Add files for radar Doppler wind processing

[ctgh]

Add files for radar Doppler wind processing:

• CX and Varobs namelists.
• UKV unit tests for the CX and VarObs writers.
• Varobs unit test for radial velocity (varobs index 63). I removed the test for radar azimuth (index 66) because it is not appropriate. Due to the setup of the VarObs writer, when radial velocity files are created they automatically include indices 64, 65, 66, 69 and 75.
• Addition of several radar variables to opsinputs_varobswriter_mod.F90. These auxiliary variables are put into the VarObs data structure as described above.
• Addition of the function opsinputs_fill_fillinteger2d in opsinputs_fill_mod.F90. This enables the variable MetaData/stationIdentification (which in this case is mapped to an integer rather than a string) to be processed correctly. NB 1: a 2D function is required here due to the OPS data structure. NB 2: I did not copy all of the functionality in opsinputs_fill_fillreal2d to this function. Instead, that can be added as necessary. Given opsinputs is quite mature I suspect that may not be required.

Fixes #211

[ctgh]

Thanks for both of your reviews! I will respond today.

[ctgh]

@ReubenHill I have implemented all of your suggestions in a662ba6.

[ctgh]

In 591e7f4 I updated deps/ops/stubs/OpsMod_Varobs/Ops_VarobPGEs.inc in order to write out the radial wind PGEs correctly. The main issue with the existing code is that the logical variable RadWind_SuperOb will never be true in opsinputs (whereas it is true in OPS). I decided to remove the test on that variable and ensure that the correct PGE is written out.

@mikecooke77 As mentioned in the comment above, I have modified one of the OPS stubs because it was leading to inconsistent behaviour relative to the actual OPS code. This change should not be ported into OPS itself. Do you foresee any issues with doing that?

[mikecooke77]

In 591e7f4 I updated deps/ops/stubs/OpsMod_Varobs/Ops_VarobPGEs.inc in order to write out the radial wind PGEs correctly. The main issue with the existing code is that the logical variable RadWind_SuperOb will never be true in opsinputs (whereas it is true in OPS). I decided to remove the test on that variable and ensure that the correct PGE is written out.

@mikecooke77 As mentioned in the comment above, I have modified one of the OPS stubs because it was leading to inconsistent behaviour relative to the actual OPS code. This change should not be ported into OPS itself. Do you foresee any issues with doing that?

The issue comes when you try and update the code for changes to the OPS e.g. we have a new datatype. If you copy straight from OPS changes like this will disappear. At one point we were thinking of having some sort of "copy fixes back in" kind of functionality but I don't think this is ever used.

[ctgh]

In 591e7f4 I updated deps/ops/stubs/OpsMod_Varobs/Ops_VarobPGEs.inc in order to write out the radial wind PGEs correctly. The main issue with the existing code is that the logical variable RadWind_SuperOb will never be true in opsinputs (whereas it is true in OPS). I decided to remove the test on that variable and ensure that the correct PGE is written out. @mikecooke77 As mentioned in the comment above, I have modified one of the OPS stubs because it was leading to inconsistent behaviour relative to the actual OPS code. This change should not be ported into OPS itself. Do you foresee any issues with doing that?

The issue comes when you try and update the code for changes to the OPS e.g. we have a new datatype. If you copy straight from OPS changes like this will disappear. At one point we were thinking of having some sort of "copy fixes back in" kind of functionality but I don't think this is ever used.

I see. Presumably I could a test to cover this eventuality (i.e. check the output PGEs are as expected). If the file is ever overwritten the test will start to fail. Do you know if this version of OPS code has been updated?

Failing that, I might be able to hack things in the varobswriter. I'll take a look at that too.

[ctgh]

In fa19be4 I removed the modification to the OPS code. Instead both Ob % RadialVelocSO and Ob % RadialVelocity are filled. It's a bit hacky but causes the PGE to be filled correctly. The unit tests pass, but I am also checking the full end-to-end test. I will let you know how it goes...

[ctgh]

In fa19be4 I removed the modification to the OPS code. Instead both Ob % RadialVelocSO and Ob % RadialVelocity are filled. It's a bit hacky but causes the PGE to be filled correctly. The unit tests pass, but I am also checking the full end-to-end test. I will let you know how it goes...

I can confirm that this change does not affect the PGE output in a full test.

[ctgh]

Thanks all for your reviews!