I also do not have any insights into whether there are any development efforts to incorporate NOAH_MP into the offline or coupled CMAQ model and what (if any) obstacles may exist for doing so.
Regarding the Campbell et al (2018) study, please note that most but not all of the developments described in that study are part of the WRF and CMAQ release codes. Notably, the windblown dust module in the released version of CMAQ continues to expect the first soil layer to have a depth of 1 cm which is not satisfied when using the NOAH LSM in WRF. For additional context, see this earlier discussion.