Spatial and temporal patterns of variability in spring onset are identified across western North America using a spring index (SI) model based on weather station minimum and maximum temperatures (Tmin and Tmax, respectively). Principal component analysis shows that two significant and independent patterns explain roughly half of the total variance in the timing of spring onset from 1920 to 2005. However, these patterns of spring onset do not appear to be linear responses to the primary modes of variability in the Northern Hemisphere: the Pacific–North American pattern (PNA) and the northern annular mode (NAM). Instead, over the period when reanalysis data and the spring index model overlap (1950–2005), the patterns of spring onset are local responses to the state of both the PNA and NAM, which together modulate the onset date of spring by 10–20 days on interannual time scales. They do so by controlling the number and intensity of warm days. There is also a regionwide trend in spring advancement of about −1.5 days decade−1 from 1950 to 2005. Trends in the NAM and PNA can only explain about one-third (−0.5 day decade−1) of this trend.