The ensemble prediction methods used for seasonal limited area forecasts were examined by comparing methods for generating ensemble simulations of seasonal precipitation. The summer 1993 model over the north-central US was used as a test case. The four methods examined included the lagged-average ensemble, perturbed physics ensemble, mixed physics ensemble, and multi-model ensemble. The results show that the similarity in spread between the mixed-physics and multi-model ensembles raises the possibility that a mixed-physics ensemble approach could be considered as a proxy for a multi-model ensemble. The mixed-physics ensemble performed well in terms of equitable threat score, especially for higher precipitation amounts.