Evaluation and application of a low-cost measurement network to study intra-urban temperature differences during summer 2018 in Bern, Switzerland

• Published in: Urban climate Vol. 37; p. 100817
• Main Authors: Gubler, Moritz, Christen, Andreas, Remund, Jan, Brönnimann, Stefan
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.05.2021
• Subjects: Air temperature; Low-cost; Measurement network; Radiation shield; Urban climate; Urban heat island; Low-cost; Radiation shield; Air temperature; Urban climate; Urban heat island; Measurement network
• ISSN: 2212-0955; 2212-0955
• DOI: 10.1016/j.uclim.2021.100817

The understanding of intra-urban air temperature variations is crucial to assess strategies for cities' adaptation to impacts of present and future anthropogenic climate change. Depending on extensive measurement networks, high-resolution air temperature measurements in urban environments are challenging due to high instrumentation and maintenance costs. Here, we present a low-cost measurement device (LCD) consisting of a temperature logger and a custom-made, naturally ventilated radiation shield. Besides intercomparisons with automated weather stations (AWS) at three reference sites during record-dry summer 2018, we tested the potential of the devices using a network of 79 LCDs to assess the intra-urban variability of urban heat island (UHI) patterns in the city of Bern, Switzerland. We found positive mean measurement biases between LCDs and AWS of 0.61 to 0.93 K (RMSE: 0.78 to 1.17 K) during daytime, of which up to 82.8% of the variance could be explained statistically by solar irradiance (radiative heating) and wind speed (insufficient ventilation). During night, average measurement biases were markedly lower and eventually negative with −0.12 to 0.23 K (RMSE: 0.19 to 0.34 K). Our results highlight the importance of sensor intercomparisons being conducted at multiple locations with differing urban land-cover, structure, and metabolism given that biases varied considerably between the reference sites. Data retrieved by the city-wide measurement network showed that the LCD approach is well suited for the analysis of spatiotemporal UHI patterns during night and adds considerable value compared to the few existing AWS in detecting fine-scale air temperature variability. In conclusion, the current LCD measurement approach represents a valuable option for cost-effective analyses of urban air temperature variability across multiple scales, which may be of particular value for the development, appliance, and monitoring of adaptation strategies to climate change in cities with restricted financial resources. •Design of low-cost measurement devices for urban air temperature mapping.•Intercomparison at three different reference sites against high-quality reference.•Better performance during night- (RMSE: 0.19–0.34 K) than daytime (0.78–1.17 K).•Up to 82.8% of variance in daytime biases explained by solar irradiance and wind speed.•Nocturnal urban heat island intensity in Bern reaches values up to 6.74 K.