We used two types of numerical model to validate the diagnosing results and estimate the relative contributions of CP El Niño and pIOD. One was an ideal linear baroclinic model (LBM) with a T42 resolution in the horizontal direction and 20 sigma levels in the vertical direction (Watanabe & Kimoto, 2000). The other was an atmospheric general circulation model (AGCM) of the NCAR Community Atmosphere Model version 5 (CAM5), which had a 1.9° × 2.5° finite volume grid on the 26 hybrid sigma levels (Neale et al., 2012). The three CAM5 numerical experiments are 1. the TSEIO run, 2. the CP run, 3. the TSEIO+CP run. The climatology variables in the control run (abbreviated as "CTL") are also provided. All the model outputs were first interpolated onto the same spatial grid as NCEP‐DOE R2. Here are the outputs of the LBM and CAM5 experiments. Each output data is written in binary format (.bin/.dat) with a data descriptor file (.ctl). The file of LBM is named as "Model name"_"output"_"variable"."format". The file of CAM5 is named as "Model name"_"Sensitivity experiment name"_"variable"_"Season (e.g., ASO)"_"Selected experiment times (e.g., 0006-0018 or Clim.)"."format".