There are 2 operational INMARSAT satellites that were within view of MH370 at various times. A 3rd satellite, INMARSAT-5 F1 was launched in December 2013, but is still undergoing on-orbit tests until mid-July 2014, per Slide 18 of:
http://www.inmarsat.com/wp-content/u...ation_2013.pdf
Nit-picking a bit, the satellites are geosynchronous, not quite geostationary. Hence their latitude and longitude will vary slightly over the day, typically tracing a figure 8 on the ground. See Figure 3 in this page by T.S. Kelso:
CelesTrak: "Basics of the Geostationary Orbit"
Note that the figure 8 is for a satellite whose orbit has eccentricity of zero, whereas non-zero eccentricity distorts the figure 8 somewhat.
Kelso lists orbital elements in TLE (two line element) format here:
http://celestrak.com/NORAD/elements/geo.txt
This is the active one for the Indian Ocean Region:
INMARSAT 3-F1
1 23839U 96020A 14077.02691924 .00000003 00000-0 10000-3 0 2792
2 23839 1.6580 73.1420 0005562 281.9751 254.5647 1.00273339 65760
Element 2 in Line 2 is the orbital inclination in degrees, in this case 1.658 deg. Thus both the latitude and longitude will vary a bit over a day, tracing a sort of figure 8
Element 4 in Line 2 is the eccentricity, with an implied decimal point at the left. The eccentricity is 0.0005562 for this satellite.
The N2YO website uses the TLE data to compute present position of the satellite:
LIVE REAL TIME SATELLITE TRACKING AND PREDICTIONS: INMARSAT 5-F1
The Pacific Ocean Region satellite is INMARSAT-4 F1.
Kelso's TLE data:
INMARSAT 4-F1
1 28628U 05009A 14077.46405487 -.00000269 00000-0 10000-3 0 3195
2 28628 2.6638 355.1900 0002955 355.6954 135.5891 1.00271237 32824
Inclination is 2.6638 deg, eccentricity is 0.0002955.
Tracking data from N2YO:
LIVE REAL TIME SATELLITE TRACKING AND PREDICTIONS: INMARSAT 4-F1
The latitude will range between plus and minus the inclination angle over one day.
Will attempt to calculate and post the latitude and longitude of both active satellites for the hourly pings.
The search agencies have certainly done this already. Which may be why they have plotted the circle for only the last ping, so they don't get bogged down in explaining to a low tech press pool why the circles aren't concentric.