But this paper specifically rules out legalization as cause because the numbers didn’t significantly increase after legalization.

How does this rule anything out? It is totally possible actual usage of THC didn't increase after legalization. It wasn't one of the hardest things to find when it was illegal.

I think you’re misunderstanding. The paper says that the rates of dead drivers having THC in their system above the threshold did not change significantly after legalization. So legalizing isn’t the culprit, exactly for the reason you cite: people could get it just fine before.

Legalization does not signify usage.