A series of linear and nonlinear barotropic models are used to interpret the extratropical response to El Niflo equatorial surface temperatures as simulated by an atmospheric general circulation model (GCM). The divergence, time-mean vorticity tendency due to transients, and the zonal mean tlow are specified from the GCM, and the deviation of the streamfunction from its zonal mean at an upper-tropospheric level is predicted. Nonlinearsteady-state models suggest that the extratropical wave train is primarily forced from the central rather than the western Pacific and that subtropical divergence anomalies are of more importance than tropical anomalies. These nonlinear solutions can be reproduced with little loss in accuracy by linearizing about the zonally asymmetric climatological flow. If one linearizes about the zonally symmetric flow, the part of the solution forced from thewestern Pacific deteriorates significantly. The solution in the tropics and subtropics also deteriorates if advection of v... Abstract A series of linear and nonlinear barotropic models are used to interpret the extratropical response to El Niflo equatorial surface temperatures as simulated by an atmospheric general circulation model (GCM). The divergence, time-mean vorticity tendency due to transients, and the zonal mean tlow are specified from the GCM, and the deviation of the streamfunction from its zonal mean at an upper-tropospheric level is predicted. Nonlinearsteady-state models suggest that the extratropical wave train is primarily forced from the central rather than the western Pacific and that subtropical divergence anomalies are of more importance than tropical anomalies. These nonlinear solutions can be reproduced with little loss in accuracy by linearizing about the zonally asymmetric climatological flow. If one linearizes about the zonally symmetric flow, the part of the solution forced from thewestern Pacific deteriorates significantly. The solution in the tropics and subtropics also deteriorates if advection of v...