A discrepancy in the International Radio Consultative Committee report 322-3 radio noise model: The probable cause


  • David B. Sailors


The cause of a discrepancy in the CCIR Report 322-3 (International Radio Consultative Committee (CCIR), 1988) radio noise model is reported. The basis for this discrepancy results from the procedure used to prepare the measured noise data for the determination of a global numerical representation of the 1-MHz data. In the development of the model, correction factors to an older CCIR model were determined for each measurement site. These corrections were interpolated for each time block and season to a grid of 100 latitude points by 84 longitude points. The correction factors at each grid point were then added to corresponding values for the older CCIR model, and finally, the resulting data for each time block and season were numerically mapped. Nineteen locations were used in the final model. Four sites used in the original CCIR model were not used. As either no correction factors were obtained for these locations or a correction factor of zero was used to maintain the status quo, the interpolation algorithm produced erroneous values near these four sites in determining the 100-by-84 grid of correction factors. For Bill, Wyoming, the result is not too serious; but for the other three sites at some seasons and time of day, the error is serious. For Thule, Greenland; for Ibadan, Nigeria; and for Byrd Station, Antarctica; the maximum and minimum errors were 10.1 and −10.8 dB; 12.5 and −1.5 dB; and 12.0 and 3.0 dB, respectively. Examination of the geographical extent of these errors reveals that the error is not confined to the measurement location but in fact can be very large. This geographical enhancement of the error is most serious during June, July, and August and least serious during December, January, and February. The error as a function of frequency was found to be diurnally dependent, being more serious during the daytime hours. The absence of the data locations as nodes in the interpolation affected the accuracy of the interpolation itself. The CCIR Report 322-3 atmospheric noise model should be used with caution, especially for locations in the northern and southern high latitudes, the Arabian Peninsula, northern Africa, and the Mid-Atlantic Ocean areas.