Good morning everyone, I hope you’re enjoying your lazy Sunday morning. In the second edition of my Weekend Weather Word, I decided I’d talk a bit about how I make my forecasts, what I base them on, what resources I use, etc.. Well, it is all rather plain, to be perfectly honest with you. There’s no secret access involved, certainly no hacking of any kind, even though that might come as a letdown to some people. All my sources are freely available on the internet.
This German website is basically the motherlode of forecasting for me. They regularly provide a wealth of forecast charts from numerous weather services from around the globe, from Germany to China. The models there have pretty much everything covered. From surface pressure, to precipitation, to high altitude winds, specifically the jetstream, to water temperatures. They also provide classic weather charts, data from radiosondes attached to weather balloons to give an indication of the structure of the atmosphere, and regular reports from weather stations all across Europe. They are pretty much a one stop shop for all things weather related.
This site might, at first glance, look like your run-of-the-mill weather site, all broken down and easily understandable, as it has to be for the Daily Mail reading UK crowd. However, once you check out the Charts & Data section, you find that there’s much more to that site. The wealth of model charts rivals that of wetterzentrale.de, and it offers some great tools on its own, a comparison view between two weather models, a lightning detector, which is great when it comes to assessing just how dangerous that bank of dark clounds actually is, and their own weather model all make this a resource that should not be overlooked.
Dundee Satellite Receiving Station
Okay, this site is a bit of a special case, as I do not use it for regular forecasts. The website is obviously aimed at the research and academic community, as well as at commercial forecasting services. It provides them with regular images from a number of weather satellites, both in geosynchronous and polar orbits. The good thing is that low-resolution versions of these images are available for free, once you’ve registered. What’s more, there’s no time delay. So, just like the paying customers, you’ll have access to the image as soon as they’ve been downloaded from the satellite. Now, not all is rosy here, and I’ve learned from experience, that the polar satellites are more often than not out of position, scanning the cloud cover over central Russia, while you need images of the North Atlantic. Still, having satellite images that are less than an hour old at your fingertips really comes in handy when you’re faced with a storm that’s moving in.
Yep, there they are again. Met Éireann have a bit of a bad reputation when it comes to the accuracy of their forecasts, but I myself find it hard to blame them. The atmosphere is a vast, chaotic system, and more often than not, small things that may have seemed insignificant turn out to be deciding factors in weather development. What’s more, they do offer quite a useful set of resources, less when it comes to forecasting, and more when it’s about tracking the progress of a storm system. Apart from the rainfall radar I mentioned in last week’s Weather Word, I’m talking about hourly reports from Irish airports, as well as from a series of weather buoys anchored off the west coast and the Irish Sea, as well as from weather stations mounted on the Kinsale Energy gas rigs.
To be perfectly honest with you, the process is pretty simple. I always start out by checking the latest runs of the GFS forecast model (A weather model run by NOAA in the US) over on wetterzentrale.de. To be precise I look at their “standard” chart, which shows sea level pressure, geopotential altitude along the 500 hpa line, and the temperatures on that line. Don’t worry if you didn’t understand what the hell I just wrote. That will come in a later edition. What this gives me is a general overview of the position of major weather systems in the atmosphere, and how pronounced they are, i.e. whether something is limited to lower altitudes, or whether we’re talking about a major system that affects the entire troposphere.
I usually look at the GFS forecasts for the next 4-6 days to get an understanding of how the different weather systems interact, whether there are any steep pressure gradients that could indicate gale or storm conditions, etc..Once I’m done with that, I do the same thing with the GFS forecasts for temperature 2m above ground, wind speeds 10 meters above the surface, and rainfall totals over a three hour period. Together with the air pressure forecasts, this produces a pretty good picture of the weather patterns for the next couple of days.
Once I’m done with the GFS model, I do similar checks with three other weather models. The first one I use is the Navy Global Environment model, or NAVGEM. This is run by the US Navy, or to be precise, the Fleet Numerical Meteorology and Oceanography Center. This gives a pretty good second opinion, and is very detailed, which is unsurprising given the global nature of naval operations, and the need to provide good forecasts. Even the most powerful warships afloat are very vulnerable to weather, after all. Read up on Admiral Halsey’s Task Force 38, and its encounter with Typhoon Cobra back in 1944 to see what I mean.
Another of my go-to models is the Canadian GEM model. It does not forecast as many details as the other models do, but it provides a pretty good short term view of conditions, especially out in the Atlantic, where most of our weather here in Ireland comes from. Finally, there’s the French ARPEGE model, a relatively new addition to the forecast models available on wetterzentrale, it does provide some rather high resolution forecasts, although it does tend to swing wildly betweeen forecasts from one model run to the other. There are of course other weather models available. I usually check those when two or more of my standard models show a severe weather event on its way, just to double check whether that’s a fluke, or something is really happening.
Once I’ve checked all these models, I then look for patterns. Do those models all forecast roughly the same thing, or are there variations between them? Each weather model has its own set of unique algorithms. Even though they are based on the same set of data, namely reports from a set of weather stations, aircraft, ships, and buoys, they might reach different conclusions. Each weather model also has its own “character”. GFS for example is a model that I like to call “wet” and “slow”, meaning that it tends to predict more precipitation than actually turns up, or that it underestimates the movement speed of weather systems. All this needs to be taken into account, and only when I see some common trends will I make a call and write that the weather is going to develop a certain way.
And finally, when a storm is inbound, I of course monitor its progress. That’s when Met Éireann’s weather station network, or those up-to-date satellite images come into their own. It isn’t perfect, especially since most of these resources have at least a 15 minute delay, so are far from realtime, but they’re still a good set of resources
There you go. Far from using voodoo spells, or black magic, it really is all about pattern recognition, learning to read the weather charts, and a decent bit of gut feeling. The important thing to remember though is that weather is, by its very nature, chaotic, and so sooner or later, everyone will make a wrong call. No one is immune from that, whether it is a national weather service like Met Éireann, or a single amateur meteoologist like me.
Thanks for staying with me and reading through this rather long post. If you’ve got any questions or feedback, just leave me a comment, that kind of stuff is always important for me. Also, feel free to subscribe to this blog in order to never miss another post.