This project holds model simulations and data analyses in support of a study developing and testing a new data assimilation technique to improve coastal water level model accuracy. The study used three different sources for the model's wind and pressure (meteorological) inputs to force the ADCIRC+SWAN coupled ocean circulation and wave model for simulations of Hurricane Matthew (2016). The study area covered the U.S. Atlantic coast from Florida through North Carolina, with some information provided throughout the Chesapeake Bay. Full details on this work are in the corresponding journal article (see Asher et al. in Related Work). Three sources of meteorological forcing were used to hindcast Hurricane Matthew. These correspond to lower- and intermediate-grade sources that are similar to what might be available during forecasting, and a third higher-grade source that is only available post-storm. Two closely related data assimilation methods were developed and applied to improve model accuracy. Results showed that the data assimilation method substantially reduced model error by correcting for gradual errors in water levels. These results also provide data on differences in performance between the three sources of meteorological inputs. These data can be used for study of model hindcast performance, work on data assimilation methodology, or model sensitivity to meteorological inputs.