Abstract. Analyzing seasonally resolved δ18O ice core data can aid the interpretation of the climate information in ice cores, also providing insights into factors governing the δ18O signal that cannot be deciphered by investigating the annual δ18O data only. However, the seasonal isotope signal has not yet been investigated in northern Greenland, e.g., at the NEEM (North Greenland Eemian Ice Drilling) ice core drill site. Here, we analyze seasonally resolved δ18O data from four shallow NEEM ice cores covering the last 150 years. Based on correlation analysis with observed temperature, we attribute about 70 and 30 % of annual accumulation to summer and winter, respectively. The NEEM summer δ18O signal correlates strongly with summer western Greenland coastal temperature and with the first principal component (PC1) of summer δ18O from multiple seasonally resolved ice cores from central/southern Greenland. However, there are no significant correlations between NEEM winter δ18O data and western Greenland coastal winter temperature or southern/central Greenland winter δ18O PC1. The stronger correlation with temperature during summer and the dominance of summer precipitation skew the annual δ18O signal in NEEM. The strong footprint of temperature in NEEM summer δ18O record also suggests that the summer δ18O record rather than the winter δ18O record is a better temperature proxy at the NEEM site. Despite the dominant signal of the North Atlantic Oscillation (NAO) and the Atlantic Multidecadal Oscillation (AMO) in the central–southern ice core data, both NAO and AMO exert weak influences on NEEM seasonal δ18O variations. The NEEM seasonal δ18O is found to be highly correlated with Baffin Bay sea ice concentration (SIC) in the satellite observation period (1979–2004), suggesting a connection of the sea ice extent with δ18O at NEEM. NEEM winter δ18O significantly correlates with SIC even for the period prior to satellite observation (1901–1978). The NEEM winter δ18O may reflect sea ice variations of Baffin Bay rather than temperature itself. This study shows that seasonally resolved δ18O records, especially for sites with a seasonal precipitation bias such as NEEM, provide a better understanding of how changing air temperature and circulation patterns are associated with the variability in the δ18O records.