The impact of the North Atlantic Oscillation on the development of ice on Lake Windermere

Windermere, the largest lake in England, seldom freezes over but the sheltered bays are usually covered with ice for several days every year. Here I analyse the meteorological factors influencing the development of ice on the lake between 1933 and 2000 and relate these to the regional and global changes in the weather. The results demonstrate that the methods used to describe the development of ice at high latitudes can also be used to predict the formation of ice on this temperate lake. The best indicator of change was the number of ice-days recorded every winter. Regression analyses based on a Poisson model showed that there were significant negative correlations between the number of ice-days and the local air temperature, the Central England Temperature and the North Atlantic Oscillation Index (NAOI). The relationship with the NAOI was particularly pronounced and explained 50% of the observed inter-annual variations. A hindcasting model based on the NAOI correctly predicted the most severe winters reported in the area between 1864 and 1910. The observed and predicted numbers of ice-days were also correlated with an index of sea ice in the Baltic. The results demonstrate that the number of ice-days reported on Windermere can be used as a proxy indicator of climate change and show that the NAO has had a major effect on the development of ice on this lake for at least a hundred and thirty years.