By
George Lane
Lane Consultant

The Ionospheric Communications Analysis and Prediction (IONCAP) program was created by the Institute for Telecommunication Sciences (ITS) for the United States Army beginning in 1975. IONCAP was "completed" in 1985 and consisted of a main-frame Fortran 77 program and a user's manual. This program was released with unlimited distribution by the National Technical Information Service at a cost of about $1,000.00. Subsequently, a number of IONCAP type of programs have been developed and distributed. The two most widely used programs in this family are: VOACAP by the Voice of America and ICEPAC by ITS/NSA. Neither of these programs are currently receiving funding for maintenance or improvement by the developing agencies. Nor does the US Army support IONCAP now that function of propagation modeling has been transferred to the US Air Force.

The Voice of America (VOA) invested over $2M in converting IONCAP to an area prediction model (1985 to 1996). In this process, they found and corrected approximately 60 errors in the original code. The VOA used the unpublished IONCAP theory manual as the guide in correcting logic and programming errors. They documented and tested all changes to the code. Considerable effort was given to improving the input and output functions for ease of use on a PC. The last major change made to VOACAP by VOA was to add a signal-to-interference prediction capability.

The NSA funded ITS to develop a PC version of IONCAP that would be useful for circuits at high latitudes. This program became ICEPAC. Unfortunately, this effort was shrouded in mystery as to exactly what changes were made to the IONCAP code. ICEPAC exists in the public sector only in compiled format so that the source code cannot be examined.

ITS has kept the input and output formats of ICEPAC and VOACAP as similar as possible. However, the two codes are quite dissimilar in terms of their subroutines. It is unclear if ICEPAC contains any or all of the errors VOA found in IONCAP. To my knowledge no one has made an extensive comparison of these two programs. Only VOACAP has been compared to measured data and those comparisons were very limited in size. The agreement between prediction and actual listener scoring was remarkably good in the absence of interference. The prevalence of interference in the international short-wave bands, however, prompted VOA to develop the interference prediction capability in VOACAP.

The only comparison of ICEPAC to measurement showed that the high latitude absorption model was inadequate. Another investigator in the Navy commented that the ICEPAC absorption model at high latitudes appears to identical to the values used in original IONCAP and VOACAP. It is truely unfortunate that ICEPAC source code is not available for inspection! The Q index function in ICEPAC developed for use by the USAF did not seem to correlate with the observations made by the VOA monitors during comparison studies made by VOA.

In my opinion, since both VOA and NSA are no longer maintaining these programs, VOACAP and ICEPAC should be consolidated into one program. This would necessitate NSA releasing the source code for ICEPAC so that if any usable or desirable subroutines exist in ICEPAC , they could be incorporated in VOACAP. The resultant code should be renamed: IONCAP. The VOA documentation of VOACAP should be consolidated and used as the program documentation for the new IONCAP.

The above suggestion raises the issue as to why the US Government should want to retain and maintain this code. Firstly, HF radio will always be needed in times of emergency or war. HF radio operation requires engineering design based on propagation predictions and the link maintenance requires knowledge of the diurnal and seasonal changes of the ionosphere. The US Government has invested a tremendous amount of money and manpower in the development of a full system performance HF radio prediction capability beginning with the fledgling efforts of the Office of the Chief Signal Officer in the War Department in 1942. No other program has proven to be more reliable or more useful than the US Government models. In the 30 years I have been using these models (ITSA-1, ITS-78, HFMUFEs, IONCAP and VOACAP), I have found that if the predicted circuit reliability is at least 90%, then the observed performance is that good or better. It would be a national shame to lose this work and its capability.

The next issue is: "If these IONCAP programs are so good, why must they be maintained?" The reason is that the world has changed since the 1970's. It use to be that IONCAP was run on a mainframe computer by people skilled in the art of making HF predictions. Now that the program is available for use on a PC, the user could be almost anyone. Further, the skills of operating and maintaining an ionospheric radio channel are no longer taught in the Signal Schools. Although either ICEPAC or VOACAP are relatively easy to use, the knowledge needed to make wise decisions is most often lacking. The documentation for using these programs is non-existent. Worse yet, the people who could write this documentation are growing old or have died. The time to retain this knowledge is running out rapidly.

If this national capability is to be maintained, what needs to be done and in what priority? First, there needs to be a Program Manager who looks at the users' requirements and prioritizes / assigns work efforts. Secondly, there needs to be a centralized point where the program can be run or the code obtained in its latest version. Thirdly, there needs to be a point of contact who will respond to user's questions. Finally, there needs to be a panel of experts who give guidance on model development to the Program Manager and answer the more complex of the users' questions.

There are a number of areas where IONCAP technology needs to be brought into the 21st century. I am listing some that come to mind.

1. Boundary conditions should be put on the input variables so that the user is warned when values outside of the normal range are being used and a explanation should be available as a help function in determining an appropriate value. Further, warning flags should be placed in the output when prediction values are suspect or if the user has input an unusual input parameter. Each input and output parameter in IONCAP should be defined in easy to understand terms and these definitions should be included in the Help function.

2. Required signal-to-noise ratios need to be developed for use in the program. Existing values were established in 1969 for equipment that is now obsolete. The same applies to Signal-to-Interference ratios.

3. Guidelines need to be developed for using the predictions when planning for Automatic Link Establishment (ALE) systems or systems with sophisticated ARQ capabilities. This may require changing the existing code to include the consideration of multipath conditions in more appropriate manner than the existing multipath probability prediction.

4. Capability of using Numerical Electromagnetic Code (NEC) generated antenna patterns or other radiation pattern sources should be improved.

5. An altitude dependent model of RF noise should be developed for use when airborne platforms are being modeled.

6. An approved and recognized groundwave model should be available in IONCAP. This model should include the capability of predicting groundwave, sky wave propagation modes, the probability of groundwave/sky wave interference and associated skip zones.

7. The long path prediction model for paths in excess of 7500 km is suspect and should be reviewed for possible improvement.

8. A theory manual should be written for the new IONCAP.

The above opinions are mine but they are in the public domain and may be borrowed or used by anyone who thinks he or she can salvage something of this effort over the past 56 years.