Content uploaded by Jaakko Mannio
Author content
All content in this area was uploaded by Jaakko Mannio
Content may be subject to copyright.
© 1999 Macmillan Magazines Ltd
NATURE
|
VOL 401
|
7 OCTOBER 1999
|
www.nature.com 575
letters to nature
.................................................................
Regional trends in aquatic recovery
from acidification in
North America and Europe
J. L. Stoddard
1
, D. S. Jeffries
2
,A.Lu
¨kewille
3
, T. A. Clair
4
, P. J. Dillon
5
,
C. T. Driscoll
6
, M. Forsius
7
, M. Johannessen
8
. J. S. Kahl
9
, J. H. Kellogg
10
,
A. Kemp
11
, J. Mannio
7
, D. T. Monteith
12
, P. S. Murdoch
13
, S. Patrick
12
,
A. Rebsdorf
14
, B. L. Skjelkva
˚le
8
, M. P. Stainton
15
, T. Traaen
8
,
H. van Dam
16
, K. E. Webster
17
, J. Wieting
18
& A. Wilander
19
1
Environmental Protection Agency, 200 SW 35th Street, Corvallis, Oregon 97333,
USA
2
Environment Canada, PO Box 5050, Burlington, Ontario, Canada L7R 4A6
3
Norwegian Institute for Air Research, PO Box 100, 2027 Kjeller, Norway
4
Environment Canada, PO Box 6227, Sackville, New Brunswick, Canada E4L 1G6
5
Ontario Ministry of the Environment, PO Box 39, Dorset, Ontario, Canada
P0A 1E0
6
Syracuse University, 220 Hinds Hall, Syracuse, New York 13244, USA
7
Finnish Environment Institute, Box 140, 00251 Helsinki, Finland
8
Norwegian Institute for Water Research, PO Box 173, 0411 Oslo, Norway
9
University of Maine, Sawyer Research Center, Orono, Maine 04469, USA
10
Vermont Department of Environmental Conservation, 103 S. Main Street,
Waterbury, Vermont 05676, USA
11
Environment Canada, 105 McGill, Montreal, Quebec, Canada H2Y 2E7
12
Environmental Change Research Centre, University College, 26 Bedford Way,
London WC1H 0AP, UK
13
US Geological Survey, 425 Jordan Road, Troy, New York 12180, USA
14
National Environmental Research Institute, PO Box 314, 8600 Silkeborg,
Denmark
15
Fisheries and Oceans Canada, 501 University Crescent, Winnipeg, Manitoba,
Canada R3T 2N6
16
Aquasence TEC, PO Box 95125, 1090 Amsterdam, The Netherlands
17
Wisconsin Department of Natural Resources, 1350 Femrite Drive, Monona,
Wisconsin 53716, USA
18
Umweltbundesamt, Seelstrasse 6– 10, 13581 Berlin, Germany
19
University of Agricultural Sciences, PO Box 7050, 750 07 Uppsala, Sweden
.................................. ......................... ......................... ......................... ......................... ........
Rates of acidic deposition from the atmosphere (‘acid rain’) have
decreased throughout the 1980s and 1990s across large portions of
North America and Europe1,2. Many recent studies have attributed
observed reversals in surface-water acidification at national3and
regional4scales to the declining deposition. To test whether
emissions regulations have led to widespread recovery in
surface-water chemistry, we analysed regional trends between
1980 and 1995 in indicators of acidification (sulphate, nitrate
and base-cation concentrations, and measured (Gran) alkalinity)
for 205 lakes and streams in eight regions of North America and
Europe. Dramatic differences in trend direction and strength for
the two decades are apparent. In concordance with general
temporal trends in acidic deposition, lake and stream sulphate
concentrations decreased in all regions with the exception of
Great Britain; all but one of these regions exhibited stronger
downward trends in the 1990s than in the 1980s. In contrast,
regional declines in lake and stream nitrate concentrations were
rare and, when detected, were very small. Recovery in alkalinity,
expected wherever strong regional declines in sulphate con-
centrations have occurred, was observed in all regions of
Europe, especially in the 1990s, but in only one region (of five)
in North America. We attribute the lack of recovery in three
regions (south/central Ontario, the Adirondack/Catskill moun-
tains and midwestern North America) to strong regional declines
in base-cation concentrations that exceed the decreases in
sulphate concentrations.
In the past two decades, the passage of national (for example, the
Clean Air Act in the United States and the Eastern Canada Acid Rain
Program in Canada) and international (for example, the United
Nations Economic Commission for Europe’s Convention on Long-
Range Transboundary Air Pollution (UN/ECE LRTAP) Sulphur
Protocols) environmental regulations and agreements has led to
widespread declines in the rates of acidic deposition, especially of
sulphur, across large regions of North America and Europe. In
north/central Europe, SO
2
concentrations in air decreased by 63%,
and SO
2−
4
in precipitation by 40%, between 1985 and 1996 (ref. 2). In
the United States and Canada, SO
2
emissions declined 28% between
1980 and 1995 (ref. 1); sulphur deposition in the same period
decreased by 29% in the northeastern United States and by 35% in
the upper midwestern United States1. In eastern Canada, decreases
in SO
2−
4
in precipitation ranged from 31% in Atlantic Canada to
36% in south/central Ontario5. In Great Britain, both emissions
(32%) and wet deposition (43%) of sulphur declined between 1979
and 1993, with most of the decrease observable in the early 1980s
(ref. 6). Few of these regions have recorded changes in the rate of
nitrogen deposition during this time period. These pervasive
declines in rates of acidic deposition create an expectation of
widespread recovery in acidified surface waters. Reasoning that
consistent recovery across a regional population of streams or lakes
would provide the strongest evidence for the success of controls on
acidic deposition, we report here on surface water trends at the
regional scale.
The primary source of data for this analysis is the International
Cooperative Programme on Assessment and Monitoring of Acidi-
fication of Rivers and Lakes (ICP-Waters), an international body
Midwestern
North America
South/Central
Ontario
Quebec/Vermont
Maine/
Atlantic
Canada
Adirondacks/
Catskills
Great Britain
Nordic Countries
North/Central
Europe
Figure 1 Lake and stream monitoring sites. Data from 205 long-term monitoring sites,
located in eight regions of Europe and North America, were used in this analysis.
© 1999 Macmillan Magazines Ltd
organized under the UN/ECE LRTAP. Additional data from
national, provincial and international monitoring programmes
(for example, in the United Kingdom7, Canada8, the United
States9and the International Cooperative Programme on Integrated
Monitoring in Europe10) were added in regions with inadequate
sample sizes. The sampling frequencies used by these programmes
varied substantially, from weekly in some cases, to quarterly or
biannually in others. In order to standardize the data used in trend
analyses, we chose quarterly sampling as a reasonable goal (achiev-
able for most sites) and calculated seasonal mean values for the
more temporally intensive data sets. We excluded sites with mean
Gran alkalinity greater than 200 microequivalents per litre (200
mequiv. l
−1
) in order to focus on those sites most likely to show
recovery11, as well as sites with fewer than three quarterly samples
(on average), less than 5 years of data per decade, or where climate-
driven trends have been documented. A total of 168 sites had
sufficient data for the 1980s, and 205 sites had sufficient data for
the 1990s. Sites were grouped into regions based on similarities in
their geochemistry, in deposition trends in the 1980s and 1990s, and
on their geographical proximity (Fig. 1).
Preliminary analyses indicated that temporal patterns in chemi-
cal variables were not monotonic; they tended to be curvilinear,
with an inflection point around 1990. For this reason, and also to
permit a comparison of trends between decades, we performed site-
specific trend tests separately on data from the 1980s and 1990s
using the non-parametric Seasonal Kendall test modified to account
for serial dependence12. In order to infer regional trends, we applied
a meta-analytical technique first described by van Belle and
Hughes13 that allows, with some restrictions, trend results from
multiple sites to be combined into a single estimate of trend.
In performing the regional trend tests, we first tested the Z-
statistics from individual Seasonal Kendall tests for homogeneity by
attributing the variance among them to each of the possible sources
of interest (sites, seasons and interactions) through an analysis of
variance, and tested the resulting sums of squares against a x
2
distribution14. There is no accepted threshold for deciding when
heterogeneity among sites is sufficient to invalidate combining the
trend results regionally. We followed the recommendation of van
Belle and Hughes13, set the threshold for tests of homogeneity
conservatively (for example, p,0:01), and examined the trend
statistics more critically if they failed this test. In some cases
(particularly for SO
2−
4
), Z-statistics failed the homogeneity test
despite having the same sign (for example, all Z,0). In these
cases, we conclude that the evidence for reporting a downward trend
is sufficient to regard the results as reliable.
At least two criteria should be fulfilled before extrapolation of
site-specific trends to the regional level11: (1) the sites must be
representative of the region or the regional subpopulation (for
example, acid-sensitive surface waters) of interest; and (2) the sites
should exhibit consistency of trend behaviour. We can test for the
latter condition through the homogeneity test described above, but
we have no current test for the former condition. Here we assume
that monitoring programmes have made good judgments in choos-
ing sites representative of acid-sensitive surface waters in each
region. All of the ICP-Waters sites were selected using published
criteria (for example, sites in headwater regions, free from local
disturbance, and located on sensitive geology)15. The national and
provincial programmes presented in this analysis used similar
criteria.
The strongest evidence for a regional surface-water response to
decreasing deposition comes from SO
2−
4
trends (Table 1). Sulphate
(we used non-marine SO
2−
4
in regions within 100 km of oceans) has
declined across most of Europe and North America, generally with
more rapid declines during the 1990s. Only Great Britain failed to
show a significant decline in SO
2−
4
in the 1990s. Further, SO
2−
4
trends
were generally homogeneous within a region, suggesting strongly
that SO
2−
4
trends are driven by changes in atmospheric deposition.
Only two subsets of data failed the homogeneity test. For south/
central Ontario during the 1980s, all of the Z-statistics were
negative, indicating decreasing SO
2−
4
, and we consider the regional
trend to be valid despite high x
2
values. Vermont/Quebec in the
1980s was the only region that exhibited heterogeneity strong
enough to invalidate a significant trend test; all of the sites with
increasing SO
2−
4
in this region and decade were located in Quebec,
although none of the positive site-specific trends were significant.
Nitrate increases were nearly universal, but largely restricted to
the 1980s. The largest increases, and probably the only ecologically
significant ones, occurred in the Adirondack/Catskill mountains
and north/central Europe. The nitrate increases in these regions
during the 1980s have received considerable attention, generating
concern that these high-deposition regions are experiencing nitro-
gen saturation16. Interestingly, these two regions also show the
largest reversals of NO
−
3
trends in the 1990s.
Both conceptual17 and mechanistic18 models of acidification
suggest that decreases in acid anion concentrations (S½SO22
4þ
NO2
3ÿ) should be balanced by smaller decreases (relative to acid
anions) in base cations (C
B
) and increases in alkalinity (which we
define as recovery). Large-scale patterns in recovery were evident
across all the European regions (north/central Europe, the Nordic
countries and Great Britain) during the 1990s (Table 1, Fig. 2). The
lack of recovery in the Nordic countries in the 1980s has been
reported elsewhere19, and declines in C
B
identified as the likely
mechanism preventing alkalinity from recovery despite large
decreases in sulphur deposition. Our analysis (Fig. 2) reinforces
this interpretation, but highlights how this pattern changed in the
1990s. Nordic trends in C
B
in the 1990s are relatively flat (Table 1),
and continued strong decreases in SO
2−
4
in this region appear to be
leading to a recovery in alkalinity. It should be noted that the
‘‘recovery’’ observed in Great Britain does not appear to be driven
by changes in acid anions (non-marine SO
2−
4
is unchanged in the
letters to nature
576 NATURE
|
VOL 401
|
7 OCTOBER 1999
|
www.nature.com
Figure 2 Regional acidification trend results. Histograms exhibit slopes of regional trends
in a, acid anions (SO22
4þNO2
3), b, measured (Gran) alkalinity, and c, base cation (C
B
)
concentrations for eight regions of Europe and North America for the 1980s and 1990s.
All slopes are in mequiv. l
−1
yr
−1
.
© 1999 Macmillan Magazines Ltd
1990s, while NO
−
3
has increased slightly), but instead results from
increasing C
B
concentrations (this is the only region to exhibit this
phenomenon). The pattern in Great Britain appears to be driven by
climate variations, particularly as they affect the transport of seasalt
aerosols across the region, rather than by changes in anthropogenic
emissions20.
Despite large decreases in acid anions in south/central Ontario,
the Adirondack/Catskill mountains and midwestern North America
in the 1990s, there was either no regional recovery in alkalinity or
continued acidification (Fig. 2); we attribute this pattern to com-
pensating negative trends in C
B
. Although some decrease in C
B
is
expected as SO
2−
4
concentrations decline, C
B
declines in excess of
SO
2−
4
declines are unexpected, and will preclude any recovery in
alkalinity21. Among the mechanisms that have been hypothesized to
produce excessive C
B
declines are (1) decreasing C
B
deposition22,
and (2) cation depletion of watershed soils23,24. While the former
mechanism has been discounted in at least some regions14,itis
possible that long-term high rates of acidic deposition have leached
enough cations from sensitive soils that adsorbed cation pools are
severely depleted. If this depletion continues even at diminished
levels of acidic deposition, it could produce the steep C
B
trends we
observe in these regions. The possibility that this mechanism will
prevent or delay recovery deserves critical attention.
It is important to consider the effects that climate variation could
have on the trends we report, especially for the relatively short 1990s
time series we use in our analysis. The effects of climate on non-
marine SO
2−
4
trends and C
B
behaviour in Great Britain have already
been mentioned. Droughts in both south/central Ontario and
midwestern North America have been shown to cause re-oxidation
of reduced sulphur in wetlands and lake littoral regions25,26. The
resulting increase in sulphate export can be relatively long-term,
temporarily delaying the recovery response in lakes with plentiful
wetlands in their watersheds; this mechanism undoubtedly con-
tributes to the lack of homogeneity in SO
2−
4
trends observed in
south/central Ontario. In midwestern North America, however,
where long water residence times create a great potential for climatic
effects27, declines in SO
2−
4
in the 1990s far exceed what is attributable
to a recovery from drought in the 1980s28; these trends doubtless
reflect declining rates of deposition. Climate may also play a role in
producing strongly decreasing C
B
trends. In lakes with long resi-
dence times, prolonged periods of drought produce increased C
B
concentrations and decreased rates of C
B
export29; as lakes recover
from drought, C
B
concentrations may decline rapidly (for example,
in drainage lakes, where increases result primarily from evapotran-
spiration) or slowly (for example, in seepage lakes, where increases
result from changes in groundwater flowpaths)27. Although data
from the current study are not sufficient to identify mechanisms
incontrovertibly, the widespread declines in SO
2−
4
we report are
completely consistent with observed changes in rates of sulphur
deposition in Europe and North America. Lack of lake and stream
recovery, and particularly the strong decreases in C
B
concentrations
observed in some regions, may be due to a number of mechanisms,
including climate fluctuations.
The three North American regions where recovery might be
expected but was not observed (south/central Ontario, Adiron-
dack/Catskill mountains and midwestern North America) exhibit
the largest North American declines in acid anions and largest
decreases in C
B
in the 1990s, and are closest to significant emissions
sources in the midwestern United States. Analysing whether the
proximity of these regions to significant sources (and therefore
higher, and possibly longer-term, inputs of acidic deposition) is
related to their C
B
behaviour is beyond the scope of this Letter.
These trend patterns are very similar to those observed in the 1980s
in the Nordic countries, where recovery is now occurring. Increas-
ing trends in alkalinity and unchanging concentrations of C
B
in the
Nordic countries during the 1990s may be the result of higher rates
of SO
2−
4
decline (Fig. 2a), or may reflect the cumulative effects of
longer-term declines in sulphur deposition (beginning in the late
1960s) in this region30. Long-term decreases in sulphur deposition
could be expected to result in the recovery of soil cation pools (as
rates of primary weathering begin to exceed loss rates due to
leaching by acid anions), and eventual recovery after a sufficient
time lag17. The recovery pattern in the Nordic countries suggests
that larger decreases in sulphur deposition and/or a longer response
time may be required before similar recovery is widely observed in
North America. This suggestion of lagged recovery, coupled with
evidence of the deleterious effects of climate variability in detecting
recovery, highlights the importance of continued coordinated
letters to nature
NATURE
|
VOL 401
|
7 OCTOBER 1999
|
www.nature.com 577
Table 1 Regional trend
Region Decade N*SO
2−
4
trend NO
−
3
trend Alkalinity trend C
B
trend
pSlope pSlope pSlope pSlope
(mequiv. l
−1
yr
−1
)(mequiv. l
−1
yr
−1
)(mequiv. l
−1
yr
−1
)(mequiv. l
−1
yr
−1
)
...................................................................................................................................................................................................................................................................................................................................................................
North/central Europe 1980s 12 ,0.001 −5.8 0.011 +1.3 0.089 +2.5 0.921 −2.0
1990s 23 p0.001 −5.9 p0.001 −2.4 ,0.001 +7.0 0.909 −0.3
...................................................................................................................................................................................................................................................................................................................................................................
Nordic countries 1980s 27 p0.001 −0.8 0.002 +0.1 ,0.001 −1.6‡p0.001 −2.6‡
1990s 33 p0.001 −3.1 0.418 −0.0 p0.001 +2.1 0.778 −0.0
...................................................................................................................................................................................................................................................................................................................................................................
Great Britain 1990s 18 0.510 +0.0 0.024 +0.1 p0.001 +1.5 p0.001 +2.4
...................................................................................................................................................................................................................................................................................................................................................................
Maine/Atlantic Canada 1980s 10 0.264 −0.0 0.829 −0.0 ,0.001 +1.2 0.180 −0.9
1990s 10 p0.001 −1.3 0.883 −0.0 0.111 −0.7 0.685 −0.1
...................................................................................................................................................................................................................................................................................................................................................................
Vermont/Quebec 1980s 47 p0.001 −1.1† ,0.001 +0.0 0.515 +0.1 ,0.001 −0.6
1990s 49 p0.001 −3.5 0.133 −0.0 0.028 +0.9 0.007 −1.1
...................................................................................................................................................................................................................................................................................................................................................................
South/central Ontario 1980s 13 p0.001 −1.2‡ 0.002 +0.0§ p0.001 +1.0† 0.030 +0.4
1990s 13 p0.001 −5.8 0.591 −0.0 0.817 −0.5 p0.001 −5.9‡
...................................................................................................................................................................................................................................................................................................................................................................
Adirondack/Catskill mountains 1980s 25 p0.001 −1.7 p0.001 +0.7k,0.001 −1.3 0.775 +0.8
1990s 25 0.012 −0.9 p0.001 −2.4 0.426 +0.1 ,0.001 −4.5
...................................................................................................................................................................................................................................................................................................................................................................
Midwestern North America 1980s 34 0.023 −1.2 ,0.001 +0.1§ 0.123 −0.9 ,0.001 +1.5
1990s 34 p0.001 −2.6 0.782 −0.0 p0.001 −1.1 p0.001 −2.8
...................................................................................................................................................................................................................................................................................................................................................................
Results of meta-analysis of trends in SO
2−
4
,NO
−
3
, alkalinity and base cation (C
B
) concentration for 1980s and 1990s in various regions. Statistics (pvalues based on x
2
tests) are from an analysis of variance of
site trend Zscores, and constitute a test of trend homogeneity within the regions. x
2
values were considered significant at p,0:01 for site and season effects. Statistics for a site– season interaction term are
not shown, as they were never significant. Significant trends ( p,0:05) are shown in bold.
* Sample sizes vary among variables. Nvalues listed are for SO
2−
4
.
† Site heterogeneity is significant; trend statistics are questionable.
‡ Site heterogeneity is significant, but all slopes are negative; trend statistics considered valid.
§ Seasonal heterogeneity significant; trend statistics questionable.
kSeasonal heterogeneity significant, but all slopes are positive; trend statistics considered valid.
© 1999 Macmillan Magazines Ltd
international monitoring to assess the success of acidic deposition
control measures. M
Received 20 April; accepted 19 August 1999.
1. Shannon, J. D. Regional trends in wet deposition of sulfate in the United States and SO
2
emissions
from 1980 through 1995. Atmos. Environ. 33, 807–816 (1999).
2. Tarrasson, L. (ed.) Transboundary Acidifying Air Pollution in Europe (International Cooperative
Programme for the Monitoring and Evaluation of Long Range Transmission of Air Pollutants in
Europe/Meteorological Synthesizing Centre-West, 1998).
3. Skjelkva
˚le, B. L., Wright, R. F. & Henriksen, A. Norwegian lakes show widespread recovery from
acidification; results from national surveys of lakewater chemistry 1986– 1997. Hydrol. Earth System
Sci. 2, 375–577 (1998).
4. Driscoll, C. T., Likens, G. E. & Church, M. R. Recovery of surface waters in the northeastern U.S. from
decreases in atmospheric deposition of sulfur. Wat. Air Soil Pollut. 105, 319–329 (1998).
5. Summers, P.W. Time trend of wet deposition acidifying potential at five ecological monitoring sites in
Eastern Canada 1981–1993. Wat. Air Soil Pollut. 85, 653–658 (1995).
6. Downing, C. E. H., Vincent, K. J., Campbell, G. W., Fowler, D. & Smith, R. I. Trends inwet and dr y
deposition of sulphur in the United Kingdom. Wat. Air Soil Pollut. 85, 659–664 (1995).
7. Patrick, S., Monteith, D. T. & Jenkins, A. UK Acid Waters Monitoring Network: The First Five Years
(Department of the Environment, ENSIS Publishing, London, 1995).
8. Jeffries, D. S. 1997 Canadian Acid Rain Assessment Vol. 3, Aquatic Effects (Environment Canada,
Ottawa, 1997).
9. Ford, J., Stoddard, J. L. & Powers, C. F.Perspectives in environmental monitoring: an introduction to
the U.S. EPA Long-Term Monitoring (LTM) project. Wat. Air Soil Pollut. 67, 247–255 (1993).
10. Kleemola, S. & So
¨derman, G. Manual for Integrated Monitoring, Programme Phase 1993– 1996
(Environmental Data Center, National Board of Waters and the Environment, Helsinki, 1993).
11. Stoddard, J. L., Driscoll, C. T.,Kahl, S. & Kellogg, J. Can site-specific trends be extrapolated to a region?
An acidification example for the Northeast. Ecol. Applic. 8, 288– 299 (1998).
12. Hirsch, R. M. & Slack, J. R. A nonparametric trend test for seasonal data with serial dependence. Wat.
Resour. Res. 20, 727–732 (1984).
13. van Belle, G. & Hughes, J. P. Nonparametrictests for trend in water quality. Wat. Resour.Res. 20, 127 –
136 (1984).
14. Stoddard, J. L., Driscoll, C. T.,Kahl, J. S. & Kellogg, J. A regional analysis of lakeacidification trends for
the northeastern U.S., 1982–1994. Environ. Monit. Assess. 51, 399 –413 (1998).
15. International Cooperative Programme on Assessment and Monitoring of Acidification of Rivers and
Lakes. Programme Manual (Programme Centre, Norwegian Inst. for Water Research, Oslo, 1995).
16. Stoddard, J. L. in Environmental Chemistry of Lakes and Reservoirs (ed. Baker, L. A.) 223– 284
(American Chemical Society, Washington DC, 1994).
17. Galloway, J. N., Norton, S. A. & Church, M. R. Freshwateracidification from atmospheric deposition
of sulfuric acid: A conceptual model. Environ. Sci. Technol. 17, 541–545 (1983).
18. Cosby, B. J., Hornberger,G. M., Galloway, J. N. & Wright,R. F. Time scales of catchment acidification:
A quantitative model for estimating freshwater acidification. Environ. Sci. Technol. 19, 1144– 1149
(1985).
19. Kirchner, J. W. & Lydersen, E. Base cation depletion and potential long-term acidification of
Norwegian catchments. Environ. Sci. Technol. 29, 1953–1960 (1995).
20. Evans, C., Monteith, D. T. & Harriman, R. Long-term variability in the deposition of marine ions at
west coast sites in the U.K. Acid Waters Monitoring Network: Impacts on surface waterchemistr y and
significance for trend determination. Sci. Tot. Environ. (in the press).
21. Driscoll, C. T., Likens, G. E., Hedin, L. O., Eaton, J.S. & Bormann, F. H. Changes in the chemistry of
surface waters: 25-year results at the Hubbard Brook Experimental Forest. Environ. Sci. Technol. 23,
137–143 (1989).
22. Hedin, L. O. et al. Steep declines in atmospheric base cations in regions of Europe and North America.
Nature 367, 351–354 (1994).
23. Likens, G. E., Driscoll, C. T. & Buso, D. C. Long-term effects of acid rain: Response and recovery ofa
forest ecosystem. Science 272, 244–246 (1996).
24. Lawrence, G. B. et al. Soil calcium status and the response of stream chemistry to changing acidic
deposition rates in the Catskill Mountains, New York. Ecol. Applic. (in the press).
25. Bayley, S. E., Schindler, D. W., Parker, B. R., Stainton, M. P. & Beaty, K. G. Effects of forest fire and
drought on the acidity of a base-poor boreal forest stream: similarities betweenclimatic warming and
acidic precipitation. Biogeochemistry 17, 191– 204 (1992).
26. Dillon, P. J., Molot, L. A. & Futter,M. A note on the effects of El Nin
˜o-related drought on the recovery
of acidified lakes. Int. J. Environ. Monit. Assess. 46, 105– 111 (1997).
27. Webster, K. E. et al. Structural features of lake districts: Landscape controls on lake chemical responses
to drought. Freshwat. Biol. (in the press).
28. Webster, K. E. & Brezonik, P. O. Climate confounds detection of chemical trends related to acid
deposition in Upper Midwest lakes in the USA. Wat. Air Soil Pollut. 85, 1575–1580 (1995).
29. Schindler, D. W.et al. The effects of climatic warming on the properties of boreal lakes and streams at
the Experimental Lakes Area, northwestern Ontario. Limnol. Oceanogr. 41, 1004–1017 (1996).
30. Mylona, S. Trends of Sulphur Dioxide Emissions, Air Concentrations and Depositions of Sulphur in
Europe since 1880 (International Cooperative Programme for the Monitoring and Evaluation of Long
Range Transmission of Air Pollutants in Europe/Meteorological Synthesizing Centre-West, 1994).
Acknowledgements
We thank the LRTAP Working Group on Effects for their support; this Working Group
supports the production of international, quality-controlled, comparable data. We also
acknowledge the work of the ICP Programme Centre at the Norwegian Institute of Water
Research (NIVA), where the data are collated, verified and archived, and thank T. J.
Sullivan and M. R. Church for comments and suggestions. This work was supported by the
US Environmental Protection Agency.
Correspondence and requests for materials should be addressed to J.L.S.
(e-mail: stoddard@mail.cor.epa.gov).
letters to nature
578 NATURE
|
VOL 401
|
7 OCTOBER 1999
|
www.nature.com
.................................................................
Sterilization and canopy modification
of a swollen thorn acacia tree
by a plant-ant
Maureen L. Stanton*†, Todd M. Palmer†‡, Truman P. Young*†‡,
Amanda Evans§& Monica L. Turner†
*Center for Population Biology, University of California, Davis, California 95616,
USA
†Mpala Research Centre, PO Box 555, Nanyuki, Kenya
‡Department of Environmental Horticulture, University of California, Davis,
California 95695, USA
§Department of Biology, University of Oregon, Eugene, Oregon 97403, USA
.......................................... ......................... ......................... ......................... .........................
Obligate symbioses between specialized arboreal ants and plants
have evolved independently in many lineages1,2. Ant-plants (myr-
mecophytes) typically provide hollow nest cavities and nutrition
to the occupying ant colony1,3–6. In turn, resident plant-ants
often protect their hosts from herbivory7–11 and/or overgrowth
by surrounding vegetation12,13. As individual plants are rarely
occupied by more than one ant colony14–17, co-occurring plant-
ant species compete intensely for hosts13,14,18,19. In such multi-
species systems, ecological interactions among potential partners
may lead to the evolution of cheating20,21. Previous studies have
revealed that some specialized plant-ants are effectively parasites
of their host-plants8,18,22,23, but the selection pressures favouring
such behaviours are poorly understood. Here we describe host
parasitism in an east African plant-ant that prunes and sterilizes
its host-tree canopies, apparently to minimize contact with
competitively dominant ants occupying neighbouring trees. We
propose that the high density of ant-trees and low diversity of
tree species in this savanna habitat have selected for induced,
parasitic pruning of host trees by this competitively subordinate
ant species.
Whistling Thorn (Acacia drepanolobium) trees dominate vast
areas of savanna on black cotton soils in upland east Africa24,25.At
each branch node, A. drepanolobium trees produce either a pair of
slender thorns or a swollen thorn pair joined by a bulbous, hollow
base 1.5–6 cm in diameter (Fig. 1). When living on a tree, ants chew
Figure 1 Close-up of
A. drepanolobium
branches showing axillary leaves, paired slender
thorns and a node occupied by a fused swollen thorn pair. To show architectural features
clearly, this photograph was taken at the beginning of the rainy season, when relatively
few axillary leaves were persistent on older growth. Resident ants had not yet chewed
entry holes into the recently formed swollen thorn (centre). Swollen thorns are produced
even in the absence of ants26.