Urban Microclimate Simulation with Humidity
With rising temperatures and more people than ever living in cities, urban heat islands (UHI) become a matter of increasing importance for communities and urban planners. With our simulations we can predict urban microclimate with wind, solar radiation and humidity, with the latter including all thermodynamic cooling effects from evaporation and evapotranspiration from green roofs and walls, water bodies, trees, ...
Before explaining our approach in greater detail, have a look at the animation below, showing local microclimate in an urban area in Vienna, Austria over a couple of days.
How does this work? For our simulation we picked the Vienna Kabelwerk, an urban development project that features a central public square with a fountain, trees and a panorama pool. For good measure we have added some grass areas and a green wall. When the sun hits a surface, the surface heats up and in turn, heats the surrounding air. This process takes a while which is why the air stays relatively cool through the morning and reaches its maximum temperature well after the sun has passed its highest point in the afternoon. Our model includes this solar radiative influence and couples it with the air velocities resulting from the interference between wind, terrain and buildings. The inlet air temperature, humidity and wind speed change in accordance to local historic day/night values throughout the simulation.
Our model also accounts for evaporation of water from the central fountain, trees, grass and the rooftop pool. Evaporation is an essential ingredient to assess the local microclimate: not only does it cool down the air directly, it also influences the local humidity levels which plays an important role for human comfort. This means that our model can predict human comfort including all four main factors: air velocity and temperature, radiation and humidity.
Just recently the University of Delft published an article about Uchimizu, a japanese tradition to reduce heat around temples and in cities, simply, by sprinkling water on the ground - this is essentially what is happening with the central fountain in our simulation, with the temperature drop of 2°C in the simulation agreeing with the measured temperature drop of ~2°C of the experiment of the University of Delft.
For urban planners and real estate developers this can answer questions like: How will people use this square? Where is the best place for an outdoor dining area? How much influence will green features (roofs, walls, ground) have on the urban heat island effect?
The image above shows a comparison of a green roof and concrete roofs (for two buildings at a height of about 35 meters). It is clearly visible that the air temperature above the green roof is about two degrees lower than above the concrete roof due to evaporation. Consequently the relative humidity is about 10% higher above the green roof. The air temperature at the right most corner of the concrete-only building is especially high. This is a nice example of what is causing the Urban Heat Island effect: the two walls in the corner shine secondary infra-red radiation onto each other which increases their surface temperature; additionally the air moves a bit slower - both effects combined result in increased air temperatures.
Our model is not restricted to outdoor areas, it can also be used for assessing indoor climate and also the interaction of both. We can fully resolve curved walls, holes, bridges, arcades and terrain. Our solution scales to huge areas, depending on the surface resolution and level of detail of surface areas up to 25km². Since the simulation can be run in parallel on hundreds of CPUs, computation is fast, in many cases better than real time (i.e. 3 wall-clock days calculation for 5 simulated days).
Currently our software is in an advanced demonstration stage, in the future we will add detailed soil models, more detailed plant models and support standardized models for architectural features like extensive and intensive green roofs. Already now all this can be simulated using manual configuration.
If you are interested in an demonstration or want to contribute to the model refinement, please contact us per Email!