Re: Tide article

[brian whatcott <inet@INTELLISYS.NET>]

At 19:04 10/21/01 -0700, John Mallinckrodt wrote:
> ... we see 31
> tidal cycles in 384.150 hours corresponding to 12.39 hours per
> tidal cycle.  This is 2 minutes *less* than half a mean lunar day

> ... at the next spring tide we find a highest high
> of 5.4 ft at 0702 on 15 November.  During this *additional* period
> we find 23 tidal cycles in 286.733 hours corresponding to 12.47
> hours per tidal cycle.  This is not quite 3 minutes *more* than
> half a mean lunar day
...
> Indeed, if we add the two periods we find a total of 54 tidal
> cycles in 670.883 hours corresponding to 12.424 hours per tidal
> cycle, about 10 seconds (!!) more than half a mean lunar day.
...
> Granted they are based on predictions rather than data as it
> sounds like you may have used.  Historical data will show larger
> specific deviations because meteorological effects, most notably
> wind and pressure, can significantly alter the heights and times
> of extreme tides.  But predictions represent a best *fit* to the
> historical data; they wash out the meteorological effects and
> reveal the astronomical influences more clearly.
>
> John Mallinckrodt


I ran a check on this month's NOAA predictions for Myrtle Beach
as John suggested. I am as uneasy as Jim evidently is about
extracting statistics from what are already cyclical models.

Still, it's not unreasonable to mention that a single frequency
sine wave can account for 95% of this month's variance in the
tidal model from NOAA for that location, with a regression value
for the periodic time of 742.15 minutes (12.37 hrs).

In view of Jim's discomfort level, it's easy to suppose that
real world factors like barometric pressure, winds, and storm surge
can really mess up these tidy numbers.


brian whatcott <inet@intellisys.net>  Altus OK
                Eureka!