Should a deprecation-warning period be established? I'm not aware of the internal methodology you guys make use of, but I've seen that changing params have been deprecated before in this project. See #773.

The previous definition was also not correct. IMO it doesn't you can't use the term "broadband" about effective irradiance which has potentially been modified for spectral effects.

g_poa_effective;broadband plane of array effective irradiance.

If you apply a dimensionless spectral factor correction to a value of broadband irradiance in Wm⁻², it's still a broadband value in Wm⁻² but just a fraction of the original broadband value. Broadband just means its the irradiance between a (wide) wavelength range, no?

Should a deprecation-warning period be established? I'm not aware of the internal methodology you guys make use of, but I've seen that changing params have been deprecated before in this project. See #773.

This would be ideal. But when it's just a rename I don't see how it can be deprecated nicely.

I don't think I agree with this change of name g_poa_effective to effective_irradiance. The current description is of a quantity that is not poa_global but also is not effective_irradiance which is supposed to be after accounting for reflections, soiling, and spectrum. I agree that the description can be improved.

Just a comment about names: sometimes variable names were chosen to more closely connect the code with the text in the reference. Didn't happen here (the reference uses :math:I_{tr} so I'm guessing the submitter of this code came up with g_poa_effective as a compromise :math:G_{POA} being common in literature, and adding "_effective" because reflections have been removed.

    g_poa_effective: numeric
    effective_irradiance: numeric
        Irradiance transmitted to the PV cells. To be
        fully consistent with PVWatts, the user must have already
        applied angle of incidence losses, but not soiling, spectral,
        etc. [W/m^2]

This would be ideal. But when it's just a rename I don't see how it can be deprecated nicely.

@AdamRJensen , we can work it out, similarly to #773 with some little variations. Hint: signature will need to be:

def pvwatts_dc(self, effective_irradiance=None, temp_cell=None, *, g_poa_effective=None):

and the warning when it's used:

if g_poa_effective:
    warnings.warn("Use 'effective_irradiance' to suppress this warning", pvlibDeprecationWarning)
    effective_irradiance=g_poa_effective
    tests of mutual exclusion

then the warning test.

because reflections have been removed

If I understand it well, then [post_]iam_irradiance, iam_modified_irradiance, irradiance_after_iam, poa_global_after_iam, poa_g_after_iam may be options that fit the description better. What do you think about it @cwhanse ?

I'm +1 to changing the parameter name and description, +1 to using a self-explanatory parameter name, +1 for poa_global_after_iam if I had to choose, not a strong opinion.

The current description is of a quantity that is not poa_global but also is not effective_irradiance which is supposed to be after accounting for reflections, soiling, and spectrum. I agree that the description can be improved.

We already have a definition of effective_irradiance, which includes both reflection and spectral effects:

So the definition proposed here (AOI correction only) would conflict with the existing definition above.

I was just reading through the pvwatts manual now --- page 9 eqn 8 --- perhaps we could just just use the same "transmitted" terminology here instead, i.e. transmitted_irradiance?

"transmitted" terminology here instead, i.e. transmitted_irradiance?

The De Soto model paper uses the term "absorbed irradiance" for this quantity, symbols $S$ and $S_{ref}$

If this parameter name changes, we should consider doing it in combination with broader changes to that function (name and module, see #1350 and #1544 (comment)).

PVWatts v5 doesn't account for spectrum (so you could say implicitly $M=1$). Soiling is applied as a generic loss factor after DC power is modeled. I agree that effective_irradiance is not quite right for the quantity in question, but that bothers me less than inventing/adopting a new term would. I suggest renaming to effective_irradiance and listing the caveats in the parameter description.

If this parameter name changes, we should consider doing it in combination with broader changes to that function (name and module, see #1350 and #1544 (comment)).

PVWatts v5 doesn't account for spectrum (so you could say implicitly M = 1 ). Soiling is applied as a generic loss factor after DC power is modeled. I agree that effective_irradiance is not quite right for the quantity in question, but that bothers me less than inventing/adopting a new term would. I suggest renaming to effective_irradiance and listing the caveats in the parameter description.

+1

I suggest renaming to effective_irradiance and listing the caveats in the parameter description.

The distinction between poa_global and effective_irradiance is probably not important to typical users of pvwatts_dc. If we are choosing to rename g_poa_effective to one of the already-used terms, I would prefer poa_global rather than effective_irradiance.

Although I'll point out the irony of this conversation at the same time we are discussing various instances of inconsistent parameter names.

I would vote for effective_irradiance because that's consistent with the other DC module models/functions.

(not sure if my earlier comment was holding this up so, just in case, a follow up comment...)

In general I am not keen on reusing an existing variable with a different definition (comment), but in this specific case I think @kandersolar's comment

(so you could say implicitly M = 1 )

addresses the overlap to a sufficient extent so I'll leave my vote neutral (not opposed) on effective_irradiance

@echedey-ls I think I fixed it as per your suggestions - any chance you could take another look? 😄

Whatever change we make here, it would be nice (though not required, since we're just deprecating, not removing) to make it in the .0 release this week.

If we are choosing to rename g_poa_effective to one of the already-used terms, I would prefer poa_global rather than effective_irradiance.

@cwhanse any chance this preference has changed since September? I think yours is the only voice of opposition to effective_irradiance.

FWIW I still think the reasoning in #2235 (comment) is sound; PVWatts v5 accounts for soiling elsewhere, and spectrum not at all, so IMHO the irradiance quantity in question is effective_irradiance in spirit as far as PVWatts v5 is concerned. I don't think calling it poa_global is suitable given that it's supposed to account for IAM.

FWIW the original PVWatts PR does have some discussion about parameter names. #195 (comment)

I suspect that many users of pvwatts_dc are applying soiling or spectral losses before calling this function. So my vote in 2025 is to call it effective_irradiance, note the discrepancy with the literature, and move on. I don't think we want to introduce something like the originally proposed irradiance_transmitted or variations.

any chance this preference has changed since September? I think yours is the only voice of opposition to effective_irradiance.

I'm ok with effective_irradiance.

I concur with the reasoning in #2235 (comment) which, in pvlib terms, is equivalent to pvwatts_dc setting spectrum_model='no_loss'. From this perspective the input is effective_irradiance.

In fact, that leads to a possible definition of effective_irradiance: broadband plane-of-array irradiance reduced by reflections, soiling and adjusted for spectrum.

That's how effective_irradiance is usually computed for PV modeling. In the context of model calibration from outdoor measurements, effective_irradiance is different: the irradiance estimated from measured Isc.

In fact, that leads to a possible definition of effective_irradiance: broadband plane-of-array irradiance reduced by reflections, soiling and adjusted for spectrum.

I'm still in favour of using this variable name here, but not convinced about the definition.