Click for next page ( 67


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 66
8 Theoretical Projections of Stratospheric Change Due to Increasing Greenhouse Gases and Changing Ozone Concentrations JERRY D. MAHLMAN Geophysical Fluid Dynamics Laboratory National Oceanic and Atmospheric Administration This talk discusses what has happened in the stratosphere and what may happen there in the future. ~ will first review the ensemble of gases present in the stratosphere and their effects: 1. ChIorofluorocarbon (CFC-ll and CFC-12) increase. The increase of CFCs already appears to be causing ozone loss through the action of chlorine. CFCs also act as greenhouse gases in the troposphere. 2. Methane (CH4) increase. We have heard in a previous pre- sentation that methane is a tropospheric greenhouse gas. The ap- proximately 1 percent per year increase in methane also has implica- tions for long-term increases of the water vapor amount in the middle and upper stratosphere. However, the methane chemistry opposes the chlorine catalysis chemistry in the lower stratosphere. Methane increases also play a role in reducing the amount of hydroxy! (OH) in the troposphere. 3. Nitrous oxide (N2O) increase. Nitrous oxide is increasing by what appears to be the comparatively modest amount of about 0.2 percent per year. However, the N2O anthropogenic source is about one-thircI of the natural source and thus is not negligible on long time scales. Nitrous oxide is a tropospheric greenhouse gas, but its small annual increases are contributing little to current increaser in infrared radiative forcing. Reaction of N2O with excited atomic 66

OCR for page 66
THEORETICAL PROJECTIONS 67 oxygen in the rniddIe stratosphere produces reactive nitrogen (NO2), which provides the major natural catalytic loss of ozone. Oddly, NOX provides an important negative feedback against the growing attack on ozone by reactive chlorine (CITY. This occurs through the formation of chlorine nitrate (ClONO2), thus inhibiting both C1x and NOX catalytic ozone destruction cycles. 4. Stratospheric carbon dioxide (CO2) increase. The increases in CO2 will lead to a strong cooling trend in the stratosphere. To some extent, this cooling effect acts to moderate the expected mid- stratospheric ozone decreases. 5. Stratospheric ozone decrease. Large ozone decreases will also result in large stratospheric cooling. There is also strong column "self-healing" of ozone, which ~ will discuss later. 6. Stratospheric water vapor increase. An increase would re- sult in increased downward infrared radiative flux, complex chemical changes if stratospheric ice clouds form, and an increased ozone loss in the 30- to 50-km layer. I will next discuss what the NASA-WMO Ozone Trends Pane! (Watson et al., 1988) has learned about recent (1979 to 1985) ozone trends in the stratosphere. SAGE satellite data suggest a decrease of about 3 percent near 40 km altitude, on an average worldwide, over the 6-year period. The ground-based Umkehr data, on the other hand, suggest a decrease of about 9 percent at 40 km. Mode! calcula- tions predict a 4 to 9 percent decrease in response to increased trace gases, primarily CFMs, and a 1 to 3 percent decrease in response to declining solar activity, for a total decrease of 5 to 12 percent. Given this range of uncertainty, the observed ozone changes near 40 km are not inconsistent with theory. The credibility of the theoretical results is enhanced by the observation that stratospheric tempera- tures have decreased globally, between 25 and 55 km altitude, by 1.7C since 1979. This decrease is consistent with decreases in up- per stratospheric ozone of up to (but not larger than) 10 percent. The vertical profile of ozone change also is in fair agreement with theoretical predictions. ~ will now turn to a discussion of future trends. Atmospheric Ozone 1985 (WMO-NASA, 1986) contains estimates for equilibrium (infinite elapsed time) changes using one-dimensional chemical mod- els and assuming that CFM emissions are held constant at the 1980 rate. At 10 ppb reactive chlorine species, equilibrium total column loss is estimated at between 5 and 9 percent. At 15 ppb chlorine

OCR for page 66
l 68 JERRY D. MAHLA~4N species, the estimated loss ~ 10 to 20 percent. Thus the equilib- rium column loss doubles, for a 50 percent increase in odd chlorine. This nonlinearity results from the progressively greater scavenging of nitrogen oxides by chlorine, leaving the excess chlorine to destroy increasingly more ozone. On the other hand, if carbon dioxide is dou- bled (with no other atmospheric changes), equilibrium total column ozone will increase by 2 to 3 percent. However, the changes at the 4~km level predicted by the same mode! are much more drastic. At a concentration of 10 ppb chlorine species, the equilibrium prediction is an ozone decrease of 60 to 80 percent, and at a concentration of 15 ppb chlorine species, a loss of 70 to 85 percent would result. For the gIobal-mean vertical profile of ozone mixing ratio, the equilibrium prediction is for an increase in ozone in the lower stratosphere, between 10 and 25 km, of a few percent, and a decrease at higher altitudes, with the maximum de- crease at about 40 km. (The feedback of higher temperatures in the lower stratosphere and lower temperatures in the higher stratosphere would reduce slightly the magnitudes of both the increase and de- crease.) The predicted increase in ozone below about 25 km would result from more ultraviolet (UV) radiation penetrating to lower alti- tudes and creating more ozone in a kind of negative feedback process that tends to limit the depletion of total column ozone. Thus the change in the total column ozone is a comparatively small difference between two large numbers, given that the atmospheric mass drops off nearly exponentially with increased altitude. Figure 8-1 shows the two-dimensional mode} prediction of percent ozone decrease by Atmospheric and Environmental Re- search, Inc. (AER) as reported in WMO-NASA (1986) at 8.2 ppb C1x species. This is about equal to the equilibrium CI2 for 1980 emission rates. The resulting stratospheric distribution shows maxi- mum ozone losses at about 40 to 45 km poleward of about 50N and 50S latitudes. In these regions, the predicted ozone loss is greater than 50 percent. On the other hand, ozone is predicted to increase by about 20 percent in the lower stratosphere (15 to 20 km) near the equator. In low latitudes, the total column self-healing effect is comparatively strong because the incident solar UV ~ strong all year, whereas the effect is much weaker at the high latitudes because of weaker UV and the presence of large downward mean advection of ozone at these latitudes. A question that is fair to ask is, what is the credibility of such tw - dimensional models? This depends, of course, on which effects

OCR for page 66
THEORETICAL PROJECTIONS OI I 1 ~r 1 ~,- -I , r 1 ~, , [---_ APRIL _ - 60 _ l - E - 10 100 1 n -50 -40 In '` 1 \ ~ / 1 ~ 10 J o ~ 1 ~1 1 , 1 1 . 1 1 1 1 69 - By LO 40 30 A: X - 20 cr CL - 1 0 - 1 000 _ - 90- 60 - 30 ED 30 60 90 S LATITUDE N FIGURE 8-1 Two-dimensional model prediction of the percent change in ozone for 8.2 ppbv of total reactive chlorine (Clay. Model results from the chemical- transport model of Atmospheric and Environmental Research, Inc. (Reprinted from WMO, 1986.) are included and which are left out. The models predict large changes in ozone concentrations; therefore, one would expect significant tem- perature and circulation changes to occur as well. It should be pointed out that tw>dimensional models used for ozone assessment all make a similar assumption about change in circulation. They implicitly assume that the stratospheric circulation does not change as absorbers are changed. This corresponds to the "fixed dynamical heating" limit; that is, the net heating of local air by dynamical processes does not change with changing absorbers, even though the temperature itself is free to change. However, when a real climate system such as the stratosphere is perturbed, the changed distri- bution of absorbers may or may not lead to a changed temperature

OCR for page 66
70 JERRY D. MAILMAN distribution in order to maintain what we call "climate balance." The sources of temperature change are not just radiative but may also be dynamical. Dynamical heating may result from advection, diffu- sion, and adiabatic compression or expansion. The radiative heating is due to short-wave solar radiation (almost independent of atmo- spheric temperature) and long-wave radiation (strongly a function of temperature). The assumption is routinely made that the strato- sphere is in radiative equilibrium, resulting in a very simple system. But we know that this is a gross oversimplification, especially in the higher latitudes, where dynamic effects are very important. Actual temperatures in the polar latitudes sometimes differ by as much as 50 to 60 K from those given by radiative equilibrium models. Thus if the distribution of absorbers is changed, the dynam- ical heating, as well as the radiative heating, is likely to change. If dyna~rucal heating is important, three-dimensional models that include dynamical response should be used in preference to the two-dimensional models that implicitly invoke the fixed dynamical heating assumption. Such a two-dimensional model uses a latitude- altitude framework that has mean mass circulation directed upward in low latitudes, downward in high latitudes, and from equator to pole at stratospheric altitudes. Superunposed on the mean meridional motion is a meridional eddy diffusion of particles that is driven by upward-propagating tropospheric disturbances. Both of these trans- port mechanisms arise in the troposphere; thus the stratospheric latitude-altitude motions (and the degree of departure from radia- tive equilibrium) are driven by the troposphere and would decay to near zero in the absence of dynamical forcing from the troposphere. The question now is, how does the stratospheric circulation re- spond when the distribution of absorbers is changed? Calculations of this type were done at the Geophysical Fluid Dynamics I,abora- tory (GFDL) in 1980 by Fels et al. In one numerical experiment, carbon dioxide was doubled but sea temperatures were held fixed. In this case, Figure ~2 shows that the three-dimensional dynamic equilibrium mode} results agree fairly closely, to within 1C, with those from a fixed dynamical heating radiative model. The dynamic mode} predicts cooling of 11C at 45 km, compared to the radiative model's prediction of about 10C. Both models show a cooling of about 2 to 4C between 20 and 25 km. Figure 8-3 shows another experiment, in which an arbitrary uniform reduction of ozone by 50 percent was assumed. (This experiment ignored the column self- healing effect.) Again, the radiative mode! agrees fairly closely with

OCR for page 66
THEORETICAL PROJECTIONS 71 the dynamic mode! in terms of magnitude of cooling, but some of the gradient details are radically different, particularly in the tropics. Cooling of 6 to 8C is predicted in the high latitudes between 15 and 45 km by both models. At the equatorial tropopause, a 12C cooling is predicted. If we more realistically assume that the ozone at the equatorial tropopause increases as in the AER model, then we may infer that a warming of about 4C will tend to occur there. Recently, we have learned much more about how poor a job the older models did of simulating the dynamics of the stratosphere. As the horizontal resolution of models has increased, predicted strato- spheric temperatures have increased and come into better agreement with observations. Predicted temperatures using a 1-latitude reso- lution dynamic mode} agree closely with temperatures observed at 62N latitude in December, whereas results for a 9 resolution dy- namic mode! were fairly close to those obtained with the radiative mode! but from 15 to 45C colder than either the observed tempera- tures or results for the high-resolution dynamic model (see MahIman and Umscheid, 1987~. Thus, the traditional higher-latitude strato- spheric cold bias of general circulation models was not the fault of radiative transfer but rather the fault of oversimplifying the dynam- ics in the models. At high resolution, the models are considerably more dynamical in the stratosphere and produce temperatures and motions that look much more like those in the real stratosphere. It appears that the tropospheric dynamical processes are strong enough to push the stratosphere some distance away from its radiative equi- librium condition. Thus dynamic modeling is essential for predicting changes cor- rectly in the stratosphere. The good news is that a mode] with suffi- ciently high resolution is capable of useful predictions. The bad news is that such models, if run to equilibrium conditions, require sum stantial computer resources. Possibly, some of the resolution can be traded for carefully devised parameterizations that are self-consistent and appropriately sensitive to climate changes. In a stratosphere that is dynamically driven, the interannual variability is quite large, thereby increasing the overall computational problem. Based on our experience thus far, ~ would like to speculate on what ~ consider to be the stratosphere climate Issues that we will have to face beyond what we already know. We think we know that the upper stratosphere will coo} by 20 to 25C, perhaps more; this makes the stratosphere a candidate for inclusion in a full climate system model. Also, there will presumably be large ozone decreases in the

OCR for page 66
72 not .02 .03 .05 .10 .20 .30 .50 1.0 20 Dd LO 5.0 10 20 _ 50 _ 100 200 300 500 1000 _ JERRY D. MAILMAN T2Xco2-TcoNTRoL (FDH) 1 1 1 ~ 1 \1 1 1 1 1 1 1 1 1/ 1 1 \ 1 1 1 -8 Am\ 6 8 0 11 v' 2 > 3 4 5 6 7 8 9 20 21 22 23 24 25 26 27 28 29 30 31 8 .5 76 5 67.5 58 5 49.5 dC.S 3! 5 22.5 13 5 4 5 -~.5 -~3.5 -22 5 -3' 5 -40 5 -~.5 -SH.S -67.5 -76.5 -8 40 LATITUDE \\ J \ ~ _10- ~- Rae - _ 55 _ 50 _ 45 3: _ 40 I _ 35 ~ Z _ ~^ ~ _ 5 _ O ._ 70 65 60 45 40 35 30 25 20 15 10 FIGURE 8-2 GFDL "SKYHI" model equilibrium temperature changes due to a doubling of carbon dioxide as determined from an annual mean model with prescribed sea-surface temperatures. On the left is a fixed dynamical heating upper stratosphere that will be partially compensated for in the lower stratosphere. But, the unpredicted antarctic ozone situation warns us to be wary of things that may be missing from current model calculations. Speculation follows on some of the things that may need to be considered in future models. The first speculation concerns the possibility of reduced strato- spheric transport circulation, that is, reduced efficiency of the merid- ional circulation. The reason is related to Robert Dickinson's pre- sentation, which follows. The projected greenhouse gas warming in

OCR for page 66
THEORETICAL PROJECTIONS oft .02 .03 .os .10 _ .20 .30 .so _ 1.0 _ 2.0 10 20 an '00 _ 200 ~_ M 300 _ soo _ 1000 1 1 11 1 1 1 1 1 id_ 85.5N 76.5 67.5 58.5 49.5 40.5 31.5 22.5 13.5 4.5 -4.5 -13.5 -22.5 -31.5 -40.5 -49.5 -58.5 -67.5 -76.5 -85.5s LATITUDE 1 1 ( T2XCO2- TCONTROL (GCM) 1 1 1 1 1 1 1 1 1 1 /1 1 1 \1 1 ~ ) Cot -~-8- 7 11 ~ 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Off 73 1 1 1 1 1 1 1 1 - Rn 7S 70 60 55 50 - 45 I V UJ 40 ~ 35 ~ z A: 30 ~ 25 IS 10 (FDH) calculation; on the right is the full general circulation model (GCM) result. (Reprinted, by permission, from Fels et al., 1980. Copyright Hi) 1980 by The American Meteorological Society.) the troposphere should result in a weaker meridional temperature gradient, thus weakening tropospheric circulation and decreasing the flux of wave activity to the stratosphere. The stratospheric transport circulation would in turn be reduced. Another speculation is that water vapor will increase in the lower stratosphere (as well as in the upper stratosphere because of methane). With the column feedback process resulting in increaser] ozone in the lower stratosphere, it ~ possible that a significant heating of the equatorial tropopause may occur. A heating of up to 4C

OCR for page 66
74 JERRY D. MAILMAN n1 no .03 .05 an .20 .50 1.0 2.0 50L 10 50 100 200 300 soot 1000 L I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ,_ 85 SN 76 5 67 5 58 5 49 5 ".5 31.5 22 5 13.5 4 5 -4 5 -13 5 -22 5 -31.5 -I 5 -49 5 -58 5 -67 5 -76 5 -85 55 LATITUDE so 75 70 ~5 .s5 50 E 45 I 40 ~ a: 35 ~ Z tan on IS In FIGURE 8-3 GFDL "SKYHIn model equilibrium temperature changes due to a uniform 50 percent ozone reduction. On the left is a fixed dynamical heating (FDH) calculation; on the right is the full general circulation model (GCM) increases the saturation vapor pressure by up to a factor of two. This temperature effect, leading to a water vapor increase in the 15- to 2~km region, may be much greater than that due to methane at these altitudes. We are beginning to understand the influence of the antarctic ozone seasonal depletion on the ozone climatology of the Southern Hemisphere. It appears that dilution is occurring, causing significant ozone decreases throughout the Southern Hemisphere, as pointed out in the presentations by Robert Watson and F. Sherwood Rowland.

OCR for page 66
THEORETICAL PROJECTIONS ~01 02 .03 .05 in n ~ n s.o 1O 20 sot 20C 30 50( 1000 it- -- 1 1 1 1 1 1 85.5N 76.5 67.5 58.5 49.5 40.5 31.5 22.5 13.5 75 HERO __ 75 70 65 60 '1 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 _ - _ 31 1 1 1 1 1 1 1 1 1 1 = 40 4.5 -4.5 -13.5 -22.5 -31.5 ~40.5 -49.5 -58.5 -67.5 -76.5 -85.5S LATITUDE _55 _ O _ 50 E - 45 ~ V _ 40 ~ C) _ 35 ~ aS _ 30- 25 20 15 10 result. (Reprinted, by permission, from Fels et al., 1980. Copyright ~ 1980 by the American Meteorological Society.) We currently know much less about the situation in the Northern Hemisphere. However, the northern high-latitude region probably contains more tote] amount of chlorine species at any particular time in winter than does the southern region in winter, because the Northern Hemisphere region is dynamically more active. Strato- spheric cooling due to the combined effects of greenhouse gases and reduced ozone will likely be accented in higher latitudes. We al- ready know of the existence of some polar stratospheric clouds in the northern polar vortex, although they do not form nearly as efficiently there as in the Southern Hemisphere. The reason is not chemical but

OCR for page 66
76 JERRY D. MAILMAN dynamical; Northern Hemisphere dynamical mixing is greater than that in the Southern Hemisphere. But as the stratosphere cools due to changed greenhouse gases, polar stratospheric clouds should become more common and widespread in the north. Also, reduced transport circulation in the Northern Hemisphere could lead to a further decrease of temperatures at high latitudes. Thus the kind of ozone depletion due to heterogeneous chemistry that is now well documented in the Southern Hemisphere may become common in the Northern Hemisphere as well. ~ think that much work needs to be done, not only in field observations, but also in theoretical studies and modeling, to quantify this change and the other possible changes that ~ have mentioned. (In answer to a question): Only GFDL and NCAR have looked in detail at dynamical modeling of stratospheric climate change. We have learned that the stratosphere has a peculiar nonlinear "switch." This has been noted for many years in sudden warming events, which can temporarily push the stratosphere far out of radiative equilibrium. When this occurs, it seems to be much easier for the system to keep the new configuration, despite the large radiative imbalance. The reason goes back to wave propagation theory, in that if the winds are very strong and the system ~ very cold, planetary waves are refracted toward the equator with great efficiency. If zone] winds decrease, then the effects of tropospheric forcing are more likely to focus toward higher latitudes. In the climate-modeling case, the models had a cold bias that made model stratospheric receptivity to planetary waves too weak, which in turn accentuated the cold bias. However, if warning is introduced from another cause, then the "switch" may be thrown, rapidly putting the stratosphere into another, quite different, quasi-equilibrium state. Such a switch could also work in the opposite direction. The antarctic polar region, for example, is strongly resistant to any wave forcing because it is on the cold, high-wind-speed side of this implied "dynamical limit." The arctic region, on the other hand, is presently not so constrained by this dynamical limit because it experiences relatively high levels of dynamical forcing. (In answer to another question): Rowland, in his talk, and others have speculated that the antarctic ozone hole-generating process may already be operating to some extent in the Northern Hemisphere polar region. However, the ejects should be much less noticeable

OCR for page 66
THEORETICAL PROJECTIONS 77 because of the greater dynamic variability there as well as likely smaller levels of ozone chemical destruction. (In answer to a question about mode} resolution and mode] ac- curacy): Fifteen years ago, we were constrained to work with 9 reso- lution because of computer and other limitations, but we knew from comparison with observations that we were in big trouble without knowing why. About ~ years ago, we progressed to finer resolution, and results began to reflect much more the real stratosphere. We also found that there were several other modeling problems to contend with. We learned that low-resolution models cannot resolve gravity waves. We also learned that we could sometimes get the right answer even though the radiation was incorrectly specified. (One way to "fix" the cold bias in the low-resolution models was to put in a bad radiation code that allowed us to get closer to the observed temper- ature.) Only recently have we learned that at least the winter half of the year is dynamical in a fundamental way. There was a tacit assumption that the stratosphere was in radiative equilibrium in the wintertime except during sudden warming events. However, we have learned that tropospheric forcing of the stratosphere does not permit radiative equilibrium to be established, a fact that was first theorized by Dickinson (1975~. To put together a stratospheric mode! that does not ~cheat," in the sense of forcing a lower boundary condition or including a bogus radiation parameterization, requires an extremely long and strong commitment of an interdisciplinary team convinced that spending a decade on the problem is worth it. As a result, there have been few sustained participants in comprehensive stratospheric modeling. We knew, even 15 years ago, that stratospheric modeling in- volved more dynamics ant} three-dimensionality than we were able to represent at the time. My chemical colleagues have challenged me with the following question: What can be said, on the basis of three-dimensional dynamical models, about the viability of one- and two-dimensional models? One thing we did learn is that one- and two-dimensional models do have a theoretically defensible fundamen- tal basis. In both cases, we have learned that the basis is trickier than had been assumed. For example, in one dimension, the eddy diffu- sion coefficients are a function of the chemistry. Two-dimensional models can capture much of what a three-dimensional mode} does in a self-consistent way as long as they stick to prescribed transport, but when dynamical adjustment of the stratosphere is a dominant factor, the two-dimensional models are completely inept. Even so,

OCR for page 66
78 JERRY D. MAILMAN this mode} hierarchy has great value provided we are also aware of the limitations of each type of model. Even the 1 dynamical mode! gives polar temperatures in the lower stratosphere that are a bit too cold. The lower stratosphere is still not dynamical enough, probably because gravity waves are not properly represented. We may have to resort to parameterizing the effects of these gravity waves. However, when a parameterization is introduced to "fix" a model, one is not really justified in perturb- ing the mode} climate unless the parameterization is also perturbed accordingly. Here the modeler is faced with a dilemma, since the na- ture of most pararneterizations is that their variation under changed climatic conditions is unknown. REFERENCES Dickinson, R.E. 1975. Energetics of the stratosphere. J. Atmos. Terr. Phys. 37:855-864. Fels, S.B., J.D. Mahlm an, M.D. Schwarzkopf, and R.W. Sinclair. 1980. Strato- sphere sensitivity to perturbations in ozone and carbon dioxide: Radiative and dynamical response. J. Atmos. Sci. 37:2265-2297. Mahlman, J.D., and L.J. Umscheid. 1987. Comprehensive modeling of the middle atmosphere: The influence of horizontal resolution. Transport Processes in the Middle Atmosphere, G. Visconti and R. Garcia teds.), NATO ASI Series C: Mathematical and Physical Sciences, Vol. 213, D. Reidel Publishing Co., 251-266. Watson, R.T., M.J. Prather, and M.J. Kurylo. 1988. Present State of Knowl- edge of the Upper Atmosphere 1988: An Assessment Report. NASA Reference Publication No. 1208, National Aeronautics and Space Admin- istration, Washington, D.C. World Meteorological Organization-National Aeronautics and Space Adminis- tration (WMO-NASA). 1986. Atmospheric Ozone 1985: Assessment of Our Understanding of the Processes Controlling Its Present Distribution and Change. Global Ozone Research and Monitoring Project, Report No. 16, 3 vole., WMO, Geneva.