I don't really understand your question as such, but perhaps this will help.

tnom is the default temperature at which the components in the circuit (i.e. the models) were characterized. This can then be overridden by having  a tnom parameter in the model definitions themselves. I suspect this is what is happening in your case, which is why it doesn't make any difference. You wouldn't normally mess around with a global tnom unless you have models without this parameter, and you know they were characterized at some unusual temperature.

temp is the simulation or analysis temperature. The models are adjusted using the difference between temp and tnom for that model - although you do also have to take into account that instances may have the "trise" parameter on them indicating a temperature rise from ambient (i.e. the temp you set) - which is a local temperature offset.

This all gets further complicated by the fact that some models include modelling for self heating...

Andrew.