The gain is precision. A term written in this scheme can be looked up, sorted, searched and cited without ambiguity, and a reader who knows Tibetan can reconstruct the spelling exactly. The cost is that the result is unreadable to anyone who does not, because the silent letters are all there: the bodhisattva of compassion is written spyan ras gzigs and is pronounced roughly Chenrezig, and no amount of goodwill will get an English reader from one to the other. Scholarship therefore runs two systems in parallel, a strict transliteration for accuracy and a loose phonetic rendering for reading aloud, and while the first is standardised the second is not, so the same name appears in half a dozen spellings across otherwise careful books.
This atlas inherits the problem and handles it the same way, giving a readable form and noting the transliteration where precision matters. Wylie other work, on Tibetan geography and on the relationship between Tibet and the Mongol court, belongs to the ordinary business of a field, and it is the two page article that turned out to be the durable contribution, which is a common enough pattern and worth recording without irony.
There is a further consequence for anyone searching a reference work. Because the strict form and the readable form share almost no letters at the start of a word, a reader who knows a name in one cannot find it in the other, and an index built on only one of them is unusable to half its audience. That is a practical reason for a reference work to record both, and it is why an atlas entry that gives a Tibetan term should say which system it is using rather than leaving the reader to guess from the shape of the letters.