search

earth-system-radiation / pyRTE-RRTMGP

pyRTE-RRTMGP provides a Python interface to the RTE+RRTMGP Fortran software package
https://pyrte-rrtmgp.readthedocs.io/
BSD 3-Clause "New" or "Revised" License
7 stars 0 forks source link

Add more usage examples #28

Open tcmetzger opened 7 months ago

tcmetzger commented 7 months ago

To make it easier to start using this library and illustrate its various use cases, we want to add usage examples highlighting the additional cloud and aerosol- related functionalities (Jupyter notebooks in the repository’s /examples folder)

tcmetzger commented 1 month ago

@RobertPincus could pull one time-step from aqua planet (a round planet without elevation)

sehnem commented 1 week ago

We need to add examples that use different paths on processing the data. Currently just two functions of RTE are used, we should implement all the functions and use them.

RobertPincus commented 1 week ago

@sehnem I can do that. Do you mean to extend the testing suite?

sehnem commented 1 week ago

@RobertPincus it would be good to have the same examples in fortran too, so we could use the same output data to test both and make sure it is consistent, it can be just examples, in python I use the examples as integration tests.

RobertPincus commented 1 week ago

@sehnem I have such examples in tests/test_variants; I can think about how to incorporate them.

sehnem commented 1 week ago

@RobertPincus I think that this is what I need, will use this as reference and implement it in python.

RobertPincus commented 1 week ago

@sehnem The most widely-used variant that's different from the clear sky is longwave clouds for which we use two-stream optical properties. For this we will need the cloud-optics in Python, as well as the ability to add two sets of optical properties together. The latter I can help with. The Python cloud optics is #53 and should be ready for Makepath to implement.

RobertPincus commented 1 week ago

@sehnem In the Fortran repo examples/all-sky would be another good example. It uses idealized profiles of temperature, humidity, and ozone; adds a bunch of other gases as scalars, and adds cloud to 2/3 of the columns. The clouds employ the 2-stream/LW logic. There are reference answers in the data repo.

The all-sky example might also be useful because it's very flexible - users can pick the number of columns and the vertical discretization, and we sometimes use it to assess timing.