Joe’s Weather World: Is the virus affecting weather model forecasts? (SUN-4/5)


The back half of the weekend will be pleasant for some…and gray and chilly for others. Clouds are the issue again today…and the I-70 corridor (right now) is on the northern edge of this cloud cover. Moisture, several thousand feet thick is just sort of hanging around…and when I show you the satellite pictures you’ll see what I mean.

The other aspect of today’s blog will be the title. It sounds like hype and something that you’d say to yourself…”oh come on what possibly can be the connection”… I’ll show you.



Today: Variable clouds with more sunshine north and more clouds south. Temperatures will range from the low 60s north to the low 50s south. KC is sort of in the middle of that mess.

Tonight: Cloudy and seasonable with lows in the 40s

Tomorrow: Variable clouds again with highs in the 60s. 20% chance of a shower

Tuesday: Still looks good with highs near 80°



Darn clouds…

The visible satellite picture this morning shows the cutoff…we’re on the northern fringe of the cloud cover for now..and there are thin spots and breaks in the clouds around the KC area. So we should moderate a bit for temperatures but it will be a struggle with an E or SE wind today blowing from the cooler air to the south.

These clouds will be a thorn in our side for the rest of the day through tomorrow it appears…they’ll thin out in time and we’ll warm up…but for now it’s a struggle.

I still think we warm up nicely on Tuesday as we get rid of this low level moisture…and we may enjoy that warm weather into Wednesday as well…depending on the progress of the next cold front.

This next front will usher in some cooler air…consider this a cool shot…but there may be a cold shot coming after that heading into next weekend. This could have some pretty important impacts into the middle of the month.

I started to alert you about this on the air a few days ago…and my concern level is a little higher. I don’t think we’re looking at record lows but there may be a few runs getting us close because records are in the low>upper 20s through mid month…

Model data has also been showing this potential…but there is something sort of interesting about the model data lately…

There is speculation that the models are having rougher times as a direct result of the virus that is in the news. Now it’s not because the models HAVE the virus…it’s because of the data that the models rely on. This is where it gets complicated…it’s because of something called data assimilation.

Weather models are only as good as the data that they have to work with…along with the equations that are being used to have the models spit out weather parameters that go into a forecast. Models with no data can’t really be helpful after a period of time. The data assimilation (information that the models use to formulate a forecast) is varied and has been exponentially improved upon over the decades. Complex weather satellite data…10s of thousands of surface observations…100s of upper air balloon soundings that are sent up at least once or twice a day and 1000s of aircraft reports that are flying around the globe all go into the model data assimilation.

There are 2 things above that have been altered because of the virus…one in particular and the 2nd as a secondary impact. Can you guess what might be affected by the virus?

IF you guessed surface weather observations…you’re correct. That is a secondary hit to the data assimilation. Why? Well we’re seeing, especially in other parts of the world…the virus reduce the number of weather observations being taken. Perhaps airports in areas of the globe are closing because of a lack of travel. We’re talking smaller airports in small places but yet critical pieces of information. Models in general smooth out data points…estimates between place A and place B where there is now data…but now you remove the data from place A and place B in the the middle of nowhere and the forecast data can become less reliable.

Again that’s a secondary hit. A bigger hit is the aircraft data. This is possibly affecting model output across a broad spectrum of places. Our short term models use aircraft reports across our hemisphere and our country extensively to get a read on the atmosphere every hour. This data has become somewhat more sparse in coverage because of all the groundings that have occurred. It’s no secret that globally and nationally airplanes are flying less and less. Some of those flights were important for measuring weather conditions as the flights were ascending and descending in addition to providing weather information at flight level (wind reports especially).

The folks at FlightAware have been keeping track of this…and it’s certainly been in the news…but there are a fraction of the flights flying now compared to just a few months ago. Flights typically peak in the high travel summer months…but you can see just how far down they’ve come.

Via FlightAware

While ALL these flights DO NOT send in weather information…some did and still do. You cut the number of flights though and the odds of cutting some of the weather information increases.

The GFS model (our American model) has seen about a 35% reduction in the aircraft information/data coming into it. This problem is even more pronounced in Europe and Asia where the flight reductions have been more severe. The European model has seen some bigger hits I believe in getting some important data into it.

A study that came from the European Center showed that IF you took away ALL off the aircraft data the model accuracy decreased some 15%.

We may be seeing some of those hits in the model output. This next chart shows the accuracy of the various models for the 18,000 foot level…the higher the lines the more accurate the models are. Now this is the time of the year of change as we all know…seasonal changes always bring about tougher forecasts…and models tend to swing a bit more…but there were some big “dropouts” in some of the model runs that verify 6 days after the forecast was issued.

GFS-American Model ECM-European Model CMC-Canadian Model

The EURO in particular, which is considered to be the gold standard…typically has higher accuracy in a 6 day forecast…and while it’s still better than the others…that’s a lower skill number than I remember.

So perhaps the lack of aircraft data is playing a role after all.

So with all that said, as I’ve mentioned the are some concerning trends in the data for the middle of the month…we’ll see how well this verifies considering everything that I wrote above.

Here is the EURO forecast for the 5,000 foot temperature anomalies…for the night of the 9th throught he night of the 14th

That’s cold air dumping into the States from the Canada. Again this is for 5,000 feet. By mid month the surface temperatures locally are in the mid 60s for highs on average. That won’t happen consistently with that look above.

Let’s take this another 5 days forward…

It gets a bit better because towards the 18th onwards we start coming out of the colder pattern and the chart about is a 5 day average.

So how does this affect things at the surface…well here are the EURO ensemble temperature forecasts for the next 14 days…

and the GFS…

I do think the mid-month cold is a big deal and with the advancing vegetation over the next week or so…this is be a nasty few days of potential hard freezes coming our way. Now remember night to night changes in winds and clouds can impact the overnight lows…and those can’t be figured out from 10-15 days out…but it’s something to start thinking about if you’re a gardener.

OK pretty thought provoking blog for you. The feature photo comes from Jessica Thompson


Tracking Coronavirus

More Tracking Coronavirus

Weather News



More News