The limitations are arbitrary limits intended to flag the calculation
as outside typical limits. There are validation tests where the model
does quite well with compartments as large as 20 000 m^3. Still, as
you have done, such large-scale calculations should be reviewed for
reasonableness of the results. It's more a function of the relative
size of the fire to the compartment; very small fires in very large
compartments or very large fires in very small compartments test the
limits of the model far more that just large compartments with a fire
typical for the compartment size.