Speaker
Report on the experience (or the proposed activity). It would be very important to mention key services which are essential for the success of your activity on the EGEE infrastructure.
Two techniques for generating an ocean ensembles were
successfully tested on the Grid.
The first approach is based on a random perturbation of the
initial conditions of
the ocean forecast. This is highly efficient because all the
members are initialized
using the same input files, only the seed numbers needed to be
transferred to the
Grid Working Nodes. The results show that a minimum number of 450
members were
successfully accomplished in 5 hours. This result was achieved
using 15 Computing
Elements and imposing a simple requirement policy to the Grid
software manager.
The second approach relies on the perturbation of the wind
forcing. All the ocean
members start from the same initial condition but are driven by
different wind
realizations. This significantly increase the amount of data that
must be transferred
through the network. Preliminary results showed that the Grid
system was able to
sustain this working load for a 100 members ensemble set up.
With a forward look to future evolution, discuss the issues you have encountered (or that you expect) in using the EGEE infrastructure. Wherever possible, point out the experience limitations (both in terms of existing services or missing functionality)
So far every member of the ocean ensemble was run on a single
Working Node. This
represents a limitation to the size of the problem that can be
treated.
While a natural evolution to this problem will be apply a Message
Passage Interface a
second strategy will also be tested. Since many events that would
benefit from an
ocean ensemble forecast are localized in a small portions of the
model domain, an
high-resolution limited-domain ocean model might be the best
compromise for the Grid
environment.
Describe the scientific/technical community and the scientific/technical activity using (planning to use) the EGEE infrastructure. A high-level description is needed (neither a detailed specialist report nor a list of references).
The Istituto Nazionale Geofisica e Vulcanologia runs an
operational oceanographic
system called Mediterranean Forecasting System.
The long term goal of this project is to collect high quality
data and provide
reliable environmental predictions for the Mediterranean Sea. End
user applications
of ocean forecast span from contaminant drift predictions to
search and rescue
operations.
An ensemble forecast system is currently under investigation to
provide estimates of
forecast uncertainties.
Describe the added value of the Grid for the scientific/technical activity you (plan to) do on the Grid. This should include the scale of the activity and of the potential user community and the relevance for other scientific or business applications
The subject of this study is the understanding and the
demonstration of the usage of
the Grid system in ensemble ocean forecasting.
Short term ocean forecast are externally driven by atmospheric
forecasts. The MFS
system collects ECMWF atmospheric forecast daily. To be effective
an ocean ensemble
forecast must be run within the time constraints imposed by the
availability of the
ECMWF products. To test the potential of the Grid for this
purpose, the time window
allowed for a 10 days ocean forecast with an ensemble of 1000
forecasts was set to 6
hours.
This experiment was run in collaboration with the Istituto
Nazionale di Fisica
Nucleare on the INFN Grid. This work demonstrated that an
extremely large ocean
ensemble forecast, that would be unfeasible on most of the other
computer
infrastructures, could be run on a Grid system under operational
forecasting working
conditions and normal Grid configuration.
