Abstract In an effort to explain observed blocking phenomena, the work of Charney and DeVore (1979) and Hart (1979) has been extended to incorporate observed zonal topography in a barotropic nonlinear channel model. Multiple stationary equilibria are obtained, one of which, for an appropriate forcing, corresponds exactly to the “normal” winter flow predicted by Charney and Eliassen (1949) from the linearized version of the model. When this forcing is applied in the nonlinear model, other equilibria, related to resonances with the wevenumber 2 and 3 Fourier components of the zonal topography, occur. Wavenumber 1 and 4 resonances could also have occurred with slight modifications of the model. For comparison with observation, semi-objective criteria are adopted for identifying blocking events from daily 500 mb observations of 15 consecutive winter seasons. Following Dole (1979), we demand that there exist sufficiently large geopotential height anomalies for a sufficient length of time. Numerical values of t...