SDN Chronicle

DOST Plans to Tap AI to Enhance Weather Forecasting of PAGASA, Seeks Php28.77-B Budget for 2025 — Secretary Renato Solidum

edd1994

DOST's PAGASA monitors and tracks clouds to forecast weather and climate happenings. (Photo: SDN)

DOST Secretary Renato U. Solidum, Jr. (Photo: SDN)

Short link: https://wp.me/paaccn-HBM

(SDN) — Weather forecasting can be as fickle as the stubborn weather itself, for nature has a way of doing things on its own. Doesn’t give a damn to you and me!

The weather has its own merry ways, destructive ways at times, but that’s nature’s nature, what God had made it to be. Extreme on its most extreme, people become fodder.

Remember Super Typhoon Yolanda (international name Haiyan) that barreled through Eastern Visayas on November 7, 2013. Earning a notoriety as one of the most powerful typhoons ever recorded to hit land, it devastated Tacloban City and other parts of the Visayas, killing at least 6,300 people, and many more missing, and destroying infrastructure and laying to waste agriculture worth billions of pesos.

Around 11 million people mostly in Samar and Leyte were harshly impacted by the super typhoon, thousands rendered homeless, people left to rue what could have been if advanced awareness and better preparations were done as Haiyan was gathering strength on its way to the Philippines, now a nation of over 119 million typhoon-weary population.

And what of the National Capital Region (NCR) frequently inundated — the streets becoming instant swimming pools, an occasion for Metro Manila residents to have fun — with the grin-and-bear it attitude in full display — as kids and adults frolic amid pouring rain and surging floodwaters. The threats of leptospirosis and other water-borne diseases even.

It is a scene that displays Filipinos’ resilience, or even foolhardiness, not minding the threats of diseases and other dangers to life.

It is no secret that the Philippines is one of the world’s most regularly exposed to 20 tropical cyclones a year, with extreme weather events slashing through this archipelagic country like someone with a mission to kill and destroy.

Think tank Germanwatch’s Global Climate Risk Index said that extreme weather events — 317 — in the period of 2000 to 2019 ranked the Philippines at No. 4 in the list of countries most impacted by climate-related catastrophes in the past 20 years, per an article in the Philippine Star newspaper.

“Countries like Haiti, the Philippines and Pakistan are repeatedly hit by extreme weather events and have no time to fully recover before the next event hits,” Star quoted Vera Kuenzel, Germanwatch’s senior advisor for climate change adaptation and human rights.

She added, “Thus, strengthening their resilience must not only address adaptation, but also provide necessary support for dealing with losses and damages.”

What if technology like Artificial Intelligence was harnessed in tracking down Super Typhoon Yolanda, charting, mapping its every turn, analyzing how destructive it could be, the amount of rain, its wind power.

What if clouds gathering high in the atmosphere could be foreseen faster and accurately, days and hours even before it becomes a threat to lives and properties? Would it matter? Does AI play a role for a better, more speedy, more accurate weather and climate forecasting?

Google’s Deep Mind has created its AI model dubbed  GraphCast, designed to be quicker and more accurate in global weather forecasting — quicker as in one-minute results and has 10 days weather prediction.

India has laid down its Mission Mausam, a two-year research study that includes AI aimed at enhancing the 1.4-billion strong population South Asian country’s ability to predict and respond to extreme weather events and climate-related challenges.

China has its Fengwu AI-powered weather forecasting tool that has proved to be more accurate, particularly with regards to Typhoon Doksuri and Typhoon Khanun, even outperforming the European Centre for Medium-Range Weather Forecasts (ECMWF) and the United States National Centers for Environmental Prediction (NCEP).

And the Philippines through the Department of Science and Technology (DOST) headed by Secretary Renato U. Solidum, Jr., Ph.D., is joining the global shift to AI/Machine Learning (ML) from the traditional weather forecasting. Of course, without abandoning the traditional ways of forecasting, but a mix of both.

Would faster and more accurate weather prediction matter in people’s lives? Perhaps, yes and no would be a more accurate answer. Why so?

“Even how good is the forecast but those to be affected will not make a timely move, lives will still be lost. So, even in the DOST, we are trying to provide better information. If the response, though, is not really appropriate, there will be those that will be affected,” the DOST head emphasizes.

AI weather forecasting won’t replace PAGASA weathermen

First of all, he pointed out, preparedness and reduction of risks or dangers start with monitoring and warning, actions like flood control, or the correct way of land use. “Of course, the correct action or correct response before dangers arrive.”

Through this, the DOST chief explained, where are the areas to be affected can be determined. He cited the recent occurrence of a landslide in Antipolo City, Rizal province, saying the fatalities there were in the areas already identified as flood- or landslide-prone. If the victims moved to safer area they could still be alive, he seemed to indicate.

Fast & accurate prediction and willingness to move out when dangers are sounded are the keys to saving lives. Won’t work without either one.

The science chief was interviewed recently on TeleRadyo. He was asked on DOST plans to harness AI for PAGASA’s use in its weather forecasts. He revealed this earlier in a budget hearing for DOST in the Senate. PAGASA (Philippine Atmospheric, Geophysical and Astronomical Services Administration), is one of the 18 attached agencies of the department.

Solidum said forecasts of the weather changes daily, or every hour, while climate is months advance.

The DOST chief, who headed in the past the Philippine Institute for Volcanology and Seismology (PHIVOLCS), another DOST agency, pointed out that with tons of accumulated data that PAGASA uses as a result of modernization through Doppler Radar, Rain Gauge, satellites, and other technologies, “we want a faster processing time through the use of Artificial Intelligence and Machine Learning and Data Analytics.”

Through the use of AI, computation (of the data) can be made faster, he said.

Given the reason above, he says, “We want to have two years research to improve PAGASA’s weather forecasting and the DOST provided funds to PAGASA through the Advanced Science and Technology Institute (ASTI) to find out if instead of three hours of computing time, it can be done in 15 minutes and the five days forecast can increase to 14 days.” ASTI is DOST lead agency on AI research.

Solidum said its importance (research on AI) can be seen because today there is a two-week lead time in the forecasts on possible typhoons happening. DOST-PAGASA, he added, will show tropical cyclone-like vortices still far from the Philippines in the Pacific Ocean as it observes weather systems, those clouds forming that may become a typhoon two weeks ahead for the preparedness of local government units (LGUs) and other government agencies.

“We still need to use the data itself of PAGASA. So, formally, this will start in October, and it will run for two years. But, of course, hopefully, in six months to a year we can have trial forecasts to observe if (forecasting with AI) will really improve PAGASA’s weather forecasting,” he says.

He said he is “very positive” in the DOST partnership with a United States firm for the research on AI, because airlines, airports of other countries have entered into partnerships with the firm. Even universities, he added, like his alma mater in California partnered with the firm to be able to zoom in on one area to enhance and make the data gathered appropriate to the local level.

“And that is our goal,” Solidum says.

An AI robot. (Credit: Pexels)

At the same time, he allayed fears on the impact of AI-powered weather forecasting on PAGASA’s weathermen.

He said they will not be affected because PAGASA will only be using algorithm, which is faster, and that the expertise of the weather bureau’s forecasters is very vital as the input in observing the local level is not all instrument based. “So, the operators are still important, AI will only help, and it will not decide in all the aspects of forecasting.”

Witn funds from the DOST, he said PAGASA and ASTI are the two agencies involved in the two years research on AI aimed at faster and more accurate weather and climate forecasting.

Before the DOST thought of harnessing the powers of AI on weather forecasting, Solidum said AI is already being used in agriculture, medicine, even business, as well as in disaster risk reduction (DRR), and interpretation of data from satellites.

“Thus, it is the DOST System together with PAGASA that will be looking at the AI application (on weather forecasting), with ASTI as the lead agency.”

DOST proposed Php28.77 billion budget in 2025 for four strategic objectives

The DOST chief did not reveal how much is the cost of the research on AI, but said it is “big” because it’s a research study. “That’s OK. If it is already operationalized, PAGASA still needs the budget with the specific purpose of bringing AI to operation.”

Once AI is operational the cost of making it work will be lesser, he added, the research costs bigger because researchers have to be paid.

He was also asked about PAGASA’s Doppler Radars. He said the number has to be increased, with DOST targeting 21 from the 11 operational. A Doppler Radar takes two years to build after a bidder wins as it has to be built and the tower to be constructed as well.

He said there’s already a budget for two Doppler Radars, and still needs two or three more.

Doppler Radar is needed in determining the exact location of a tyhoon, strength of rain and wind because satellites can only observe from space down to the lower level, but cannot see from the ground and up in space. “So, Doppler Radars and satellites must be complimentary,” he explains.

In the past, Solidum recalled, DOST only has a few Doppler Radars so there were no warnings on strong rains. Today, he added, aside from rainfall forecast, there’s already heavy rainfall warning that comes every three hours because of Doppler Radars.

He was apparently referring to the alerts sent as SMS by the National Disaster Risk Reduction and Management Center (NDRRMC) to Filipinos, colored coded in Yellow, Orange, and Red to indicate the volume of rainfall.

Solidum also spoke about the DOST’s proposed 2025 proposed budget to Congress worth Php28.77 billion.

He laid down where the budget would be spent, mentioning that the DOST program for it is anchored on four strategic objectives, such as:

Let us hope Solidum’s vision and plans for the nation and the Filipinos through DOST and its various agencies, support of its partners in the private and public sectors see the light of day and improve lives overall. (♡)

Exit mobile version