The topic of heat in the city is becoming more and more present in Germany. The so-called tropical nights (when the temperature does not fall below 20° C at night) occur rather rarely in Germany, but in the "heat summers" of 2003, 2015 and 2018, the number of tropical nights rose to double digits in many cities. Dense development, the expansion of transport infrastructure, and the sealing and building up of climatically important open and green spaces are making cities even more vulnerable to the consequences of climate change. Large concrete, glass and asphalt surfaces store far more heat than parks and other green spaces. The result is heat islands whose temperature difference compared to their surroundings can be as much as 10 Kelvin. Heat islands can lead to health impairments for particularly vulnerable parts of the population and cause high damage and costs to the infrastructure, for example through burst asphalt ceilings and warped rail tracks.
The goal of KLIPS is the development of an AI-based information platform for the localization and simulation of urban heat islands (UHI). In order to obtain a detailed area-wide overview of the temperature development within the city, a close-meshed temperature sensor network will be established in the two pilot cities Dresden and Langenfeld. The collected data will be combined with other relevant data on an information platform in order to gain important insights into the topic of heat islands in the city. With the help of the KLIPS information platform, the effect of measures against overheating of these areas can then be simulated.
How can urban heat islands be detected, predicted, and simulated? How can high-density sensor swarms be used to predict heat islands and simulate the influence of mitigation measures with the help of an AI-based algorithm? Hypothesis: A deep neural network (DNN/RNN) trained with meteorological and settlement structure mass data is able to forecast urban heat island effects and to simulate UHI under various urban layouts.
In order to obtain a detailed understanding of the temperature distribution and course within the urban area, high-density temperature sensor networks will be set up in the two study cities of Dresden and Langenfeld. The data will be linked with other data sets, including remote sensing, urban structure, traffic and other weather and climate data. The combined use of innovative swarm sensors and artificial intelligence will not only localize heat islands, but also create a tool to simulate possible mitigation measures in advance for long-term planning. The project will result in a set of methods as well as a web-based tool to support municipal planning in the investigation and long-term prevention of urban heat islands. The tool will be designed together with local stakeholders Dresden and Langenfeld city councils) and adapted to the specific requirements of the respective urban areas.