OCR Text |
Show REPORTS GIVEN Forecasts Through Nation Are 88 Per Cent Verified. Says Bureau Chief The daily weather forecast Issued every morning for the past five years " the United states Weather Bureau, for the entire 1'nited States, including the minor subdivisions of each state show ;i srrund average verification of B84 per cent accuracy, according tu the annual report oi Professor Charles :F. .Marvin, chief of the Weather Bureau. Just received at the local of-flce of-flce i the Salt Lakfcj bureau. These torecasts are based on the niomlng reports shown un the daily weather map, and cover the period ending lp 3t hours. It has long beer, 'recognised, according to 1'roi'essor Siarvtll, thai the complete verlfieatlon of weather fOrecaatS, by which js meant tne Finding of the exact rels-tlon rels-tlon between the conditions forecast and those which actually occur. ln- vi'i,-M Insurmountable difficulties and) that uiii approximate results are pos-i slble, according to U series of relative-! ly arbitrary roles. j KO I i Is l li.l RBD. As verifications are based on the result- u two obserutlong u day (J a. in. and ij. mi. mountain time ac Hunt Is taken onlj ol precipitation or its absence uurmg li'-hour periods J'or example, the forecast "Rain to-nlght, to-nlght, fair Saturday" woiilO be counted h failure In the second period if some I rain fell aft. i tne morning observa-1 tlon, although the weather was fair the rest of the day; It would receive i a credit of only 50 per cent by the rules of verification, but the public would undoubtedly regard the forecast) as n complete success. In the same manner, a rain forecast for hothi pi rfods would fail if the rain stopped just before the first observation of the second period. While general rules Sot the verification verifi-cation of weather forecasts will always) fall and succeed, moro or less, in 111-1 (llviduil cases of the kind cited, it Is nevertheless generally admitted that! In such cases the failures and sue- ' cesses balance each other in the long run, so that after all the general aver-f' ages under empirical rilles fairly rep-' resent thi true results " I i ! ( ss I IDKD. The forecasts are divided for vertli-j Cation purposes into ' weather" and "temperature," the " weather" fore-1 easts for the five-year period for the I niud M ites being erifled VCf, perl) hiii ui uic inn, .iiiii in,- i , ii i " i , i i 1 1 i j 90.2 per cent. The highest verification! !for "weather" Is made in California, I Nevada, Arizona and western Texas, where the values are above 90 per cent the highest being per cent In southern California The lowest weather weath-er verifications arc in Michigan. Wisconsin, Wis-consin, Iowa, Ohio. West Virginia,! western portions of New York and I Pennsylvania, and New ESiigland where 'the average is less than So per cent; the lowest Is 813 per cent in the upper up-per peninsula of .Michigan, S3.1 per cent in lower lower Michigan and; S3. 8 per ent lii northern New England. Eng-land. The iest temperature forecasts are mad"- In California, Louisiana, Ala-! bama, Mississippi, Florida. North and) South Carolina, Maryland and Delaware Dela-ware and the Lhatrict of Columbia, where th,,- verification is above 1 pei I .cent In southern California. The poorest poor-est temperature forecasts are made for I olorado. Wyoming, Montana and the i Dakotas, where the verification is less, thaiigS6 per cent, the lowest being $4.4 per cent In Montana. I UTAH l 'Hi i Wl s. i The forecast verification for Utah, which Is In the Denver forecast district, dis-trict, has been S! -I per cent for both .weather and temperature for the past five years combined and for the state las i whole The highest weather verl-! Ification was In 1917 when the average' was 91.7 per cent, and th- lowest was In 1918 with 87.6 percent. The highest high-est temperature verification was in 1918 with 91.1 per cent the lowest 87 3 per cent In 1 9 1 6. For forecasting purposes the state! is divided Into quarters about midway I through the State, 'northern portion" I Comprising Juab, northern Sanpete, i Carbon and Uinta counties and the re-fgion re-fgion to the northward western por-ition" por-ition" comprises the region west of a line running south from Ihe inner! northeast corner However, as a mat-j tor of tact the Uinta and Wasatch .Mountain crests ai tually form the division di-vision moro accurately. for storms that prevail, in "eastern I'tah" are I usually rather general eostfljbf the Wasatch ami south of the Cintas. Salt 1-ike City, Ogden and Provo are In j "northern." "western" and "uorth-l western' Utah The Idaho Weather forecasts are Verified s.8 per cent of the time, and temperature forecasts per cent of the time, as shown by the past t i -years average. The highest verification verifica-tion of weather forecasts occurred in 1917, when it was k7.s per cent, the lowest being 84.2 per cent in l!)lo, The highest temperature verification was 89.2 per cehl in 191$, and the lowest was 85.4 per cent in 191C. 'Northern Idaho'' includes the panhandle pan-handle from Idaho county northward; 'southeastern Idaho" comprises that region east of the I tah-Nevada boundary bound-ary line, projected, Burley beliiK ill southeastern and Twin Palls In southwestern south-western Idaho though these lines are arbitrary divisions, separating only the major portions of the state. A storm that would give any Important precipt-j tatlon in southwestern Idaho would be, expected to extend rather generally to nil regions similarly exposed topo-' graphically In that district. Forecasts are printed and dlstrlbUt-j e free to all who can use them, every, week'-day morning, being placed in the malls by 9:30 usually, at Provo, Salt Lake City, Qgdeu, focatello and I'oise. the published reports at Sail Lake City and 1-toise including current weather and temperature data for 50 ar 60 places over the country All the lolly newspapers In these states publish pub-lish the dally forecasts and many of i :hem publish the additional weather: lata for other states. |