. Smithsonian miscellaneous collections. owsome correlation with the event. Early in November 1941, at the request of an Army officer, we 8 Smithsonian Misc. Coll., vol. 101, No. 1, 1941. 9 I include here the sunspot cycle, making 15 periods in all. The sunspot cycle,although not appearing in solar-constant variation, yet, as many have shown, isof meteorological significance. 32 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IO4 used this method to predict the average precipitation over the Ten-nessee Valley area for the 3 months, Novemher. December, andJanuary. From studies of the 23-year cycles,
. Smithsonian miscellaneous collections. owsome correlation with the event. Early in November 1941, at the request of an Army officer, we 8 Smithsonian Misc. Coll., vol. 101, No. 1, 1941. 9 I include here the sunspot cycle, making 15 periods in all. The sunspot cycle,although not appearing in solar-constant variation, yet, as many have shown, isof meteorological significance. 32 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. IO4 used this method to predict the average precipitation over the Ten-nessee Valley area for the 3 months, Novemher. December, andJanuary. From studies of the 23-year cycles, at 10 stations distributedover the Tennessee Valley area, we predicted the average total pre-cipitation of the area for the 3 months as 84 to 87 percent of normal. Itturned out to be 87 percent normal. A critic objected that the averagedeviation between prediction and event for the 10 individual stationswas percent, and their average deviation between the event and100 percent was only 16 percent. Hence, he said, there was not much. Copenhagen Vienna Fig. 21.—A periodicity of 8-g- months in temperatures at Copenhagen,Vienna, and New Haven, Conn., since the year 1700 when seasonal phasedisturbances are excluded. value in the prediction. I replied that as the probable error of a meanis proportional to the square root of the number of observations, 10in our case, the probable error of our prediction for the area wasonly about 4 percent, and the prediction was verified to within thismargin. I also pointed out that the deviations from 100 percent haveno such significance as he implied. For suppose the weather had beenvery wet during those 3 months, so that the average precipitationtherein was 200 percent of normal. Then instead of 16 percent ourcritics result would be 100 percent. There is no reason to say thatour prediction might not still have given an average deviation of percent, and a probable error of 4 percent. More recently, and still using 5-month running me
Size: 2093px × 1193px
Photo credit: © Reading Room 2020 / Alamy / Afripics
License: Licensed
Model Released: No
Keywords: ., bookauthorsm, bookcentury1800, bookdecade1860, booksubjectscience
