TEPS: Targeted EPS for Norway

Principal Investigator

Dr Inger-Lise Frogner
The Norwegian Meteorological Institute
P O Box 43
Blindern, Oslo 3
Norway

inger.lise.frogner@met.no

Other researchers: Marit Helene Jensen, Trond Iversen

Project description

The requested computer time is for experimentation with targeted EPS – TEPS – for the European area. This project is related to the project spnoteps that currently is a special project at ECMWF. The title of the present project “NORLAMEPS: Limited Area Ensemble Prediction System for Norway” is no longer applicable, hence the request for a new special project to improve TEPS.

At the Norwegian meteorological Institute we have experience from running TEPS for some years, as TEPS is part of our regional ensemble prediction system NORLAMEPS. Although TEPS is running every day at ECMWF with good scores, there is a need to improve the system further. For details regarding the NORLAMEPS system, please see bottom of this page.

The computer time applied for in this special project does not cover the cost of running TEPS at a daily basis, for this the Norwegian quota is used. After the last update on the EPS system, running TEPS costs about 1.3 Million SBUs per year.

In addition to the daily running of the TEPS system, we will continue to develop the system. The computer time applied for is for such experiments.

We are planning to test:

• Diabatic singular vectors
• Hessian singular vectors
• a combination of targeted SVs and hemispheric SVs
• find the best combination of amplitudes from targeted and hemispheric SVs
• test different resolutions and optimization times for the targeted SVs.

This work also has the potential to improve EPS. From ECMWF Martin Leutbecher is involved in the development of TEPS through ideas for improving the system and also in offering some scientific and technical assistance.

NORLAMEPS

Since February 2005 NORLAMEPS has been running daily at the Norwegian Meteorological Institute (met.no). NORLAMEPS is a combination ensemble of two ensemble prediction systems: 1) TEPS; a special run with the ECMWF ensemble prediction system (running at ECMWF). It uses the same set-up and model version as the operational EPS, except for a targeting over Northern Europe and the bordering sea areas and a reduced number of ensemble members (20 + 1 members for TEPS). 2) LAMEPS; an ensemble prediction system run with the Norwegian version of the limited area model HIRLAM (running at met.no). The horizontal resolution is 12 km and it has 60 levels in the vertical. It runs up to 60 hours. TEPS is used for perturbing both the initial and lateral boundary conditions of LAMEPS, which gives a 20 + 1 LAMEPS ensemble. TEPS and LAMEPS are then combined to an ensemble consisting of 41 +1 ensemble members (NORLAMEPS).

NORLAMEPS is a supplement to ECMWF EPS as it is designed for optimal performance in our area of interest and because it is valid for day 1 to 3, LAMEPS also have a better resolution than EPS.

The system is crucially dependant on TEPS, both as boundary and initial conditions to LAMEPS and as a part of the combination ensemble NORLAMEPS.

Additional information

New Project for 2009

Would accept support for 1 year only, if necessary.