Skip to content

Tidal Time Series Analysis in R

2 messages · William H. Asquith, Spencer Graves

William H. Asquith
# 
I am looking at using R to analyze time series data containing a tidal 
component.  I need to remove the tidal signal to extract the time 
series of the phenomena I seek to study.  A browse of R-project search 
engines has not been too fruitful?  I've found 'hoa' and 'Rwave', but 
need further help getting started.  THANKS.

-wa
5 days later
Spencer Graves
# 
I haven't seen any replies, so I will offer a few thoughts:

	  If this were my problem, I'd start using "lm" with terms that predict 
the tide in terms of the positions of the sun and moon.  If other things 
were thought to be important, I'd include them also in the "lm" model. 
However, I would not believe the answer from "lm" until the 
autocorrelation function (acf) of the residuals was plausibly white 
noise.  If for some reason I couldn't get there, I'd use "arima" with 
its "xreg" argument to describe the plausible deterministic effects plus 
some ARMA model identified in the residuals.  For documentation, I 
suggest you start with Venables and Ripley (2002) Modern Applied 
Statistics with S, 4th ed. (Springer), especially ch. 14.  There may be 
better books on R available today, but among the books I've used, this 
is the premier general reference on R, and ch. 14 on time series should 
help you get started.  I'm not familiar with "hoa" and "Rwave".  They 
may ultimately be the best tools for your need, but I don't know that. 
If it were my project, I would likely try "lm" and "arima" as discussed 
in Venables and Ripley first.  Then I'd look at "hoa", "Rwave" and 
anything else that might look promising.

	  Should you desire other help from this listserve, I suggest you 
PLEASE do read the posting guide! 
"www.R-project.org/posting-guide.html".  Anecdotal evidence suggests 
that people who follow more closely that guide tend to get more useful 
replies quicker.

	  hope this helps.
	  spencer graves
William H. Asquith wrote: