- Changes in glacier length reflect the integrated response to local fluctuations in temperature and precipitation, but when do such changes indicate forced climate change, and when do they indicate natural variability? In this study, we simulate the past ~1000 years of glacier length variability across the globe using the 3-stage linear glacier model of Roe and Baker (2014), forced with temperature and precipitation anomalies from the CESM Last Millennium Ensemble. Comparing the magnitude of the ensemble-mean length anomaly to the ensemble-spread allows us to quantify the relative importance of forced vs. natural variability. For most small glaciers, we find that length variations over the last millennium have mostly resulted from natural variability. However, at large spatial scales, and for large glaciers with long response timescales, radiative forcing has a greater influence than natural variability. Single-forcing simulations indicate that most of the forced response over the last millennium has been driven by global-scale temperature change associated with volcanic aerosols. So far, anthropogenic warming has primarily affected small-to-medium-sized glaciers, which are more responsive than large glaciers to decadal-scale climate change. In future decades, however, large glaciers will experience the greatest retreat, as they integrate the effects of both past and continued warming.