–––––––––
Modern weather work underway relies on “grib data.” This is sailor’s jargon for surface analyses and various forecasts presented in a gridded binary format called “grib.” These are vector data given for each point on a Lat-Lon grid. High resolution data can be as fine as 0.02º (1.3 nmi) between points, on up to 1.0º (60 nmi) between points. A grid spacing of 0.5º often meets ocean route planning needs, but but for computed optimum global routing 0.25º is common. For inland work we need the best resolution possible.
Wind speed, wind direction, and sea level pressure are the primary datasets needed for planning, but also available and often useful include the height of the 500 mb surface and wind speed at that level, sea surface temperature, precipitation (rate or accumulated), wind gusts, and simulated radar for squall forecasting. There are also dozens of other atmospheric parameters that could be useful in special circumstances such as humidity at 850 mb level to help identify frontal boundaries, or wind at that level to anticipate squall motions.
Most of this grib data now in use are the direct computer output of numerical weather prediction models, the most popular of which is the US Global Forecast System (GFS) and as of 2023 we can also get surface wind from the ECMWF model. Thus we are looking at the same data the professionals use to make their official forecasts. The professional forecasters at the NWS, however, also look at other models we might not have access to. They also have more varied ways to look at the GFS data itself.
Thus it is our obligation underway to check the grib data we use, i.e. compare pure GFS model forecasts with the final manual forecasts of the NWS. These two forecasts (GFS and NWS) will be very close more often than very different, but for crucial decisions, or for fine tuning routes as when racing, we must rely on the NWS forecasts, which brings us to one of the main points of this note.
There is increasing awareness and availability of the National Digital Forecast Database (NDFD) in grib format. Unlike all of the other grib datasets, this one is actually created by professional meteorologists who are looking at all model predictions before they create the NDFD forecast. The NDFD grib data are the same we see in the NWS graphic maps. The first entry in the NDFD grib set (h00) is the surface analysis, h24 is the 24-hr forecast, h48 the 48-hr forecast and so on. The difference now is we have these every 3 hours, not just as the traditional synoptic valid times shown on the graphic maps. Unfortunately, the global coverage of the NDFD is limited, but for extended US coastal waters we have good coverage (called NDFD oceanic). In the near future we will get easier access to the National Blend of Models, NBM, which has high resolution and covers most of the Atlantic and Pacific. See Modern Marine Weather.
***You can get the NDFD oceanic grib data with an email request to saildocs (saildocs.com), and now we are pleased to see this appear in commercial products like WeatherNet for PC from Ocens (ocens.com). They offer both standard and high resolution NDFD data in the traditional convenience of the WeatherNet interface. Resolution becomes a key factor underway when we are downloading by sat phone or HF radio. The high resolution of NDFD (0.12º) is not needed for open ocean sailing and it can make the files very large. The high res data are best reserved for forecasts near tropical systems or along coastal waters where you might want to detect diurnal changes in the wind closer to shore or within bays.
The other exciting news from Ocens is they offer the ASCAT scatterometer data in grib format through their WeatherNet service. This wind data is not a forecast; it is the true wind speed and direction measured on the ocean surface as the ASCAT satellite passed over it—normalized to a height of 10m, the same as the GFS winds are. We only get this unique data about once a day for any one location, but it is a definitive evaluation of a forecast at the time, which we need in crucial decisions.
It is not a surprise that Ocens would be the first to offer this ASCAT grib data as they were indeed the pioneer in this field. In collaboration with the NWS, they offered the first ever grib scatterometer data from the QuickSCAT instrument back when that scatterometer was working. Unfortunately, it stopped transmitting the type of wind data we need in 2009, which was rather before many mariners got to know the power of the data source. So scatterometer grib data is back, and we hope to see more of it. In the interim, the Indian Met office launched the OSCAT instrument, which provided very good data for a while, but it too has failed.
The latest scatterometer (past) news is the US RapidSCAT instrument on the International Space Station. For a while it provided a broad swath of data (as QuikSCAT and OSCAT did) but it too has been taken offline. We can, however, count on the ASCAT instruments (on satellites MetopA and MetopB), which are part of the European Space Agency. You can see and access graphic images of all scatterometer data at Ocean Surface Winds Team website (google ASCAT). And this data is available in GRIB format from more than just Ocens. The LuckGrib apps (Mac and iOS) have beautiful presentations of ASCAT data and we can get this data directly from NOAA ourselves (see Modern Marine Weather for details.)
Another new development in grib viewing is a low-cost, high-quality grib viewer and data source for Mac and iOS is called LuckGrib (luckgrib.com/blog). It was developed by Pacific NW sailor Craig McPheeters, created in large part while underway in the South Pacific. LuckGrib has an elegant design and convenient interface, with several unique features added to the standard functionality we expect from a high-quality weather tool. Luckgrib has the most extensive source of weather models of any grib app. It also includes a sailboat weather routing option. At present LuckGrib and Expedition are the only sources of ASCAT data in grib format.
For more background on the important role of grib viewing software and the acquisition of grib files underway in Mariner's Checklist Before Departure.
There are many grib viewers for computers and tablets with as many different styles of data presentation and unique features, so I have focused here on viewers that can show NDFD grib data, regardless of where the data files were obtained. It is not uncommon to download a file from one source or viewer and then display it in another viewer. We hope to eventually see more sources offering the NDFD grib files, because these are likely to be the best global data available for the waters it covers. It will take a lot more study to decide if a regional model such as COAMPS or NAM might be better for near coastal waters in some cases. Discussions that favor these over GFS (an almost certainly a valid conclusion near shore) may not have made the same comparison with NDFD.
Special features of several grib viewers are shown in the graphics below.
(The once popular ugrib viewer also had a meteogram function, but it did not show NDFD data and it did not have any route planning features. As of mid 2016 or so, Ugrib is no longer available.)
Meteogram display from ugrib. |
Figure 5. Coastal Explorer from rosepointnav.com is a popular echart program that also shows grib weather files obtained from other sources, and indeed it will display the NDFD data. Shown here is one easy way for a quick overview of weather conditions. We set multiple range rings on the h00 vessel position at 50-nmi intervals (6h run at 8.3 kts), then we can step the forecast forward with an estimate of where we would be at that time. This program and the others illustrated here will also show ocean currents to help with route planning. The RTOFS or OSCAR ocean current data are readily available in grib format by email request underway.
OpenCPN for Navigation and Weather Work
Figure 6. Here is an example of the OpenCPN weather routing plugin finding a good route through the Gulf Stream. The light blue-green is the strongest current.
RapidSCAT was relatively short-lived. It turned out to be too complex having it attached to the ISS. It provided data from September 2014 and operated until August 2016, but is not planned to come online again.
ReplyDelete