Beta

Evaluation of MACC Reanalysis: comparison with ACE-FTS observations

You are here

Inness et al. (ACP, 2013) described in detail this reanalysis, including validation against independent observations. The validation of stratospheric ozone with independent satellite observations (section 3.2.1, figures 13 and 14) uses ACE-FTS. Here is an extensive view of this comparison.

ACE-FTS data were available from January 2004 to September 2010, with a gap from December 2009 to May 2010.

Monthly mean relative bias between MACC reanalysis and ACE-FTS

The following figures show time series of the monthly mean relative biases between the MACC reanalysis and ACE-FTS, at 3 selected pressure levels (10, 46.42, and 100 hPa) for extrapolar (left) and polar (right) regions.

Click on the figures to enlarge them.

Extrapolar regions Polar regions
at 10 hPa
at 46.42 hPa
at 100 hPa

Discussion

The MACC reanalysis generally has a small positive bias.

  • 10 hPa:
    At 10hPa, we clearly see the positive impact of the bias correction performed in January 2008 in the polar and even, to a lesser extent, in the midlatitude regions. Biases at the poles are going down from about 25 to 5 %.

  • 46.42 Pa:
    Biases are relatively constant, mostly less than 10% in extrapolar regions and less than 5% in polar regions.

  • 100 hPa:
    Biases are less than 10% in most regions. Larger biases are observed for the period April-August 2004, when no O3 profile data were assimilated in the reanalysis (gap between MIPAS and MLS profiles), and for the post-March 2009 period, when the reanalysis started to assimilate NRT MLS data for which ozone values below 64 hPa are not reliable and can therefore not be used.

Monthly mean standard deviations between MACC reanalysis and ACE-FTS

The following figures show time series of the monthly mean standard deviations between the MACC reanalysis and ACE-FTS, at 3 selected pressure levels (10, 46.42, and 100 hPa) for extrapolar (left) and polar (right) regions.

Click on the figures to enlarge them.

Extrapolar regions Polar regions
at 10 hPa
at 46.42 hPa
at 100 hPa

Discussion

The monthly mean standard deviations between MACC and ACE-FTS ozone are generally between 5 and 10% at 10 and 46.42 hPa, but twice as large at 100 hPa. Standard deviations are largest above the South Pole region, during the Antarctic ozone hole season (especially September months).

Seasonal mean relative bias and standard deviation between MACC reanalysis and ACE-FTS

The following figures show profiles of the seasonal mean relative biases between the MACC reanalysis and ACE-FTS (with months DJF=winter, MAM=spring, JJA=summer, and SON=autumn).

Click on the figures to enlarge them.

Mean vmr Mean bias Mean standard deviation
DJF
MAM
JJA
SON

Discussion

Biases are mainly positive between 3 and 100 hPa, and negative outside this pressure region (i.e. above 3 hPa and below 100 hPa). Biases are largest around 10 hPa (up to 15%) and smallest between 20 and 50 hPa (less than 5% for most regions).

Evaluation against HALOE

Bias and standard deviation during mid-ozone hole season 2003

Bias and standard deviation of the MACC IFS-MOZ reanalysis run fbov and the BASCOE reanalysis MIP_v4q30a w.r.t. HALOE satellite data, for the period 15 September - 15 October 2003, for 5 latitude bands.

The above figure shows the global results for the bias and standard deviation of the IFS-MOZART (fbov) and BASCOE O3 reanalyses versus HALOE satellite profiles for the Antarctic ozone hole period 2003. The agreement is generally quite good outside the South Pole region. The larger bias around 10 hPa at the South Pole shows that the coupled system has problems in predicting the ozone hole.