About GRIBS and forecasting
>> Reliability of the GRIBs: GRIB files are raw data plucked from national oceanic and Atmospheric Administration databases without human interpretation. As unedited raw data, they require analysis by the user. though the models have improved in the decade that we’ve been using them, the forecasts diverge from reality at least half the time in the periods between 72 and 96 hours. We heavily weight the first 48 hours and tend to discount the last 48 hours unless the systems shown are large and well defined.
>> North/south passages vs. east/west passages: We’ve found the GRIBs to be most useful on passages in the temperate and high latitudes that have a large north/south component, like the one described in this article, and less useful on downwind passages made within a degree or so of the same latitude. that’s because when running down the latitude, large systems overtake the vessel from behind with generally following winds, making it less necessary and less useful to maneuver a bit north or south as a system approaches. When sailing north/south, the crew has many more options and can choose a course 90 degrees from the route while still making progress toward the destination, which is enough to maneuver with respect to smaller systems. Also, headwinds are much more likely on these north/south routes, and small changes in position, as in the final example in this article, can position the vessel to avoid them.
>> Wind speeds: The GRIBs forecast average sustained winds, which means that gusts can easily be 20 percent to 30 percent higher. In gales and storms, we tend to set the boat up to handle the gusts, so when looking at wind speeds around storm systems, we add 5 to 10 knots to the forecast wind speed when deciding what sails to carry or how closely we should approach the low center.
>> Small, compact systems: The GRIBs shown in this article use a reporting area based on a latitude/longitude grid with 5 degrees on an edge. these can fail to show small, compact, and intense systems such as meteorological bombs or tropical depressions. GRIBs can be downloaded with a resolution of 1 degree on an edge but even then may not capture these dangerous systems. Throughout this passage, we were also downloading text weather forecasts off our Inmarsat-C for the next 24 hours, and those would’ve warned us of something small but intense that the GRIBs might miss.
>> Timing of weather systems: When the GRIBs get things wrong in the first 24 hours, it’s almost always because systems have moved faster or slower than forecast or have moved slightly north or south of their forecast track. This was the case with our October 5 forecast, when the low-pressure system came right over the top of us instead of staying south. By tracking changes in the wind direction, cloud cover, and barometric pressure, we can tell where we are with respect to an approaching system, and we can tell when the system passes over us — even if it’s significantly before or after the GRIBs had forecast. This is one of the main reasons why we download the barometric pressure as part of the GRIBs, though it makes for a larger file and costs more to download. With this information, we can easily tell how close we are to the center of a system and how quickly it’s moving using our barometer.
>> “Mushy” weather patterns: Strong, well-defined systems tend to develop as forecast. But when many weak, poorly defined systems and slack areas of little wind cover the forecasting area, we don’t place much confidence in the forecast. In our experience, one system will almost always begin to dominate and drive things in a different direction than the GRIBs are forecasting. In these situations, we often download GRIBS every 12 hours to see if anything is developing.
>> Margin of error: Part of our decision-making is based on what will give us the greatest margin of error and keep the most options open should the end of the GRIB forecast period be way off. When passing close to the center of a low-pressure system, we tend to aim for a band of 20-knot winds so that if the system moves closer to us, we’ll still be out of the strongest winds. In my last example, we chose to stay on the back of the low in case another low should follow in its track — which is exactly what happened. If we hadn’t done that, we would’ve been pinned on the Chilean coast in a strong southerly flow facing a major beat against wind and current to reach the Canal de Chacao.