Regional operational prediction requires the development, implementati
on and steady improvement of a modelling system which comprises differ
ent components like the data assimilation scheme, numerical forecast m
odel, postprocessing and application modules as well as packages for m
onitoring, diagnostics and verification. The text highlights the impac
t of some of those components like lateral boundary conditions, data a
ssimilation and statistical postprocessing on the quality of operation
al regional prediction systems and proposes some important research di
rections for an improvement of the systems from the operational point
of view.