资源环境科技发展态势分析平台

Spectral analysis of global-mean precipitation, P, evaporation, E, precipitable water, W, and surface temperature, T-s, revealed significant variability from sub-daily to multi-decadal time-scales, superposed on high-amplitude diurnal and yearly peaks. Two distinct regimes emerged from a transition in the spectral exponents, . The weather regime covering time-scales<10days with 1; and the macroweather regime extending from a few months to a few decades with 0<<1. Additionally, the spectra showed a generally good statistical agreement amongst several different model- and satellite-based datasets. Detrended cross-correlation analysis (DCCA) revealed three important results which are robust across all datasets: (1) Clausius-Clapeyron (C-C) relationship is the dominant mechanism of W non-periodic variability at multi-year time-scales; (2) C-C is not the dominant control of W, P or E non-periodic variability at time-scales below about 6months, where the weather regime is approached and other mechanisms become important; (3) C-C is not a dominant control for P or E over land throughout the entire time-scale range considered. Furthermore, it is suggested that the atmosphere and oceans start to act as a single coupled system at time-scales>1-2 years, while at time-scales<6months they are not the dominant drivers of each other. For global-ocean and full-globe averages, (DCCA) showed large spread of the C-C importance for P and E variability amongst different datasets at multi-year time-scales, ranging from negligible (<0.3) to high (0.6-0.8) values. Hence, state-of-the-art climate datasets have significant uncertainties in the representation of macroweather precipitation and evaporation variability and its governing mechanisms.