Question about chart datum

[mel80]

This isn't really an issue of much practical use for navigation, but which has been puzzling me none the less.

Now, we all know that chart datum is aproximately the same as the lowest astronomical tide. My question is: how do hydrographers calculate this from their soundings, when accurate tidal predictions (both times and heights) are only available for a very few locations?

One way that comes to mind is that they sould simply sit in one position, continualy making soundings at approximately the time of low tide. The least depth would be the depth at low tide, and the ratio between that tides range and the range at LAT (at a standard port) could be used, along with corrections for atmospheric pressure etc, to calucate depth at LAT. However, that would seem to be a painfully slow method (allowing only two observations a day) and not really of any practical use.

Any hydrographers (or anyone else for that matter) care to comment how its done in practice, or am I just expecting too high a degree of accuracy?

 

[tome]

LAT is a calculated value based on astronomical predictions and local observations. The vertical datum for land heights is the Ordnance Datum Newlyn, which approximates to MSL at Newlyn

The relationship between Newlyn and CD is not constant: UKHO provide the following differences

Torquay -2.8m
Lowestoft -1.5m
Barry -6.1m
Barnstaple +0.6m

There is no such thing as a 'seamless' vertical datum, although this will start to happen as WGS-84 becomes more widely adapted

 

[mel80]

Thanks for the reply. I understand all that. I guess my question could be more clearly stated as: how does a survey vessel know it's height relative to chart datum at the time it is making a sounding?

In particular, how does it do this when it is far away from an area for which tidal observations have been reduced?

 

[tome]

If we are doing a bathymetry survey (not hydrographic, I hasten to add) we will set up a tide gauge at a surveyed benchmark - there are plenty scattered about so we will chose a convenient one closest to our survey area. The tide gauge records the height of tide and is linked to CD via the benchmark, and transmits this data in real-time

On board the survey vessel, the recorded depths are adjusted to chart datum on the fly. If the telemetry fails, we can achieve the same by recovering the tide gauge data from the local recorder and post-processing our bathymetry

Edit: Should add that when working offshore, tidal variations are insignificant but swell becomes a factor and we may fit a heave compensator

 

[Twister_Ken]

Rather like the sound of a heave compensator. Do they work after a night of ale and vindaloo?

 

[mel80]

Ah Ok; that makes sense, thankyou.

Presumably you could also come up with some sort of predicted error in the sounding based on the distance from the 'benchmark' (and differences between two adjacent benchmarks), thus keeping the whole survey within acceptable bounds of accuracy.

Has a similar process always been followed? I ask because I have a 'modern' chart that uses some survey data from the 19th century, and have often wondered how rigourous survey methods were back then and what sort of accuracy they managed to achieve with (relatively) basic equipment.

 

[Sailfree]

If I understand you correctly CD is therefore always local and not a theoretical plane! Is that correct?

 

[peterb]

[ QUOTE ]
If I understand you correctly CD is therefore always local and not a theoretical plane! Is that correct?

[/ QUOTE ]

As I understand it, the closest thing to a theoretical plane (more accurately, a theoretical spheroid) is mean sea level, although even this has bumps and hollows due to local gravitational anomalies. CD will differ from this by roughly half the tidal range. Since the tidal range is local, CD will be local

 

[SteveGorst]

[ QUOTE ]
even this has bumps and hollows due to local gravitational anomalies.

[/ QUOTE ]

Whats this? Do we weigh more in some places on earth than others? I have been trying to persuade my wife that we have developed one of these underneath the scales in our bathroom for years, and it appears to be getting worse!!

 

[tome]

Mean sea level is an approximation of the geoid which defines the planet. The geoid links points of equipotential gravity, but is an irregular shape which is difficult to model mathematically

So the earth (even at MSL) is an irregular bumpy object, which is why there are so many datums in use around the earth (eg OSGB36 here in the UK, ED50 in Europe, NAD27 America etc). Each datum is an attempt to make a speroid which best fits the local geoidal bump - spheroids being much easier to model mathematically

WGS84 is, by necessity, a whole earth datum since GPS is a global positioning system. This means that it only fits the earth in odd places, being above MSL in some and below in others. That's why heights always look a bit odd when operating at sea level

 

[Woodlouse]

When I was on a vessel doing a bathymetry survey on the Bristol Channel a couple of years back we had a tide gauge set up like you say. However with the echosounder and the RTK dgps we were able to basically measure the distance of the seabed from the satellite. This meant that tide and sea state were not really a problem.

 

[Twister_Ken]

[ QUOTE ]
Mean sea level is an approximation of the geoid which defines the planet. The geoid links points of equipotential gravity, but is an irregular shape which is difficult to model mathematically

So the earth (even at MSL) is an irregular bumpy object, which is why there are so many datums in use around the earth (eg OSGB36 here in the UK, ED50 in Europe, NAD27 America etc). Each datum is an attempt to make a speroid which best fits the local geoidal bump - spheroids being much easier to model mathematically

WGS84 is, by necessity, a whole earth datum since GPS is a global positioning system. This means that it only fits the earth in odd places, being above MSL in some and below in others. That's why heights always look a bit odd when operating at sea level

[/ QUOTE ]

And this is from a man who can't find his way from Petersfield to Lymington, following a motorway!

 

[tome]

Harrumph! You can see why I joined the marine division rather than the land surveyors

 

[tome]

Yes, you can use RTK to get up 2cm horizontal/ 5cm verical accuracy but only in very limited areas such as estuaries and close to shore. More than 10km from base station it drops off rapidly in accuracy and it's back to vanilla dgps

 

[mel80]

[ QUOTE ]
...with the echosounder and the RTK dgps we were able to basically measure the distance of the seabed from the satellite. This meant that tide and sea state were not really a problem.

[/ QUOTE ]

Interesting. It wouldn't be much use to a hydrographer though, would it; he would also need to know how far chart datum was from the satellite.

 

[jamesjermain]

And did you know that the WGS84 model puts the GPS prime meridian 107 meters (approx) east of the physical meridian at Greenwich?

 

[Woodlouse]

Thats all in the post processing. But chart datum will be known, as will the hight of the satellite. So not to hard to work out.

 

[tome]

[ QUOTE ]
It wouldn't be much use to a hydrographer though, would it; he would also need to know how far chart datum was from the satellite

[/ QUOTE ]

GPS solves for height as well as lat/lon. Provided you know the local geoidal separation and it's relationship to CD you don't need satellite ranges to CD, but you do need the tide gauge

 

[tome]

Bloody cheek, isn't it?

 

[Woodlouse]

I was told that the RTK would give you a fix to within 3mm on the horizontal and no more than 2 or 3 cm on the vertical. It is however, as you say limited as the gps relies on either a VHF or GSM link to another unit in a fixed location that has to be as to the survey area as possible.