Urban heat through smartphone battery temperatures

“Sensors in smartphones allow atmospheric scientists to monitor the environmental quality in data sparse regions!” (Gert-Jan Steeneveld, WUR Project Leader)

A snapshot of the project

The main scientific problem that motivated this research is the emerging problem that the environmental quality of cities is under pressure due to the urban heat island effect, i.e. the phenomenon that cities are warmer than the countryside. Especially on hot summer days this may result in an urban heat island effect of 5-8 K which has negative effects on human health, labour productivity, and energy demand for cooling. In addition, climate change is creating an increased heatload. Hence it is tempting to measure the urban heatload, its spatial extent and relation to the urban morphology (greenness, building density, building height – street width ratio). At the same time air temperature observations are scarce in cities, which motivated us to develop this research line.

The project aim, therefore, is to explore whether smartphone environmental data as collected via the OpenSignal App can be used to monitor the environment in data-sparse regions, i.e., cities where official weather stations are rare because of high costs and technical challenges, in these cases smartphone observations fill a knowledge gap. Thus, the aim is to monitor urban air temperatures to diagnose the urban heat island effect and its spatiotemporal behaviour using battery temperature observations. To do so, a model has been developed that relates the smartphone battery temperatures to air temperatures. In a later stage also cloud cover was successfully estimated from the smartphone’s light sensor.

In this method massively collected smartphone battery temperatures as recorded by the OpenSignal smartphone App is used to estimate air temperatures in urban areas to study urban heat islands. Using a citizen science approach, data is collected from citizens who download the OpenSignal smartphone App on their phone. Citizens simply move around the city following their schedules, so they are not required to make specific moves or do particular activities. The OpenSignal App records the battery temperature which is used in a later stage to estimate the air temperature in the city. The principle behind the method is that the power produced by a smartphone that is carried by a person result in a battery temperature and the smartphone want to release its heat towards conduction through clothing towards the air. A statistical model was developed to estimate the air temperature from the battery temperature, and this was illustrated and works well for multiple cities across the globe when validated against official and traditional observations by the World Meteorological Organization.

Figure 1. Availability of smartphone temperature records for a warm episode in 2015 in Amsterdam
Figure 1. Availability of smartphone temperature records for a warm episode in 2015 in Amsterdam

The results of the project

The project has been resulting, so far, mainly in scientific impact by creating a method to estimate air temperatures from smartphone data which was tested for eight cities worldwide (see Overeem et al., 2013). It was shown that already about 800 smartphone recordings per day are sufficient to estimate the daily mean urban heat island effect. This method has been refined to neighbourhood scale in a case study for Sao Paolo (Droste et al., 2017). Therein we successfully linked the urban heat island effect as estimated from the smartphone data to so called local climate zones, i.e., archetypes of urban morphology. Certain local climate zones clearly have different urban heat island effects than others. This is a major achievement using free data only, which provides useful information for spatial planning and design, energy demand planning, and health services. Moreover, taking benefit of the smartphone’s light sensor cloud cover could be well estimated, which helps to fill the spatial data gap of cloud cover, which then can assist applications in agriculture.