We apply a recently developed and validated numerical model of tsunami propagation and runup to study the inundation of Resurrection Bay and the town of Seward by the 1964 Alaska tsunami. Seward was hit by both tectonic and landslide-generated tsunami waves during the Mw 9.2 1964 mega thrust earthquake. The earthquake triggered a series of submarine mass failures around the fjord, which resulted in land sliding of part of the coastline into the water, along with the loss of the port facilities. These submarine mass failures generated local waves in the bay within 5 min of the beginning of strong ground motion. Recent studies estimate the total volume of underwater slide material that moved in Resurrection Bay to be about 211 million m3 (Haeussler et al. in Submarine mass movements and their consequences, pp 269-278, 2007). The first tectonic tsunami wave arrived in Resurrection Bay about 30 min after the main shock and was about the same height as the local landslide-generated waves. Our previous numerical study, which focused only on the local land slide generated waves in Resurrection Bay, demonstrated that they were produced by a number of different slope failures, and estimated relative contributions of different submarine slide complexes into tsunami amplitudes (Suleimani et al. in Pure Appl Geophys 166:131-152, 2009). This work extends the previous study by calculating tsunami inundation in Resurrection Bay caused by the combined impact of landslide-generated waves and the tectonic tsunami, and comparing the composite inundation area with observations. To simulate landslide tsunami runup in Seward, we use a viscous slide model of Jiang and LeBlond (J Phys Oceanogr 24(3):559-572, 1994) coupled with nonlinear shallow water equations. The input data set includes a high resolution multibeam bathymetry and LIDAR topography grid of Resurrection Bay, and an initial thickness of slide material based on pre- and post-earthquake bathymetry difference maps. For simulation of tectonic tsunami runup, we derive the 1964 coseismic deformations from detailed slip distribution in the rupture area, and use them as an initial condition for propagation of the tectonic tsunami. The numerical model employs nonlinear shallow water equations formulated for depth-averaged water fluxes, and calculates a temporal position of the shoreline using a free-surface moving boundary algorithm. We find that the calculated tsunami runup in Seward caused first by local submarine landslide-generated waves, and later by a tectonic tsunami, is in good agreement with observations of the inundation zone. The analysis of inundation caused by two different tsunami sources improves our understanding of their relative contributions, and supports tsunami risk mitigation in south-central Alaska. The record of the 1964 earthquake, tsunami, and submarine landslides, combined with the high-resolution topography and bathymetry of Resurrection Bay make it an ideal location for studying tectonic tsunamis in coastal regions susceptible to underwater landslides. ?? 2010 Springer Basel AG.