SCIENTIFIC RESULTS 



189 



A service of ieebero- forecastiii<>% covering the spring season for 

 the Avestern Xorth Atlantic based upon the above correlations, was 

 instituted in 1I>20 and has been continued through 1929. Forecasts 

 compared with actual records have been as follows: 



ACTX'AL 



Berss 



345 

 390 

 515 



,1926, below normal 150 



1927, normal 386 



1928, above normal 500 



1929, below normal 350 



1930, iibove normal 520 



T.t2G. below normal 



i;»27. normal 



1928. above normal 



1929 (estimate(P^), far above 



normal 1, 300 



1'.>M0. above normal 475 



A comparison of the actual number of icebergs with those fore- 

 liisted a month prior to the inauguration of the berg season, shows 

 four successes out of five predictions. We are able, apparently, to 

 predict quite accurately the smaller deviations from normal from 

 year to year, but the occasional great iceberg seasons such as 1909, 

 1912. and 1929 come without meteorological warning. As our knowl- 

 (■(Ige of ice, meteorology, and oceanography of polar conditions 

 increases we have good cause to expect improvements in iceberg 

 forecasting. 



ANNUAL AMOUNTS OF GLACIAL ICE AND SEA ICE 



Considerable (juantitative data have been compiled by comparing 

 (lie relatives volumes of glacial ice and of sea ice discharged into the 

 northwestern North Atlantic annually. The material discharged 

 from the iceberg fjords, set forth in the foregoing discussion (p. 95), 

 at J:.6 cubic (nautical) miles (29^ kilometers), more or less is estimated 

 to amount to 70 per cent of the annual outj^ut. The total wastage for 

 the entire coast of west Greenland is something like 7 to 10 cubic 

 miles (42" to G^"* kilometers). The American shores are believed to 

 contribute al)out 0.3 cubic mile (1.9- kilometers). 



The total area of sea ice which is fed into the northwestern Xorth 

 -Vtlantic during any one season, and which melts annually, may be 

 calculated quite closely from a map of these regions (see fig. 121, p. 

 200). upon which is plotted the ice and water temperature records of 

 the northwestern North Atlantic. The shelf area included between the 

 coast line on one side and the 200-fathom depth contour on the other 

 ajjproximates the bounds of melting sea ice. The area in which we 

 j'.re particularly interested is equal to 467,300 square miles, and on 

 Figure 121 it is the shaded portion })ounded by the full black line. 

 Occupying the same general inclosure but spread out somewhat 

 farther on its outer side, to and including the dotted line, lies the 

 Hieltincr area. The fact that sea ice remains on the whole wnthin the 



^1 The single failure occurred in the year 1020. The critical months for Davis Strait— 

 I>i'<ember to March — showed December pressure normal, but .Tanuar.v was made notable 

 by one of the greatest excesses of atmospheric pressure over Iceland ever recorded. 

 Kurope. in consequence, suffered a severel.v cold winter and Davis Strait one of the 

 w.umest Decembers on record. This s.vstem of winds distinctly forbade the southward 

 transportation of ice. February witnessed a "high" over the Canadian Maritime Prov- 

 imes and a "low" southeast of Cape Farewell, thus initiating a northwesterly air stream 

 w liich proved as favorable for the invasion of ice as conditions the previous month were 

 1' I] favorable. The fact, however, that December is given twice the weight of .Tanu- 

 aiy in the equation employed makes the former much outweigh the latter in the forecast. 

 March was only moderately favorable to ice importation and therefore it failed materially 

 tn effect the final prediction of a berg year below normal. The presence of over 1,300 

 ii I bergs south of Newfoundland in 1929 can not, therefore, be explained by meteorological 

 factors, but must have resulted from some other cause, perhaps one of those listed on 

 ;.. 188. 



