Not a nuclear test because there's no initial transient.
This is more like a forced resonance - the geological equivalent of an organ pipe.
So I'd guess there's a pipe-like feature or a chamber in the area (an old magma tunnel? - it doesn't have to be empty, it just needs to have a constant density significantly different to its surroundings) and "noise" from moving magma made it to ring.
With a 17s period and 6km/s velocity, for lambda/4 resonance the pipe/chamber would be around 30km long - which looks not-completely-insanely-wrong, possibly.
As you can see, there are seismic station for detecting underground tests (the signal differs from earthquake signals if you look at stations around the world), acoustic and hydroacoustic stations for detecting tests in the atmosphere and oceans, and isotope detection to differentiate between any kind of explosion and a nuclear one (which leave the same signal in the other detectors).
The stations have a fantastic global coverage and have been measuring for quite a while. The data set is now also given to the scientific community for all sorts of purposes.
