Mismatch between gravitysphericalharmonic and the acceleration derived from Orbit Propagator Numerical

Hi!

I have been trying to solve this issue for several days, but I am not able to get the error.

When I run the following code:

r_ECEF = [-1175208.0 -6573913.0 9786.0]; 
[gx gy gz] = gravitysphericalharmonic(r_ECEF,'EGM96',21)

I obtain: [1.5752 8.8112 -0.0131] m/s^2. According to the gravitysphericalharmonic documentation, the output is in Fixed Coordinates (ECEF), ITRF in this case. I am aware I am setting the order of the model to just 21, it is irrelevant for the purpose of this example.

To me, this is the gravity vector, in coordinates fixed to the Earth, so if I run the following code:

r_ECEF = [-1175208.0 -6573913.0 9786.0]; 
v_ECEF = [6205.323421, -1103.830260, 3686.076361];
[gx gy gz] = gravitysphericalharmonic(r_ECEF,'EGM96',21)
a_ECEF = [gx gy gz];
utc = [2015 1 1 0 0 0];
mjd = mjuliandate(utc);
pm = polarMotion(mjd, 'action', 'none')*180/pi;
[r_eci,v_eci,a_eci] = ecef2eci(utc, r_ECEF, v_ECEF, a_ECEF,'pm',pm,'dUT1',-0.45816)

I would be able to obtain the gravity vector (a_ECI) in an inertial frame (ICRF in this case). For the previous example, a_ECI = [-9.9056 -0.0007 0.0013] m/s^2. From this point, I would be able to integrate the acceleration to obtain the position and velocity. I would expect this inertial gravity vector to be equivalent to the one obtain from the Orbital Propagator Numerical block:

However, when setting the block accordingly (i.e., no Sun pressure, no drag, gravity set to EGM96 with up to order 21, and the same initial conditions), the acceleration for the initial moments is roughly [-8.951, 0, 0] m/s^2. There is almost an extra 1 m/s^2. I know this is the right value (I have compared it with other propagators), so I must have confused some concept along the way but I do not know which one.

Could you help me here?

All the best!!