Solomon and Newman (GRL, 2011) quantified the impact of ENSO on the decadal predictability of tropical Indo-Pacific Ocean trends in a very large ensemble of NCAR CCSM3 anthropogenically-forced (A1B scenario) simulations, by decomposing upper ocean temperatures into “ENSO” and “non-ENSO” variability. They found the non-ENSO component to be much more robust across a large model ensemble than both the ENSO component and the total trend pattern, suggesting that ENSO variability may complicate the determination of tropical Pacific ocean trends even over many decades. Their technique is here applied to the 20th century AR4 models to similarly show that differences amongst the models' simulated 20th century tropical Pacific trends are also largely the result of differences in ENSO simulation. In this case, we use the corresponding pre-industrial control runs to define each model's “natural” variability, defined using the leading 4 EOFs of annual mean (July-June) ocean (surface to 300m depth) temperature variability within each control run. The output of each ensemble member of each model for the period 1900-1999 is then projected on the corresponding control EOFs of that model, and this “natural” variability is removed from the 20th century simulations. EOFs of the resulting “residual” dataset are then calculated. Comparing the leading PC for each 20th century model ensemble member to the leading PC for each residual dataset shows that not only is the externally forced residual trend in the PC time series much more distinct (including variations corresponding to volcanic eruptions in those models with volcanic forcing) but the spread across both models and ensemble members is greatly reduced. The similarity of the residual PC time series extends to the corresponding SST patterns as well; for all models, both the ensemble mean of the residual trend and the leading residual EOF has a similar pattern to the “non-ENSO” pattern earlier found by Solomon and Newman. In particular, while the total trend in some of the models has “El Niño-like” characteristics, in that the warming is greatest in the central equatorial Pacific and is similar to three-dimensional structures associated with ENSO, the non-ENSO component of the trend has a structure distinct from ENSO with cooling in the South Pacific due to increased southeast trades, warming of the warm pool, and strengthening of the equatorial Pacific near-surface temperature gradient superimposed upon uniform warming. The robustness of this structure across models suggests a physically meaningful response to external forcing. Still, it is important to bear in mind that while this pattern represents perhaps the most robust portion of the trend, or at least the most consistent portion from model to model, it is not the total trend since an ENSO-like trend component -- whose amplitude cannot be quantified in the limited ensembles available -- is removed along with the ENSO variability. A similar analysis is also performed for corresponding simulations made by the CMIP5 models.