One of the questions I’m asked most often is how I create my weather forecasts.
I have discussed some of the weather websites I use, but I don’t think I’ve ever gone into detail. If I do a good job, I might put myself out of business!
To create your own forecast, you first need a good understanding of basic meteorology. I have tried to write, talk, and teach about this over the last 20 years of writing weather articles.
Read Also
Farmer gift idea: How about a weather station?
The 2025 holiday season is looming, and a home weather station might make a great Christmas gift for farmers
You should know about high and low pressure areas and how the air flows around these systems. You should know about contour lines such as isobars (lines of equal pressure), isotherms (equal temperature), and even isopleths (lines of equal geopotential height or atmospheric thickness).
You should also understand that the atmosphere is three dimensional and what is happening at one layer or height can impact the layers around them.
For example, a surface low can strengthen quickly if it interacts with an upper level low or it can weaken quickly if the upper-level low disconnects with the surface low. Another example is how upper-level winds can drive what is happening at a lower level in the atmosphere.
There are plenty of different courses and websites you can explore if you are interested in learning more. I do not have one in particular to recommend. Personally, I am still a bit of a textbook kind of guy.
If I had to pick a textbook, I would go with Meteorology Today. It is an expensive book new, but I have seen it for as little as $15 used and it doesn’t matter what edition you get.
A good website that I like to use when trying to visualize what is going on in the atmosphere is earth.nullschool.net.
This site shows the whole planet, and you can zoom in and out and rotate to see any part. The default map is a wind map, and it uses moving lines and colours to show the direction and how fast the wind is moving.
From this view, you can see how the air is flowing either into an area of low pressure or out of a region of high pressure. You can see how airflows converge. This one view alone is worth checking out the site.
You can also change the height of the wind from the surface all the way up to 10 hPa, which is near the top of the atmosphere. Clicking through the different heights, you can see how the wind pattern changes with height.
Then you can look at temperatures, relative humidity, precipitable water, plus a few other parameters. This site is invaluable for visual learners.
The next step is to look at weather models. The secret is two-fold. First, you should check them regularly and track what they are doing. This sounds like a big time investment, and as you are learning to read the weather models, it might be. But eventually it will only take 10 minutes or so in the morning and again later in the day.
The two main weather models I use generate new forecasts, or model runs, every six hours or four times a day.
There are different ways to access these weather models but I use two main websites:
These sites offer similar access to model data. They present the data a little differently and sometimes I find one form works better than the other – it is bit of personal preference.
For medium range forecasting, I look at two models, the American GFS and the European ECMWF. Both are available on the Nexlab site, while the Tropical Tidbit site only has a few options for the ECMWF model.
There are many weather model maps, ranging from sea level pressure (SLP) and precipitation (with or without showing whether it is rain or snow), to different upper-level winds, cloud cover and temperatures.
I start by looking at SLP and precipitation and go through the animation of the model run, which typically goes out 240 to 384 hours, depending on the model. I compare the two models and see how much they are alike or different. Because I watch them every day, I also keep track of which model seems to be doing a better job.
If these two models don’t agree, I go to a third model, especially in winter, and this is the Canadian model or CMC. It is only available on the Tropical Tidbits site.
This is where science and experience come together in the art of forecasting.
Sometimes I go with a particular model and agree with its predictions. If it has been doing a good job and the prediction seems reasonable, I don’t argue. More often than not, I look at something the weather model predicts and ask myself whether it looks reasonable.
For example, one weather model may show development of a strong area of low pressure, while other models show a weak area of low pressure. Is this one model getting it right and the others are not? Does the setup of the atmosphere support development of this feature?
If it does, maybe it’s something to keep an eye on. Then, in the next model run later that day, I will see if the model still predicts this. Are the other models changing to match?
If we see a shift in the other models toward one model and the models are consistent from model run to model run, then confidence rises.
I’ve surprised myself a little bit on just how much is involved in creating a forecast.
