Is there any interactions between landunits with in a gridcell?

[Yuan Sun]

Hi all,

When I conducted a control run and an experiment run in a land-only component set, the only change was urban parameters.

However, rural outputs like TSA_R and TG_R are different in the two runs. For example, the global-mean TG_R is similar in two runs, but in some grids, the difference of TG_R between two runs reaches 2 °C.

I am not sure if the changes in rural variables should be attributed to model internal variability or urban-induced. This also influenced when I assessed changes in UHI as I could not distinguish whether the UHI changes come from urban factors or not. I could survive using global-mean value ignoring some abnormal grids.

Thanks for any comments.

Best,
Yuan

 

[oleson]

That is surprising. There shouldn't be any interaction between the urban and rural landunits in a land-only simulation. Although I haven't checked that behavior in recent versions of the model. I guess I would check for any differences in forcing, your model configuration (e.g., the lnd_in), etc

 

[Yuan Sun]

Hi, sorry for my previous confusing information. I ran land-only simulations with the same initialization so the changes are not internal variability, but from PE layouts. Thanks, Keith.