ROSE  3D Chemistry-Transport Model

DLR Version - Model Description

 

Model resolution  
Horizontal: Euler grid with 5° x 5.6° resolution (latitude/longitude)
Vertical: 37 levels between 316 and 0.316 hPa (8-55 km) giving 1.3 km spacing
Temporal: 15 min. for chemistry; 1.5 hours for transport
Output resolution: 1h for validation purposes; 00, 06, 12, 18 UTC in operational use.

Transport scheme

Semi-Lagrange transport driven by daily UKMO wind fields (Smolarkiewicz, 1995).

Chemistry

Gas-phase chemistry is based on the family-concept as formulated by Rose and Brasseur, 1989 (revised 1996). Reaction rates are calculated using daily temperature fields from UKMO and pre-calculated photolysis frequencies depending on local insolation and ozone column. Differential equations are solved by a semi-implicit scheme.

Transported species Equilibrium species
Ox HO2NO2
(O(1D)+O(3P)+O3) CO
H2O2 BrOx(Br+BrO)
H2O² HBr
NOx (NO+NO2) HOBr
NOy (N + NO) BrONO2
N2O H2SO4³
N2O5
HNO3²
HNO4
ClOx (Cl+ClO)
Cl2O2
OClO
Cl2²
ClNO2²
ClONO2
HCl
HOCl²
CH4
O(1D), O(3P)
O3
O2*
H
OH
HO2
NO
N
NO2
NO3
N2*
Cl
ClO
Br
BrO
BrCl
H2*
CFC-11, CFC-12

² PSCs: Chipperfield, 1994
³ Aerosols

*constant 2-D field

Assimilation scheme

Sequential assimilation of level 2 trace gas concentrations  at the moment only GOME ozone columns planned:
ENVISAT trace gas columns and vertical profiles.

Differences between ROSE/DLR and NCAR-ROSE

GOME ozone columns are assimilated sequentially using optimal interpolation to model-grid into the ROSE/DLR version:

A factor between the internal model ozone column and the incoming GOME total ozone measurement is derived and applied as a constant factor to the modelled ozone concentration at all vertical levels. 

ROSE/DLR version 2.4 contains an empirical correction to cope with observed altitude dependent systematic deviations between model and measurements. ROSE/DLR ozone profiles (with assimilated GOME measurements) vary systematically if compared with satellite and sonde measurements (see validation page).
As a first, pragmatic approach to correct this model behaviour SAGE II measurements are used. A height and latitude dependent factor between ROSE/DLR ozone profiles and SAGE II measurements is calculated for four seasons. This factor is applied to the modeled ozone profiles as a last step before the data output.  During that 'correction' the total ozone column amount is kept constant to avoid inconsistencies in total column ozone. Comparison with CRISTA measurements show significant improvements if the 'SAGE correction factor' is applied.
Explanations for this model behaviour are under examination.

The horizontal resolution  is set to a fixed value. Modifications are necessary for an operational use in the DLR-DFD environment. A  more flexible controlling of model runs through parameter files and a different output scheme is necessary.

NCAR-ROSE

Information about the original NCAR-ROSE model version can be found on the NCAR website http://www.ncar.ucar.edu/ncar.

Papers on ROSE

Brasseur G., K. Rose, C. M. Smythe, The behavior of ozone and other trace gases in the statosphere during dynamically perturbed situations: A three-dimensional model, Ozone in the Atmosphere, R. D. Bojkov, P. Fabian (Eds.), DEEPAK Publishing, 1989, ISBN 0-937194-15-8

Riese, M., X. Tie, G. Brasseur and D. Offermann, Three-dimensional simulation of stratospheric trace gas distributions measured by CRISTA, J. Geophys. Res., 104, 16419-16435, 1999

Smolarkiewicz, P.K. and P.J. Rasch, Monotone advection on the sphere: An Eulerian versus Semi-Lagrangian approach, J. Geophys. Res., 100, 1253-1269, 1995

Please send questions, comments, ideas or criticism to Frank Baier (DLR-DFD).