In this approach at night, it is absolutely essential to fly the aircraft vertical path manually using the PAPI for visual guidance for any descent below the MDA. Descent below the MDA using the automation on a procedure that was not suited for vertical guidance in the first place is different than flying a GS on an ILS. A 3 degree ILS has a clear visual segment (34 to 1) and pilots are advised to follow the GS to the runway in the visual segment. This visual segment was not clear on a 20 to 1 slope and is why the PAPI was required (3.2 degrees is 17.9 to 1 slope). My question is why was the autopilot still on below the MDA, even if it is treated as a DA by Opspec? Is using the autopilot below the MDA(DA) on such an approach normal procedure in airline operation?
As a side note, if the aircraft had advisory vertical guidance that follows the VDA on the chart and the autopilot was operating properly and programmed to follow the advisory vertical guidance, the guidance would not have resulted in a CFIT.