory Coast	equator	yes	no	O. Obrou
Jicamarca, Peru	equator	yes	yes	B. Reinisch
Forteleza, Brasil	equator	yes	yes	I. Batista
Arequipa, Peru	equator	no	yes	M. Mosert
Ibadan, Nigeria	equator	yes	?	J. Adeniyi

(4) Variability Studies presented during the TFA

Variability data for foF2, hmF2 and TEC measured above Havana, Cuba were presented by B. Lazo and K. Alazo. R. Ezquer, M. Mosert, G. Miro, and D. Buresova studied the variability patterns of foF2, foF1, foE, M(300)F2 recorded at several Argentine and European stations representing low, middle and high latitudes. R. Ezquer, C. Brunini, M. Mosert, and S. Radicella used data from 10 Argentine GPS stations to investigate the TEC variability in this longitude sector during high solar activity. They find a variability factor v of 46 – 78 % during nighttime and 34 – 47 % during daytime and maximun values during the sunrise period LT=5-7. foF2 variability at equatorial latitudes was reported by J. Adeniyi based on data from Ougadougou. M. Hernandez-Pajares and S. Pulinets studied the equator anomaly variability as a function of local time and longitude. The results of all of these investigations will be combined into a first preliminary model describing the variability of foF2, hmF2 and TEC. This will be accomplished in the form of a table of representative values similar to what was done for the shape parameter B0 in the bottomside density profile.

(5) Topside and Total Electron Content

A correction function for the IRI topside was presented based on Alouette and ISIS topside sounder data. The new topside model (old+correction) needs now to be tested with other data sources, e.g. IC-19 topside sounder data. A comparison of the current IRI topside model with ISIS topside sounder profiles illustrated the shortcomings of this model (K. Alazo, P. Coisson, S. Radicella). The ISIS profiles were a selection of data now produced at NSSDC from the digitized ISIS ionograms using the TOPIST automated scaling and inversion software. Different functions for representing the topside electron density profile were evaluated by C. Fonda, B. Nava, and P. Coisson. R. Leitinger discussed and explained the parameters used in his suite of topside models and profilers. The plasmaspheric electron content was the topic of presentations by B. Reinisch and by L. Ciraolo. B. Reinisch presented results from his Radio Probe Instrument (RPI), one of the instruments one the IMAGE satellite, and reported about the model that is now being developed at UML based on these data. L. Ciraolo has worked with data from the GPS receiver on the TOPEX satellite. These measurements provide the electron content from the TOPEX orbit altitude of 1340 km up to the GPS satellite. He compared the data with the prediction by the Gallagher-88 plasmasphere model and finds good agreement.

(6) Related Studies

The new model for the bottomside electron density profile that was developed during earlier TFAs and is part of IRI-2000, was tested with data from Argentine and European ionosonde stations (M. Mosert) and with data from Havana (K. Alazo, B. Lazo) finding overall good agreement. Data from China’s low latitude Digisonde station in Hainan (Dip=22.8) were compared with IRI for a high solar activity time period (2002) (M. Zhang and S. Radicella). Discrepancies were found for hmF2. Ionospheric effects of tropospheric thunderstorms were reported by S. Pulinets. These effects are best seen during the pre-sunrise time period because of the very low ionospheric densities during this time. A validation study of the STROM model in IRI for several storm periods showed an improvement of close to 50% over the IRI model without the STORM model (E. Araujo-Pradere and D. Bilitza).