How else might the Latin 'torrere' have shifted semantically to mean 'rushing, roaring (of streams)'?

Question

[Etymonline] [...] from Latin torrentem (nominative torrens) "rushing, roaring" (of streams), also "a rushing stream," originally as an adjective "roaring, boiling, burning, parching, hot, inflamed," present participle of torrere "to parch" (see terrain).

Google revealed only one conjecture:

So, what could be the connection between a rushing stream and something which is dried out, thirsting for some water?

Supposedly, the contrasting meanings are linked by the image of a mountain creek or river which may be a raging current in spring, but dries out in summer.

I cannot diagnose why, but the above feels unconvincing, probably because:

1. most raging streams do not dry;

2. it is deceptive to describe a dried raging stream as a 'raging stream', because a stream is dried, if and only if, it is NOT raging. To wit, 'dry' and 'rushing, roaring' are polarly opposite adjectives if water is the underlying semantic notion.

I know that etymology is subjective and speculative, but are there any other conjectures of more conviction and common sense?

Answer 1

Indeed, this conexion between a dried-out creek and a torrent of water seems unlikely. The Proto-Indo-European root supposedly meant "make/be dry". De Vaan simply says that there may have been a stative present verb *trs-eh1- "to be dried out" in Proto-Indo-European, from which a participle torrens "dry" logically followed.¹

The only etymological path that makes any sense to me would be dry → hot → flaming → raging with flames → raging → torrential. But that is quite the semantic distance to cross, and we need intermediate evidence.

---

¹) Michiel de Vaan, Etymological Dictionary of Latin and the other Italic Languages (Leiden 2008), s.v.