Inconsistent solar zenith parameter naming in decomposition functions

[ramaroesilva]

Describe the bug
Among the decomposition models, solar zenith assumes two parameter names depending on the function:

• zenith (erbs, erbs_driesse, orgill_hollands, campbell_norman, dirindex )
• solar_zenith (louche, disc, dirint, boland).

This demands adjustments in the code when a user wants to run different decomposition models for comparison.

Expected behavior
Assume a harmonized parameter naming across all functions, solar_zenith seems good to me :-)

Additional context
Relates to #1403. Can be a good issue for first-timers.