Forecasting low water levels months in advance possible using global climate data – research

WNM | Aug 7, 2020 at 6:59 AM
Rhine (Thomas B. from Pixabay)

BREMERHAVEN, August 6 (WNM/Scientific Reports/Sebastian Grote) - Using conventional methods, the water levels in German rivers can be predicted roughly six weeks in advance. That is why the drought in the summer of 2018, characterised by extremely low flow conditions in the Rhine and Elbe, caught not only ships’ crews on the inland waterways by surprise, but also most of those running refineries, steelworks and chemical companies located along the course of the rivers. Many of the companies that rely on water transport soon reported delivery bottlenecks and production downtimes. The economic damages could certainly have been avoided if modern forecasting methods had been used. That is the conclusion researchers at the Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research (AWI) reached after testing their recently developed model for long-term predictions of flow volumes in rivers using the conditions from the drought summer 2018. The result: using global ocean and climate data, they were able to reliably forecast the low waters in the Elbe and Rhine as much as three months in advance. The AWI experts’ detailed analysis was published in a freely available article in the Nature online journal Scientific Reports.

AWI climate researcher Dr Monica Ionita still shakes her head in disbelief when she thinks about the drought in the summer of 2018 and the headlines the same autumn. At the time, newspapers and news programmes reported on unexpectedly low water levels in German rivers and how transport and supply chains had been thrown into disarray as a result. Companies like BASF and ThyssenKrupp had to limit their production; petrol and diesel prices rose due to supply bottlenecks. “Personally, I was amazed that the extreme water levels took so many stakeholders by surprise,” recounts the AWI researcher. “I have been providing the Hamburg Port Authority with regular forecasts on the water level in the Elbe for the past three years, and informed our partners in late May 2018 that the river level would be extremely low due to the long drought period – particularly in the late summer and autumn,” Ionita explains.

Five years ago, accurately forecasting the water level of rivers for periods longer than six weeks wasn’t possible. But in 2015, Monica Ionita developed a statistical model that, based on ocean and climate data, allowed her to more accurately estimate how much water would flow through various points of a selected river in the future. But would this method work even in extreme situations like the 2018 summer drought? “For us, the extreme summer was a welcome test, which gave us the opportunity to once again check our forecasting method down to the last detail,” says Ionita.

The findings of the new study speak for themselves: using their statistical methods, Ionita and her AWI colleague Dr Viorica Nagavciuc were able to precisely forecast the water volume trends in the Elbe and Rhine over a period of up to three months – even the extremely low water levels in summer 2018. “In practice that means we can say by late May of a given year approximately how the flow volume – and with it the water level – will develop up to the end of September that year. We can then evaluate whether, compared to the previous year, it will be lower, similar or higher. We can now calculate the expected discharge three months in advance – that means from this point on, we can provide concrete figures,” the researcher explains.

The analysis is based on global ocean and climate data going back almost seven decades. The two AWI scientists search these datasets for relationships between the water volume in selected rivers and key weather, ocean or soil parameters, like the surface temperature and sea level pressure in specific ocean regions. Temperature, precipitation and soil moisture at the source and in the drainage basin of each river also play an important part. “For the water levels in the Rhine and Elbe, the sea surface temperature in the North Atlantic is the crucial factor. It has a significant influence on the weather in Central Europe and, simply put, determines which course storms and rain will take,” reports Monica Ionita.

Relationships that are identified as ‘reliable’ are fed into the model – not for all rivers at once, but for each one individually. Lastly the researchers adapt their model to the respective season, since completely different independent parameters may be relevant for the same river in spring than in autumn.

For concrete forecasts, the AWI experts collect real-time data on all relevant weather and environmental parameters for the previous months and feed them into their statistical model. Using computers, they then calculate how much water will flow through a particular point of the river in a particular period. Based on this figure, the researchers can estimate how the water level will develop.

Reliable long-term water-level forecasts are as valuable for inland shipping and the companies that rely on it as accurate weather forecasts are for farming. The sooner those responsible receive information, the more effectively and more economically they can plan ahead. “We hope that with our new study, we will be able to convince more stakeholders of the value of our work. In the future, there will be no need for unwelcome surprises and production downtimes like those at the end of the 2018 summer drought. With the help of our models, we can determine almost exactly how the water levels in Germany’s major rivers will develop at any time,” says Monica Ionita.