Application of Z-numbers to monitor drought using large-scale oceanic-atmospheric parameters

Abstract

The sea surface temperature (SST), Southern Oscillation Index (SOI) and North Atlantic Oscillation (NAO) as the oceanic-atmospheric parameters could be used as input parameters for long-term drought modeling. This study, for the first time in hydro-climatic studies, applies the concept of the Z-number valued if-then rules to predict the monthly dry, wet, and normal periods. In contrast to the classic fuzzy logic, which does not talk about the confidence and reliability of data and information, Z-numbers consist of restraint and reliability and have significant potential to describe the uncertainty of human knowledge. To highlight this approach, classified monthly standardized precipitation index (SPI) of two synoptic stations of Tabriz and Kermanshah in Iran as predictands and different lags of SST time series of surrounding seas (Black, Mediterranean and Red Seas), NAO and SOI indices as predictors (for 1955–2019) were used to construct teleconnection between inputs and outputs and the obtained results were compared with the results of the conventional fuzzy model. The results indicated that the approach using Z-numbers could lead to more comprehensive and accurate results averagely up to 146% and 590%, respectively, for the Kermanshah and Tabriz owing to its ability to consider uncertainty and reliability of information and allocating appropriate weights to the rules.