Autumn Forecast 2018

Forecast Factors

Autumn 2017

This time last year ENSO had briefly turned positive, prompting speculation over a possible El Nino, but then headed into a weak La Nina state towards the winter. This is much the opposite of the current situation, with ENSO forecast to head into a positive state towards winter 2018/19 giving a weak to moderate El Nino. Both last year and this year had similarly low summer sea ice extent in the Arctic but this year has less over the Atlantic side of the Arctic and more over the East Siberian side.

The autumn of 2017 had a cool cloudy changeable September, a warm October which was dry for many but also cloudy, and then November was near average to fairly cool with frequent northerly and north-westerly winds, but rather warmer than might have been expected in previous decades given the synoptic setups (the Arctic was anomalously warm). Nonetheless, snow was fairly widespread around 25 and 30 November.

ENSO Signal

There is a strong consensus among the forecast models from NMME, NOAA, ECMWF and the UK Met Office that the autumn of 2018 will see ENSO move into a positive state, most likely around +1, indicating a significant possibility of an El Nino setting in during late autumn or into winter, but few model ensembles are suggesting anything comparable to the 2015/16 El Nino (out of the NMME ensemble, CFS is the only model going for a particularly strong El Nino).

Thus, when checking analogues for previous years the nearest matches are those which started out with a near-neutral to slightly positive ENSO state and moved towards a positive state of approximately +1 by November/December. 1951, 1963, 1969, 2004 and 2006 are the nearest approaches to the ENSO state we have at this time of this year.

The overall signal from these results is weak, with a weak tendency for below-average sea level pressure over the British Isles. The Septembers show a tendency for a stronger than usual Icelandic low but with close to average sea level pressure over the UK, the Octobers show a weak anticyclonic/southerly tendency, and the Novembers tend to be anomalously cyclonic and north-westerly, with pressure much below normal over southern Scandinavia and above over south-western Greenland. These anomaly centres would point towards warm weather in September and October and relatively dry weather in October, but with the odds favouring an unsettled and fairly cold November.

Stratospheric conditions

The QBO (Quasi-Biennial Oscillation) has been in an easterly phase since the summer of 2017 and the strength of the easterly phase has continued to increase during 2018. In winter the easterly phase of the QBO is correlated with a weak jet stream and more high latitude blocking than usual, a factor which may come into play later in the autumn.

MJO

The Madden-Julian Oscillation is currently in a neutral phase but the forecast models are suggesting a shift to phases 7 and 8 into September. Phase 7 is correlated with above average sea level pressure over the British Isles. Phase 8 is correlated with a stronger than average Icelandic low, broadly matching the anomaly pattern from the ENSO composites.

Phases 6 and 7 are associated with enhanced chances of cold wintry spells in November associated mostly with an above average likelihood of northerlies, but at this range it is too far out to be able to predict whether phases 6 and 7 of the MJO will feature prominently in November.

Arctic Sea Ice

Arctic sea ice is anomalously low, about the same as in 2017 but with some regional differences, notably less ice on the Atlantic side of the Arctic this year but more on the East Siberian side. Record breaking low sea ice extent affected the Beaufort and Chukchi Seas for most of winter and spring 2018 but a sluggish melt in this region during the summer has meant that the annual minimum extent there has been unremarkable by recent standards. The sea ice extent is not as low as in the exceptional years of 2007, 2012 and 2016.

Composites for years with similarly low sea ice extent at the end of summer show very little evidence of a link with mean autumn sea level pressure patterns. Breaking it down by month, however, we see a weak link with above average pressure over Scandinavia in September, an east Atlantic trough in October, and cyclonic/westerly weather in November but with high pressure in the mid-eastern Atlantic extending towards Greenland. Most of the long-term forecast models are going for another warm autumn over the majority of the Arctic, but not over the Greenland/Iceland region, which means that any polar north-westerly incursions could be quite potent, but that any direct northerlies or north-easterlies are likely to be anomalously warm.

North Atlantic Oscillation and Arctic Oscillation

The above analogues point towards a positive North Atlantic Oscillation during September, associated with a stronger than normal Icelandic low, near neutral conditions during October and then an increasing chance of the NAO turning negative during November, associated with high pressure in the eastern North Atlantic extending up towards Greenland. The Arctic Oscillation signal is weak, and the main forecast centres' model outputs disagree on the seasonal AO but do not show a clear cut signal, with the Met Office ensemble mean going for a weakly positive Arctic Oscillation and the ECMWF going for a very weakly negative Arctic Oscillation.

Temperatures and rainfall

The Met Office long range model ensemble, which was quite accurate in identifying the mean sea level pressure anomaly distribution for the summer, is going for a moderately warm autumn overall but with Britain influenced by the persistent cool anomaly in the North Atlantic, meaning closer to average temperatures in western Britain than in eastern Britain. The ensemble mean pressure anomaly points towards a relatively anticyclonic autumn over the British Isles with below average pressure around Svalbard, with a likelihood of below average rainfall.

The Netweather long-range model is suggesting a relatively anticyclonic September (mean sea level pressure 0.5 to 1hPa above the long-term normal) and with an anomalous southerly flow in October, and then a slightly more cyclonic than average November with an anomalous north-easterly flow over the British Isles. The October and November anomaly patterns are broadly consistent with many of the analogues that I have examined. The CFS long range model outputs also appear to be broadly in agreement with this.

The ECMWF long range guidance is suggesting slightly below average pressure over the British Isles overall with slightly above average pressure over Greenland, but nothing particularly remarkable. There is no signal for mean temperature, implying that mean temperatures for the autumn quarter are most likely to be near normal, and again the cold anomaly over the North Atlantic extends its influence eastwards close to the British Isles. As in many recent autumns the ECMWF model is forecasting unusual warmth in the Arctic and there are hints that the anomalous warmth may be most concentrated in November. There is no signal for above or below average precipitation.

The JAMSTEC model outlook has no signal for above or below average precipitation over the UK, but goes for a slightly warmer than average autumn over England, Wales and Scotland, and slightly cooler than average over Northern Ireland.

Conclusions

Taking the evidence together as a whole, it looks probable that September will be more anticyclonic than usual over the British Isles, with a stronger than usual Icelandic low, which means that the north and west of Scotland will be more prone to bouts of wet and windy weather but for most of the UK a warmer and drier than average September is likely. For October the signals are weak but there is a general theme of pressure being above normal across central and northern Europe, with southerly winds slightly more frequent than normal over the British Isles. For November there is a consistent signal for cyclonic and fairly cold weather, with a region of above average pressure extending from the eastern North Atlantic to southern Greenland, suggesting that northerly and north-westerly winds will blow more frequently than usual for the third November in a row. The persistent and strong easterly phase of the Quasi-Biennial Oscillation may also promote cold blasts, most likely from the north, during mid to late November.

Temperatures

Temperatures for the autumn of 2018 are expected to be above the 1981-2010 average for England, Wales and Scotland (about 80% chance) but not exceptionally so, with positive anomalies of about 0.5C most likely. For Northern Ireland, near average temperatures is the most likely outcome, thanks to the proximity of the cold anomaly in the North Atlantic.

Both September and October are most likely to be warmer than average, albeit not exceptionally so, mainly due to frequent anticyclonic weather in September and southerly winds blowing more frequently than average in October. September is 80% likely to be warmer than average, October about 60-70% likely to be warm.

November is looking about 60-70% likely to be colder than normal, with frequent northerly and north-westerly winds. The negative temperature anomaly will probably be largest over Scotland and Northern Ireland. However, severe cold outbreaks look unlikely, because of a high likelihood that the Arctic will be exceptionally warm, in view of the fact that most recent Novembers have been unusually warm in the Arctic and so was the analogous November of 1996.

Rainfall

Rainfall for Autumn 2018 is expected to be close to normal overall (about 50% chance) but with some uncertainty because of the likelihood of a dry September for the majority of the country, and a wet November. September is about 70% likely to be a dry month away from the north and west of Scotland, where near average rainfall is more likely. There is no strong signal for a particularly wet or dry October, with near average rainfall therefore the most likely outcome for October. November is about 70% likely to be wetter than normal over England and Wales, but with rainfall probably closer to normal over much of Scotland and Northern Ireland.