The maturing stages of GeoWeather
Learning the structure of the weather data
The first idea was a visualisation of 3D weather, which was only a
3D cloud picture without any data connection.With this Regina St÷ber Yurdakul applied for a job at DWD.
On the background of a 12-year professional experience in 2D and 3D
computer animation, the communication designer worked for several years
at the German National Weather Service DWD as a digital artist, where
she had been responsible for the creation of weather forecast animations
for several TV -stations.
This was an important condition for the
founder to gain deep insight into the structure of weather data.
She had been also involved in the further development of the 3D weather
visualisation system at DWD, which had been designed by meteorologists and programmers.
As a communication designer with a 12-year professional experience in 2D and 3D computer animation,
she tried, to convince the system architects to solve some obvious weaknesses of that system.
She propagated, that the measure for a visualisation system is the communication task and not the technology.
3D can never be an end in itself. The user don't care if something is rendered in 2D or 3D, especially,
if he does not see the difference.
Her rule, which is valid until today:
The look and feel should be cutting edge - as realistsic, as possible,
but the system should use always the cheapest visualisation technology in terms of
required computing power and rendering time..
The local TV-Weather:
With a clear concept of a Weather-Visualisation system based on 2D
compositing , she started in 1999 the founding of IP-Wetter (the name was WeatherArts at that time).
Weather films have to be local, to
communicate 7 km (today 2,8 km for germany) resoluted numerical weather
This is a local weather, visualizing
the german NWP model LM (7km), ready for daily production on the
compositing system “Shake”
This is an example of 3D local weather with realistic clouds and landscapes from the year 2001.
It shows a 6 hour forecast of summerly convective clouds, produced as 3D-Particles
with the system“Maya” and
composited on a 3D digital terrain background.
Developing new concepts of
processing: 3D Visualisation as starting point
Nowcast ( forecast range up to + 6 hours ) is the most interesting
forecast range for Location based weather and moreover the most
precise, with a resolution of 1 km and better.
According to new developments in
satellite picture analysis,
developed a concept and applied for a funding of a research project “
Nowcast -Project”, where the design of data processing of the forecast
is based on the visualisation. The simulation of
realistic 3D cloud objects and conceptual models of weather
parametrisations, with a preproduced
lifcycle, was conceived as the starting point for designing the
predict the further development of the observated weather.
this classes are visualised in 2D in the
prototype of Geo-meteo
Go to clouds, timepoint “jetzt”.
This is the visualisation of different cloudetypes, which become more
and more realsitic, the more you zoom in.
Static Internet version
A Flash based internet version of the local weather was finished in
The internet enables much more
The user has the possibility to see the animation or to navigate
through the timeline to focus on timesteps. He/she can choose
also the parameter.
The weather films needs not to be produced before transfer, but are
composed and rendered on the client side in the Flash-Player.
This is a big advantage , because a daily production of a big amount of
local weatherfilms for TV is a huge and expensive effort.
version was a cost effective
A new data processing was implemented at the serverside.
5. Stage :
Discover the world of Web-mapping and
open source tools
2005 The BBK (Federal Office for civil protection and disaster control)
assigned IP-Wetter a feasibility study for the better supply of
disaster control with Nowcast, in the case of chemical transportation
A concept of an internetbased and Gis
based nowcast processing ( OGC Web-Processing Service) and
visualisation was developed.
The task was to process a precise nowcast
, which can be combined with a windprediction on scales with a much
higher level of detail ( streets and houses). ( see also
The main result of that investigations
for the progress of GeoWeather was the discovering and study of web
To treat weather as a geografic
was the last breakthrough for Geo-weather.
Weather forecasts have always been a
geografic feature, because it is the science of the Atmosphere, which
belongs to the earth.
Weather was for ever visualised on a map. But the
map was always not more than a
simple, static BG.
This was the last hurdle to be
resolved, the transfer of the
advanced web mapping technologies with easy and fast pan and
zoom to the weather area
was the last step, to bring the local weather to fly.
Because weather is characterized in difference to geografical maps by the time dimension, fast animations on
had to be enabled. For that some special
required, which geografical maps do not have.
This task had been accomplished with GeoWeather.
A worldwide zoom with fast animations on each zoomlevel is a main USP of Geo-wind.
But this is only the beginning of GeoWeather. The technical conditions
for 3D interactive weather
Zoom in the time dimension
Similar to the different level of detail on a zooming map
, there are also different scales in the time dimension.
For example the observations of lightnings are done in splits of seconds, other observations are normaly in 15 min
intervals, the 48 forecast has hourly intervals, the 7 day forecast has 6 hourly intervals 12 hourly or daily intervals.
In climatic views there are daily, weekly, monthly intervals.
This different scales are to be considered in data processing and are visualised by a zoom in the time dimension. First
Solutions for the latter are prototyped, improved concepts are existing.
Dataflow in Nowcast for roadweather.
This is the place to express much gratitude to the unknown developers of all the free software tools, which were used, and to all the known contributors, who actively advanced the project with consultation and coding.
The founder herself, thrilled with the triumphal of free software, visited a lot of conferences and exhibitions in this field. Without that knowledge, nothing would have been achieved.
Special thanks to the FSF member Volker Dormeyer, who developed in 2002 the first serverside data processing and the frontend developer┬á Dirk Friedenberger, who developed in 2006 a preliminary stage of the prototype geo-meteo. The latter has been developed 2009/2010 during the incubation, funded by┬á the ESA incubation center Cesah.
Without all that groundwork, the many tests and gaining experiences over all the years, GeoWeather would have not become, what it is today.