Real-time Cartography in Operational Hydrology
Christophe
LIENERT1, Rolf WEINGARTNER2, Lorenz HURNI1
1Institute
of Cartography, ETH Zurich, Switzerland,
lienert@karto.baug.ethz.ch,
hurni@karto.baug.ethz.ch
2Institute
of Geography, University of Berne, Switzerland,
wein@giub.unibe.ch
The increase of
damaging flood events in recent times demands for enhanced monitoring
capabilities in operational hydrology. Indeed, a working monitoring
infrastructure can help to detect at an early stage a looming flood event and
may improve the immediate preparedness to anticipate and counter it. However,
in order that decision makers can constantly revaluate actual hydrological
situations by comparing and retracing them, easy to grasp data and data
visualizations must be readily available in real-time. Translating real-time
requirements of hydrology into cartography implies that new cartographic
concepts have to be examined. Up to the present, cartographic map production
work steps were mostly accomplished off-line, with human supervision. For a
real-time cartographic application, acquisition, storage, processing,
visualization and archiving of data must be achieved on-line and with the
highest degree of automation possible.
The present paper
outlines a prototype application which features three different visualization
layers. Particularly the first of these layers takes up these new
cartographic concepts by letting users explore the most actual hydrological
situation of a catchment. A series of functions have to be devised for
automatic and repeated operations on different data models. Initially, the data
must be acquired from the data providers, checked for their quality and
deposited in a transaction-oriented database from where it is exported for
cartographic processing. Geometric, semantic-topological and temporal
properties of real-time date have to be linked with geo-referenced base data
and rule-based cartographic design is then to be applied. Automatic insertion
of new data into, and removal out of, the transaction-oriented system is
accomplished according to a “first-in-first-out” principle. Old, processed data
is moved and archived in an analysis-oriented system. For all these steps,
error handling must be granted high priority. The second layer allows a
user to retrace the hydrological developments on the event-scale from a chosen
moment up to the present. Methodically, animations and multi-views will be
applied to visualize developments and different development stages. The third
prototype layer offers functionalities to compare real-time data with
historical data deposited in the analysis-oriented archive. That way, users may
instantaneously classify real-time data and learn from former events and
experiences. Historical data, if resembling actual data, is retrieved from the
warehouse and re-visualized, along with statistical and graphical information.
The methodical challenge of modelling historical data
is to meaningfully select search keys, indexing them and provide fast
retrieval.