‘Even the most se­vere weather can be pre­dicted with rea­son­able ac­cu­racy’

The Hindu Business Line - - COMMODITIES - VINSON KURIAN

Glob­ally, weather causes close to 90 per cent of disasters, im­pact­ing mil­lions of peo­ple ev­ery year and cost­ing billions of dollars in dam­age. 2017 was one of the worst on record for catastrophic nat­u­ral disasters, cost­ing mil­lions of lives and billions of dollars of dam­age across the world, says Hi­man­shu Goyal, In­dia sales and al­liances leader, The Weather Com­pany, an IBM Busi­ness.

Even the In­dian sub­con­ti­nent is highly vul­ner­a­ble to nat­u­ral disasters af­fect­ing mil­lions of lives and im­pact­ing the econ­omy. Com­bin­ing ac­cu­rate fore­cast data with AI, an­a­lyt­ics, ma­chine learn­ing with smart sen­sors and In­ter­net of Things (IoT) can help con­sumers and busi­nesses make faster, smarter de­ci­sions based on up­com­ing weather. These tech­nolo­gies have the po­ten­tial to pre­dict weather more ac­cu­rately, so gov­ern­ments, com­mu­ni­ties and peo­ple can bet­ter an­tic­i­pate and act on se­vere weather per­ils, says Goyal. Ex­cerpts:

Is there a way tech­nol­ogy can deal ef­fec­tively with nat­u­ral disasters?

The Weather Com­pany, an IBM busi­ness, de­liv­ers per­son­alised, ac­tion­able in­sights to con­sumers and busi­nesses across the globe by com­bin­ing the world’s most ac­cu­rate weather data with in­dus­trylead­ing AI, IoT and an­a­lyt­ics tech­nolo­gies.

It de­liv­ers billions of fore­casts a day around the world via seam­less cross-plat­form tech­nol­ogy. The Weather Com­pany con­nects news­cast­ers, air­line pilots, en­ergy traders, in­sur­ance ex­ec­u­tives, state agency em­ploy­ees, re­tail man­age­ment and more, to the weather in­tel they need, on any de­vice. Ad­di­tion­ally, we also ag­gre­gate the deep­est, rich­est data sets both busi­ness and con­sumer to de­liver per­sonal, re­li­able and ac­tion­able weather in­for­ma­tion, an­a­lyt­ics and in­sight.

This is a great tes­ti­mony of how de­ci­pher­ing data us­ing the right tech­nol­ogy can re­ally trans­form the way we can deal with nat­u­ral disasters.

OWhat is the ac­cu­racy with which events can be pre­dicted?

With the ad­vances in weather ob­ser­va­tion, nu­mer­i­cal weather mod­els and in­crease in com­put­ing power, even the most se­vere weather can be pre­dicted with rea­son­able ac­cu­racy. How­ever, the depths of the in­sight are de­pen­dent on the scale of the phe­nom­ena. For ex­am­ple, we are more likely to fore­cast a trop­i­cal cy­clone devel­op­ment at least three-four days in ad­vance, in com­par­i­son to pin­point­ing the ex­act lo­ca­tion of a tor­nado with the same lead time.

Does In­dia have enough data that can be mined to de­velop ad­vanced weather fore­cast­ing mod­els?

It is re­ally hard to de­fine what ‘enough’ is. From an ob­ser­va­tion stand­point, the more the va­ri­ety, ve­loc­ity and vol­ume of data, the bet­ter it will be. How­ever, a very cru­cial as­pect of data is also its qual­ity and rel­e­vance which adds im­mense value. Com­pa­nies need ad­vanced ca­pa­bil­i­ties and a strong tech­nol­ogy plat­form to mine the data.

If you look at The Weather Com­pany, our data story is about scale. We use more than 2.5 lakh per­sonal weather sta­tions, at­mo­spheric data from 50,000 flights per day, more than 40 mil­lion pres­sure read­ings from mo­bile de­vices, and close to 162 fore­cast mod­els.

All of this helps us de­ter­mine and pre­dict data with pre­ci­sion, ac­cu­racy and speed. There are sev­eral open-source nu­mer­i­cal weather mod­els which are avail­able freely to­day.

How­ever, their model con­fig­u­ra­tion and data as­sim­i­la­tion re­quires strong skills and ex­pe­ri­ence, apart from a good ob­ser­va­tion net­work which is well cal­i­brated and main­tained.

How big is the chal­lenge to mine and crunch the hu­mon­gous data to make pre­dic­tions?

There is a lot of data which is avail­able on the in­ter­net to­day. How­ever, 80 per cent of the world’s data is not search­able. It lives be­hind cor­po­rate fire­walls.

For or­gan­i­sa­tions or in­di­vid­u­als to mine this data for mean­ing­ful in­sights and to em­power busi­nesses, they need tech­nolo­gies such as AI or an­a­lyt­ics.. So stand­alone raw data might not help, but one needs pow­er­ful tools to de­ci­pher it as well.

For in­stance, cloud­bursts are usu­ally caused by strong con­vec­tive ac­tiv­i­ties on a lo­cal scale, with short du­ra­tion. Nu­mer­i­cal weather pre­dic­tion can cap­ture the en­vi­ron­ment con­di­tions con­ducive for con­vec­tive devel­op­ment quite well, but to pin­point the ex­act lo­ca­tion and tim­ing of the down­pour is not easy some­times.

While tech­nol­ogy pro­vides the right in­sights, it is im­por­tant that mul­ti­ple stake­hold­ers come to­gether to take prompt and timely ac­tion in the event of a nat­u­ral calamity.

ZY HI­MAN­SHU GOYAL In­dia sales and al­liances leader, The Weather Com­pany

We use more than 2.5 lakh per­sonal weather sta­tions, at­mo­spheric data from 50,000 flights per day, more than 40 mil­lion pres­sure read­ings from mo­bile de­vices, and close to 162 fore­cast mod­els

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.