Probabilistic innovative trend analysis

Abstract

This paper proposes new methodology for trend analysis, named as the probabilistic innovative trend analysis (PITA), whereby instead of statistical parameters probability distribution functions play the major role. The basis of the methodology is to compare the cumulative probability distribution functions (CDF) for each half of a given hydro-meteorological records. The methodology is without any restrictive assumption. The application of the PITA methodology is presented for Danube River discharge, New Jersey statewise precipitation and Istanbul minimum temperature records.