search

nanograv / PINT

PINT is not TEMPO3 -- Software for high-precision pulsar timing
Other
119 stars 101 forks source link

`PLDMNoise` doesn't work correctly with wideband TOAs #1660

Open abhisrkckl opened 11 months ago

abhisrkckl commented 11 months ago

PLDMNoise is a GP representation of DM variations. Hence, while computing the wideband residuals, it should be applied to both TOAs and DMs. However, we don't have the framework to do this properly for a correlated noise component, and thus PLDMNoise is only applied to TOAs. This will lead to wrong results while analyzing wideband data.

abhisrkckl commented 11 months ago

@blarsen10

blarsen10 commented 11 months ago

My limited understanding of the other DM modeling features in PINT is that the DispersionDM and DispersionDMX classes apply only to ToAs, while the ScaleDmError and DispersionJump classes apply only to wideband DMs. My intention when adding PLDMNoise was that it would behave similarly to DispersionDM and DispersionDMX, which means applying to ToAs but not wideband DMs. However, if my understanding is wrong and the DispersionDM and DispersionDMX classes also work with wideband ToAs, then maybe PLDMNoise should also do so in a similar way. Is that the case here? Or is there some other reason why this PLDMNoise should be made to work with wideband DMs? Apologies that I am not up to speed on the requirements for wideband analyses right now.