We tend to assume that when (optimistically) we finally locate a planet that looks and smells an awful lot like the Earth, it will be in a 'normal' state - not in the throes of global change. This could be a very dangerous assumption - 'normal' may actually be a state of flux. This is a drum that I'm fond of beating; Earth today is just a snapshot of a bewildering array of possible conditions.
A paper by Napier a few weeks ago raises the possibility of an astronomical explanation for a fascinating period of terrestrial global change some 12,000 years ago - known as the Younger Dryas cooling (the Figure here shows this, taken from the GRID-Arendal website). He sets forth an intriguing case for the large scale impact of debris from a still dispersing comet - now known as the Taurid Complex - across much of the northern hemisphere at this time. These impacts could have ignited surface forestation, consistent with evidence of clear carbon deposits in the geological record, among other pointers. Subsequent atmospheric soot, presumably along with cometary dust, could have then helped cool much of the planet by 5-10 Celsius for the next thousand years or so.
What's interesting here is less the mechanism of the cooling - most theories involve not comets but the shutdown of major ocean thermal transport loops in the Atlantic - but the fact that the planet could experience such an abrupt climate change without a colossal smoking gun. Whole biota got bumped and reshuffled, several dozen species of large mammals are seen to vanish during this episode. It has even been suggested that this could have helped push humans to an agriculture based society - fewer juicy bison to hunt, altered local climates, and dramatic changes of forests into tundra.
Suppose an unsuspecting observer chanced across the Earth in the midst of the Younger Dryas. They would still deduce the presence of a biosphere (from atmospheric chemistry among other things), they would measure the global temperature, perhaps even the extent of photosynthetic life and pigmentation. They would then - if like us - construct an elaborate model of this planet, tweaking it to match the observations, and then categorizing it. The problem is that if they had happened to looked a few thousand years earlier or later then their model would be very different. The Younger Dryas was a mild fluctuation, but what about the big ones, like the mass extinction of perhaps 90% of surface life 250 million years ago? Snapshots of planets carry an inherent sampling error, it's not clear to me how we allow for that in our models, but we should definitely be aware of the problem.