Identify (based on literature / analysis) typical length (or distribution) of the length of the disturbances in the signal
vc1492a opened this issue · 3 comments
In order to set the proper parameters for nonparametric dynamic thresholding (with a slight modification on our part), we will want to know (based on previous work in literature / known science) what is the typical length of a disturbance in the signal from tsunami waves (e.g. 20 minutes, 30 minutes, 1 hour, etc.).
If we know the distribution of this, we can do an even more in depth analysis as well.
@MichelaRavanelli do you have any thoughts / knowledge you can share and / or could you do some background literature on this?
I'd like the specification of window length and other smoothing parameters used in the classification of anomalous time periods based on the model error to be perhaps rooted in knowledge already known to scientists in this context. Let me know if that makes sense!
Theoretically, periods periods for the tsunami wave in the ionosphere are similar to ones of the tsunami ocean waves, which can range from 5 min up to an hour with the typical deep ocean period of only 10–30 wavelengths around 400 km, and the velocity approximately 200 m/s (Peltier and Hines).
Generally, literature reports typical periods for the tsunami wave in the ionosphere in the range from 10 minutes to 30 minutes (Makela, Grawe).
Got it, thank you! 👍 will use a range between 10 and 30 minutes in our experiments.