UCL Research Reveals Limitations in Crowdsourced Urban Weather Data
The research, published in Nature Communications, found a significant difference in the distribution of personal weather stations, with greater prevalence in more affluent areas compared to deprived areas. This lack of data could lead to environmental inequalities as these deprived areas are generally home to higher proportions of ethnic minorities, younger populations, lower vegetation coverage and higher building heights and densities.
These personal weather stations are helpful for monitoring local microclimates, particularly during extreme weather events. The researchers are concerned that a data bias towards more affluent areas could skew policies and responses and not fully address the needs of deprived areas.
Personal weather stations, erected by hobbyists and weather aficionados, have become an important tool for climate and weather scientists looking to fill in data gaps where official sources don’t cover. Official networks like the UK’s MIDAS are managed by the UK Met Office and make up the national weather station network. Though each station collects comprehensive data, the distribution of these stations is sparse, and scientists often augment this data with crowdsourced personal weather station data, especially for urban climate studies.
Personal stations are usually smaller and more basic than official government stations, most often collecting only basic data on temperature, humidity, precipitation, wind speed and direction. There are numerous different types, and supply their data into several different data networks. Because of the availability of data, the researchers looked at the Netatmo personal weather station network.
By correlating the locations of the weather stations with openly accessible socio-economic, demographic and environmental data across England and Wales, the team was able to map where personal weather stations were most prevalent, and the characteristics of these areas.
They found that across England and Wales, only 3% of people in the most deprived socio-economic decile have a personal weather station covering their region, while 24% of people in the least deprived socio-economic decile have one.
In addition, the team found evidence of a rural-urban divide as well, with about 22% of the population in natural areas having a personal weather station in their area, while only 10% of built-up city populations being covered. The team termed this deficit, the “urban sensor desert.”
Moreover, the authors believe that this trend isn’t localised solely in the UK, but extends broadly around the world. They project that developing nations, many of whom are facing some of the most severe impacts of a warming climate, are also the ones with the least weather station coverage.
Lead author, Dr Oscar Brousse, Research Fellow in Urban Meteorology and Environmental Modelling, (UCL Bartlett School Environment, Energy & Resources), said: “The more granular weather data we have, the better we can understand the localised effects of climate change and extreme weather events as well as the local impacts of cities in which the vast majority of people live. However, our research shows that the limited data available from deprived areas could hamper our ability to identify and properly respond to climatic changes and extreme weather events, especially for those that have the least capacity to face them.”
The team also looked at age demographic data, and found that while people over 65 years of age were more vulnerable to heat-related health hazards, they tended to live in areas relatively well covered by personal weather stations. This could offer them, as well as local stakeholders, greater capacities to cope with local climate challenges.
Dr Brousse added: “Filling in these data gaps is important to understanding the full picture of heat risk for cities and other disadvantaged regions. Public or private incentives from local stakeholders could encourage wider adoption of such type of weather stations for underserved areas; other cheaper or more accurate options exist depending on the needs. This way, no one would be left behind.”
This research was a collaboration between UCL and KU Leuven University and supported by the Wellcome Trust and NERC.