Quality control and correction method for air temperature data from a citizen science weather station network in Leuven, Belgium

Abstract. The growing trend toward urbanisation and the increasingly frequent occurrence of extreme weather events emphasise the need for further monitoring and understanding of weather in cities. In order to gain information on these intra-urban weather patterns, dense high-quality atmospheric measurements are needed. Crowdsourced weather stations (CWSs) could be a promising solution to realise such monitoring networks in a cost-efficient way. However, due to their nontraditional measuring equipment and installation settings, the quality of datasets from these networks remains an issue. This paper presents crowdsourced data from the “Leuven.cool” network, a citizen science network of around 100 low-cost weather stations (Fine Offset WH2600) distributed across Leuven, Belgium (50∘52′ N, 4∘42′ E). The dataset is accompanied by a newly developed station-specific temperature quality control (QC) and correction procedure. The procedure consists of three levels that remove implausible measurements while also correcting for inter-station (between-station) and intra-station (station-specific) temperature biases by means of a random forest approach. The QC method is evaluated using data fr