Add more usage examples

[tcmetzger]

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]

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

[sehnem]

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]

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

[sehnem]

@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]

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

[sehnem]

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

[RobertPincus]

@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]

@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.

[tcmetzger]

From the 12/1/2024 call: We want to find a well-known, standardized example dataset. We then want to show how pyRTE can be used to process this data efficiently (using dask #54)

[RobertPincus]

ERA5 reanalysis would be compelling. The issue is that it lacks some fields so we'd need to interpolate.