Year 2000 Problem for TLEs??

Bruno Tilgner (Bruno_Tilgner@compuserve.com)
Thu, 21 Aug 1997 18:59:32 -0400

Robert Sheaffer wrote:

>In other words: will all our satellite prediction programs come to a
>screeching halt on Jan. 1, 2000, or will we be OK? (And if we will
>be OK, for HOW LONG will we be OK? Maybe the catastrophe is postponed to=

>the year 2057, the 100th anniversary of Sputnik, or some such date?)


I would expect they are ok. The NORAD model calculates satellite
position and velocity as a function of a time variable which is the
difference in minutes between time T and the time of the TLE epoch.
The date does not enter explicitly into the calculations.

But there could be a problem in the calculation of that time variable
just around the turn of the century. 00.xxx - 99.yyy would yield a
very large number when expressed in minutes which presumably crashes
most prediction programs anyway because of arithmetic overflow. Since
this number would be negative, all those programs that refuse to
calculate satellite positions for a point in time before the TLE
epoch (although there is no justification for this behaviour in the
NORAD model) would simply stop.


Bruno Tilgner
48.85N 2.02E