Climate Data

Stratospheric Ozone: Randel & Wu 2007

This is a monthly, zonal-mean ozone profile data set spanning 1979–2005 for altitudes from the tropopause up to 50 km. Interannual variability is based on a combination of SAGE (Stratospheric Aerosol and Gas Experiment) satellite profile measurements and polar ozonesondes.  The interannual anomalies were added to a climatology from Fortuin and Kelder (1998). This dataset was used as boundary conditions (forcing) for many CMIP3 historical experiments. For CMIP5, the ACC/SPARC project formed a global ozone data set by combining a similar data set for the historical stratosphere with other measurements and modeling to include the troposphere and extend the data set to 2099. Both Randel and Wu (2007) and the ACC/SPARC data are considered to have relatively conservative trends in lower stratospheric ozone over 1979-2005 (see Hassler et al., 2013 and Solomon et al., 2012).

Key Strengths:

  • SAGE solar occultation data are stable and have high vertical resolution
  • Low-frequency variations due to trends, solar cycle and QBO are well-characterized by SAGE measurements
  • Estimated ozone variability and trends over midlatitudes are in good agreement with ozonesonde data.

Key Limitations:

  • Gaps in SAGE coverage at the time of El Chichon (1982) and Pinatubo (1991) eruptions means that ozone anomalies associated with these events are not included in the data
  • Less SAGE coverage in the polar regions. While this was mitigated by inclusion of Resolute and Syowa radiosonde data for the NH and SH, respectively, the Antarctic ozone depletion may be underestimated

Expert Developer Guidance

#The following was contributed by Bill Randel (NCAR), February, 2013:

The Randel and Wu (RW) stratospheric ozone data set was constructed by combining interannual ozone anomalies (derived from SAGE I+II satellite measurements and polar ozonesondes) with a background ozone climatology from Fortuin and Kelder (1998), to give a monthly global ozone data set covering 1979-2005. Description of the data and comparison of the interannual anomalies with independently measured column ozone variability is described in Randel and Wu (2007). The interannual anomalies are constructed using a multivariate regression fit to the gappy satellite measurements.  The data are intended to accurately capture seasonal and interannual variability associated with long-term decadal trends (including development of the Antarctic ozone hole), along with the 11-year solar cycle and stratospheric quasi-biennial oscillation (QBO).  

Strengths: Interannual changes are based on high quality, high vertical resolution satellite measurements from the SAGE I (covering 1979-1982) and SAGE II (1984-2005) satellites. These solar occultation measurements are the acknowledged ‘gold standard’ for quantifying stratospheric ozone changes during 1979-2005. Because SAGE measurements do not cover polar regions throughout the year, polar ozone is based on measurements from two high quality stations with near-continuous observations during 1979-2005 (Syowa at 69o S, and Resolute at 75o N).

Weaknesses: The data are focused on low-frequency interannual variability that is well-characterized by SAGE measurements (trends, solar cycle and QBO fits). The effects of the large volcanic eruptions of El Chichon (1982) and Pinatubo (1991) on stratospheric ozone are not included, as there were no observations during these periods (El Chichon occurred during the break between SAGE I and SAGE II, and SAGE II measurements after Pinatubo were unavailable due to excessive aerosol contamination). Also, month-to-month ‘natural’ variability of ozone is not included in these data.

Applications: The seamless global data are useful for forcing model simulations and characterizing low frequency variability in free-running models.

Common mistakes for users: unknown

Spurious features:  none

Corrections applied: no

Conversions for comparisons to model output: no

Comparable data sets: The historical, stratospheric component of the ACC/SPARC ozone data set (Cionni et al, 2011) is based on a data set very similar (but not identical) to RW.

Uncertainty characterized: Not explicitly. Comparisons to the global column ozone measurements (shown in RW) suggest an overall reasonable depiction of large-scale ozone variability and trends during 1979-2005.

Summary: The RW data set was intended to provide a simple global compilation of ozone variability and trends during 1979-2005 for use in forcing or evaluating global models. It has been highly utilized in the community.##

Years of Record

1979/01 to 2005/12
temporal metadataID:




Data Time Period Extended?

no, data set not being extended


Spatial Resolution

5° lat zonal mean x 1Km vertical up to 50Km

Ocean or Land


Missing Data Flag

spatially complete

Vertical Levels

Input Data

SAGE profiles and polar ozonesondes

Earth system components and main variables

Data Access: Please Cite data sources, following the data providers' instructions.

Key Figures

Click the thumbnails to view larger sizes



Climatological range of monthly mean O3 (Randel & Wu, 2007).
Monthly ozone climatology (1979-2005).

Cite this page

Randel, Bill & National Center for Atmospheric Research Staff (Eds). Last modified 20 Aug 2013. "The Climate Data Guide: Stratospheric Ozone: Randel & Wu 2007." Retrieved from

Acknowledgement of any material taken from this page is appreciated. On behalf of experts who have contributed data, advice, and/or figures, please cite their work as well.