On the relationship between global warming, local warming in the Netherlands and changes in circulation in the 20th century

The temperature in De Bilt in the Netherlands has risen by 1 K over the 20th century. This rise parallels the rise in global temperature quite closely, albeit with a slightly higher amplitude. A linear relationship between the two, with a regression coefficient close to one, is an obvious first-order approximation. This is supported by the spatial homogeneity of global warming during the 20th century, the lack of seasonality in the temperature rise, and the residuals being almost white in time. The wind direction is used as a proxy for circulation type. Locally measured wind direction gives the same results as geostrophic wind direction from pressure stations, so that systematic errors are not likely to be large. The temperature in the Netherlands, on the edge of the continent, strongly depends on the wind direction. For most wind directions and seasons the average temperature per wind direction has increased. The exception is northeasterly winds in winter, in which the variability is too large to observe a trend. The increased temperature for each wind direction can explain the observed temperature rise in all seasons within the 95% error estimates. Changes in the distribution of wind directions explain most of the interannual variability of temperature. On longer time scales, these changes have led to cooler weather in the middle of the century, but no trend is discernible over the whole century. However, in late winter and spring there is clear evidence for a change in the frequency distribution of circulation patterns affecting the Netherlands over the second half of the 20th century. During the months of February to April, more days with southwesterly wind and fewer with northeasterlies have increased the temperature even more than the observed increase in temperature per wind direction. Copyright (C) 2003 Royal Meteorological Society.