In most scientific fields it is ususual to have a separate table of variables/nomenclature for a paper but it is common in books. The reason is that books are much longer, and in this case it is useful for the reader to be able to find all the important variables/special terms in the work defined in one place.
You may have noticed that when you consult a large scientific book (e.g., in mathematics) and go to a later section, it can be difficult to track down the meaning of the variables from where they are first used in the body. Having a table of variables assists with this, so in long books it is fairly common. (Similarly, long scientific books often have a list of figures, and index, etc.
A table of variables/nomenclature might be useful in a paper if it is long and has a large amount of notation/special terms. However, most papers are sufficiently short and succinct that such a table would be unecessary. You are right that this is usually too much. Even if it is necessary to include such a table, it would be usual to define the variables as you go through the material as normal, but then have a table to summarise and collect this information in addition to defining things as you go in the body.
Some fields/journals may have a special style where such a table is expected. You should be able to determine this by browsing some papers in the journal of interest, and/or reading their style guide. Unless there is a particular expectation from the journal, always apply the Golden Rule of academic writing --- what would be useful to you in the paper if you were the reader?