Copula-based non-stationarity detection of the precipitation-temperature dependency structure dynamics and possible driving mechanism
Haixia Dong,Shengzhi Huang,Wei Fang,Guoyong Leng,Hao Wang,Kang Ren,Jing Zhao,Chuanhui Ma
DOI: https://doi.org/10.1016/j.atmosres.2020.105280
IF: 5.965
2021-02-01
Atmospheric Research
Abstract:<p>In the context of global warming, precipitation (P) and temperature (T) are the most important climate indicators playing important roles in the hydrological cycle. Nevertheless, the response of their dependency structures to the changing environment is not clearly revealed on a regional or global scale. To this end, the non-stationarity of the precipitation-temperature (P-T) dependency structure was identified via the Copula-based Likelihood-ratio (CLR) method, which further verified through the frequently used double mass curve method. Furthermore, local meteorological factors (e.g. wind speed (WS), sunshine duration (SD), relative humidity (RH) and vapour pressure (VP)) and teleconnection factors (e.g. the Arctic Oscillation (AO), Pacific Decadal Oscillation (PDO), El Niño-Southern Oscillation (ENSO) and Sunspots) were selected to explore possible driving forces and mechanism of the P-T dependency structure dynamics. The Datong River Basin (DRB), located in the Qinghai-Tibet Plateau, one of the climate change-sensitive and eco-sensitive areas worldwide, was selected as a case study. Results showed that: (1) the CLR method simultaneously capturing bivariate linear and nonlinear information is more superior than the double mass curve in detecting the non-stationarity of bivariate dependency structure; (2) change points of P-T dependency structure were identified at Qilian and Minhe stations, indicating that its non-stationarity occurred in the DRB; (3) in terms of local meteorological factors, the P-T dependency structure dynamics were directly driven by the VP, which was closely associated with the Clausius-Clapeyron (CC) equation where P and T would be theoretically linked by the atmospheric moisture; (4) in terms of teleconnection factors, the impacts of AO and PDO on local meteorological (VP, WS, and RH) are dominant, which further leads to the change in the P-T dependency structure dynamics. Generally, this study provides important insights into the response of the P-T dependency structure dynamics to a changing environment, where the proposed research framework could be extended to any other watershed and any bivariate hydro-meteorological elements.</p>
meteorology & atmospheric sciences