Coupling of Nitrous Oxide and Methane by Global Atmospheric Chemistry

Lingering Atmospheric Perturbations Nitrous oxide and methane are chemically active greenhouse gases whose atmospheric abundances are greatly influenced by anthropogenic emissions. Prather and Hsu (p. 952) used an atmospheric chemistry model to show how nitrous oxide emissions lower the concentration of tropospheric methane through a chain of chemical reactions that include stratospheric ozone depletion, changes in solar ultraviolet radiation fluxes, altered fluxes of ozone transport from the stratosphere to the troposphere, and increases in the amount of tropospheric hydroxyl radicals. This mechanism acts on a 108-year-long time scale—10 times longer than the atmospheric residence time of methane. Nitrous oxide emissions affect atmospheric methane concentrations on a time scale that is 10 times longer than the methane lifetime. Nitrous oxide (N2O) and methane (CH4) are chemically reactive greenhouse gases with well-documented atmospheric concentration increases that are attributable to anthropogenic activities. We quantified the link between N2O and CH4 emissions through the coupled chemistries of the stratosphere and troposphere. Specifically, we simulated the coupled perturbations of increased N2O abundance, leading to stratospheric ozone (O3) depletion, altered solar ultraviolet radiation, altered stratosphere-to-troposphere O3 flux, increased tropospheric hydroxyl radical concentration, and finally lower concentrations of CH4. The ratio of CH4 per N2O change, –36% by mole fraction, offsets a fraction of the greenhouse effect attributable to N2O emissions. These CH4 decreases are tied to the 108-year chemical mode of N2O, which is nine times longer than the residence time of direct CH4 emissions.

M ethane (CH 4 ) has environmental impacts beyond those of a direct greenhouse gas, through atmospheric chemistry that enhances the abundance of tropospheric ozone (O 3 ) and decreases that of hydroxyl radicals (OH) and hence the atmospheric lifetime of many other pollutants (1,2). Likewise, nitrous oxide (N 2 O) is a known O 3 -depleting substance (3,4). Both CH 4 and N 2 O interact directly in the chemistry of stratospheric O 3 , where global CH 4 concentration increases drive proportional but much smaller N 2 O increases (5). The pathway of how N 2 O affects global CH 4 is more complex, involving the coupling of stratospheric O 3 depletion with global tropospheric chemistry through OH, and consequently the lifetime of CH 4 .
This chemical coupling of the stratosphere and troposphere alters one's expectations about the amplitude and persistence of anthropogenic perturbations. Here we describe four multidecade numerical simulation experiments-one control and three perturbations runs-which have been designed to find such coupling between N 2 O and CH 4 through analysis of long-lived chemical modes. We define modes as the perturbation patterns in the abundances of all chemical species resulting from, for example, the addition of one species (i.e., eigenvectors of the linearized system). In atmospheric chemistry, the most important modes are the long-lived, slowly decaying ones associated with N 2 O and CH 4 , because these have the largest environmental impacts.
The University of California Irvine's threedimensional chemistry-transport model incorporates algorithms for both stratospheric and tropospheric chemistry (5)(6)(7). The meteorology and trace gas emissions in this study are representative of conditions in the year 2005 (7). For our control run, the CH 4 and N 2 O emissions are prescribed to achieve steady-state abundances of 1775 and 320 parts per billion (ppb), respectively. The annual budgets for N 2 O, CH 4 , and carbon monoxide (CO) from year 80 of this run are summarized in Table 1 The chemical modes, using an annually repeating meteorology, are five-dimensional (species, time of year, and three spatial dimensions) quantities, as shown in previous studies (5,9,10). Here we focus on the global burden on 1 January, thus ignoring the spatial and seasonal variations shown in (9). Modes are calculated from the difference between a perturbation and the control simulation. The same initialization and 80-year control simulation (C1) is used in all cases. The N 2 O perturbation (simulation C2) introduces a pulse of 10 Tg of N from N 2 O in the lowest four model layers on top of the initial conditions; the CO perturbation (simulation C3) introduces 100 Tg of CO; and the CH 4 perturbation (simulation C4) introduces 10 Tg of CH 4 . In the perturbation-minus-control differences, the shorter-lived modes (those associated with other chemical species, interhemispheric mixing, and stratosphere-troposphere exchange) decay within a few years, leaving behind a combination of the N 2 O-driven mode 1 (108 years) and the CH 4driven mode 2 (12.5 years).
The pulse of N 2 O alone (C2 -C1) induces perturbations in almost all species as shown in Fig. 1A. The relative perturbations in total atmospheric burden (in kilograms) eventually lock into the pattern of the 108-year mode 1: positive perturbations (solid lines) for N 2 O, odd nitrogen species (NOy), and upper tropospheric O 3 ; and negative (dashed lines) for CH 4 , CO, total O 3 , and lower tropospheric O 3 . The relative amplitudes, in terms of percent of the steady-state control C1, are shown in Fig. 2. For the long-lived gases N 2 O and CH 4 , these have only slight variations with altitude, but for O 3 , the pattern changes sign and varies greatly with latitude.
The CO pulse (C3 -C1) is shown in Fig. 1B. Results are plotted beginning 1 year after the pulse, when the initial 10 11 kg of CO and its related perturbation of tropospheric O 3 have mostly decayed. The CH 4 perturbation, caused by the COdriven decrease in OH, is established by year 1 and decays with the mode 2 time of 12.5 years, which is longer than the budget lifetime of 7.8 years. Perturbations of CO and total O 3 also occur in mode 2 and are apparent for at least the first 60 years. The N 2 O perturbation is predominantly in mode 1 and is caused by increased CH 4 in the stratosphere [see the stratospheric chemistry modes in (5)]. Following the chemical perturbations by the difference between two runs, C3 -C1, becomes difficult when the amplitudes in modes 1 and 2 become comparable as they do near the end of the simulation in Fig. 1B. The control C1 does not begin in an exact steady state and thus has a drift that is also a combination of the two modes, with the transition in O 3 and CO amplitudes from 2 to 1 occurring near year 50. This interference of the modes causes the derived amplitudes of CO and O 3 to diverge from a simple decay curve (e-fold), plus some noise in the subtraction (thin lines in Fig. 1B). This transition does not affect Fig. 1A, where the amplitudes in the longer-lived mode 1 are much larger.
In these perturbations, modes 1 and 2 can be separated and then traced back to a value at time zero, knowing that exponential decay is fixed globally for each mode. These time-zero amplitudes are reported in Table 2 for the three perturbations and the primary species of interest here (N 2 O, NOy, CH 4 , total O 3 , and tropospheric O 3 ). Chemical modes, like eigenvectors, are scalable patterns; and thus in Table 2, mode 1 perturbations can be characterized by the total mass of N 2 O, and mode 2 perturbations by that of CH 4 Table 1. Annual mean chemical budget terms (in teragrams/year) and budget lifetimes. Surface emissions, production (prod), and loss are in teragrams/year except for N 2 O, which are in teragrams of N/year. Terms are split into stratosphere (strat) and troposphere (trop). CO surface deposition is included in trop loss. CH 4 surface deposition is not included, and it would have shortened the CH 4 lifetime to 7.44 years, which is still within the standard deviation of models in (8) destruction. Thus, the avoided loss, about 0.2 Tg of N, adds to the amplitude of mode 1, amplifying the integrated radiative forcing from the N 2 O pulse. Such an increase of about 2% should be applied to all greenhouse gases with stratospheric loss, such as chlorofluorocarbons, but not to those with tropospheric loss, such as hydrofluorocarbons. In a similar but opposite vein, the pulse of 10 Tg of CH 4 (C4) has an amplitude of only 9.85 Tg in mode 2, because the most rapid loss occurs in the lower atmosphere where it is emitted. The N 2 O pulse (C2) produces a CH 4 amplitude of -2.1 Tg in mode 1 and +2.8 Tg in mode 2. This pulse must also excite other, short-lived modes (not seen here) with a net CH 4 amplitude of -0.7 Tg, because the initial CH 4 perturbation is zero. The overall negative perturbation of CH 4 appears within the first year, and within two decades, the CH 4 perturbation parallels that of N 2 O in mode1. This study looks at the coupled chemistry of stratosphere and troposphere, and thus we examine the mode patterns in Fig. 2  The use of chemical modes to deconstruct such perturbations allows for better understanding and predictions. Here we go beyond a onedimensional (1D) representation of the N 2 O chemical modes (14) by following the perturbations into 3D tropospheric chemistry. A relatively large reduction in CH 4 concentration (-3.6 ppb) is tied to an increase in N 2 O (+10 ppb) that decays with a 108-year time scale. Thus, N 2 O's climate impact through radiative forcing is diminished: (i) -8.4% because the decay of a pulse is faster than the e-fold of the steady-state lifetime used previously (15), and (ii) a further -4.5% to account for the decrease in the concentration of CH 4 . For CH 4 , the chemical feedbacks are opposite, lengthening the time scale of perturbations and increasing the greenhouse impact by about +40% (16)(17)(18).
Using the mode 1 coupling across species, one can readily compute that the N 2 O concentration increase since preindustrial times, from 270 to 320 ppb, has caused a 2-Dobson unit (DU) (0.7%) decrease in total O 3 abundance, as compared with the 3.5% decrease since 1980 that is attributable primarily to halocarbons (19). The corresponding mode 1 offset in CH 4 since preindustrial times, -18 ppb, is small relative to the overall CH 4 concentration increase of 1000 ppb.
The agreed-on atmospheric residence time for CH 4 used to weight emissions in international treaties is based on the mode 2 time scale, currently estimated as 12 years (20). Our derived mode 2 time scale of 12.5 years is well within the uncertainty, but the ratio of residence to budget lifetime in our model is 1.6, at the upper end of the model range in that assessment. Our model thus predicts much stronger chemical feedbacks from CH 4   This N 2 O-CH 4 coupling will shift with climate change over the 21st century. For example, the upper stratosphere cools as CO 2 increases, and this temperature change alters the N 2 O-NOy-O 3 chemistry, reducing the impact of N 2 O on O 3 (21). The importance of N 2 O as an ozone-depleting substance (22) will thus be reduced, weakening the coupling described here but still maintaining the negative greenhouse feedback effect of CH 4 on N 2 O emissions.
Penguin feathers are highly modified in form and function, but there have been no fossils to inform their evolution. A giant penguin with feathers was recovered from the late Eocene (~36 million years ago) of Peru. The fossil reveals that key feathering features, including undifferentiated primary wing feathers and broad body contour feather shafts, evolved early in the penguin lineage. Analyses of fossilized color-imparting melanosomes reveal that their dimensions were similar to those of non-penguin avian taxa and that the feathering may have been predominantly gray and reddish-brown. In contrast, the dark black-brown color of extant penguin feathers is generated by large, ellipsoidal melanosomes previously unknown for birds. The nanostructure of penguin feathers was thus modified after earlier macrostructural modifications of feather shape linked to aquatic flight. D uring wing-propelled diving, penguins generate propulsive forces in a fluid en-vironment~800 times more dense and 70 times more viscous than air (1). Recent fossil discoveries have yielded information on the sequence of early osteological changes in penguins accompanying the evolution of aquatic flight (2)(3)(4)(5), but these specimens have not included feathers. Living penguin melanosome morphologies have not been described, although the melanin they contain is generally known to provide both color and wear-resistance to bird feathers (6)(7)(8). Here, we describe a giant fossil penguin with feathers recording preserved melanosome morphologies (9) and discuss the pattern and timing of major events in the evolution of penguin integument. Reserve, Peru (10). Diagnosis: Inkayacu paracasensis is diagnosed by the following combination of characters (autapomorphies within Sphenisciformes demarcated by an asterisk): paired grooves meeting at midline on dorsal surface of premaxilla* (Fig. 1, GR), articular surfaces of otic and squamosal head of quadrate contacting one another,* furcula with blade-like hypocleidium,* conspicuous n. coracobrachialis sulcus developed on the humerus (Fig. 1, CNS), femur with widened and sharply distally tapering medial condyle* and tab-like process projecting from posterior intramuscular ridge at midshaft* (Fig. 1, MC and T), and weak proximal projection of cnemial crests of tibiotarsus (Fig. 1, CNE). See (10) for additional diagnosis, figures, and description. Table 2. Chemical modes and amplitudes excited by pulses of N 2 O, CO, and CH 4 . The modes are identified with perturbation minus control (P -C) simulations. Amplitudes (A) are in globally integrated teragrams of N for N 2 O and NOy and in teragrams for other species. The amplitudes of these modes are characterized by the amounts of N 2 O (mode 1) and CH 4 (mode 2), shown in bold. Relatively small amplitudes are not shown (-).