Dr. R. F. Griggs, the man credited with naming Novarupta (and the leader of the 1915 expedition to "study the re-establishment of plant and animal life after the devastating and almost unprecedented volcanic eruptions of Katmai in 1912" as well as five other expeditions in the are), was trained in geology, zoology, and taxonomy. However, he was a botanist.
This choice must have been intentional, if he was following the rules set forth in the International Association for Plant Taxonomy's (IAPT) 1912 Brussels Rules version of The International Code of Nomenclature for algae, fungi, and plants (ICN), as a botanist trained in taxonomy. Unfortunately there is very little information on the naming process for Novarupta itself, even from the late Dr. Griggs. On page 192 of his book The Valley of Ten Thousand Smokes, he simply states:
We did not take time to examine the conditions to the west, for we beheld in the opposite direction a prodigious steam column from a volcano we later named Novarupta.
(The we seems to indicate consensus on the naming among that expedition, which may have been the 1919 one - the largest of the six expeditions, at 19 members. One member was a petrologist and two were topographers, meaning that there would have been the opinion of a geologist on the naming as well as topographers putting names on maps. How much they played a role in the naming is unknown)
These are some relevant quotes from the Brussels Rules:
(Chapter 1, Article 4) 2. to avoid or to reject the use of forms and names which may cause error or ambiguity or throw science into confusion.
Next in importance is the avoidance of all useless creation of names.
Other considerations, such as absolute grammatical correctness, regularity or
euphony of names, more or less prevailing custom, respect for persons, etc., notwithstanding their undeniable importance are relatively accessory.
(Chapter 3, Section 1, § 1) XIII. In the formation of specific names composed of two or several roots and taken from latin or greek, the vowel placed between the two roots becomes a connecting vowel, in latin i, in greek o; thus we write menthifolia, salviifolia, not menthaefolia, salviaefolia. When the second root begins with a vowel and euphony demands, the connecting vowel is eliminated (e. g. calliantha, lepidantha).
The connecting ae is legitimate only when etymology demands (e. g. caricaeformis from Carica, may be retained along with cariciformis from Carex).
(emphasis mine again)
It's also worth noting that the translation is always given as "new eruption", never as "new broken thing" (the above question is unique in that translation), meaning that rupta could come from "eruptio" as was referenced in the first answer - which almost makes the most sense, given the common translation. Adding to that is that 'rupta' is apparently a participle (meaning you're not going to be combining it with another adjective - unless there is a noun version that means 'broken thing' instead of the adjectival 'broken').
So how 'Novarupta'?
In 'novus', '-us' is counted as a connecting vowel, being replaced by an 'i'. However, this would become 'novi' normally (but the numbers and case don't make sense for that). 'Eruptio' becomes 'ruptio', both of which are feminine, apparently (that's beyond my ken, but it also makes sense). But how does that become 'rupta'? It really can't. Which means 'rupta' would just come from 'erupta', meaning something very similar to eruptio. That 'ruptio' and 'rupta' are already words was probably ignored. 'Novus', as an adjective, instead becomes 'nova' to match the gender (Feminine), number (singular), and case (nominative) of 'erupta' (which is likely still referencing 'eruptio').
Novarupta was named by a botanist. Botanical species nomenclature sometimes involves combining two root words. While this may not have been done perfectly, it's likely the basis for Novarupta would be Nova Erupta.