working with the RETICS research group, we routinely do and deliver (by free) operative forecasts at 3km resolution over Spain. Other domains are avaiable upon request.
The lite version is here.
We work mainly with WRF, both the standard version and in-the-house 'avatars' of the code (the WRF-UCLM model). The reason for using community models is threefold: firstly, the code is publicly available so the bugs and errors are minimized and auditing is made possible; secondly, having thousands of scientists worldwide working together ensures a quick and full validation of any new development; and latest but not last, WRF model implements the most advanced physics, avoiding the risk of quick obsolescence and the pitfalls arising from the use of incomplete models.
we have been measuring the spatial structure of precipitation at small and medium scales. Our disdrometer network is complemented by the x-band radar used in dedicated campaigns.
we have been performing some exercises analyzing hurricanes such as Vince, severe weather events, and case-studies of particular precipitation events.
We have also been analyzing the effects of several parameterizations on precipitation forecasts, and found that any ensemble approach requires a careful consideration of the individual performances of the ensemble members -adding a deficient or incomplete model won't add much value to the overall result.