Long-Term Ozone Trends from Satellite and Ground-Based Data: General Approach, Results and Uncertainties.

Abstract:

Long-term trends are an important aspect of atmospheric variability. Examples are:

• Temperature trends indicating climatic changes
• Trace gas trends like increasing CO₂
• Stratospheric ozone trends (declining from the 1970s to the late 1990s, possibly increasing since about 2000) due to changes in atmospheric chlorine and bromine loading. 

Often these trends are small compared to other modes of variability. Multilinear regression is the standard method to account for the various modes of variability (e.g. annual cycle, quasi-biennial oscillation, solar cycle, meteorological influences) and to estimate the trends. Using ozone as an example, we will discuss this approach, its uncertainties and their estimation.

Another major aspect is availabilty of the necessary long-term data sets. Instrument lifetimes, as well as changes in instruments and methods usually mean that long-term records have to be constructed from multiple and inhomogeneous data sources. Again using ozone, we will discuss uncertainties arising from this, and explain a simple practical approach.

The Figure shows, as an example, anomalies (=deviation from average annual cycle) of ozone in the upper stratosphere as measured by various instruments at/near five stations of the Network for the Detection of Atmospheric Composition Change. Ozone at these altitudes decreases/ increases due to the long-term increase/decrease of effective stratospheric chlorine (ESC, inverted magenta line near the bottom). Ozone is also affected by the quasibiennial oscillation of stratospheric winds near the Equator (black line at the top), and by the 11-year cycle of solar activity (black line near the bottom). The grey range in the background shows results from chemistry climate models. The challenge of trend dectection is to separate the long-term trends (e.g. due to ESC) from the other variations and from the underlying uncertainties.