Smoothed time series of winter (December to February) snowfall in cm at Napoleon, North Dakota. Data from 1902 through 1987 have been smoothed with a nine-point binomial filter; the only points plotted are those where 9 years are available for filtering (i.e., plotted year ± 4 yr).

this associated with what has been happening in the North Pacific? We know that the North Pacific was rather cool in the 1980s compared to most of the rest of the globe.

I would like to show a few time series of snowfall. The stations are from a network of about 1,000 stations—about 300 of which go back to the turn of the century—over the United States, which I have assembled and quality-controlled in cooperation with NCDC. Figure 1 looks at Napoleon, North Dakota for the past century. An important point here is that these are stick measurements of snowfall and snow depth. Thus, they have no biases due to a change of gauge. We have not yet applied filters and looked for decadal variations. But you can see the suggestion of an increase in snowfall at Napoleon during the century.

Figure 2 is a record of New Brunswick, New Jersey snow going back to 1860. This is not total snowfall, but the number of 5-inch-plus and 10-inch-plus snowfall events. You can see quite a bit of variation across time. I do not see anything in the way of trends or cycles. This small study was inspired by the fact that until March 18, 1992, it was the third least snowy winter in New Brunswick in the last 135 years. Then it snowed 10 inches in 4 days, shooting the near-record to pieces.

It is interesting to note from this figure that the 1973-1974 to 1990-1991 period was the longest interval in which not a single year missed having a 5-inch-plus snow event. The winter of 1991-1992 ended that string. Another item of interest is that in the past 135 years, no year has had more than five 5-inch-plus snow events in central New Jersey.

Another way of examining this time series is to sum the number of 5-inch (12.7 cm) snow events by decade, as in Table 1. Here you can see that the 1980s had the most 5 inch-plus snowfall events of any decade on record. I have shown this chart to people throughout New Jersey, and no one ever believes me. However, the subject of one's


Number of =5-inch to <10-inch ( =12.7-cm to <25.4-cm) and =10-inch snowfall events per winter (October through April) from 1859/1860 to 1990/1991 at New Brunswick, New Jersey. (A snowfall event may be longer than one day, but may not be interrupted by a day without snowfall; no such events were observed for winter 1991/1992.)

perception of climate change is best left for discussion at another time.

Finally, Dr. Groisman talked about associations between snowfall, precipitation, and other climatic variables, and noticed that from 45°N to 55°N there was an inverse rela-

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement