Homogenization of a combined hourly air temperature dataset over Romania

Abstract

Daily and sub‐daily homogenization of climate variables have been intensively investigated in the last decades, but to the best of our knowledge, this is the first study on homogenization of hourly temperature in Romania. This paper describes the creation of a homogenized hourly air temperature data set at a country scale by combining data from four independent meteorological networks. The air temperature measurements for the period 2009 and 2017 were obtained from the following networks: Romanian National Meteorological Administration (ANM), National Network for Monitoring Air Quality (RNMCA), Regional Basic Synoptic Network (RBSN), and Meteorological Terminal Aviation Routine Weather Report network (METAR). The climatological limits, persistence, temporal variation (step test), and spatial consistency were the quality control tests used to isolate the errors due to malfunctioning of the temperature sensors, data coding or transmission. The Climatol homogenization method was successfully applied for identifying and correcting any suspicious values. The missing data were filled by considering the similarities between each station and the reference series. Comparing the output with the original data, it is apparent that the removal of the break points, correction and homogenization resulted in a new data set with statistical properties very similar to the raw data, but more reliable for climate research due to the increased homogeneity. Eventually, the procedure can be implemented in operational use for collecting more data from other networks.

Uncertainty evaluation of Climatol's adjustment algorithm applied to daily air temperature time series

“…The R package Climatol is a homogenization software that has been widely used in recent years for removing inhomogeneities from collections of raw time series of different climate variables and different time resolution (e.g., Mamara et al ., 2013; Sanchez‐Lorenzo et al ., 2015; Guijarro et al ., 2018; Meseguer‐Ruiz et al ., 2018; Azorin‐Molina et al ., 2019; Coll et al ., 2020; Dumitrescu et al ., 2020). The effectiveness of the software has been evaluated in several benchmark tests (Venema et al ., 2012; Killick, 2016; Guijarro et al ., 2017; Guijarro et al ., 2019) where it demonstrated good results, which are comparable in terms of accuracy to other well established and tested homogenization algorithms.…”

“…and approximate the data to the real climate signal, that took place in some area. Usually the homogenization procedure allows to improve the consistency of the data, which can be seen in the process of a statistical comparison of the raw and homogenized time series (e.g., Mamara et al ., 2014; Prohom et al ., 2016; Osadchyi et al ., 2018; Yosef et al ., 2018; Fioravanti et al ., 2019; Skrynyk et al ., 2019; Dumitrescu et al ., 2020). However, the question that may remain unclear is: how far are the homogenized data from the true climate signal?…”

GSDM-WBT: Global station-based daily maximum wet-bulb temperature data for 1981–2020

“…RHtests (Brugnara et al, 2020), HOMER (Coll et al, 2020), and Climatol (Dumitrescu et al, 2020). These algorithms differ in methods of detecting break points, applicable variables and their resolutions, the number of series to be processed, and the ability of automation.…”

Meteorological and Ancillary Data Resources for Climate Research in Urban Areas

“…Ideally, the observations should be performed in constant locations, quasi-continuously over time, with limited and isolated gaps. The shortcomings related to missing data or changes in station locations can be successfully secured by homogenization procedures [10,11]. The period covered by satellite images can be too short and contain too many gaps for developing climatic studies, but the remote sensing products are valuable for meteorological applications as much as they are consistent temporally and spatially.…”