Assessing Multi-site Drought Connections in Iran Using Empirical Copula

Jenq-Tzong Shiau, Reza Modarres, Saralees Nadarajah

Research output: Contribution to journalArticlepeer-review

18 Citations (Scopus)

Abstract

Drought is a multi-dimensional natural hazard with stochastic characteristics usually related to each other. Separate univariate statistical models cannot capture the important relationships among drought characteristics, that is, severity and duration. In this study, an empirical copula is employed to construct a bivariate model of droughts, where droughts are defined as continuously negative standardized precipitation index (SPI) periods with one SPI value reaching -1 or less. Bivariate frequency analyses in terms of recurrence intervals are performed using the established empirical copula-based bivariate drought model. The inter-connection among different regions of droughts is explored by a lower tail dependence coefficient. A nonparametric estimation based on an empirical copula is employed pairwisely to calculate the lower tail dependence coefficient among stations. The proposed method is applied to six rainfall gauge stations in Iran to explore drought properties of single sites as well as the inter-connection among multi-sites. The results show that greater mean drought severity and duration are associated with the least arrival rate of drought events, which occurs at the Ahwaz station. The tail dependence analysis reveals that distance between stations is not a key parameter. Generally, the Ahwaz and Isfahan stations have the highest probability of simultaneous droughts among the six stations.

Original languageEnglish
Pages (from-to)469-482
Number of pages14
JournalEnvironmental Modeling and Assessment
Volume17
Issue number5
DOIs
Publication statusPublished - 2012 Sep 1

All Science Journal Classification (ASJC) codes

  • Environmental Science(all)

Fingerprint Dive into the research topics of 'Assessing Multi-site Drought Connections in Iran Using Empirical Copula'. Together they form a unique fingerprint.

Cite this