I know that CS journals usually accept extended version of a conference paper with some, say 10%, new contents. But is it still true for non-CS journals?
I think it would be a better strategy to submit the extended version to a CS journal and a different approach of the problem with a different point of view (not just 10% extension) to the non-CS journal.
For the reader of the non-CS journal having a non-CS way of expressing the problem/solution would be more beneficial. For the reader that sees both submissions there will not be a feeling of simply trying to get more publications with the same paper (not meaning to be rude or anything!).
Most journals expect at least 30% difference from any previous published version of your work. Some CS journals do not require such a change, or do not enforce it. In other journals this requirement may be strictly enforced or less strictly — this depends on the editor and the specific journal.