There is a famous saying with computer software GIGO – garbage in garbage out.
So this review of software based routing calculations starts with the data that is being input.
There are 5 main sources of “data”:
- The expert assessments and factors stored in the software by the authors, and their expert consultants such as Michel Desjoyeaux
- The instruments on the boat itself feeding data in real – time
- The behaviour of the boat – its specific sailing characteristics – especially as represented by a polar diagram(s)
- The adjustments made by the skipper/navigator based on experience, or experiment.
- Data from external sources especially GRIB data.
It is this GRIB data that I will look at first.
Nearly all ECS / ECDIS systems no incorporate a facility to display weather data – at least as an overlay to the chart area in view. The suppliers either have facilities to download GRIB files from some of the major services such as:
- Grib.US which also provides a free viewer for the data called UGrib
- Global Marine Networks – 7 day GRIB forecasts for the entire world. The GRIB files are generated at 0015, 0615, 1215, 1815 GMT everyday.
- NOAA MMAB – Incorporating their Operational Wave Model (wwave3)
Or, they take this data and re-format; filter; transform; compress the data into suitable form for their software and distribute it using their own owners networks; clubs; fora and so on – MaxSea Chopper is one such service.
For the purposes of this review of routing software there are some issues you should note when checking out a suppliers claims about GRIB files:
a) – There is plenty of weather data available, some of it pretty impressive – BUT, forecasts are forecasts – and there are clear warnings on all the links above that tell the sailor to listen to official weather forecasts and to use their own local experience and reading of the weather actually where they are. You should never go to sea without receiving and studying forecasts issued under the GMDSS. They have the benefit of experienced human interpretation. Amendments to GMDSS forecasts are broadcast on radio, NAVTEX and INMARSAT-C. There is no equivalent broadcast of GRIB amendment when at sea
b) – Not all GRIB file data is incorporated into routing calculations. Some systems will read only the wind speed and direction, others may also incorporate wave height and wave direction for example. GRIB data can include all of the following:
- Sea Temp
- Wave height
- Wave direction
- Wind Speed and Barometric Pressure
These data are from NOAA’s GFS model. The Global Forecast System (GFS) is a global numerical weather prediction model. The model is run 4 times a day and produces forecasts up to 16 days in advance. Data are generally available on a 0.5 x 0.5 degree grid out to 180 hours and then on a 2.5×2.5 degree grid out to 16 days.
- 500mb Wind and Height
Wind and height of 500mb pressure (about 18,000 feet) from the GFS model.
- COAMPS Detailed Regional Model
A more detailed regional model of wind and pressure provided by the U.S. Navy. Data are available for three regions: Eastern Pacific, Western Atlantic, and Caribbean/Central America. Forecast data are available on a 0.2 x 0.2 degree grid at 6 hour intervals out to 48 hours (72 hours for the Western-Atlantic). For best results use a grid of 0.2 degrees or multiples of that to avoid interpolation.
- Sea Temperature
Sea temperature from the GFS model.
- Air Temperature
Air temperature from the GFS model.
- WW3 NOAA’s Wave Model
This is NOAA’s WW3 wave model. Data are available on a 1 x 1.25 degree grid (Lat x Lon). Forecasts are made for every 3 hours out to 180 hours. The forecasts are updated every 6 hours. The data reported in Fugawi are the Significant Wave Height (average height of the highest third of waves), the Peak Wave Direction and the Peak Wave Period.
- QSCAT Satellite Wind Observations
Current wind conditions calculated from radar observations. The data are available on a 0.33 x 0.33 degree grid. The data are updated every few hours, but some of the data points may be up to a day old. The age of the data is reported in Fugawi Marine ENC. If the radar observes rain, the wind data is not shown.
- RTOFS Atlantic Current Model
The Real-time Ocean Forecast System (RTOFS) model operated by NOAA provides high-resolution forecasts for currents and sea temperature. Data are available on a 0.25 x 0.25 degree grid for the entire Atlantic and a 0.05 x 0.06 degree grid for the Gulf Stream area. There is one model run per day. Forecasts are provided at 12 hour intervals out to 5 days for the Atlantic data and at 24 hour intervals for the Gulf Stream. A fine grid of 0.2 degrees or better is necessary to see the Gulf Stream.
c) – Not all GRIB file data will be at the same resolution i.e. the level of detail can vary. You might want a low resolution of detail if you are planning an Ocean voyage for example as opposed to a crossing of the English Channel (Medium) or a race in the Solent (High).
“Low Resolution” would be a native resolution of 1.0 degrees (100km x 100km) GFS Model with 3 or 6 hr time steps
“Medium resolution” would be a native resolution of 0.5 degrees (50km x 50km) GFS Model with 3 or 6 hr time steps
“High resolution” would be a native resolution of 0.1 degrees (2km x 2km) WRF-NMM Model with 1, 3 or 6 hr time steps
Some systems allow the explicit selection of the level of detail others do not.
d) – Finally connectivity… if you are on a long voyage you would need to get access to email or other services to be able to download your up to date GRIB data. You have to think this through. As Will Oxley says “..The first thing is to ensure that you have a reliable onboard internet link. For races along the Australian east coast this is best achieved by a Telstra Next-G connection. Teething problems in the change from CDMA to Next-G have been well publicised, but you should now be able to get a good connection up to 40nms offshore with a good setup. I have had the most success with a professionally installed Ericcson W-25 fixed wireless terminal and a decent marine antenna. Further offshore you will need another solution to continue to access the internet…”
B – Wind
- Apparent wind speed is the wind velocity that is measured directly across the deck
- of the boat—it is the wind you feel on your face.
- Apparent wind angle is the angle of the wind that is blowing across the deck
of the boat relative to the boat’s centre-line.
The boat sails in the apparent wind not the true wind. But, The true wind angle and speed are amongst the most useful numbers on the instrument system; they tell you what the wind is doing independent of the boat’s own motion, and they are calculated from the numbers in the wind triangle that we can measure: the boat speed, apparent wind speed and angle. This calculation is the most important thing your instrument system does and it is the key to all the second stage functions.
The apparent wind angle and apparent wind speed are usually measured by the masthead unit, and they can be viewed as the vector sum of three different components of wind:
- The ground wind which is the wind blowing across the land—the wind you will see on weather forecasts and feel when you stand still on land – Note that the forecast wind direction in the UK shipping forecast is the direction expected at 10m above sea level
- The tide wind which is produced by the motion of the water relative to the land. It is equal in strength and opposite in direction to the water flow
- The motion wind which is produced by the motion of the boat relative to the water. It is equal in velocity to the boat speed and blows directly onto the bow
Most commonly the true wind speed will read higher downwind than upwind— or the true wind direction will be different on starboard tack to port—without there being any change in the real wind. The reason for the true wind direction being wrong is error in the apparent wind angle.
There are three causes of this error. The first is that the apparent wind angle you feel and measure on the boat is distorted by the flow of air over the sails. We can most easily understand this by imagining a motor boat driving parallel to, and a hundred yards away from, a sail boat. The sail boat would measure a different apparent wind to the motor boat because of the distortion of the wind flow around the sails. But the apparent wind angle we need in the Wind Triangle calculation is the one measured by the motor boat. This error is usually called the upwash effect.
The second problem is that the tip of the mast will twist because of the loads imposed by the sails. Although you may think the masthead unit is lined up with the centreline because the apparent wind angle is the same on both tacks, in fact it is pulled off a few degrees both sides by the mast tip twisting when you are sailing.
The third problem is that, upwind at least, the measurements are taken with the masthead unit heeled into the airflow.
Pretty much all of mainstream software packages reviewed incorporated wind speed and direction data. But only racing packages like Deckman and Expedition try to calibrate to counter these impacts – it also helps if you have B&G instruments of course!
You can go even further to try and represent wind twist or wind sheer. This is of interest to the racing community only, since most cruising boats will have a lot more problems to sort out to do with their rig and sails before the worry about this level of detail.
There is an excellent article here at OceanSail about the effects of the wind on a sailing yacht. As you can see in reality you would need considerably more data than just wind speed and direction if you were racing. Some packages such as Expedition, allow you to enter data manually as well as automatically, to try and fine tune the wind data with your own experience on the spot as you sail or for your boat characteristics in general. Clearly most cruising sailors don’t have the instruments or the knowledge to be able to capture this data.
C – Pressure
Again, pretty much all of the mainstream software incorporated overlays of barometric pressure charts. This is useful, since even a slight knowledge of weather forecasting can help you to see the gradient between isobars and the direction of wind and the direction of cyclonic or anti-cyclonic weather formations. This would be critical to appreciate the severity of a storm and also its general direction if the software allows you to page forward or animate the trends over time.
D – Tidal streams / currents
The National Oceanographic Data Centre shows the work that is being done to track the surface temperature of the worlds Oceans. Using this data as well as the data from the data buoy network – pretty accurate predictions of Ocean currents can be made and these in turn impact on wind speed and sheer.
Some of the systems reviewed used tidal flow data – same as the tidal diamond data on a normal chart – and took the nearest tidal diamond data to any particular point, in order to assess an applicable tidal vector and speed. Some like Advantage used more detailed data that the supplier had generated for specific coastal areas – for example well known race areas such as SF bay or the Solent or the Sydney/Hobart Race.
Seatrack claims to have the most accurate tides for the racing area of the Solent for example…”Winning Tides is the most accurate and detailed tidal database available for the Solent and adjacent waters. It provides much greater detail than standard Admiralty charts and conventional tidal atlases…”
There is s graphic illustration of the impact of routing software upon decision making in the 2009 Sydney-Hobart race which many say was won or lost on the first night…here is an extract from the Sail World site…
“….Fickle winds and complex eddies in the East Australian Current mean that the 2009 Hobart Race will go down in the history books as a navigator’s nightmare.
As the fleet departed Sydney Heads, the forecast showed a moderate Southerly breeze, backing slowly overnight to the South East. This meant that pretty much everyone headed offshore on starboard tack in anticipation of the shift. But when to tack? Those who had Tidetech’s current data from the CSIRO for the race knew that they would have to ‘bang the corner’ and go as much as 70 miles offshore to get into the best of the South flowing current. But as the night drew on and boats on the inside started to tack away, those who had elected to stay on starboard were feeling decidedly nervous. In the end it was the ones that stayed out wide, trusted the information and most importantly, held their nerve that came out on top. ….”
Local Knowledge – as used in the Advantage Software also claims to have and edge due to its hydrodynamic current prediction model and its proprietary current data for certain key sailing areas e.g. UK Region ….”Other programs display currents at isolated points, taken from NOAA publications which are often outdated. There is no way to interpolate between these points, so they are useless from the standpoint of tactical computations. We build computer models which compute current at EVERY POINT, so you can customize the display by adding display points anywhere you like and perform calculations which require continuous, point-by-point coverage…“
The major systems shown in TableA all have specific tidal databases available see Deckman Tidal Data for example.
E – Sea temperature
Why is this important – apart from fishing! – It is used to monitor Ocean currents. The National Data Buoy Network also produces data and ouputs it onto Google maps to make it easier to understand – at least for the North American Continent anyway. Cross-correlation analyses of successive satellite imagery of the ocean surface can provide estimates of surface currents by tracking features of temperature or ocean colour.
F – Wave height / direction / period
The data available is well represented at the the many surfing sites like Windfinder site or Surforecasts.com. A major source of data is the NOAA Wave Watch project.
There is a very good article on the Jordan Series Drogue site that gives an insight into the mathematics of wave behaviour. For non breaking waves this maths is suitable for incorporationg into routing software to fine tune the output even further.
A real life example….
Pictured is Josh Warren-White talking of his experience of using GRIB file data on an ocean passage…“While on passage, we receive daily weather grib files via email. Thanks to a simple program created by the good folks at Saildocs, any voyager with offshore email capabilities can access free weather data by sending requests to Saildocs (a similar program is also available from MailASail.com). All one needs to do is email Saildocs with simple text commands, which automatically uploads certain files from the Saildocs server that are then provided to the user in an automatic email response...After sending the query and receiving our grib file via email, our next step is reading it. We often use the software program MaxSea to navigate on our laptop computer. Our Raymarine GPS provides the program with data via a serial cable, so in MaxSea we already have our current position, course, hull speed and route plotted on our navigational charts. We simply take the grib files we received via email, import them into MaxSea and overlay them on top of our navigational charts. This gives us a picture of what the NOAA GFS forecast is for our projected position over the following hours and days. There are also numerous free grib-reading software programs available online…”
This excellent article from Compusail gives you an insight into the possible differences between a leisure boat and a race boat’s instruments. This company specialises in sail boat instruments and it is of note what a difference the quality of instruments and their siting on the boat can make – well worth flicking through this article.
It is not unusual on a normal amateur boat for instruments to flick on and off or just not work at some crucial moment – for no apparent reason. The impact on auto routing software is obvious…
With all this data you would be right in thinking the devil is truly in the detail – remember GIGO. However, there is such a thing as spurious accuracy too. In other words the tendency to assume that the output of a calculation must be all the better for the quantity and apparent accuracy of the data going in. While inputting the best available data is obviously critical – it does not in itself guarantee the best result.
A software system that does not input all the data above is not necessarily inferior to one that does. On the other hand as I said at the outset of this journey, for a leisure sailor there is a huge educational aspect to using and understanding this data.
Simply having access to a suitable computer program to manipulate all this data easily, efficiently, and repeatedly is a huge advantage.
…watch this space for further episodes in the quest for optimal routing calculations …