In an update posted on Sunday (Feb. 18), ESA said that the rentry ERS-2 is expected to take place on Wednesday (Feb. 21) at 10:19 a.m. ET (1519 GMT), plus or minus around 19 hours. This uncertainty is due to the "influence of unpredictable solar activity, which affects the density of Earth's atmosphere" and can therefore change how much drag pulls on the satellite on its way down, ESA wrote.
Plus or minus 19hrs due to the sun's effect on the density of the atmosphere. Mind blown.
I don't know all of the details of this mission, but it seems like they've just lowered the lowest point in its orbit - called periapsis - until it sits low enough in the atmosphere to get enough drag that the orbit slowly decays over a decade.
The lowest part of the orbit would only drop a little bit, but the highest part of the orbit woukd reduce more with each orbit. If you do it slowly enough, the orbit would circularise and then it would begin to decay more evenly. As it falls deeper into the atmosphere the orbit would decay faster and faster until it can no longer sustain orbit, and then it falls deeper into the atmosphere and burns up in just a few minutes.
The reason for this I can only guess at - it wouldn't take a whole lot more fuel to just deorbit all at once. My best guess is that it has something to do with reentering at the lowest possible speed. If you fall from a high orbit and reenter, you have a lot more speed and have to dissipate more energy all at once. It's possible this increases the risk that the satellite will fail to deobrit, and break up and send pieces off in less predictable orbits. If it breaks up from a low circular orbit, there's no chance of any parts escaping back into orbit.
