{"id":171191,"date":"2026-09-05T11:05:11","date_gmt":"2026-09-05T11:05:11","guid":{"rendered":"https:\/\/www.europesays.com\/sk\/171191\/"},"modified":"2026-09-05T11:05:11","modified_gmt":"2026-09-05T11:05:11","slug":"predpoved-pocasia-dostava-velky-upgrade-google-nasadil-ai-ktora-pracuje-s-cerstvymi-satelitnymi-datami-vosveteit-sk","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/sk\/171191\/","title":{"rendered":"Predpove\u010f po\u010dasia dost\u00e1va ve\u013ek\u00fd upgrade. Google nasadil AI, ktor\u00e1 pracuje s \u010derstv\u00fdmi satelitn\u00fdmi d\u00e1tami | Vosveteit.sk"},"content":{"rendered":"

Predpove\u010f po\u010dasia patr\u00ed medzi inform\u00e1cie, ktor\u00e9 ovplyv\u0148uj\u00fa na\u0161e rozhodnutia takmer ka\u017ed\u00fd de\u0148. Niekedy ide len o to, \u010di si r\u00e1no vezme\u0161 d\u00e1\u017ednik, no presn\u00e9 \u00fadaje o vetre, da\u017edi \u010di teplote maj\u00fa v\u00fdznam aj pre po\u013enohospod\u00e1rstvo, energetiku, dopravu alebo z\u00e1sobovanie.<\/p>\n

Pr\u00e1ve v tejto oblasti chce Google<\/a> posun\u00fa\u0165 svoju umel\u00fa inteligenciu \u010falej. Google DeepMind a Google Research predstavili WeatherNext 3, nov\u00fd glob\u00e1lny AI model na predpovedanie po\u010dasia. Pod\u013ea Googlu ide o jeho najpokro\u010dilej\u0161\u00ed a najpresnej\u0161\u00ed model svojho druhu. V\u00fdraznou novinkou je najm\u00e4 sp\u00f4sob, ak\u00fdm WeatherNext 3 z\u00edskava \u00fadaje. Model ne\u010derp\u00e1 iba z historick\u00fdch meteorologick\u00fdch d\u00e1t, ale dok\u00e1\u017ee pracova\u0165 aj s aktu\u00e1lnymi satelitn\u00fdmi pozorovaniami.<\/p>\n

<\/p>\n

<\/p>\n

Odoberaj Vosveteit.sk cez Telegram a prihl\u00e1s sa k odberu spr\u00e1v
\n <\/a><\/p>\n

Nov\u00fd model dok\u00e1\u017ee predpoveda\u0165 po\u010dasie podrobnej\u0161ie<\/p>\n

Jedn\u00fdm z probl\u00e9mov star\u0161\u00edch modelov je rozl\u00ed\u0161enie. WeatherNext 2 pracoval s glob\u00e1lnou sie\u0165ou s rozl\u00ed\u0161en\u00edm 25 kilometrov a nov\u00e9 predpovede vytv\u00e1ral ka\u017ed\u00fdch \u0161es\u0165 hod\u00edn. WeatherNext 3 ide v\u00fdrazne \u010falej. Teplotu a vlhkos\u0165 pri zemskom povrchu dok\u00e1\u017ee predpoveda\u0165 v rozl\u00ed\u0161en\u00ed a\u017e 5 kilometrov. Pri \u010fal\u0161\u00edch povrchov\u00fdch veli\u010din\u00e1ch pou\u017e\u00edva rozl\u00ed\u0161enie 10 kilometrov a atmosf\u00e9rick\u00e9 \u00fadaje, napr\u00edklad r\u00fdchlos\u0165 vetra, predpoved\u00e1 v rozl\u00ed\u0161en\u00ed 25 kilometrov. Google tak z\u00edskal glob\u00e1lnu predpove\u010f s pribli\u017ene p\u00e4\u0165kr\u00e1t jemnej\u0161ou mrie\u017ekou ako pri predch\u00e1dzaj\u00facej gener\u00e1cii.<\/p>\n

Pre pou\u017e\u00edvate\u013ea to m\u00f4\u017ee znamena\u0165 presnej\u0161ie inform\u00e1cie aj v oblastiach, kde dok\u00e1\u017ee po\u010dasie v\u00fdrazne meni\u0165 charakter na malej vzdialenosti. Typick\u00fdm pr\u00edkladom s\u00fa horsk\u00e9 oblasti, \u00fadolia alebo pobre\u017eie.<\/p>\n

\u201eWeatherNext 3 sa u\u010d\u00ed priamo z pozorovan\u00ed v re\u00e1lnom \u010dase,\u201c uv\u00e1dza Google.<\/p>\n

Pr\u00e1ve \u017eiv\u00e9 d\u00e1ta zo satelitov patria medzi najv\u00e4\u010d\u0161ie zmeny oproti predch\u00e1dzaj\u00facemu modelu. WeatherNext 3 vyu\u017e\u00edva glob\u00e1lnu mozaiku \u00fadajov z geostacion\u00e1rnych satelitov, ktor\u00e9 poskytuj\u00fa nov\u00fd poh\u013ead na aktu\u00e1lny stav atmosf\u00e9ry<\/a>. Model tak m\u00f4\u017ee vytvori\u0165 nov\u00fa predpove\u010f ka\u017ed\u00fa hodinu.<\/p>\n

Neprehliadni
\n <\/p>\n

\"V\"V <\/p>\n

\n\t\t\t\t\tVIDEO: Tak\u00fdto z\u00e1pas neuvid\u00ed\u0161 ani v Mortal Kombat. Robot odkopol s\u00faperovi hlavu, ten bojoval \u010falej <\/p>\n

<\/a><\/p>\n

Po\u010dasie sa dok\u00e1\u017ee zmeni\u0165 ve\u013emi r\u00fdchlo<\/p>\n

Hodinov\u00fd cyklus m\u00e1 v\u00fdznam najm\u00e4 pri javoch, ktor\u00e9 sa vyv\u00edjaj\u00fa v kr\u00e1tkom \u010dase. B\u00farky<\/a>, fronty alebo intenz\u00edvne zr\u00e1\u017eky sa m\u00f4\u017eu vytvori\u0165 a zmeni\u0165 v priebehu nieko\u013ek\u00fdch hod\u00edn.<\/p>\n

Star\u0161ie modely pritom \u010dasto pracuj\u00fa s d\u00e1tami, ktor\u00e9 maj\u00fa ur\u010dit\u00fd \u010dasov\u00fd odstup. Google upozor\u0148uje, \u017ee tradi\u010dn\u00e9 numerick\u00e9 modely predpovede po\u010dasia pou\u017e\u00edvaj\u00fa fyzik\u00e1lne simul\u00e1cie na superpo\u010d\u00edta\u010doch a ich d\u00e1ta m\u00f4\u017eu ma\u0165 pribli\u017ene \u0161es\u0165hodinov\u00fd posun.<\/p>\n

WeatherNext 3 chce tento probl\u00e9m obmedzi\u0165 pr\u00e1ve pravideln\u00fdm vyu\u017e\u00edvan\u00edm aktu\u00e1lnych satelitn\u00fdch pozorovan\u00ed.<\/p>\n

E\u0161te zauj\u00edmavej\u0161ia je snaha modelu lep\u0161ie zachyti\u0165 lok\u00e1lne rozdiely. Teplota<\/a> alebo vlhkos\u0165 sa toti\u017e m\u00f4\u017ee v\u00fdrazne zmeni\u0165 u\u017e po\u010das nieko\u013ek\u00fdch kilometrov. Klasick\u00e1 glob\u00e1lna predpove\u010f preto nemus\u00ed v\u017edy vystihn\u00fa\u0165 rozdiel medzi \u00fadol\u00edm a kopcom alebo medzi pobre\u017e\u00edm a vn\u00fatrozem\u00edm.<\/p>\n

\"Google\"GoogleZdroj: Google<\/a><\/p>\n

Google do tr\u00e9ningu WeatherNext 3 zapojil aj \u00fadaje z meteorologick\u00fdch stan\u00edc. Model tak dok\u00e1\u017ee pracova\u0165 s podrobnej\u0161\u00edmi inform\u00e1ciami o podmienkach pri zemskom povrchu.<\/p>\n

AI m\u00e1 presnej\u0161ie odhadn\u00fa\u0165 aj d\u00e1\u017e\u010f a sneh<\/p>\n

Predpoveda\u0165 zr\u00e1\u017eky patr\u00ed medzi naj\u0165a\u017e\u0161ie \u00falohy meteorol\u00f3gie. D\u00e1\u017e\u010f a sneh s\u00favisia s procesmi v oblakoch, ktor\u00e9 sa odohr\u00e1vaj\u00fa na ve\u013emi mal\u00fdch priestorov\u00fdch \u0161k\u00e1lach. V\u00fdsledkom m\u00f4\u017ee by\u0165 situ\u00e1cia, ke\u010f model s\u00edce spr\u00e1vne odhadne oblas\u0165 s nepriazniv\u00fdm po\u010das\u00edm, no nedok\u00e1\u017ee presne ur\u010di\u0165 hranicu intenz\u00edvnych zr\u00e1\u017eok. Google preto WeatherNext 3 tr\u00e9noval aj na \u0161pecializovan\u00fdch zdrojoch \u00fadajov o zr\u00e1\u017ekach. Jedn\u00fdm z nich je datab\u00e1za IMERG od NASA, ktor\u00e1 vyu\u017e\u00edva satelitn\u00e9 merania v r\u00e1mci syst\u00e9mu Global Precipitation Measurement. Druh\u00fd zdroj predstavuje glob\u00e1lna anal\u00fdza zr\u00e1\u017eok vytvoren\u00e1 Googlom na z\u00e1klade satelitn\u00fdch a radarov\u00fdch \u00fadajov.<\/p>\n

V\u00fdsledky pod\u013ea spolo\u010dnosti ukazuj\u00fa v\u00fdrazn\u00e9 zlep\u0161enie. Pri strednodob\u00fdch glob\u00e1lnych predpovediach dosiahol WeatherNext 3 v ur\u010dit\u00fdch testoch a\u017e o 60 percent lep\u0161\u00ed v\u00fdsledok oproti IMERG, o 30 percent oproti MRMS a o 10 percent oproti meraniam zo zr\u00e1\u017ekomerov pri krat\u0161\u00edch predpovedn\u00fdch intervaloch. Nejde pritom len o to, \u017ee AI povie, \u010di bude pr\u0161a\u0165. Nov\u00fd model m\u00e1 lep\u0161ie zachyti\u0165 aj konkr\u00e9tny tvar a hranice zr\u00e1\u017ekov\u00fdch syst\u00e9mov.<\/p>\n

\"Google\"GoogleZdroj: Google<\/a>
\nGoogle mysl\u00ed aj na sol\u00e1rne a vetern\u00e9 elektr\u00e1rne<\/p>\n

WeatherNext 3 nem\u00e1 sl\u00fa\u017ei\u0165 iba na be\u017en\u00fa predpove\u010f po\u010dasia. Google do modelu pridal aj premenn\u00e9, ktor\u00e9 m\u00f4\u017eu vyu\u017ei\u0165 v\u00fdrobcovia energie z obnovite\u013en\u00fdch zdrojov. Model dok\u00e1\u017ee predpoveda\u0165 napr\u00edklad r\u00fdchlos\u0165 vetra vo v\u00fd\u0161ke pribli\u017ene 100 metrov, teda pribli\u017ene v oblasti, kde pracuj\u00fa vetern\u00e9 turb\u00edny. Prev\u00e1dzkovatelia vetern\u00fdch elektr\u00e1rn\u00ed tak m\u00f4\u017eu lep\u0161ie odhadn\u00fa\u0165 bud\u00facu v\u00fdrobu elektriny. Podobn\u00fd princ\u00edp plat\u00ed aj pri sol\u00e1rnej energetike. WeatherNext 3 pracuje s obla\u010dnos\u0165ou a intenzitou slne\u010dn\u00e9ho \u017eiarenia, \u010do m\u00f4\u017ee pom\u00f4c\u0165 pri odhade toho, ko\u013eko energie sol\u00e1rne elektr\u00e1rne vyrobia.<\/p>\n

Presnej\u0161ie predpovede preto nemusia ma\u0165 v\u00fdznam iba pre \u010dloveka, ktor\u00fd si r\u00e1no kontroluje po\u010dasie v mobile. Energetick\u00e9 spolo\u010dnosti m\u00f4\u017eu tieto inform\u00e1cie vyu\u017ei\u0165 pri pl\u00e1novan\u00ed v\u00fdroby a elektrick\u00e9 siete pri vyrovn\u00e1van\u00ed ponuky s dopytom.<\/p>\n

WeatherNext 3 u\u017e Google nasadzuje do svojich slu\u017eieb<\/p>\n

Google nov\u00fd model nechce necha\u0165 iba vo v\u00fdskumnom prostred\u00ed. WeatherNext 3 za\u010dal nasadzova\u0165 do svojho ekosyst\u00e9mu vr\u00e1tane Vyh\u013ead\u00e1vania, aplik\u00e1cie Gemini, Google Maps a Google Earth Engine. D\u00e1ta spr\u00edstup\u0148uje aj v\u00fdvoj\u00e1rom, v\u00fdskumn\u00edkom a firm\u00e1m prostredn\u00edctvom Google Cloud. Pou\u017e\u00edvatelia ich m\u00f4\u017eu vyu\u017ei\u0165 napr\u00edklad cez BigQuery alebo Earth Engine.<\/p>\n

Pre be\u017en\u00e9ho pou\u017e\u00edvate\u013ea je v\u0161ak najzauj\u00edmavej\u0161ia zmena v samotnej predpovedi. Google tvrd\u00ed, \u017ee pri pl\u00e1novan\u00ed jeden alebo viac dn\u00ed dopredu m\u00f4\u017ee nov\u00fd model prinies\u0165 a\u017e o 50 percent presnej\u0161ie predpovede zr\u00e1\u017eok. Najv\u00e4\u010d\u0161\u00ed pr\u00ednos m\u00e1 prinies\u0165 v oblastiach, kde b\u00fdvaj\u00fa klasick\u00e9 predpovede menej spo\u013eahliv\u00e9.<\/p>\n

\u201eAtmosf\u00e9ra si v\u017edy zachov\u00e1 ur\u010dit\u00fa mieru nepredv\u00eddate\u013enosti,\u201c upozor\u0148uje Google.<\/p>\n

Aj preto WeatherNext 3 nemo\u017eno bra\u0165 ako z\u00e1ruku toho, \u010do sa v konkr\u00e9tnej lokalite stane. Model v\u0161ak ukazuje smer, ktor\u00fdm sa predpove\u010f po\u010dasia m\u00f4\u017ee ubera\u0165. Namiesto spoliehania sa iba na rozsiahle historick\u00e9 v\u00fdpo\u010dty dok\u00e1\u017ee AI pracova\u0165 aj s \u010derstv\u00fdmi pozorovaniami zo satelitov a meteorologick\u00fdch stan\u00edc.<\/p>\n

To m\u00f4\u017ee by\u0165 u\u017eito\u010dn\u00e9 najm\u00e4 pri r\u00fdchlo sa meniacom po\u010das\u00ed, presnej\u0161om odhade zr\u00e1\u017eok alebo pl\u00e1novan\u00ed v\u00fdroby elektriny z vetra a slnka. Ako presne sa jeho v\u00fdhody prejavia v ka\u017edodenn\u00fdch predpovediach, uk\u00e1\u017eu a\u017e \u010fal\u0161ie d\u00e1ta a re\u00e1lne pou\u017e\u00edvanie modelu.<\/p>\n


\n
\n
\n
\n
\n
\n P\u00e1\u010dil sa v\u00e1m \u010dl\u00e1nok? Sledujte n\u00e1s na Facebooku
\n
\n <\/a> <\/p>\n

<\/p>\n","protected":false},"excerpt":{"rendered":"Predpove\u010f po\u010dasia patr\u00ed medzi inform\u00e1cie, ktor\u00e9 ovplyv\u0148uj\u00fa na\u0161e rozhodnutia takmer ka\u017ed\u00fd de\u0148. Niekedy ide len o to, \u010di…\n","protected":false},"author":2,"featured_media":171192,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":"","_share_on_mastodon":"0"},"categories":[11],"tags":[48338,64,65,66,48339,48340,15712,1654,48341,43,40,39,42,41,234],"class_list":["post-171191","post","type-post","status-publish","format-standard","has-post-thumbnail","category-ekonomika","tag-ai-pocasie","tag-business","tag-economic","tag-ekonomika","tag-google-deepmind","tag-google-weathernext-3","tag-meteorologia","tag-predpoved-pocasia","tag-satelitne-data","tag-sk","tag-slovak","tag-slovakia","tag-slovencina","tag-slovensko","tag-umela-inteligencia"],"share_on_mastodon":{"url":"https:\/\/pubeurope.com\/@sk\/117218110782773706","error":""},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/sk\/wp-json\/wp\/v2\/posts\/171191","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/sk\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/sk\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/sk\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/sk\/wp-json\/wp\/v2\/comments?post=171191"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/sk\/wp-json\/wp\/v2\/posts\/171191\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/sk\/wp-json\/wp\/v2\/media\/171192"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/sk\/wp-json\/wp\/v2\/media?parent=171191"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/sk\/wp-json\/wp\/v2\/categories?post=171191"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/sk\/wp-json\/wp\/v2\/tags?post=171191"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}