Predicting what the climate will look like four months from now is more difficult than predicting how it will change in a hundred years in response to rising greenhousegas concentrations. Yet that’s precisely what seasonal forecasts do; they gauge the likelihood that the next few months will be warmer or cooler, drier or wetter than usual for the given season. In Germany, climate researchers at Universität Hamburg’s Center for Earth System Research and Sustainability (CEN) and the Max Planck Institute for Meteorology (MPI-M), working in close collaboration with experts at Germany’s National Meteorological Service (DWD), have taken on this daunting task.
Monthly updated seasonal forecasts
The researchers spent over five years developing a system based on a climate model (MPI-ESM) used by the Max Planck Institute for Meteorology, but specially adapted to provide global seasonal forecasts. Today the DWD can routinely prepare global seasonal forecasts once a month. As of today, the forecasts are freely accessible online: www.dwd.de/jahreszeitenvorhersage
To mark the occasion, at a press conference held at the German Climate Consortium (DKK) Prof. Johanna Baehr (CEN), Dr. Kristina Fröhlich (DWD) and Dr. Wolfgang Müller (MPI-M) unveiled the new forecasting system, which employs a fundamentally different methodology than standard weather forecasts; instead of detailed predictions, it provides information on averaged seasonal trends, the goal being to estimate the probability of deviations from the longterm data. As such, the maps on the website show visitors how much temperatures are expected to differ from the average temperature in the next few months.
El Niño as the best example
A further area of the website is dedicated to the natural climate phenomenon El Niño Southern Oscillation (ENSO) in the tropical Pacific. Time and again, its warmphase El Niño has produced weather extremes around the globe, which can be catastrophic – like they were last winter. Prof. Johanna Baehr showed that the El Niño forecast for winter 2015/2016 was sufficiently precise, which shows that operational seasonal forecasts are already working quite well in the equatorial latitudes most affected by the ENSO. The current outlook for the next several months indicates fairly neutral conditions in the tropical Pacific, which means a cold La Niña event is relatively unlikely.
In contrast, Europe’s climate is primarily shaped by other, partly chaotic influences, like the interplay of the Icelandic Low and Azores High. As a result, the prognosis for the winter in Berlin isn’t reliable enough to serve as the basis for purchasing a warmer winter coat or deciding how much deicing salt to buy. Since the demands of government agencies, the business sector and the public at large can’t currently be met, further research is urgently needed, especially given the limitations of seasonal forecasting.
What sets seasonal forecasts apart
Dr. Wolfgang Müller subsequently explained how the researchers arrive at their forecasts. In order to predict seasonal trends, the interactions between various components of the climate system are taken into account. The status of the stratosphere – the “second floor” of our atmosphere –, the soil, the ocean and the sea ice has a much greater influence on weather developments than in a standard weather forecast. Accordingly, the latest observational data on these components is fed into the climate model, which also allows the starting point for the prognosis to be set as precisely as possible. Then the model calculates several potential seasonal trends (referred to as an “ensemble”). In other words, the model produces several different forecasts, which reflect the uncertainties in both the observational data and the model itself.
Reliability of seasonal forecasts
One of the greatest challenges in connection with the seasonal forecasts is communicating their value and reliability. Though the system can easily determine how reliable its own forecasts are, it can be difficult for nonexperts to interpret the data. Dr. Kristina Fröhlich presented the method that researchers use to assess forecast reliability. They use the “hindcasting” approach, which involves entering data from the past 30 years into the climate model and comparing the results with the corresponding actual weather data. The accuracy tends to vary considerably by region and over time. As such, on the website’s maps only those regions without any crosshatching indicate that the “hindcasted” data was accurate.
Combining models to reach the goal
The global seasonal forecasting system developed by researchers at three of the DKK’s German member institutions, dubbed the German Climate Forecast System, paves the way for a range of international collaborations in the area of seasonal forecasting. The main difference from the systems used by e.g. the US National Weather Service or the British Weather Services is the choice of model. Similar to the climate models used to identify the climate trend through the end of the century, the German system consciously employs a mix of models to ensure more reliable results.
Climate Model MPI-ESM
Prof. Johanna Baehr
Dr. Wolfgang Müller
DFG-Video zur Jahreszeitenvorhersage mit Prof. Johanna Baehr und Dr. Wolfang Müller (only German)
Articel by Prof. Johanna Baehr for predicting El Niño in Winter 2015/2016
Press release by German Climate Consortium (Deutsches Klima-Konsortium, DKK)
Elisabeth Weidinger, Press and Public Information Officer
German Climate Consortium (Deutsches Klima-Konsortium, DKK)
Wissenschaftsforum, Markgrafenstraße 37, 10117 Berlin
Tel.: +49 (0)30 76 77 18 69-4 | Fax: +49 (0)30 76 77 18 69-9
E-Mail: firstname.lastname@example.org | Internet: www.deutsches-klima-konsortium.de
The German Climate Consortium (Deutsches Klima-Konsortium, DKK) brings together major actors in German climate research and climate impact research. These include universities, research institutes outside university and higher federal authorities. Under the motto “research for the society, the economy and the environment”, the DKK promotes policy relevant climate science.