Email
Password
Remember meForgot password?
    Log in with Twitter

article imageOp-Ed: Can AI deal with climate change AND human stupidity?

By Paul Wallis     Jan 28, 2018 in Environment
Sydney - It’s official; predicting the weather is now very difficult, thanks to climate change. The weird jet streams, heat, floods and droughts are hitting hard, and knowing what happens next is crucial.
As Cape Town faces “The End”, a massive water crisis, America gets hit with a weather bomb and fantastic snow, prediction is required to protect resources and crops. AI is now being touted as the way to figure out how.
People brave cold weather in New York as the National Weather Service said very cold temperatures an...
People brave cold weather in New York as the National Weather Service said very cold temperatures and wind chills will follow for much of the eastern third of the US through the weekend
Jewel SAMAD, AFP
The current state of AI in weather prediction is that it’s a multi-billion dollar industry. Major corporations are investing, and so is NOAA. The question, of course, is how good the predictions CAN be.
Estimates of weather changes have historically been under the mark. The weather patterns are so complex, and the rising sea temperatures aren’t helping. They drive the world's climate, and they're pretty restless
This is what climate change CAN mean:
• Crop failures through heat, excessive rain, and loss of water availability.
• Water resource failures like California and now Cape Town.
• Ice melt dry-ups; one theory, which seems fantastic is that the Rhine could actually run dry. Not proven, but think of what’s required to achieve that.
• Massive fires in dried-out areas, as per California recently
• Huge hurricanes, typhoons and cyclones
• Trade disruption as commodities respond to shortages in food
• Crashing fish stocks through climate induced population crashes, coral bleaching, etc.
AI is the solution? Eventually, maybe
The demand for weather information is therefore super-hot. It affects everything from health to agriculture, basic travel and shipping.
The new systems include:
250,000 plus weather stations in the US alone.
1000 weather satellites
Cognitive computing to learn, analyze and assess weather data.
Modelling storms to predict serious risk, as per Hurricane Irma last year. (Check out the link, it’s like watching a fire hose hit Florida.)
…And so what, you ask? Plenty. The fan hasn’t really been hit yet.
The main issue is that this tech is still relatively new. “Managing” isn’t the same thing as “solving”. Learning how to deal with the gigantic amounts of data involved is the first step. Prediction is OK for short term stuff but not for long term, for example. That’s the huge problem; what happens next? Nobody’s too sure.
Desiccated tree trunks protrude from the floor of Theewaterskloof Dam  a major water supply for Cape...
Desiccated tree trunks protrude from the floor of Theewaterskloof Dam, a major water supply for Cape Town. The city is grappling with its worst drought in a century
Rodger BOSCH, AFP/File
Consider this: In Cape Town “water is gold” is the new saying. Citizens get 50 litres of water per day for EVERYTHING. That’s washing, cooking, toilets, bathing, etc.
In 2015, the UN predicted major global water shortages by 2030. Mega droughts like those in the US could become common. That’s not good for an increasing population which consumes water like, well, water.
Check out Cornell’s 2014 mega drought risk map and the link. If that comes true, and the South African prediction is now coming true, as did the Californian super drought, things aren’t looking good.
Cornell University drought predictions:  As it turns out  the US southwest is really only on the edg...
Cornell University drought predictions: "As it turns out, the US southwest is really only on the edge of the region likely to be hardest hit with high levels of risk, and compared to other parts of the world has only a moderate projected change."
Cornell University
So prediction means a lot more than some academic exercise. Billions of lives could be at risk, and soon. As it is now, millions already are. New systems, machine learning, and artificial intelligence able to assess the situations are critical.
This what AI has to learn:
• Arctic and Antarctic melt rates
• Rain and snow volumes
• Locations and local variants
• Atmospheric behaviour.
• Ocean temperature fluctuations
• Ocean pH levels
• Sea temperature changes
Heat waves which have killed thousands of people
• Jet stream movements
• Major ocean current issues, like the Gulf Stream, which is crucial to the US eastern seaboard climate.
• Sea level changes
Seasonal events like hurricanes, bomb cyclones, etc.
This information even in part, uses up hundreds of terabytes for single systems. Human intelligence isn’t enough to analyze this vast amount of data.
That’s not all – Humans have to understand AI findings
The other issue, the sheer senility of the uncoordinated, incompetent response to climate change at “leadership” level, is an additional, self-inflicted human problem. Can AI penetrate human stupidity to the extent of forcing change?
Given that any amount of incontrovertible information was ridiculed by corporate culture for so many years, to the point of hitting a crisis, AI will have to learn how to make unchallengeable arguments. Data will have to be rock solid.
The moral of the story is that this isn’t a fake world. Fake news is for fake people. Maybe a type of intelligence for which fakery isn’t an option will be enough. I hope so.
This opinion article was written by an independent writer. The opinions and views expressed herein are those of the author and are not necessarily intended to reflect those of DigitalJournal.com
More about Artificial intelligence, AI and weather prediction, Heat waves, Climate change, Floods
More news from
Latest News
Top News