The geometric similarity between two sets of coordinates should be assumed for frame alignment. When aligning Global Navigation Satellite System (GNSS) positions with seasonal variations to the linear Terrestrial Reference Frame (TRF), non-negligible aliasing errors of transformation parameters can appear and bias the station positions. The recent release of the seasonally augmented International Terrestrial Reference Frame 2020 (ITRF2020) necessitates an investigation into the impacts of seasonal terms on the alignment of regional networks. Using four-year of GPS data for Crustal Movement Observation Network of China (CMONOC) stations, we realize five sets of regional reference frames (RFs) aligned to the linear ITRF2020 (REF-LIN), the seasonal ITRF2020 (REF-FRQ), the linear ITRF2020 augmented either with seasonal signals or with full signals from loading models (REF-FIT and REF-NTL), and the combination of seasonal displacements from ITRF2020 and sub-seasonal signals from loading model (REF-CMB). Our analysis revealed that the augmented ITRF2020 had significant impact on the global stations used for frame alignment. Aligning daily solutions to the augmented ITRF2020 instead of the linear ITRF2020 resulted in a noticeable root mean square (RMS) decrease for over 70% of global stations. The REF-CMB method was the most preferred for frame alignment of global networks, with the largest decreased RMS of 0.3 and 1.0 mm in horizontal and vertical components. This indicates that most true signals are preserved in the aligned daily solutions. For regional network, we found only marginal impacts on regional stations in terms of linear velocity, residuals of time series stacking between the augmented solutions and REF-LIN. However, the larger discrepancies are noticed between annual estimates and loading model for the vertical component of REF-CMB and REF-FRQ, compared to RER-LIN solutions.