HARDBACK
\$99.95

Page 817

### Appendix Q Geoengineering Options

This appendix is divided into four sections: (1) naval rifle system, (2)- balloon system, (3) multiple balloon system, (4) changing cloud abundance. Each section either describes the system or indicates how the costs were computed.

#### Naval Rifle System

The current cost of a naval projectile weighing 1900 pounds (lb) is \$7000 to \$8000. The cost of propellant alone (if the shell is furnished) is \$900. It seems that a reasonable estimate for a 1-t shell, dust (commercial aluminum oxide can be obtained for \$0.25/lb), and a propellant for each shot is \$10,000. An efficiency of one-half is assumed: one-half of the shell is dust, and the other half consists of the packaging, dispersal mechanisms, and so on, necessary to make the shell function. Thus the cost of the ammunition for 40 years will be

The number of shots required in the 40 years is

If a single rifle can fire 5 shots per hour (naval rifles can fire faster than this, but cooling intervals between shots can lengthen the barrel life) and the rifle operates 250 working days per year, then a rifle can fire 5 shots/hour × 24 hours/day × 250 days/yr = 3 × 104 shots/yr per rifle.

The National Academies of Sciences, Engineering, and Medicine
500 Fifth St. N.W. | Washington, D.C. 20001

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 817
Page 817 Appendix Q Geoengineering Options This appendix is divided into four sections: (1) naval rifle system, (2)- balloon system, (3) multiple balloon system, (4) changing cloud abundance. Each section either describes the system or indicates how the costs were computed. Naval Rifle System The current cost of a naval projectile weighing 1900 pounds (lb) is \$7000 to \$8000. The cost of propellant alone (if the shell is furnished) is \$900. It seems that a reasonable estimate for a 1-t shell, dust (commercial aluminum oxide can be obtained for \$0.25/lb), and a propellant for each shot is \$10,000. An efficiency of one-half is assumed: one-half of the shell is dust, and the other half consists of the packaging, dispersal mechanisms, and so on, necessary to make the shell function. Thus the cost of the ammunition for 40 years will be The number of shots required in the 40 years is If a single rifle can fire 5 shots per hour (naval rifles can fire faster than this, but cooling intervals between shots can lengthen the barrel life) and the rifle operates 250 working days per year, then a rifle can fire 5 shots/hour × 24 hours/day × 250 days/yr = 3 × 104 shots/yr per rifle.

OCR for page 817
Page 818 Thus are required. Therefore, operating inventory of 4 × 102 riflescan be assumedat any time. A gun barrel will have to be replaced approximately every 1500 shots; thus over the 40 years, will be needed. A gun barrel probably would cost (in continuous production several hundred thousand dollars—say a million dollars. The total cost of rifle barrels is thus3 × 105 barrels × 106 \$/barrel = \$3 ×1011 for barrels. If the rifles are organized into 10-barrel stations, on land or at sea, and a billion dollars is allocated for the capital cost of each station, one might expect to buy 40 10-barrel stations to keep 350 barrels operating at a time, thus giving a cost for stations of 40 stations × 109 \$/station = \$4 × 1010. This should probably be doubled, at least; to allow for overhead, power, maintenance, replacement, and so on. Multiplying by 5 gives \$2 × 1011 for stations. Finally, people are needed to operate the system. Although the system would probably be highly automated, assume that it will work like current operations. Then allocate 10 people/barrel × 4 × 102 barrels × 3 shifts × \$105/person/yr × 40 years = \$48 × 109 \$5 × 1010, which can be doubled to include indirect personnel, overhead, and so on, giving \$1011 for operators. Therefore, 24,000 people are assumed to be involved at any time. To sum up, Ammunition \$4 × 1012 = 4.0 × 1012 Rifle barrels \$3 × 1011 = 0.3 × 1012 Stations \$2 × 1011 = 0.2 × 1012 People \$1 × 1011 = 0.1 × 1012   TOTAL \$4.6 × 1012 \$5 × 1012 for 40 years, giving an annual undiscounted cost of \$50/40 × 1011 = \$100billion.

OCR for page 817
Page 819 Clearly, the cost of the project is dominated by ammunition, and the number of stations and rifles is reasonable, as is the amount of activity, considered on a large industrial scale. The rifles could be deployed at sea or in empty areas (e.g., military reservations) where the noise of the shots and the fallback of expended shells could be managed. Balloon System Consider a hydrogen balloon floating at 20 km, using the Archimedes principle and noting that the density of hydrogen-gas is one-fourteenth that of air: md(isplaced) = mg(as inside balloon) + mb(alloon) + mp(ayload) 4/3pr3ro=4/3pr31/14ro + 4pr2Drrs(kin) + mp mp=4/3pr3ro13/14-4pr2Drrs =4pr3[13/(3x14)ro-(Dr/)/rs] If r = 100 m (radius of balloon) ro = 88 g/m3 = 8.8 × 10-2 kg/m3 (density of air at 20 km) Dr = 1 mm = 10-3 m (thickness of balloon skin) ro = 1.15 g/cm3 (nylon) × 10-3 kg/g × [102 cm/m]3 = 1.15 × 103 kg/m3. Then mp = 1.26 × 107 (2.7 × 10-2 - 1.15 × 10-2) = 1.26 × 105 (1.55) = 1.95 × 105 2 × 105 kg. The mass of the balloon for a 1-mm thickness is 4pr2Drr = 12.6 × 104 × 10-3 × 1.15 × 103 = 1.26 × 1.15 × 105 × 10-3 × 103 kg = 1.5 × 105 kg. If the balloon is 2/3-mm-thick (assumed for convenience), its mass from the previous computation is 1.5 × 105 kg and the mass of dust lifted, if a 50 percent efficiency factor is used to account for instruments, dust dispenser, container, and so on (this is conservative), is 105 kg. Nylon of the appropriate gauge for weaving into a 2/3-mm-thick fabric (1050 denier is about 0.3 mm) costs \$2/lb = \$4.4/kg. If this is tripled for fabric and balloon manufacture

OCR for page 817
Page 820 (the cost of parachute fabric is about 3 times the cost of the yarn, based on information from a colleague at Du Pont Industrial Fabrics), cost of controls, dust dispensing, and so on, \$15/kg can be estimated or 1.5 × 105 kg/balloon × \$15/kg = 2.25 × 106 \$/balloon. Twenty lifts are necessary in 40 years: 2 × 106 balloons × 2.25 × 106\$/balloon = \$4.25 × 1012. Consider the additional costs of infrastructure and support: there will be 2 × 106 lifts in 40 years or If there are 100 crews (each responsible for 2 lifts per day on 250 days a year) and each crew has 100 people, 104 people × 105 \$/person/yr × 40 years = \$4 × 1010 \$1011 with an overhead of 150%. If each station is capitalized at \$109, another \$1011 is required, but this infrastructure barely affects costs, as does the cost of dust even at \$0.50/kg or hydrogen at \$10/kg. Hydrogen can currently be purchased as liquid hydrogen in 1500-gallon lots (equivalent to 169,000 standard cubic feet) for \$2.5/100 ft3. For conversion, 1 kg of hydrogen = 432.3 standard cubic feet. Thus the cost is In quantities of 100 × 106 ft3/day, Ogden and Williams (1989) quote costs lower than \$30/GJ. This is Each balloon has a mass of 4.2 × 106 m3 × 1/14 × 8.8 × 10-2 kg/m3 = 2.6 × 104 kg of hydrogen. At 5 × 104 balloon lifts per year, the annual quantity is 13.2 × 108 kg 109 kg = 423 × 109 ft3 109 ft3/day = 102 × 106 ft3/day.

OCR for page 817
Page 821 The total mass of hydrogen required for 40 years is 2.6 × 104 kg/balloon × 2 × 106 balloons = 5.2 × 1010 kg. At \$10/kg, this costs \$5.2 × 1011 = \$0.52 × 1012. [Design note: The breaking strength of 1200 denier ( 0.4 mm) nylon is over 25 lb (Du Pont, 1988). The equatorial circumference of the balloon is 2pr = 6.3 × 102 m × 103 mm/m = 6.3 × 105 mm; therefore, the payload will be suspended from a double (actually 2.5) set of nylon strings 0.4 mm in diameter: 6.3 × 105 mm × [25 lb/(2.2 lb/kg)] × 2 = 142 × 105 kg. Because the payload weighs 1.18 × 105 kg, the safety factor = 121 times!] By using hydrogen at \$10/kg, costs may be summarized as Balloons \$4.25 × 1012 Infrastructure and personnel \$0.10 × 1012 Capital for launch stations \$0.10 × 1012 Hydrogen \$0.52 × 1012   TOTAL \$4.97 × 1012     \$5 × 1012. This mitigates 1012 t of carbon or 4 × 1012 t of CO2. An undiscounted cost the same as that for the naval rifle system is obtained: \$5/t C = \$1.25/t CO2 \$5/40/t C/yr = \$0.125/t C/yr = \$0.03/t CO2/yr. All of the above material assumes no reuse of balloons, and no allowance is made for the automation of launch, and so on. The possibility of some reuse, and of automation, probably reduces the total cost. If not controlled to land for reuse, balloons could be "chased" and controlled to land for collection and disposal, or to land in the deep ocean and sink promptly. Consider hot air balloons. Again by using the Archimedes principle, mdisplaced = mgas + mballoon + mpayload Vro = Vri + mballoon + mpayload Using the perfect gas law piVi = miRTi pi = RriTi poVo = moRTo po = RroTo where m = mass, V = volume, po = outside pressure, pi = inside pressure, ro = density of air outside, and ri = density of gas inside. At floating equilibrium,po=pri, because the balloon is limp. Therefore,

OCR for page 817
Page 822 If r and dr are expressed in meters, and roand rs are in each r must be multiplied by 1/10-3 = 103: mp = 12.6 × 103 r3 { (ro/3)[(Ti - To)/Ti] - (Dr/r)rs} = 1.26 × 104 r3 { (ro/3)[(Ti - To)/Ti] - (Dr/r)rs} where r (specific gravity) is expressed in grams per cubic centimeter, r in meters, and mp in kilograms. At 20 km, ro = (88 g/m3)(102 cm/m)3 = 88 × 10-6 g/cm3 for To = -58.5°C = 217 K (Kelvin) Ti = 104°C = 377 K (Kelvin) r = 102 m Dr = 1 mm r s = 1.15 = 1.26 × 1010 (1.23 × 10-5 - 1.15 × 10-5) = 0.1 × 105 kg = 104 kg. If 2/3-mm nylon is used, mp = 6 ×4 kg. Thus the costs of a hot air balloon system can be expected to be at least 4 to 10 times higher than the cost of a hydrogen balloon system. These costs

OCR for page 817
Page 823 could be decreased by running the balloon at higher temperature, but to get 105 kg of payload per balloon with 1-mm nylon a temperature of 658 K (385°C) is required, and with 2/3-mm nylon 475 K (202°C), which seems difficult to manage. The breaking strength of nylon goes to zero percent of its room temperature value by 250°C. While the skin temperature of a hot air balloon is well below the core gas temperature, the management of temperature to guarantee skin strength with so large a differential between average and skin temperature seems rather difficult, although the skin might be insulated as some weight penalty. The results are sensitive to the factors. Hot air balloons seem to be nearly competitive with hydrogen balloons. This question would have to be explored further before choices between hydrogen and hot air systems could finally be made. Multiple Balloon System The mass of a bubble filled with hydrogen is one-fourteenth the mass of the air displaced. The total mass of the hydrogen-filled balloon will be (at any altitude) At floating equilibrium, we have 1/14·4/3pr3ra=4pr2Drrs=4/3pr3ra Drrs=13/14rra Drrs=13/3x14rra=3x10-1 rra If plastic with density of 1 g/cm3 and a skin thickness of Dr = 10-1 mm = 10-4 m = 10-2 cm (which is plausible) is used, then At 19 km = 12 miles 62,000 feet, ra 10-4 and r3x10-2/10-4=300cm=3. Such a balloon has a disk area of pr2 = 9p = 28 m2 = 3 × 10 m2. Thus, 5x1012/3x10 2x1011 = 200x109 balloons of 3 -m radius

OCR for page 817
Page 824 are required. If the balloon is 10-mm material, a balloon of 3 × 10-1 m (30-cm) radius is obtained and 20,000 × 109 balloons are needed. Hydrogen will diffuse through the skin of the balloons, which probably means that the system must be refreshed annually. The fall of collapsed balloons might be an annoying form of trash rain. Because the area of the material required for a balloon is 4pr2, the material requirement is of material for any size balloon. At \$0.10/m2 (20 m2 of wrapping plastic can be bought in the supermarket for about \$2), this is \$2 × 1012. Over 40 years, this amounts to \$80 × 1012. It offsets 1012 t of carbon, so the cost is \$80/t C or \$80/40 = \$2/t C/yr or \$0.50/t CO2/yr. A reasonable cost range of \$0.50 to \$5/t CO2/yr can be assumed. Changing Cloud Abundance A study was undertaken to consider the various factors that would be required to increase the albedo effect of global cloud cover sufficiently to balance the temperature increase that is projected to occur with a doubling of CO2. Toward this end, the temperature sensitivity to different (high, middle, and low) cloud layer properties was calculated by using a radiative-convective atmospheric model. In addition, cost estimates have been made. These amelioration processes are reversible and inexpensive. If they were determined to be deleterious or if cost-competitive programs were developed, these measures could be discontinued immediately. At the outset it cannot be emphasized too strongly that there are tremendous uncertainties associated with these intellectual exercises. As a case in point, circumstantial evidence teaches that we have a very limited understanding of the role of cloud abundance because a warming accompanied the measured increase in cloud coverage over the past century. Consequently, a much better understanding of the system is necessary before any large-scale operations could reasonably be proposed. The Climatic Effect of Clouds Earlier, Reck (1978) studied the effect of increases in cloud cover and, using a radiative-convective atmospheric model, found that a 4 to 5 percent increase in low-level cloud cover would be sufficient to offset the warming predicted from a doubling of preindustrial CO2. This value is in reasonable agreement with Randall et al. (1984), who estimated that a 4 percent increase was required in the amount of marine stratocumulus, which comprises the bulk of the low clouds on a global basis. Unfortunately, many

OCR for page 817
Page 825 assumptions are contained in these estimates, and to understand those assumptions and the role that clouds could play, cloud sensitivity calculations have been made to illustrate the range of surface temperature for various assumptions of cloud properties. In these calculations, the Mitigation Panel used the assumed abundances and optical properties shown in Table Q.1 and a global surface albedo of 15.4 percent. The model has three layers of clouds under global average conditions. It is assumed that clouds, once formed, will have the same effects over their entire lifetimes and that they will have optical properties identical to those of current low-level clouds, which are assumed to be unchanging during the seeding process. Unfortunately, these assumptions contain many uncertainties. These sensitivity calculations show that the effects of clouds depend not only on the fraction of a given cloud type, but also on the surface albedo beneath the clouds. The special role of the low-level cloud and its varying effect as the surface albedo changes add considerable complication because the surface albedo varies from about 4 to 20 percent over some water to as high as 90 percent over pure snow or ice (Hummel and Reck, 1979). This means that once a cloud is formed it may start with a cooling effect and end up in an area where it could produce either greater or lesser cooling, with the slight possibility of even a heating effect. Albrecht (1989) (see also Twomey and Wojciechowski, 1969) suggests that the average low-cloud reflectivity would increase if the abundance of cloud condensation nuclei (CCN) were to increase through emission of SO2. TABLE Q.1 Assumed Properties of Average Global Clouds   Cloud Type   High Middle Low Cloud Abundances       Fraction of shortwave cloud cover 0.181 0.079 0.302 Fraction of longwave cloud cover 0.181 0.079 0.302 Cloud Optical Properties       Solar albedo of cloud cover 0.21 0.48 0.69 Solar absorptivity of cloud cover 0.005 0.02 0.035 Infrared absorptivity of cloud cover 0.50 1.00 1.00

OCR for page 817
Page 826 To test for the sensitivity to this part of the problem, the surface temperature changes with varying optical properties were calculated and are shown in Table Q.2. For comparison purposes the sensitivity of high and middle clouds was also included. Clearly, the estimate depends strongly on the value of assumed low-cloud solar reflectivity. For example, a change of 4 percent in the reflectivity value (low-cloud abundance—see Table Q.2) would be sufficient to cause the calculated surface temperature to change by 3°C. With a sensitivity of this magnitude, clearly a large potential exists for forced changes provided they could be controlled, and provided large regional anomalies and uncontrolled long-distance effects are not created. There is also a height dependence in the radiation field that varies greatly with latitude and altitude (Ramanathan et al., 1987). The cloud fraction variation with latitude is shown in Table Q.3. In the present environment, there is a greater probability of having clouds over water than over land, with more clouds over land in the afternoon and more clouds over water in the morning. This occurs because cloud height and optical properties are intimately related to humidity and physical conditions. For example, the role of a cloud at a given latitude is controlled by the zenith angle of the sun. If the cloud were to move to a more northern latitude, its cooling effect would be expected to diminish in proportion to the change in the cosine of the sun's zenith angle. As can be noted from the cosines listed in Table Q.3, a cloud at 5° latitude could have about twice as large a contribution as the same cloud at 65° latitude. Many less predictable features are also crucial (such as the degree of evaporation). Reck (1978, 1979), using a model based on that of Manabe and Wetherald (1967), has also illustrated cloud height effect. These calculations show heating from high-level clouds and cooling from middle- and lower-level TABLE Q.2 Calculated Surface Temperature Sensitivity to Changes in Cloud Properties   Cloud Type   High Middle Low Sensitivity (°C) per percent change in cloud abundance 0.36 -0.35 -0.66 Sensitivity (°C) per percent change in cloud albedo -0.16 -0.06 -0.35 Sensitivity (°C) per percent change in cloud absorptivity -0.062 0.048 0.045

OCR for page 817
Page 827 TABLE Q.3 Latitudinal Variation of Assumed Annual Cloud Cover     Fraction of Cloud Cover Latitude (degrees) Cosine of Zenith Angle Upper Cloud Middle Cloud Lower Cloud 5 0.61 0.225 0.075 0.317 15 0.593 0.181 0.064 0.264 25 0.560 0.160 0.063 0.248 35 0.512 0.181 0.079 0.302 45 0.450 0.210 0.110 0.388 55 0.381 0.242 0.131 0.438 65 0.309 0.254 0.119 0.444 75 0.259 0.252 0.111 0.424 85 0.243 0.205 0.092 0.375 ones. One possible error in the estimates presented here is the assumptionof either a fixed cloud altitude or a fixed cloud temperature. Reck (1979)has shown a greater model sensitivity to a fixed cloud temperature. Mixedbehavior might be observed in the real atmosphere. Clearly with all thepossible heating or cooling effects, the presence of naturally occuring cloudscould complicate the analysis of data obtained to test the role of humanintervention. See, for example, the cloud experiments suggested below. With all the above assumptions in mind, it is proposed both that CCN emissions should be done over the oceans at an altitude that will produce an increase in the stratocumulus cloud albedo only, and that the clouds will remain at the same latitudes over the ocean where the surface albedo is relatively constant and low. As noted in Figure Q.1, an increase in surface albedo, should the cloud float over land, would only enhance its cooling effect. This is true provided the latitude of the cloud does not change, as discussed previously. How Cloud Condensation Nuclei Can Change Climate Despite the lack of knowledge about cloud processes, the possibility of altering clouds has been considered for a long time. The idea of cloud seeding for agricultural purposes became popular in the 1950s and 1960s, but because of the lack of precision and the litigation that resulted, it has not been very succesful (see, for example, Todd and Howell, 1985; and Kerr, 1982). Changes in cloudiness on a regional scale were also proposed some time ago by Russian scientist, who considered decreasing the cloudiness

OCR for page 817
Page 828 FIGURE Q.1 Calculated surface temperature variation with changes in low-cloud cover and surface albedo. in the arctic region to promote ice melting and improved growing conditions in Siberia. Before the more recent satellite measurements, most of what was known about cloud processes and how they contribute to the global radiative balance came from climate modeling, and in climate models, most of the details of the cloud processes were not included. Certainly, no individual clouds were included on the grid scale of the general circulation models (GCM); thus specific details of the microphysics, as it might involve seeding or CCN, could not be studied within the concept of GCMs. Proposed Change in Low-Cloud Albedo Through Emissions of Cloud Condensation Nuclei In a recent paper, Albrecht (1989), following a hypothesis of Twomey and Wojciechowski (1969), grossly estimated the additional CCN that would be necessary to increase the fractional cloudiness or albedo of marine stratocumulus clouds by 4 percent. He estimates that this increase in low-level fractional cloudiness would be equivalent to that attributed to a 30 percent increase in CCN. As noted from Table Q.3, this 4 percent increase, if it were strictly in lower-level cloud abundance at global average conditions (35° latitude), would be more or less equivalent to the cloudiness at 4° latitude further north. Albrecht's idealized stratocumulus cloud, which he argues is typical, has a thickness of 375 m, a drizzle rate of 1 mm per day,

OCR for page 817
Page 829 and a mean droplet radius of 100 mm; he also assumes that each droplet is formed by the coalescence of 1000 smaller droplets. The rate at which CCN are depleted by this model is 1000/cm3 per day. Consequently, about 300/cm3 per day (30 percent of 1000) of CCN would be needed to be discharged at the base of the cloud to maintain a 4 percent increase in cloudiness. This assumes that the perturbed atmosphere would remain sufficiently close to saturation in the vicinity of the CCN that additional cloud cover would be formed every time the number of CCN increased. Now an extrapolation will be made to the entire globe, while keeping in mind Albrecht's assumption that cloudiness in a typical ocean region is limited by the small number of CCN. On the average, 31.2 percent of the globe is covered by marine stratiform clouds (Charlson et al., 1987). If no high-level clouds are present, the number of CCN that must be added per day is = 4p × (radius of earth)2 × (cloud-layer thickness) × 31.2 percent × CCN/volume = 4p × (6.37 × 108 cm)2 × (3.75 × 104 cm) × 0.312 × 300/cm3/d = 1.8 × 1025 CCN per day. The three materials that have been used for cloud seeding are silver iodide (AgI), lead iodide (PbI), and dry ice. Dry ice is not applicable to this situation because it does not create CCN. It is used because of its precipitation-enhancing properties. Lead iodide will not be considered because it was used before full awareness of the environmental problems associated with lead. Although adverse environmental consequences will probably also be associated with AgI, a calculation will be made anyway. Calculations will also be performed using sulfuric acid (H2SO4), because most of the CCN that occur naturally over the oceans are believed to be H2SO4 CCN arising from the oxidation of dimethyl sulfide (DMS) produced by planktonic algae in the seawater (Charlson et al., 1987). The mass of a CCN is (4/3 pr3 × density), and it is assumed that the mean radius r = 0.07 × 10-4 (Charlson et al., 1987). Because the density of AgI is 5.7 g/cm3, the CCN mass is = 4/3p × (0.07 × 10-4 cm)3 × 5.7 g/cm3 = 8.2 × 10-15 g. The total weight of AgI to be added per day is = (total number to be added) × (weight of average CCN) = 1.8 × 1025/day × 8.2 × 10-15 g = 1.5 × 1011 g/day or about 1.5 × 105 t per day. Worldwide silver production in 1985 was 420 × 106 ounces (U.S. Bureau of the Census, 1987). This is converted to metric tons:

OCR for page 817
Page 830 420 × 106 oz/yr × 28.35 g/oz × 1 t/106 g = 11.9 × 103 t/yr of Ag, or = 25.5 × 103 t/yr AgI. Clearly there is not enough silver or AgI to consider this experiment. For H2SO4, with a density of 1.841 g/cm3, the total weight to be added per day = 1.841/5.7 × 1.5 × 105 t/day = 48 × 103 t/day H2SO4 = 31 × 103 t/day SO2, if all the SO2 is converted to H2SO4 CCN. To put this number in perspective, a medium-sized coal-fired U.S. power plant emits about this much SO2 in a year; the equivalent emissions of 365 U.S. coal-burning power plants (50 percent of present U.S. SO2 emissions) would produce sufficient CCN. To estimate the value of the sulfur directly, the total weight of SO2 to be added per day is 32 × 103 t or about 16 × 103 t of sulfur, which is equivalent to about 6 megatons (Mt; 1 Mt = 1 million tons) of sulfur per year. Given the average market price of sulfur for 1983–1987 (f.o.b. mine or plant)—\$96.90 (U.S. Bureau of the Census, 1988)—the minimum yearly cost would be at least \$580 × 106/yr. Equating this yearly cost to the 300 parts per million by volume (ppmv) of CO2 necessary for full compensation gives \$580 × 106/(2840 Mt C/ppmv CO2 × 300 ppmv CO2), or about a fraction of a cent per ton of CO2. To obtain an equivalence to conserved carbon, known emissions of carbon in 1978, 1979, and 1980 have been compared with the total measured increase of CO2 to obtain the equivalence: 3890 Mt C 1 ppmv CO2. A 4 percent increase in cloudiness was then equated to a 300-ppmv CO2 decrease, which translates into a reduction of 1200 gigatons (Gt; 1 Gt = 1 billion tons) of carbon, or 4400 Gt of CO2. The primary cost of this process involves the mechanism for distributing SO2 in the atmosphere at the correct location. Assume a fleet of ships each carrying sulfur and a suitable incinerator. The ships are dedicated to roaming the subtropical Pacific and Atlantic oceans far upwind of land while they burn sulfur. They are vectored on paths to cloud-covered areas by a control center that uses weather satellite data to plan the campaign. In addition to choosing areas that contain clouds, it is important to distribute the ships and their burning pattern so as not to create major regional changes, or the kind of change with a time or space pattern likely to force unwanted wave patterns. These restrictions (which we may not know how to define) could be a difficult problem for such a system to solve. From the above, 16 × 103 t/day, or 6 Mt/yr of sulfur must be burned. If 102 t per ship per day are allocated, and a ship stays out 300 days each year, roughly 200 ships of 10,000-ton capacity are needed (one reprovisioning stop every 150 days). At a cost of \$100 × 106 per ship (surely generous),

OCR for page 817
Page 831 the capital cost of the fleet is \$2 × 1010. Amortized over 20 years, an annual capital cost of \$1 × 109 may be used. The sulfur will cost another \$0.6 × 109 per year, and \$2 × 106 per ship per year may be allocated for operating costs (\$10,000 per operating day), to give a total cost of \$2 × 109 annually. Over 40 years (until 2030) this means \$8 × 1010, or \$1011. This continuously mitigates ˜103 Gt = 1012 t for a cost of \$0.10/t of CO2. Of course, this continues to be a yearly cost of \$1 × 109/yr. The SO2 could also be emitted from power plants. These plants could be built in the Pacific Ocean near the equator (hopefully on small deserted islands) and would serve to furnish power for nearby locations (e.g., South America). Transmission or use of the power in the form of refined materials could be considered, or possibly the use of superconducting power transmission systems. It is estimated that eight large power plants using spiked coal would be required (with 4 times the normal amount of sulfur) at a cost of \$2 to \$2.5 × 106 per plant. Most of the cost would be borne by those buying the power, so the cost might be at most 10 percent per year (the interest on the investment), or a total of \$2 × 109 per year (with the above conversion, \$2 × 109/3890 × 106 \$0.0005/t CO2). Comparison of the Cloudiness and Proposed Cloud Condensation Nuclei Emissions with Current Estimates in the Real Atmospher Total U.S. SO2 emissions are 65.7 × 103 t per day, which is roughly 2 times the amount calculated in the previous paragraph. Consequently, there should already be some cloud-enhancing effects evident in the northern hemisphere if Twomey and Wojciechowski's hypothesis, as implemented by Albrecht, is correct. An examination of available CCN data shows that the mean CCN concentration at oceanic locations in the northern Atlantic is about 5 times higher than at remote locations in the southern Pacific (see Schwartz (1988), who, however, concludes that there is no discernible contribution of anthropogenic SO2 emissions to the global cloud cover effect on planetary albedo or temperature). Furthermore, several studies have examined trends in cloudiness in the northern hemisphere and have all come to the same conclusion: The total cloud amount has been increasing in the northern hemisphere (study areas include United States, North America, the North Atlantic, and Europe) since the early 1900s (Henderson-Sellers, 1986, 1989; Changnon, 1981; Angell et al., 1984; Warren et al., 1988). The largest increases in cloudiness in the United States occurred from the 1930s to about 1950 and from the mid-1960s to about 1980. The first period corresponds to a period of rapid growth of U.S. SO2 emissions after the Depression and extends to the end of World War II; the second period corresponds to the proliferation of tall stacks. From 1965 to 1980 the mean effective stack height (physical height of stack plus plume rise) of SO2

OCR for page 817
Page 832 emissions doubled from about 300 to 600 m. This, of course, increased the lifetime of discharged emissions in the atmosphere and transformed the SO2 pollution problem from primarily a local issue in many localities to a long-range transport issue. Between 1900 and 1980 the mean cloud cover over the conterminous United States has increased about 10 percent (Henderson-Sellers, 1989), which should be more than sufficient to compensate for an equivalent doubling of CO2. Because CO2 increased only about 12 percent during the same period, the net effect should have been a cooling. However, analyses of temperature data in the northern hemisphere over the same periods consistently indicate that the mean temperature has risen about 0.5° to 0.7°C overall, but no trend was evident in the conterminous United States (Jones et al., 1986; Hansen and Lebedeff, 1987; Hanson et al., 1989). This suggests either that the effects of clouds are not understood, or that other factors, such as the very poor data reliability for cloudiness and the effect of cloud height, need to be considered. Wigley (1989) presents some crude calculations suggesting that SO2/CCN-derived forcing could be large enough to have offset any temperature increase due to CO2 in the northern hemisphere. Schneider (1972) points out that SO2 emissions are regionally heterogeneous, which would lead to long-wave forcing anomalies that in turn could lead to long-wave anomalies plus teleconnections. In any event, all of this is quite speculative and underscores the fact that much is yet to be understood about the causes of climate variations during the last century. Impacts of Enhanced Acid Deposition One must now consider whether the injection of this much additional SO2 into the atmosphere will cause an acid deposition problem. It should be kept in mind that the principal component of naturally occurring CCN is sulfate formed from DMS emission from marine algae. Schwartz (1988) quotes estimates of 16 to 40 × 1012 g/yr or perhaps about 25 × 109 kg/yr emitted from this source. The addition about 6 × 109 kg/yr is being considered, approximately 25 percent of the natural amount, although locally much more than 30 percent may be added to the amount naturally present. The oceans have an enormous buffering capacity (Stumm and Morgan, 1970), so the additional rainout of sulfate (especially after dilution through cloud dispersal and droplet coalescence) seems unlikely to have any effect, even locally, although there is clear disagreement on this point. The principal concern is to avoid additional sulfate deposition over land. With a 30 percent rainout per day, this could be ensured to a 90 percent level by operating about a week upwind of land. Such a constraint would have to be added to the others already stated.

OCR for page 817
Page 833 Another possible way of dealing with the acid rain concern would be to introduce sulfate in the form of ammonium sulfate or bisulfate, each of which is a neutral salt. This would avoid the acid question from the start. Both salts are used frequently as fertilizers and in the dilutions to be seen here would have a mild fertilizing effect locally. These salts can be made by reacting ammonia with sulfuric acid. The price of ammonia is about \$100/t, so the cost of the CCN might double, and there would be an additional cost for equipment to run the reaction at sea. These additional costs might increase the total by as much as 50 percent to \$0.15/t of carbon mitigated per year or \$0.04/t CO2. Necessary Cloud Condensation Nuclei Experiments If global-scale CCN emissions were to be considered in a serious way a number of fundamental studies would need to be performed. Among these would be the following: • Exploratory studies of the effectiveness of CCN for enhancing stratocumulus cloud cover, with a full statistical analysis of covariates, and so on. • Determination of CCN properties: (1) lifetimes of CCN at various altitudes; (2) effectiveness in cloud enhancement; and (3) effect of their precipitation on oceans. • Determination of the fraction of SO2 emissions converted to CCN and the resulting particle size distribution. • Extension of the idea of CCN enhancement from local and regional to global dimensions: a careful study of the scale dependence of the effectiveness of cloud enhancement processes and the interaction of clouds with the radiation field. • Full confirmatory analysis of the effectiveness of CCN on fractional cloudiness with carefully selected test statistics. A multiplicity of analysis would have to take into account all variables such as the humidity profile, convective processes, and CCN count, along with methods for the study of precipitation processes. Note 1. Throughout this report, tons (t) are metric; 1 Mt = 1 million tons; and 1 Gt = 1 billion tons. References Albrecht, B. A. 1989. Aerosols, cloud microphysics, and fractional cloudiness. Science 245:1227–1230. Angell, J. K., J. Korshover, and G. F. Cotton. 1984. Variation in United States

OCR for page 817
Page 834 cloudiness and sunshine, 1950–82. Journal of Climatology and Applied Meteorology 23:752–761. Changnon, S. A. 1981. Midwestern cloud, sunshine and temperature trends since 1901: Possible evidence of jet contrail effects. Journal of Applied Meteorology 20:496–508. Charlson, R. J., J. E. Lovelock, M. R. Andreae, and S. G. Warren. 1987. Oceanic phytoplankton, atmospheric sulfur, cloud albedo and climate. Nature 326:655–661. Du Pont. 1988. Properties of Du Pont Industrial Filament Yarns. Du Pont Fibers Technical Information Multifiber Bulletin X-272. Wilmington, Del.: Du Pont. Hansen, J., and S. Lebedeff. 1987. Global trends of measured surface air temperature. Journal of Geophysical Research 92:13345–13372. Hanson, K., G. A. Maul, and T. R. Karl. 1989. Are atmospheric ''greenhouse" effects apparent in the climatic record of the contiguous US (1895–1987)? Geophysical Research Letters 16:49–52. Henderson-Sellers, A. 1986. Cloud changes in a warmer Europe. Climatic Change 8:25–52. Henderson-Sellers, A. 1989. North American total cloud amount variations this century. Palaeogeography, Palaeoclimatology, Palaeoecology 75:175–194. Hummel, J. R., and R. A. Reck. 1979. A global surface albedo model. Journal of Applied Meteorology 18:239–253. Jones, P. D., T. M. L. Wigley, and P. B. Wright. 1986. Global temperature variation between 1861 and 1984. Nature 322:430–434. Kerr, R. A. 1982. Cloud seeding: One success in 35 years. Science 217:519–521. Manabe, S., and R. T. Wetherald. 1967. Thermal equilibrium of the atmosphere with a given distribution of relative humidity. Journal of the Atmospheric Sciences 24(3):241–259. Ogden, J. M., and R. H. Williams. 1989. Solar Hydrogen: Moving Beyond Fossil Fuels. Washington, D.C.: World Resources Institute. Ramanathan, V., L. Callis, R. Cess, J. Hansen, I. Isaksen, W. Kuhn, A. Lacis, F. Luther, J. Mahlman, R. Reck, and M. Schlesinger. 1987. Climate-chemical interactions and effects of changing atmospheric trace gases. Review of Geophysics 25(7):1441–1482. Randall, D. A., J. A. Coakley, C. W. Fairall, R. A. Kropfli, and D. H. Lenschow. 1984. Outlook for research on sub-tropical marine stratiform clouds. Bulletin of the American Meteorological Society 65:1290–1301. Reck, R. A. 1978. Thermal Effects of Cloud Parameter Variations in the Manabe-Wetherald Radiative-Convective Atmospheric Model. Report GMR-2820. Warren, Mich.: General Motors Research Laboratories. Reck, R. A. 1979. Comparison of fixed cloud-top temperature and fixed cloud-top altitude approximations in the Manabe-Wetherald radiative-convective atmospheric model. Tellus 31:400–405. Reck, R. A. 1984. Climatic Impact of Jet Engine Distribution of Alumina (Al2O3): Theoretical Evidence for Moderation of Carbon Dioxide (CO2) Effects. Report GMR-4740. Warren, Mich.: General Motors Research Laboratories. Rowland, F. S. 1987. Can we close the ozone hole? Technology Review 21:51–58. Schneider, S. H. 1972. Cloudiness as a global climatic feedback mechanism: The

OCR for page 817
Page 835 effects on the radiation balance and surface temperature of variations in cloudiness. Journal of the Atmospheric Sciences 29:1413–1422. Schwartz, S. E. 1988. Are global cloud albedo and climate controlled by marine phytoplankton? Nature 336:441–445. Stumm, W., and J. J. Morgan. 1970. Aquatic Chemistry. New York: Wiley-Interscience. Todd, C. J., and Howell, W. E. 1985. Weather modification. In Handbook of Applied Meteorology, D. D. Houghton, ed. New York: John Wiley & Sons. Twomey, S., and T. A. Wojciechowski. 1969. Observations of geographical variation of cloud nuclei. Journal of Atmospheric Sciences 26:1413–1422. U.S. Bureau of the Census. 1987. Statistical Abstracts of the United States 1988. Washington, D.C.: U.S. Bureau of the Census, U.S. Department of Commerce. U.S. Bureau of the Census. 1988. Statistical Abstracts of the United States 1989. Washington, D.C.: U.S. Bureau of the Census, U.S. Department of Commerce. Warren, S. G., C. J. Hahn, J. London, R. M. Chervin, and R. L. Jenne. 1988. Global Distribution of Total Cloud Cover and Cloud Type Amounts over the Ocean. Report DOE/ER-0406. Washington, D.C.: U.S. Department of Energy. Wigley, T. M. L. 1989. Possible climate change due to SO2-derived cloud condensation nuclei. Letter. Nature (6223) 339:365–367.