Add Montornes's solar eclipse modelling

[irowebbn]

TYPE: new feature

KEYWORDS: radiation, eclipse

SOURCE: Isaac Rowe (University of Kentucky); original source: Alex Montornès (University of Barcelona), internal

DESCRIPTION OF CHANGES:
Alex Montornès developed a modification to the radiation schemes for WRF 3.6 to model the effect of eclipse based on
the Bessel's method. This PR aims to bring the work back into the latest version of WRF for four shortwave schemes:
RRTMG, Dudhia, Goddard, and old Goddard.

LIST OF MODIFIED FILES:
Makefile
Registry/Registry.EM_COMMON
Registry/Registry.NMM
clean
dyn_em/module_first_rk_step_part1.F
dyn_nmm/module_PHYSICS_CALLS.F
phys/Makefile
phys/module_ra_eclipse.F
phys/module_ra_goddard.F
phys/module_ra_gsfcsw.F
phys/module_ra_rrtmg_sw.F
phys/module_ra_sw.F
phys/module_radiation_driver.F
run/eclipse_besselian_elements.dat
run/README.namelist
share/module_check_a_mundo.F

TESTS CONDUCTED:

1. Jenkins tests all passed.
2. The model is run with and without the eclipse mods for the August 2017 eclipse event over US. A figure is shown
   below: the left is SWDOWN valid at 1800 UTC without eclipse mod, and right is with eclipse mod:
   

RELEASE NOTE: Add a new option (ra_sw_eclipse) to model the effect of eclipses based on the the Bessel's method in four radiation schemes: RRTMG, Dudhia, Goddard, and old Goddard. Additional output includes the location of the eclipse: ELAT_TRACK and ELON_TRACK. For reference, see this paper: Montornès, A., Codina, B., Zack, J. W., and Sola, Y.: Implementation of Bessel's method for solar eclipses prediction in the WRF-ARW model, Atmos. Chem. Phys., 16, 5949–5967, https://doi.org/10.5194/acp-16-5949-2016, 2016. (Thanks to Alex Montornès of University of Barcelona and Issac Rowe of University of Kentucky.)

[dudhia]

Since the changes to rrtmg_swf were more extensive than the others, I would prefer to leave that out of this set of changes, because that is hard to review. Even rrtmg_sw has much simpler changes and is OK. The other schemes have easier to understand changes.

[dudhia]

Thanks. Also remove from list of modified files in PR.

[davegill]

@irowebbn
Isaac,
Since you have added this file run/eclipse_besselian_elements.dat, there should be a couple of modifications to additional files.

1. The top-level Makefile links files from the run directory into the test/em_real directory. Look for places where this file is mentioned: CAMtr_volume_mixing_ratio.RCP4.5. Basically you will see a template to follow.
2. The script that cleans up after a build should also have this new file mentioned, in this case to remove it from the test/em_real directory. Look in the top-level clean script for the file name CAMtr_volume_mixing_ratio.RCP4.5. Again, just follow the implied template.

What we want after the end of these two mods is that your new file is now automatically placed in the test/em_real directory during a build and also removed from the test/em_real directory during the clean-up step.

[davegill]

@irowebbn
Isaac,
What is the status of the jenkins testing? Your should have received an email with a text-based table. Would you please post that email in one of these comment boxes (not the attachment, just the text of the email).

1. Is the data file binary or just text? This impacts the real4 vs real8 builds.

[weiwangncar]

@davegill The data file is a text file, about 80 Kb in size.

[vikramjakhr]

reopening for testing

[weiwangncar]

@irowebbn I think there may be a chance you mixed up the branches you use to push up changes. In your local fork, you have a develop branch which you used to push up some changes. But your original development may be in branch wrf-eclipse. When I used that branch, you have code in dyn_em/module_first_rk_step_part1.F to pass new eclipse variables and namelist to radiation driver. But this code is missing in your develop branch. If you try to compile code in your develop branch, it will likely fail. Can you check please?

[weiwangncar]

@davegill Would you also like to approve this PR?