In recent years, different vertical land motion (VLM) datasets based on the Global Navigation Satellite System (GNSS) have been published (e.g., Schumacher et al., 2018; Argus et al., 2021; Vardić et al., 2022). These datasets provide valuable constraints for testing and calibrating models of glacial isostatic adjustment (GIA). This presentation will compare recent GNSS datasets of VLM for Canadian stations for which contemporary elastic signals associated with recent ice mass variations and lake level changes have been removed. Each dataset will be used to infer optimal parameters in two classes of GIA models: one with spherically-symmetric (1D) Earth structure and one with a more realistic 3D Earth structure. A suite of ~30 3D Earth viscosity structures will be considered based on 4 different global seismic velocity models. Two different ice history models will be used in this comparison: ICE-6G (Peltier et al., 2015) and the model developed at the Australian National University (Lambeck et al., 2014). The results of this work will be used to address three questions: (1) does the choice of GNSS-VLM dataset impact inferences of optimal GIA model parameters? (2) Are the more complex/realistic 3D Earth models more accurate and able to produce improved data-model fits compared to the less complex 1D models? (3) If the 3D models are more accurate, what are the implications for interpreting other geodetic datasets for contemporary climate signals (e.g., hydrology signals from GRACE, ice melt or steric signals in tide gauge records)?