I'm not sure why you think they are enantiomers. They are actuallually diastereomers as stated:
As depicted, the compound to the left is $(2R,3S)$- while the compound to the right is $(2R,3R)$-. Thus, they are pair of diastereomers.
To be enantiomers, either the compound to the left should be $(2S,3S)$- or the compound to the right should be $(2S,3R)$-.
Let me speak of geometrical isomerism as such the discussion is more general.
Geometrical isomerism of the cis-trans / E-Z type can happen about a plane containing two atoms each with two different substituents (though in cyclic compounds the latter requisite is merely formal). In fact the two identical substituents or the pair with high priority can be at ...
The (cis)/(trans) nomenclature (or, more universally applicable, (E)/(Z) nomenclature) works fine for isolated and conjugated double bonds. In the case of 2,3-pentadiene, however, where double bonds literally are adjacent to each other, you have a cumulene. As a result, the mutual orientation of two terminal methyl groups is close to perpendicular.
To put ...