The Trieste Workshop continued the series of meetings organized by the IRI Working Group to advance specific aspects of the model. With this year's workshop the group has begun the difficult task of establishing quantitative descriptions of ionospheric variability. IRI currently provides monthly average values for the undisturbed ionosphere. The intend is to also predict the expected standard deviations and to issue warnings when the occurrence probability for specific ionospheric irregularities (e.g., Spread-F, Sporadic-E) is above certain critical levels.
The Workshop was co-sponsored by COSPAR, URSI, IAGA, ICTP and ISF. 47 participants from 20 countries attended the 42 presentations spread over 4 days and 9 sessions. As Chairman of the URSI/COSPAR Working Group on IRI, I was the main organizer of the Trieste meeting. Excellent support before, during and after the meeting was provided by the Local Organizer, S. Radicella, and his team at ICTP. During the morning of the first workshop day a special session was held in celebration of Karl Rawer's (the father of IRI) 80th birthday. His colleagues, co-workers, and students described Rawer's many contributions to ionospheric physics in particular his leadership in advancing the capabilities of the ionosonde instrument to study ionospheric dynamics and his important role in guiding the development of insitu instruments (the German/US satellites AEROS A and B carried four experiments from his group). I chaired the session and as one of his students I also gave a talk entitled "International Rawer Initiative" about Rawer's involvement in the IRI effort. The laudatio was given by B. Reinisch (U. Mass. at Lowell). On the evening of the second workshop day a special computer session was held to familiarize the participants with the PC version of IRI and related software.
As a result of this workshop several improvements and extensions will be incorporated into the next version of the IRI model (IRI-94). Most importantly, it was decided to open IRI to regional F-peak models. Using a regional model (e.g. for the European or Asian sector) can increase the accuracy considerably. Other new feature coming with the next IRI include: (i) an explicit representation of the auroral oval, (ii) a description of the subauroral temperature enhancement, and (iii) an improved model for the ion composition below 300 km including the ratio of molecular to cluster ions. Another interesting outcome of this workshop is the plan to convene a special task force at ICTP next summer that will focus on much needed improvements of the IRI electron density profile in the bottomside. Experimenters will bring their data and theoretical modellers their software codes. If successful this may be a venue for future improvements of critical parts of the model.
Four teams are currently involved in D-Region modelling for IRI; all four presented status reports at this meeting. Singer's (Germany) effort focuses on the most characteristic D-Region point, NmD, where a sharp change in gradient signals the transition from molecular to cluster ions. He deduced an improved seasonal and solar cycle dependence of NmD for the IRI model based on his ground radiowave propagation measurements on several frequencies. Friedrich (Austria) uses the neutral density as scaling parameter and developed his model with about 80 rocket profiles measured by Faraday rotation/differential absorption and insitu probes. Both authors find discrepancies up to a factor of 10 with the current IRI model below 80 km; agreement is much better at higher altitudes. Comparisons with the results of an aeronomical simulation by Kopp (Switzerland) and his colleagues come to similar conclusions. Danilov (Russia) and his team have used more than 60 rocket profiles to point out the difficulties in modelling the D-Region variability in particular during winter, when the Winter Absorption Anomaly (WAA) and stratospheric influences can lead to a variability range of two orders of magnitude. A collaborative effort between the four teams was initiated in Trieste, which if all goes well should lead to a joint recommendation for the new IRI D-Region model (median plus variability estimates).
Bottomside F Region
Two characteristic points are currently used to define the IRI bottomside profile shape: (A) NmF1 is the electron density of the F1 ledge that is clearly seen in ionograms (F1 cusp), but that is difficult to identify in incoherent scatter radar measurements; the change in gradient at this point seems to be correlated to the transition from atomic to molecular ions; the F1 feature disappears for large solar zenith angles (nighttime) because of the dominant solar control of the F1 region; (B) h0.5 is the height where the density has decreased to half the F2 peak value; IRI actual uses the ratio h0.5/hmF2 often called the G-factor. For both parameters the present IRI applies solar zenith angle dependent formulas developed with ionosonde data; the NmF1 formula depends also on magnetic latitude. Ionosonde data presented at this meeting showed that the F1 parameter modelling needs to be revisited, in particular the latitude dependence and the critical upper limits (of solar zenith angle) for F1 occurrence. The group strongly endorsed a proposal by S. Radicella (ICTP) to invite a team of experts to ICTP for a one-month empirical and theoretical modeling study of the F1 feature. This team would also work towards a quantitative description of ionospheric variability in the E-F region based on the preliminary efforts presented during the workshop (Mosert de Gonzalez, Argentina). Mahajan (India) investigated the diurnal variation of h0.5 with more than 2000 profiles from the incoherent scatter facility in Arecibo, Puerto Rico. His results can be used to extend the G-factor formula to low latitudes; the current formula is based on mid-latitude ionosonde measurements (Gulyaeva, Russia).
Topside and Plasmasphere
Following popular demand the IRI group decided to add the total ionospheric electron content (TEC) as a new IRI output parameter. TEC will be obtained by numerically integrating the IRI electron density profile with a user-specified upper height limit.
Kimura (Japan) presented a diffusive equilibrium model for the plasmaspheric electron density distribution based on observations by the Akebono satellite. If consolidated with the work by Rycroft (U.K.) and colleagues, this new study may lead to a reliable plasmaspheric extension of the IRI model.
Mapping of E and F Peak Parameters
The URSI/VIM chairmen (Reinisch, ULCAR and Anderson, PL) agreed to include the E and F1 peak heights in their Verification of Ionospheric Models (VIM) effort. The currently used simple solar-zenith-angle-dependent formulas for these parameters have never been fully verified with data and leave much room for improvement. Gupta's (India) compilation of insitu rocket measurements could be of help at low latitudes. An accurate global representation of the F peak height, hmF2, remains as one of the most pressing mapping issue. To pursue this goal more direct (incoherent scatter) and indirect (ionosonde) hmF2 measurements are needed on a worldwide basis.
It was decided that future versions of the IRI model should also accommodate regional models in addition to the global CCIR and URSI maps for the F peak parameters. Regionals maps exists for different parts of the globe (e.g., Australia, China, Europe). They have proven to represent the regional ionosphere much better than the global models. As a test case the Chinese regional maps will be incorporated into IRI (Dai, China/ICTP). A good candidate are also the PRIME maps for Europe (Mikailov, Russia); PRIME is the Parameterized Regional Ionosphere Model for Europe project (P. Bradley, Chair). Through regional models it may be also easier to introduce storm effects into IRI, since the ionospheric documentation of the different storm phases may vary from region to region. PRIME, for example, intends to publish maps for the different storm phases which could than be readily implemented into IRI.
Accepting my suggestion, the group decided that the next version of the IRI model will include a representation of the auroral oval. As a start this will the Holzworth parametrization of the Feldstein ovals for different magnetic activity.
A number of papers dealt with the day-to-day variability of F peak parameters, which can be as high as 25% and more. Gravity wave activity (20 to 120 min) and magnetic storm effects are the major causes for this deviations from a monthly median value. A better understanding and possibly modeling of expected standard deviations will require considerable effort by the IRI group and related URSI, COSPAR and IAGA groups.
Oyama (Japan) compared his Hinotori electron temperature probe measurements with IRI. He finds that at the Hintori altitude of 600 km IRI does not include the early morning and late afternoon peaks seen in the data. IRI in this altitude range is based on AEROS data which only covered the noon and midnight time periods. Substitution of the current global maps at 600 km with maps based on the Hinotori data base is now planned for the next edition of the IRI model.
An improved model for the ion composition below 300 km was presented by Danilov (Russia); this Danilov&Smirnova model will replace the older Danilov&Semenov model implemented in IRI-90. Use of the new model leads to more reliable transition heights (molecular to atomic ions) which is one of the most important parameter in the bottomside ionosphere.
An evaluation of the different formulas for the ratio between molecular and cluster ions was identified as one of the most important tasks in modeling the D-region ion composition for IRI. The current IRI formula is based on Danilov's (Russia) assessment of a large number of rocket measurements. Kopp (Switzerland) has presented a formula based on his aeronomical computations. Friedrich and Torkar (Austria) have established a correlation between the neutral density and temperature at the transition height from molecular to cluster ions. Of interest for future IRI work is also the suggestion by Danilov (Russia) to represent the ratio between the two dominant cluster classes: proton-hydrated and non-proton-hydrated cluster ions.
Membership and Meetings
We mourn the death of our dear colleague C. Serafimov, who passed away earlier this year. Cyril was with the IRI group almost from the very beginning and has been very effective in promoting the IRI goals. He will be sadly missed by all of us. Four new members were accepted into the Working Group: (1) B. Ward, DSTO, Salisbury, Australia; (2) K.K. Mahajan, NPL, New Dehli, India; (3) W. Hoegy, NASA, GSFC, USA; (4) M. Mosert de Gonzalez, Universidad Nacional de Tucuman, Argentina. Longtime member L. Brace decided to retire from the group.
K. Oyama has proposed to produce and distribute a IRI Newsletter at ISAS, Japan. The Working Group supports and endorses this activity as a way to keep the IRI user and producer community informed about current developments and future plans, and to promote the IRI activities in general. The presentations from the IRI sessions during the World Space Congress in Washington, DC, last summer have been submitted to Pergamon Press for publication in an upcoming volume of Advance in Space Research. A.P. Mitra and K.K. Mahajan have invited the group to hold the 1995 IRI meeting at NPL in New Dehli, India most likely in early January; the topic will be the low-latitude and equatorial regions in IRI.