SPACE SOLAR POWER REVIEW Volume 3, Number 3, 1982
SPACE SOLAR POWER REVIEW Published under the auspices of the SUNSAT Energy Council Editor-in-Chief Dr. John W. Freeman Space Solar Power Research Program Rice University, P.O. Box 1892 Houston, TX 77001, USA Associate Editors Dr. Eleanor A. Blakely Lawrence Berkeley Laboratory Colonel Gerald P. Carr Bovay Engineers, Inc. Dr. M. Claverie Centre National de la Recherche Scientifique Dr. David Criswell California Space Institute Mr. Leonard David PRC Energy Analysis Company Mr. Hubert P. Davis Eagle Engineering Professor Alex J. Dessler Rice University Mr. Gerald W. Driggers L-5 Society Mr. Arthur M. Dula Attorney; Houston, Texas Professor Arthur A. Few Rice University Mr. I.V. Franklin British Aerospace, Dynamics Group Dr. Owen K. Garriott National Aeronautics and Space Administration Professor Norman E. Gary University of California, Davis Dr. Peter E. Glaser Arthur D. Little Inc. Professor Chad Gordon Rice University Dean William E. Gordon Rice University Dr. Arthur Kantrowitz Dartmouth College Mr. Richard L. Kline Grumman Aerospace Corporation Dr. Harold Liemohn Boeing Aerospace Company Dr. James W. Moyer Southern California Edison Company Professor Gerard K. O'Neill Princeton University Dr. Eckehard F. Schmidt AEG—Telef unken Dr. Klaus Schroeder Rockwell International Professor George L. Siscoe University of California, Los Angeles Professor Harlan J. Smith University of Texas Mr. Gordon R. Woodcock Boeing Aerospace Company Dr. John Zinn Los Alamos Scientific Laboratories Editorial Assistant: Edith R. Mahone
0191 -9067/82/030193-02$03.00/0 Copyright ® 1982 SUNSAT Energy Council EDITORIAL SECOND UNITED NATIONS CONFERENCE ON THE EXPLORATION AND PEACEFUL USES OF OUTER SPACE The conference (UNISPACE '82) is being held 9-21 August 1982, Vienna, Austria. In addition to the plenary and committee meetings of the conference where specific agenda items will be addressed, technical presentations by nongovernmental organizations will be made in conjunction with UNISPACE '82. Under the auspices of the International Committee on Energy from Space, a symposium has been organized which will be held on August 11 and 12. The purpose of the symposium is to present information to the delegates to UNISPACE '82 on the Framework for Energy from Space, the Technology and Its Implications, Steps Towards Implementation, and Technical Issues. An important agenda item at the symposium will be the discussion of the following resolution. UNITED NATIONS CONFERENCE ON OUTER SPACE NGOs AT UNISPACE '82 INTERNATIONAL COMMITTEE ON ENERGY FROM SPACE Resolution Recognizing that • The world is facing formidable challenges posed by population pressures and the depradation of nonrenewable energy, resulting in environmental degradation and resource exhaustion. • Huge energy supplies are needed on a global scale for developing nations to approach the economic levels of industrialized nations. • The global demand for energy requires a mix of energy resources and conversion technologies for both decentralized and centralized systems during the transition to renewable sources. • Significant and continuing efforts will have to be devoted to the development of new and renewable energy sources to meet future energy requirements. • There are no limits to the evolution of planet Earth's civilization if the limitless energy and material resources available in space are used for the benefit of humanity. Further, considering that • Significant advances are resulting from space missions which are leading to further improvement in the use of satellites for telecommunication, Earth observation, and for scientific studies of the solar system and the planets. • Plans for industrial uses of space and a permanent manned presence in space represent a new direction for the use of extraterrestrial resources. • As a result of expanded orbital operations and consequent emerging opportunities to overcome the impending crises represented by finite global resources, destruc-
tive weapons and unmet requirements, the time has come for the community of nations to plan together to develop technologies capable of further utilizing the energy and material resources of space, to consider new enterprises which could lead to peaceful uses of these resources for the benefit of mankind and to bequeath to future generations the means to build a better world. Therefore, be it resolved that • Obtaining energy from space for use on Earth is of common interest to both developed and developing countries and is one of the major options for obtaining renewable energy sources. • The United Nations and its competent agencies should in their deliberations consider how energy from space could meet the requirements of interested nations and especially developing nations. • Interested nations working in cooperation with the United Nations and its competent agencies should seek to establish a framework for an international program to develop technologies for the beneficial uses of energy from space on Earth, and to provide a forum for the exploration of the technical, economic, environmental and societal issues associated with the development of such technologies. • The United Nations Committee on the Peaceful Uses of Outer Space (COPUOS) should establish a task group to obtain all necessary information to enable COPUOS to plan and encourage cooperative programs on energy from space between developed and developing countries, consult with organizations developing technologies for obtaining energy from space and report on the progress achieved to the United Nations and its competent agencies and other interested public bodies. The resolution will be submitted to the delegations of countries participating in UNISPACE '82 for their consideration. Peter E. Glaser
0191 -9067/82/030195-27$03.00/0 1982 SUN SAT Energy Council THE SATELLITE POWER SYSTEM: ASSESSMENT OF THE ENVIRONMENTAL IMPACT ON MIDDLE ATMOSPHERE COMPOSITION AND ON CLIMATE R. C. WHITTEN, W. J. BORUCKI, C. PARK, L. PFISTER, and H. T. WOODWARD Ames Research Center, NASA Moffett Field, California 94035, USA R. P. TURCO R&D Associates Marina del Rey, California L. A. CAPONE and C. A. RIEGEL San Jose State University San Jose, California T. KROPP Informatics, Inc. Palo Alto, California Abstract — The heavy-lift launch vehicles (HLLVs) proposed for use in constructing a satellite power system (SPS) would deposit various contaminants in the middle atmosphere, contaminants that could have adverse effects on upper air structure and climate. These contaminants consist primarily of water vapor, hydrogen, carbon dioxide, carbon monoxide, and traces of sulfur dioxide and nitric oxide. Large quantities of nitric oxide are also formed during reentry. To assess the effects of such effluents, we constructed new models or modified existing models of the upper atmosphere: a one-dimensional and a two-dimensional photochemical model, a rocket plume model, and a reentry model. All are described here. Using an SPS scenario of 400 launches per year for 10 years, our calculations lead to the following conclusions: (1] the buildup of water vapor, nitric oxide,-CO2, CO, or sulfur dioxide, including a possible “corridor” effect (zonal enhancement centered on the launch latitude) is not likely to be significant; [2] ozone perturbations should not be significant — the ozone total column density decreases would probably be less than 0.1%; [3] although significant perturbations of odd-hydrogen (H, OH, HO2) are not predicted for the stratosphere and mesosphere, thermospheric hydrogen could be doubled in concentration; and [4] with respect to climate, none of the SPS-induced changes mentioned would lead to measurable changes in climate. 1. INTRODUCTION Heavy-lift launch vehicles (HLLVs), which have been proposed for use in the construction of satellite power systems, would be about 2 to 5 times as large as a Space Shuttle (1). Because of the large projected launch frequency (400 per year), modifi-
cations to the composition and perhaps the temperature structure of the stratosphere and mesosphere (middle atmosphere) might be expected. The purpose of the study reported here is to assess possible effects on the middle atmosphere of continued HLLV activity for 10 years or more. The principal emissions of the liquid-fueled HLLV launch motors that might modify atmospheric composition are water vapor, hydrogen, and carbon dioxide (Table 1); copious amounts of nitric oxide are also formed during reentry. In addition, small amounts of nitric oxide and sulfur dioxide are expected to be formed in the combustion process and to be deposited during launch. Figure 1 presents water vapor and hydrogen deposition profiles for a single launch in units of molecules per centimeter; note the discontinuity at an altitude of 56 km (35 mi), where the first stage, which uses methane as fuel, cuts off and the second stage, which used hydrogen as fuel, ignites. Also note the large increase in the deposition rate as the vehicle approaches an altitude of 120 km (75 mi); this is a result of nearly horizontal flight near the end of the trajectory and burnout at 120 km. Carbon dioxide, a product of methane combustion, is also emitted up to first-stage burnout; its deposition rate, by number, has been estimated to be 0.4 of the water vapor rate. The emission rate of nitric oxide during the launch phase is also shown in Fig. 1. The emission rate of sulfur dioxide is estimated by assuming that the first-stage fuel contains 0.05% sulfur (K. L. Brubaker, Argonne National Laboratory, private communication, 1980). Because the production of nitric oxide during reentry is the subject of a rather lengthy calculation, we defer its discussion to a later section. The complexity of the physical processes occurring in the middle atmosphere requires mathematical models to obtain quantitative estimates of the effects of HLLV operations. General circulation models that consider the interactions of radiation, chemistry, and dynamics are beyond current capabilities. Hence, a range of more limited models must be used to describe each area of concern. In the following sections, we discuss the atmospheric photochemical models used in this assessment, including a model for the calculation of nitric oxide production during HLLV reentry. We then apply the models to the assessment of HLLV effects on the middle atmosphere, evaluating possible global effects and short-term local effects. Specialized topics (i.e., “corridor” effects and contrail and cloud formation) are treated in accompanying papers. Finally, we summarize our conclusions and identify those effects that could be significant and thus require further study.
2. ATMOSPHERIC PHOTOCHEMICAL MODELS To carry out the required assessments most efficiently and realistically, two photochemical models are used: a one-dimensional (1-D) model and a two- dimensional (2-D) model. The 1-D model, which contains over 50 species, is most useful for investigating globally or hemispherically averaged changes in composition; it is also capable of simulating a much larger system of species and reactions than the 2-D model, which is limited to about 25 chemical species at altitudes below 55 km (34 mi) and 9 species above 55 km. The 2-D model, on the other hand, can simulate meridional variations in composition changes and is thus able to treat possible seasonal and “corridor” effects; the latter refers to enhanced changes in a latitudinal zone containing the launch or reentry windows. In the present 2-D model, the bulk velocities were obtained by zonally averaging the wind fields predicted by three- dimensional dynamic models; the eddy diffusion tensor elements were obtained by an adjustment procedure based on matching computed tracer distributions with their observed distributions. Because both models have been described in detail in the literature (2-7), only a brief description is given of them here. Both models are based on the continuity equations for trace constituents, where P, and L, are production and loss rates, respectively, for the /th constituent, v is the meridional bulk velocity, </>, is the “eddy flux” representing large-scale eddy motions, and nt is the number density of the zth constituent. In the 1-D model the velocity v is, of course, absent, and the eddy flux is written as
where ez is a vertical unit vector, Ke is the eddy diffusivity, and T and H are the temperature and mean atmospheric scale height, respectively. Because the 1-D model extends into the lower thermosphere (i.e., to an altitude of 120 km), molecular diffusion cannot be neglected; a term 0? representing molecular diffusive flux must therefore be added to </>,: mt and m are the molecular mass of the /th constituent and mean molecular mass, respectively, k is the Boltzmann constant, and g is the gravitational acceleration. The 2-D model, which extends only to an altitude of 90 km (56 mi), ignores molecular diffusion. However, because it is a tensor, the eddy diffusion coefficient is more complicated than in the 1-D model; that is, Ke contains “off-diagonal” elements which are important below 25 km (16 mi); the eddy flux is then written The chemical rate coefficients used in the model are essentially those tabulated by Hudson and Reed (8) except that the photolysis rate of NO is taken from the work of Nicolet (9) and Nicolet and Cieslik (10) and three of the odd-hydrogen reaction rate coefficients have been changed (see Sec. 5). Furthermore, we include a source of mesospheric N(2D) (first excited state of N) due to the ionization of NO by solar Lyman-a radiation: In the model, diurnal averaging of photodissociation rates and reaction rate coefficients is performed using, respectively, the techniques reported in Refs. 11 and 12. For each species we specify a lower boundary flux, Ojo, and an upper boundary flux, <I>f„, which may be fixed or have a specified time dependence; at the lower boundary we also include a flux component, by defining the “velocity” at the boundary, vBj, and concentration, nB. (if nB. = 0, our code sets 0B. = 0 automatically); nt is computed by the model. When we use boundary condition specified in Eq. 7, we set the boundary velocity, vB., equal to 1 cm/sec, which is estimated from typical tropospheric mixing rates. A justification for this approach, together with its mathematical application, is given by Turco and Whitten (3). Upper boundary conditions are imposed in a somewhat similar manner, except that allowance must be made for the escape of hydrogen. The mean escape flux of H atoms is about 108 cm-2 sec-1 (13), which must be included in the computa-
tions of both H and H2. At 120 km (75 mi) both H and H2 have upward fluxes. Therefore, we specify the upper boundary fluxes as The constant h is determined from simultaneous estimates of the escape flux and nH, and the constant n is estimated from the flux of H2 at 120 km (75 mi). This H2 is decomposed into H above 120 km. In Fig. 2, 1-D-model predictions for the ambient atmosphere are compared with observations. The agreement between the mesospheric prediction of atomic oxygen, O(3P), and its observation is reasonable, considering the large zenith angle of observation; the computed concentrations are perhaps too large by a factor of 2 between 90 and 100 km (56-62 mi). Computed ozone concentrations are generally within the error bars of the Krueger-Minzner empirical model (18), although the high-altitude [above 45 km (28 mi)] predicted concentrations are somewhat too small. Predictions for OH and water vapor appear to lie within the range of observed values; however, it should be remarked that the range of observed water vapor is quite broad, extending from as low as 2 parts per billion by volume (ppbv) to as high as 7 ppbv (20). Also, the OH observations were made at large solar zenith angles whereas the computed values are for average midlatitude daytime conditions; to correspond to the same conditions, the computed values would have to be reduced by a factor of ~2. The predicted nitric oxide abundance in the stratosphere lies well within the range of measurements. However, in the upper mesosphere and lower thermosphere, there is disagreement, which may be due in part to the large zenith angle of the observations and in part to adjustments needed in the mesospheric NO photochemical source [production of NO from N(2D) may have been overestimated] and sink and in the diffusion rates at these heights. The 2-D model extends from the surface to an altitude of 90 km (56 mi) in 2.5-km (1.6-mi) steps and from latitude 80 °N to latitude 80 °S in 5° steps; the time steps are fixed at 1 day. End boundaries are taken at latitude 80 °S and latitude 80. °N because meridional fluxes are expected to be small at those latitudes. Hence, the end boundary conditions are taken to be zero flux of all constituents across the vertical boundaries. The upper boundary conditions are given by setting the fluxes equal to zero for all species except H, NO, and O(3P). We have checked the upper boundary conditions [at 90 km (56 mi)] for the various species against results from a one-dimensional model that extended up to 120 km (75 mi). It was found that the effect of choosing a mixing equilibrium condition at the upper boundary had very little effect on any of the constituents at altitudes below 50 km (31 mi). The lower boundary condition used for all species, except N2O, CH,, HNO:S, NO2, O3, HC1, H2O2, and the halocarbons, is chemical equilibrium because of their short lifetimes against chemical loss. Because HNO:1, HC1, and H2O2 are water soluble, their number densities are set equal to zero at the lower boundary. The number densities of CH.(, NO2, and N2O are fixed at
3.7 x 1013 cm-3, 3 x 109 cm-3, and 7.5 x 1012 cm-3, respectively, at the lower boundary, and that of O3 is fixed at 6 x 10" cm-3 in order to conform to measured values. To assess the influence of ozone on the temperature structure (and thus the chemical composition) of the upper stratosphere and lower mesosphere, we have included a heating and cooling code, which is not described in our earlier papers. The heating algorithm is developed from the work of Lacis and Hansen (21). Their method, a parametric treatment based on accurate multiple-scattering computations, includes the effects of Earth's albedo. (The planetary albedo is assumed to be constant and equal to 0.34; see Ref. 22, p. 2-2.) The vertical distributions of ozone both above and below the calculation point are accounted for in the calculation of the absorbed and scattered solar radiation. Because the heating rate is dependent on the zenith angle and on the number of hours of daylight, diumally averaged heating rates are calculated, as described by Cogley and Borucki (11). The heating rates are calculated at each latitude and altitude every 7 days so that the changing solar position and number of hours of daylight are taken into account. The cooling rates are computed from a Newtonian cooling model (23) such that where R is the cooling rate (in degrees per day), AT is the model-calculated temperature minus the standard atmosphere reference temperature, and A(0,P) and B(P) are latitude-dependent (0) and pressure-dependent (P) parameters. The were determined by requiring the radiative cooling to exactly balance the radiative heating for the standard reference temperature (AT = 0) and ozone profiles. The B(P) were taken from the work of Dickinson (23). In Fig. 3, 2-D-model predictions for the ambient atmosphere (autumn, latitude 40 °N) are shown. The agreement with observational data is about the same as that obtained with the 1-D model, except for NO; the mesospheric concentrations appear to be about an order of magnitude too small between 55 and 85 km (34-53 mi). We attribute the low predicted values to the absence of a mesospheric source of NO |see processes (6)] in the model and to slow upward transport of NO from the stratosphere; the vertical eddy diffusivity in the mesosphere is probably a factor of 2 to 3 too low, a state of affairs that is not readily correctable because of limitations in calcula- tional stability. To complement our one-dimensional and two-dimensional analyses of the widespread photochemical effects caused by rocket emissions, we utilize a simple model of an expanding rocket plume to study potentially important local effects of the rocket plume. The model is applied for both rocket launch and reentry events, and is predicted on several approximations. The simulated launch-reentry plume is oriented vertically and has cylindrical symmetry. The 1-D model is used to sample the entire plume, although the plume may be truncated in altitude extent to simulate a flight trajectory. The plume is given an initial width corresponding to a short period of expansion. The rocket ejecta at each height are averaged over the initial area of the trail, and these concentrations, when added to the ambient concentrations, represent the starting conditions for the calculations. The area of the rocket trail is assumed to increase with time according to the relation (24) where A is the area (cm2), K is the effective horizontal diffusion coefficient (cm2
sec-1) and time, t, is measured from the initial expansion time of the plume. Obviously, the initial area and time can be related by Au = (n/2)Kt0. The constituents under study are always taken to be uniformly distributed across the plume. For a chemically active species in a uniform expanding plume, we can estimate the rate of change in the average plume concentration due to expansion as where na is the ambient concentration. Equation 12 presumes that the plume expands by mixing in air containing an ambient abundance of each species, and that mixing within the plume is instantaneous. If K is constant, Eq. 12, by virtue of Eq. 11, becomes The expansion term, Eq. 13, is readily incorporated into the model calculation as a “pseudochemicar' loss process.
3. CALCULATION OF NOX PRODUCTION DURING HLLV REENTRY The average mass of the HLLV at reentry is 452.6 tonnes (499 tons) (25). The total projected area of the craft is determined from the figures in Ref. 25 as approximately 2280 m2 (24,530 ft2); the initial orbit height of the craft is given as 450 km (280 mi). Reference 25 also specifies that the deceleration of the HLLV during reentry shall be less than 1.4 g. The report does not specify, however, the reentry flight trajectory or the lift and drag characteristics of the gliding body needed in a trajectory calculation. Presumably, such aerodynamic data have not yet been generated. In order to estimate the amounts and spatial distributions of NOX produced during the reentry flight, therefore, the following assumptions have been made. [1] The functional relationship between the drag coefficient and the angle of attack of the craft is the same as for the Space Shuttle Orbiter, as given in the design report (26). [2] The lift coefficients of the vehicle are expressible as a product of the lift coefficient values of the Space Shuttle Orbiter given by Ref. 26 and an arbitrary fixed constant not greater than unity. Controlling lift coefficients in this manner would be possible, for instance, by varying the angle of wing flap. [3] The angle of attack remains fixed at a high value initially. When the craft reaches a certain intermediate altitude [of about 70 to 80 km (43-50 mi)], the angle of attack is changed at a steady rate to a low value.
[4] The orbit of the HLLV is inclined 45° to the equator. [5] The end point of the entry flight defined as the point where the altitude is 50 km is latitude 40 °N and longitude 120 °W, so that the craft could land on an airfield in central California. [6] The atmosphere rotates with Earth without any slip. The entry trajectory can be varied by varying [1] the duration of the retrorocket bum at the orbit (the amount of fuel required for this purpose would be very small, and so its mass is ignored), [2] the lift coefficient (i.e., flap angle), [3] the initial angle of attack, [4] the final angle of attack, |5] the altitude at which the angle of attack begins to decrease, and [6] the rate at which the angle of attack is changed. In the present work, these variable parameters were changed over a wide range in order to generate a family of possible entry trajectories that satisfies the condition of maximum allowed deceleration of 1.4 g. The calculation requires a straightforward numerical integration of the Newtonian equations of motion in two dimensions. Earth was assumed to be a perfect sphere, and the atmospheric data of the U.S. Standard Atmosphere (27) were used for the calculation. Figure 4 shows the profiles of such flight trajectories. As indicated, the trajectories are produced by varying retro-AV, that is, the velocity decrement achieved by firing the retrorockets, between -0.105 and -0.110 km/sec (-344 and -360 ft/sec). At the altitude of 120 km (75 mi), these two AV values produce entry velocities of 7.588 and 7.583 km/sec (4.715 and 4.712 mi/sec), and entry angles of -0.896° and -1.103°, respectively. Of these, three typical trajectories were selected for computation of the NOj. production rate. According to the distance required for the entry flight to end, they are referred to here as “shallow,” “nominal,” and “steep” trajectories, as shown. The rate of NOZ production along these trajectories was calculated by the method of Park (28). In brief, the method assumes that the amounts of NO^ produced by a spacecraft are the same as that produced by a circular cone having the same overall chord length, overall surface area, and angle of attack. Such a method is shown (28) to yield approximately the same amount of NO^as a real spacecraft with a blunt nose, at least up to the point where such an “exact” calculation was possible. The computer code used by Park (28) was run in the present study. The code calculates
the thermodynamic variables simultaneously with the chemical rate equations to compute the concentrations of N and NO in the flow around a given cone. The computation is carried out until the flow leaves the trailing edge of the spacecraft. Then the assumption is made that the NO molecules are chemically frozen, but that all N atoms convert to NO in the wake. This assumption slightly overestimates the total NO.r concentration in the wake because it ignores the mutual destruction process in which N reacts with NO. But such an error would be small because the total concentration of N is quite small, and there are far more O2 than NO molecules in the wake. The calculation yields the number of NOj. molecules deposited along the wake per unit length of the wake for each entry flight. By integrating the number of the molecules along the trajectory, one obtains the integrated (regionally or globally) number of NOj. molecules produced. The overall efficiency of NOj. formation is calculated from the relation
where e is the formation energy of one NO molecule (e = 1.515 x 10-12 erg or 0.95 eV), M is the mass of the HLLV, and V, is the entry velocity. To be consistent with the two-dimensional atmospheric model adopted in the present work, the global atmosphere was partitioned by the latitude increment of 5° and the altitude increment of 2.5 km (1.6 mi). Figure 5 shows the distribution of NO^ by altitude. The total amounts of NOj. produced per reentry, in terms of the mass equivalent of NO, and the efficiency of conversion of the initial kinetic energy into NO are given in Table 2. The efficiency values obtained here are nearly equal to those for the Space Shuttle Orbiter (28) and for large meteoroids (29). Figure 6 shows the production of NO molecules as a function of altitude and latitude for the nominal entry case (45° angle of attack). 4. THE ACCUMULATION AND DISPERSION OF ROCKET EXHAUST CONTAMINANTS The models described above have been used to simulate the accumulation and dispersion of the substances deposited during HLLV launch (water vapor, nitric
oxide, carbon dioxide, carbon monoxide, sulfur dioxide) and created during reentry (nitric oxide). In order to simulate the long-term effects, the 1-D and 2-D photochemical models were run for 10 years of continuous HLLV operations at the rate of 400 launches per year with the contaminants averaged over different horizontal extents (Northern Hemisphere, the globe, etc.). The 1-D model simulations appeared to have reached a steady state. The 2-D model simulations had closely approached a steady state in the mesosphere, but had not quite attained steady state in the stratosphere. The 1-D model directly simulated water vapor deposition at all altitudes up to 120 km (75 mi). Water vapor deposition at higher altitudes was simulated by inserting it into the model at 120 km altitude from where it could be transported downward or converted photochemically to H2 and flow upward as well as downward. The hydrogen could then escape as discussed below. In the 2-D model, which did not compute H2, all water deposited above 90 km (56 mi) altitude was assumed to be injected at 90 km; the upper boundary condition was constructed such that water vapor could move across the boundary, seeking diffusive equilibrium of high-altitude water vapor (i.e., above 90 km) with the water vapor below. The short-term injection effects are treated with the plume (dispersion) model discussed in Sec. 2. Rocket contrails are treated in a companion paper (30). In the present section we discuss the widespread accumulation of ejecta under quasi-steady-state conditions and the dispersion reentry nitric oxide plumes for single events. Water Vapor The predicted increases in the water vapor abundances obtained with the 1-D and
2-D photochemical models for 10 years of HLLV operations are shown in Fig. 7. It is apparent from the results of the 2-D model that a significant "corridor effect” (enhanced accumulation in the vicinity of the launch latitude) is likely only at altitudes above 75 km (45 mi). At an altitude of 85 km (the highest altitude at which the 2-D model yields reliable results) the computed corridor enhancement is —50%. Corridor effects will be discussed in detail elsewhere. The globally averaged increase in H2O obtained with the 1-D model is shown by the broken curve in Fig. 7. In the stratosphere [i.e., at altitudes below 50 km (31 mi)] the two models predict nearly the same increase in water vapor. However, between 50 and 75 km (31-47 mi) (lower mesosphere), the increase obtained with the 2-D model is about double that obtained with the 1-D model. The difference is a consequence of the vertical “eddy diffusion” coefficient in the 2-D model being somewhat smaller than that in the 1-D model. On the other hand, the actual value of that coefficient is uncertain and our results give an indication of the effects of this parameter on the H2O prediction. Many meteorologists believe that the influx of water vapor to the stratosphere is controlled by the low-temperature “cold trap” of the tropical tropopause (16,19,31,32). The mechanisms that remove water vapor from the stratosphere are still unknown, but are probably dominated by subsidence at midlatitudes and high latitudes; we have made such an assumption in our model calculations. Another but less important loss mechanism for water vapor is photochemical conversion to relatively inert molecular hydrogen in the mesosphere, mainly through the reaction sequence some of the H2 diffuses to the troposphere, and some is dissociated at high altitude, contributing to the hydrogen escape flux to space (for discussion of the escape mechanisms, see Ref. 33). According to recent measurements the hydrogen escape rate is —108 atoms cm-2 sec-1 (34). An escape flux of this magnitude was imposed as an upper boundary condition on the hydrogen in our 1-D model. Molecular hydrogen may also be converted back into water vapor by reactions such as However, in the mesosphere, H2 is produced at the expense of H2O, and diffuses into the stratosphere and eventually the troposphere. Our results are in substantial agreement with those of Forbes (35). The effect of the deposition of water vapor and hydrogen at high altitudes [up to 120 km (75 mi)] is a substantial buildup of thermospheric molecular and atomic hydrogen. The H2 dominates below about 120 km and the H above 120 km (e.g., see Refs. 34 and 36). The computed increase in total hydrogen above 100 km (62 mi) (globally averaged) is about a factor of 2 (i.e., a doubling of thermospheric H and H2).
The implication of such a large increase in worldwide thermospheric hydrogen is beyond the scope of the present study, but must be boldly underscored. It has been suggested to the authors by B. M. McCormac (Lockheed Palo Alto Research Laboratory, private communication, 1980) that the population of ring-current protons in the magnetosphere could be substantially altered by charge exchange. Nitric Oxide Nitric oxide deposited in the stratosphere during the launch phase of the HLLV leads to an increase in the global NO abundance of less than 1% over ambient concentrations for 400 launches per year. Nitric oxide is not normally produced in the mesosphere. Rather, it is produced by photochemical processes in the stratosphere and the thermosphere and then transported upward or downward, respectively. The stratospheric source results from the reaction of O('D) with nitrous oxide, and thermospheric nitric oxide is produced through the dissociation of N2, mainly via ionic processes. Intermittent direct mesospheric sources include meteoritic bombardment and auroral activity. However, NO created in the mesosphere and lower thermosphere during HLLV reentry can lead to increases in its upper-atmospheric abundance. The relative magnitude of the increase depends on the background NO concentrations, which are quite uncertain. Figure 8 shows the predicted absolute increases in NO concentrations caused by 10 years of HLLV launches and reentries on the nominal trajectory at the rate of 400 per year. These NO increases are less than 40% at all heights. The “corridor effect” for NO is much more pronounced than was the case for water vapor. The reason for this can be expressed as follows: nitric oxide has a rather short lifetime in the mesosphere, about 4 days at an altitude of 70 km (43 mi) and even shorter at higher altitudes. This is a consequence of rapid photolysis by solar ultraviolet radiation (9,37), followed by the odd-nitrogen destruction reaction, Of course, reaction 25 must compete with the NO recycling reaction whith inhibits NO loss; the rates of reactions 25 and 26 are comparable at an altitude of 70 km (43 mi). The NO lifetime can therefore be expressed as In Eq. 27, k denotes a rate coefficient and J a photolysis rate; values of the reaction rate coefficients were taken from Hudson and Reed (8) and photolysis rates from Nicolet (9). Since the concentration of O2 decreases rapidly with ascending altitude, the lifetime also decreases, reaching an asymptotic value of ~2 days at 90 km (56 mi). One can account for the relatively strong enhancement of NO in the region of rocket reentry by noting that the NO is likely to be destroyed by the sequence of reactions
24 and 25 before it can be transported meridionally over a significant distance. For the transport parameter values used in the 2-D model, the mean distance through which material is transported meridionally in 4 days in only —400 km (249 mi). The mesospheric source of nitric oxide includes photochemical processes and reactions that dissociate N2, both EUV and X-rays (38), as well as solar protons, meteorites (28), etc. The 1-D model includes such sources of NO below 120 km (75 mi) and a flux source due to photochemistry of —108 cm“2 sec-1. The NO predictions from the 1-D and 2-D models are in rough qualitative agreement, but differ in details. Because it is not possible to simulate the reentry trajectory in the 1-D model, substantial differences between the results from the two models are expected. Nitric oxide also cools the thermosphere in the 5.3-p.m emission band (39), but the amount of NO produced above 120 km (75 mi) by HLLV reentry is very small, and this effect is expected to be negligible. We have also simulated the dispersion of an NO reentry trail with the aid of the plume model discussed in Sec. 2. Figure 9 illustrates the dispersal of NO at points along the reentry flightpath where NO is generated in large quantities (see Figs. 5 and 6). It is important to note that the profiles in Fig. 9 represent sections along a trajectory that is inclined to the vertical. The horizontal eddy diffusivity employed for the calculations was 108 cm2 sec-1; actually, K is scale-dependent and hence is time-dependent (40), its effective value increasing with time. For periods of less than 1 day, K ~ IO8 cm2 sec-1 is a reasonable value, but it increases rapidly thereafter,
with the result that the trail undergoes rapid dispersal to concentrations approaching ambient values. The effect of such large NO enhancements on ionospheric structure will be discussed elsewhere. Carbon Dioxide and Carbon Monoxide The first stage of the HLLV bums methane fuel and carbon dioxide is a principal product of combustion. Since the methane molecule contains carbon and hydrogen atoms in the ratio of 1 to 4, assuming stoichiometric combustion, one expects the volume rate of CO2 emission to be one-half that of water vapor at altitudes below 56 km (35 mi). Actually, according to Ref. 1, the combustion is not stoichiometric; the CO2 volume emission rate is about 0.4 that of water vapor and that for CO is about 0.3 that for water (1). Although CO2 deposition ceases at an altitude of 56 km (35 mi), it is effectively mixed upward from that height; diffusion to the troposphere is a slower process. As a result, the absolute increase in the CO2 content should roughly approximately one-half that for water vapor. Figure 10 shows the computed change in the mixing ratio of CO2 for several altitudes. The increase in the mixing ratio of CO2 at an altitude of 50 km (31 mi) ranges from about 3 x 10“3 ppmv to about 7 x 10-3 ppmv, which should be compared with the mixing ratio in the ambient atmosphere of ~300 ppmv. The fractional change is only about 1 to 2 parts in 105,
which seems to be much too small to cause a significant increase in the middle atmospheric cooling rate. If all of the cooling in this region were due to emission of infrared radiation by CO2, a very rough estimate of the resulting temperature change is ~10-3°K or less. The heating-cooling models described in Sec. 2 were used to confirm this estimate. Carbon monoxide emitted by HLLVs cannot accumulate like CO2 because CO is rapidly oxidized to CO2. Afterburning in the rocket plume should convert most of the emitted CO into CO2 (K. L. Brubaker, Argonne National Laboratory, private communication, 1980). Otherwise, the CO will mix into the atmosphere and be oxidized, principally by reacting with OH, with a rate coefficient —1.4 x 10-13 cm3 sec-1 (8) corresponding to an upper- atmospheric lifetime of tc0 < 1 month. Reactions of CO with O2, O3, HO2, etc., are extremely slow and can be ignored. Inasmuch as stratospheric OH concentrations are as large as — 1 x 107 molecules cm-3 (15), the lifetime of injected CO could be as short as 2 weeks, assuming that OH levels are not greatly disturbed. Carbon monoxide concentration increases would have to exceed 10'° molecules/cm~3 to affect OH concentrations [background amounts of CO in the stratosphere vary from —2 x 1010 molecules cm 3 near 20 km (12 mi) to — 1 x 109 molecules/cm near 40 km (25 mi), based on recent model calculations]. Thus, if significant amounts of CO remained in the rocket plume, the OH abundance near the plume could be reduced by reaction (28). However, OH would also be influenced by the large quantity of water contained in the plume, which might compensate the CO effect. Because the plume would be rapidly diluted with ambient air as it disperses, the OH reaction cycle would begin to deplete CO quite efficiently, perhaps at a rate approaching (or locally exceeding) the global oxidation rate(s) estimated above. It follows that CO perturbations by satellite power system (SPS) rockets will very likely be negligible. This is confirmed by
calculations with the 1-D model that for the globally averaged case, indicate that the CO abundance would be increased by less than 10%, with no effect on the distributions of other constituents. Sulfur Dioxide It has been estimated (K. L. Brubaker, Argonne National Laboratory, private communication, 1980) that the first-stage fuel of the HLLV could contain 0.05% sulfur by mass. The yearly deposition of sulfur dioxide in the stratosphere is therefore estimated to be about 200 tonnes yr-1 (220 tons yr-1) or less, as compared with an estimated ~ x 105 tonnes yr-1 (1 x 105 tons yr-1) of SO2 produced in the stratosphere from COS of tropospheric origin (41). Based on the work reported by Turco et al. (41), we estimate that the resulting fractional increase in the mass concentration of the stratospheric aerosol layer would be negligible and not affect Earth's albedo and mean surface temperature. 5. PERTURBATIONS TO ODD-OXYGEN AND ODD-HYDROGEN The computed accumulations of excess water vapor and nitric oxide lead to predicted changes in the distribution of odd-oxygen and odd-hydrogen (HOJ species. The odd-hydrogen concentration is enhanced mainly through the reactions where water vapor photolysis occurs at high altitudes [above 70 km (43 mi)]. The odd-hydrogen generated is distributed among several species, including H, OH, HO2, and, at low altitudes, H2O2 and HNO3. Our computations show that the increases in HOj. abundances are significant only in the mesosphere and thermosphere; in the stratosphere, concentrations of OH and HO2 increase only about 0.5% at 50 km (31 mi), and decrease with descending altitude. Of course, such stratospheric perturbations will influence the ozone column density somewhat, as is discussed below. Figure 11 shows computed absolute increases in HOX species concentrations. The calculations were carried out with the 1-D model for globally averaged conditions, which are close to the 2-D results. In the region near the mesopause, excess injected atomic hydrogen may influence the thermal balance. The chemiluminescent reaction with ozone, would enhance hydroxyl (Meinel) vibrational band emission and cool the atmosphere. More importantly, hydrogen catalyzes odd-oxygen recombination through reaction 30 followed by The net reaction cycle releases as heat the chemical energy in atomic oxygen, providing an additional local heating source. We have not been able to quantitatively assess
the importance of this heat source because of complications involving turbulent heat transfer. In the mesosphere and upper stratosphere, odd-oxygen species, O3, O(3P), and OC/)), are strongly influenced by catalytic reaction sequences, in addition to the Chapman oxygen reactions. Important cycles are The stratosphere is more complex than the mesosphere, but several key odd-oxygen reactions cycles have been identified: Interestingly, nitric oxide deposited in the lower stratosphere may lead to ozone increases (42) through the decomposition of peroxy compounds, for example, The critical reaction 35 also interferes with the odd-oxygen destructive cycles 16 and 31, and 16 and 32. Recently, important new rate-coefficient data have been obtained for several key HOj. reactions. The data are listed in Table 3. Accordingly, the importance of the odd-oxygen generating reactions 35 to 37 has greatly diminished. We have computed ozone column changes due to water vapor and nitric oxide injection by HLLV launch and reentry operations. The computed globally averaged
ozone reductions due to water vapor deposition (400 launches per year) are less than 0.01% with no apparent corridor effect. This is less than one-tenth the ozone reduction computed for the Space Shuttle (48). Similar calculations have been made for nitric oxide during atmospheric reentry, which yield globally averaged reductions of about 0.02% to 0.04% (depending on the season); the corresponding result obtained with the 1-D model if ~0.02%. It is important to note that the 2-D model simulates horizontal transport, whereas the 1-D model does not. Also, there are differences in the way in which the mean sun angle is estimated (the 1-D model uses a mean solar zenith angle that is time-invariant; in the 2-D model, the solar zenith angle is latitude- and time-dependent) and in the lower boundary conditions. In the 2-D model, it turns out that NO is efficiently transported downward into the stratosphere where it is not
photodissociated. The computed latitude variation in total ozone is shown in Fig. 12 for two seasons. We note that the predicted changes are highly uncertain at this time (49). Figure 13 shows corresponding changes in the ozone concentration as a function of altitude. Figure 14 shows the computed changes in ozone and atomic oxygen concentrations above 40 km (25 mi). There is reasonably good agreement between the globally averaged 1-D model results and the 2-D model results. The predicted changes in mesospheric odd-oxygen species do not seem to be especially important, except perhaps with respect to chemical heating near the mesopause; reductions in O(3P) at the mesopause may lead to slightly lower mesospheric temperatures, but the magnitude of the cooling, although difficult to estimate is probably small (recall that the O concentration is smaller because the H concentration is larger, so that the overall rate of oxygen recombination via reaction cycle 17 and 30 is not altered significantly). 6. POSSIBLE CLIMATIC EFFECTS OF LARGE ROCKET OPERATIONS Changes in the mixing ratios of radiatively active minor constituents caused by HLLVs could conceivably alter atmospheric heating rates and produce climatic changes (climate here is defined narrowly as the time-averaged meteorological state at Earth's surface). At least a 10-year average would be required to obtain a representative climate that is not polluted by sampling error. Since compositional changes due to HLLVs would be confined to levels above the midstratosphere, we must look for mechanisms that could couple these higher levels with the troposphere. Such mechanisms could be radiative or dynamical. Radiative mechanisms involve perturbations of upper atmospheric composition or temperature that affect heating at the
surface, either through changes in the extinction of visible solar radiation or through changes in the transfer of infrared radiation. Dynamical mechanisms involve high- level circulation changes, including changes in reflection and absorption of planetary waves, that force changes in tropospheric winds, pressure, and temperature. Radiative mechanisms for climatic change have been examined in a number of radiative-convective one-dimensional models, including those of Ramanathan et al. (50), Schneider and Coakley (51), and Manabe and Wetherald (52). Such models examine the globally averaged radiative balance for various atmospheric compositions, with dynamics essentially excluded, except for the constraint that vertical temperature gradients cannot exceed some critical value, sometimes taken as the adiabatic lapse rate. An examination of the results of these models indicates that the
most important changes in composition associated with HLLVs, namely, decreases in ozone and increases in H2O and NO^., would not have a significant effect on climate, at least as measured by changes in surface temperature. For example, Ramanathan et al. (50) computed a surface temperature increase of only 0.06 °K for a 10% increase in stratospheric water vapor. Since HLLVs are expected to produce at most a 0.5% increase in water vapor, their globally averaged effect through this mechanism can be ignored. Ozone decreases should be even less important, since they are concentrated strongly at levels well below the maximum in ozone heating at 50 km (31 mi). Even if ozone were to be decreased at all heights by the maximum value of 1%, surface temperature decreases would amount to no more than 0.01 °K (50). This decrease would be made even smaller by the expected increase in NOn which acts to warm the troposphere by absorbing solar radiation reflected from the surface. Thus, on a globally averaged basis, radiative mechanisms would be ineffective in causing significant climatic changes associated with HLLVs. It is conceivable that the radiative effects could be enhanced by the formation of “corridors” of H2O and NOj. enhancement (or ozone reduction). For example, if all the water vapor from the HLLVs were concentrated in a 5°-wide latitudinal band, water vapor increases of 10% within that band would be possible. However, this is not likely to result in a local 0.06 °K increase in surface temperature, for horizontal transport will smear out any synoptic scale heating on a much more rapid time scale than the several months it takes to build up significant water vapor increases. In any case, it was shown in Sec. 4 that the establishment of a water vapor “corridor” is highly unlikely below 60 km (36 mi). It should be noted that cloud cover is an extremely important parameter in the assessment of climate — and a critically uncertain one as well. However, it appears that changes in cloud cover associated with HLLVs will be insignificant. In the mesosphere, noctilucent cloud incidence is not likely to increase greatly, as shown in
another paper (30). As for the troposphere, changes in water vapor will be less than 0.1%, hardly important relevant to climatic changes. Moreover, very little sulfur or particulate materials are expected from HLLVs, and condensation or nucleation processes are not likely to be affected significantly. Dynamical interactions between the troposphere and upper atmosphere have traditionally been thought of as a “one-way street,'' that is, the troposphere performs work on the stratosphere, but does not respond in turn to stratospheric motions. This is because stratospheric motions contain much less energy than tropospheric motions, a result of the much lower mass of the stratosphere. The mean meridional circulation of the stratosphere, for example, would not have much effect on the troposphere since meridoneal wind speeds are comparable in magnitude to tropospheric wind speeds and thus much less energetic because of lower air density in the stratosphere. However, eddy motions, such as large-scale planetary waves, generally amplify as they propagate upward, maintaining a nearly constant energy per unit volume. If one allows for these waves to be reflected in the stratosphere by some mechanism, with the amplitude and the phase of the reflected waves being governed by stratospheric conditions, it is possible for the stratosphere to have a feedback effect on tropospheric planetary waves. In fact, as Charney and Drazin (53) pointed out, the westerly jet of the winter upper stratosphere does act as just such a reflection mechanism. The implications of this for climatic changes are significant, for if changes in the zonal jet could be reduced by compositional changes at high altitudes, planetary wave reflection characteristics would also be affected, thus affecting the phase or amplitude of planetary waves in the troposphere (54). A number of investigators have explored this mechanism, including Avery and Geller (55) and Geller and Alpert (56). Their studies focused on the linear planetary wave response of a number of zonally averaged atmospheres to topographic and adiabatic forcing. Comparison of results for different cases could then show whether different types of changes in the stratospheric zonal wind had significant effects on tropospheric climate. The major result of the studies is as follows: below 35 km (22 mi), zonal wind changes of 20% or more must occur before planetary wave phase or amplitude changes at the surface approach the interannual variability (56). Of course, this result does not take into account the highly nonlinear nature of stratospheric motions; still, it is a good indication of the magnitudes necessary to cause significant dynamically induced climate changes. The implications of this result for SPS operations can be crudely assessed by making some estimates of the effects of HLLV water vapor emission. The radiative- convective model of Manabe and Wetherald (52) suggests a 10 °K decrease in stratospheric temperatures associated with quintupling the water vapor content. Scaling this down to an expected 10% maximum H2O increase if HLLV water vapor remains at low latitudes, we obtain temperature decreases of —0.2 °K. When these decreases are compared with the equator-pole temperature gradient of about 40 °K (which is proportional to the zonal wind strength), it can be seen that the maximum expected change in the zonal wind is only 0.5%, well below the threshold for significant effects on reflected planetary waves. As before, HLLV-induced ozone changes would have even less effect, since they are confined to levels above 40 km (25 mi). Based on these estimates, we conclude that no significant effects on Earth's climate should result from the projected launch schedule of HLLVs (400 yr-1). That is, we expect no temperature changes, locally or globally averaged, in excess of 0.1 °K and no dynamical variations exceeding 5% of the interannual variability in planetary wave activity.
RkJQdWJsaXNoZXIy MTU5NjU0Mg==