Reference evapotranspiration data from atmometers at three locations on the Delmarva Peninsula (USA) were compared to Penman-Monteith reference evapotranspiration (ETo) data across two growing seasons. Atmometer reference evapotranspiration (ETa) was found to underestimate ETo by 22.8% in 2016 and 30.4% in 2017. Stepwise linear regression was used to examine the relationship between both ET datasets and local meteorological conditions measured by Delaware Environmental Observing System (DEOS) mesonet stations that were co-located with the atmometers. Variability of ETa and ETo are well explained (R2 equal to 0.890 and 0.956, respectively) by a combination of meteorological variables, though the R2 in 2017 (R2 = 0.754) was notably lower than in 2016 (R2 = 0.890). The ET datasets were further examined by partitioning the data into days with similar synoptic conditions using a temporal synoptic index (TSI). Using the TSI results, three dominant synoptic categories were defined during the study period: High Pressure (HP), Southwest Flow (SW), and Cold Fronts (CF). Overall, the similarity between ETa and ETo was greatest on HP days, followed by CF and SW days. This relationship was primarily driven by wind speed, which had the greatest influence on ETa-ETo differences under all synoptic weather patterns. The 2016 growing season consisted of more days with synoptic conditions that are associated with smaller ETa -ETo differences than the 2017 growing season. Thus, changes in synoptic category frequency impact the nature of the ETa-ETo relationship from season to season. This study improves upon previous atmometer comparison studies by associating atmometer correction factors with synoptic weather patterns and descriptions in order to improve the utility of atmometers and remove the need for expensive meteorological equipment to correct atmometer data.