|Part Of The Keel Damage|
Fact is, we were lucky. The mast and rig stayed up, the hull was not breached and the damage could be repaired fairly simply. However a boat built less heavily would have been in severe trouble. The story that follows is intended to serve as a warning for others, to help prevent further accidents. So how did it happen ?
First off, I made the mistake of trusting Navionics Sonar Charts, which I'm unlikely to do again. Secondly, I failed to check any secondary navigation aids or references. So with more care I could have avoided the reef. Dodging lobster pots, early in the morning, I was busy at the helm and trusted the charts that had brought us into harbor the afternoon before. Never again...
It's ironic that I was actually looking at the charts (on the iPad) as we connected with the reef - when Ley stumbled up the companionway and I picked myself up off the wheel, we both looked at the chart on the iPad - it showed over 13 meters of water depth. Wrong. In fact very wrong.
Crystal Blues was at that point bobbing on the gentle swell, afloat, but occasionally colliding with the reef on the bottom of the swells and occasionally hitting something as she rocked from side to side. Not wanting to start the engine in unknown waters, we lowered the dinghy and towed her away to deep water using the outboard motor. We lifted the floor boards and inspected the bilges for damage or leaks, and then (more than a little shaken) proceeded to Booth Bay, where I dived on the boat and captured the image above.
|Pulling Away From The Reef|
So now, over three weeks later, we're hauled out of the water at Journey's End Marina in Rockland, Maine, working towards repairing and refinishing the keel. In the time since the accident both we and others have had time to look at the Navionics charts for the area in some detail, and the results are not good.
The Missing Reef
First of all, this problem has nothing to do with zoom levels on the screen. No matter how close in you zoom, that reef was not shown. So here is what we were viewing on screen right when the collision occurred :
With the wisdom of hindsight, it does look weird. The yellow line is our track, and it shows just over 13 meters where we came to a sudden stop. When we checked the standard Navionics chart (not the Sonar Chart version), a serious chart error became apparent :
You can see that this version shows a depth of 5.4 meters nearby, and the dark blue zone adds to the message that we shouldn't be there. However it gets worse, much worse.
The official government chart, above, shows a minimum depth on that reef of less than 1 meter. Yoiks ! So, not only the Navionics Sonar Charts were wrong, but also the "standard" Navionics charts. How can this be ? How widespread is this problem ?
Soon after the accident, we anchored in Rockland Harbour to arrange for the haulout, repairs and a marine surveyor to inspect the damage. Other cruising friends, also OCC members, came to visit and share a meal with us. They explained that they too were using the Sonar Charts when navigating in harbors, as they valued the extra detail provided, and were shocked to hear of our collision. They were nervous enough to examine our accident site in detail, and then to cast more widely around the region looking for similar charting errors. Apparently it took only 10 minutes for them to find several more in the vicinity of Potts Harbour. As our friends said, these things are dangerous !
I had quickly notified Navionics of the chart error, and they proudly committed to correcting the error before the end of August. We checked recently, updating the charts on just one of our systems, and yes the Sonar Charts now appear to be correct for that reef, showing a minimum depth of 1.2 meters :
My follow up response to Navionics was more specific - I explained that we (and others) had since found numerous examples of dangerous errors in their charts, and suggested that they had a systemic problem that needed fixing - were they going to simply correct the charts when advised of errors, or were they going to address the wider issue of fixing the system that created them? Navionics response was less than pleasing :
"Dear Neil Langford,
I'm sorry you feel this way. We investigate the reports as they are brought to our attention. We have guidelines that we follow for the logs to process; however there can be issues with the processing so we welcome any feedback provided. If you have found numerous instances, please feel free to provide the coordinates of the locations so we can further investigate as we cannot be in all places at once and will be happy to work with anyone who would like to report this information.
For additional support, access Navionics online Help or Chat Now with a live agent.
Best regards,Jennifer Pereira
Navionics Customer Service"
OK, so they can't be in all places at once ? What does that mean ? And there can be issues with processing ? And they'll be happy to correct their mistakes if only we'll report them.... I had hoped for a more positive response!
Navionics web site advertising says :
"Navionics integrates the sonar logs from individual users with existing data to ensure the SonarChart™ becomes more and more precise"
What they haven't yet explained is how often they actually lose large chunks of the existing data - like entire reefs.
Loss Of The Leopard Catamaran Tanda Malaika, French Polynesia
Just a few days ago I learned of another accident where Navionics charting errors were possibly a contributing factor. The catamaran was a total loss, though very fortunately the family on board survived. You can read about it here, including a range of comments from many cruisers.
For me the interesting comments came from one observer who took a good look at the Navionics charts and found errors that look similar to those we found. These are his images. He said :
"I have never seen an error like this before. It is not missing data, and it is not old/obsolete survey data, and it is not a data layer issue (the three typical errors). I don't really understand how the sounding got warped that way."
From that sad story, we can see that the Navionic charting problems may not be associated only with the rocky coast of Potts Harbour, Maine.
Take A Look At Other Areas
When I learned from friends that they had found other significant chart errors, near Potts Harbour, I was naturally concerned. But when I read of the errors in French Polynesia, I started to think that Navionics may have a wider systemic problem. Since we're in Rockland as I write this, I started looking at the coast close to this harbor and it took only 10 minutes to find yet another set of significant and dangerous errors.
|Navionics Sonar Chart|
What a very dangerous representation - the Gangway Ledge (1.5 meters) is shown as being over 16 meters deep! Inner Grindstone is shown as 5 meters deep, when in reality it is around 2.4 meters. If I can find this area in just a few minutes of local searching, how many other reefs, ledges and rocks are missing (or wrongly plotted) on these charts ? This really is dangerous charting.
The Raymarine Version
If you've followed the story to this point, you're probably wondering if the problem is limited just to Navionics charts when displayed on the iPad - and I was wondering the same thing. So, immediately after our collision I checked the Navionics data on our Raymarine chart display. Raymarine and Navionics are closely linked - yes, you can run many other types of charts on Raymarine displays, but the most common charts supplied with Raymarine systems are made by Navionics. Additionally, Raymarine displays are used to collect the crowd sourced data that Navionics uses to create their Sonar Charts.
On our system (to their eternal credit) Raymarine displayed a large warning when I selected the Sonar Charts - Not For Navigation it said. To be fair, a similar warning is also available on the iPad, covering all Navionics charts - though it is hidden and you can start the software without reading it - as I'm sure most people do.
Sure enough, the missing reef problem existed on the Raymarine display system as well - the reef we hit was clearly displayed on the standard charts, but not displayed on the Sonar charts, when viewed on our C97 multi function display.
There is a sense of un-reality about this situation. Here we have a long established and reputable manufacturer, Navionics, who publish electronic chart data aimed at the marine industry.
Navionics marketing includes promotion of Sonar Charts for use on marine chart plotter displays, claiming extraordinary detail.
So check the wording in the email I received today, at right. See any warnings there about "Not For Navigation"?
On the Navionics website (see it here) Sonar Charts are promoted like this :
"Get the most up-to-date charts every day. Navionics integrates the sonar logs from individual users with existing data to ensure the SonarChart becomes more and more precise, even in the ever-changing conditions of sea, lake and river bottoms."
Once again, I don't see any warnings in that text saying these charts are actually not for navigation. So I have to ask, if I can't use them for navigation, what will I actually do with these Sonar Charts ?
Fact is, the marketing of Navionics Sonar Charts places them on a pedestal, supposedly better than the standard charts - after all, the idea of crowd sourcing depth data and using it to improve the government charts is exceptionally attractive. If only they could be trusted. If only they could be used for actual navigation. If only Navionics could integrate the crowd sourced data, without actually losing significant and well known dangerous features.
So how do we deal with this? Specially if you have a Raymarine system supplied with Navionics charts? Well, first up I won't use Sonar Charts at all from this point. Secondly, I'll use the government charts whenever possible, as even the standard Navionics Chart showed the depth at our accident site incorrectly as 5.4 meters - the 0.9 meter shallow spot was completely missing (they've now fixed that since my report). Note that the official government charts are available in the Navionics iPad app.
Also, for US coastal navigation we'll use the US NOAA vector charts that can be downloaded for free and run in Open CPN. Or, for $9.99 you can license the NOAA vector charts for the US Eastern Seaboard, from the Raymarine chart store online. Better still, purchase the C-Map charts to run on your Raymarine system - I suspect you'll sleep more soundly knowing the data is actually intended for navigation.