On 19/08/26 at 15:11 +0000, Zlatan Todoric wrote: > On 2026-08-19 10:27, Matthias Urlichs wrote: > > On 18.08.26 23:21, Zlatan Todoric wrote: > > > >> (some cities reported already an average increase of 4C > >> because of AI datacenters), > > Source? IMHO that number is highly suspect. > > Good call, it is 4 Fahrenheit (2.2°C) instead of 4 Celsius: > https://asmedigitalcollection.asme.org/sustainablebuildings/article/7/2/024501/1233035/Data-Center-Waste-Heat-as-an-Emerging-Urban
According to the article, 2.2°C is the difference between the lowest single measurement (upwind), and the highest single measurement (downwind). See Fig 2, and Table 1. 44.3 - 42.1 = 2.2 It's not a city-wide average, it's about measurements made directly near the DC (< 500 meters away), and if you want to talk about averages, the difference between the averages of the upwind and downwind areas on the map in Fig 2.a is 0.8°C. Also, if you look at Fig 2.b, the article explains that there's a downwind area that is actually colder than the upwind area because of a park between the DC and the area (red cross on Fig 2.b). So the negative effect of the DC is actually weaker than the positive effect of the park. Lucas
