Evaluating weather maps
Weather work at sea typically involves three components: knowing what the resources are and how to get them onto your vessel, evaluating the dependability of the data given to you, and finally, using this data to make decisions. When it comes to doing the best job possible, all three parts are equally important. We’ll address the middle step (evaluating the weather map), which is too often taken for granted.
Here is the crux of the matter: There will always be a forecast, and they are not marked good or bad. On land we get 30 percent chance of rain, but at sea we do not get 30 percent chance of gale. We get gale, or no gale. Probabilistic forecasting has not yet reached into the realm of marine weather as we receive it while underway. A notable exception is the wind forecasts for tropical storms from the National Hurricane Center, but we need Internet access to get this wonderful resource underway. For more routine extratropical weather analysis, it is up to us to evaluate the reliability of official forecasts we receive and act accordingly.
The most basic test of any weather map in hand is to ask if it reproduces the actual conditions we had? Notice this is past tense. The most recent surface analysis map is at least two hours old the moment we download it or receive it by radio facsimile. These maps are valid for the synoptic times of 00z, 06z, 12z, and 18z (z is shorthand for UTC), but they are not available underway until some two or three hours later. Thus the first key point in checking the weather maps is to be sure we record all pertinent weather data at the synoptic times.
For weather analysis and planning, we need a logbook entry at all synoptic times of the true wind speed and direction plus the barometer reading. Clouds, sea state, temperature, and a note on precipitation and visibility can be useful, but wind and pressure are the key factors. We also need a GPS position at the synoptic times, along with course and speed. In short, even if the watch is not changing, nor any thing else notable occurs, we need a full logbook entry at 00z, 06z, 12z, and 18z.
Comparing surface analyses to ground truth
Ultimately we are going to make our decisions from the forecast maps, not surface analyses, but the forecasts are based on conditions represented in the latest surface analysis, so it is mandatory to make the surface analysis comparisons to actual observations as carefully as possible. If the maps report exactly what we observed, we can have more confidence in the forecasts based on them. If they do not agree, then we must be more careful with decisions we base upon them.
Start by plotting your valid-map-time position on the surface analysis map as carefully as possible, which is not trivial, because weather maps have meridians and parallels marked only every 10°. Some interpolation technique is called for. If your maps are on a computer there are various digital solutions.
Once your position is plotted, interpolate the isobars to find the predicted pressure for your location. This can typically be estimated to within a few tenths of a millibar (mb). This pressure can then be compared to your logbook pressure for that time. Needless to say, it is crucial that you have a calibrated barometer on your boat.
If you have not yet checked your barometer you can do so at www.starpath. com/barometers. This site gives you the 10 closest places to your Lat/Long that offer accurate pressure online at the moment, including the range and bearing to each. It also gives your ground-level elevation to within a fraction of a foot. It works worldwide.
If you get underway without having made this crucial test, then you can still learn about your barometer underway by sending an e-mail to firstname.lastname@example.org with the word “help” in the subject line. The return e-mail explains how to obtain by e-mail all ship reports of the Voluntary Observing Ship program within 300 nm of your location over the past six hours. This gives you wind, pressure, and sea state, and the range and bearing to each ship. We are not guaranteed the ships all have accurate barometers, but if you compare what they report for a while, you will zero in on the accuracy of your instrument.
Once we have the pressure comparison, we can do the wind comparison, and then figure what to do with the results. The wind is not quite as easy as the pressure, but it also, of course, requires calibrated wind instruments. Most sailors have this under control, even if the barometer has not been checked. We do, though, need true wind directions. If you recorded only apparent wind, then step one is to take your recorded apparent wind speed and direction and convert this to a true wind direction and speed. This is why you need your speed and heading at the synoptic times.
Getting the wind from the map could be easy, if there happens to be a nice wind arrow right at your location at map time. But this is unlikely. You make it rather more likely by also downloading the Wind and Wave Analysis maps, but these special maps are not very dependable. They are only offered twice a day and the coverage is sketchy.
Note, too, that you cannot use digital GRIB data for this comparison, no matter how tempting it is. The 00 GRIB forecast is a rough approximation of the corresponding surface analysis, but it will always be one map cycle late in real time. An 18z GRIB map, for example, is a 6-hour forecast based on the 12z base analysis. In short, you cannot get the latest surface analysis in GRIB format — not to mention that they are not surface analyses to begin with, but unvetted numerical model output.
Figuring winds from isobars
U The best solution to figuring winds from surface analysis maps is to do it yourself from the isobars. The process is fundamental to weather map reading, and its value comes up time and again. The spacing of the isobars determines the predicted wind speed; the orientation of the isobars determines the wind direction.
Now in principle we have in hand our wind and pressure observations and the corresponding surface analysis depictions of these. The next question is: How close should they be to be in agreement? Both our measurements and our interpretation of the maps have some uncertainty.
A good onboard barometer should be able to tell you the pressure to within 1 mb. A very good one could do 0.5 mb. We can usually interpolate the plotted isobars to within about 0.5 mb. Ship report data are generally accurate to within about 1 mb. Thus we have to consider pressures that differ by just 1 to 1.5 mb to be essentially equal. Pressures that differ by 2 mb or more are reflecting actual disagreements. A difference of 4 mb is a full isobar. That is a huge difference. Again, however, you will not see this, nor be able to use these crucial observations if your barometer is not calibrated. It is not at all uncommon to have a typical (uncalibrated) barometer be wrong by more than 4 mb.
If you are judging wind speed from a weather map by just looking at wind arrows, then you have an inherent uncertainty of about 2.5 knots. This is part of the reason it is best to figure the wind yourself from the isobars rather than use wind arrows, even when they are present. Two feathers means 17.5 to 22.5 knots, or 20 plus or minus 2.5 knots, and likewise for all wind arrows. Some European meteorological offices assume winds can only be known to within one Beaufort force number, and thus they do not give reports or forecasts any more accurate than that. But since the actual force of the wind is proportional to the square of the wind speed, we would like to do better than that if possible, and often we can.
Winds we record from our own instruments depend on the calibration of the instruments and our care in recording the data. We have to make some mental average of the right wind speed to record in the logbook. Chances are this will be within 10 to 15 percent correct. For example, when you record 10 knots, are you confident it is between 8.5 and 11.5, etc. We can fairly consider anything within 15 percent as essentially in agreement.
Wind direction can be interpreted from isobars to within a few degrees in many cases. They are not that accurate, for sure, but that is not the point at hand; that is what we are trying to find out. Our own wind direction accuracy again depends on the care we put into the measurement, our records of it, and our conversion to true from apparent. There is no reason this could not be within some 10° if we are careful. In other words, when you record the wind direction as 270, are you confident it is between 260 and 280? Then of course the conversion to true wind must be done correctly, which calls for good average speeds and headings in the logbook. In short, wind directions agree if they are within some 10° of each other for straight isobars, and maybe up to twice that for curved isobars. We can use 15° as a working guide.
We get surface analysis maps every six hours, so the typical procedure is to make the comparison, guess what might be the cause of any discrepancy, guess how this might affect the forecasts, make an appropriate tactical decision, then in 6 hours check to see how things are evolving.
Along the perimeter of a High, an observed pressure higher than mapped, means the High has moved more toward you than mapped, or the High has built to higher pressure. If your recorded wind speed is higher than mapped, it means the true isobars are closer together than the mapped ones. If the wind direction is different than mapped it means the isobars are not oriented as mapped. Then comes the guessing game, depending on what is going on around you.
A Low on the other side of you that is deepening or moving toward you could rotate the isobars at your location, but this would generally increase the wind as well. Both could happen without the actual pressure changing at your location. If the High is just closer to you than plotted, then chances are the wind speed and direction are about the same, but just the pressure is wrong.
The first step toward understanding a discrepancy is to identify Highs, Lows, and fronts around you on the map, and vary their speeds of motion to learn if that accounts for discrepancies you see. Generally the weather service has a good idea of what is out there, but they may be off some on the timing. Furthermore, if a Low slows down, the National Weather Service can’t tell us that until the next map comes out.
You can sometimes discover a change in a Low’s speed, or that of a High, by comparing your measured pressure tendency with that predicted on the maps. The Lows and Highs are marked at their valid map time positions, as well as at their forecasted positions, 24 hours later. This can be used to compute their predicted speeds, and from this you can compute the expected rate of pressure drop at your location. Comparing this pressure change over time with what you observed can tell you about the speed of the system.
If you have doubts about what is taking place based on the latest map, then on the tactical side you might consider doing just half of what you want to do until you can see the next map.
If you want to gybe now, for example, if the map is right, but this would be a mistake if the map is wrong, then you might run another three hours (halfway until the next map) and then gybe. If you are reaching and want to change course by 60° to meet some favorable winds ahead if the map is right, but this would be a waste if the map is wrong, then consider turning just 30° until the next map arrives.
“Do half of what you want to do” is obviously a figure of speech that has to be adapted to the conditions. If you are fully confident in the map, then you can base your maneuvers on it, but if you are less certain, then compromise is called for.
More ship reports
Confidence in the map and forecasts comes not just from a comparison of past performance as outlined above. We can also use other criteria. A simple one is to observe the number of ship reports in your region. If there are a lot, then chances are the data is better than areas were there are few or none. Sailors often care about the shape of the High along a corner as they go around it. This is likely to be a region of few reports, and because the air is light in these regions, the reports might not be as good as others. You may want to see the corner in two successive maps to believe the shape is right.
Also note on the surface analysis maps that ship reports can definitely influence the lay of the isobars. Do not hesitate to check the spacing to be sure the report is consistent with the isobars and consider the implications when they are not.
We can also make reasonable judgments based on the shape of the High we might be sailing around. If the High is indeed high (>1,030 mb) and it is located in its climatic average home for the season, and it has at least two round isobars around it, then you can feel confident that the map tomorrow is going to look much like today, as the forecasts will confirm. This is a blocking High, likely to be tucked under an Omega block in the winds aloft. Once any one of these conditions change, the map will look progressively different and less dependable.
Finally, we can also get more generic insight into the dependability of the forecasts based on the winds aloft, meaning the winds seen on the 500-mb level. Some 500-mb wind patterns lead to more dependable surface forecasts than others. When the winds aloft are flowing in smooth, broad waves across the ocean we can be more confident that the surface forecasts are dependable. These winds are flowing around the natural climatic action centers on the surface, and the atmospheric models are more accustomed to the influences they have on surface systems.
When these winds are flowing fast along a taut straight path across the oceans, or in the other extreme along erratic or split irregular paths, their influences on the surface patterns can be less predictable, and the forecasts might not be as good.
Granted, this is not a quantitative analysis, but it is one way to add an extra input to the decision: Should I do all of what I want to do now, or just half of it? The payoff for the voyager may mean a more comfortable, efficient, and safer voyage.
David Burch is the founder of Starpath Navigation in Seattle.