24 
contributions of DWD to a wide range of users. 
German GOOS Secretariat 
The German GOOS Secretariat hosted by the DWD since October 1992 supports the German 
representatives in the GCOS bodies, helps to co-ordinate the German contribution to the 
climate module of GOOS and provides information. The focus of its work is on the Joint Data 
and Information Management Panel (J-DIMP). 
DWD-P3 Operational marine meteorological applications and services 
Objectives 
to supply customers with marine meteorological forecasts, warnings, advice, expert 
opinions, and real-time observation data 
to compile marine climatological basic data, partly on behalf of WMO, for use by marine 
meteorological and oceanographic services, scientific and nautical institutions and the 
public. 
Present scientific knowledge and methods used 
The operational marine meteorological services make forecasts and issue warnings which are 
mainly based on the results of model computations. Empirical methods prevail only in 
nowcasting (0-2 hours) and short-term forecasts (up to about 12 hours). Owing to the 
availability of latest state-of-the-art mainframe computers, the capabilities of numerical 
weather predictions are continually improved. 
For example, the Global Model of the German Weather Service computes forecasts with a 
horizontal resolution of 1.1° and 19 vertical layers. The prediction period covers 7 days. The 
model performance will be further improved shortly by refining the gridsize to 50 km. Periods 
up to about 10 days can be covered by deterministic predictions. This period can be extended 
even further by using the ensemble technique. 
Nested models provide much higher resolutions for shorter prediction periods. The Germany 
model of DWD uses resolutions of 14 km and 30 layers, with a prediction period of 48 hours. 
A reduction of the grid resolution to 8 km will be carried out shortly, and a further reduction 
is being planned. In the new model, within the framework of non-hydrostatic equations, 
vertical motions will be ascertained directly. The deterministic output is improved by filtering 
techniques such as Kalman filtering and statistical methods like the perfect-prog method or 
model output statistics (MOS). 
A crucial requirement in numerical weather forecasting is the consideration of ocean- 
atmosphere coupling. The manifold momentum, energy, and mass transfer processes at the sea 
surface play an important role, especially in medium- and long-term predictions. Recently 
models have been introduced taking into account these processes. 
For wave predictions, models are used which are forced by the predicted windfields from 
numerical weather prediction models. The DWD, e.g., uses a hybrid 2 nd generation wave 
model. Computations are made for the North Atlantic (130 km / 7 days), the European Shelf 
(30 km / 78 hours), and the Baltic Sea (15 km / 78 hours). The follow-up model, the 3 rd 
generation wave model, will be applied on a global scale and will also cover the Mediterranean 
and North and Baltic Seas, and it will be coupled to atmospheric models and additionally, in