Appendix E — Retrieval: Input Uncertainty Model

Implementation status

Model	DITL	PACE V3	PACE V4	Evaluation	Planned
HARP2	x	x	x	—	—
SPEXone	—	x	x	—	—

As discussed in Chapter 2 and later in Appendix Appendix D, measurement uncertainties are used to weight the observations in the retrieval cost function. The uncertainty model accounts for contributions from instrument measurement uncertainty, NN forward-model approximation, and numerical uncertainty in the radiative transfer (RT) simulations used to generate the NN training dataset.

The RT and NN contributions are evaluated based on factors such as the RT model configuration, the number of Gaussian quadrature points and scattering orders, and the size and performance of the NN training dataset (Gao et al. 2021).

Instrument measurement uncertainties depend on the characteristics and calibration performance of each instrument and are based on information provided by the instrument teams. For the current FastMAPOL retrievals used in all the previous versions and studies [Gao et al. (2021); Gao:2023aa; Gao et al. (2026)], the following measurement uncertainties are adopted:

Instrument	Reflectance uncertainty	DoLP uncertainty
HARP2	3%	0.005
SPEXone	2%	0.003

The reflectance uncertainty is specified as a relative uncertainty, whereas the DoLP uncertainty is specified as an absolute uncertainty. These values are used in constructing the measurement uncertainty terms that weight the corresponding observations in the retrieval cost function.

Note that above table denotes the current instrument uncertainty model used in the retrievals. These uncertainty models may be updated in the future to reflect changes in instrument characterization or calibration provided by the instrument teams.

Gao et al. 2021. Efficient multi-angle polarimetric inversion of aerosols and ocean color powered by a deep neural network forward model. Atmospheric Measurement Techniques, 14, (6), 4083–4110. https://doi.org/10.5194/amt-14-4083-2021.

Gao et al. 2026. Where sky meets sea: Integrated aerosol and ocean color retrieval from PACE SPEXone multi-angle polarimetry. Remote Sensing of Environment, 345, 115603. https://doi.org/https://doi.org/10.1016/j.rse.2026.115603.