Tides

[Michael Rogers]

Funny things, tides.

As a retired lay-about, apart from the (considerable) constraints of domestic and other commitments, to some extent I can plan my sailing days around such variables as wind direction and tide times. Thus, while looking ahead last week I came across the following tide times for Studland (Knoll Beach), which is where I keep my boat, for the week beginning Monday 22nd May. I don't think I've seen the like before -

                          High                Low                 High
Mon 22               0910               1351                2113
Tues 23              0914               1437                2114
Wed 24              0916               1522                 2120
Thurs 25            0926               1606                 2136

That's four high tides on consecutive days within 16 minutes of each other, and three of the four following highs 12 hours later as tightly grouped; while the intervening lows 'progress' as one might expect.

I also have, in rather small scale graph form, information for tides at Poole (Town Quay). This is at the top of Poole Harbour, where tides are presumably later than those at or near the entrance (e.g. at Studland). These seem to indicate that the corresponding highs there were expected at about 0620 (22), 0700 (23), 0830 (24), and 0930 (25). These 'progress' as one would expect: all but one of them is EARLIER than those I have for Studland. (It's New Moon on 25th, I don't know if that's relevant.)

I think it's amazing, and to me incomprehensible, that clever people can calculate tide times, usually with astonishing accuracy, ages (years?) in advance. But what would lead to this sort of 'grouping' of tide times? Local factors? - if so, what? Is this unusual? And why apparently out of synch with other information about the area?

BIG subject!! Over to you, Peter Taylor (and Julian/ - who also knows those waters): and others, of course.

Michael Rogers

[Peter Taylor]

Hi Michael, yes local factors! I can't exactly explain your tide times for Studland but I can demonstrate what gives rise to them. In what follows you only really need to read the last paragraph!

To calculate tide times you have to take into account astronomical factors such as the orbits of the moon and earth around the sun and the rate the earth spins and also geographical factors such as the depth of water, shape of the coastline etc. This is done using "Tidal constituents" of which the largest are the semidiurnal effects of the moon and sun, a further semidiurnal lunar term, and two lunar diurnal terms. That's 6 constituents, but to get a reasonably accurate prediction you need to consider around 60 or more (to be really accurate,hundreds of them)!

The importance of each constituent varies from place to place, so for any one place each each constituent has a scaling factor and a time lag applied (known as "Tidal constants"). So to forecast a tide for a place and time period you calculate the relative positions of the moon, sun, and earth, then calculate each of the tidal constituents and scale them using the tidal constants for that place, then add them together to get the tide height.

Tidal forecasts for a particular place may differ because they take into account a different number of tidal constituents or they use a different set of tidal constants. The latter are well known for places where tides have long been measured ("Primary Ports" such as Newlyn, Portsmouth, etc.) but have to be interpolated elsewhere.  In all, this a a major computing task requiring a powerful computer (or, in the past, a machine with a very complicated set of wheels and pulleys). Tidal predictions run on smaller computers will cut corners and be less accurate!

When you've done all that you get a tidal curve like the one attached which is for Poole Harbour in the period before and during the time you have highlighted.

Now, if you want to produce a tide table containing high and low water times you have to devise a way of picking off the heights and times at which the water level is at a maximum or minimum. This becomes more difficult if your tide table only has space for two high tides and two low tides a day and the area is one like the Dorset to Portsmouth region where the period of high water contains more than one maximum.

Thus in the figure the low water time is relatively easy to define. But there are two possible high water times and, during the period of neaps the second high water is sometimes the greater whereas sometimes it is the first high water. On the morning of the 21st the first high water is higher but in the evening it's the second one.

I can only assume that, for the tidal predictions for Studland which you quote, at the start the second high water was the highest and, as the days progress, the highest level transfers from the second to the first high water. Thus the time of high water does not move ahead in the way that it normally does!

Cheers,
Peter

[Michael Rogers]

Thanks, Peter, yet again for explaining something complex so clearly. This, together with your manifest enthusiasm(s), would have made you a brilliant science teacher - perhaps you were, in years now past? Anyway, you are a big asset to this forum, and I for one am very grateful!

I understood most of it, and I was aware it was complicated. Did all that calculating really go on, at least to any useful effect, in the days before computers? - amazing!

Just one or two relatively trivial points. The tide times I listed came from the Met Office detailed forecast: I don't know where they get them from. I take your point that those with the 'biggest' computers are likely to have the best forecasts, and vice versa.

Your suggestion that low water times (perhaps especially along the 'two highs' south coast) are easier to define is a useful practical point.

With regard to the two highs phenomenon, the graphs I referred to are in the 'Tidal Planner', produced by Crest Publications, for Poole (Town Quay) 2017. The tide curves for the days in question (May 22 - 25) all show the 'first high' as higher than any later surge. So, for inside Poole Harbour at least, the suggestion in your last paragraph doesn't really stand up. And therefore I STILL don't understand why those highs at Studland stacked up the way they did!

I did actually go sailing on Monday 22nd. I got afloat after the predicted morning high, and sailed into Poole Harbour on a dead run past Sandbanks with a southerly breeze. This was wind against a fairish ebb which produced interesting sea conditions - 1 metre waves to surf down even with a mere upper F3 breeze. The critical thing there is enough speed over the ground to get past the chain ferry, which has a very loud horn it uses to make small craft feel even smaller if it thinks they might impede its imperious 200 metre voyages. We made it! Then, as mentioned elsewhere, I had fun chasing BC26 006 for a while. Then, visiting the waters south of Brownsea Island, I was reminded of what I already knew - that huge areas of sparklingly inviting water there are, at lowish tide, covered by about 9 inches of water, and that at least some of the stakes marking channels are nowhere near the channels in question. So I anchored and had a snooze, which was very pleasant and lasted long enough for me to find the tide flooding nicely when I came to. Then it was out past the ferry again, this time against both wind and tide - flattish, gurgly-type water, and bow waves round all the channel buoys. It took quite a while, but we made it, again without incurring the ire of the ferry skipper. A long beat up along the Training Bank (which seemed extra long that afternoon), and round the end beacon into Studland Bay, and home. To borrow from a well-known personality, 'SUCH fun!!"

[Peter Taylor]

The tide times I listed came from the Met Office detailed forecast: I don't know where they get them from. I take your point that those with the 'biggest' computers are likely to have the best forecasts, and vice versa.
...
With regard to the two highs phenomenon, the graphs I referred to are in the 'Tidal Planner', produced by Crest Publications, for Poole (Town Quay) 2017. The tide curves for the days in question (May 22 - 25) all show the 'first high' as higher than any later surge. So, for inside Poole Harbour at least, the suggestion in your last paragraph doesn't really stand up. And therefore I STILL don't understand why those highs at Studland stacked up the way they did!

Well, the Met Office certainly have powerful enough computers. It is a somewhat surprising fact that the development of the world's most powerful computers has been prompted by the needs of weather and climate forecasting and also for designing nuclear bombs. At at least one establishment in the US (Lawrence Livermore Labs) the same computer complex is used for both!

Indeed, as mentioned elsewhere on the forum, the Met Office is responsible for running a numerical model of the area around the UK which calculates both the astronomical tidal predictions and also how they will be modified by the forecast weather conditions which, in extreme circumstances, might produce a "storm surge" and flooding.

For the curious Studland tide times my guess is as follows.  To produce a "detailed forecast" for places all around the UK the Met Office web site has to interpolate tidal forecasts for places where the tidal constants are known to the rest of the coastline. There are different ways of doing that, I don't know what they choose. However at some point in the process they will have a tidal curve similar to that which I attached to my last post, my example was for somewhere in Poole Harbour, we don't know what their Studland curve looked like.

They also have to use a method to reduce the double peaked high tide curve to a single time and height. Most mariners would choose the highest peak. However I suspect the Met Office computer guys have fitted a smooth curve to the entire high tide period and taken the peak of that curve. During the period in question the first peak is always the larger, and the difference between the first and second peaks increases with successive tides. Thus the peak of a fitted smooth curve will progressively shift towards the first peak  as time goes on. That would cause the quoted high tide time to advance less rapidly from day to day than is normally expected. It does not mean that the true high tide time varies that way!

The order in which you do the geographical interpolation versus the reduction to a single high tide time will alter the result you get, and somehow on this occasion it resulted in the rather odd looking tidal progression which you noted.

Some of your other points:

For how tides used to be computed look at http://www.ntslf.org/about-tides/doodson-machine and there are also videos, e.g. https://www.youtube.com/watch?v=AxC770lpSLw (look at about 1min 40sec onwards where some of the components are labelled, I find it fascinating).

The time of low water is indeed taken as the reference time for tide curves in Reed's Almanac for the region Christchurch to Selsey Bill and a similar set of "special curves" is also shown for Poole Quay.

In some ways I wished I had taught science since explaining something to someone else is a good way of testing one's own understanding. However University lecturers nowadays are forced to do so much teaching that their research suffers; being purely a research scientist (as I was) is something of a luxury. And I did at least get to supervise a few Ph.D. students.

Sounds like you had a good sail in the Harbour!

Peter