Discrete Correlation Function/ Bayesian Blocks
I have some time series data sets with gaps, and would like to stitch the data together to see seasonal trends.
I have four sites where I collected underwater noise data and porpoise presence data for ~ 1 week each month for over a year. The noise samples are separated by ~30 minutes, and the porpoise samples are each minute.
I would like to separately analyze these data sets. I have the autocorrelation functions for each trial (site/month), but I do not believe you can indicate separate acf's within one model. I have been reading about discrete correlation functions, which as far as I can tell line up all of the time series at the beginning and compute a multiplicative correlation function which can then be used in a model to remove temporal correlation issues.
I was hoping someone who is familiar with these techniques could help me.
My specific questions are:
1) Am I interpreting DCFs properly?
2) What about time series of different lengths?