Secular trend of global drought since 1950

Abstract

Drought is a recurring extreme climate event over most parts of the world featured by long duration and low predictability. The secular trend of drought is of particular interest for investigators in agriculture, climate change and sustainability domains. In this study, we applied the ensemble empirical mode decomposition (EEMD) method and analyzed the spatio-temporal characteristics of the secular trends of meteorological drought over global land surface during the period 1950–2015 using a self-calibrating Palmer Drought Severity Index (PDSI) product. We found that there were 25.98% PDSI samples had turning point, namely the shift of trend, in the corresponding secular trend series; the probability distribution of the turning points position (period) extracted by EEMD closely follows a normal distribution with mean value at Nov. 1981. We showed that there is large discrepancy in the secular trend types extracted by EEMD and Mann–Kendall test, and exemplified the risk of using a monotonic trend to capture the changes of the intrinsic secular trend of PDSI series. We suggested that there was an accelerated drying trend over global land surface as a whole, but large areas with wetting trend existed in the meantime, especially at the high latitudes in the northern hemisphere. Additionally, we found that the PDSI secular trend change rate exhibits a multidecadal variability of about 50 years or so and it implies a potential relationship with periodic variations of the oceanic and atmospheric current. We showed that the secular trend of PDSI series extracted by EEMD could provide more detailed spatio-temporal characteristics, featured by the shifts of trend and nonlinear property of the secular trend, of global drought than that of the non-parametric or linear regression methods. The secular trend of PDSI could present more insights about the transition and progress of wetting/drying trend over global land surface at multidecadal scale.

The strengthened impact of water availability at interannual and decadal time scales on vegetation GPP

“…Our findings revealed that the dynamics of global WA are predominantly influenced by interannual variations, particularly those characterized by 3‐ and 7‐year time scales with large variance contribution rates. This is similar to the previous regional studies emphasizing the dominant role of interannual oscillations in WA (Jin et al., 2016; Li et al., 2019; Song et al., 2020). This is possible from that interannual oscillations in the global climate system, such as ENSO, influenced the distribution and intensity of global precipitation and evaporation (Chen, Zhu, et al., 2019; Feliks et al., 2021).…”

“…Meanwhile, numerous studies employing ensemble empirical mode decomposition (EEMD; Wu & Huang, 2009) elucidated the presence of multiple time scales in WA (Li et al, 2019;Liu et al, 2018;Song et al, 2020). However, these studies mainly focus on smaller regions, lacking an understanding of the global multi-time scale characteristics of WA.…”

“…The dynamic nature of climate change has led to fluctuations in WA across various time scales, ranging from daily to interannual, interdecadal, and longer‐term periods (du Plessis, 2017; Lobos‐Roco et al., 2022; Wang & Qin, 2017; Woolway et al., 2020). Meanwhile, numerous studies employing ensemble empirical mode decomposition (EEMD; Wu & Huang, 2009) elucidated the presence of multiple time scales in WA (Li et al., 2019; Liu et al., 2018; Song et al., 2020). However, these studies mainly focus on smaller regions, lacking an understanding of the global multi‐time scale characteristics of WA.…”

Phase variations of the summer and winter seasons in the Bohai Sea during the last four decades

“…Many recent studies have revealed that the SST time series and the associated length time series of the seasons are non-linear and non-stationary under dramatic climate change (Park et al, 2018;Xu et al, 2021;Wang et al, 2021a). In various climate change applications, the ensemble empirical mode decomposition (EEMD) method has been an objective and insightful means in revealing the secular trends from the non-stationary time series, including the timing and length time series of each season over drylands of the Northern Hemisphere midlatitudes (Wu et al, 2011;Franzke, 2012;Song et al, 2020;Wang et al, 2021a). Thus, we conducted the EEMD method to extract the secular trends of the lengths and timing time series of the summer and winter seasons in the Bohai Sea for further investigation on the phase variations of the seasonal cycle.…”