The EM field is used both for the function, and the values of that function. (I think it's actually a skew-symmetric linear operator on the tangent space T_x M at a given point. This can be phrased in terms of a bivector at that point. A "bundle" TM = Union_x T_x M talks about an extended manifold connecting tangent spaces at a different points.) I think it's entirely reasonable in common language to use "where" to mean "where it's non-negligible". Consider that physical objects are also fields. It's entirely reasonable to ask "where an electron is" even though the electron field is a function of spacetime. Once we're able to ask the right questions, this becomes a less-useful question, as it only applicable in cases where the field is concentrated. The EM field case just breaks down much sooner.