NOTE: The galeon mailing list is no longer active. The list archives are made available for historical reasons.
Hi,
One of my action items resulting from last week's WCS.SWG was to contribute
use cases representing the needs of the GALEON community. Of course there
are many such use cases and they can be described in several ways. I have
chosen a scenario that embodies the great variety of data types required, a
need for real-time access, and a situation in which the use depends on
having the data in the full 3 spatial dimensions as well as several time
lines: observation times, model run times, and the times the forecasts
apply.
I have attempted to describe the use case in terms that a non-expert, end
user of the system, can understand. But I realize that it may eventually
need to be translated into the language of the standards community. My
sense is that it will end up being many use cases at that point and will
perhaps lose some of the coherence it has in this form. My hope is that
this will be a useful starting point for discussion and will stimulate more
and better GALEON-related use cases.
Your comments, corrections, and suggestions will be greatly appreciated.
-- Ben
==================================================================================
Outline of Coverages Use Case for High Res Weather
Forecast Models
Draft by Ben Domenico
To improved air traffic efficiency and safety, an organization wants to
assemble the data needed to run high resolution local weather forecast
models for the region surrounding an airport and to serve the input data as
well as the output of the forecast model. The goal is to base the data
services for both input and output datasets on standard data models,
interfaces, and protocols wherever possible.
On the input side, this includes real-time data from a wide variety of
sources and many different types:
-- point data from lightning strike observations
-- "station" observations from fixed weather stations in the region
-- vertical profiles from nearby balloon soundings and wind profilers
-- trajectory data obtained from other aircraft which have taken off and
landed recently
-- volumetric scans from ground-based radars
-- visible, infrared, and water-vapor (and possibly other wavelength)
satellite imagery
-- gridded output from national or hemispheric weather forecasts (typically
run at centers like NCEP* and ECMWF**) to be used as boundary conditions for
the local forecast model.
The observational data (point, station, vertical profile, trajectory, radar,
and satellite) is run through a process atmospheric scientists call "data
assimilation." Assimilation can be thought of as a highly sophisticated
"interpolation" to the grid on which the local forecast model will be run.
It involves solving equations representing the physics of the situation in
addition to the usual geometric and statistical interpolations. The grids
resulting from assimilation are to be served in such a way that the local
model can use them for the initial conditions. Thus additional datasets
have to be served via standard interfaces
-- gridded datasets resulting from assimilation of observational data
Using the assimilated observational data for initial conditions and the
hemispheric model output for boundary conditions, the local forecast model
runs to produce its own high-resolution gridded output which is another
dataset to be served via standard interfaces.
-- gridded data output from local weather forecast model in region of
airport
There are a few things to note.
1. Local forecasting systems such as the one described above are already
being run on a regular basis, but the OGC and ISO data services standards
are not far enough along to be used in the existing systems.
2. All the datasets in the scenario above can be mapped into the ISO 19123
abstract coverage specification, but it is not clear whether the WCS is the
appropriate protocol specification for serving these datasets.
3. In addition to mapping such datasets to the ISO abstract model, it is
important to have "application profiles" for encoding these types of data if
they are to be served via standard protocols.
* NCEP is the US National Centers for Environmental Prediction
** ECMWF is the European Center for Medium Range Weather Forecasting
galeon archives: