Event Spotlight: Urban Heat Island Hackathon

Earth Hacks
2 min readOct 20, 2020

--

Map of areas surrounding Los Angeles, CA

Urban heat is an intensifying global problem as temperatures continue to rise and the world becomes increasingly urbanised. In particular, the urban heat island effect — when urban areas are much hotter than rural areas around them — is a multifaceted issue, leading to threats to public health, excess energy consumption, the increase of smog and even violent crime. Extreme heat is the cause of more fatalities than any other weather related event.

Thankfully, a startup called Urban Canopy, led by the brilliant team from Hello World Labs, is working on a series of solutions to the effects of urban heat, starting with just helping people see it. Earth Hacks worked together with them to put together a hackathon focused on identifying and mapping urban heat islands across the world using data from ECOSTRESS, a NASA Jet Propulsion Lab mission on the International Space Station.

Hackathon attendees were asked to search through ECOSTRESS surface temperature data to find high quality images of cities during heat waves, treat the images in a geographic information systems (GIS) program called QGIS, then to put the data on a map product using Mapbox.

The group was also asked to explore the data more deeply, looking into new relationships between surface temperature, tree cover, and albedo, and using machine learning algorithms to increase the resolution of images.

This hackathon produced the world’s first public map of urban heat islands in the world. Teams worked together to create maps of urban land surface temperatures to reveal urban heat islands, exploring a total of 20 cities so far.

Hackathon attendees feeling “faint from the heat” as they explore elevated urban temperatures

This hackathon was meant to serve as a starting point for global urban heat mapping, so if you would like to create one or get involved, check out the Urban Canopy website here and the mapmaking tutorial here.

Happy mapping!

--

--