try to stick as much as possible to native model files and settings

openearth / glofrim

Globally Applicable Framework for Integrated Hydrological-Hydrodynamic Modelling (GLOFRIM)

GNU General Public License v3.0

46 stars 27 forks source link

try to stick as much as possible to native model files and settings #81

Closed DirkEilander closed 5 years ago

DirkEilander commented 5 years ago

right now CMF still needs a tif file for the nextxy data and LFP needs a refdate settingsin the par file which are not native to the models. I think we should try to minimize this kind of non-native data and settings in the setup. Are there other non-native data / settings that we should have a look at?

[x] replace required nextxy.tif with reading directly from binary file
[x] make LFP refdate setting optional. glofrim v2 does need a reference data, but we can provide a warning and use (random) default if it is not provided. If coupled we can set the default form upstream model.
[x] don't realy on PCR landsmak file for grid props. Use fallback to ldd if it is not given.

JannisHoch commented 5 years ago

Hi @DirkEilander that's definitely a valuable remark. Cannot think of another one at the moment. Think if we solve these two (particularly the first one!), we should be fine for now and see how further sharing/testing/application goes.

What's a realistic time line to achieve this?

DirkEilander commented 5 years ago

@hii600 @ChippChapp I've worked on the issues above in the new use-native branch. Could you guys double check if the new branch is working for you?

DirkEilander commented 5 years ago

we need to especially test the CMF updates on a global model!

JannisHoch commented 5 years ago

@DirkEilander Do you mean with "testing the CMF updated on a global model" that we should run the global CMF model via GLOFRIM stand-alone mode?

DirkEilander commented 5 years ago

@ChippChapp With a stand alone model the updated part is not used. We need to test with a coupled global CMF to nested LISFLOOD model. The thing is that for a global model CMF uses multiple files to index hires gridcells to the unit catchments. It would be good to test if that works with GLOFRIM.

[ ] test with a coupled global CMF to nested LISFLOOD (regional/local) model