Differences between true mean temperatures and means calculated with four different approaches: a case study from three Croatian stations

Different countries use varied methods for daily mean temperature calculation. None of them assesses precisely the true daily mean temperature, which is defined as the integral of continuous temperature measurements in a day. Of special scientific as well as practical importance is to find out how temperatures calculated by different methods and approaches deviate from the true daily mean temperature. Five mean daily temperatures were calculated (T0, T1, T2, T3, T4) using five different equations. The mean of 24-h temperature observations during the calendar day is accepted to represent the true, daily mean T0. The differences Δi between T0 and four other mean daily temperatures T1, T2, T3, and T4 were calculated and analysed. In the paper, analyses were done with hourly data measured in a period from 1 January 1999 to 31 December 2014 (149,016 h, 192 months and 16 years) at three Croatian meteorological stations. The stations are situated in distinct climatological areas: Zagreb Grič in a mild climate, Zavižan in the cold mountain region and Dubrovnik in the hot Mediterranean. Influence of fog on the temperature is analysed. Special attention is given to analyses of extreme (maximum and minimum) daily differences occurred at three analysed stations. Selection of the fixed local hours, which is in use for calculation of mean daily temperature, plays a crucial role in diminishing of bias from the true daily temperature.