Continue below for the in depth forecast and explanation of the factors which are expected to influence the weather this Summer.
The strongest signal that emerges from our seasonal forecast for the summer is the persistence of high pressure over northern Europe while there is a tendency for low pressure over Azores, Iberia and into the western Mediterranean basin. This would bring dry weather to Scandinavia, and wetter than normal in the southern Europe, more average rainfall for northwest Europe including the UK and Ireland in July and August following a dry June.
For the June-July-August period, temperatures for the UK overall could be higher than the seasonal averages (based on the last 30 years) with a difference close to +1.5°C. This hot trend is expected later in June onwards, peaking in July, which is shown quite unanimously by the seasonal NWP models. So at least one or two heat waves will be inevitable, though perhaps temperatures not to the extreme of last summer.
Unlike last summer though, rainfall is forecast to be most likely end up closer to the seasonal average by the end of the summer, after a dry start to the summer. However, confidence for rainfall is lower than for temperatures. Indeed, some numerical models present sometimes different scenarios concerning rainfall. High pressure may continue to dominate June, so a drier-than-average month is expected. However, after a dry start to July, there may be a change away from persistent high pressure – which may allow more rainfall, particularly through August – with lower pressure close to the west or southwest.
Remaining dry for the first 10 days with warm sunny days but cool nights thanks to an easterly flow, with high pressure close to northern Britain. A breakdown of the settled conditions looks possible round or from the weekend of 10th/11th, as high pressure retreats east for a time allowing low pressure to move in from the southwest, with a chance of rain or storms pushing up from the south. But also turning very warm or hot for a time. Period 12th-16th, perhaps starting with unsettled conditions for a time in the south, with showers or storms, north may stay dry. High pressure may gain control across all parts with dry conditions returning. Generally warm.
For the second half of the month, confidence drops, but blocking high latitude high pressure may re-assert close to the north, with low pressure to the west of Portugal. The low in the means to the southwest may act as a "heat pump" and bring hot air with the threat of a heat wave in the second half of June, but any hot sunny conditions could be interspersed with thunderstorms. There is a lower chance of more unsettled and less hot conditions from the west at times too, depending on how far north low pressure over the Atlantis shifts.
Temperatures overall for the month likely to be above average, perhaps as much as 1 to 1.5C above the 1981-2010 long-term average overall, though more above in the south with hotter spells later in the month, compared to the north. Rainfall will likely be below average, thanks to little rain in the first 10 days and rain in the form of occasional showers or storms thereafter and mostly in the south.
Probability for temperatures against 1991-2020 average: 50% chance of above average, 40% chance of average, or 10% below average.
Probabilities for rainfall: 50% of chance for below average, 30% chance of average, 20% chance of above average.
Almost all global forecast models point to a dominant high pressure over Northern Europe, but the models disagree somewhat about the location. However, analogs and some models suggest blocking high further east over Scandinavia. Meanwhile, relatively low pressure looks like persisting from Portugal to the Mediterranean basin. High pressure close to the northeast and low pressure to the southwest could cause heat plumes but with some thundery weather too, especially as the heat breaks down at times. So, we envisage a hot month of July, with temperatures above the seasonal average between +1.5 and +2°C. Highest temperature anomalies in the south, closer to average in the north. Most rainfall contributed by thunderstorms which could break out, especially across southern and central areas of Britain. Scotland could be anomalously dry though, thanks to high pressure close by.
Confidence is quite high concerning the temperatures, but lower confidence on the occurrence of rainfall, which remains the most sensitive parameter to model - especially when it is likely to be convective in nature. Rainfall around average for England and Wales, but perhaps below for Scotland.
Probability for temperatures against 1991-2020 average: 70% chance of above average, 20% chance of average, or 10% below average.
Probabilities for rainfall: 50% of chance for average, 30% chance of below average, 20% chance of above average.
The evolution for the month of August, as would be expected two months away, is less clear than for June and July. The predominant areas of high pressure and low pressure that characterise June and July could remain but with high pressure further east over Scandinavia, while low pressures could be more pronounced over the North Atlantic close to western Europe as well as southern Europe, meaning more unsettled and less hot conditions than July, but more humid too. This would mean the least settled and least warm month of the summer - but temperatures still likely above average, but not as much as July.
Probability for temperatures against 1991-2020 average: 40% chance of above average, 40% chance of average, or 20% below average.
Probabilities for rainfall: 50% of chance for average, 30% chance of average, 20% chance of below average.
Producing a summer forecast is more challenging than producing one for winter, with less impact from global atmospheric drivers that influence weather patterns than during the winter. Global temperatures are rising due to climate change, so it is perhaps no surprise when a seasonal forecast goes with above average temperatures. Rainfall more susceptible to local scale short-term changes and regional variations, particularly in summer, when localised convective rainfall events can make a difference.
Conventional forecasts and seasonal forecasts are two completely different things. Conventional forecasts, produced by supercomputers from meteorological data collected from all over the globe on the ground and at altitude, make it possible to predict the weather in detail up to a week or even 10-12 days in the case of stable situations. Even though computers are still evolving and improving, this limit of around ten days can only be exceeded with difficulty or low confidence.
Seasonal forecasts are not deterministic forecasts. We can never predict what the weather will be like in Bournemouth in the evening in 1 month and three days time.
However, large-scale phenomena that happen over long periods of time can be quite a reliable predictor to how the atmosphere and thus our weather patterns can be driven and play out over long periods. Ocean temperatures which can impact global weather patterns, such as El Niño Southern Oscillation (ENSO), tend to change very slowly, so can be a useful predictive tool over longer periods and this one of the main drivers looked at when making this summer forecast. While the movements and temperatures of air masses vary greatly from one day to the next, those of the oceans are much slower due to inertia. However, there is a constant exchange of energy between the atmosphere and the oceans. A warmer than normal ocean will tend to warm and humidify the atmosphere more than usual and therefore influence the climate. This can be local (like the North Sea, the near Atlantic for our regions) but also remote: the surface waters of the equatorial Pacific off the coast of Peru, when they are warmer than normal, is a known as "El Niño" - which affects the climate of the entire planet.
As well as ENSO, there are other parameters we look at that may drive weather on a seasonal basis, such as certain cycles, such as Atmospheric Angular Momentum (AAM), the North Atlantic Oscillation, North Atlantic and Mediterranean Sea temperatures, the Madden-Julian Oscillation (MJO) cycle. Some cycles are multi-annual, others monthly, etc. The knowledge and integration of these cycles helps us refine the seasonal forecasts.
But there are also numerical weather prediction (NWP) supercomputers, that take all global atmospheric drivers into account. Numerical models, calculate from data from observations on a global scale, then make it possible to predict seasonal trends. There are about ten models offering forecasts of this type (American, European, British, Japanese and even French for the most reliable).
Then there is looking at past weather by using analogs of months or seasons that had similar patterns or cycles, such as La Nina or El Nino.
There is a range of seasonal models that have forecast charts of mean temperature, precipitation and pressure /heights available for the each of the 3 summer months. These get updated every month, the latest was issued earlier in May.
Below is a snapshot of what the main NWP seasonal models are showing at the 500mb level. Generally, for June, there is a signal for lower heights to the west and southwest of Europe, higher heights over northern Europe. For July - a mixed signal but towards + heights over much of Europe, lowest heights SW and E Europe. Signal for August is very uncertain. So, given the mixed look at 500mb, we will rely more on analogs for height tendencies and sensible weather at the surface.
Warmer than average temperatures are unanimous across all NWP seasonal output for all 3 summer months. So, confidence in warmer-than-average summer is higher than height / pressure tendencies and thus rainfall.
Standard weather forecasts are only really accurate for a week or so ahead. However, these seasonal models attempt to make forecasts over several months. They are based less on current weather events and more on major climate drivers.
These long-term weather forecasts are getting better and better. For example, they made it possible to predict the heat wave of last summer fairly accurately. But their inaccuracy and the many uncertainties associated with such long-time intervals should not be underestimated.
Many variables within an extremely complex system, such as the position and motion of the jet stream as well as different atmospheric pressure systems, serve as the basis for long-term model predictions. The “El Niño-Southern Oscillation” (ENSO) model also plays a decisive role, even if the phenomenon occurs 1000s of miles away from Europe.
The phase of El Niño Southern Oscillation (ENSO) that drives tropical weather patterns can also have some influence on the extra-tropical weather patterns at higher latitudes by forcing large-scale patterns across the globe.
A warming in the equatorial Pacific is taking place, with SSTs now in positive anomaly territory. The start of El Niño is expected to be officially announced soon because the deviation of the water temperatures in the central part of the tropical Pacific (Nino Region 3.4) is already at +0.4C degrees. That is close to the +0.5C threshold for the start of an El Niño.
El Nino causes droughts and heat waves in parts of the world, e.g. Australia in 2019, as well as in Indonesia. Other consequences: cyclones are more numerous in the central Pacific Ocean while hurricanes are generally less numerous in the Atlantic.
Global temperatures also tend to be higher than average during El Niño and lower during La Niña. But it appears that global warming is altering these observations as UK and Europe recorded its hottest summer last year, in the middle of La Nina.
Nevertheless, ENSO will be one of the main factors in compiling our summer forecasts, the temperature variations of the Pacific Ocean are carefully monitored, as this influences the global climate. On a global scale, the La Niña phenomenon, which cooled the tropical Pacific Ocean for 3 years, gave way to the gradual emergence of a warming towards El Niño conditions since March. This positive temperature anomaly of the tropical Pacific waters should peak from this summer. Currently, the phenomenon is already marked off Peru. Some models even indicate the possibility of a "super El Nino" developing at the end of the year, however with continuing cooling Trade Winds over the tropical Pacific, at the moment, keeping Nino 3.4 region at +0.4C rather than higher, it seems far-fetched for now, with a weak or moderate Nino seeming more likely. Either way, 'El Nino' could have repercussions for the second half of the year (more likely between late summer and next winter).
One of the consequences of developing El Nino on European weather this summer, we think, will be to produce a notable negative pressure anomaly over southern Europe, thanks to an enhanced sub-tropical jet stream typical of El Nino. This could bring storms and showers here that are somewhat more active than normal.
So, this summer, an area of very strong high pressure anomalies is forecast in northwestern Europe initially, before migrating to Scandinavia. While negative pressure anomalies will probably be located in the southwest of Europe and penetrate the Western and Central Mediterranean.
As well as ENSO / tropical Pacific Ocean temperature - atmosphere interactions, this forecast uses a few signals from the surface temperature of the seas and oceans nearer to Europe. Temperatures are expected to remain higher than usual in the Mediterranean and the nearby Atlantic Ocean. There is an important interaction between the ocean and the atmosphere, so if the sea around and upwind of Europe are warmer than normal, it necessarily spreads to the atmosphere.
In addition, there are studies that suggest a link between North Atlantic Ocean anomalies and jet stream behaviour. Studies have shown variations in the jet stream between cold and warm anomalies are continuous over the North Atlantic. The jet stream has been shown to exhibit a steadily southward displacement west of the meridian during cold North Atlantic SST events, corresponding to a preference of a trough over the North Atlantic and high over Europe.
During summer 2015 as well as summer 2018 cold SST anomalies were present in the North Atlantic region. In summer 2018 the North Atlantic ocean exhibited SST seasonal mean anomalies of up to -2.5◦C, whereas Europe experienced 2m-air temperature seasonal mean anomalies of about +2◦C.
European heat waves, particularly the occurrences during 2015 and 2018 both reveal a widespread area of negative SST anomalies across the eastern North Atlantic and positive surface temperature anomalies in central to northern Europe. Last year also exhibited widespread negative SST anomalies over the North Atlantic, with an enhanced undulation of the jet stream with increased double jet occurrence, phase-locking the North-Atlantic.
European sector into a trough-ridge pattern with extreme heat pulled north from North Africa across western Europe leading to many records to be broken.
As in 2015 and 2018, widespread cold SST anomalies over N Atlantic last summer likely led to trougingh here and downstream ridge over Europe – increasing risk of extreme heat plumes.
This year the North Atlantic has widespread positive surface temperature anomalies, so we can perhaps expect the reverse of the pressure patterns seen last summer, with lower pressure over mainland Europe and higher pressure over the NE Atlantic. So, a repeat of last summer’s extreme heat and also persistent heat of 2018 heat seems less likely. Indeed - some seasonal NWP are indicating this. But given the warmth over the North Atlantic and Mediterranean, we could still expect above normal temperatures, with some hot spells.
Dry soils encourage anomalously high temperatures during the summer months, because hardly any heat from the sun is lost through the evaporation of soil moisture. In Iberia it has already been very hot in April and May. However, heavy rain and storms recently are likely to continue this month across southern Europe, particularly Iberia.
So, soil moisture may recover from the recent severe drought that has plagued not just Iberia but also other parts of southern Europe. Should rain continue to moisten soils in Iberia this month, our main source of heat in summer, as models suggest through lower pressure than normal in June and July over southern Europe, then it may be less likely that extreme heat will build and move to the UK compared to last summer. But it doesn't mean spells of hot weather are unlikely, they may occur but with it more humidity and without extreme dry heat temperatures that were experienced last July.
Atmospheric angular momentum (AAM) is a complicated subject to discuss in layman's terms, so I will try to make it simple. Basically, it is a measure of the rotation of the atmosphere around the Earth's axis, is a useful quantity to investigate changes in the global atmospheric circulation. Atmospheric angular momentum (AAM) can be a useful tool to predict sub-seasonal weather-climate phenomena and is a starting point for monitoring their regional impacts. It was discovered over 30 years ago that there are 50-day variations in length of day and global AAM. These variations were linked to the MJO, discussed below, which drives a coherent poleward and equatorward propagation of zonal mean zonal wind anomalies. Global AAM anomalies peak as the zonal mean zonal wind anomalies move into the subtropics. Global AAM is largest when MJO convection anomalies are weakening near the date line. Madden (1987, 1988) proposed that frictional torque anomalies over the Pacific Ocean basin to the east of convection anomalies were responsible for the exchange of angular momentum between the atmosphere and the earth. There tends to be an upward trend in AAM when MJO wave moves east and convection increases over the central Indian Ocean. As convection moves to the west Pacific, positive mountain torques from east Asian topography provide for a continued increase in AAM. AAM reaches a maximum as convection weakens near the dateline.
An increase in AAM as the MJO wave of tropical convection reaches the western Pacific (MJO phases 7-8-1) and the frictional and mountain torques they induce are responsible for transport of AAM anomalies from the tropics to the mid-latitudes which can lead to amplification of the upper flow in the mid-latitudes in winter by slowing the jet stream. A few models publish AAM forecast for a few weeks to a month ahead.
CFSv2 AAM forecast
During the last 3 weeks of May, the MJO reached the Western Pacific and we saw corresponding increase in AAM since late May, with high AAM forecast by CFSv2 for much of June - so we are likely to see a continuation of northern / high latitude blocking and troughing over southern Europe (associated with enhanced sub-tropical jet and weak polar front jet) through June. Beyond that, it looks like AAM will decrease and perhaps even become negative in July - which suggests a pattern change that could involve troughing near Azores / Iberia shifting further north to the west of Ireland with high latitude blocking shifting east to Scandinavia.
The MJO is a large-scale coupling between atmospheric circulation and tropical convection. It differs from ENSO in being a travelling pattern across the warm tropical oceanic areas of the globe rather than the standing pattern of El Nino / La Nina over the tropical Pacific. It is characterised by an eastward progression of large regions of either enhanced or supressed tropical rainfall, this anomalous rainfall area is mostly evident over the western Indian Ocean, then the warm tropical western and central Pacific as the MJO progresses eastward. The MJO wave of enhanced or supressed tropical rainfall is usually less evident when it moves over the eastern tropical Pacific and tropical Atlantic.
The MJO affects the Indian and Monsoon, plays a role in the onset of ENSO events, has an impact on tropical cyclogenesis but also has more far-reaching impacts on northern hemisphere extratropical weather through Rossby Wave propagation. The Rossby Wave propagation depends on the longitude of where the enhanced convection associated with the MJO wave takes place. As the MJO has a significant impact on northern hemisphere weather patterns, including the North America, Atlantic and Europe, there are correlations that can be made between the 8 different phases of the MJO and the lagged impacts on the upper air patterns that might be expected over the North Atlantic and Europe based on composites of previous events.
The MJO tends to be more active during El Nino than La Nina, so we may see increasing forcing on global weather patterns from coherent MJO waves over the coming few months as El Nino develops.
At the moment, the recent MJO pulse, which had been moving steadily across the Western Pacific region in late May, has recently weakened and is forecast to become indiscernible in the coming days. High AAM, as discussed above, will continue to drive a strong sub-tropical jet stream beneath high latitude blocking – bringing unsettled southern Europe / settled northern Europe through June.
However, there is a signal from some longer-range model output, by mid-June, for an MJO pulse to develop over the Indian Ocean then progress towards the Indonesia area before reaching the western Pacific by late June or early July. Should this occur, we may see troughing develop further north and east across western Europe in early July, bringing a greater threat of unsettled conditions to the UK, but probably of the hot/humid and thundery variety. Beyond late July, confidence on MJO impact is very low.
Years with similar La Nina to El Nino transition: With a developing El Nino predicted this summer, we use previous years (analogs) where similar Oceanic Nino Index (ONI) increases were noted between previous winter and the upcoming summer. The ONI is based on SST departures from average in the Niño 3.4 region, and is a principal measure for monitoring, assessing, and predicting ENSO.
The ONI for DJF (December-January-February) was -0.7C (La Nina). The ONI forecast for JJA (June-July-August) is 1.3C from ECMWF and UKMO models, 1.4C for NCEP
So, this is an increase of around 1.6C between DJF and JJA.
We can then use the table in the above link of ONI in previous years to see which years showed a similar increase in ONI of over 1C from La Nina into neural to El Nino conditions. We will use years as far back as 1990 since which the world has warmed considerably, the years selected for the ONI analog are 1997, 2006, 2009, 2012 and 2018.
For June, the composite of analog years is for an upper high close to the NW and upper trough just west of Iberia.
For July, the composite of analog years is for an upper high over Scandinavia and upper trough close to SW Europe.
For August, the composite of analog years is for an upper high over Scandinavia (centred further east than June) and upper trough to the west of British Isles.
Early Junes that are closest match with SSTs to the start of June this year and were years where there was a transition from La Nina to El Nino are 1997 and 2009.
Early June 1997
Early June 2009
The composite of all three month suggests trough to the southwest and upper high over Scandinavia.
The forecast for March stated a probability for temperatures against 1991-2020 average of being a 50% chance of below average, 30% chance average, or 20% above average. Probabilities for rainfall: 40% of chance for average, 30% chance of below average, 30% chance of above average.
The provisional UK mean temperature was 5.7°C, which is equal to the 1991-2020 average. For England and Wales it was a month of two halves temperature-wise, colder-than-average first half, mild second half - which meant overall temperature for the month was close to the long-term average. Scotland, however, was cooler-than-average. England and Wales had the wettest March since 1981, with Northern Ireland also seeing one of its wettest Marches on record.
TThe forecast for April stated a probability for temperatures against 1991-2020 average being a 40% chance of average, 30% chance of below average or 30% chance of above average. Probabilities for rainfall: 40% of chance for average, 40% chance of below average, 20% chance of above average.
The provisional UK mean temperature was 7.8°C, which is 0.1°C below the 1991-2020 long-term average. So, we were right with forecasting a higher probability of close to average (then average or below) temperature for the month. April was an unsettled month, with rainfall close to average overall, but with regional variations, most parts of Scotland being drier than average, but southern and eastern parts of England being rather wet, most notably in Kent, and for the UK overall rainfall was 97% of average.
The forecast for May stated a probability for temperatures against 1991-2020 average of being 40% chance of above average, 40% chance of average or 20% chance of below average. Probabilities for rainfall: 40% of chance for average, 30% chance of above average, 30% chance of below average.
The provisional UK mean temperature was 11.6°C, which is 1.0 °C above the 1991-2020 long-term average. Much of Scotland was dry throughout the month, England and Wales had a wet first half but little or no rain after mid-month, so UK average rainfall was 39mm, 55% of the average, so below average for the UK as-a-whole.
Stay upto date with the long range forecast throughout the year with the Month Ahead Forecast - Updated Every Monday.