Climate Change 2001:
Working Group I: The Scientific Basis
Other reports in this collection

6.4 Stratospheric Ozone

6.4.1 Introduction

The observed stratospheric O3 losses over the past two decades have caused a negative forcing of the surface-troposphere system (IPCC, 1992, 1994; SAR). In general, the sign and magnitude of the forcing due to stratospheric O3 loss are governed by the vertical profile of the O3 loss from the lower through to the upper stratosphere (WMO, 1999). Ozone depletion in the lower stratosphere, which occurs mainly in the mid- to high latitudes is the principal component of the forcing. It causes an increase in the solar forcing of the surface-troposphere system. However, the long-wave effects consist of a reduction of the emission from the stratosphere to the troposphere. This comes about due to the O3 loss, coupled with a cooling of the stratospheric temperatures in the stratospheric adjustment process, with a colder stratosphere emitting less radiation. The long-wave effects, after adjustment of the stratospheric temperatures to the imposed perturbation, overwhelm the solar effect i.e., the negative long-wave forcing prevails over the positive solar to lead to a net negative radiative forcing of the surface-troposphere system (IPCC, 1992). The magnitude of the forcing is dependent on the loss in the lower stratosphere, with the estimates subject to some uncertainties in view of the fact that detailed observations on the vertical profile in this region of the atmosphere are difficult to obtain.

Typically, model-based estimates involve a local (i.e., over the grid box of the model) adjustment of the stratosphere (Section 6.1) assuming the dynamical heating to be fixed (FDH approximation; see also Appendix 6.1). An improved version of this scheme is the so-called seasonally evolving fixed dynamical heating (SEFDH; Forster et al., 1997; Kiehl et al., 1999). The adjustment of the stratosphere to a new thermal equilibrium state is a critical element for estimating the sign and magnitude of the forcing due to stratospheric O3 loss (WMO, 1992, 1995). While the computational procedures are well established for the FDH and SEFDH approximations in the context of the surface-troposphere forcing, one test of the approximations lies in the comparison of the computed with observed temperature changes, since it is this factor that plays a large role in the estimate of the forcing. While the temperature changes going into the determination of the forcing are broadly consistent with the observations, there are challenges in comparing quantitatively the actual temperature changes (which undoubtedly are affected by other influences and may even contain feedbacks due to O3 and other forcings) with the FDH or SEFDH model simulations (which necessarily do not contain feedback effects other than the strato-spheric temperature response due to the essentially radiative adjustment process).

We reiterate both the concept of the forcing for stratospheric O3 changes and the fact that this has led to a negative radiative forcing since the late 1970s. Further, the model-based estimates that necessarily rely on satellite observations of O3 losses are likely the most reliable means to derive the forcing, notwithstanding the uncertainty in the vertical profile of loss in the vicinity of the tropopause. Since several model estimates have employed the Total Ozone Mapping Spectrometer (TOMS) observations as one of the inputs for the calculations, there is the likelihood of a small tropospheric O3 change component contaminating the stratospheric O3 loss amounts, especially for the lowermost regions of the lower stratosphere (Hansen et al., 1997a; Shine and Forster, 1999). Both the estimates derived in the earlier IPCC assessments and the studies since the SAR show that the forcing pattern increases from the mid- to high latitudes consistent with the O3 loss amounts. Seasonally, the winter/springtime forcings are the largest, again consistent with the temporal nature of the observed O3 depletion.

It is logical to enquire into the realism of the computed coolings with the available observations using models more realistic than FDH/SEFDH, namely GCMs. Furthermore, comparison of the FDH and SEFDH derived temperature changes with those from a GCM constitutes another test of the approximations. WMO (1999) concluded, on the basis of intercomparisons of the temperature records as measured by different instruments, that there has been a distinct cooling of the global mean temperature of the lower stratosphere over the past two decades, with a value of about 0.5°C/decade. Model simulations from GCMs using the observed O3 losses yield global mean temperature changes that are approximately consistent with the observations. Such a cooling is also much larger than that due to the well-mixed greenhouse gases taken together over the same time period. Although the possibility of other trace species also contributing to this cooling cannot be ruled out, the consistency between observations and model simulations enhances the general principle of an O3-induced cooling of the lower stratosphere, and thus the negativity of the radiative forcing due to the O3 loss. Going from global, annual mean to zonal, seasonal mean changes in the lower stratosphere, the agreement between models and observations tends to be less strong than for the global mean values, but the suggestion of an O3-induced signal exists. Note though that water vapour changes could also be contributing (see Section 6.6.4; Forster and Shine, 1999), complicating the quantitative attribution of the cooling solely due to O3. As far as the FDH models that have been employed to derive the forcing are concerned, their temperature changes are broadly consistent with the GCMs and the observed cooling. However, the mid- to high latitude cooling in FDH tends to be stronger than in the GCMs and is more than that observed. The SEFDH approximation tends to do better than the FDH calculation when compared against observations (Forster et al., 1997).

Other reports in this collection