This data set presents the reconstructed vegetation cover for 1115 North American sites based on harmonized pollen data from the data set LegacyPollen 2.0 (https://doi.pangaea.de/10.1594/PANGAEA.929773) and optimized RPP values. Sugita's REVEALS model (2000) was applied to all pollen records using REVEALSinR from the DISQOVER package (Theuerkauf et al. 2016). Pollen counts were translated into vegetation cover by taking into account taxon-specific pollen productivity and fall speed. Additionally, relevant source areas of pollen were also calculated using the aforementioned taxon-specific parameters and a gaussian plume model for deposition and dispersal and forest cover was reconstructed.In this optimized reconstruction, relative pollen productivity estimates for the ten most common taxa were first optimized by using reconstructed tree cover from modern pollen samples and LANDSAT remotely sensed tree cover (Townshend 2016) for North America. Values for non-optimized taxa for relative pollen productivity and fall speed were taken from the synthesis from Wiezcorek and Herzschuh (2010). The average values from all Northern Hemisphere values were used where taxon-specific continental values were not available. We present tables with optimized reconstructed vegetation cover for all records in North America. As further details we list a table with the taxon-specific parameters used and a list of parameters adjusted in the default version of REVEALSinR.