 |
| Titel |
OpenMI 2, an extended open standard for integrating models and tools |
| VerfasserIn |
Peter Schade, Rob Knapen, Stanislav Vaneçek, Jesper Groos, Stef Hummel, Gennadii Donchyts, Adrian Harper, Jan Gregersen, Peter Gijsbers, Andrea Antonella |
| Konferenz |
EGU General Assembly 2011
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 13 (2011) |
| Datensatznummer |
250052661
|
|
|
|
|
|
| Zusammenfassung |
| Linking models from different domains so that they can complement each other and thereby
provide comprehensive and balanced assessments requires both technical and semantic
integration. This contribution will focus on the Open Modeling Interface OpenMI, an open
standard that offers interfaces and tools for linking qualitative and quantitative models into
modeling systems. It is being developed and maintained with the vision to become the
accepted standard for model linking both in the hydrological and the broader environmental
domain.
After the initial release of the OpenMI version 1 in 2005 the OpenMI-LIFE project
demonstrated its usability at the operational level on real world scale problems. Several use
cases of the river Scheldt rising in France and running through Belgium and the Netherlands
will be shortly described as an example for an international and multi-disciplinary study.
River, sewer, water quality and estuarine models have been linked together to different
modeling networks according to the stakeholders’ interests. Use of the OpenMI to integrate
components outside of the hydrological domain will be illustrated with a description
of the generic GEI reader capable of delivering meteorological wind fields into a
three-dimensional estuarine modeling system, and a short section about the SEAMLESS
project integrating agricultural and economical models from farm field to world
scale.
The OpenMI is seeing a significant increase in the number of applications that leads to
new requirements. For example integration of non-model components like the web-based
hydrologic information system HydroDesktop or the decision support system AM-DSS was
possible but required some effort. To improve the OpenMI and advance it to the next
version several European research institutes bundled activities and, after a modeling
community review period, released the OpenMI standard version 2 in December
2010.
Some of the key ideas of the new version that will be explained include:
A geographical representation for exchanged data approaching the common
representation in the GIS-world.
The Use of adapters, e.g. interpolating in space or time, to transform data into
a requested form and of series connections of several adapters offering a piping
and filtering mechanism.
Removal of the restriction to time step based models enabling the integration of
new kinds of models, e.g. neuron networks, in the future.
Setting and varying boundary conditions for individual models for running
comparative simulations, simplifying the use of the OpenMI in DSSs and tools
for calibration, optimisation and data assimilation.
The introduction of a set of mandatory base interfaces and sets of optional interfaces has
the aim to make the OpenMI fit for future requirements of integrated modelling. The current
edition includes the extension for time and space dependent components. Future extensions
could support compliance with the standards of the Open Geospatial Consortium (OGC) or
request and deliver data in terms of dictionaries or ontologies.
The OpenMI is available under the terms of the open source license LGPL with the aim of
easier dissemination. It is implemented in C# and Java and offers ways for wrapping code
from other languages. There are guidelines for migrating existing models to OpenMI
compliancy in order to maintain approved legacy code.
With the results from operational use and the explanation of its extended feature set this
contribution will show that the OpenMI is becoming a valuable tool for integrating models
and related tools, both for single domain applications or across domains supporting
multi-disciplinary studies. |
|
|
|
|
|
|