The surface station sees its own sky and the radiosonde its own column; this lesson adds the three eyes that see everyone's weather: the satellite looking down, the radar looking out from the ground and from the aircraft's own nose, and the aircraft themselves, reporting back what they actually met.
The two satellite families
Weather satellites come in two orbits with complementary talents:
Geostationary satellites — parked about 36,000 km above the equator, matching the earth's rotation, so they keep the same full-disc view continuously: perfect for rapid image sequences and loops, but with a poor, oblique view of the polar regions.
Polar-orbiting satellites — flying low, around 850 km, in a sun-synchronous orbit that crosses each place at the same two local times every day, one daytime pass and one night-time pass. They scan the whole earth in strips at high resolution, each place roughly twice a day: sharper eyes, slower revisit.
What the pictures are used for
For aviation, the satellite's gift is coverage where nothing else observes:
cloud location and extent over oceans and data-sparse regions;
fronts and depressions, tracked as they move;
tropical revolving storms, kept under continuous surveillance;
fog and low stratus, detected over wide areas;
cumulonimbus tops and anvils, monitored as they grow.
Image loops then turn the still pictures into motion.
Visible, infrared, water vapour
Three kinds of picture, three different physical quantities:
Visible imagery — records reflected sunlight: the brighter, the thicker the cloud. It works only by day.
Infrared imagery — records temperature: the brighter (colder), the higher the cloud tops. It works around the clock.
Water-vapour imagery — sees the moisture of the upper troposphere: bright means moist air, dark means dry, sinking air. Cloud is not required.
From image pairs the processors also derive atmospheric motion vectors: winds computed by tracking cloud and moisture elements between successive pictures.