based on the terrain following sigma coordinate. As regards vertical
turbulent exchange of heat, moisture and momentum, the Mellor-Yamada
closure schemes level 2.5 and 2, for the Boundary Layer and Surface Layer,
respectively, are used. Precipitation is parametrized as of large scale
and convective origin; deep (precipitating) and shallow (nonprecipitating)
convection is parametrized using the Betts/Miller adjustment scheme. The
model includes also a fourth-order non-linear horizontal diffusion scheme,
with the diffusion coefficient dependent on flow deformation and on
turbulent kinetic energy. The soil is parametrized using a two layer
model. LAMBO is daily running operationally at SMR/ER since the 1st of
September 1993, at two different horizontal and vertical resolutions. The
first run, referred to as "father", is performed at approximately 40 km
horizontal resolution and with 20 vertical levels. Initial conditions are
taken from 00GMT ECMWF operational analysis and boundary conditions, are
extracted from the 00GMT ECMWF forecast. The second higher-resolution run,
referred to as "son", is performed at approximately 20 km horizontal
resolution with 32 vertical levels. This last run takes the boundary
conditions from the previous lower-resolution father run using a suitable
nesting procedure (Paccagnella et al.,1992). The numerical integrations
are carried on for 48 hours.
Model forecasts
In this section we discuss the forecasts produced by the two resolution
father and son runs. Forecasted precipitation is cumulated over 24 hours
to be directly comparable with observations for the 5h and the 6th.
Observed precipitation collected during the 6th is shown in
Fig. 1 since this data was not discussed in
BA95. In the rest of this paper, as sketched in the following table, the
forecasts of both the father and the son runs will be referenced with short
names indicating also the date of emission and the range of validity:
The high-resolution run S4-48 forecasted intense precipitation for the 5th
overall of Northern Italy and particularly over the Alpine chain (Fig.
2b). Comparing with observations (Fig. 3 of BA95), there is a general
underestimation over the western part of the Alps, although the model
produces a maximum of 141 mm in the area south of Cuneo (the southernmost
part of Piedmont) where 286 mm have been observed. Just north of this
maximum, there is a local minimum which is also present in the
observations, but certainly with a notable underestimation. Apart from
this local minimum, there is a good reproduction of the east-west gradient
in the precipitation field within the Po Valley, while precipitation is
overestimated in the eastern part of the Alps. Although available
precipitation data do not allow any detailed verification over the
Apennine, the obcode
MAP Newsletter. October '00