10th International Conference on Hydroinformatics
HIC 2012, Hamburg, GERMANY
INTEROPERABLE MARINE WIND DATA FOR THE GERMAN
MICHAEL BAUER (1), RAINER LEHFELDT (1)
(1): Federal Waterways Engineering and Research Institute, Wedeler Landstr. 157, Hamburg, 22559,
For many environmental models, wind data is an important parameter. This has been well-known for a
long time, but data acquisition and transfer could be quite problematic. Thus in 2007 a digital wind
atlas for the German sea was established. Unfortunately the technology did not age well, leading to a
relaunch, using sustainable web service technology. In the course of this article we will recap the
original design and show our efforts to reestablish the wind atlas in its new garb as well as outline our
future plans for enhancing the application and integrating it in vaster information networks.
Digital wind data is a vital input parameter for marine and climate modeling as well as for
engineering projects such as harbor design, dike planning or coastal defense works. Pre-
calculated wind scenarios can be used as a fast alternative for time consuming on-demand
modeling. A former research project concluded in a digital wind atlas that consisted of pre-
calculated coastal wind scenarios well documented with metadata.
Wind fields in the German Bight at a height of 10m above sea level and the
corresponding shear coefficients have been calculated, which can be used as input data for
further analyses. The input parameters for this systematic study were specified at a height
of 800m at the upper boundary of the wind model. Here the wind velocity, wind direction
and water elevation were prescribed in typical ranges, see Niemeyer 2005 .
When this work was conducted, data infrastructures and spatial web services were, of
course, still developing concepts, so the previous implementation with stand-alone java
clients and Java Webstart applications can rightfully be called dated.
Still being convinced by the concept and our hand being forced by a fatal server crash
which took the application offline for good early last year, it has been decided that the
digital wind atlas should be reborn in the shape of an up-to-date and sustainable web
mapping application, that can be integrated into spatial data infrastructures (SDI). This
naturally leads to the use of OGC (Open Geospatial Consortium) web services as
foundation of interoperable spatial data infrastructures.
MAKING THE TECHNOLOGICAL TRANSITION
When the digital wind atlas has been designed in 2007, digital atlases were a quite
challenging exercise due to the turn-around times for map creation from extensive datasets.
Software design always reflects current trends in software engineering and must take into
consideration the contemporary hardware. This applies especially for software solution
tailored to govermental agencies, as these tend to use a wide range of soft- and hardware,
from high-end computers and up-to-date programs to rather outdated machines and
products at the end of their life cycle.
For these reasons, a platform indepent software was the aim of the original approach, with a
server shouldering the main workload. This led to a three-tier client-server architecture, see
Figure 1. The technology of choice on the client side was Java as it is platform independent
by definition. It was used to create Java applets as well as stand-alone applications to access
the wind data and execute user-defined data manipulation operations such as interpolations.
Figure 1. Original Architecture Lehfeldt R. 2008 
These operations were handled in the application layer, situated between the client and
the server that also took care of secure communications between client and server by
utilizing the Remote Method Invocation (RMI) standard.
Heart of the system was the object-oriented db4o database on the server side, see
Paterson 2006 . It hosted the model metadata and simulation data as well as the selection
and interpolation methods.
For the new installation of the wind atlas we profit from the advancements in web
browser software, thus eliminating the need for such a complex architecture or client
In the new installation, all the visualizing and interaction components are accessible
through the browser, relying on free and open source libraries like GeoExt 1 and
OpenLayers2. Generally we opted for open source software and the use of open standards to
ensure a sustainable software environment. With the use of freely obtainable open source
tools, the wind data could be migrated into a modern spatial database (PostGIS3) and made
available via OGC Web Services.
The OGC (Open Geospatial Consortium) is a formation of the most influential
corporate, academic and governmental bodies with the goal to define open and
interoperable geo-standards. Their main objective is the standardization of services for geo-
data access, management, manipulation, representation and sharing over computer-
For the implementation of a new wind atlas, we utilized two of the most basic OGC
services: the Web Map Service (WMS) and the Web Feature Service (WFS), see La
Beaujardiere 2002 , . A Web Map Service, as the name indicates, provides maps over
a network. In order to do this, it accesses the data in a database, shape file or other geo-
format, computes a map thereof and renders it as an image. This image can be influenced
by providing the WMS with styling information in form of a Styled Layer Descriptor
(SLD) document, see Y. Coene et al. 2007 . While the WMS renders the data into a geo-
referenced image that can be viewed in a browser window or appropriate viewer, a WFS
provides the download functionality for the underlying data.
The metainformation is stored in NOKIS, a separate, metadata management tool,
developed especially for marine data that provides a Catalogue Service for the Web (CS-W)
interface, see Lehfeldt et al. 2008 . Our choice for NOKIS is rooted in its availability in
German coastal zone agencies and a long term development cooperation with the software-
producer. There are several open source products in existence that can handle the necessary
Figure 2. Revised Architecture
The WMS, WFS, and CS-W services provide the basic functionalities for map creation,
data downloading and metadata provision in a standardized way, see Figure 2. For easy
access a small web portal will be set up, where data can be searched via the metadata,
displayed as interactive maps and downloaded for further use.
As mentioned above, the original system worked on a custom-made architecture and
proprietary tools while the new incarnation focuses on using open standards and
standardized interfaces. So apart from NOKIS, which is a running stand-alone application
also employed in other contexts, little could be re-used. The wind data itself was salvaged
in the form of Fortran-generated text files. These had to be converted by a tool, especially
written for this purpose.
It consisted of several Java classes, parsing the text files and transferring the data from
a complex, multidimensional storage matrix to simple geo-objects. While the original
format was densely packed and needed some logic to extract the exact wind values
according to the boundary conditions, the new data model consists of a geo-object for each
possible combination of boundary conditions. This leads to a significant increase in storage
space needed but also enhances the performance, especially for the WMS if a tile cache is
used. With the decline in prices for storage media and the performance requirements
imposed by the European spatial information infrastructure INSPIRE and other directives,
this is a trade-off worth paying.
USING THE DIGITAL ATLAS
Users of the digital wind atlas usually have two goals: to visualize a certain state and/or to
download one or several datasets. Therefore a web interface catering to those needs doesn’t
have to be overly fancy.
The interface of the digital atlas therefore will be rather simple and consist of only few
fields to choose the boundary conditions from pre-set values for wind velocity, wind
direction and water elevation. Once the choice is made, the user has the options to view or
download the data, inspect more datasets, view the model metadata or download all of the
wind data in a compressed file format.
If the user chooses to view the data, it will be provided by the WMS and displayed on-
site in a map viewer powered by GeoExt and OpenLayers. Does the user opt to download
datasets, the WFS then provides different file formats like gml  and shapefile for the user
to save onto his local machine.
Thanks to the above described overhaul and relaunch, the digital wind atlas is again up-to-
date on the technical side, while the data itself never lost it relevance. Thanks to the use of
standard compliant tools, the pre-calculated wind scenarios from the German North Sea
coast are now ready to be integrated in data infrastructures or downloaded for local
processing. Data acquisition procedures are simplified for those in need of wind as an input
parameter for numerical models.
This new, standardized implementation provides interoperable data access on a
sustainable technology as required by current national and European spatial data
The work presented here is funded by the Federal Ministry for Education and Research
(BMBF) through Project Management Jülich (PtJ) under grant number 03KIS089 (Marine
Information Infrastructure for Germany MDI-DE). The authors gratefully acknowledge this
support as well as the contributions from co-workers and other partners to this research
 La Beaujardiere, J. de,OpenGIS Consortium Web Mapping Server Implementation
Specification 1.3. In OGC Document, pp. 04-024. Available online at
 Lehfeldt R., Milbradt P. HöckerM,Coastal Scenarios documented with Digital Atlases.
Computational Modeling and Metadas. In Jane McKee Smith (Ed.): Prodeedings of the
31th Conference on Coastal Engineering - Vol 5. Held on Hamburg, Germany, 31.
August - 05. September 2008, vol. 5. New York (2008), pp. 4633–4644.
 Lehfeldt, R.; Reimers, H.-C; Kohlus, J.; Sellerhof, F.,A Network of Metadata and Web
Services forIntegrated Coastal Zone Management Paper No. 207. In 7th International
Conference on Coastal and Port Engineering in Developing Countries PIANC-
COPEDEC, pp. 395–396. (2008).
 Niemeyer, H. D. Modellierung des mittelfristigen Seegangsklimas im deutschen
Nordseeküstengebiet (MOSES). Statusbericht Anlage I, Deutscher Wetterdienst,
Geschäftsfeld Seeschifffahrt, 29 S. Edited by KFKI. NLWKN. (2005)
 Ogc,OpenGIS Web Feature Service (WFS) Implementation Specification. Open
Geospatial Consortium, Inc. Available online at
 Open Geospatial Consortium, Inc,OpenGIS Geography Markup Language (GML)
Encoding Standard. Edited by Open Geospatial Consortium. Open Geospatial
Consortium. Available online at http://www.opengeospatial.org/standards/gml,
checked on 31/01/2012.
 Paterson, Jim; Edlich, StefanThe definitive guide to db4o. [the first comprehensive
guide to db4o, the open source native object database for.NET and Java]. Berkeley,
Calif: Apress (The expert's voice in open source) (2006).
 Y. Coene; P.G. Marchetti; S. Smolders,Architecture and Standards for a Distributed
Digital Library of Geospatial Services. In Maristella Agosti, Floriana Esposito,
Costantino Thanos (Eds.): Post-proceedings of the Third Italian Research Conference
on Digital Library Systems, IRCDL 2007, Padova, 29-30 January 2007: DELOS: a
Network of Excellence on Digital Libraries (2007).