on march 30, 2022, i will give a guest lecture at the urban water interfaces research training group, tu berlin:

climate emergency calls for process-based modelling

30 march 2022, 2 pm, online


work in progress. simple graphical interface to view files:

+ written in c, because their graphics library is easy to use.
+ can filter by tags. tags have to be assigned in the bibtex file.
+ can export a single reference to bibtex format and formatted text.
+ to do: capability to filter by document type.
+ to do: open file by clicking on it, using the plumber.

things i don't need/want:
+ looking up citations online, cloud storage.

still particle tracking. finished implementing capability for infiltration into the subsurface and exfiltration to the surface.

Show thread

i'm back to particle tracking. i found a way to significantly speed up everything.

finalised the implementation of a wavelet-based multiresolution mesh generator for rainfall-runoff simulations. multiresolution meshes try to guess where flow in the region will be complex. usually we use topography and assume that smooth regions can be kept coarse, because flow in these regions will be smooth and can be captured accurately at coarser resolution.

+ implemented in c
+ depends on the triangle mesh generator library
+ shared memory parallelised

second entry in our series.

we watched "sunset boulevard", a bizarre film noir from 1950. a failing screenwriter is found shot dead in the swimming pool of a has-been silent movie film star.

LA films watched: (1) miracle mile; (2) sunset boulevard.

We just binge-watched We Are Lady Parts and it's great! Watch it if you have the chance!

Working version of wavelet analysis-based unstructured mesh generation. I carry out a wavelet analysis of flow metrics to get the mesh below. Regions with smooth flow are kept coarse, regions with rapid changes are refined.

Just watched Miracle Mile. It's a weird little 80s film that starts as a romantic comedy and ends with nuclear war.

Almost finished with the particle tracker. I can now track across surface and subsurface domains (left) and get the residence time along the paths (right).

We observe that in this case, once a particle exfiltrates, it's quickly advected to the stream and does not re-enter the subsurface.

First "working" version of the particle tracker in the subsurface.

Particles follow the velocity field computed with a Richards solver.

I had to use local time stepping per particle to cope with the large differences in the time scales. There's a lot of infiltration and exfiltration going on.

They have a lot of bonsai trees at Huntington Botanic Garden. I liked this one the most.

Particle tracking again.

Figure below shows flow paths in a subcatchment of East River Watershed in Colorado. Not all particles reach the outlet.

Update on particle tracker — Can now track if the particle is leaving the domain. Basic shared memory parallelisation implemented for the outer loop.

Wrote a small particle tracker.

Figure below shows the trajectories of 20 particles released in Thies Catchment, Senegal.

Next step is coupling with subsurface flow, so I can track particles infiltrating into the soil.

Show older
Scholar Social

Scholar Social is a microblogging platform for researchers, grad students, librarians, archivists, undergrads, academically inclined high schoolers, educators of all levels, journal editors, research assistants, professors, administrators—anyone involved in academia who is willing to engage with others respectfully.