Modified Crossing Empirical Trend Analysis (MCETA) for meteorological time series

Abstract

Trend analysis of extreme events is an efficient method for assessment of climate change effects on hydro-meteorological variables. However, famous methods like Mann-Kendall test are able to detect only one trend slope for all of data series. The crossing empirical trend analysis (CETA) method, recently presented, can detect trend slope for every risk level of data. The aim of this study is to present a modified methodology for CETA method (MCETA) to strengthen it in the cases with non-monotonic peaks in high and low values of data, or with different trend directions in high and low risk level of data. The MCETA modifies the search ranges of slopes and the location of pivot points of the CETA test. The ability of MCETA compared to CETA and Mann-Kendall tests was assessed for trend analysis of spring rainfall for the period 1980–2010 in eastern Iran for 5%, 50%, and 95% risk levels of data. The results showed that the MCETA changed considerably the slope values of 5% and 95% risk levels of data compared to CETA specially for Semnan from 3.65 to −0.55 (for 5%), and from 5.80 to 7.06 (for 95%), as well as Zahedan from −1.80 to −0.59 (for 5%) and from −1.87 to −0.95 (for 95%). Moreover, the slope of 50% in MCETA for almost all stations was lower than that of the M−K test. The MCETA also provides a slope range for each risk level of data that implies the reliability of slopes and improves the test flexibility.