I have a need for a comprehensive database of a certain type of event described in the scientific literature. For what it's worth, the event is a 'paleoearthquake', which is a historic or prehistoric earthquake that is found in the geologic record, usually by digging a trench across a fault line and identifying the disturbances in the geologic strata across or adjacent to the fault and, if possible, dating them via radiocarbon or other geochronological methods. However I don't think the specifics are particularly important.

The issue is that these are generally reported in the literature from local investigations of one or two faults, yielding a few events. These studies are done wherever there are earthquakes on land, so we have a global scope and language issues. Even limiting the results to the English peer-reviewed literature, however, it's a huge distributed search.

I estimate that there are on the order of 10,000 published events, and a mean of 2-3 events per publication.

For my immediate use of the database, it is very important for the database to be as complete as possible--I'm not looking for a sort of statistically representative sample. The literature itself is quite incomplete of course, but we're limited to what exists for now.

Starting with the first step of collating publications, what tools would one use? I have access to most journals through various university affiliations. Are there particular APIs? Web scraping tools? LLMs?

Thanks!