|
NinJo
High Performance Meteorological Visualization
In the year 2000 the German Meteorological Service (Deutscher Wetterdienst
DWD) and the
German Military Geophysical Service (GMGO) jointly started a project for a
new
meteorological workstation called NinJo. Shortly thereafter MeteoSwiss and
the Danish Meteorological
Institute (DMI) joined the project. In mid 2003 the Meteorological Service
of Canada (MSC)
decided to become a member of the NinJo consortium.
The aim of the project is to replace aging workstation systems and to unify
the existing software structure
supporting the forecasting and warning operations. The technical goals are:
performance, stability and portability
Since NinJo is a international project there is a strong demand for an open
architecture to allow the adaption to various hardware platforms, local
software structures and special needs of the partners. At the very beginning
of the project after prudent evaluation, the JAVA API was chosen, due to its
platform independency. NinJo is
completely written in Java, even the servers and numerical computations.
NinJo is an configurable framework, which uses layers to add special
functionalities. NinJo displays all kinds of meteorological data, for
example Grid-Data (output from numerical meteorological models) or
Point-Data (surface observations from weather stations). The visualization
of NinJo heavily depends on the Java Advanced Imaging API to display huge
amounts of data. JAI is used to load, scale and project Raster Data (height
data based on gtopo30, Landsat, and more) as tiled images. This tiling makes
the visualization independent from the actual image size. The calculation of
geodetic
projections is done in real-time. JAI's built-in support for tiled TIFF and
multiresolution images helped
to speed up the development of base components. NinJo always displays the
optimal resolution for the given
viewing area and is able to visualize composits of multiple satellite
channels (e.g. visible and IR)
NinJo is to be operational by the end of 2004.
|