digital video terminology.
It's easy to see what's the easiest way of reducing the video stream's size: by simply compressing each frame, you can already get pretty small video files. This is what video formats like AVI and Quicktime do when using most of the codecs. An example of this sort of codec would be MJPEG.
Temporal compression goes somewhat further: the compressor takes into account the differences between frames, that is, the changes that come to the picture over time (that's where the "temporal" comes to the name). If some part of the picture hasn't been updated (for example, the camera stays still, and the subject moves over a static background), that can be compressed away.
One of the codecs/video formats that do this sort of thing is MPEG. MPEG only stores keyframes (or "I-frames", "intra-frames") in full every n frames, filling the remainder with "P-frames" ("predicted frames") and "B-frames" ("Bidirectional frames") to store temporal changes.