Chart Datum and negative tidal heights

[bdh198]

This is my understanding:

Chart Datum is the plane used on navigational charts to show the depth of water before accounting for the height of the tide. By calculating the tide height in a particular location and adding it to the Chart Datum the true depth of water at a particular time can be determined. Chart Datum is calculated locally and is dependent on the tidal range in that location. As such it is close to the lowest possible tide that can be predicted, although I appreciate that storm surges and unusually high atmospheric pressure can create the occasional tide that will fall below the lowest astronomical tide. Nevertheless it is rare for the tide to fall below the Chart Datum shown on a navigational chart because it is calculated locally depending on the tidal range in that location.

If that is correct then why are there some locations where the tide regularly falls below the Chart Datum? One location I currently have in mind is Bembridge Harbour. There, the low water is often calculated to be negative in the days surrounding the spring tide. This Saturday’s spring tide (which is not a particularly heavy spring) is preceded by three days of negative low waters and followed by a further three days of negative low waters (seven days of negative lows in total). The greatest negative is 0.31 meters, but for a heavy spring it can be as much as a meter or more.

Consequently, the low water in Bembridge is frequently predicted to be lower than Chart Datum, which itself is supposed to be fairly close to the lowest predictable tide. Is it the Chart Datum in Bembridge that is not as accurate as it should be, or have I missed something else completely?

 

[maby]

Isn't it down to the fact that tide levels published are for a limited set of tide stations and in most places we have to apply corrections to arrive at a local tide height? These corrections are, at best, a bodge and I guess they are not always very accurate.

 

[LadyInBed]

As above, CD is set at Standard Ports - Portsmouth for Solent.
In these cost cutting days it would be too much effort to change!

 

[bitbaltic]

Nevertheless it is rare for the tide to fall below the Chart Datum shown on a navigational chart because it is calculated locally depending on the tidal range in that location.

If that is correct then why are there some locations where the tide regularly falls below the Chart Datum?

The simple answer is because it is a model, and all models are an approximation. I don't know Bembridge particularly but the question you are asking is effectively about the spatial resolution of the model which produces the LAT/CD surface.

I can't find exactly how the UKHO make their model for the zero contour line, but they will have to do it something like this:

1) choose a number of places to make a series of long-term observations of tidal rise and fall, in order to make a calculation of LAT at that location (the calculation itself will be an approximation to some degree, but perhaps quite a small one). Each calculated LAT at each place therefore becomes a data point. Apparently LAT is calculated over a 19-year cycle.

2) interpolate between those data points all the way around the UK coast to produce a continuous Chart Datum zero contour. This is where the real approximations are made, because this page suggests that there are only around 700 definitions (=datum points) to interpolate between in the whole of the British Isles. (I don't know how many standard ports there are, but I suspect this is more than just them alone).

If you look at the page I linked above, you can see that the researchers were trying (for their own purposes) to create a computer model of chart datum. Making a model of the whole of the UK and Irish coastline based on only 700 scattered data points is necessarily going to mean that large areas of coastline are filled in with interpolated, not measured values. Anywhere you go on the coast, you are more likely to be at a locality where the CD value was arrived at by interpolation rather than by direct measurement, and the interpolation will be an approximation. i.e. reality will be slightly different from the published CD at the harbour you've arrived at. This presumably is the case at Bembridge.

Whether the actual differences you are talking about are significant enough to be acceptable or not (ie whether they are within the unresolveable error of the model or whether there is something duff at Bembridge) requires more knowledge of the variance in calculated CD and the sort of methods used to make the model interpolations, which is outside the scope of a morning's google.

 

[AntarcticPilot]

Any datum - horizontal or vertical - is an arbitrary fixed coordinate system from which other depths/heights/latitude/longitude are measured. Different maps use different vertical datums; for example the Ordnance Survey use a fixed datum for the whole of the UK. Charts tend to use a tide related datum that is "fixed" to tide gauges, and is usually chosen to be LAT or LWMS. Your GPS (if it gives a "height" reading; it's not useful in a marine environment!) will give it with respect to a the WGS84 global datum.

The point is that a datum is arbitrary; it can be anything the chart maker decides. I've worked on a set of charts of Deception Island surveyed at different times and by different organizations where a major part of my effort was to find out and correct for about 5 different vertical and horizontal datums! Also, tide-gauges are usually located at major ports, and places in between them may well find that the LAT datum interpolated between tide-gauges is exceeded by local tidal effects.

 

[John Barry]

Antarctic Pilot is right - vertical datum is arbitrary.

The IMO stated (30 years ago) that charts soundings should be reduced to a level that the water 'rarely falls below'. The UKHO use 'Approximately LAT', the French normally use 'Lowest Predicted Tide'., the Americans often use Mean Lower Low Water'. In the Baltic and the Med it's normally 'Mean Water Level'

The UKHO publish details of the difference between LAT, CD and OD Newlyn for each port (standard and secondary) in the Admiralty Tide Tables, which also give details of the period of observations used to create the local model. Yes it should be 19 years, in practice it isn't. It should also be at least 29 day, in practice it sometimes isn't (although this would be very rare in the UK).

Oh yes, OD Newlyn is currently about 0.20m below actual Mean Sea level, and 0.80m below the EMG2008 Geoid, and something like 50m below the WGS84 Ellipsoid.

John.

 

[bdh198]

Thanks for the replies and clarification as to the reliability of Chart Datum.

I suppose from a practical point of view it highlights the importance of always knowing (or having some idea) of the actual tide height in unfamiliar shallow water and not simply relying on the Chart Datum level shown on a chart or GPS contour to assume 'well lowest predicted tide is xxx so I should be ok'. A secondary tide calculation from Portsmouth to Bembridge at low water on a heavy spring would turn a small tide height at Portsmouth into a considerable reduction from the Chart Datum at Bembridge.

 

[John Barry]

Time and height differences for Bembridge are only approximate (see comments in ATT).
Also they are heavily affected by wind and pressure.

John

 

[MountainGoat]

Isn't it down to the fact that tide levels published are for a limited set of tide stations and in most places we have to apply corrections to arrive at a local tide height? These corrections are, at best, a bodge and I guess they are not always very accurate.

If they were published hyper-locally, you'd end up with large scale charts where the chart datum varies within the chart! Not ideal.

Where we are in Burnham on Crouch, low springs often go below chart datum, but it's an area with lots of estuaries which amplify tides, so what can you do?

 

[John Barry]

If they were published hyper-locally, you'd end up with large scale charts where the chart datum varies within the chart! Not ideal.

Where we are in Burnham on Crouch, low springs often go below chart datum, but it's an area with lots of estuaries which amplify tides, so what can you do?

The difference between CD and OD Newlyn does vary across charts. OD Newlyn can be considered as the local Geoid (a surface level with mean sea level across the UK).
The CD differences are quoted in ATT, for example - Differences on OD Newlyn (ATT (NP2101 2015):

121b Rochford -0.40
122 Burnham on Crouch -2.35
122c Battlesbridge +0.50
123 Bradwell Waterside -2.68
123a Osea Island -2.63
123b Maldon +0.11

The local corrections allow for the differences in the 'level' differences in charted datum. It's only really a problem when you try to relate it to land heights that are referred to OD Newlyn - everything on charts is related to CD for depths/drying heights or (normally) HAT for land heights.
CD is chart 0. LAT and HAT heights above chart 0 are in the ATT for standard ports (HAT is given for all ports, Burnham is +5.8). e.g.:

Walton on the Naze LAT -0.1 HAT +4.7
Harwich LAT -0.1 HAT +4.4
Lowestoft LAT +0.1 HAT +2.9.

From the Burnham harmonics, Z0 = 2.50, M2H = 1.88, S2H=0.53, so mean spring low will be 2.50 - 1.88 - 0.53 = 0.09 (standard corrections give 0.4 - 0.2 - 0.2), so it's not surprising that spring low is often below CD. Burnham HAT is 5.8, so with Z0 of 2.5 you get LAT of -0.8.

John