How did DSCOVR capture the Moon's shadow on Earth during the August 21 2017 solar eclipse without the Moon itself also being visible?

Question

In this gif you can clearly see the Moon's shadow moving from the West Coast of the U.S. towards the East Coast during the August 21 2017 eclipse.

DSCOVR (the satellite that imaged this) is located at the L1 point, which is exactly between the Earth and the Sun, meaning it always sees the sun-lit part of Earth.

But for a Solar eclipse to happen, the Moon also has to be exactly between the Earth and the Sun, and thus it should be visible in this gif.

At first I thought it was because DSCOVR was closer to Earth than the Moon was, but it's actually more than 3 times farther away.

Geometry dictates the Moon should be between the Earth and DSCOVR since DSCOVR itself is between the Earth and the Sun, so why isn't the Moon visible?

Answer 1

DSCOVR is in a large Lissajous orbit around the Sun-Earth L1 point, about 1,500,000 kilometers from the Earth. At this point it is far from the line drawn between the Sun and Earth; if the Sun were behind you and you were looking at the Earth-Moon-DSCOVR system, you would say that the spacecraft is down and to the right by several hundred thousand kilometers.

I've used the Python package Skyfield plus a location for DSCOVR from JPL's Horizons to plot their positions. Earth is large and blue, the Moon is green (cheese) and smaller, and the very tiny black spot is DSCOVR. In each case I've annotated DSCOVR's position.

I've done some rotations so that my X-axis passes through the centers of the Sun and the Earth during the eclipse.

If I get a chance I'll make a 3D plot of DSCOVR's orbit. The script I used can be found in https://pastebin.com/AXsJM8J1