User derived bathymetry

[boomerangben]

Navionics and I think others can collect bathymetry data from users (who presumably are able to opt in/out) which can then be used to update their database and then passed back to all users.

What I find interesting is how they account for states of the tide, barometric pressure, storm surges and so on.

The concept is, I think, excellent, passing on potentially accurate and detailed information of otherwise poorly surveyed areas. But just how accurate is that information that is passed back to users?

 

[Deleted member 478]

I am under the impression that Navionics average out the changes in the data received from their users. If I eventually remember where I read this, I'll post a link.

 

[Sandy]

I am sure @AntarcticPilot has posted about this before.

 

[GHA]

too unreliable imho to rely on. Sasplanet comparing satellite images to navionics/cmap can be very useful to check survey data/datums. Though datums seem far better on navionics now compared to not too many years ago when sailing down high streets was much more common even just a bit off the beaten track.

 

[AntarcticPilot]

I am sure @AntarcticPilot has posted about this before.

I have; for those who don't know I am a retired GIS Manager at a well-known polar research institution. My view is basically as follows:

1. In principle, it's a good idea, but it suffers from many drawbacks and Navionic's implementation has some deficiencies.
2. Data are collected by uncalibrated equipment and are subject to the introduction of errors arising from the settings and operation of the equipment.
3. As noted, while the astronomical component of the tide can be accounted for, meteorological components are unknown. Marine survey uses tide-gauge data; this is not available for soundings collected at arbitrary times and locations.
4. Where data are plentiful, none of the above is a problem; outliers are readily identified, and random variability averages out. However, in many places data are NOT plentiful, and a single track of bad data can result in spurious features (this has been reported here).
5. Navionics apply interpolation techniques to the data, assuming that the shoreline is a zero contour. Again, this can (and often does) produce spurious depth contours that bear no relation to reality. Many examples have been posted here.; some have been positively dangerous in that dangerous features were absent from the data because no one sails over a reef! In my view, this is the most dangerous aspect of the system. Navionics should not extrapolate the data beyond the extent of the data.
6. Isolated dangers are avoided by sailors, and so do not appear in the data.
7. It appears that Navionics apply little or no QA to the data (many of the issues above could be avoided by appropriate QA). But QA is time-consuming and requires skilled staff - Hydrographic organizations spend at least as much time and skill on QA as they do on compiling data (I have been part of the QA for Antarctic charts).

To sum up - in well-frequented areas with many contributors it can be useful - but such areas are likely to be well-charted by hydrographic agencies. If using it, you should ALWAYS be aware that it may not show important features and that it can be in serious error. Please note that in well-surveyed areas the HOs have much more data than is printed on the chart; the soundings on the chart are usually a representative selection of those available.

A general point: Crowd-sourced data can be excellent (e.g. OpenStreetMap), but they require a mechanism by which spurious, inaccurate, or malicious data can be removed. They are only useful in situations where there IS a crowd! For example, one of the first examples of crowd-sourcing (Seti@home) required that a set number of results for a particular data parcel were received before the data were accepted into their database. In areas where only one or two persons are reporting, it is very easy for error to accumulate without any means of identifying it.

 

[boomerangben]

Thank you Antarctic Pilot .

 

[GHA]

bearing in mind that all charts are wrong, but quite often close enough to be useful... more important knowing where you are not rather than exactly where you are..

 

[Deleted member 478]

The Navionics chart where this catamaran has met its fate says it is 10 metres deep right there. I had Satellite image charts for the area, so was informed sufficiently to go around the reef.

As @GHA advises, get your head around making your own alternative charts using SasPlanet.

 

[Chiara’s slave]

I’d have thought that surface offset was more of an issue than calibration for any relevant data, ie it’s shallow. Most of us don’t care as long as it’s deep enough, so whether someone's sounder is set to surface or keel bottom is a big issue. If your sounder is 20% out, at the highly relevant 2m contour, the 2m ‘error’ introduced by not knowing the sounder offset is way worse than it being hysterically inaccurate.

 

[AntarcticPilot]

I’d have thought that surface offset was more of an issue than calibration for any relevant data, ie it’s shallow. Most of us don’t care as long as it’s deep enough, so whether someone's sounder is set to surface or keel bottom is a big issue. If your sounder is 20% out, at the highly relevant 2m contour, the 2m ‘error’ introduced by not knowing the sounder offset is way worse than it being hysterically inaccurate.

The main problem is that Navionics extrapolate outside measured data, producing nice, authoritative looking contours that have no basis in reality. Many examples have been posted on here. The measured data are usually ok, but you don't know where they are!

 

[lustyd]

One of their most useful features is recording your own depth onto the chart in real time, allowing you to survey an anchorage. In this instance you know it’s accurate and can see it on your track so super useful compared to driving in circles and remembering where that deep bit was

 

[Chiara’s slave]

One of their most useful features is recording your own depth onto the chart in real time, allowing you to survey an anchorage. In this instance you know it’s accurate and can see it on your track so super useful compared to driving in circles and remembering where that deep bit was

We use it for exactly that. Drive around at high water, you then know what it’s like 6 hours later.

 

[GHA]

We use it for exactly that. Drive around at high water, you then know what it’s like 6 hours later.

can you export the data in any form from navionics? Sort of thing that can be handy to share as a GPX.

 

[Chiara’s slave]

can you export the data in any form from navionics? Sort of thing that can be handy to share as a GPX.

I have ticked the ‘share my data’ box. Mine is Garmin based, it does the exporting for me. There may well be a way to write the data to my second SD card, I have not investigated.

 

[dunedin]

We have to remember that the UK is a small and often largely irrelevant country for these sorts of international products. There are massively more Navionics users in the USA and Med than here.
I believe that the US recreational lakes used by sports fishermen etc are not well charted. But lots of boaters means user bathymetry data is really important and easily obtained. (And no tides.) Hence these were what this feature was invented for.

Regarding allowance for tide heights, keel offsets etc, I suspect people are looking at this the wrong way. A chart publisher doesn't need to worry about that. Instead if the track goes over basepoints with known depths it is easy to convert the other soundings accurately to chart datum using this. Not 100% certain that is how it is done but that is how it can be done easily by computer.

Still keeping it switched off as has big flaw in that if few boaters go there can't get accurate data. Fine for US lakes and lake Solent or Poole, but for remote Scotland I will stick with Antares Charts.

 

[GHA]

I have ticked the ‘share my data’ box. Mine is Garmin based, it does the exporting for me. There may well be a way to write the data to my second SD card, I have not investigated.

Sounds like a no then. Wouldn´t be surprising. Commercial company actually letting you get the data you´ve just recorded for them...
No matter, wouldn´t trust it anyway. With sasplanet & opencpn there´s masses of data to use as charts now, sasplanet is great for checking navionics charts against cmap against arcgis satellite etc

 

[Chiara’s slave]

Sounds like a no then. Wouldn´t be surprising. Commercial company actually letting you get the data you´ve just recorded for them...
No matter, wouldn´t trust it anyway. With sasplanet & opencpn there´s masses of data to use as charts now, sasplanet is great for checking navionics charts against cmap against arcgis satellite etc

Since the data isn’t sound enough to have commercial value, it’s quite selfish of them. We can see other people’s data, but have not attempted to make use of it. The practical use of it is really only as Lustyd says. A quick survey of your prospective anchorage. Though I might be tempted to motor slowly along the Yarmouth/Bouldnor sea front, close in, for my personal use later

 

[Roberto]

In case one does not want to risk too much to bump into the bottom, there is the portable solution. This is a Garmin Striker 4 CV on top of a hopefully waterproof box with a 12V battery inside. I put the transducer on a pole holder used by carp fishermen, it has a clamp and can be tied to the dinghy transom or to my boat aft swimming platform.
This fishfinder has what they call "Quickdraw contours", one zigzags over an area and it automatically draws 30cm spaced contours. Also, looking at the type or return on the display one should be able to tell the nature of the bottom, or so they say.

 

[GHA]

Since the data isn’t sound enough to have commercial value, it’s quite selfish of them. We can see other people’s data, but have not attempted to make use of it. The practical use of it is really only as Lustyd says. A quick survey of your prospective anchorage. Though I might be tempted to motor slowly along the Yarmouth/Bouldnor sea front, close in, for my personal use later

The user data on navionics charts is a big selling point so not surprising they won´t let you get at any it.

So how do you get to the data from your just surveyed anchorage?

Vaguely similar I´ve had a sounder working connected to a raspberry pi which with python creates a GPX file with spot depths as waypoints or track marks to import into opencpn.

 

[ylop]

I’d have thought that surface offset was more of an issue than calibration for any relevant data, ie it’s shallow. Most of us don’t care as long as it’s deep enough, so whether someone's sounder is set to surface or keel bottom is a big issue. If your sounder is 20% out, at the highly relevant 2m contour, the 2m ‘error’ introduced by not knowing the sounder offset is way worse than it being hysterically inaccurate.

I've always assumed this was a tool for American fishermen (their core market) who are more interested in the shape of the bottom for finding good places to fish than we are about knowing if its 10 or 100m deep.