Spatial codecs (also known as intraframe schemes) deal with each frame independently. That is to say, they don't apply calculations across a set of frames, even when much of the information is repeated.

Temporal codecs (also known as interframe, of which H.264 is the most popular example) achieve vastly lower file sizes by applying samples from one image across a set (aka a 'group of pictures, or GOP). This is fantastically useful for broadcast and streaming, since you get high visual quality in an efficient package. But if you try to DO anything with that video (reformatting, enlarging, compositing, etc.) the downside will become very clear very quickly.

Typically, professional shows are going to demand a lot of flexibility in post, meaning they're likely to insist on spatial codecs whenever possible.