A first look at how to use 500mb anomaly charts to help try and predict the weather
These are notes from about 8 years ago. I have tried to check through to avoid any mistakes in the original text!
First off what are they, well as they say the heading on the tin explains.
They are at 500mb, roughly 18,000ft, and show the predicted contour lines, green in the case of NOAA charts and black (by and large) for ECMWF-GFS charts. The dashed red or blue lines, on the NOAA charts, with numbers on them are the predicted anomalies, hence the title, of heights in the areas shown by the dashed lines.
A word of warning, the anomaly lines are, in my view, best ignored in any attempt to use these charts to suggest what the upper air pattern at 500mb may look like in the time scale they show. There are few charts that show what the ACTUAL height is for the time scale given for the anomaly charts so knowing what the anomaly may be at T+00 let alone at T+240 or more is even more difficult to work out. With care and experience they can be used to help ‘see’ what is being predicted but do be careful.
On the NOAA charts the anomalies are indicated by dashed red (positive) or blue (negative) heights compared to the long term average for the dates shown.
Why 500mb? Well in the days prior to computer models the 500mb height was the main ‘tool’ used by forecasters to try and get an idea of what the upper air might look like 24 hours ahead and thus what the surface pattern might be.
Today with hugely complex weather computers all these and many more calculations at a large number of levels from the surface to way out into the upper atmosphere are routinely done by computers. They still follow the same basic rules forecasters of old (like me!) used 40-50 years ago. Largely very complex mathematics to solve the Laws of Thermodynamics. That said the idea of using the 500mb level is still very useful to a forecaster, hence the charts we are looking at. Predicting the upper air, difficult though that is, is much easier than that at the surface. Much of this has to do with the complication caused by moisture (water/ice vapour). By 18000ft a lot of this is no longer available at least over the northern hemisphere and especially so in winter as temperatures at all levels are much lower than in the summer. This does enable the models to fairly accurately deal with the expected airflow at this height. Of course these charts, just like any forecast chart, are far from infallible. But they do, from tests over a 3 year period I have carried out, lead me to believe that in over 70% of occasions they predict fairly accurately the 500mb flow.
By that I am referring to the major wave pattern in the atmosphere. Around the globe, north and south, long wave patterns, sometimes referred to as Rossby waves are permanently in the atmosphere. In both hemispheres the general flow is from west to east, not so around the equator where an easterly is often the direction. Within this broad westerly flow are the ridges and troughs that largely govern how the weather is at the surface. Thus if we can get the upper wind direction and pattern near enough correct then it does give a forecaster a fighting chance of getting the weather at the surface reasonably accurate. Unlike some comments you will see during the winter about ‘short wave’ developments ‘scuppering things’ they cannot develop at the surface unless the upper air pattern allows them not the other way round. We then get into the discussion as to which comes first, surface or upper air. Hugely complex and with no easy answer but the upper air does govern the surface on MOST occasions and not the other way round. Trust me.
What are ridges and troughs?
As a VERY rough guide where the ridges are generally the weather is more settled and where the troughs are is often unsettled.
Bends in the contour charts is a very simple way, the ridge is the bend at the top, see below, and the trough is the bend at the bottom is as simple as I can make it.
Sorry if the diagram is not very clear, not very good at paint and insert! (2012 period) So I hope that shows as simply as possible what I mean.
In very simple terms the ridges and troughs are caused by differences in heating north to south in the hemispheres and the mountains chains which occur over the world surface. Do NOT get too hung up on the detailed explanation for the time being.
To try and use them they need to be consistent run to run with themselves and if comparing one type of chart (say NOAA) with another set (ECMWF-GFS) then consistent with one another. I take a run of 3 days minimum as being consistent with a 70-75% probability that the pattern predicted will occur in the 6-10 day period. With NOAA we get charts for 8-14, so similar with itself. At this range then about 65% probability of accuracy.
The charts can be used to predict major pattern changes in the wave lengths, or a continuation of the current pattern. Accuracy seems about the same for either.
It is 8-10 years since I ran a 12 month check on their accuracy (chiefly for NOAA) but I have not detected any noticeable change either way recently.
They can be used to give a fair idea, with a lot of practice, to decide where the main centres may occur at the surface. The type of air mass can also be predicted from the general flow being shown. What they do not show is day to day surface variations or indeed surface detail.
500mb anomaly charts and are they useful for forecast guidance?
These are examples from way back in 2013. They show that these charts can be used to predict marked changes in weather patterns. Be that from blocked (meridional) to unsettled and vice verca. Why not show examples from recent weather. Good question but my ability to explain recently has become not very good, words I want to use, phrases, names etc, so these examples from 7 or 8 years ago should be better explained than any attempt by me now.
First of all you will have to take my word to some extent on what was showing on these charts in the prior days and whether the ones in the example are similar or not as it would take up too much space showing charts for the past 4 or 5 days. So just the ones that I feel are relevant for what I am showing.
So again showing the NOAA chart I showed in part 1 from last evening (Wednesday 7 February valid, depending whether we take the 6-10 valid for 12-16 or the 8-14 valid for 14-20 February).
I must first make the comment that close scrutiny over many months, even several years, shows that the 8-14 is often a less ‘gung-ho’ if that is the correct term than the 6-10. The 6-10 will quite often show a stronger flow, more meridional flows and larger anomalies than the 8-14. Whether these differences are real or not I have never done any research on them. The major thing is to see between the two how any changes are being predicted and do they look realistic over several days. I then compare them to the ECMWF outputs valid for the 10 days ahead, and latterly the NAEFS outputs. ECMWF-GFS is output using 00 and 12z data as is the NAEFS, NOAA do one only (mid evening our time).
Back to the 6-10 and 8-14 NOAA; 6-10 on the left
The 8-14 shows a bit less 500mb flow in the Iceland area than the 6-10 but not by much and it also shows the trough from that area as being a little less marked, a smoother flow almost. Beyond that the major pattern seems pretty similar in all aspects.
So what does the ECMWF-GFS from that same morning show?
The charts are actually both means from T+168 – T+240 so they are for the same period as the 6-10 day NOAA outlook, ending 17 February. To be totally coincident it would need the issue from the previous morning but from what I recall it makes little difference in this case.
We now have two interpretations for the same time from differing models which can be both frustrating and useful, probably in about equal measure!
ECMWF on the left shows a fairly similar pattern to NOAA in the area of Iceland and not that different in the UK area for the predicted 500mb flow. Let’s leave the anomaly parts out for the moment. GFS has the ridge more pronounced further ENE than both the other models but otherwise suggests a similar suggestion of ridging towards Greenland and not too different for where the European trough is shown. Indeed the eastern US trough and the western US pattern are again not identical but fairly similar in positions, with GFS favouring a more pronounced ridge in the far west possibly there more a touch towards the NOAA idea?
So we have considerable similarity on the 500mb patterns on all 3 models. How have they changed over the past few days? This will give us an idea of is the pattern consistent and when did this start for some idea on when the pattern shown may be said to be going to show on actual upper air charts?
Just to show how they do change let’s look at what was predicted on 19 January for 6-10 days ahead, that is out to about 2 February.
I am sure most of you remember what the UK weather was generally at that time of issue? Cold, frosts and snow in places with a good deal of discussion about how long it might last.
As you can see from the predicted pattern the change to a more unsettled westerly was showing up on the 500mb anomaly charts so that would come as no surprise to those who use them. Just briefly turning to the anomalies; note the reasonable sized –ve anomaly west of the UK. This would suggest that the main upper low, within the overall westerly pattern would probably be bringing upper winds from a fairly SW’ly direction, thus at the surface somewhat milder air might well be expected over the UK. How about what the height for southern UK on the prediction? Over the Channel area heights of about 546DM so nothing much out of the ordinary. (That comes from working for very many years and having a very rough idea of what 500mb heights translate to surface values at different times of the year). Certainly not cold but not unusually mild. Looking at my own weather station data for the end of January, say 25th onwards and maximum temperatures were, 2.3, 9.4, 9.4, 10.2, 14.1 11.4, 9.9 degrees C with my average maximum for the end of January around 6C. So the expected change to milder weather was well indicated but the actual degree of mildness was less well predicted. I think this probably was fairly similar across much of the country?
So when did these charts start to suggest another change to more blocked type and thus colder?
This was part of my post into the forum on, I think 19 January before I went on holiday.
Overall then a less cold spell but only ‘milder’ in meteorological terminology for the south west.
Beyond very early February then other links are needed to give a clue about that
So not 100% for sure but it certainly suggested that the very cold and quite snowy spell was not going to last.
I hope you will believe me when I say I am not manipulating things to fit but as they were.
On to the first charts that started to suggest another return to a more blocked type. This occurred whilst I was on holiday so I am not able to show the actual charts but from notes I made from my hotel each evening after looking at the outputs, something along the lines, again from my general notes that I keep on my pc
Up to about 25 jan all 3 showed less cold wly type, ridge over Europe=prob wsw/sw flow even but then slowly the 3 changed to a suggestion of a blocking type in the 10 day time frame, not Greenland but Azores ridging north turning flow n of west over uk on all e with a trough about 10-15 east of uk
So from 25 jan=4 feb poss start colder type, it could be as early as 2-3 days before this
30th=9th for it to start
More prob 4th and then 6-10 days perhaps at this range, could be longer?
Below is noaa from sun 2 feb 6-10 and 8-14
The chart above is the first one I saw at home so I was able to copy this. Hopefully it is quite obvious that the pattern and consistency I mentioned in rough notes on holiday continued and the pattern on the last chart is quite different from that of 19 January. Please believe me when I say that all 3 of the anomaly charts had shown this change and a similar consistency. So I stuck my neck out for the start of another cold spell on 25 January to start around 4 February. Got the date wrong but the change in pattern did occur.
So back to what we have in the past 24 hours from all 3 and also a look at what NAEFS showed this morning out to T+240.
NOAA 6-10 on left and ECMWF-GFS on the right with NAEFS below.
NAEFS at 240 hours
Remember that NAEFS shows anomaly height differences only
To me all 4 versions are not that far out from one another-you may disagree of course.
To my eyes and understanding having followed this kind of chart for 3-4 years I would say that I would side with the UK Met 6-15 day outlook. There is no sign, yet (!), of any major shift in the upper air pattern. They all to me also indicate again the possibility with heights being higher NW of the country than NE, (this does not mean the surface high will not be NE of the UK) that there continues to be the risk of disturbances moving down from the NW in a SE direction. Looking at the jet stream predictions over a similar time scale is another factor that tends to support this idea. Just where the surface high will spend much of its time is difficult to be firm about. Certainly not south of 50N. Somewhere north of 55N and probably west of rather than east of the meridian. But I would suggest a 30% probability that we may see it east, more like NE of the meridian for some of the time.
Overall no mild weather for at least two weeks with the details as always from model consensus in the
T+24 to T+144 time frames.
I have not spent much time on the anomaly part of this for a very good reason. I would rather newcomers to this idea of upper air charts become familiar with and have a good understanding of what the actual contour lines can tell us before venturing into the anomaly side of things.
I hope you enjoyed reading this and equally that you may be persuaded that this aspect of meteorology to us on the web with limited data access can give a good insight into general weather patterns even on the surface.