By Sam Wong
This week, social media users have been sharing photos of themselves from 10 years ago alongside pictures from today and the hashtag #10YearChallenge. That got us thinking about how the world of science has changed in that time. From retreating glaciers on our home planet to the deepest reaches of our solar system, the last decade has seen great advances in our knowledge as well as stark changes in our natural world. Meanwhile, computers have learned to create completely imaginary but photorealistic images. We chose four pairs of images that tell different stories about the last 10 years in science.
Columbia Glacier, Alaska
The Columbia Glacier is one of the most rapidly changing glaciers in the world. Since 1980, it has been in retreat, as shown in these false-colour images from NASA’s Landsat satellites. Its terminus has moved more than 20 kilometres to the north in the last three decades. Chunks of ice have broken off at the bottom, contributing to global sea level rise. At the same time, the glacier has thinned substantially, expanding the area of brown bedrock visible in the pictures.
Ten years ago, Pluto was at a low point. The International Astronomical Union stripped it of planet status in 2006, reclassifying Pluto as a dwarf planet. Pluto was out of favour, and the best images we had of it were blurry photos from the Hubble Space Telescope.
In 2015, that changed, as NASA’s New Horizons spacecraft gave us a closeup view of Pluto for the first time. The beautiful pictures it sent back enchanted and surprised us with its hazy atmosphere, smooth expanses and frozen mountains, winning Pluto a new generation of admirers.
Fake faces generated by neural networks
In the last 10 years, computers have learned to create life-like photos of entirely imagined people. Neural networks work by analysing a large set of data, such as photographs, and looking for patterns. In a particular type called a generative adversarial network or GAN, part of the system generates fake data while another part tries to work out if it is real or not. By pitting the network against itself, it learns to produce better images. As the images below show, the rate of progress in this field has been astonishing.
In 2008, New Scientist reported on the first forecasts to try and predict how the climate would change over a 10-year timescale. A model built by the Met Office Hadley Centre, the UK’s official centre for climate change research, predicted that surface air temperatures would remain steady for around six years as cool sea surface temperatures kept the atmosphere cool. But, it expected surface temperatures to rise again by 2014 and hit “a string of record highs at the end of the next decade”. That has proven to be pretty close to the mark.
Despite the wealth of knowledge we have about climate change, the intervening years saw precious little progress on steering us away from impending catastrophe. On a cover in 2018, we asked whether humanity is getting more stupid, but we answered the question with a no. “If we perceive the world to be dumbing down, perhaps that is because, as science expands our knowledge, we see a widening gap between the rational solutions it suggests and the messy reality of the world,” we wrote.
More on these topics: