Quantifying the response of vegetation to climate change across timescales

Abstract

Climate and ecosystems exhibit dynamic behavior across various timescales. Existing studies often focus on individual timescales when examining ecosystem responses to climate. To address this, we develop a conceptual and analytical framework using spectral analysis that examines a continuum of timescales, from hundreds to hundreds of thousands of years. By comparing power spectra of vegetation turnover of the last 1.1 million years with climate, we find that vegetation closely tracks climate at intermediate frequencies (872$^{-1}$ to 4,789$^{-1}$ years$^{-1}$) but is decoupled from climate at high frequencies (> 152$^{-1}$ years$^{-1}$) and low frequencies (<4,789$^{-1}$ years$^{-1}$). We provide striking evidence that the relationship between plant communities and climate is timescale dependent; they may co-vary on one timescale but differ in their dynamics and drivers on other timescales.