Exploring the Links in Monthly to Decadal Variability of the Atmospheric Water Balance Over the Wettest Regions in ERA-20C
Journal Geophysical Research: Atmospheres, 122,10,560–10,577. https://doi.org/10.1002/2017JD027012
Monthly-to-decadal variability of the regional precipitation over Intertropical Convergence Zone and north-Atlantic and north-Pacific storm tracks was investigated using ERA-20C reanalysis. Satellite-based precipitation (P) and evaporation (E) climatological patterns were well reproduced by ERA-20C. Regional P and E monthly time series displayed ~20% differences, but these decreased rapidly with time scale (~10% at yearly time scales). Spectral analysis showed good scale-by-scale statistical agreement between ERA-20C and observations. Using ERA-Interim showed no improvement despite the much wider range of information assimilated (including satellites). Remarkably high Detrended Cross-Correlation Analysis coefficients (ρDCCA > 0.7 and often ρDCCA > 0.9) revealed tight links between the nonperiodic variability of P, moisture divergence (DIV), and pressure velocity (ω) at monthly-to-decadal time scales over all the wet regions. In contrast, ρDCCA was essentially nonsignificant between nonperiodic P and E or sea surface temperature (SST). Thus, the nonperiodic monthly-to-decadal variability of precipitation in these regions is almost fully controlled by dynamics and not by local E or SST (suggested by Clausius-Clapeyron relation). Analysis of regional nonperiodic standard deviations and power spectra (and respective spectral exponents, β) provided further robustness to this conclusion. Finally, clear transitions in &beta: for P, DIV, and ω between tropical and storm track regions were found. The latter is dominated by transient storms, with energy accumulation at synoptic scales and β < 0.1 at monthly-to-decadal time scales, implying that variability and information creation decrease rapidly with time scale. Larger β values (0.2 to 0.4) were found in the tropics, implying longer-range autocorrelations and slower decreasing variability and information creation with time scale, consistent with the important forcing from internal modes of variability (e.g., El Niño-Southern Oscillation).