Complexity and Predictability of Daily Actual Evapotranspiration Across Climate Regimes

Abstract

Actual evapotranspiration (ET a ) is a key terrestrial ecosystem process that links hydrosphere with atmosphere and biosphere through modulating a range of land surface and ecohydrological processes (Fisher et al., 2017;Wang & Dickinson, 2012). Despite the rapid developments of sensing and modeling technologies in recent decades (Hansen & Wei, 2021;Zhao et al., 2019), it still remains challenging to accurately estimate ET a due to its complex spatiotemporal ET a patterns across scales (Han et al., 2021;Hu et al., 2018). To improve the predictability of ET a , it requires a thorough evaluation of the complexity of ET a dynamics along with its influencing factors across diverse climate regimes and ecosystems (Sang et al., 2015).In general, the predictability of hydrometeorological variables decreases with increasing system complexity, which in turn is tightly related to the degree of system stochasticity (Koutsoyiannis, 2010;Rajagopalan et al., 2019). It is worth noting that the predictability investigated here is different from the one frequently used for hydrological models (Haughton et al., 2018), the latter of which is commonly defined as the prediction skills that can be compromised by uncertainties in initial value, boundary value, and parameter estimation problems (Kumar, 2011). Instead, the predictability is specifically defined here as the estimation of the deterministic (predictable) versus the stochastic (unpredictable) nature of ET a dynamics (Westerhold et al., 2020).The complexity and predictability of system variables can be investigated with the aid of the chaos theory. Previous studies have revealed that hydrological systems exhibit chaotic features, which display both deterministic patterns that are dictated by physical laws and stochastic behaviors that determine the complexity of hydrological systems and subsequently the predictability of hydrological variables (Sivakumar & Singh, 2012). Conceptually, a chaotic system is controlled by few independent variables that are sensitive to initial conditions, which results in random-looking processes (

Detecting the Non‐Separable Causality in Soil Moisture‐Precipitation Coupling With Convergent Cross‐Mapping

“…This is also proven by P in PRB, where P has an integer CD (2.00) but with λ > 0 (0.68). Also, different variables have different natures, which is another the reason why various methods should be applied in combination instead of only one (Di et al., 2023). But limited by the length and scope of this study, only these two methods are employed to investigate chaos in this research.…”