1 of 20

BASCOE Reanlysis of Aura MLS (BRAM): released version 2, unreleased version 3 and plans for version 4

Quentin Errera, Daniele Minganti, Simon Chabrillat, Marc Op de beeck and Sarah Vervalcke

1-Jul-2021

2 of 20

Outline

  • BRAM2
  • BASCOE
  • Who is BRAM3, how will be BRAM4
  • Lessons learned from BRAM2&3
  • Summary
  • Points for discussion

2

3 of 20

BRAM2 (BASCOE Reanalysis of Aura MLS, version 2)

  • A 15-year reanalysis (Sep 2004-Aug 2019) of stratospheric chemical composition based on the BASCOE assimilation of MLS v4 profiles of O3, H2O, HNO3, N2O, HCl, ClO, CO and CH3Cl
  • BRAM2 is done at 3.75°lon x 2.5° lat x 37 levels (0.1 hPa-surface)
  • BRAM2 ends in Sep 2019 as ERA-Interim which is used to drive BASCOE
  • BRAM2 is saved in the space of independent observations (ACE-FTS, MIPAS-IMK, O3sondes, SMILES ClO) using the BASCOE observation operator for a posteriori validation
  • BRAM2 produced by 4 experiments (streams) running in parallel with one month overlap

​

​

​

​

  • BRAM2 is stored in BIRA-IASB ftp server. Access is available after registration on strato.aeronomie.be (=>Datasets=>BRAM2), 28 registered users.
  • BRAM2 delivered products are 6 hourly analyses and ensemble spread of MLS assimilated species + Cl2O2 + temperature + surface pressure (285 Gb)
  • BRAM2 is published and characterized: Errera et al., ACP, 2019, doi.org/10.5194/acp-19-13647-2019

​

​

​

​

​

​

​

​

3

1015*Nobs

4 of 20

Belgian Assimilation System for Chemical ObsErvations (BASCOE)

  • A Chemistry Transport Model + a Data Assimilation system (see Errera et al., 2008, 2012, 2019; Skachko et al., 2014, 2016; Chabrillat et al., 2018)
  • Focus on the stratospheric composition
  • CTM:
    • Advection resolved by Lin and Rood (1996)
    • Includes ~60 chemical species and ~200 chemical reactions
  • DA methods:
    • 4D-Var (Errera et al., 2008, Errera and Ménard, 2016)
    • EnKF (Skachko et al., 2014, 2016, Errera et al., 2019)
  • Driven meteorology:
    • Flexible in choice, usually ERA-Interim or ERA5

​

​

5 of 20

BASCOE chemistry

  • BASCOE has a detailed representation of stratospheric chemistry (gas phase, photolysis, heterogeneous)
  • SF6 and its mesospheric loss have been recently included
  • Sulfate aerosols, used to compute heterogeneous reaction rates are taken from CMIP6 recommendations
  • Surface emissions are from CMIP6 recommendations
  • Currently, no UBC. For BRAM4, plan is to include CO&H2O from MLS, NO from MIPAS

5

CCl4, CFC11, CFC12, CFC113 , CFC114, CFC115, HA1301, HA1211, HCFC22

​

CH3Cl, CH3CCl3

​

CHBr3, CH2Br2, CH3Br

​

N2O, CH4, SF6, (CO2)

Tropospheric sources

O3, O, O1D

H2O, OH, H2, H2O2, H, HO2

CH3, CH3O, HCO, CH3OOH, CO, CH3O2, CH2O

BrO, BrONO2, BrCl, Br, Br2, HBr, HOBr

HNO3, NO, NO2, NO3, N2O5, HNO4, N

HCl, ClONO2, Cl2O2, Cl, Cl2, ClO, ClOO, OClO, ClNO2, HOCl

HF, SF6

Meso-Strato Prod and Loss

6 of 20

What is BRAM2, who is BRAM3, why BRAM4?

6

​

BRAM2

BRAM3

BRAM4

Period

Aug 2004 – Aug 2019

Aug 2004 – Dec 2023 (will be cont.)

Aug 2004-end of MLS

Chemistry

sb18

sb20c

sb23b (with SF6)

Spatial resolution

2.5° lat, 3.75° lon, 37 lev (96x73x37)

2° lat, 2.5° lon, 42 lev (144x91x42)

2° lat, 2.5° lon, 48 lev (144x91x48)

Dynamic

ERA-Interim at T31

ERA5 at T47

ERA5 at T47

Surface boundary conditions

None

CMIP6 (Meinshausen et al., 2016)

Up: None

CMIP6

NO from MIPAS, CO&H2O from MLS

Upper boundary conditions

None

None

NO from MIPAS, CO&H2O from MLS

Strato. sulfate aerosols

v4q30 (i.e. Hitchman et al., 1994, divided by 5)

CMIP6

TBD

7 of 20

BRAM vertical resolutions

  • BRAM4 vertical resolution: Likely 48 levels out of ERA5

7

8 of 20

BRAM vertical resolutions

8

Vertical resolution defined by 20 years (2-years perpetual) BASCOE Age-of-Air simulation

​

AoA shown here in July (monthly zonal average) of the last year of simulation

100 hPa

10 hPa

1 hPa

0.1 hPa

9 of 20

DA & obs setup, compared to BRAM2

9

​

BRAM2

BRAM3

BRAM4

Version

07.00.00

09.00.00

10.xx.xx

EnKF vertical localization cutoff

1.55 level

3 km except for H2O (1.55 km)

as BRAM2

MLS version

v4.2

v4.2

v5.0

MLS N2O

MLS 190 GHz (std prod)

MLS 640 GHz (up to Aug 2013 only)

TBD

MLS CH3Cl

minval=1e-18

minval=-1e-9 allowing negative value assimilation

TBD

MLS obs quality flag

0, 18 (low clouds), 34 (high clouds)

0, 18 (low clouds), 34 (high clouds), 50 (low and high clouds)

As BRAM3

Use of AK

No

No

TBD

MIPAS SF6

No

No

TBD

Indep. obs.

ACE-FTS, MIPAS, O3Sondes, SMILES

As BRAM2 with version updates

As BRAM31 + SAGE-III/ISS and OMPS-LP

1 With latest available version

10 of 20

Outputs, compared to BRAM2

10

​

BRAM2

BRAM3

BRAM4 (& CTRL4BRAM4)

Format

HDF/GDFB (but distributed as NetCDF-CF)

NetCDF-CF

NetCDF-CF

Products for assim species1

6 hourly analysis and ensemble spread

N/A

as BRAM2 + daily mean & spread + daily mean increment

Products for non-assimilated species2

Cl2O2 only

N/A

6 hourly snapshot and daily mean

1 O3, H2O, N2O, HNO3, HCl, CH3Cl, ClO, CO

2 TBD, CH4, NOy species, Cly species, SF6, CFCs

11 of 20

Control run

  • CTRL4BRAM4a is also a model control run in preparation for BRAM4
    • It uses a corrected STS sticking coefficient for BrONO2+HCl(c) which has a strong impact on BrONO2 and ClONO2 (see later)
  • BRAM4 will be provided with its control run (same model, no DA)
    • BRAM4 will be run with the latest version of BASCOE CTM including SF6
    • MIPAS SF6 assimilation will be tested and implemented if valuable
    • BASCOE under development for new photolysis computation which will be used if timely available (end of 2024)

11

12 of 20

Some lessons learned from BRAM2, BRAM3 and CTRL4BRAM4a

Comparison of BRAM2, BRAM3 and CTRL4BRAM4a vs ACE-FTS v4.1

  • NH Stratospheric Mid-Latitudes
  • SH upper stratospheric Polar Vortex

​

​

  • SH lower stratospheric Polar Vortex and Tropic => see backup slides

​

​

12

13 of 20

NH Stratospheric Mid-Latitudes

Period: Jan 2005-Aug 2019 (Jul 2013 for N2O)

Latitudes: 30°N-60°N

Observations: ACE-FTS v4.1

Values are:

(Model-Obs)/Mean(Mod)

Color code:

Observations, BRAM2, BRAM3, CTRL4BRAM4a

For relevant assim species

In the lower strato, BRAM2 is better than BRAM3 => because of >cutoff in vertical correlation?

Do we need to assim N2O below 10 hPa and CH3Cl? CTRL better than DA

Note the ≠ x-range in plots

14 of 20

NH Stratospheric Mid-Latitudes

Period: Jan 2005-Aug 2019 Latitudes: 30°N-60°N

Observations: ACE-FTS 4.1

Values are:

(Model-Obs)/Mean(Obs)

Color code:

Observations, BRAM2, BRAM3, CTRL4BRAM4a

Very good CH4, lower bias with BRAM2 (?), higher correl with BRAM3 and CTRL.

For relevant non-assim species

BRAM2 ClONO2 much better than BRAM3 (better het. chem.).

CTRL better than BRAM2

Excess of NOx. JNO2?

15 of 20

Upper Stratospheric SH Polar Vortex

15

Color code:

Observations, BRAM2, BRAM3, CTRL4BRAM4a

For relevant species

Enhanced NO due to mesospheric descend not in BASCOE

Period: Jul-Aug-Sep-Oct 2005-2019

Latitudes: 90°S-60°S

Observations: ACE-FTS v4.1

Values are:

(Model-Obs)/Mean(Obs)

16 of 20

Summary

  • BRAM3 not so better than BRAM2 despites better meteorology and higher resolution
    • vcutoff should be updated/back to BRAM2 setup
    • NOx still too high => hope to be improved with updated photolysis
  • CTRL is not so bad, do we really need assimilation of N2O and CH3Cl?
  • BRAM4:
    • Increasing spatial resolution and period using ERA5
    • Update chemistry and DA setup
    • UBC of CO, NO and H2O
    • Assim. N2O only above 10 hPa? No assim for CH3Cl?
    • Provide its control run

​

16

17 of 20

Points for discussion

  • Would be great to have a common repository to store BRAM2/4 with other A-RIP reanalyses
  • What are the user needs?
    • BRAM2 has 28 registered users but not yet any publication using BRAM2. How to get user communities (i.e. A-PARC activities) engaged?
  • When is BRAM4 expected to be available for A-RIP?

17

18 of 20

​

​

Backup slide

18

19 of 20

Southern Polar Vortex

19

Period: Jul-Aug-Sep-Oct 2005-2019 (2013 for N2O)

Eq. Latitudes: 90°S-75°S

Observations: ACE-FTS v4.1

Values are:

(Model-Obs)/Mean(Obs)

Color code:

Observations, BRAM2, BRAM3, CTRL4BRAM4a

Do we really need to assim. N2O?

For relevant assim species

K

K

K

20 of 20

Tropics

20

Period: Jul-Aug-Sep-Oct 2005-2019 (2013 for N2O)

Latitudes: 30°S-30°N

Observations: ACE-FTS v3.6

Values are:

(Model-Obs)/Mean(Obs)

Color code:

Observations, BRAM2, BRAM3, CTRL4BRAM4a

For relevant species

Ozone deficit in BASCOE (like in all CTMs/CCMs), long-standing issues

  • Here, BRAM2 ≠ BRAM3 because T° ≠ ERA-I vs ERA5
  • Deficit of O3 linked to excess of NOx