I have recently started working on calculating a Geopotential model based on GGM05C. In order to obtain a model as accurate as possible, some corrections must be applied to the normalized Stokes' coefficients provided in GGM05C (Cnm and Snm).
Chapter 6 of this IERS note describes in detail the process. From my understanding, it comprises 5 types of corrections: secular (long-term) corrections, solid Earth tides corrections, ocean tides corrections, solid Earth pole tides corrections and ocean pole tides corrections.
I have been able to grasp the idea behind each of these (or so I hope!), except for the case of ocean tides. The main difference I see with all the other corrections (including solid Earth tides) is that the corrections due to each frequency of ocean tides comprises a positive and negative component (both on the Cnm and the Snm components), which, if I understand correctly, must be added before applying the total, resulting correction to the corresponding Cnm/Snm. See for example equation 6.15 here. This will then lead to partial cancellation between the 2 (positive and negative) parts. But why is this the case? Both solid Earth tides and ocean tides seem to be an ensemble of periodic effects with different frequencies and amplitudes, but they seem to be treated differently. Why don't we have directly just a single term for each Cnm/Snm, as is the case for solid Earth tides?
I am also puzzled by the fact that ocean pole tides do not comprise the same combination of positive and negative terms, and are instead modeled as a single component as can be seen in equations 6.24 here, in a very similar way as done for the solid Earth pole tides.