Hacker News new | ask | show | jobs
by qewrffewqwfqew 3779 days ago
Huh, yesterday's submission of the same thing got no attention at all:

https://news.ycombinator.com/item?id=11125050

Transcribing my comment from there:

I really want to see some papers, models or simulations to illustrate some of these effects - starting with the gravitational influence of ice sheets on sea level. This shouldn't be a difficult thing to illustrate numerically, but wow it would have a big impact on how (at least I) perceive the ice sheets.

So far, I haven't been able to dig anything up - there's some prose at [1] but nothing hard. A poster at [2]. Mitrovica's website [3] doesn't seem to have anything. It's late, so I'll have to postpone the search for now, but here's hoping other readers can help me :-).

[1]: http://sealevelstudy.org/sea-change-science/whats-in-a-numbe...

[2]: http://geo.orst.edu/files/geo/Mitrovica-2009-Science.pdf

[3]: http://environment.harvard.edu/about/faculty/jerry-x-mitrovi...

2 comments

If you still have some FORTRAN compiler installed you can run some simulations yourself [1]. There are two difficulties in 'easily' solving this problem. First of all the high level of feedback: changes in the sea level anywhere influence all other sea levels (conserving mass!). So you cannot solve this problem locally. Secondly, solving problems on spheres is difficult and expensive, so most sea level models solve the problem in spherical harmonics which is faster. (It's like the Fourier Transform, but instead of decomposing 1D signals into sines/cosines, you transform spherical shapes in more fundamental 'blobs'. See the examples here: [2] , after the header 'Jouons un peu avec les coeeficients harmoniques sphériques de la topography Terrestre'. You can see the individual first 0-6 'blobs', which when added form the first image reconstruction de l = 0 à 6. Which kind of resembles the Earth's topography already.) Spherical harmonics have some desirable mathematical properties and allow for quick simulations, but obviously formulating a problem in a different domain masks the actual physics being modeled. So I doubt you'll become wiser by studying these models...

Maybe you can be more convinced through [3] . Paolo Stocchi developed for the SELEN FORTRAN program.

[1] http://www.fis.uniurb.it/spada/SELEN_minipage.html (nb. this requires GMT, nearly impossible to install on Windows, and needs to be built from source in Linux. But OS X's homebrew has got it easily.)

[2] http://www.geologie.ens.fr/~vigny/cours/chp-gphy-2.html

[3] http://www.tudelft.nl/en/current/latest-news/article/detail/...

There's a guy over at http://blogdredd.blogspot.com/ doing some modelling in C++. I have no idea how good or bad his modelling may be.