It seems that the UWG of DF doesn’t take into account any of the parameters that we input.
I downloaded the example of @chris to create an urban weather file from building footprint and updated all the DF component with the latest version of march 2020. At first, when I compare the original epw and the urban epw, I see a difference, which makes sens.
But when I try to play with the parameters, I have almost the same urban epw file. I tried with a city with 0.04 building ratio, 50% grass and 50% trees, almost no heat gain, wich should look like the rural configuration and I have almost the same weather file as a dense city without any vegetation. See below comparaison where all of the urban epw file are the same:
The underlying assumptions in the UWG don’t lend itself towards modeling a rural scenario like you are describing, since it’s assuming its modeling a city. Without modifying advanced parameters governing the more geographic-scale parameters, you won’t be able to get the rural/airport EPW to match the transformed EPW just by changing the urban density, and vegetation coverage.
There was some recent bug fixes regarding the vegetation coverage which I’m not sure will be integrated into the dragonfly-legacy model. Can you transition to the [+] libraries?
How long is your simulation? I only see three days in your image. If it’s that constrained you may not be giving the simulation enough time to warm-up. I would suggest modeling at minimum a month.
If you’ve done all that, I would finally suggest taking a look at Aiko Nakano’s thesis which indicates which variables have the most impact on the UHI effect.
I think this is mostly likely the reason for this - UWG makes some assumptions about urban and rural heat exchange / mixing through urban boundary layer (UBL) that wouldn’t quite cancel out even if the urban surface is completely covered by vegetation. The assumption is that urban heat island effect simulation would only be used for relatively densely urbanized area.
