Project Noah cool, laugh at clouds, sing in the rain

Screengrab of Project Noah website

For many Filipinos, rain comes and goes in wildly unforeseeable ways, especially in these uncertain climes. But the people behind the government’s multibillion-peso weather forecasting and warning system, “Project Noah,” beg to differ.

“We refuse to believe that rainfall is a random event,” said University of the Philippines scientist Carlos Primo “CP” David, one of the project leaders of Noah, which stands for Nationwide Operational Assessment of Hazards.

“And if rainfall is not random, then we should be able to predict it—by looking at the indicators, trends, and so on and so forth,” he told Inquirer editors, reporters, and columnists on Tuesday night.

Through Project Noah, a two-year public-private venture, and its rain forecasting tool, the “Climate Experiment,” David boasts of a 90-percent success rate in predicting rainfall in specific areas of the country.

Website

Since its launch in July, the Noah website (noah.dost.gov.ph) has been giving the public the unique chance to access data taken from the weather bureau’s Doppler radars, satellite imagery, rain and stream gauges, and other forecasting tools in real or near-real time.

The grand objective is to improve rain and flood forecasting through speedier and more accurate reporting and dissemination of weather information.

By navigating Noah’s rainfall prediction indicator, Filipinos can tell if rain will fall over certain towns or cities in the next hour, and even in the second, third or fourth hour. Other tools will show water levels in major river systems and these can alert riverine communities to potential flash floods.

90-percent accuracy

“Every day we look at our accuracy and I can tell you that when we say there’s rain, the accuracy is, in one hour, 90 percent,” David said.

But what about the 10 percent?

“That’s God changing the settings,” joked the UP and Stanford-educated geologist, who is the son of Inquirer columnist and UP professor Randy David.

“Sometimes, you can see a rain cloud from Doppler data, (so) you project it will rain. Then for some reason, God will move the cloud,” he said.

“If it falls within the 10 percent, that’s God already. All of us know it’s going to be a sunny day, then suddenly it [rains, throwing] our statistics off,” David said.

Inexact science

The fact of the matter is, he said, meteorology is an inexact scientific discipline, unlike pure sciences such as physics and chemistry.

“My personal level is 95 percent,” David said. “That’s the limit, meaning when all indicators say from four sources— radar, satellites, statistics, rain gauges—in all areas, all point to the forecast, it will most probably rain in one hour,” he said.

But it is also important to explain to the public, especially to those unfamiliar with the concept of probability, what Project Noah actually means by its forecasts, David said.

“If we say 80 percent chance of rain, there is still a 20 percent chance that it will not rain. It just means eight out of 10 times, it will rain. But don’t blame me if it doesn’t. Don’t say,  ‘Hey, you said 80 percent, that’s more than 50. But it didn’t rain.’ That’s the 20 percent,” he said.

“Meteorology is really like that. There is a margin of error. We need to teach that to the public,” David said.

Too technical?

Noah’s executive director, Alfredo Mahar Lagmay, also from UP’s National Institute of Geological Sciences, reported significant strides in the system’s efforts to reach and educate more people about the warning system.

“We receive about five-six requests per second (on the website), and sometimes, it says bad gateway because there are too many requests, but we’re trying to improve that by getting as many host servers as possible,” he said.

Lagmay also dealt with criticism that Noah’s website is too technical and hard to understand for the ordinary Internet user.

“Sure, you’re making information available to the public, but it takes a while for the mind to process … But that’s also what they said about computers,” he said. “They said, or IBM said, they won’t be able to sell computers to every household because it’s too difficult to use. But now every household has a computer.”

At present, Noah uses four Doppler radars of the Philippine Atmospheric, Geophysical and Astronomical Services Administration (Pagasa), as well as 200 weather stations spread out across the Philippines.

Human factor

Lagmay announced plans to add 1,000 more stations with sensors that would send information to Noah’s servers either every 10 minutes or 15 minutes, “so that means we’re getting near real-time information from all parts of the Philippines.”

He said Noah was more accurate because of its use of different sources of data, including satellite sensors, Doppler radars, and ground stations, which he, speaking in Filipino, described as “cheating a bit.”

“You need ground verification,” he said. “As in any investigation, if you use remotely sensed imagery to explain something on the ground, you always have to have ground validation. Otherwise, it’s just an interpretation that anybody can do.”

David said the human factor was still quite important in weather forecasting.

“You can’t replace human experience, like 30 years of looking at the clouds,” he said. “If there’s a cloud, there are people who can tell that, no, it’s not going to rain. They know that but that kind of knowledge can’t be replicated. We can combine forecasting experience with automatic tools.”

Pagasa reticent

David admitted there was some reticence on the part of Pagasa about Noah’s automatic computations.

“Pagasa is insisting that the automatic computations we  make, their people need to validate and to check first,” he said. But this, he said, defeats the purpose of forecasting.

“If you really want to be sure, then don’t do forecast. Because you will just wait … So with the color system of Pagasa, for red warning (to be raised) in Metro Manila, you’ll wait for that amount of rain to fall, then issue a red warning. But the rain is there already,” he said.

“When I get the [red warning] message, I’m already stuck. The warning should come before, maybe a few hours before … Then that’s going to be useful. We can already do that,” David said.

Still, David acknowledged improvements in Pagasa’s outlook.

Not to replace Pagasa

“One thing that’s different about Pagasa right now is they’re more receptive to changes and technology. Before, they relied only on manual. But they have realized this system is okay,” David said.

Noah, he stressed, will not replace Pagasa as the country’s weather bureau.

“In fact, Noah was conceived to augment the functions of Pagasa, because by 2014, all of this will be transferred to [them] … and one objective is to be able to provide weather information directly to the people,” he said.

A parallel objective, David said, is to make forecasting “a little bit more high-tech.”

“We want to bring forecasting of weather, of rainfall in particular, into the 21st century. That’s the idea,” he said.

Lagmay said local governments, especially in Marikina City, now rely on Noah to warn residents of imminent floods.

“Marikina City has officials on standby near the Marikina River to look at the water level in the river, if it reaches a critical point,” Lagmay said. “However, the city also has a war room where data from Project Noah are accessed.”

Lagmay said  Marikina has rules that specify different levels of alarm depending on the water level in the river. If the water level is 15 meters above sea level, he said, people are put on alert. If it rises to 16 meters, they are told to prepare for evacuation. If the water level rises to 17 meters, people are evacuated, and if it reaches 18 meters, the city government orders forced evacuation.

Mesh of technologies

The technologies that make up Project Noah are a mesh of instruments, such as weather satellites and Doppler radars that give information on rain-bearing clouds; all-automatic weather stations, rain gauges and stream gauges that track the weather, amount of rainfall and depth of rivers; and manned ground weather stations that validate the data culled from these technologies.

All of the automated weather stations, stream gauges and rain gauges are locally developed, and they transmit data to the Department of Science and Technology (DOST) every 10 to 15 minutes, allowing forecasters to see the weather situation in specific areas in near-real time.

Data from the  instruments and observations are put together on the Noah website, which is open to all Internet users.

Easy to learn

The website looks intimidating at first but it can be easily learned and, with practice, users will soon find it easy to navigate.

At one point in his presentation, Lagmay turned on the rainfall contour option on the overview menu and it showed the map of the Philippines with color overlays.

“The rainfall contour map shows me where in the country it rained in the past hour,” he said. The colors correspond to different amounts of rainfall: light blue for light to no rain; blue for moderate rain; dark blue for heavy rain; yellow, intense rain; and red, torrential rain.

“If you wake up in the morning and want to know where it has been raining, or if you just want to check which areas are hazardous, check the rainfall contour map,” Lagmay said.

Knowing what the colors stand for, you can avoid flooded places. Or tell whether your place is likely to be inundated within hours. With a report from Nathaniel R. Melican