Regime Shifts and Trends in the Baltic Sea area: a statistical approach

During the late 1980s air and sea surface temperature started to increase in the Baltic Sea area. Sea ice broke up earlier and overall ice coverage declined. This resulted in a longer growing season and in increases in phytoplankton biomass as well as changes in the zooplankton and fish communities. These changes are often considered to represent a regime shift in the ecology of the central Baltic Sea, which according to some authors could be caused by a related sign change in the North Atlantic Oscillation (NAO). The aim of this investigation is to inspect relevant physical and ecosystem variables for trends and structural breakpoints in the concerned time series. We use sound statistical methods that include confidence tests at the 5% error probability level. For this purpose we investigated a broad range of physical variables including air temperature, wind speed, sea surface temperature, ice cover, precipitation, oxygen as well as ecosystem variables such as phytoplankton biomass, zooplankton and several fish species from the Baltic Sea region. Most of these time series do exhibit a statistical significant linear trend. But tests for structural breakpoints in these time series reveal only for some investigated variables the existence of a breakpoint in the 70-80ties of the last century. In contradiction to the seemingly well established “regime shift” hypothesis in the Baltic Sea no clear breakpoint can be identified in many physical variables and also not in most ecosystem variables including fish. Finally also the proposed reason for the supposed ecological regime shift in the Baltic Sea, the change of the NAO sign at around 1987, is not statistically significant. Therefore we conclude that the Baltic Sea regime shift does remain a hypothesis and most physical and ecosystem time series data from the Baltic Sea region are statistically best described by a linear trend.JRC.H.1-Water Resource