See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/12606075

Suppression of Rain and Snow by Urban and Industrial Air Pollution

Article  in  Science · April 2000

DOI: 10.1126/science.287.5459.1793 · Source: PubMed

CITATIONS READS
1,106 2,839

1 author:

Daniel Rosenfeld
Hebrew University of Jerusalem
355 PUBLICATIONS   21,700 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Aerosol Cloud Precipitatio Climate initiative (ACPC) View project

Aerosol-Cloud Interactions View project

All content following this page was uploaded by Daniel Rosenfeld on 19 May 2014.

The user has requested enhancement of the downloaded file.

 Suppression of Rain and Snow by Urban and
Industrial Air Pollution
Daniel Rosenfeld, et al.
Science 287, 1793 (2000);
DOI: 10.1126/science.287.5459.1793

The following resources related to this article are available online at

www.sciencemag.org (this information is current as of December 19, 2006 ):

Updated information and services, including high-resolution figures, can be found in the online
version of this article at:
http://www.sciencemag.org/cgi/content/full/287/5459/1793
A list of selected additional articles on the Science Web sites related to this article can be

Downloaded from www.sciencemag.org on December 19, 2006

found at:
http://www.sciencemag.org/cgi/content/full/287/5459/1793#related-content
This article has been cited by 204 article(s) on the ISI Web of Science.

This article has been cited by 12 articles hosted by HighWire Press; see:
http://www.sciencemag.org/cgi/content/full/287/5459/1793#otherarticles

This article appears in the following subject collections:

Atmospheric Science
http://www.sciencemag.org/cgi/collection/atmos

Information about obtaining reprints of this article or about obtaining permission to reproduce
this article in whole or in part can be found at:
http://www.sciencemag.org/help/about/permissions.dtl

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by the
American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. Copyright
c 2005 by the American Association for the Advancement of Science; all rights reserved. The title SCIENCE is a
registered trademark of AAAS.
 REPORTS
11. D. E. Smith et al., Science 284, 1495 (1999). constructing models of bending stress that are con- 47. S. Zhong and M. T. Zuber, J. Geophys. Res., in press.
12. M. H. Carr, The Surface of Mars (Yale Univ. Press, New sistent with a best-fit lithospheric strength envelope 48. D. L. Turcotte, R. J. Willemann, W. F. Haxby, J. Nor-
Haven, CT, 1981). (50). berry, J. Geophys. Res. 86, 3951 (1981).
13. H. Frey, S. E. Sakimoto, J. Roark, Geophys. Res. Lett. 34. D. W. Forsyth, J. Geophys. Res. 98, 16073 (1993). 49. C. Goetze, Philos. Trans. R. Soc. London Ser. A 288, 99
25, 4409 (1998). 35. N. H. Sleep, J. Geophys. Res. 99, 5639 (1994). (1978); Y. Caristan, J. Geophys. Res. 87, 6781 (1982).
14. D. E. Wilhelms and S. W. Squyres, Nature 309, 138 36. F. Nimmo and D. J. Stevenson, J. Geophys. Res., in 50. W. F. Brace and D. L. Kohlstedt, J. Geophys. Res. 85,
(1984); G. E. McGill, J. Geophys. Res. 94, 2753 (1989). press. 6248 (1980).
15. H. V. Frey and R. A. Schultz, Geophys. Res. Lett. 15, 37. J. W. Head III et al., Science 286, 2134 (1999). 51. R. J. Phillips, M. T. Zuber, S. A. Hauck, R. M. Williams,
229 (1988). 38. T. J. Parker, D. S. Gorsline, R. S. Saunders, D. C. Pieri, K. B. Portle, Lunar Planet. Sci. 31 (abstr. no. 1303)
16. S. R. Bratt, S. C. Solomon, J. W. Head, C. H. Thurber, D. M. Schneeberger, J. Geophys. Res. 98, 11061 (2000).
J. Geophys. Res. 90, 3049 (1985); M. T. Zuber, D. E. (1993). 52. We thank the MGS spacecraft and mission operations
Smith, F. G. Lemoine, G. A. Neumann, Science 266, 39. The proposed buried channels are situated within the teams at the Jet Propulsion Laboratory and Lockheed-
1839 (1994). Tharsis negative gravity ring, which is likely a mem- Martin Astronautics for their contributions to this
17. S. C. Solomon and J. W. Head, J. Geophys. Res. 82, brane response to Tharsis loading (51). The linear effort. We also acknowledge R. Follas, J. Abshire, and
9755 (1982). anomalies are superposed on the relatively broad the MOLA Instrument Team for laser performance
18. W. B. Banerdt, M. P. Golombek, K. L. Tanaka, in Mars, membrane gravity signal. information; M. Torrence and J. Schott for assistance
H. H. Kieffer, B. M. Jakosky, C. W. Snyder, M. S. 40. I. Klaucke, R. Hesse, W. B. F. Ryan, Sedimentology 44, in altimetry processing; R. Simpson, P. Priest, S. As-
Matthews, Eds. (Univ. of Arizona Press, Tucson, AZ, 1093 (1997); Geol. Soc. Am. Bull. 110, 22 (1998). mar, and J. Twicken for assistance in tracking data
1992), pp. 249 –297. 41. F. G. Lemoine et al., Eos (Fall Suppl.) 80, F618 (1999); acquisition and processing; S. Fricke for help with
19. J. Morgan and M. Warner, Geology 27, 407 (1999). W. L. Sjogren, D. Yuan, A. S. Konopliv, Eos (Fall orbit determination; M. Wieczorek for compilation of
20. F. Sohl and T. Spohn, J. Geophys. Res. 102, 1613 Suppl.) 80, F618 (1999). martian meteorite densities; M. Simons and F. Simons
(1997). 42. F. G. Lemoine et al., in preparation. (no relation) for codes used in the lithosphere thick-
43. W. M. Kaula, Theory of Satellite Geodesy (Blaisdell, ness inversions; and D. Stevenson and N. Sleep for
21. Using the method of (47), we estimated the relax-

Downloaded from www.sciencemag.org on December 19, 2006

Waltham, MA, 1966). helpful reviews. The MGS Radio Science and MOLA
ation time for a degree-1 crustal thickness variation
44. S. S. C. Wu, USGS Map I-2160 (1991). investigations are supported by the NASA Mars Ex-
on a self-gravitating spherical planet. We assumed a
45. M. A. Wieczorek and R. J. Phillips, J. Geophys. Res. ploration Program.
mechanically strong upper crust of thickness 20 km
and a uniform-thickness viscous lower crust. Densi- 103, 1715 (1998).
ties are as in (8). 46. H. Y. McSween Jr., Rev. Geophys. 23, 391 (1985). 20 January 2000; accepted 18 February 2000
22. G. Schubert and T. Spohn, J. Geophys. Res. 95, 14095
(1990).
23. O. Grasset and E. M. Parmentier, J. Geophys. Res.
103, 18171 (1998); G. Choblet, O. Grasset, E. M.
Suppression of Rain and Snow
Parmentier, C. Sotin, Lunar Planet. Sci. 30, 1556
(1998); T. Spohn, D. Breuer, V. Conzelmann, Eos
(Fall Suppl.) 80, F619 (1999).
by Urban and Industrial Air
24. M. Simons, S. C. Solomon, B. H. Hager, Geophys.
J. Int. 131, 24 (1997); F. J. Simons, M. T. Zuber, J.
Korenaga, J. Geophys. Res., in press. The admittance
Pollution
is the response function between spectral gravity and
topography, and the coherence is the square of the
Daniel Rosenfeld
correlation coefficient between normalized spectral
amplitudes of gravity and topography. The admit- Direct evidence demonstrates that urban and industrial air pollution can com-
tance approach assumes surface loading of an elastic pletely shut off precipitation from clouds that have temperatures at their tops
spherical shell, whereas the coherence approach as-
sumes surface and subsurface loading of an elastic of about –10°C over large areas. Satellite data reveal plumes of reduced cloud
plate. Both methods use densities as in (8). particle size and suppressed precipitation originating from major urban areas
25. M. H. Acuña et al., Science 284, 790 (1999). and from industrial facilities such as power plants. Measurements obtained by
26. J. E. P. Connerney et al., Science 284, 794 (1999).
the Tropical Rainfall Measuring Mission satellite reveal that both cloud droplet
27. Common magnetic minerals on Earth carry remanent
magnetization only at temperatures less than the coalescence and ice precipitation formation are inhibited in polluted clouds.
Curie temperature of 580° (magnetite) to 680°C
(hematite) [D. J. Dunlop and O. Ozdemir, Rock Mag- The precipitation-forming processes in clouds ed that giant CCN enhanced precipitation
netism (Cambridge Univ. Press, Cambridge, 1997)]. depend to a large extent on the presence of (12), but attempts to correlate the urban-en-
28. D. E. Smith et al., Science 279, 1686 (1998).
29. D. W. Forsyth, J. Geophys. Res. 90, 12623 (1985).
aerosols, specifically cloud condensation nuclei hanced rainfall to the air pollution sources
30. Given the similar diameters of Hellas and Utopia (CCN) and ice nuclei. The large concentrations failed to show any relation (13). Another
(11), we used the current gravitational and topo- of small CCN in the smoke from burning veg- explanation for the urban rain enhancement
graphic signatures of Hellas and calculated the trade- etation nucleate many small cloud droplets (1, invoked the heat-island effect and increased
off between Te and fill density that is required to
explain Utopia. For Hellas, we used a depth of 11 km, 2) that coalesce inefficiently into raindrops (3, friction, both of which would tend to increase
a degree-6 (corresponding to the basin diameter) 4). Although this effect has been suspected for the surface convergence, resulting in more
gravity anomaly of –160 mGal, and a crust-mantle many years (5, 6), conclusive evidence that cloud growth and rainfall over and downwind
boundary as in Fig. 1C. For Utopia, we used a depth
of 2.5 km and a gravity anomaly of 270 mGal. We smoke from burning vegetation suppresses pre- of the urban areas (14). Furthermore, the
calculated Te as in (48). If the primary contributor to cipitation was obtained recently with the obser- recent suggestion (15) that air pollution might
Utopia basin fill is sedimentary (density ⫽ 2200 kg vations of the Tropical Rainfall Measuring Mis- enhance precipitation on a large scale in
m⫺3), then for a local crustal density comparable to
that of shergottites (3200 kg m⫺3), the same elastic sion (TRMM) (7) satellite (8). northeastern America and the accompanying
thickness (100 km) is obtained as for volcanic filling Much less is known, however, about the speculative explanations demonstrate how lit-
of a crust having the density assumed for the model impact of aerosols from urban and industrial tle is known about the subject.
of Fig. 1C.
31. K. L. Tanaka, D. H. Scott, R. Greeley, in (18), pp.
air pollution on precipitation. It was assumed Space-borne (16) and in situ aircraft (17)
345–382. initially that industrial and urban pollution measurements of ship tracks in marine
32. A. B. Watts, J. H. Bodine, M. S. Steckler, J. Geophys. inhibited precipitation, similar to the smoke stratocumulus clouds provided the first evi-
Res. 85, 5369 (1980). from burning vegetation (9). Later, reports of dence that effluents from ship stacks change
33. S. C. Solomon and J. W. Head, J. Geophys. Res. 95,
11073 (1990); M. K. McNutt, J. Geophys. Res. 89,
enhanced rainfall downwind of paper mills cloud microstructure by redistributing their
11180 (1984). We converted effective elastic thick- (10) and over major urban areas (11) suggest- water into a larger number of smaller drop-
ness (Te) to heat flow via computation of the me- lets. Albrecht (18) suggested that the drizzle,
chanical thickness of an elastic-plastic plate (Tmech). which normally occurs in marine stratocumu-
The approach entails adopting a range of represen- Institute of Earth Sciences, The Hebrew University of
tative strain rates (10⫺19 s ⫺1 and 10 ⫺15 s–1) and Jerusalem, Givat Ram, Jerusalem, Israel. E-mail: lus clouds in clean air, would be inhibited
ductile flow laws for the lower lithosphere (49) and [email protected] from the clouds with reduced droplet size,

www.sciencemag.org SCIENCE VOL 287 10 MARCH 2000 1793

 REPORTS
thereby increasing the cloud water content
and longevity. Extrapolation to clouds that
are sufficiently thick for raining (i.e., at least
2 km from base to top) would mean that the
effluents have the potential to suppress pre-
cipitation over ocean and over land. Howev-
er, pollution tracks in any clouds over land
were not reported in previous studies. Appli-
cation of the imaging scheme of Rosenfeld
and Lensky (6 ) to the Advanced Very High
Resolution Radiometer (AVHRR) on board
the National Oceanic and Atmospheric Ad-
ministration (NOAA) orbiting weather satel- Fig. 1. Satellite visualization of NOAA AVHRR
lites revealed numerous ship track–like fea- images, showing the microstructure of clouds
tures in clouds over land, created by major for three cases over three different continents
with streaks of visibly smaller drops due to
urban and industrial pollution sources. Illus- ingestion of pollution originating from known
trations of such tracks from Turkey (Fig. 1A), pollution sources that are marked by white
Canada (Fig. 1B), and Australia (Fig. 1C) are numbered asterisks. (A) A 300 km by 200 km

Downloaded from www.sciencemag.org on December 19, 2006

shown. Because the tracks clearly originate cloudy area containing yellow streaks originat-
from pollution sources, they will be called ing from the urban air pollution of Istanbul (*1),
hereafter “pollution tracks.” Izmit (*2), and Bursa (*3) on 25 December
1998 at 12:43 UT. (B) A 150 km by 100 km
The pollution tracks in Turkey (Fig. 1A) cloudy area containing yellow streaks showing
originate from several sources in and near the the impact of the effluents from the Hudson
cities of Istanbul, Izmit, and Bursa. Bay Mining and Smelting compound at Flin-
The pollution track in Canada (Fig. 1B) Flon (*4) in Manitoba, Canada (54°46⬘N
originates from Flin-Flon, Manitoba, the home 102°06⬘W), on 4 June 1998 at 20 :19 UT. (C)
of the Hudson Bay Mining and Smelting Com- An area of about 350 km by 450 km containing
pollution tracks over South Australia on 12
pany. That location has been a frequent source August 1997 at 05:26 UT originating from the
for such tracks. Other sources in Canada have Port Augusta power plant (*5), the Port Pirie
been observed, but not reported here. lead smelter (*6), Adelaide port (*7), and the
Study of the pollution tracks emanating oil refineries (*8). All images are oriented with
from the region of Adelaide, South Australia, north at the top. The images are color compos-
Australia, is especially interesting. They re- ites, where the red is modulated by the visible
channel; blue is modulated by the thermal in-
ceived special attention because of their in- frared (IR); and green is modulated by the solar
tensity and frequent occurrence. These pollu- reflectance component of the 3.7-␮m channel,
tion tracks were identified in the clouds of all where larger (greener) reflectance indicates smaller droplets. The composition of the channels
47 AVHRR images on different days exam- determines the color of the clouds, where red represents clouds with large drops and yellow
ined in which stratocumulus and cumulus represents clouds with small drops. The blue background represents the ground surface below the
clouds with tops warmer than about ⫺12°C clouds. A full description of the color palettes and their meaning is provided by Rosenfeld and
Lensky (6).
existed over the region. The pollution tracks
in Fig. 1C coincide with the following major
industrial and urban areas: (i) Port Augusta kilometers farther downwind from the pollu- lite. The TRMM instruments used here are
has a 520-MW power plant operating on tion source, mainly in the form of sulfates. the visible and infrared sensor (VIRS), the
brown coal, providing electricity to the near- The AVHRR data were used to retrieve precipitation radar (PR), and the TRMM pas-
by mines and to the adjacent large steel in- the dependence of the indicated effective ra- sive microwave imager (TMI). The VIRS is
dustry in Whyalla. (ii) Port Pirie is the home dius re ⫽ 具r 3典/具r 2典, where r is the radius of the similar to the NOAA AVHRR, but it uses a
of the world’s largest lead smelter and refin- cloud droplets in the measurement volume, 2-km subsatellite resolution instead of the
ery. (iii) Adelaide has industry for processing on cloud temperature T. The method of 1.1-km resolution used by the AVHRR to
minerals mined in the vicinity. Among these Rosenfeld and Lensky (6) was used to derive obtain the T-re relations. PR is a 2.2-cm radar
are Australia’s largest cement plant, located the T-re relations for inference of the precip- with a subsatellite resolution of 250 m verti-
on the Port Adelaide River. A major oil re- itation-forming processes in the clouds. cally by 4 km horizontally. The minimum
finery and a power plant are located 20 km to The median re of the cloud tops in the detectable signal is about 17 dBZ [decibel of
the south of the city near the origin of the pollution plumes (Figs. 1 through 3) was Z (mm6m⫺3)], which is equivalent to about
strongest pollution track in Fig. 1C. considerably less than the precipitation 0.7 mm hour⫺1. The PR is used to measure
The 1998 –99 annual average of effluents threshold of 14 ␮m (20). Outside the plumes, the precipitation that forms in the clouds. The
from the stack of Port Augusta power plant, however, re increased steeply with decreasing TMI uses a 85-GHz vertical polarization
which is equipped with an electrostatic pre- T to more than 25 ␮m, indicating that the brightness temperature (T85) to detect the wa-
cipitator, was 43 kg hour⫺1 of submicrometer cloud droplets in the general area were coa- ter in nonprecipitating clouds.
ash particles with modal diameter of 0.14 lescing into precipitation. At the same time, The TRMM measurements are validated
␮m. The gaseous annual average of effluents little growth of re with decreasing T was by an extensive ground validation program
for the same time period is 1108 kg hour⫺1 of indicated within the pollution plumes, indi- (21). Preliminary results show variability of
SO2 and 1655 kg hour⫺1 of NOx (19). Ap- cating a lack of coalescence and, thus, sup- about 25% between rain gauges and TRMM
parently, part of the ash particles act as CCN pressed precipitation. rainfall over large areas, with some TRMM
at short range, and chemical reactions of the These inferences are validated using the underestimation with the heavier rainfall (22).
gases produce additional CCN hundreds of additional sensors onboard the TRMM satel- The simultaneous spaceborne measurements of

1794 10 MARCH 2000 VOL 287 SCIENCE www.sciencemag.org

 REPORTS

Downloaded from www.sciencemag.org on December 19, 2006

Fig. 2. Satellite visualization of the TRMM VIRS, with the precipitation right, respectively). The image shows pollution plumes in the clouds
information in the white overlay. The white patches denote precipita- over southeastern Australia on 21 October 1998 at 04:44 UT. The lines
tion echoes as observed by the TRMM PR. The two parallel lines delimit AB and CD show the locations of the vertical cross sections presented in
the 230-km PR swath. The swath is oriented from west to east (left to Fig. 5.

(box 4). The re increased gradually farther

downwind but did not reach the precipitation
threshold. Farther downwind and to the east of
the plume (box 5) only a few isolated showers
had formed in some of the clouds that reached
the ⫺9°C isotherm and barely exceeded the
14-␮m precipitation threshold.
Clouds over the Sydney area (box 6) also
had a much reduced re, apparently because of
the pollution there. Clouds over the sea some
150 km away from Sydney (box 7) had a much
larger re, indicative of strong coalescence and
Fig. 4. Vertical profiles of the precipitation precipitation in the apparently pristine marine
echo intensities as measured by the TRMM PR air. The TMI-based TRMM rainfall algorithm
for the various boxes in Fig. 2. The peak near 2 identified those clouds as precipitating.
km corresponds to enhanced echoes from
snowmelt just below the 0°C isotherm. Box 2 The vertical cross sections (Fig. 5) of the
had no detectable precipitation. Boxes 4 and 7 plumes show no obvious differences in the
are outside the radar swath. H, cloud height. cloud top heights and horizontal dimensions
in and outside the areas of suppressed precip-
The TRMM overpass selected for analysis itation. Furthermore, the depression of the
here is from 21 October 1998 at 04:41 UT, a TMI-measured T85 in the nonprecipitating
time of strong pollution-track signatures in the clouds indicates that lack of cloud water was
clouds throughout southeastern Australia (Fig. not the reason for the lack of precipitation
2). Seven boxes within the image were delim- from these clouds.
ited for analyses of the dependence of re (Fig. 3) The vertical profile of the precipitation
and the PR reflectivities (Fig. 4) as a function of echoes as measured by the PR has a distinct
T. Box 2 encloses a pollution plume that is maximum near the 0°C isotherm, between 2
clearly visible by the yellow coloring, which and 2.5 km above sea level (Fig. 4). This
Fig. 3. Analysis of the T-re relations, where T is the indicates the small re of the cloud particles maximum is caused by the enhanced radar
temperature and re (reff) is the cloud particle downwind of Adelaide. According to the T-re returns from melting snowflakes, known as
effective radius for the clouds in the seven boxes relations presented in Fig. 3A, the clouds in the the “bright band.” The existence of the bright
(traces numbered 1 through 7) in Fig. 2. Plotted plume had little coalescence, had not glaciated, band shows that much of the precipitation
are the 15th (long dashed line), 50th (solid line), and were without precipitation, whereas the was initiated as snow in the upper parts of the
and 85th (short dashed line) percentiles of re for
unpolluted clouds in boxes 1 and 3, side-wind clouds. That means that the pollution sup-
each 1°C interval. The black lines correspond to
the boxed areas of pollution. The vertical green of box 2, had strong coalescence and were pressed the precipitation not only by inhibit-
line marks the 14-␮m precipitation threshold. A precipitating. That observation is corroborated ing the coalescence of the cloud droplets into
full description of the T-re charts and their mean- by the PR, which recorded precipitation echoes raindrops but also by preventing the forma-
ing is provided by Rosenfeld and Lensky (6). in the clouds outside, but not within, the plume. tion of ice particles and cold-precipitation
The PR-measured rain intensities exceeded 10 processes of the clouds. A likely explanation
cloud microphysics (VIRS), cloud water (TMI), mm hour⫺1. Another large area of clouds with is that the pollution reduces the radius of the
and precipitation (PR) make it possible to relate extremely small re existed in the Melbourne largest cloud droplets below the threshold of
the precipitation to cloud microstructure. area and extended downwind to the northeast 12 ␮m, which is required for both primary

www.sciencemag.org SCIENCE VOL 287 10 MARCH 2000 1795

 REPORTS
Fig. 5. Vertical cross
section along the lines
AB and CD in Fig. 2.
The dark gray areas
represent clouds. The
vertical extent of the
clouds is converted
from the VIRS-mea-
sured cloud-top temperatures. The colors represent the precipitation reflectivity in dBZ as measured by the TRMM radar. The white line is the
brightness temperature of the TMI 85-GHz vertical polarization (T85), plotted at the altitude of that temperature. A lower T85 value is represented as
greater height of the white line, and in nonprecipitating clouds it means greater cloud water content. The T85 and actual cloud-top temperature have
different physical meanings.

and secondary ice generation in clouds (23, Australia because it is seen against a back- 15. R. S. Cerveny and R. C. Balling Jr., Nature 394, 561
(1998).
24). Air pollution must be an important factor ground of pristine clouds, whereas in most
16. J. A. Coakley Jr., R. L. Bernstein, P. A. Durkee, Science
in determining the precipitation amounts in other places the clouds are already polluted on 237, 1020 (1987).
the Snowy Mountains (east of box 4 in Fig. 2) a very large scale. Such results might indicate 17. L. F. Radke, J. A. Coakley Jr., M. D. King, Science 246,
because it has been observed that most of the that human activity may be altering clouds and 1146 (1989).
winter precipitation events in that region natural precipitation on a global scale. 18. B. A. Albrecht, Science 245, 1227 (1989).
19. Emission data obtained from Flinders Power, Port
come from clouds with temperature at the

Downloaded from www.sciencemag.org on December 19, 2006

Augusta, South Australia, Australia.
tops between ⫺4° and ⫺13°C (25). Interest- 20. D. Rosenfeld and G. Gutman, J. Atmos. Res. 34, 259
References and Notes
ingly, a decreasing trend of the snow cover in 1. P. V. Hobbs and L. F. Radke, Science 163, 279 (1969). (1994).
the Snowy Mountains was reported for the 2. Y. J. Kaufman and R. S. Fraser, Science 277, 1636 21. Information about the TRMM ground validation pro-
(1997). gram is available at http://trmm.gsfc.nasa.gov/trmm_
period 1897–1991 (25). However, trend anal- office/field_campaigns/field_campaigns.html.
3. P. Squires, Tellus 10, 256 (1958).
yses of snow, winter temperature, and total 4. P. R. Jonas and B. J. Mason, Q. J. R. Meteorol. Soc. 22. R. Oki et al., Mar. Technol. Soc. J. 32, 13 (1999).
winter rainfall for the period 1910 –91 100, 286 (1974). 23. S. C. Mossop and J. Hallett, Science 186, 632 (1974).
showed statistically insignificant decreases in 5. J. Warner, J. Appl. Meteorol. 7, 247 (1968). 24. A. Rangno and P. V. Hobbs, Q. J. R. Meteorol. Soc.
6. D. Rosenfeld and I. Lensky, Bull. Am. Meteorol. Soc. 120, 573 (1994).
all three parameters (26 ). 79, 2457 (1998). 25. B. Harasymiw and J. McGee, Snowy Precipitation En-
The satellite data provide evidence con- 7. TRMM was launched on 28 November 1997 as a hancement Project: A Proposal to Evaluate Feasibility
necting urban and industrial air pollution to cooperative project of the National Aeronautics and of Increasing Snow Precipitation over the Snowy
Space and Administration (NASA) and the National Mountains Area (Snowy Mountains Hydroelectric Au-
the reduction of precipitation, pinpointing thority, Cooma, Australia 1993), appendix B.
Space Development Agency of Japan (NASDA). The
both the sources and the affected clouds. This TRMM data are available at http://daac.gsfc.nasa.gov/ 26. A. L. Duus, Aust. Meteorol. Mag. 40, 195 (1992).
has become possible with the newly acquired CAMPAIGN_DOCS/hydrology/hd_trmm_intro.html. 27. I thank all the members of the TRMM team, too
capabilities to observe both cloud microstruc- 8. D. Rosenfeld, Geophys. Res. Lett. 26, 3105 (1999). numerous to mention individually, for all their hard
9. R. Gunn and B. B. Phillips, J. Meteorol. 14, 272 (1957). work to make the satellite a reality and the data of
ture and precipitation over large areas with 10. R. C. Eagen, P. V. Hobbs, L. F. Radke, J. Appl. Meteorol. such high quality. The NOAA AVHRR data were
TRMM satellite observations. It might seem 13, 535 (1974). obtained from the NOAA Satellite Active Archive. I
strange that some of the most prominent pol- 11. R. R. Braham Jr., Meteorol. Monogr. 18, 141 (1981). also thank A. Gingis of Australian Management Con-
12. D. B. Johnson, J. Atmos. Sci. 39, 448 (1982). solidated Pty., Ltd. for assisting in this study and W. L.
lution signatures occur in Australia, which is Woodley for help with the manuscript.
13. D. F. Gatz, J. Appl. Meteorol. 18, 1245 (1979).
probably the least polluted inhabited conti- 14. E. Jauregui and E. Romales, Atmos. Environ. 30, 3383
nent. The pollution is perhaps most evident in (1996). 5 November 1999; accepted 10 January 2000

An Archaeal Iron-Oxidizing ⬃150 ⫻ 1012 g of sulfur per year, contribut-

ing ⬃50% to the net river transport of sulfate

Extreme Acidophile Important to the ocean, which is about half of the sulfate
input into the ocean (2).
Microorganisms accelerate the rate of py-
in Acid Mine Drainage rite dissolution through regeneration of Fe3⫹
(1, 3), the primary pyrite oxidant at low pH
Katrina J. Edwards,1,2* Philip L. Bond,1 Thomas M. Gihring,1 (4–6). At Iron Mountain, an AMD site in
Jillian F. Banfield1 northern California, the iron-oxidizing bacte-
rium Thiobacillus ferrooxidans, previously
A new species of Archaea grows at pH ⬃0.5 and ⬃40°C in slime streamers and thought to be the most important iron-oxidiz-
attached to pyrite surfaces at a sulfide ore body, Iron Mountain, California. This ing species, played a minor role in pyrite
iron-oxidizing Archaeon is capable of growth at pH 0. This species represents oxidation (7, 8). Instead, Archaea constituted
a dominant prokaryote in the environment studied (slimes and sediments) and a large proportion (⬎50%) of the prokaryote
constituted up to 85% of the microbial community when solution concentra- population at important sites of acid genera-
tions were high (conductivity of 100 to 160 millisiemens per centimeter). The tion during the dry summer and fall months
presence of this and other closely related Thermoplasmales suggests that these
acidophiles are important contributors to acid mine drainage and may sub- 1
Department of Geology and Geophysics, University
stantially impact iron and sulfur cycles. of Wisconsin–Madison, 1215 West Dayton Street,
Madison, WI 53706, USA. 2Woods Hole Oceano-
graphic Institute, Department of Marine Chemistry
The oxidative dissolution of metal sulfide and water through geological or mining ac- and Geochemistry, Woods Hole, MA 02543, USA.
minerals causes the formation of acid mine tivities undergo oxidative dissolution and
*To whom correspondence should be addressed at
drainage (AMD) and plays an important role generate sulfuric acid by the reaction FeS2 ⫹ Woods Hole Oceanographic Institute, Department of
in the geochemical sulfur cycle. Sulfides ( pri- 14Fe3⫹ ⫹ 8H2O 3 15Fe2⫹ ⫹ 2SO42⫺ ⫹ Marine Chemistry and Geochemistry, Woods Hole,
marily pyrite, FeS2) that are exposed to air 16H⫹ (1). Mining and extraction mobilizes MA 02543, USA. E-mail: [email protected]

1796 10 MARCH 2000 VOL 287 SCIENCE www.sciencemag.org

View publication stats