Decadal variability analysis of extreme precipitation in Turkey and its relationship with teleconnection patterns

• Published in: Hydrological Processes
• Main Authors: Duzenli, Eren, Tabari, Hossein, Willems, Patrick, Yilmaz, Mustafa Tugrul
• Format: Article in Journal/Newspaper
• Language: English
• Published: Wiley 2018
• Subjects: Arctic; Soi; North Atlantic; North Atlantic oscillation
• Online Access: http://dx.doi.org/10.1002/hyp.13275; https://api.wiley.com/onlinelibrary/tdm/v1/articles/10.1002%2Fhyp.13275; https://onlinelibrary.wiley.com/doi/pdf/10.1002/hyp.13275; https://onlinelibrary.wiley.com/doi/full-xml/10.1002/hyp.13275

Abstract Natural disasters such as droughts and floods originate as a consequence of excessive high or low precipitation amount and/or frequency. Due to the temporal persistence of the latter, the disasters tend to cluster in time. Because global ocean–atmosphere teleconnection patterns with (multi‐)decadal oscillations are tightly related with the precipitation variability, it is useful to analyse precipitation variability at the same timescale to understand any possible connection between them. In this study, decadal oscillations of daily extreme precipitation are investigated using quantile perturbation method for 65 stations in Turkey for the period 1955–2014. Arctic Oscillation (AO), North Atlantic Oscillation (NAO), Southern Oscillation Index (SOI), and Western Mediterranean Oscillation teleconnection patterns are examined for their relation with the extreme precipitation variability. The analyses are conducted for four climatic seasons and seven subregions. According to the analysis based on single drivers, NAO is identified as the most effective driver of Turkish extreme precipitation variability in winter, especially over regions influenced by the Mediterranean climate, whereas AO has a similar effect. When the teleconnection patterns are investigated in pairs, the combination of the NAO and SOI results in the strongest influence on the winter extremes. Though obvious differences are not recorded between the results of linear and nonlinear methods for single driver analysis, the nonlinear method is superior for the multiple driver analysis.