Comment on: Stratospheric Ozone Depletion at northern mid‐latitudes in the 21st century: The importance of future concentrations of greenhouse gases nitrous oxide and methane

[1] Recently, Randeniya et al. [2002] (hereafter R2002) presented predictions of the future evolution of stratospheric O3 over 2000–2100 based on two-dimensional (2D) model calculations. They investigated two IPCC scenarios covering high (A1F1) and low (B1) greenhouse gas (GHG) emissions, and a third based on A1F1 but with reduced CH4 emissions. Among their results, they concluded that stratospheric O3 may only partially recover towards pre-1980 values by around 2050 and argued that after this date it may again decline due to increases in stratospheric NOx produced from increasing N2O. However, R2002 did not include the coupled effects of GHG-induced temperature (T) changes on future O3 trends. They did estimate the effect of future CO2-induced stratospheric cooling on column O3 using results of a single GCM simulation based on an unspecified scenario. When these GCM T changes were imposed on their 2D model the increase in column O3 at northern mid-latitudes in 2100 was around 1%. [2] Future cooling from increases in GHGs (most importantly CO2) is a well-established feedback in the stratosphere [e.g. Haigh and Pyle, 1979]. We have performed similar 2D model calculations to R2002 but which include this feedback. This makes qualitative changes to the predicted future O3 and we would therefore like to expand on the conclusions of R2002 by discussing our coupled calculations. [3] We have used a global 2D dynamical-radiative model Kinnersley [1996] which extends from the ground to ∼80km with a resolution of 9.5° (horizontal) and ∼3.5 km (vertical). The model calculates its own temperatures using the model O3, CO2, CH4, N2O and NO2. The model uses the stratospheric chemical module from the SLIMCAT 3D model [Chipperfield, 1999]. (Note that the model does not contain detailed tropospheric chemistry such as non-CH4 hydrocarbons). Photochemical data is from Sander et al. [2000]. The model includes the effects of liquid aerosols in the lower stratosphere (LS) although in the runs presented here it did not include the effect of chlorine activation on cold solid/liquid particles as the model-calculated temperatures are not realistic enough to ensure a realistic timing/extent of polar stratospheric clouds (PSCs). [4] Note that in the polar LS cooling may enhance PSC-induced O3 loss, although stronger planetary wave activity may actually warm the Arctic vortex [e.g. Austin et al., 2002]. However, 3D models are required to capture the interaction of the polar vortex and PSC chemistry realistically. Such 3D models are computationally expensive while 2D models are much cheaper but can still capture many of the chemical-radiative feedbacks of the global stratosphere. [5] Six 2D model integrations were performed covering the period 1970–2100 (Table 1). These runs used either the A1F1 or B1 scenarios for GHGs. The model CO2 (only used in the T calculation) was either time-varying (from the appropriate IPCC scenario) or fixed at the year-2000 value. The halogen scenario was taken from WMO [1999] scenario A3. The aerosol loading was taken from WMO [1999] for the period 1979–1995. Before 1979 and after 1995 the aerosol loading was held constant at these initial/final values. [6] Figure 1 shows the evolution of column O3 at 35°N–60°N from the six model runs. Over the period 1980–2000 the model produces a downward trend in O3 and the enhanced loss after the eruption of Mt Pinatubo in June 1991. (Due the small modelled changes in the troposphere (see Figure 2), Figure 1 shows essentially only stratospheric O3 changes). The model underestimates the extent of the overall decrease and early 1990s' dip due to the neglect of some polar processing and the inability of the model to capture dynamical effects [see Hadjinicolaou et al., 1997]. However, the model performs reasonably over this period in line with other 2D models, though by 2000 our model indicates less column depletion than that of R2002. [7] After 2000, the model runs with fixed CO2 (and hence largely fixed T) show similar features to the calculations of R2002. Runs FA1, FB1 and FA1C all show an increase in O3 towards the middle of the century followed by a decrease again. No run shows column O3 returning to 1980 levels. Compared to R2002 our model shows only a slight enhancement of O3 loss in run FA1 compared to FB1 in the year 2100, though our A1F1 run with low CH4 (FA1C) also shows significantly more O3 loss. [8] The evolution of column O3 changes dramatically when the T feedback effects of CO2 are included. For all scenarios the O3 recovery is initially more rapid and reaches larger values. For the low-GHG scenario B1 (run VB1) column O3 returns to 1980 values while for the high-GHG scenario A1F1 (run VA1) column O3 is predicted to exceed 1980 values by over 3% towards the end of the century. Note that in the high GHG scenario the T effect of CO2 outweighs the chemical effect of increased NOx from N2O: the inclusion of the T feedback has reversed the order of which scenario gives the largest O3 loss. [9] The CO2-induced stratospheric cooling increases column O3 by reducing the rate of gas-phase loss processes. (Note that late in this century the stratospheric halogen loading will have returned to pre-1980 values and enhanced O3 loss in the LS due to cooling and more PSCs should not be an important effect.) Model profiles of O3, ΔO3 and ΔT are shown in Figure 2. The model runs with fixed CO2 also show a small cooling trend due to the increases in other GHGs, but the largest effect is from CO2. The changes in the mid stratosphere from 2000–2050 in the coupled run VA1 are similar to the 5K cooling reported by Rosenfield et al. [2002] and this increases to a 12 K cooling by 2100. The runs with CO2 feedback have more O3 throughout the mid-upper stratosphere in 2100. The fixed CO2 runs show a similar profile change from 2000–2100 as R2002, with a minimum near 10 hPa due to increase NOx-catalysed loss. The effect of the stratospheric cooling is to increase O3 throughout the mid-upper stratosphere in 2100 relative to 2000. [10] There is also a small feedback of the cooling on a reduced yield of NOx from N2O [Rosenfield and Douglass, 1998], which further reduces the impact of increased N2O. Figure 3 shows that by 2100 the high GHG scenario A1F1 has increased tropospheric N2O with respect to B1 by 22%. In the runs with fixed CO2 this increases mid-stratosphere NOy by 18%. However, when the effect of CO2-induced cooling is considered the increase in NOy is only around 12%. [11] In summary, for predictions of the future evolution of the O3 layer it is essential to include the well-known feedback of stratospheric cooling caused by increasing GHGs [see also Rosenfield et al., 2002]. Studies exploring different scenarios of chemically important species (e.g. CH4, N2O) also need to consider the possible future scenarios of CO2. CO2-induced cooling may cause stratospheric O3 to recover to values greater than 1980 levels during this century.