This is an excellent question we get quite frequently regarding the different performance numbers generated by the FMS, Runway Analysis and Climb Gradient Analysis.

FMS

The FMS is not calculating a climb analysis, nor a runway analysis. It "knows" nothing of obstacles, close-in or otherwise. Its just raw data from the AFM or tab data based on OAT, Elevation, etc.

Runway analysis (Foreflight, APG, et al)

Runway analysis is not adhering strictly to the AFM. It will often instruct the user to not level-off at acceleration heights, exceed engine time limits for max power or use bank angles beyond that prescribed in the AFM. There is a video demonstrating this on our youtube channel here.

There is a second, even more prominent, issue. If an operator manually enters the fixes of a runway analysis derived departure (ie escape) procedure into the FMS (all off-the-shelf Runway Analysis apps require this); the navigation sensitivity will default to RNP 1.0, (TERM or "terminal" on some PFDs).

Due to the narrower departure paths created by all AC120-91 procedures (a width one third of TERPS), the RNP must be RNP 0.3 or less.

This can NOT be overwritten in the vast majority of FMSs and even on those that allow it, the process is cumbersome and may cause navigational issues downrange [sidebar: we had one user who reported they were able to change the RNP internally, with some difficulties navigating multitude menus; but noticed that subsequent phases of flight generated warnings and error messages regarding improper RNP] . Only FMS databases that contain the escape procedure by name, and through an LOA, does the appropriate RNP occur; the RNP of 1.0 correlates to TERPS (and PanOps) departure procedures. Manually overriding the RNP (if allowed at all) does not update the sensitivity of the GPS or autopilot, it just resets the alert level.

To put it bluntly, it is unlikely that the aircraft can maintain the centerline within the narrower obstacle clearance area of a runway analysis path, even if on the magenta line. It would be like flying an LPV approach using a VOR. This was the experience of Gulfstream who did a study in their simulators using Gulfstream demo and test pilots, under the direction of Randy Gaston, Director of Aviation (Ret).

This issue has largely been ignored by the FAA as these are "third party" procedures allowed by operators. FAA POIs are virtually unaware of this situation even though they are mandated to provide approvals to 135 operators to use these procedures.

A brief survey of sim instructors reveals that when clients request to fly a particular "escape procedure", the simulator is set to RNP deviation of zero, thus the simulated aircraft is always physically and precisely at the location represented by the CDI. This is not real world. When Gulfstream ran tests, they incorporated "real world" RNP deviations for aircraft low to terrain, at speed and GPSS availability.

This kind of "training" provides nothing but a false sense of confidence.

Climb Gradient Analysis (EFB-Pro)

Climb Gradient Analysis duplicates the calculation within the AFM; adhering to acceleration heights, max power limits and bank angles, keeping the aircraft's Net Takeoff Flight path above the obstacle plane by 35 ft.

The climb analysis differs from runway analysis in that we use the existing departure path, which is flight-tested, using TERPS/PanOps lateral margins of safety from obstacles and probably is pre-loaded by name within your Nav DB. CGA is not trying to clear AN obstacle, it is staying above a climb plane that contains ALL obstacles below, thus the name Climb Analysis. In other words, the solution is the same as the procedure found in the AFM.