We present a field survey and a number of simulations of the local Persian Gulf tsunami of 19 March 2017 at Bandar Dayyer, Iran, which resulted in one death, five persons missing and significant damage to the port. The field survey defined the inundated area as extending ?40 km along the coast, with major effects concentrated on an ?8 km stretch immediately west of Dayyer, a maximum run-up of 3 m and maximum inundation reaching 600 m. In the absence of significant earthquakes on that day, we first test the possibility of generation of a landslide; however, our simulations for legitimate sources fail to reproduce the distribution of run-up along the coast. We prefer the model of a meteorological tsunami, triggered by Proudman resonance with a hypothetical weather front moving at 10 m/s in a NNW azimuth, which could be an ancillary phenomenon to a major shamal wind system present over the Persian Gulf on that day. More detailed simulations of the Dayyer tsunami would require an improved bathymetric grid in the vicinity of the relevant coastal segment.