Many dangerous errors in new Navionics cartography

[Bru]

The problem here Tiller Girl, my dear chap, is that you're thinking like a cartographer and surveyor, rather than like a programmer!

Interpolation between two tidal data sources for a given lat,long would not be a particularly complex programming problem and here's the thing, the programmer only has to solve it once

I suspect, however, that they haven't in fact bothered with that level of sophistication. My educated guess is that they reduce to chart Datum using the nearest predicted tide data (information that they already have in their data set) and leave it that

 

[Tiller Girl]

Judging by the example at #4, I doubt they reduce it at all!

 

[CalicoJack]

An interesting point made here. I am very suspicious of relying on anything that has come from user reports. For example in Rye Bay many of the wrecks that people fish have been mis-named. I don't think that this is done in error, as they are well known to those who take the time to find out, but done deliberately in order to mislead. Imagine the scene, you come ashore having had a great days fishing and someone asks where you caught all those fish, you tell them that you had them on The German. They look at their Navionic charts and see a wreck named as The German so that's where you go next time and catch nothing. The problem is that the wreck on the Navionics charts called The German is in fact another wreck altogether.

Okay so misleading people about where to fish isn't going to be dangerous, but if people are prepared to do this what else might they do either by accident or deliberate action?

 

[drakes drum]

Shock, horror! On Sunday night, coming up the Medway in the dark, navigating with my iPad and watching out for the big iron mooring buoys, I noticed some weird features on the Navionics charts, for which I have downloaded the latest updates.

You can check the Navionics cartography for yourself on their free web app here: http://webapp.navionics.com/#@13&key=ip{xHigdC
a. Supposedly there are three mooring buoys (one is a “group of 3”). The one labelled "B.7" just NW of Stangate Spit shows at moderate zoom, to see the other two you have to zoom in a bit. I'm pretty sure none of these exist and never have.
b. The drying patch at bottom left (by Sharp Ness) does not exist and the depths along the shore of Burntwick Island are shown much too shallow.

Pan to the south and east a bit, around Stangate/Queenborough area, and you will see many “rocks” close to the HW mark. They are suspiciously evenly spaced. I don’t believe any of these exist, even allowing for a wide interpretation of what a “rock” might mean in the local context.

It’s only because of my detailed local knowledge that I noticed these anomalies. In an unfamiliar place it could be very misleading, which kind of defeats the object of having charts, does it not?

I have some experience of geographic data and map making but can’t understand how such errors could occur since the cartography must surely come from UKHO and the sources (harbour authority, Trinity House etc) would be the same as for other charts. Navionics does have a “community edit” feature but such additions are shown with a small “+” sign.

However Navionics have recently introduced an add-on called “Sonar Charts” which I understand is bathymetry based on user-reported depths, Though I have not knowingly opted for this feature it might explain the potentially dangerous incorrect depths shown. A friend who has an older Navionics app on his Android phone seems to get more believable cartography than I see on my Android phone, iPad and PC.

There does not seem to be any way to find out the source of Navionics cartography or to report errors, other than “community edit” features. Does anyone else know?

I have not seen anyone report this kind of error yet on YBW.COM but the French have (see St Vaast/Tatihou example):
http://blog.francis-fustier.fr/en/la-cartographie-navionics-devient-elle-dangereuse/

If youcare using an ipad and navionics as your sole means of navigation then you are navigating dangerously. Even navionics themselves say you should not use their products as the primary means of navigation but should also use pukka charts properlu updated.

Even then you should be cautious. We sail in waters that are too shallow for big ships and therefore rarely if ever surveyed . ukho charts before the update to wgs 84 were as far out as half a mile in some places not least because those old suveys were triangulated from land which was found with satellite mapping not to be wherecpeople thought it was.

Ipads are toys and not for serious navigating

 

[Tintin]

It's lucky you don't sail in Venzuela the coast is half a nautical mile out and the out islands three quarters of a nautical milbe out, as is the island Klein Curacao and most lights don't work because they are not maintained. The windward reefs are littered with yachts and coasters who believed their chart/plotter at night. A yacht hit Klein Curacao the night before we arrived and was a good demonstration of the light not working. The north west corner of a Greek island is also out by three quarters of a nautical mile on the Admiralty chart so it happens in Europe too.

Which Greek Island?

 

[RupertW]

If youcare using an ipad and navionics as your sole means of navigation then you are navigating dangerously. Even navionics themselves say you should not use their products as the primary means of navigation but should also use pukka charts properlu updated.

Even then you should be cautious. We sail in waters that are too shallow for big ships and therefore rarely if ever surveyed . ukho charts before the update to wgs 84 were as far out as half a mile in some places not least because those old suveys were triangulated from land which was found with satellite mapping not to be wherecpeople thought it was.

Ipads are toys and not for serious navigating

I'm not sure I buy that - if you mean electronic charts are just toys then that's a point of view, but iPads are just another way of displaying them. I agree you should have a backup paper chart but if you are on a long meandering trip like our month just spent going from Italy to Chalkadiki via whether the wind went, then I wasn't going to source detailed paper for every possibility. I don't think there is anything inherently unsafe about that providing you treat every info source with a pinch of salt - and we certainly found some innacuracies in rarely used anchorages.

 

[bobgarrett]

I agree that electronic devices are just another way to view charts. Of course they carry different risks: the device can fail, they may not show all the data at some zoom levels, they may not show as wide a view as a chart, etc. However, they are more easy to keep up to date, they can show more detail as you can zoom in.
I use both for different applications.
The issue raised here is whether the source data differs, and if so why.

 

[Hadenough]

All the charts and pilot books refer to a set of visitors' moorings in the mouth of Langstone Harbour that don't exist either.

Pete

Very good point. Sir Toms channel pilot is due for update this year, not before time.

 

[prv]

Sir Toms channel pilot is due for update this year

Pretty sure I bought the latest one this winter just gone, doubt there's another coming up just yet

That edition is on the boat so I can't check now whether it still shows the Langstone buoys. I'll try to remember this weekend.

I've had a look at the Admiralty chart on VisitMyHarbour, and the moorings aren't shown. Although, neither are the Eastney Cruising Association ones which did exist on the other side of the entrance, so maybe they never were. I'll try to remember to look at the latest Imrays along with the pilot.

Pete

 

[Giblets]

Pete the ECA buoys on the western side of the entrance are technically private moorings so I wonder if they would be shown on official charts. Certainly the moorings for the Langstone Sailing Club up on either side of the Hayling Island road bridge are not shown on any charts. The so called Visitor's Buoys on the eastern side of Langstoe entrance have not been there for quite a few years.

Let's hope in the next edition of Sir Tom's pilot he manages to get the correct procedure for the entrance into Portsmouth Harbour right.

 

[Arcady]

I have never understood how Navionics achieve a reduction to chart datum with their user supplied data. Seems to me to be a pretty massive job.

Pretty easy, actually. I'm not familiar with the technicalities of Navionics own data collection processes, but assuming the user data is automatically logged to include a date/time stamp, (and GPS, obviously) then really the only question is whether or not Navionics are reducing data from predicted heights or actual recorded tidal heights from the nearest available tide gauge station. One hopes it's the latter.

Don't decry the integration of user-supplied data. A lot of old surveys had incredibly sparse data-lines. More recent surveys of 20 or 30 years ago still involved much manual intervention in reducing the data. It's only in recent years that technology has converged to put accurate GPS (no more Selective Availability degradation of civilian signals by Uncle Sam since 2001), and cheap and reliable electronics with adequate storage for data logging into the hands of your average yachtsman. And 'average yachtsmen' go to some interesting places in their search for a quiet anchorage or interesting pilotage. The information thus acquired would have seemed an impossibly rich source of data to a professional hydrographer of a few years ago!

 

[prv]

Let's hope in the next edition of Sir Tom's pilot he manages to get the correct procedure for the entrance into Portsmouth Harbour right.

Heh. There was at least a correction issued for that cockup.

Pete

 

[dunedin]

So can anybody please clarify, related to the OP's original example - is there material error in the main Navionics cartography for the area in question, or is it just an issue with the extra user supplied data (that many of us are careful to switch OFF as not reliable)?

 

[tony_lavelle]

Just to be absolutely clear, in the example that started off this topic, I was referring only to the "normal" Navionics cartography, not the "sonar" (user-supplied depths) version. The depths are very obviously wrong and there are many spurious buoys and rocks.
This is in a major port area (Sheerness/Thamesport) which has been accurately and frequently surveyed by the Admiralty and the port authority. I think the strange errors have only appeared recently, via automatic internet updates.
The French example shows that the errors are not just in my local area (River Medway, Kent) but elsewhere too.
So the question is: where are Navionics getting their data from?

 

[gjgm]

An interesting point made here. I am very suspicious of relying on anything that has come from user reports. For example in Rye Bay many of the wrecks that people fish have been mis-named. I don't think that this is done in error, as they are well known to those who take the time to find out, but done deliberately in order to mislead. Imagine the scene, you come ashore having had a great days fishing and someone asks where you caught all those fish, you tell them that you had them on The German. They look at their Navionic charts and see a wreck named as The German so that's where you go next time and catch nothing. The problem is that the wreck on the Navionics charts called The German is in fact another wreck altogether.

Okay so misleading people about where to fish isn't going to be dangerous, but if people are prepared to do this what else might they do either by accident or deliberate action?

I haven't gone back to check, but I thought their concept of Sonar Charts was indeed for leisure fishing, not navigation.

 

[Cantata]

So can anybody please clarify, related to the OP's original example - is there material error in the main Navionics cartography for the area in question, or is it just an issue with the extra user supplied data (that many of us are careful to switch OFF as not reliable)?

I know the area in question (and I know the OP well too) and I can verify that there are many 'features' (rocks?? - in a mud creek and salt-marsh?!) on his updated Navionics chart which simply do not exist.
(Later edit - afterthought - and I also have the latest UKHO paper charts, and the 'features' do not exist on those either!)

 

[Angele]

So can anybody please clarify, related to the OP's original example - is there material error in the main Navionics cartography for the area in question, or is it just an issue with the extra user supplied data (that many of us are careful to switch OFF as not reliable)?

The changes made to the Navionics charts for Saint Vaast (see my post #20) were applied to both the sonar and non-sonar versions - i.e. they had overwritten the existing chart information that had, presumably, all come from official sources. There was no way you could therefore "switch off" these amendments.

These changes for Saint Vaast have now been removed, so we are back to what was there before. Presumably because the channels that had been put onto the Navionics chart had never existed. So, for a brief period (days or weeks) the Navionics charts were providing false (and potentially dangerous) information. :disgust:

 

[AntarcticPilot]

I find this deeply disturbing, because it suggests that Navionics QA procedures have broken down. As others have noted, the raw data for ALL charts comes from the UKHO, and in the vicinity of major ports it is reliably and frequently updated. Navionics should have taken these data and applied their own reformatting to meet their internal standards, but the data should have remained unchanged. That is not to say that errors won't creep in - for example, closure of outlines of drying areas might not take place, or a closed area might be miscoded. But the basic geometry should be from the UKHO and should be reliable (not necessarily correct, but reliable in that if you navigate assuming it's true, you'll be safe).

This strongly suggests that Navionics have taken crowd-sourced data and used it to "update" the data they already have. Sadly, the major problem with crowd-sourcing is dealing with error, which may be introduced by bad measurement or (sadly) malicious submissions. And it can be very difficult to identify error in map data; the volume of data and the low number of contributors in some areas mean that a single "bad" source can contaminate the map. I've had datasets from reputable sources where a bored (or mischievous) worker has added spurious features (Santa Claus!); they weren't caught by the QA procedures of my source, and were only caught because I applied a different set of QA procedures.

If someone suffers damage or loss from these charts, I doubt very much that a "Not for Navigation" notice would suffice to keep Navionics from being liable to pay damages.

 

[prv]

That edition is on the boat so I can't check now whether it still shows the Langstone buoys. I'll try to remember this weekend.

Down on the boat now. For what it's worth, the 2013 edition Imray chart no longer shows the Langstone visitors' buoys - but the 2014 Pilot (also published by Imray!) has them both on the chartlet and in the text. Poor show there, Sir Tom!

Pete

 

[Lucky Duck]

This strongly suggests that Navionics have taken crowd-sourced data and used it to "update" the data they already have. Sadly, the major problem with crowd-sourcing is dealing with error, which may be introduced by bad measurement or (sadly) malicious submissions. And it can be very difficult to identify error in map data; the volume of data and the low number of contributors in some areas mean that a single "bad" source can contaminate the map. I've had datasets from reputable sources where a bored (or mischievous) worker has added spurious features (Santa Claus!); they weren't caught by the QA procedures of my source, and were only caught because I applied a different set of QA procedures.

If someone suffers damage or loss from these charts, I doubt very much that a "Not for Navigation" notice would suffice to keep Navionics from being liable to pay damages.

Navionics gave a presentation at my club at the beginning of the season and they stated that the "base" charts were only updated from official sources and wouldn't make changes based on user reports.

The Sonar Charts seem to be a big thing for them to the extend they were loaning out combined echo sounders and GPS units which could be fitted to smaller craft without integrated electronics.