Good observation John about Celestrak use of "epoch date"! I also did some reading through the details on celestrak.com and "epoch date" is the correct terminology and the time fraction after "eppoch date begins at 00:00:00 UTC. This clears up the confusion of "epoch" day" vs "Julian day" in my post above. Celestrak also specifically states that "epoch time format" the day number changes at midnight and not at noon as with Julian Day used for centuries in astronomical time calculations: http://celestrak.com/columns/v04n03/#FAQ02 Thus I believe it is NASA that has mislabeled their explanation of how to read TLE data on their web page by calling this a "Julian Day Fraction": http://spaceflight.nasa.gov/realdata/sightings/SSapplications/Post/JavaSSOP/SSOP_Help/tle_def.html Whoever prepared this explanation has everything correct and even their example of how to convert the decimal day fraction to h , m, s is correct. They should have used "epoch day fraction" instead of Julian Day fraction which changes day numbers at 12:00:00 UTC. Quote from NASA page which is not correct terminology: This is a surprising error for NASA which has plenty of of people who know better, but apparently not the person who prepared the web page graphic.:nono2: Edit comment: This orbital stuff is hard enough for a beginner like me to try to learn without having NASA mislabel their explanation! BTW astronomers use this Julian Day number system since almost all data is taken at night. Thus dating an observation takes place on a single "Julian day" number and does not divide the data into two different day numbers by changing day numbers at midnight, however it still leads to confusion in time zones far away from Greenwich where midnight UT may be in the daytime locally.