Live Science | Science & Technology
“Scientists have created a ‘digital twin’ of our planet that can be used to predict weather far faster than traditional services.
The technology could help prevent some of the catastrophic impacts of disasters such as typhoons and flooding. The intensive data-crunching system could also give us a more detailed view of the future effects of climate change and reveal clues about how to mitigate it.”
From Live Science.
“As Hurricane Melissa exploded into a category-5 storm over the weekend, scientists forecast its trajectory and growth with a powerful new tool — helping to inform warnings to Jamaica and other nations that the storm has devasted. That tool, an artificial intelligence (AI) forecast model developed by Google DeepMind, is successfully predicting how Melissa and other dangerous storms arise and evolve…
DeepMind’s developers trained the model on two data sets: a large database of global weather observations and a smaller database of observations that included nearly 5,000 cyclones from the past 45 years1. Adding that second, cyclone-specific database might be the reason that the DeepMind model performs better on hurricane forecasts than do other AI-based forecast models, Franklin says. In particular, scientists have long struggled to improve their forecasts of a storm’s intensity — but the DeepMind model seems to capture this well.”
From Nature.
“The potential for AI to improve weather forecasting and climate modelling (which also takes a long time and uses a lot of energy) has been known for several years now…
But a huge trial in India this year has taken a huge step forward. The Indian Ministry of Agriculture partnered with teams of scientists from the Human-Centred Weather Forecasts Initiative, the University of Chicago, California, Berkeley, Bombay, Bangalore, and others.
They sent weekly AI-powered forecasts about the monsoon to 38 million farmers across 13 states in India. These AI forecasts predicted changes in the monsoon that all other ones missed. The forecasts of the timing of the monsoon were sent up to four weeks in advance of its arrival; conventional physics-based modelling usually can’t do it more than five days in advance.
This year’s monsoon was a weird one. It hit Southern India in early June (which the AI model predicted), but then stopped temporarily for 20 days. No conventional model predicted this stall, but the AI-based one did…
In a self-reported survey, around one-quarter of the 38 million farmers adjusted their plans in response to the forecast.”
“In 1970, Cyclone Bhola hit Bangladesh, killing more than 300,000 people. It was a strong cyclone, but not unprecedented. What made it so deadly was the lack of any early detection systems, alarms, or mass evacuation procedures. A huge storm surged into a densely populated area, and hundreds of thousands of people drowned in their homes.
Since then, Bangladesh has become much more resilient to these events…
Throughout the 1980s and early 1990s, a few large events claimed many lives. But in recent decades, the death toll has been low. That’s despite Bangladesh experiencing some extremely powerful cyclones. Cyclone Amphan (2020) and Mocha (2023) were both Category 5 — the strongest rating.”
From Our World in Data.
“Fire is still a threat, especially to older buildings. But beneath the sound of those sirens is a story of underappreciated progress toward ever greater safety. Compared with 1980, the per-capita civilian fire death rate [in the United States] has fallen by roughly two-thirds — from about 28.6 deaths per million people to around 11 per million in 2023. Total reported fires are also down by half over that time period, and injuries have fallen by more than half.”
From Vox.
