
It led to a cold spell lasting more than 1,000 years known as the Younger Dryas or 'Big Freeze', during which temperatures in parts of the northern hemisphere fell to about 10 degrees C colder than they are today.
'This episode was the last time the Earth underwent a major cooling, so understanding exactly what caused it is very important for understanding how our modern-day climate might change in the future,' said Alan Condron, of the University of Massachusetts Amherst.
The cooling began after Lake Agassiz, at the southern edge of the Laurentide ice sheet covering much of the Canadian Arctic, broke through an ice dam and dumped thousands of cubic kilometers of cold water into the ocean.

The original hypothesis, proposed 1989 by Wally Broecker of Columbia University, suggested that Lake Aggasiz drained into the North Atlantic down the St Lawrence River.
But using the new model, Dr Condron, working with Peter Winsor of the University of Alaska, found that this proposed route would have weakened the oceans' thermohaline circulation by less than 15 per cent.
That level of weakening, they say, is unlikely to have accounted for the 1,000-year cold climate event that followed the flood.
Meltwater from the St Lawrence River actually ends up almost 1,900 miles south of the deep water formation regions, too far south to have any significant impact on the sinking of surface waters.
By contrast, Dr Condron and Dr Winsor's model shows that if the meltwater first drains into the Arctic Ocean, narrow coastal boundary currents efficiently deliver it to the deep water formation regions of the sub-polar north Atlantic, weakening the thermohaline circulation by more than 30 per cent.
They conclude that this scenario, showing meltwater discharged first into the Arctic rather than down the St. Lawrence valley, is 'more likely to have triggered the Younger Dryas cooling.'
Dr Condron and Dr Winsor ran their simulations on one of the world's top supercomputers at the National Energy Research Science Computing Centre in Berkeley, California.
Reporting their findings in the journal Proceedings of the National Academy of Sciences, they wrote: 'With this higher resolution modelling, our ability to capture narrow ocean currents dramatically improves our understanding of where the fresh water may be going.'
Dr Condron added: 'The results we obtain are only possible by using a much higher computational power available with faster computers.
'Older models weren't powerful enough to model the different pathways because they contained too few data points to capture smaller-scale, faster-moving coastal currents.'
'Our results are particularly relevant for how we model the melting of the Greenland and Antarctic Ice sheets now and in the future.
'It is apparent from our results that climate scientists are artificially introducing fresh water into their models over large parts of the ocean that freshwater would never have reached.
'In addition, our work points to the Arctic as a primary trigger for climate change. This is especially relevant considering the rapid changes that have been occurring in this region in the last 10 years.'
This is a very interesting remark, as it highlights the problem with scientific global warming myths. Scientists make assumptions about their models, as they can only predict from a limited set of variables in a limited amount of time. The problem occurs when they try to conform their models to their expectations, regardless of reality, often using all kinds of statistical trickery to make it fit. There is a nice book written on the subject: [Link]. Refreshing to see that there are scientists that are hitting the nail!