Application of Z-numbers to teleconnection modeling between monthly precipitation and large scale sea surface temperature

Abstract

The teleconnection modeling of hydro-climatic events is a complex problem with highly uncertain circumstances. In contrast to the classic fuzzy logic methods, by using the Z-number in addition to the constraint of information, and by evaluating the data reliability, it is possible to characterize the degree of ambiguity of data. In this regard, this study investigates the performance of the Z-number-based model (ZBM) in prediction of classified monthly precipitation (MP) events of two synoptic stations in Iran (up to five months in advance). To this end, the sea surface temperature (SST) of adjacent seas was used as a predictor. The suggested model, by using Z-number directly and applying fuzzy Hausdorff distance to determine weights of if-then rules, predicted MP events of both the stations with over 70% confidence. Analysis of the results in the test step showed that the ZBM compared to the traditional fuzzy approach improved the results by 69% for Kermanshah and 112% for Tabriz. Overall, the Z-number concept by assessing events reliability can be used in various sectors of water resources management such as decision-making and drought monitoring. HIGHLIGHTS In this study, the performance of the Z-number-based model in teleconnection modeling is investigated.; In contrast to classic fuzzy logic, Z-numbers consist of both restraint and reliability of data.; The performance of the Z-number-based model and the conventional fuzzy model are compared.; The Z-number-based model can predict classified monthly precipitation based on SST variations.;

Keywords

• data mining
• fuzzy modeling
• hydro-climatologic processes
• monthly precipitation
• sea surface temperature
• z-numbers