7.4.5.2 Consequences for Lifetimes

7.4.5.2.1 Lifetime definition

The global instantaneous atmospheric lifetime of a trace gas in the atmosphere is obtained by integrating the loss frequency l over the atmospheric domain considered. The integral must be weighted by the distribution of the trace gas on which the sink processes act. Considering a distribution of the trace gas C(x,y,z,t), a global instantaneous lifetime derived from the budget can be defined as:

τ_global = ∫ C dv / ∫ C l dv (7.4)

where dv is an atmospheric volume element. This expression can be averaged over one year to determine the global and annual mean lifetime. The global atmospheric lifetime (also called ‘burden lifetime’ or ‘turnover lifetime’) characterises the time required to turn over the global atmospheric burden.

The global atmospheric lifetime characterises the time to achieve an e-fold decrease of the global atmospheric burden. Unfortunately τ_global is a constant only in very limited circumstances. In the case where the loss rate depends on the burden, the perturbation or pulse decay lifetime (τ_pert) is introduced (see Velders et al., 2005). The perturbation lifetime is used to determine how a one-time pulse emission may decay as a function of time as needed for the calculation of Global Warming Potentials (GWPs). The perturbation lifetime can be distinctly different from the global atmospheric lifetime. For example, if the CH₄ abundance increases above its present-day value due to a one-time emission, the time it takes for CH₄ to decay back to its background value is longer than its global unperturbed atmospheric lifetime. This delay occurs because the added CH₄ will cause a suppression of OH, in turn increasing the background CH₄. Such feedbacks cause the decay time of a perturbation (τ_pert) to differ from the global atmospheric lifetime (τ_global). In the limit of small perturbations, the relation between the perturbation lifetime of a gas and its global atmospheric lifetime can be derived from a simple budget relationship as τ_pert = τ_global / (1– f), where the sensitivity coefficient f = dln(τ_global) / dln(B). Prather et al. (2001) estimated the feedback of CH₄ to tropospheric OH and its lifetime and determined a sensitivity coefficient f = 0.28, giving a ratio τ_pert / τ_global of 1.4. Stevenson et al. (2006), from 25 CTMs, calculate an ensemble mean and 1 standard deviation uncertainty in present-day CH₄ global lifetime τ_global of 8.7 ± 1.3 years, which is the AR4 updated value. The corresponding perturbation lifetime that should be used in the GWP calculation is 12 ± 1.8 years.

Perturbation lifetimes can be estimated from global models by simulating the injection of a pulse of gas and tracking the decay of the added amount. The pulse of added CO, HCFCs or hydrocarbons, by causing the concentration of OH to decrease and thus the lifetime of CH₄ to increase temporarily, causes a buildup of CH₄ while the added burden of the gas persists. Thus, changes in the emissions of short-lived gases can generate long-lived perturbations as shown in global models (Derwent et al., 2001; Wild et al., 2001; Collins et al., 2002). Changes in tropospheric ozone accompany the CH₄ decay on a 12-year time scale as an inherent component of this mode, a key example of chemical coupling in the troposphere. Any chemically reactive gas, whether a greenhouse gas or not, will produce some level of indirect greenhouse effect through its impact on atmospheric chemistry.

7.4.5.2.2 Changes in lifetime

Since OH is the primary oxidant in the atmosphere of many greenhouse gases including CH₄ and hydrogenated halogen species, changes in OH will directly affect their lifetime in the atmosphere and hence their impact on the climate system. Recent studies show that interannual variations in the chemical removal of CH₄ by OH have an important impact on the variability of the CH₄ growth rate (Johnson et al., 2002; Warwick et al., 2002; J. Wang et al., 2004). Variations in CH₄ oxidation by OH contribute to a significant fraction of the observed variations in the annual accumulation rate of CH₄ in the atmosphere. In particular, the 1992 to 1993 anomaly in the CH₄ growth rate can be explained by fluctuations in OH and wetland emissions after the eruption of Mt. Pinatubo (J. Wang et al., 2004). CH₄ variability simulated by Johnson et al. (2002), resulting only from OH sink processes, also indicates that the ENSO cycle is the largest component of that variability. These findings are consistent with the variability of global OH reconstructed by Prinn et al. (2005), Manning and Keeling (2006) and Bousquet et al. (2005), which is strongly affected by large-scale wildfires as in 1997 to 1998, by El Niño events and by the Mt. Pinatubo eruption.

The effect of climate change on tropospheric chemistry has been investigated in several studies. In most cases, the future CH₄ lifetime increases when emissions increase and climate change is ignored (Brasseur et al., 1998; Stevenson et al., 2000; Hauglustaine and Brasseur, 2001; Prather et al., 2001; Hauglustaine et al., 2005). This reflects the fact that increased levels of CH₄ and CO depress OH, reducing the CH₄ sink. However, climate warming increases the temperature-dependent CH₄ oxidation rate coefficient (Johnson et al., 1999), and increases in water vapour and NO_x concentrations tend to increase OH. In most cases, these effects partly offset or exceed the CH₄ lifetime increase due to emissions. As a consequence, the future CH₄ lifetime calculated by Brasseur et al. (1998), Stevenson et al. (2000) and Hauglustaine et al. (2005) remains relatively constant (within a few percent) over the 21st century. In their transient simulation over the period 1990 to 2100, Johnson et al. (2001) find a dominant effect of climate change on OH in the free troposphere so that the global CH₄ lifetime declines from about 9 years in 1990 to about 8.3 years by 2025 but does not change significantly thereafter. Hence the evolution of the CH₄ lifetime depends on the relative timing of NO_x and hydrocarbon emission changes in the emission scenarios, causing the calculated CH₄ increase in 2100 to be reduced by 27% when climate change is considered. Stevenson et al. (2006) reach a similar conclusion about the relatively constant CH₄ lifetime. As a result of future changes in emissions, the CH₄ steady-state lifetime simulated by 25 state-of-the-art CTMs increases by 2.7 ± 2.3% in 2030 from an ensemble mean of 8.7 ± 1.3 years for the present day (mean ± 1 standard deviation) for a current legislation scenario of future emissions of ozone precursors. Under the 2030 warmer climate scenario, the lifetime is reduced by 4.0 ± 1.8%: the total effect of both emission and climate changes reduces the CH₄ lifetime by only 1.3%.