Since looking into chart accuracy following the Orkney cruise (see previous article) I’ve spent a couple of months in the Mediterranean. It’s much worse there.
I vaguely knew that, but previous sailing in the Mediterranean had been without chartplotters and on the earliest occasions without GPS. More recently, chartplotters seem to be standard on charter boats as well as private yachts. We amused ourselves by doing some checking.
In the Ionian and South West Turkey (but only rarely on a cruise round the Peloponnese) we noted numerous position errors, often a tenth of a mile, and a couple of times as much as a quarter of a mile; there were very noticeable – almost random – variations in the size of position errors, measured by GPS, within the area of a single paper chart, so it was not a simple matter of the datum being wrongly set (which of course we checked, as well as satellite coverage.) The errors were clearly the result of using an accurate GPS on inaccurate paper and electronic charts based on old surveys.
Rod Heikell, the pilot book author, in a talk at the Cruising Association in London in February 2013, amused the audience with a plot of a yacht track running up the hill behind a Mediterranean town as the boat approached the harbour. That was exactly the kind of anomaly we picked up, in one case in the Ionian with the boat moving down a road on the island we were coasting.
Now none of this is new to old hands in the Mediterranean, with some of whom I have recently discussed it on a Cruising Association forum. Charts based on old leadline surveys can miss underwater dangers and sometimes have large position errors, mainly of longitude.
The advice, of course, is that GPS should always come second to older methods of pilotage, using eyes, handbearing compass, clearing lines and depth sounders, because while the longitude on charts may be wrong, the relative positions within a single area of a chart – for example a bay and harbour – tend to be accurate. The best a chartplotter and GPS can do is get you out of somewhere by using a plot of the track, once you have safely got in by conventional means. The better equipped Med sailors also talk of radar and forward looking echo sounders. Pilot book authors give GPS waypoints they have checked themselves.
However, it is very different for new arrivals in an area, and for holiday charterers it can be positively dangerous. A couple of recent trips were on bareboat charters with a large UK operator. After a series of chartplotter and GPS checks at known chart positions for a week, we found large and variable errors. But neither the technical staff at base nor a flotilla skipper we spoke to had any idea of the scale of the local chart errors.
When we checked the equipment on arrival at one base in Turkey, we found the plotter showed the boat nearly a quarter of a mile from the marina. The company sent a technician to rebase the chartplotter’s boat icon onto our pontoon in the marina. It was an older type of instrument which we had not come across before, which allowed recalibration. (This was not a datum adjustment, which we also checked). The technician announced that the plotter was now fine.
That night, 15 miles away, the plotter was a tenth of a mile out; for the rest of the cruise, the charted positions within a cruising range of no more than 50 miles fluctuated from zero to a tenth of a mile and in one case approaching a quarter of a mile, simply because the GPS was more accurate than the charts. Back at base at the end of the week, the error was zero, so there was nothing to show the next customer that there was a problem.
The base clearly thought we were imagining it. I got the impression they were not even conscious of Rod Heikell’s pilot book warning of the size of longitude errors in Turkey. Skipper briefing at the base did warn about chartplotters, but it was not specific to the area, and was the kind of routine mention you hear in the UK. Visiting charterers might well assume it meant the relatively modest range of inaccuracy found at home, rarely more than 100 metres.
One lesson is that specific warnings with numbers attached are vital for getting the message across: for example the Imray Ionian chart used to have a general note warning that there were longitude inaccuracies, and I suspect not much notice was taken. A quantified warning was recently added saying the errors could be up to a minute of longitude, which has a much greater impact. Suddenly you realise what you are up against. The errors could be as big as the gaps between some of the islands. In these circumstances, chartplotters are an accident waiting to happen for holiday charterers. It would be safer to switch them off.
Next, how do we find out more about survey accuracy and chart errors?