You can define an indexed collection of tables, for example using a statement like this before the loop:
table TabScenStab {nscen in ScenCard} IN "ODBC"
"../scenario generation/scenari tip1 stability.xlsx" ("Tab"&nscen):
[j ~ Nodo, s ~ Scenario], , {t in 1..T} <g_out[j,"I",t,s] ~ (t)>;
and then reading one table in each pass through the loop, with the statement
read table TabScenStab[nscen];
The string expression ("Tab"&nscen) will evaluate to Tab60 for table TabScenStab[60], to Tab100 for TabScenStab[100], etc. So in your spreadsheet the data for TabScenStab[60] should be in a range named Tab60, the data for TabScenStab[100] should be in a range named Tab100, and so forth. Because SC_out is changed in each pass through the loop, these data ranges will be of different sizes.