A compact and customisable street-level sensor system for real-time weather monitoring and outreach in Freiburg, Germany

Climate adaptation and emergency management are major challenges in cities, that benefit from the incorporation of real-time weather, air quality, differential exposure and vulnerability data. We therefore need systems that allow us to map, for example, localised thermal heat stress, heavy precipitation events or air quality spatially resolved across cities at high temporal resolution. Key to the assessment of average conditions and weather extremes in cities are systems that are capable of resolving intra-urban variabilities and microclimates at the level of people, hence in the urban canopy layer at street-level. Placing sensors at street-level, however, is challenging: Sensors need to be small, rugged, safe, and they must measure a number of quantities within limited space. Such systems may ideally require little or no external power, provide remote accessibility, sensor interoperability and real-time data transfer and must be cost-effective for mass deployment. However, these characteristics as well as a wide spectrum of observed variables are not available in current commercial sensor network solutions, hence we designed and implemented a custom partly in-house developed two-tiered sensor system for mounting and installation at 3 m height on city-owned street lights in Freiburg, Germany.

Our partly in-house developed two-tiered sensor network, consisting of fifteen fully self-developed, cost-effective “Tier-I stations” and 35 commercial “Tier-II stations” (LoRAIN, Pessl Instruments GmbH), aims to fill these gaps and to provide a modular, user-friendly WSN with a high spatial density and temporal resolution for research, practical applications and the general public. The Tier-I stations were designed and optimised from the ground up, including the printed circuit board (PCB), for temporally high-resolution WSNs that support wide ranges of sensors and that is expandable. The core of the system is a low-power embedded computer (Raspberry Pi Zero) running a custom multithreaded generic logging and remote control software that locally stores the data and transmits it to a custom vapor-based TCP server via GSM. The software also features system monitoring and error detection functions, as well as remote logging. The setup can easily be expanded on the fly by adding predefined sensors to a configuration file. For better modularity, each station registers itself on the server and will be automatically integrated in all further processes and vice versa. Custom frontends as well as bidirectional communication and task distribution protocols enable remote access and across node interaction, resulting in a more easy-to-maintain system. 

In addition to air temperature, humidity and precipitation measured by the Tier II stations, the Tier-I station feature a ClimaVUE 50 all-in-one weather sensor and a BlackGlobe (Campbell Scientific, Inc.) that provides data on wind, radiation, pressure, lightning, solar radiation and black globe temperatures. That allows for calculation of thermal comfort indices in real-time. A webpage and the self-developed “uniWeather” (iOS-App, API) offers near-realtime data access and data interpretation for stakeholders and public outreach.