A measurement of frequency against time produces a Doppler curve. As the satellite passes, the received frequency appears to fall but not in a constant manner. The rate of change starts off slow, is greatest at the time of closest approach and then tails off towards the end of the transit. This is because we are measuring the rate of change in the component of a satellite's velocity along a line joining the satellite to the receiver. The rate of change is greatest at the time of closest approach and, if measured, can be used to determine passing distance.

Below is a Doppler curve produced from the GFO pass around 1600 GMT on 2000 December 8. THis is an early attempt at producing a curve - from before the days of plentiful availability of the hardware and software that allowed frequency data to be harvested and plotted directly. It was made by reading-off the frequency from a radio receiver at intervals and then entering the results into a spreadsheet.

In the space of seven minutes between 1558 and 1605, the received frequency drops from 400.041 MHz to 400.024 MHz. The closest approach computes as 873 kilometres against an orbit height of 790 kilometers, so the pass is nearly overhead. Analysis of a prediction based on Mike McCants' element set produces 840 kilometres but given that the margin of error in the Doppler calculation is 50 kilometres, then they agree.

Doppler measurements of the previous and subsequent passes both produced a closest approach of 1850 kilometres.

With the advent of Software Derived Radios, Doppler measurement has become somwhat easier. First, a standard display immediately shows the shape of the Doppler curve. Below is a classic curve from an satellite in LEO, in this case - Progress M-63.

Software also allows measurements of actual frequency to be extracted from the receiver, somewhat simplifying Doppler calculations.