The identification of historical events by geological and archaeological evidence is often ambiguous and conflicting, undermining the enormous potential for sub-annual precision in dating. The ruin of one of the largest pottery factories in the Middle East during Byzantine times, recently excavated in Yavneh (central Israel), exemplifies this: aligned fallen walls and columns and a kiln that collapsed while still in operation, with dozens of ceramic storage jars in articulation. Archaeological dating, which limits the time of the collapse to the seventh century CE, cannot distinguish between two large documented earthquakes that occurred during this century. By using pollen grains trapped by the collapse, we were able to distinguish, for the first time, between the two candidate earthquakes: September 634 CE and early June 659 CE. The pollen was extracted from the dust captured on the floor of the kiln during the cooling process of the vessels. The dust was collected only from below in situ whole vessels, and based on our reconstruction had been accumulated for about several days (after the heating process ended and before the collapse). Since the palynological assemblages included spring-blooming plants (such as Olea europaea and Sarcopoterium spinosum) and no common regional autumn bloomers (e.g. Artemisia), it is proposed that the kiln went out of use due to the early June 659 CE earthquake. We also propose that the recovery of the Yavneh workshops was no longer economically worthwhile, maybe in part due to changes in economic and political conditions in the region following the Muslim conquest.