The Coming Flood: A Data Error Is Corrected, and Our Future Is Rewritten
What’s happening in California right now — blackouts, wildfires, mass evacuations — should tell you everything you need to know about the urgency of the climate crisis. But the hard truth is, the flames in California are just one aspect of life on a superheated planet. Every once in a while, climate scientists publish a new study that clarifies just how dire the climate crisis really is for millions of people around the world.
One such paper was published in Nature Communications this week by scientists at Climate Central. The paper corrects what was essentially a data error in previous calculations about how many people are at risk from rising seas. The results stunned even some of the world’s top climate scientists:
This study by colleagues from Climate Central shocks even me, as I was not aware of this problem. It’s not about higher projections of sea-level rise; it is more accurate land elevation data, showing that many areas are closer to sea level and thus more vulnerable than thought. https://t.co/sNvVXn8HDf
— Stefan Rahmstorf (@rahmstorf) October 31, 2019
According to the new analysis, roughly three times as many people are at risk from being inundated by rising seas than previously reported — 150 million people are now living on land that will be below high tide by 2050. And this is a hopeful scenario, where warming is held to 2 degrees C and the ice sheets don’t collapse in the near future. In a more pessimistic scenario, the numbers double to 300 million people flooded out by 2050. That is roughly equivalent to the entire population of the United States going underwater in the next three decades.
The study finds that about 70 percent of the people currently at risk are in eight Asian countries: China, Bangladesh, India, Vietnam, Indonesia, Thailand, the Philippines, and Japan. As early as 2050, most of Southern Vietnam could disappear at high tide, putting 20 million people underwater. In Thailand, 10 percent of the population lives on land that will be flooded, compared to 1 percent now. Large parts of Bangkok and Mumbai, both financial centers in Asia and home to tens of millions of people, are at risk of being inundated in coming decades.
Why this big revision in our understanding of the risk of sea level rise?
Until now, calculating the risk of sea level rise has largely depended on two variables: first, how fast carbon emissions will continue to rise, which impacts how quickly the climate continues to warm; second, how sensitive the ice sheets are to that warming – especially West Antarctica, which is the wildcard in future sea level rise projections (there’s lots more about this in The Big Melt, a new audiobook I recorded about my recent trip to Antarctica).
But one largely overlooked variable is the actual elevation of land along the coast, as well as how many people live there. After all, when it comes to sea level rise, what matters is not just how high the water rises, but how high the land is where the water rises. Three feet of sea level rise in South Florida, which is low and flat and densely populated, will have a much greater impact than three feet of sea level rise along the high, rocky, sparsely-inhabited coast of Oregon or Maine.
For years, scientists have known that coastal elevation measurements have been approximate. The elevations are based on data from the Space Shuttle Endeavor mission in 2000, which used space-based radar to basically create a topographical map of the civilized world (the Arctic and Antarctica were left out). It worked fine in some regions, but the radar often misread the tops of buildings and treetops for the land surface, which lead to elevation measurements that were off by 10 feet or more. And that is a big problem when trying to calculate flood risk, where even 6 feet of elevation has big implications.
In the U.S., Europe, and Australia, problems with coastal elevation measurements were corrected with LIDAR (Light Detection and Ranging), which is basically a method of shooting lasers from airplanes to measure the contours of the ground. It’s far more precise than space-based radar, but it is also expensive and time-consuming.
But there was no publicly available LIDAR data for Asia, Africa, and the Middle East, making it difficult for scientists to accurately calculate risks of sea level rise in those regions. It was especially a problem in Asia, where the land is low and flat and heavy vegetation made it easy to mistake treetops for the land beneath.
“As early as 2014, we knew that elevation data was leading to a big underestimate in sea level rise risk,” says Ben Strauss, a co-author on the report. “We spent a couple of years looking for better data. Then about three years ago, ran out of patience. Instead of waiting for better data, we decided to make it ourselves.”
Strauss and Scott Kulp, a computer scientist at Climate Central, essentially figured out a way to use artificial intelligence to mimic the work of LIDAR. “We looked at 50 million data points across the coastal U.S. and told the computer, ‘Here is the elevation, here is what the space shuttle radar measurement thinks it is, and go make a model that fixes the error.’” Strauss and Kulp compared their model to data in other parts of the U.S. and Australia where they have accurate elevation numbers and discovered the AI model (known as CoastalDEM) accurately corrected elevations in some places from 10 feet down to one inch.
When they re-analyzed the space shuttle data with the new model, the results were astonishing. And not in a good way.
For one thing, Strauss and Kulp discovered that changes in risk due to previous mistakes in elevation data “exceeds the combined effects of emissions level, Antarctic behavior, and incorporation of annual flooding, as assessed using [space shuttle data].” In other words, a simple correction of elevation data dwarfed every calculation of how much seas will rise as a result of how much fossil fuels we burn, or how much they will rise if the big ice sheets turn out to be as fragile as some scientists fear. For example, the previous analysis of flood risk using the space shuttle data, under high emissions scenarios and Antarctic instability, found that 95 million people will be living below mean high water line in 2100. In the new analysis, even with a low emissions scenario and fairly stable ice sheets, 190 million people will be living below the water line.
To put it another way, with this new data, even the best-case scenario for future flood risk is twice as bad as the worst case scenario before.
The study is not all bad news. On the upside, Strauss and Kulp found that 110 million people already live in areas below high tide, mostly behind seawalls and other barriers. “That gives me some hope about our ability to adapt,” says Strauss, although he also stresses that those barriers and sea walls are unlikely to be sufficient as sea level rise accelerates in the coming years.
Important as this study may be, says Strauss, “it is not the last word on any of this.” New satellites with better technology will be launched in the coming years, opening the door for more precise space-based elevation measurements. And the study doesn’t consider the dynamic nature of the land itself, such as the erosion that occurs as sea rise. Or the fact that in many places land is sinking fast, often due to the pumping of groundwater. (That’s what is happening in Jakarta, where Indonesian President Joko Widodo recently announced he has decided to move the capital to higher ground in Borneo.) The study also ignores the fact that sea level rise is not equal everywhere. For example, from 2011 to 2015, a study co-authored by MacArthur Prize winner Andrea Dutton found that seas rose three times faster than the global average on parts of the Atlantic coast, in part due to the impacts of El Nino.
Like most data-driven climate science, the biggest thing the study leaves out is the human element. It does not consider what kind of new defenses might be built to protect against rising seas – walls, barriers, dikes. Or how people might adapt in other ways, with Venice-like canals or floating structures. Nor does it consider political and economic factors, such as how fast property values will fall as the risk of flooding rises and how that will impact tax revenues and local governments. Or how hard people will fight to stay on their drowning land, even if it makes more sense to flee. Or, most importantly, how many millions of people will be displaced, and where they will go, and how they will be treated by people on higher ground. It’s impossible to think about the ground flooding beneath hundreds of millions of people and not think about hardship, suffering, and war. In the end, the real X factor in this study has nothing to do with ground elevation or rising seas, and everything to do with the vast mysteries of the human heart and mind.