Advance in spring plant phenology over the last several decades has been found in all continents of the Northern Hemisphere. Compared to the studies detecting phenological trends, the studies investigating the geographical pattern of phenological variability (including mean date and magnitude of variability) are rather limited. In this study, we analyzed spatial pattern of mean date and standard deviation (SD) of first bloom date (FBD) time series (≥15 years) for black locust (Robinia pseudoacacia) at 22 stations in China, common lilac (Syringa vulgaris) at 79 stations in the Western US and Chinese lilac (Syringa chinensis) at 45 stations in the Eastern US. Subsequently, the impact of geographical factors (latitude, longitude, and altitude) on the mean date and SD was quantified by using the multiple regression analysis method. Meanwhile, the relationship between FBD variability and temperature sensitivity of FBD was examined. Results showed that the mean FBD highly depended on geographical factors for all the three species. Compared to the mean date, the dependence of SD of FBD time series on geographical factors was weaker. The geographical factors could only explain 13 to 31 % of spatial variance in SD of FBD. The negative regression coefficients of latitude (P < 0.05 except black locust) indicated that FBD is more variable at lower latitude. At most of stations, significant and negative correlations between FBD and preseason temperature on interannual scale were found, but the temperature sensitivity varied among different stations. The magnitude of temperature sensitivity decreased with increasing latitude. In general, the locations at lower latitude had earlier and more variable spring phenophase and showed stronger phenological response to climate change than the locations at higher latitude.