Environ. Sci. Technol. 2009, 43, 2784–2786
Effect of Stratospheric Aerosols on Direct Sunlight and Implications for Concentrating Solar Power DANIEL M. MURPHY* Earth System Research Laboratory, Chemical Sciences Division, National Oceanic and Atmospheric Administration, 325 Broadway, Boulder, Colorado 80305
Received August 6, 2008. Revised manuscript received November 20, 2008. Accepted January 26, 2009.
Light scattering calculations and data show that stratospheric aerosols reduce direct sunlight by about 4 W for every watt reflected to outer space. The balance becomes diffuse sunlight. One consequence of deliberate enhancement of the stratospheric aerosol layer would be a significant reduction in the efficiency of solar power generation systems using parabolic or other concentrating optics. There also would be a reduction in the effectiveness of passive solar design.
Introduction Crutzen (1) revived discussion of using deliberate enhancement of the stratospheric aerosol layer to minimize climate change during the years before atmospheric carbon dioxide concentrations can be stabilized. The postulated aerosol loading could be comparable to that following the eruption of Mt. Pinatubo in 1991. Besides scattering sunlight to outer space, aerosol particles also scatter sunlight to the earth as diffuse radiation. The reduction in direct sunlight is therefore always greater than the reduction in total solar irradiance. The reduction in direct sunlight is calculated and verified with data from Mauna Loa, HI. The impact on concentrating solar power is surprisingly large: peak power output was reduced by about 20% during a year when the stratospheric aerosol layer from the Mt. Pinatubo eruption reduced total sunlight by less than 3%.
Light Scattering by Stratospheric Aerosols In the visible region where absorption is small, sunlight scattered by particles will either go to outer space or reach the surface of the earth as diffuse sunlight. The fraction of light going to either fate depends on both the angle by which the light is scattered and the solar zenith angle. Ignoring small effects such as refraction near the horizon, for isotropic scattering half the scattered light reaches the earth and half goes to space. However, aerosol particles preferentially scatter light in the forward direction. This leads to a smaller fraction of light going back to outer space and a larger fraction of diffuse light. At a given solar zenith angle (SZA) the fraction going back to space is given by * Corresponding author; phone: (303) 497-5640; fax: 303-497-5373 e-mail:
[email protected]. 2784 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 8, 2009
B(SZA) )
{
1 2π
∫
π
0
f(θ)S(θ) sin(θ) dθ
f (θ) ) 1;θ < π/2 - SZA 1 cos-1(tan(π/2 - SZA) cot θ);π/2 - SZA < θ < π/2 + SZA π 0;θ > π/2 + SZA
}
(1)
j is the average fraction of total scattered light that where B goes to outer space, θ is the scattering angle in radians, and S(θ) is the scattered intensity. A derivation is shown in the Supporting Information. Since half of the earth is always illuminated by the sun, the global average is independent of the time and season. The global average fraction of light scattered to space by a uniform, optically thin layer was derived by Wiscombe and Grams (2): j) 1 B 2π
∫
π
0
θS(θ) sin(θ) dθ
(2)
These equations are valid for each wavelength. For the earth’s energy budget they must be integrated over the solar spectrum. This integration also smooths out the Mie resonances in scattered light that happen when particles are similar in size to the wavelength of light. Light scattered by extremely small angles will be equivalent to direct sunlight for most purposes. For Figure 1, light scattered less than 2.8° was considered to remain as direct sunlight. This value was chosen to match the angular response of the instrumentation at Mauna Loa. The choice of 0° to about 4° is within about a line width in Figure 1 except for particles larger than 1 µm, as shown by the difference of the heavy curve with the dashed curve that diverges above 1 µm. Solar parabolic trough collectors often have smaller acceptance angles (3), so their response will be between the curves with the 2.8° cutoff and no cutoff. Figure 1 shows how much direct sunlight is lost compared to total sunlight for various sizes of spherical particles in clear sky conditions. In the size range of natural stratospheric particles, between 3.5 and 4.8 W is lost from the direct solar beam for every watt reflected to outer space. The difference is due to scattering into diffuse sunlight, much of which still reaches the earth’s surface. The reason for the size dependence is that larger particles scatter more sunlight in forward directions that still reach the earth’s surface. These calculations can be verified with data from Mauna Loa, a site with high-quality instrumentation and a data record spanning the eruption of Mt. Pinatubo. Figure 2a shows direct, diffuse, and total sunlight measured at Mauna Loa. Those data are for a 60° solar zenith angle that approximates average incoming sunlight weighted by the slant path length. Regressions among transmission, direct, and diffuse radiation at Mauna Loa (Table 1 in ref 4) show that for the stratospheric aerosol loadings following the eruption of Mt. Pinatubo the ratio of changes in direct to total sunlight was about 4.6, in agreement with calculations for particles of the size known to be present in the stratosphere after the eruption (5). Similar changes in diffuse sunlight were also observed after the eruption of El Chicho´n (6). Any intentional enhancement of the stratospheric aerosol layer would need to produce particles between about 0.2 and 1 µm in diameter: smaller particles do not efficiently scatter light, and larger particles are quickly lost from the stratosphere due to gravitational settling (Figure 1b). Atmospheric transport, coagulation, and gravitational settling will all limit the stratospheric lifetime. In the 0.2-1 µm size 10.1021/es802206b
Not subject to U.S. Copyright. Publ. 2009 Am. Chem. Soc.
Published on Web 03/11/2009
FIGURE 1. (a) Changes due to aerosols in globally averaged direct and diffuse sunlight (thick line) as calculated from scattering theory. Selected solar zenith angles are also shown. For example, in the global average 0.4 µm diameter particles scatter about 4 times as much energy out of the direct solar beam as they reflect to outer space. The diameters labeled “nonvolcanic” and “volcanic” are the median scattering-weighted diameters of the stratospheric aerosol layer between and after major volcanic eruptions. That is, half of the light is scattered by particles smaller or larger than these diameters. For this figure light scattered less than 2.8° was considered to remain direct sunlight. The thin dashed line diverging from the heavy global line shows the effect of having no cutoff angle instead of 2.8°. (b) Mass scattering efficiency for light scattered to outer space and fall velocity for standard conditions at 20 km. range globally averaged direct sunlight is reduced 2.7-4.8 times as much as total sunlight. The calculations shown here are for spherical particles with the refractive index of sulfuric acid produced after injection into the stratosphere of either carbonyl sulfide or sulfur dioxide. Designer particles have also been mentioned (1). Equations 1 and 2 would still be valid, but the phase functions S(θ) would be different from those for spherical particles. Such phase functions would change with time as the particles were coated with and/or dissolved by naturally occurring sulfuric acid in the stratosphere, as is the case for ablated meteor debris in the stratosphere (7).
Discussion There are many consequences of large changes from direct to diffuse sunlight. Diffuse solar radiation affects ecosystem productivity (8). An effect considered here is the reduction in output from central solar electric power generation. As an alternative to flat photovoltaic panels, concentrating systems focus sunlight either onto high-efficiency photovoltaic cells or onto tubes to produce steam or a hot fluid. This heat is used to generate power using a conventional generator. In
FIGURE 2. (a) Sunlight measured at Mauna Loa. The reduction in direct sunlight and increase in diffuse sunlight following the eruption of Mt. Pinatubo in 1991 are readily apparent. The reduction in total sunlight was much smaller. (b) Output of the solar electric generating systems (SEGS) solar thermal power plants in California (data are from ref 9). The SEGS plants had significant reductions in on-peak capacity and total output following the eruption of Mt. Pinatubo. sunny locations centralized thermal plants using concentrated sunlight generate electricity at lower cost than electricity from photovoltaic cells (9). There are nine SEGS plants in California with a combined capacity of 354 peak MW as well as plants operating and under construction elsewhere in the world. An important characteristic of concentrating solar collectors is that they only utilize direct sunlight. The power output of the SEGS plants therefore dropped significantly after the eruption of Mt. Pinatubo (10, 11). The percentage reduction in total power generation was similar to the reduction in direct sunlight. For example, both the Mauna Loa direct sunlight and the total power generation increased by 13% between 1992 and 1993 as the stratosphere aerosol layer recovered from the eruption (Figure 2). The on-peak solar capacity was more sensitive to the decrease in direct sunlight than was the total power generation (Figure 2b). This is because “on-peak” is defined by peak electricity demand (afternoons in late summer) rather than peak solar performance. The slant path through the stratosphere is longer at times of peak electrical demand than it is at local noon, causing a larger reduction in on-peak power even though the relative amount of diffuse sunlight is smaller at large solar zenith angles (thin lines in Figure 1a). Combining the ratio of changes in direct to total sunlight with operation at various solar angles, each 1% reduction in VOL. 43, NO. 8, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2785
total sunlight reaching the earth from enhancement of stratospheric aerosols will cause a 4-10% loss in output from concentrated solar power applications depending on what measure is used for electrical output. Because flat solar hot water and photovoltaic panels utilize diffuse as well as direct sunlight, the performance losses are not as large as for concentrating solar collectors. However, the performance loss will still exceed the reduction in total sunlight because a tilted panel does not capture diffuse sunlight as efficiently as direct sunlight. A potentially important effect is that any shift from direct to diffuse sunlight makes a passive solar design less effective: in winter diffuse sunlight is harder to capture with south-facing windows, and in summer shading windows with overhangs is less effective. Although only stratospheric aerosols are explicitly considered here, any cooling of the earth that relies on light scattering by particles, including tropospheric aerosol scattering and increased cloudiness, will also result in reductions in direct sunlight that are several times the reductions in total sunlight.
Acknowledgments We thank J. T. McKinnon for locating a reference and E. G. Dutton for providing the Mauna Loa data in digital form.
Supporting Information Available Figure showing the derivation of eq 1. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Crutzen, P. J. Albedo enhancement by stratospheric sulfur injections: A contribution to resolve a policy dilemma. Clim. Change 2006, 77, 211–219.
2786
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 8, 2009
(2) Wiscombe, W. J.; Grams, G. W. The backscattered fraction in two-stream approximations. J. Atmos. Sci. 1976, 33, 24402451. (3) Rabi, A. Active Solar Collectors and Their Applications; Oxford University Press USA: New York, 1985. (4) Dutton, E. G.; Bodhaine, B. B. Solar irradiance anomalies caused by clear-sky transmission variations above Mauna Loa:195899. J. Clim. 2001, 14, 3255–3262. (5) Deshler, T.; Hervig, M. E.; Hofmann, D. J.; Rosen, J. M.; Liley, J. B. Thirty years of in situ stratospheric aerosol size distribution measurement from Laramie, Wyoming (41° N), using balloonborne instruments. J. Geophys. Res. 2003, 108, No. 4167, DOI: 10.1029/2202JD002514. (6) Olmo, F. J.; Tovar, J.; Alados-Arboledas, L.; Okulov, O.; Ohvril, H. O. A comparison of ground level solar radiative effects of recent volcanic eruptions. Atmos. Environ. 1999, 33, 4589– 4596. (7) Murphy, D. M.; Thomson, D. S.; Mahoney, M. J. In situ measurements of organics, meteoritic material, mercury, and other elements in aerosols at 5 to 19 kilometers. Science 1998, 282, 1664–1669. (8) Gu, L. H.; Baldocchi, D.; Verma, S. B.; Black, T. A.; Vesala, T.; Falge, E. M.; Dowty, P. R. Advantages of diffuse radiation for terrestrial ecosystem productivity. J. Geophys. Res. 2002, 107 (D6), No. 4050, DOI: 10.1029/2001JD001242. (9) Quaschning, V. Technical and economical system comparison of photovoltaic and concentrating solar thermal power systems depending on annual global irradiation. Sol. Energy 2004, 77, 171–178. (10) Michalsky, J. J.; Perez, R.; Seals, R.; Ineichen, P. Degradation of solar concentrator performance in the aftermath of Mount Pinatubo. Sol. Energy 1994, 52, 215–213. (11) Kearney, D.; Price, H. Recent advances in parabolic trough solar power plant technology. Adv. Sol. Energy 2005, 16, 155–232.
ES802206B
