The Madden Julian Oscillation (MJO) is the primary mode of intraseasonal, convectively coupled wave variability in the tropics, and has teleconnection impacts around the globe. In the current Community Earth System Model (CESM2) it is a well simulated feature, in terms of propagation speed, strength and regional coherence, but only in the fully coupled model version with an interactive ocean. Much research exists that ascribes successful MJO simulation to frequent interactive flux coupling between the atmosphere and ocean enabling intraseasonal SST anomalies to propagate into the West Pacific. When observed SSTs are prescribed in the CESM2 atmosphere model (CAM6) the MJO shows no improvement compared to the previous CESM atmosphere model (CAM5). However, the imposition of the SSTs from the fully coupled model in an AMIP configuration (i.e., using monthly averaged SSTs) restores much of the MJO signal, implying the requirement for surface coupling may be secondary. Surprisingly, this also leads to similar improvements if CAM5 is used with the same SST anomalies. Similar sensitivities are further demonstrated in response to modifying characteristics of the Maritime Continent surface land units, including modification to orographic height and replacement of the land surface. Although there is a significant mean climate sensitivity to modification of the island heights, in terms of increased precipitation and vertical ascent for a doubling of orographic height, for example, there is limited response in restricting or allowing MJO propagation; the so called barrier effect. However, the replacement of land units by a prescribed SST increases surface latent heat fluxes and enables MJO propagation over the region and well into the Pacific. The results presented here demonstrate that, although we strive to advance atmospheric processes in models in order to improve climate, including the MJO, there are sensitivities to surface characteristics that have to be considered on an equal footing. In this vein, future work will focus on examining the response of the MJO to prescribing SSTs from other CMIP6 coupled models. Finally, this analysis uses, and aims to contribute to, the NOAA Model Diagnostics Task Force process-oriented diagnostics package.