European Summer Forecast Success

One of the core products that the World Climate Service delivers to customers is a once-a-month seasonal forecast report focused on expected climate conditions in the next three months in Europe and North America. The report includes comprehensive discussion of factors that are likely to influence the seasonal climate, and the forecast is summarized in subjective forecast graphics for temperature and precipitation. The graphics depict regions where significant departures from the 1981-2010 normal are regarded as likely, and these maps represent the “bottom line” World Climate Service seasonal forecast.

To illustrate, the summer 2019 forecast maps for Europe are shown below. The forecast was issued on May 13th, and confidence was classified as “moderate to high”, which is higher than usual; in the 10-year history of WCS forecasts, only two other summer forecasts for Europe were issued with “moderate to high” confidence (2015 and 2018).

The forecast for summer 2019 pointed to unusually warm conditions across much of mainland Europe from the Pyrenees north to the Low Countries and across central and eastern Europe, and a broad zone of below-normal rainfall was also expected in this area. Below-normal temperatures were indicated in far northwestern Russia and the northern Middle East, and a wet summer was expected along the northwestern Atlantic margin of Europe as well as from southern Italy to Greece and Turkey.

How did the forecast fare? According to the verification maps below, the forecast was quite successful. The main warm anomaly that occurred over central Europe was aligned well with the predicted warm anomaly, and dry conditions did prevail from western mainland Europe to the Baltic states and eastern Europe. Greece and western Turkey were unusually wet, as predicted, and the cool anomaly indicated for northwestern Russia proved to be prescient (although much too conservative). However, the forecast also had some notable weak points; the British Isles were wetter than expected, northern Scandinavia was dry rather than wet, and the northern Middle East was hotter than normal instead of cool and damp.

A key component of the WCS process is a careful analysis of recent and historical climate patterns in order to uncover historical analogs that often provide useful indications of the likely future pattern. Traditionally this is a subjective process requiring expert interpretation, but the World Climate Service has developed an expanding set of statistical and analog forecast tools that provide more objective guidance. As just one example of the WCS guidance, the maps below show the summer 2019 temperature and precipitation signals arising from a super-ensemble of analogs; this analog forecast alone was mostly successful in anticipating the eventual outcome.

In light of these positive results, it is natural to ask whether this was a “one-off” success. While WCS scientists are the first to admit that seasonal forecasting is enormously challenging, and forecasts occasionally go badly wrong, summer 2019 was not a unique success for the World Climate Service. In particular, the previous summer was also well-predicted by WCS methods, as demonstrated below in the forecast and verification maps for June-August 2018.

A key contrast to note about the two summer forecasts is that the predicted warm and dry anomalies extended farther north and west in 2018 than in 2019, and in particular the British Isles and southern Scandinavia were included in the prediction of “warm and dry” in 2018, but not in 2019. This contrast was explicitly noted in the 2019 forecast discussion, where we stated, “it is unlikely that the warm and dry regime will extend as far north and west as last summer”; and this proved to be correct, especially in regard to temperature (but central and eastern Scandinavia were dry in both years). It is also worth noting that the 2018 forecast showed unusual warmth extending into western Russia, whereas the 2019 forecast did not, and this too was correct.

A final point to make is that both of these forecasts were issued with relatively high confidence, indicating that there was sufficient agreement within the guidance, and sufficient amplitude in the forecast signals, that the outcome was relatively certain. Unfortunately this is not always the case, and when signals are mixed or weak, then confidence is necessarily lower. World Climate Service verification statistics show that the forecasts are more successful when confidence is high, as we would expect.

In summary, the World Climate Service summer forecasts for Europe in 2018 and 2019 were both quite successful and illustrate that modern long-range climate forecasting is a viable and valuable enterprise. As data and computing resources continue to grow, and as the WCS continues to develop more sophisticated prediction methods, there is every reason to believe that forecast performance will continue to improve, bringing increasingly tangible benefits to diverse sectors of modern society.