Declining fish stocks could be partly responsible for algal blooms in the oceans, researchers have found.

Scientists found that the fall in cod stocks in the Baltic Sea in recent decades increased numbers of the tiny marine plants that produce the blooms.

Algal blooms - sometimes known as "toxic tides" - can be poisonous to people, fish and other wildlife, and may be on the increase worldwide.

The research is reported in the Royal Society's journal Proceedings B.

"In recent years, the frequency of intense blooms (in the Baltic Sea) seems to have increased, and the level in summer has also been increasing," said Michele Casini from the Swedish Board of Fisheries in Lysekil, lead scientist on the new research.

Low on oxygen

The main cause of the blooms has been thought to be increasing levels of nutrients in the sea, with a second factor being sea temperatures driven higher by climate change.

Nutrients such as nitrogen and phosphorus wash into the seas from agricultural land, and are also produced by some types of industry - a particular problem in largely enclosed waters such as the Baltic.

These nutrients stimulate the growth of types of phytoplankton - varieties of algae - that can form blooms.

As well as the toxins they produce, the process takes oxygen out of the water.

The scientific team - which also involved researchers from Germany and Latvia - assessed three decades of data on the Baltic Sea food web.

Basically, zooplankton (tiny marine animals) eat phytoplankton, and sprat (small fish) eat zooplankton. Finally, cod eat the sprat.

"Right now, in the last 30 years, cod have been the top predators in the Baltic, after populations of seals and other marine mammals declined because of hunting," explained Dr Casini.

The data showed a simple correlation. As the cod population declined sharply from the early 1980s, the sprat population rose; zooplankton declined, and phytoplankton increased.

Wider horizons

Many other factors could have been involved; cod do not exclusively eat sprat, and sprat are also fished.

But these do not appear to have had an impact - a statistical analysis ruled out, for example, the possibility that changing herring stocks (which are also eaten by cod) were playing a role.

Instead, the influence of the cod population (which decreased by about three-quarters in a decade) emerged as the dominant factor.

The relative importance of over-fishing to algal blooms outside the Baltic is another issue. Food webs in other parts of the oceans are more complex, and data less available; so even doing the research would prove problematic.

Nevertheless, the idea that changes in top predators percolate down through the food web is well accepted.

It is simply that here, the percolation appears to have an impact at the very lowest level that is significant, visible and potentially problematic.