Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Sure, but why didn't they come up with it 5-10-20 years ago? Technological reasons?


The short answer is yes. To characterize sources like this you need a very small pixel footprint size on the earth, which reduces the magnitude of the signal you actually observe, leading to higher technological requirements.


It's a long road. I'll enlarge a bit on @dannyz's explanation you can see nearby.

I happen to know one thread of work on the space-based CH4 detection problem...here is one early publication, from 2011:

https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/201...

Not sure if the paper is behind a paywall for you, but you can see the xCH4 maps in Figs. 14 and following, on a 30km (1/3 degreee) grid. xCH4 = total air-column concentration of CH4, integrated from ground to top-of-atmosphere. The first author of the above paper, who is remarkably talented, continues to work on this problem all these years later.

The maps you see there average many years of measurements to decrease noise. The signal is there, but you have to average a lot of measurements to extract it, and then your resolution is not helpful for the "real-time-detection" (leak) problem (small spatial/temporal footprint => averaging washes it out).

The xCH4 signal is noisy partly because the CH4 is being measured spectroscopically, so you have to get a (bunch of) very fine slices of spectrum to see the CH4 absorption lines. There just aren't that many photons, and the detectors weren't that efficient. So yes, that's a technological/fabrication problem. Lots of people have worked to improve spectroscopic detectors over the meantime.

Additionally, you have to disambiguate the CH4 signal from others that are much more impactful on spectra, like H2O vapor and CO2. You end up with a nonlinear inversion problem ("fiddle with the gas concentrations to optimize the fit of an atmospheric radiative transfer model to the observed spectrum"). This is a technical/modeling and algorithms problem, and again, lots of people have worked to improve models and methods, and on validation that they work.

There has been a succession of instruments of higher capability launched to address this problem (OCO-2 by NASA, Tropomi by ESA as seen in OP). Usually you fly them in planes (https://avirisng.jpl.nasa.gov) and move them to space as they are shown to work.

And also a succession of very impactful papers in PNAS, Science, and Nature that discuss advances. And there will be more.


need somebody with money. oil and gas companies are certainly interested in not paying for this.




Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: