BASCOE NRT implementation

You are here

BIRA-IASB is in charge of the monitoring and evaluation of the stratospheric composition products delivered by CAMS and its pre-operational predecessor MACC. In this context, the BASCOE assimilation system was setup to deliver near real-time analyses and forecasts of ozone and related species for the stratosphere.

The validation of BASCOE in the frame of MACC-III is documented in the BASCOE Validation Report (Deliverable D38.7).

From 2009-07-01 until 2012-10-20 the BASCOE system run in NRT for MACC was completely stable (i.e. version q02 assimilating Aura-MLS v2.2 offline dataset). Lefever et al. (ACP, 2015) described in detail this version and its NRT configuration. They also compared it with independent observations and with the other systems used for the MACC stratospheric ozone service. The version q02 was based on work realized for the past GSE-PROMOTE program of ESA. These original versions are described and validated by Errera et al. (ACP, 2008) and Viscardy et al. (J-STARS, 2010).

This MACC-III deliverable documents the performance of the system (version04.03 at the end of the MACC-III phase) through comparison with assimilated observations and provides all technical details and bibliographical references.

Current configuration

The BASCOE data assimilation system assimilates the offline dataset (level-2) retrieved (v4.2) from the Aura-MLS instrument. The offline dataset is delivered a few days later than the NRT stream but it includes more species.

  • The following species are assimilated: O3, ClO, H2O, HNO3, HCl, CO and N2O .
  • The assimilation window is 24 hours.
  • Processing lags by typically 4 days, mainly due to latency for arrival of the Aura-MLS offline dataset.
  • BASCOE version is 05.07.
  • Global horizontal grid with a 3.75° longitude by 2.5° latitude resolution (nlon=96, nlat=73).
  • Winds, temperature and surface pressure are interpolated in the ECMWF Atmospheric Model high resolution 6-hourly analyses.
  • Vertical grid is of hybrid-pressure type, comprising 86 levels and extending from 0.01 hPa to the surface.
  • Timesteps of 20 minutes, output every 3 hours.
  • After an analysis, a chemical forecast of 10 days is run using winds, temperature and surface pressure interpolated in the ECMWF Atmospheric Model high resolution 10-day forecasts.

One should use only the values at stratospheric gridpoints, i.e. approx the 70 first levels (more precisely, those gridpoints where p < p_tropop where p, p_tropop are explained here).

Note: when Aura-MLS observations are not available to BASCOE within 4 days, it falls back to CTM mode for a 1-day integration. Such episodes with no assimilation are not reported here, unless they last for more than 30 consecutive days. To identify the exact dates in CTM mode, look this list of logfiles from the model run.


Valid date


2016-07-01 Preparation of Aura-MLS offline v4.2 has been updated to v04q (following more closely the AURA-MLS recommendations for observation sceeening)
2016-03-01 Major upgrade: switching from version 04.03 to 05.07; preparation of Aura-MLS offline v4.2 has been updated to v04p; a new vertical grid with 86 levels has been adopted
2015-07-01 Assimilated dataset switches from Aura-MLS offline v3.4 to v4.2 (using preparation v04b).
2014-02-01 Major upgrade: switching from version 03.11 to 04.03.
2013-08-19 ClO is included in the list of species actively assimilated by BASCOE. ClO is part of the Aura-MLS offline dataset v3.3 but was pure model output until now.
2013-06-25 version 03.11 : Vertical grid upgraded from 91 to 137 hybrid-pressure levels, extending from 0.01 hPa to the surface, following upgrade at ECMWF.
2013-05-01 Assimilated dataset switches from Aura-MLS offline v3.3 to v3.4.
2013-01-12 version 03.09 : H2O replaced by ECMWF meteo analysis between surface and 2 km above tropopause.
2013-01-01 Major upgrade: switching from version q02.05 to 03.08.
2012-10-20 Assimilated dataset switches from Aura-MLS offline v2.2 to v3.3. See the EOS MLS Data Documentation.
2011-07-09 Assimilation not used poleward of 85°. This inhibits a bug which could be seen in ozone analyses at 50 hPa (polar regions), e.g. after 2011-06-20.
2011-05-18 Bug fixed in pre-processing of ECMWF wind fields. The impact of this bug could be seen in ozone analyses at 50 hPa since start of operations ("checkerboard pattern").
2009-07-01 Operations start. BASCOE version: BASCOP q02.04.00.