Paleoelevation reconstructions of the North American Cordillera inferred from the oxygen ( δ18O) and hydrogen ( δD) isotope ratios of terrestrial paleoclimate proxy materials (soils, ashes, lake sediments) suggest rapid north-tosouth migration of topography in the early Cenozoic (pre-49 Ma to 28 Ma). The validation of this reconstruction relies on an accurate understanding of the 18Op and the associated regional climate change in response to the uplift of the western North America. Here we study this response using a global climate model (GCM) with explicit 18Op diagnostics (ECHAM5-wiso) focusing on the isotopic effects of different types of precipitation, vapor mixing, recycling and moisture source and compare the response against estimates made using a Rayleigh distillation models of moist adiabatic condensation (RDM). Four experiments are performed with Eocene topography inferred from terrestrial stable isotope paleoaltimetry records to investigate how southward propagation of topography affects regional climate (temperature, precipitation and circulation pattern) and δ18Op over North America. Our experiments predict δ18Op patterns that are broadly consistent with maps of temporally binned proxy δ18O and generally support an early Cenozoic north-to-south propagation of high topography in the North American Cordillera. They do not support the commonly made assumption that isotopic fractionation occurs primarily through rainout following Rayleigh distillation nor the application of modern empirical δ18Op lapse rates to past environments. In our GCM simulations, precipitation processes and climate changes that are not captured by RDMs substantially affect δ18Op. These processes include shifts in local precipitation type between convective and large-scale rain and between rain and snow; intensification of low-level vapor recycling particularly on leeward slopes; development of air mass mixing and changes in wind direction and moisture source. Each of these processes can have significant (>2%) influences on δ18Op that are comparable in magnitude to surface uplift of hundreds or even thousands of meters. In many regions, these processes fortuitously compensate each other, explaining the apparent agreement between ECHAM5-wiso and proxy δ18O and, more broadly, between RDM estimates and observed δ18O-elevation relationships. In some regions, compensation is incomplete, and as a result, ECHAM5-wiso δ18Op does not agree with estimates from the RDM. In these regions, including the interior of the northern cordillera and the eastern flank of the southern Cordillera, moderate adjustments of paleoelevations may be in order.