Weather buoy 42060

Figure 1 – Courtesy of NOAA & NDBC

I am always impressed when a weather buoy survives an encounter with a hurricane as did 42060 yesterday with category 5 hurricane Maria. 42060 reported mean winds of 64.1 knots (74 mph) with gusts to 81.6 knots (94 mph) at 1510 UTC yesterday (19 September) afternoon, with a minimum pressure of 955 hPa at that time (fig 1). Looking at the wind directions reported by the buoy it looked like the eye of Maria passed close by to the northeast of 42060. Although the AWS was undamaged and still reporting, 42060 According to NOAA, has now broken a drift from her sea anchors (fig 2). I still can’t understand why the same reliable AWS that are used on these weather buoys aren’t used on land stations?

Figure 2 – Courtesy of NOAA & NDBC


41043 takes a direct hit from Jose

Figure 1

Even with a couple of bogus observations the contouring can’t handle either Irma or Jose at all well in this mornings plotted SYNOP chart, but without some kind of background field of gridded MSLP values from the T+3 of some handy NWP model, it’s the best I can do. Here are the observations from Key West (fig 2), the AWS stopped reporting gusts at 21 UTC yesterday, the eye of category four hurricane Irma is now close by at 10 UTC according to the NHC.

Figure 2

Category four Hurricane Jose was very close to the weather buoy 41043 at 09 UTC, and as far as I can see the missing pressure tendency at 09 UTC should have been something like 192 (19.2 hpa lower than 3 hours before). An hour later, 41043 only reported a sea temperature (10 UTC), maybe the hurricane has damaged it too severely, but they usually survive.

Figure 3

Ignore the very high gusts from yesterday, they are spurious. I have a problem parsing the American Weather Buoys because of the exotic groups that they use in them. Here’s an example from the 11 UTC observation on the 9th of September (fig 4). They report wind speeds are in metres per second. I obviously parse the final group 91058 as the gust, and convert the 58 mps to 113 knots. Looking at this one it maybe that it’s not my parsing at all, but this has become quite a problem recently, drop me a line if you can see where exactly I’m going wrong

Figure 4

Weather buoy observations

Figure 1

There are a number of weather buoys and light vessels scattered around the British Isles and in the eastern Atlantic (fig 1). The markers in the map of the North Sea are from oil or gas rigs. I thought that I’d just look at some of the data that they collect over a year and graph it. The sensors typically collect the following hourly information:

  • Air Pressure
  • Wind Speed
  • Wind maximum gusts
  • Wind Direction
  • Temperature
  • Dewpoint
  • Sea Surface Temperature

At times the onboard sensors do go U/S, which can be for extended periods, because repairs have to wait for the quieter summer months to get done, but usually they are very reliable. This is a record of the SST at K7 (64045) which is tethered at 59.1° north 11.7° west (fig 2), and as you can see the SST doesn’t vary that much 500 miles out to the west of the Hebrides, peaking at just over 14°C in early September and dropping to just over 9°C by March.

Figure 2

This chart is of the instrumental wave height measured by the K7 buoy, the vertical pink bands on the graph are for the named storms of last Winter (fig 3). I’ve plotted this 6 hourly, which still means processing almost 1500 files. The maximum wave height measured by the K7 buoy, was 11.7 metres (38 feet) at 18 UTC on the 23rd of December 2016 during Storm Barbara.

Figure 3

Closer to land the range in the SST is much higher, here is the graph for the Sandettie Lightship in the eastern end of the English Channel (fig 4). SST range from a high in September of over 19°C to a very early low of just over 6°C in the last week of January.

Figure 4


The application can access and visualise weather data from any weather buoy in the world, and that includes not only all the worlds oceans, but also the Great Lakes as well (fig 5).

Figure 5

As far as I know all weather buoys deployed on the Great Lakes are taken in for the Winter during November because the lakes can ice up, and put out again at the beginning of June. An amazing amount of warming goes into the Great Lakes in just a few short months, I’m sure people must have thought of turning them into some kind of huge heat pump to generate electricity with. These are the SSTs, or should that be the LSTs, from weather buoy 45006 in Lake Superior, ENE of Duluth (fig 6).

Figure 6

Occasionally a weather buoy will get in the way of the occasional hurricane, as was the case with 42058 last October with Hurricane Matthew (fig 7)

Figure 7

Weather buoys are also very handy for monitoring ENSO conditions in the central Pacific. Having said that, I think most of the ones strung out along the equator don’t report a SYNOP, the closest buoy with SST that I could find was 51004, southeast of Hawaii (fig 8) but 17.6° north of the equator, and seems to show the final stages of the above average SST from the recent El Niño event which ended last year.

Figure 8

The UK weather buoy network is far from perfect as I pointed out in an article I wrote in January, even at the best of times the coverage to the west is quite thin, the Americans in comparison make extensive use of weather buoys to monitor their coastline from the threat of hurricanes, and seem to freely dot weather buoys across the Caribbean and Pacific.

The state of the Met Office’s old buoy network

Figure 1 – Courtesy of the Met Office

As you all must well know by now, I love plotting weather charts, and regularly download and plot SYNOP observations from around the world. The Marine weather buoy network around the UK is a God send, because although over the years as the number of reports from ships continued to dwindle, they have always been there, around the UK, to give a heads up to depressions that may be developing unexpectedly rapidly of our western coasts. They provide an hourly observation that contains details of wind, pressure, temperature, dewpoint, sea surface temperature and wave heights and period of the swell. Weather buoys are not a new idea, and the Germans used them in the second world war in the Atlantic, in the Wikipedia article about them they estimate that there are 1250 drifting, as opposed to tethered, weather buoys deployed and active across the World’s oceans.

O Romeo, Romeo, wherefore art thou Romeo?

I believe that the existing marine buoy network may have sprung into life after the demise of the weather ship network, although I can’t be sure. Weather ships were never going to be a cheap idea, but by 1965 there were 21 scattered across the oceans of the world (fig 2 & 3). That slowly reduced, and by January 1982 the Met Office brought home to port the last two frigates that we had.  The last weather ship to be withdrawn from service was station Mike, run by the Norwegian Meteorological Institute on January 1, 2010. As an aside at this point, there is a great website that pays tribute to the men who manned our Ocean Weather Ships during that time which I encourage you to peruse.

Figure 2 – Courtesy of Wikipedia

Figure 3 – Courtesy of Wikipedia

As far as I can see from scanning the SYNOP data that I have, the Met Office hadn’t replaced weather ships with buoys by the autumn of 1987. So in the autopsy of the storm in October 1987, the fact that Ocean Weather Ship Romeo (I’m sure that we always used to plot it as OWS Kilo) had been removed (it was actually a weather ship run by the French), combined with the lack of reports from any commercial ships was blamed for the failure to locate the centre and intensity of the embryonic low on the 15th that was to cause all the trouble on the 16th (fig 4).

Figure 4 – Courtesy of Google Books and the New Scientist

The Met Office on their website say in the aftermath of the storm that they had increased “the quality and quantity of weather observations from [buoys]” (fig 7), which when you think about it wasn’t that difficult do since there were no weather buoys before. Here (fig 5 & 6) are a couple of plotted charts from 12 & 18 UTC on the 15th of October 1987. The charts aren’t totally devoid of any ship reports, but of course the question is where they available on that date?

Figure 5

Figure 6

Figure 7 – Courtesy of the Met Office

The Current state of play

In the last few years that network of weather buoys has become a little thin, and the buoys that I’ve marked with a red circle (fig 8) have been missing for a considerable length of time, and as you can see they are in a fairly strategic position of the southwest coast of Ireland. So the network is compromised, it’s not totally without observations from buoys, K1 and Brittany are still present, but they do become faulty at times, and sometimes they break away from their sea anchors and go adrift, so having six missing designated weather buoys M1, M3, M6, K2, Pap, and Celtic Sea is far from ideal. It can’t imagine that these buoys I’ve highlighted have never been active, otherwise why place a marker on the website for each of them? They certainly are not available at the moment either on the website or from OGIMET where I download my SYNOP observations. Perhaps they are private observations and not available to the public? That might be a possibility, but surely they would simply not display them. You can check for yourself the latest observations from any of the buoys on the Met Office website.

Figure 8 – Marine Buoy Network – Courtesy of the Met Office

Here’s the network from back in 2011 and looking in a bit better shape than it does today (fig 9).

Figure 9

Lesson Learnt?

Could we see the lack of observations from these appointed sentinels let another rapidly deepening low catch us out? Well, as John Hoghton admitted in that New Scientist article, ‘There is that danger we are caught in that gap‘. So possibly it could, especially if that approaching low was positioned just to the west of Valentia, Southwest Ireland, and was tracking eastward between o60° and 120°. I’ve sat through a couple of wash-up meetings after a project that I’ve been involved had finally finished, and talked about the section that usually concludes any project of – ‘lessons learnt‘. I should imagine that we did learn the lessons after the storm of October 1987, and the equally fierce one that occurred in January 1990. But maybe because that was 25 years or more ago, the people who remember that have now moved on or retired, and now we have to find a bit more to pay for that massive Cray super computer we’ve just bought. I’m just guessing of course, and maybe the Met Office are planning to overhaul their marine buoy network in 2017, I don’t know. But if they aren’t planning to do it sometime soon, then more fool them.