Resolution problems in postprecessing CMS output

[lalv...@uabcs.mx]

Hello to All

I'm having some troubles with the Landmask Boundary Module, I've change in the runconf.list file the option to avoid land (.true.) but when I plot my results there seems to be no effects on the particles trajectory. Ahh and I'm running backards

Does it has anything to do  with the data bases I am using?

I have made my on matlab function to plot trajectories and in it, I use gebco bathymetry data and gshhslandmask with full resolution.

3 files are attached

What are data bases used by the CMS to make the plots, can this be changed in a CMS script?

Thanks in advance

[bexeross]

Hi,

The landmask depends on the topography inside the ocean model, and not something within CMS. The best way to plot the coastline is to use the nest file fill values as they reflect exactly where the model thinks there is land (see the post proc file draw_traj_netcdf.m supplied with CMS for the code to do this - you just nest any one nest file as a template for plotting where the land is). 

Gebco may not match up with the coastline in the model you are using (it definitely doesn't line up well with Hycom) so you may find particles on land according to Gebco, but within your model they are not yet touching land. See my attached image I have used to show this in the past - the grey shows GEBCO (this is the west coast of ireland) the red is my best interpretation of what HYCOM sees as water mapping out the much coarser coastline that is inside the model.

I have definitely found this a problem in the past, especially when planning release points near land and having to reposition them several times to make sure that they are in the water in the model. I don't have a perfect solution for this at present - the hycom forum talks about ways of interpolating the hycom bathymetry using multiple files they supply, but there is no .tif to work with. I have now got a .tif that approximates the HYCOM bathymetry (as seen in the attached image, and spanning the globe) and can try to attach that too, but it is still not perfect (projection issues) and I honestly cannot remember where I got it from!

I also think this is a major issue when modelling and assuming you will get a realistic result out the other end. The coarse bathymatry is a huge simplification and modelled results should be interpreted with that in mind.

All of this is to say that CMS is probably working and your points probably are avoiding the land, but you just have the wrong map of where the land is inside the ocean model.

All the best,

Bex

[bexeross]

Here is a link to real_hycom.tif (on my onedrive for now) - a raster of something that looks a lot closer to hycom's bathymetry for those who need it. 

This is still not perfect, but will better approximate what is inside the model than e.g. GEBCO. 

Please note I cannot remember where I got this or whether I made it, let me know if you have it ir find it somewhere else and I'll attribute it properly. It might just be the ETOPO1 dataset? I must label these files better!

All the best,

Bex