Copula-based non-stationarity identification of watershed water and energy dependency structure and possible driving forces

Abstract

Watershed water-energy balance plays an important role in affecting ecological environment dynamics. The stationarity of watershed water and energy dependency structure is likely to be invalid in a changing environment, which may influence the feasibility of the Budyko hypothesis extensively applied in the hydrological domain. However, little attention has been paid to the identification of their non-stationarity and possible driving forces. To this end, the Copula-based likelihood ratio (CLR) method was employed to identify the non-stationarity of precipitation (P) and potential evaporation (PET) dependency structure. Additionally, meteorological and teleconnection factors were used to explore the dominant driving forces of the dependency structure dynamics. The Wei River Basin belonging to the Yellow River Basin was selected as a case study. Results showed that: (1) change points are observed in the P-PET dependency structure in the southeast region of the Wei River Basin, indicating that the stationarity of their dependency structure is invalid in a changing environment; (2) in terms of meteorological factors, the substantially decreasing wind speed and increasing temperature are the dominant driving forces of the P-PET dependency structure dynamics; (3) in terms of teleconnection factors, Pacific Decadal Oscillation strongly affects the P-PET dependency structure dynamics generally through exerting effects on temperature and relative humidity, whereas other teleconnection factors such as Arctic Oscillation, El Nino-Southern Oscillation, and sunspot exhibit weak impacts. These findings could shed important insights into water-energy balance dynamics in a changing environment