Investigating ecological turnover in the frequency domain offers a novel perspective on classic problems in ecology. We use this approach to investigate the relationship between climate and vegetation turnover in a global compilation of fossil pollen assemblages that span the last 1.1 million years. Vegetation turnover follows a power-law scaling relationship (S(f) ∝ f -𝛃) globally that closely tracks simulated temperature and precipitation from a climate model. Spectral power between temperature, precipitation, and vegetation turnover are similar across all frequencies and indicate that the magnitude of vegetation turnover is closely linked to total climate variability. Dynamic climate equilibrium holds at high and low latitudes despite less climate variability and evidence of greater biotic interactions in the low latitudes. At a local scale, vegetation turnover appears to have greater memory (larger 𝛃) than temperature or precipitation and may be a signature of abrupt changes that are common in fossil pollen assemblages. The projected spatial structure of changes to climate variability differs from mean climate change and presents a unique threat to biological systems.