Alternative Trend Analysis: Half Time Series Methodology

Abstract

This paper attempts to determine and visualize the trends by considering the value via Sen’s innovative trend analysis methodology, which has been suggested instead of the well-known approaches called Mann-Kendall (MK), Spearman’s rho (SR), Sen’s slope, and linear regression methodologies. Another attempt is to show that an additional trend approach by using half time series methodology (HTSM) including MK’s Z value, and to compare it with the innovative method. Although, Sen’s method is practical, easy and clear representation of data variation, but some information is not evident on time series. This study comprises 66-year rain gage records in Adana, Izmir, and Rize stations located in different regions of Turkey. The comparison results indicated different trend conditions for these stations. Based on this information, representative figures are obtained according to Sen’s and HTSM for the precipitation records. According to Sen’s method, Adana and Rize stations’ measurements have no trend component but Izmir data have monotonically increasing trend. HTSM results reveal some additional information on Sen’s half time series approach. Adana station time series halves include opposite trends (increasing-decreasing) detected by HTSM in spite of no trend. Different extreme cases are encountered in Izmir and Rize stations through HTSM. As a result, if the proposed method is combined with Sen’s method, better trend analysis results can be obtained for detailed interpretations.