Conceptualization of the indirect link between climate variability and lake water level using conditional heteroscedasticity

Abstract

This study investigates the indirect effect of large-scale climate oscillations and the corresponding teleconnection with lake water level (WL) oscillations. For this, the effect of the Southern Oscillation Index (SOI), and North Atlantic Oscillation (NAO) on the Lake Urmia WL during 1966–2016 is investigated using cross-correlation, cross-wavelet (XWT), wavelet-coherence (CWT), and nonlinear multivariate generalized autoregressive conditional heteroscedasticity (GARCH) models. Based on the XWT and CWT analyses, a temporal phase of WL leading by the SOI and an anti-phase of WL-NAO linkage within 11 years is evident. The models also depict a long-term persistence and effectiveness of the conditional covariance on both SOI–WL and NAO–WL links. It is concluded that the SOI and NAO have no immediate impact on the WL while the magnitude of the impact is exacerbated after the year 2000; however, the establishment of the SOI–WL link is found to be more relevant than that of the NAO–WL link.