Discussion keeps the world turning.
Flood season is hitting harder than ever.
More rain, stronger storms, unpredictable patterns, and the need for better prevention has never been more critical.
Things like faster warnings and smarter tech and less guesswork.
When every hour counts, the right measures aren't a luxury.
They're the difference between reaction and prevention.
The fight against extreme weather is getting an upgrade and it couldn't come soon enough.
Coming to you live from our studios in Beijing, this is Roundtable.
I'm Steve Hatherly.
Thank you very much for joining us today.
And for the show, I'm with Fei-Fei and Yushan.
First up.
China is entering into a complex flood season, with rainfall, typhoons and growing climate risks, putting many regions on alert for floods, landslides and urban waterlogging.
But this year, the country's flood control response is showing a major transformation.
What once relied heavily on experience and reactive emergency measures is increasingly becoming faster and smarter, and more predictive.
Today, we'll explore how new technological breakthroughs are reshaping the country's flood prevention system and offering new hope in the fight against extreme weather.
Fei-Fei and Yusheng, good afternoon to you both.
Flood control has become an urgent issue for China, not only this year, but exactly right now.
Yeah, and also when it comes to flood seasons, it's also coming earlier every year.
I remember in the past when we were talking about flood controls.
In China we have this saying called 七上八下, meaning the peak of flood seasons will come in July and August.
That is the sort of prime time to take flood control measures.
But now in May and we're talking about in the middle and even earlier May we are seeing floods happening not only in southern China, but also in central China.
And in fact on Sunday I think, the country issued red alerts for flash flood disasters in multiple regions.
So, for example, in southern and central regions, and red alerts is the highest level for flash flood warnings.
There are different levels, right?
They are color coordinated.
There are four levels.
There are blue, which is the lowest, and then yellow and orange and then coming to the red alert, which is the most serious one, which means there are over 95 of the chance of dangerous flash flooding will happen and with potential very devastating impacts.
And that also leads to you know corresponding response measures, for example evacuation, suspend all outdoor activities or closed off some mountain or water-related tourist sites, and also enhanced monitoring which, in today's world, is heavily relying on smart technologies.
And also when it comes to overnight.
Authorities also issued some very forecasting when it comes to red alerts, flash floods in, for example, Anhui province, Hubei province or even in Zhejiang province.
And they are also issued Orange Alert, which is the third level when it comes to flood, the chances of happening serious flooding.
Flood control has become an urgent issue in the country this year due to a combination of things Yushan.
Yeah, a combination of complex weather patterns.
There's also extreme environmental conditions.
And also there's official warnings of severe flooding across multiple regions like Fei-Fei, mentioned.
And yeah, you mentioned flash flood.
I specifically looked that up.
So flash floods refers to...
The sudden and rapid flooding of low-lying areas, meaning that its impact is huge.
And once it arrives, wherever it reaches and hits, there's going to be damage, more or less.
And there's no time, very limited time to react.
Indeed.
So the Ministry of Water Resources warned as early as this April that this year could bring severe flooding in some areas, alongside drought in others.
And we're calling this year, 2026, the complex flood season.
So yeah, really, there's a lot of things that we are preparing, if not already prepared.
Specifically, northern regions, we're also expected to face unusually strong floods.
Because when we think about floods, we always think about the coastal regions or southwestern China, where there's more basin than like.
Typhoons as well, right?
Typically restricted to those areas.
But this time around, the typhoons are projected to move further inland during the summertime.
So what type of impact is that going to have on flood control?
The typhoons specifically, I mean.
First of all, it means there will be more rainfalls, especially when traditionally, for example in Beijing, we are not known to be affected by typhoons.
But in recent years we've.
But the minute that it made sort of a landfall in a big land area will fastly diminish.
So most of the time in the past, I think 10 years ago, for example, these kind of storms brought by typhoons can only be seen in coastal provinces like Guangdong or Fujian or Zhejiang.
But now we're seeing I think last year or the year before we're seeing central China, for example in Hubei province.
And even in Beijing and even my hometown, you know, it's a very, very inland region in Mongolia.
We are also affected by typhoons.
Really?
Yeah.
And that is because of the you know, climate change issue that we've said multiple times on the show before is because the environment is warming up.
That means more water is circulating in the air that contributes to stronger winds.
More rainfalls.
And so the air is carrying more water and they are able to travel, so long that inland regions are affected by typhoons.
Inland regions.
Also regions like Chongqing, for example, going through a difficult time right now.
Chongqing has very complex topography.
They have large mountains.
They have river valleys there.
And this, combined with more frequent extreme events, has made traditional monitoring, the way that we would traditionally monitor these types of extreme weather events, less effective.
And that not only that area, but anywhere around the world.
And this has become, of course, an extreme global problem.
But it necessitates a change towards more advanced and tech-driven responses.
There's one phrase that's been appearing repeatedly in recent reports, and it's talking about the shift, or the change from relying on experience to relying on predictions.
Yeah, because experiences, it's something that we've come to learn ever since thousands and thousands of years ago.
Even during prehistorical times there have been records here in China of how ancient Chinese people used to manage floods because of the I think it was.
Yellow River when it changes its droughts, and that's by nature.
Now we're talking about the increasing of rainfall.
So it's making this phrase that you mentioned Steve, from relying on experience to relying on prediction even more important nowadays, because that represents a fundamental change in how this country manages flood risks, moving just away from subjective manual estimation toward a more data-driven real-time simulation and precise digital modeling.
So in practice, it looks like This, it can be explained from several different aspects.
First of all, it's benefiting for the managing of the reservoirs.
So the reservoir acts as a regulatory valve for a river basin, balancing a lot of different things the immediate safety within long-term economic needs as well, because it's capable of many things too.
It's capable of shaving off the flood peak.
That's the first time I've ever heard of this kind of metaphor, too.
It's shaving off a flood peak when it comes.
And what that means is that the sudden surge of water that comes during a storm, that's where the reservoir can intercept it, that storm, and then store that water there.
And then that's the job of a reservoir to prevent downstream areas and river embankments from being overwhelmed by a single massive flood peak.
You can kind of think of a reservoir as a bathtub filling up if the drain is partially closed.
Closed, it's not completely blocking off the water um, but it's it's, it's reserving, it's reserving the water um, it's holding it, um.
Yeah, they play obviously a very, very important uh role.
It also helps to kind of pre-discharge the water basin, or rather the reservoir.
Yeah, especially nowadays when we're talking about rainfalls and storms, they're bringing water.
You know the level of rainfalls that's often record-breaking.
For example, this time in Chongqing, the reason that they're being so devastating is because they've seen historical record broken level of rainfall.
This is an unprecedented amount of rainfall.
Yeah they're talking about.
This kind of record has never been broken in the history when they have any meteorological records.
Unbelievable.
And it only happens within one day or even half a day.
So not only there are more rains, but also they're happening in a very short period period of time that leaves people very limited capacity when how you can respond to this level of rainfalls.
And when it comes to reservoirs, you know some of the reservoirs were built like decades ago and they were built based on historical numbers and that often the times that, For example, for some coming storm, these reservoirs must release some water to create room or capacity to absorb more rainfalls coming into the future.
And that's another way of managing floods in the past, I think.
But now I think I saw it on the report that, for example, in the city of Chongqing this time, When they're talking about monitoring floods and monitoring rainfalls, it's very data-driven.
They're relying on a lot of screens, a lot of data.
They're not only talking about the rainfalls coming to the city, so meteorological authorities are being part of the forecast.
But also they also need to measuring, for example, how fast waters in rivers are traveling.
So they can prepare, for example, which reservoir needs to be prepared if certain reservoirs needs to release some of the waters in preparation.
Yeah, because in the past, this would be a reactive process, right?
You would look at the current weather and then make a decision based on what was happening at that particular moment.
It's raining hard, so let's release some water from the reservoir.
Now what they're doing is they're using predictive measures so that they can pre-release, pre-discharge that water to make more room in the reservoir for an upcoming storm.
Yeah, and this always reminds me of this whole paradox of just how much of the ways of the nature can we as human really take control and alter in our very capable nature resources and capabilities that we now have?
That's called technology, but it's not that we're saying that we're creating a perfect system that's against all unknown possibilities, but that we have it so that we have at least something to protect us and really predict things before it happens.
For example, There's one very interesting example that I did put hard efforts on in discovering what it is.
It's called Dongxixi system.
It's located in Tong'an district of Xiamen in Southeast China's Fujian province.
So what they are known for is the...
3D virtual mirror of the physical river basin.
So it's also called digital twin system.
So by twin, it means there's the natural, actual lakes, reservoir, and rivers.
And then there is the digital replica of it.
So one real one, one digital one, that's the twin, the digital twin.
And the twin is for the what-if situation.
Exactly because the whole system refers to the deep integration of advanced technologies such as big data, cloud computing, as well as ai with water conservancy operation.
So what it does is that it creates a real-time digital replica of all the, the lakes, the river, the hydraulic infrastructure too, via digital methods.
And The reason why they really prioritize this whole system locally is because previously, the local reservoirs, such as the Ting Sea Reservoir, which is a very major one locally, heavily rely on human experiences and management, but now, with the twin system,
They have the capability of map the real time water levels and infrastructure conditions and manage to predict things quote, unquote predict things before it happens.
Well, yeah.
I mean, with a digital twin, they can put any scenario into that digital twin.
So they can test whatever situations they can come up with and see what the responses are going to be before that situation arises.
Chongqing is well known, actually, internationally.
It got a lot of recognition here in China, of course, but also internationally as well, as a leading example in this particular field that we're talking about today.
And this is Exactly what the world needs is this type of tech.
But as you just mentioned, Yushan, no system is perfect.
It's very, very difficult to predict what is going to happen next.
But with advances in precision flood warning, it could help save a whole lot of lives in the future.
And that's exactly what it's designed to do.
Yeah.
And when it comes to, for example Chongqing, I think in the last year, in 2025, the city actually also faced very heavy rainfalls.
And according to the city, they faced 119 floods above warning levels.
So it's essential to have this kind of a digital twin network so that the city can predict and prepare and respond faster, as we discussed, because these days, when it comes to rainfall, when it comes to extreme weather events, is not only about the amount of rainfall, for instance, it's also about the time.
And I think, when we're talking about the alert system, the blue yellow, orange and red alert system is another way to simplify the procedure.
When certain things happen, how should we respond to that?
And we need a very clear system so that everyone knows their job when things happen.
Because when it comes, for example, floods, you probably have a few hours to get prepared.
So you need to be very clear about, for example, I'm with the meteorological authority.
What should I do in this kind of scenario?
And when it comes to transport, when it comes to civil affairs, when it comes to rescue efforts, who should be doing what at what time is also very important.
So I think that is another way of also, on top of technology, another way of dealing and mitigating with a lot of challenges brought by extreme weather events.
The extreme weather events.
It seems like it's happening more and more often around the world, which is what makes predicting things so hard, because if you don't know what's going to happen next, you can rely all you want on things that have happened in the past.
But if what has happened in the past is not likely to happen in the future, meaning it's about to be more extreme and more severe and more dangerous, it makes planning almost impossible in some situations.
There was a case earlier this year.
It was between January and February.
There were nine named storms that hit Spain and Portugal and Morocco, and some parts of Spain received a year's worth of rain in a few days.
So there are different reasons for that.
Scientists go into it.
They say that the world weather attributions.
Their scientists found that warming made the rainfall more intense, up to 36 stronger in southern Spain.
Different percentages for different regions.
But the problem with models and this is the troubling part is that climate models they failed in that situation to reproduce the observed trend.
So they showed only an 11 increase in the north and nothing in the south, even though the real world data showed much higher.
So that means current models may be underestimating future risks.
So that's kind of my point is that the models that are being designed.
It's very difficult to design to predict something that is unpredictable.
Because models were built based on past experiences.
So if things are going to happen in the future, never happened in the past.
For example, I remember a couple of years ago in Beijing, we also saw record level flooding.
And when it comes to this kind of reservoir and flood control system in place, I vividly remember one of the scientists said yeah, we were prepared.
We were prepared for 200 years.
We sort of accumulated 200 or even 1,000 years of records, and then we built this reservoir.
But this time, when it comes to that rainfall, it's 1,000 or a few hundred years level of record.
That's not been seen in the city for several hundred years.
So there is no way that you can be 100% prepared.
And also when it comes to extreme weather events.
For example in India, I think very recently, is seen also record level heat waves.
Some regions are approaching 50 degrees Celsius.
58.
Yeah, so how should they predict that in this summer I'm going to see a 50 degree Celsius weather out there?
And even if they knew, Even if India and a lot of Indian people know that we're going to embrace extreme heat waves this year, then we're talking about response measures, for example, installing ACs, building more roads that can survive in these extreme heat weathers.
It's not a few months of efforts.
It probably takes years of planning and also technological upgrades.
So that's another story as well.
So it's not as easy as we think that you know living in a more extreme weather conditions is could be on the book.
Yeah, and even for countries that are stereotypically known for receiving a lot of rainfall, such as the UK, it's not always the same, and they're
The website used it as a level of wet.
Weathers we have nowadays are also increasing, and they're making their preparations too.
But it's been hard even for them.
So the latest state of the UK climate report indicates that the UK has become wetter than over the last few decades, although with significant annual variations too.
So yeah, that agrees with what you said, too steve.
Even though we have all the preparations done, there's models constructed, digital avatars used for predictions.
Not every year has the same conditions.
This year could be higher than next year.
It's probably not as high.
So they are kind of finding this tricky because in the meantime they think tracking flooding events within UK that's caused by extreme rainfall is much trickier, according to their National Meteorological Service for the UK too.
Because flooding there can be affected by just many different things, many factors including, but not limited to, river flow rates, local soil type and the presence of flood defenses.
So no, this is not just a China issue.
It's happening worldwide.
Sure is.
And we don't mean to say either that these AI advances or tech advances are a waste of time.
No, not at all.
These are extremely, extremely helpful.
I mean, just that one example, the digital twin example.
If you can put any type of scenario into that digital twin to see how it's going to react.
This is going to help people.
This is going to help governments and scientists and communities better prepare for extreme situations when they do come in the future.
But beyond the faster warnings and the smarter monitoring, how advanced has China become in holding, or building, excuse me, the underlying AI, the satellite infrastructure that powers help to monitor disaster response here in the country?
Yeah, and Steve is quite spot on when we're talking about satellites, because when we're talking about building a digital twin system, the first thing is we need to have eyes in the system.
And who can be the eyes in the satellite?
The satellites, some sensors.
And that is why we heard a lot about spatial temporal industry, meaning industries that not only can cover what's happening on the ground, but also have capacities in the space.
So you can get a full lens of what's really happening.
So that also means, for example, we need a high-precision space-ground integrated network.
That means, for example, we need technologies like the BADL satellite navigation system, which can empower technologies like 5G IoT, Internet of Things and AI.
And that leads to we heard a lot of news reports, of course, also talking about Beto system now provides centimeter level accuracy for monitoring infrastructure.
And that refers to monitoring what's happening on the rivers.
Yeah, the river embankment soil erosion situation.
Things like that.
You said centimeter level accuracy.
That's impressive technology.
Yeah, each centimeter counts because when it comes to rainfall.
And there's not only Beidou, but also the Fengyun 4 satellite infrastructure that now provides the high resolution atmospheric data necessary for AI models to improve heavy rainfall forecasting accuracy by more than 15.
So that means, in the meantime, these systems, they're also used not only as just national data collecting, but also affordable or accessible For normal day-to-day people, you and I, because as of the end of 2025, there are over 2.2 billion beta-enabled terminals, and that includes smartphones, and 100 million vehicles in use, meaning that people can see weather reports and the very accurate prediction of weather from their phone or in their smart car systems that can just showcase before you go out or things like that.
So it's really creating a ubiquitous ground sensing network that feeds real-time data back into disaster response systems and that's relatable to all of us.
Yeah, and on top of that, with a lot of satellites and sensors that can enable us to build, for example, ai powered platforms to run models to make prediction of what's going to happen, for example, 12 hours from now, if the rainfalls continue like this.
And also, I think earlier this year China also launched a AI-powered sort of early warning platform that's basically open source to the whole world, called MaZu, which is named after a goddess in southern China, and yeah.
So, for example, they are collaborating with countries like Mongolia and some Pacific island countries so that they can also get the data that's been collected by the Fengyun Satellites or the Beidou Satellite Navigation System and then run some of the models on that MAZU AI platform to see if, for example, I saw a case in Mongolia They sort of predict a certain storm is going to lash out a certain region at a certain time.
And they then issued alert warnings to the local harassment and that can prevent them and saving them a lot of losses because of the storms.
So that is what also I think align with what the United Nations been calling for early warnings for all.
Basically meaning for a lot of countries, especially developing countries, they don't have the power to have a big satellite network in space.
Of course not, yeah.
And China's been helping out a lot with that.
The country's satellite and disaster response systems have been exported to over 140 countries, which is a huge help indeed.
AI weather forecasting is helping.
Digital Twin Rivers is helping.
Real-time evacuation systems, satellite-powered monitoring networks, all helping.
China's flood control strategy is really undergoing a very significant transformation.
And, as we discussed before, it relied heavily in the past on human experience, but it's very much increasingly becoming a data-driven system.
It's built around prediction and simulation and rapid coordination.
And as climate change makes extreme weather more frequent and less predictable all around the world, the challenge facing governments all around the world is no longer simply responding to disasters after they happen, but anticipating them before they strike.
That is the goal.
