Look into whether the experimental analysis quantities can be assimilated

[stvdwtt]

IR analysis needs an emissivity to convert the intensity to temperature. Can we learn the emissivity with data assimilation? Or does having a parameter both in the state and the observation matrix cause problems?