Content uploaded by Amy T Austin
Author content
All content in this area was uploaded by Amy T Austin
Content may be subject to copyright.
Plant litter decomposition in a semi-arid ecosystem
controlled by photodegradation
Amy T. Austin
1
& Lucı
´
a Vivanco
1
The carbon balance in terrestrial ecosystems is determined by the
difference between inputs from primary production and the
return of carbon to the atmosphere through decomposition of
organic matter
1
. Our understanding of the factors that control
carbon turnover in water-limited ecosystems is limited, however,
as studies of litter decomposition have shown contradictory
results and only a modest correlation with precipitation
2–5
. Here
we evaluate the influence of solar radiation, soil biotic activity and
soil resource availability on litter decomposition in the semi-arid
Patagonian steppe using the results of manipulative experiments
carried out under ambient conditions of rainfall and temperature.
We show that intercepted solar radiation was the only factor that
had a significant effect on the decomposition of organic matter,
with attenuation of ultraviolet-B and total radiation causing a 33
and 60 per cent reduction in decomposition, respectively. We
conclude that photodegradation is a dominant control on above-
ground litter decomposition in this semi-arid ecosystem. Losses
through photochemical mineralization may represent a short-
circuit in the carbon cycle, with a substantial fraction of carbon
fixed in plant biomass being lost directly to the atmosphere
without cycling through soil organic matter pools. Furthermore,
future changes in radiation interception due to decreased cloudi-
ness, increased stratospheric ozone depletion, or reduced vegeta-
tive cover may have a more significant effect on the carbon balance
in these water-limited ecosystems than changes in temperature or
precipitation.
Traditional models of the controls on litter decomposition in
terrestrial ecosystems have focused on how soil organisms interact
with climate and litter quality to control mass loss and nutrient
release of senescent plant material
6–8
. Although litter decomposition
does positively correlate with annual precipitation at the regional
scale
6,9,10
, these empirical models cannot account for the rapid turn-
over of organic material observed in arid ecosystems
2,3,11
, or the lack
of correlation of decomposition with rainfall in deserts
4,5
. These
conflicting results suggest that there are unique factors affecting
decomposition in arid and semi-arid ecosystems, which may include
abiotic controls such as photodegradation or limitation on biotic
activity from low soil resource availability. At the same time, the net
effect on carbon sequestration in terrestrial ecosystems will result
from the balance between changes in primary production and
decomposition
12,13
, and as such, the controls on decomposition in
water-limited ecosystems are crucial parameters for predicting future
impacts of global change.
Photochemical mineralization of organic material results in a
reduction of the average molecular mass of organic compounds,
alteration of the capacity to absorb light both in the ultraviolet and
visible spectrum, and the formation of novel photoproducts
14,15
.
Solar radiation plays an important role in the turnover of organic
matter in aquatic ecosystems and oceans: photochemical reactions
change the quality of dissolved organic matter (DOM)
14,16
and
produce dissolved inorganic carbon and volatile CO
2
, CO and
carbonyl sulphides
15,17
. Photochemical production of pyruvate
from DOM in ocean surface waters, for example, is believed to be a
critical transformation to permit biological degradation of recalcitrant
DOM
14
. Although dissolved inorganic carbon (in the form of CO) is
quantitatively one of the most important photoproducts observed
during photochemical mineralization of DOM
16
, most studies suggest
that the indirect effects on lability of organic matter, and not direct
production of photoproducts, are the most important influences of
solar radiation on the carbon cycle in aquatic ecosystems
14,16
.
There are very few studies of the direct effects of solar radiation on
carbon turnover in terrestrial ecosystems, although CO
2
production
from sterilized litter subjected to radiation treatments has been
observed in emergent macrophytic vegetation
18
and carbon monoxide
has been detected as a photoproduct from live and senescent plant
material in short-term incubations in grasslands
17,19
. Ultraviolet-B
radiation (UV-B) has been shown to affect litter decomposition
through changes in the chemical composition of the litter, or through
changes in the microbial community characteristics
20,21
. However,
large direct UV-B effects on carbon turnover are undocumented, and
only small negative and positive direct effects have been observed
21–23
.
In contrast, simulations for deserts using the ecosystem model
CENTURY suggest that direct effects of UV-B through increased
photodegradation of litter could be important
24
. Photodegradation
has been mentioned as a possible control on carbon turnover in
studies of litter decomposition in water-limited ecosystems
2,11
,
although this effect has never been quantified empirically. As such,
the quantitative importance at the ecosystem scale of photodegrada-
tion effects on carbon turnover in arid and semi-arid ecosystems,
from both direct effects on the production of volatile compounds,
and indirect effects on the quality of organic matter, is unknown.
We conducted two factorial experiments under ambient con-
ditions of rainfall and temperature designed to disentangle the
abiotic and biotic controls on litter decomposition in a semi-arid
ecosystem. In the first experiment, recently senesced litter of mixed
native grasses was decomposed under different light regimes (full
sunlight, reduced UV-B radiation, and blocked total radiation) in
combination with a treatment of soil biocide to reduce biotic activity.
In a second parallel experiment, the same litter mixture was decom-
posed, treating underlying soils with additions of labile carbon and
inorganic nitrogen to stimulate biotic activity and remove possible
resource limitation for decomposer organisms.
Manipulations of biotic activity and resource availability produced
dramatic changes in soil characteristics. There was a significant
decline in the soil microbial biomass C (P , 0.05), potential soil
respiration (P , 0.001) and soil nitrate concentrations (P , 0.05) of
LETTERS
1
Instituto de Investigaciones Fisiolo
´
gicas y Ecolo
´
gicas Vinculadas a la Agricultura (IFEVA) and Consejo Nacional de Investigaciones Cientı
´
ficas y Te
´
cnicas (CONICET), Facultad
de Agronomı
´
a, Universidad de Buenos Aires, Av. San Martı
´
n 4453, Buenos Aires (C1417DSE), Argentina.
Vol 442|3 August 2006|doi:10.1038/nature05038
555
© 2006 Nature Publishing Group
biocide-treated soils (Table 1). In addition, there was a significant
effect of biocide application on microbial colonization of litter for
both cultivable fungal and bacterial populations (Table 2,
P , 0.001). The biocide application eliminated cultivable litter
fungal populations entirely and reduced cultivable litter bacterial
colonization by an order of magnitude. In the substrate addition
experiment, soil additions of labile C and inorganic N resulted in
increased soil microbial biomass (P , 0.05), increased NO
3
2
and
NH
4
þ
concentrations in the N addition treatment (P , 0.0001), and
increased potential C mineralization with both C and C þ N
additions (Table 1, P , 0.01).
In spite of considerable changes in biotic activity and soil resource
availability, solar radiation was the only variable that significantly
affected litter decomposition (Fig. 1, P , 0.001). Decomposition,
expressed as a constant (k) that integrates the rates of organic matter
loss over time, showed a 33% and 60% reduction in litter decompo-
sition in the reduced UV-B and blocked total radiation treatments,
respectively (Fig. 1b, P , 0.0001). In contrast, there was no effect of
biocide application on rates of organic matter loss or k constants of
decomposition (Fig. 1, P . 0.05); moreover, soil substrate additions
had no effect on litter decomposition (Fig. 2, P . 0.05). Finally, there
was no interaction between radiation and biocide treatments for
organic matter loss (P . 0.05), suggesting that the principal effect of
radiation exclusion was due to direct effects on photochemical
mineralization of organic matter and not interaction with increased
lability of substrates for soil biota. The organic matter loss observed
in the total blocked radiation treatments (10% at the end of the
incubation period, Fig. 1a) could be the result of abiotic processes
other than photodegradation, such as physical fragmentation or
leaching. At the same time, these possible effects do not appear to
have been biased by soil temperature differences in the radiation
treatments. Overall mean temperature among radiation treatments
from all sampling dates did not differ significantly (P . 0.05, see
Supplementary Information).
Taken together, these results suggest that photodegradation is a
dominant control of above-ground litter decomposition in this semi-
arid ecosystem, and that UV-B radiation alone can account for
almost 50% of the carbon lost due to photochemical mineralization
of litter, a much larger effect of UV-B than has been previously
reported
22–25
. Ultraviolet-A (UV-A) and/or short wavelength visible
radiation have been shown to affect gaseous carbon losses from
terrestrial vegetation
18,19
, and the results from our study support the
observation that solar radiation other than UV-B also affected
photodegradation, seen in the significant differences in organic
matter loss between the attenuated UV-B and total blocked radiation
treatments. Surprisingly, alteration of biotic activity, through inhi-
bition with biocide and stimulation with labile carbon and nitrogen
Table 1 | Soil characteristics of litter decomposition experiments
C mineralization
(
m
g C per g soil d
21
)
Microbial biomass
(
m
g C per g dry soil)
Soil NH
4
-N
(
m
g per g dry soil)
Soil NO
3
-N
(
m
g per g dry soil)
Soil H
2
O
(%)*
Radiation and biocide
(2)Biocide 5.3 ^ 0.50
a
121 ^ 15
a
3.31 ^ 0.35
a
0.64 ^ 0.09
a
2.5 ^ 0.21
a
(þ)Biocide 1.5 ^ 0.87
b
64 ^ 18
b
9.26 ^ 1.40
b
0.20 ^ 0.08
b
2.8 ^ 0.25
a
Substrate addition
Control 6.1 ^ 0.60
a
111 ^ 20
a
3.99 ^ 0.66
a
0.76 ^ 0.09
a
3.7 ^ 0.24
a
Carbon 18.1 ^ 2.47
b
223 ^ 40
b
0.86 ^ 0.36
a
0.03 ^ 0.01
a
2.8 ^ 0.28
ab
Nitrogen 6.4 ^ 0.68
a
96 ^ 19
a
42.39 ^ 4.99
b
3.97 ^ 0.70
b
2.5 ^ 0.08
b
C and N 13.6 ^ 0.57
b
386 ^ 74
b
7.38 ^ 5.35
a
0.58 ^ 0.37
a
2.7 ^ 0.31
b
Experiments were performed in the Patagonian steppe; data shown were obtained at the last litter harvest. Data shown for potential C mineralization are laboratory incubations of 7 days
(n ¼ 5) and microbial biomass C, soil inorganic N and soil water from field measurements (n ¼ 5). Mean values are shown (^s.e.m.). Different letters indicate significant differences among
treatments for Tukey post-hoc comparisons (P , 0.05).
*Calculated as g per g dry soil.
Table 2 | Effect of biocide and substrate addition on microbial coloniza-
tion of litter
Experiment Treatment/substrate Fungi (CFU d
21
) Bacteria (CFU d
21
)
(2)Biocide Full sun 9,050 ^ 3,460 195 ^ 92
(2)Biocide (2)UV-B 3,037 ^ 545 153 ^ 66
(2)Biocide Blocked total 22,470 ^ 3,865 361 ^ 104
(þ)Biocide Full sun 0 3 ^ 1
(þ)Biocide (2)UV-B 0 20 ^ 9
(þ)Biocide Blocked total 0 22 ^ 10
Substrate addition Control 1,125 ^ 982 245 ^ 56
Substrate addition C 1,235 ^ 745 162 ^ 23
Substrate addition N 604 ^ 150 325 ^ 84
Substrate addition C and N 770 ^ 438 301 ^ 176
Mean values (n ¼ 5) are shown (^s.e.m.) for colony forming units (CFUs) of fungi and
bacteria from litter extracts after 24 h of cultivation. Litter from substrate addition
experiments was not subject to either biocide or radiation treatments.
Figure 1 | Effect of solar radiation and biocide on litter decomposition in the
Patagonian steppe.
a, Organic matter remaining over time. Symbols
indicate different treatments of radiation exclusion and biocide application
(n ¼ 5); see Methods for details. Error bars, ^s.e.m. b, Constants of
decomposition, k. These were obtained by plotting ln(organic matter
remaining/initial organic matter) against time; k is the slope of the
regression (n ¼ 5). Error bars, ^s.e.m.
LETTERS NATURE|Vol 442|3 August 2006
556
© 2006 Nature Publishing Group
additions, had no effect on litter decomposition. These results
suggest that biotic activity exerts very little influence over rates of
organic matter loss of above-ground litter in this ecosystem. More-
over, indirect effects of photodegradation resulting in changes in the
lability of organic compounds, although it cannot be discounted
entirely, appear to be secondary to direct photochemical mineraliza-
tion of organic matter, a contrast with the relative importance of the
two effects observed in studies of aquatic ecosystems
15,16
.
The importance of photodegradation for organic matter turnover
in the Patagonian steppe suggests a ‘short-circuit’ in the carbon cycle,
with implications for the functioning of semi-arid ecosystems. The
loss of carbon through inorganic photoproducts may be a major
pathway of carbon loss from above-ground plant biomass in this
ecosystem, with a direct return of inorganic carbon to the atmos-
phere without cycling through soil organic matter pools. Patchy
ecosystem structure with over 50% exposed bare soil
26,27
,high
radiation interception at the soil surface, large amounts of standing
dead material and a low number of rainy days are typical of arid and
semi-arid ecosystems. These characteristics combine a set of unique
conditions under which photodegradation can dominate decompo-
sition of above-ground litter, and could explain the direct, large
effects of UV-B radiation that have not been observed in more humid
ecosystems at higher latitudes with contrasting dominant vegetation
and litter quality
21–23
. A quantification at the ecosystem scale of this
‘short-circuit’ in the carbon cycle could rectify current discrepancies
in modelled carbon balances of arid and semi-arid ecosystems, and
explain the lack of correlation with traditional models of biotic
controls on decomposition based on variation in climate and litter
quality
6
.
Ecosystem-scale processes governing the carbon balance in water-
limited ecosystems may respond differently to global change owing to
differential controls on production and decomposition
9
. Effects of
global change on carbon fixation in arid and semi-arid ecosystems
include direct effects of elevated carbon dioxide on net primary
production
28
, as well as indirect effects on shrub/grass ratio, vegetative
cover and climate patterns
29,30
. Our results suggest that the direct
effects of solar radiation on above-ground decomposition over-
shadow the importance of other controls such as water availability,
which have shown little or no correlation with litter decomposition
in deserts
2,4
, and smaller effects on organic mass loss in rainfall
manipulation experiments
5,27
. Changes in radiation dose due to
decreased cloud cover, increased ozone depletion or reduced vege-
tative cover could directly affect carbon loss
19
, and could be more
important than changes in rainfall amount or climatic variability. As
close to 40% of the terrestrial land surface is currently classified as
arid or semi-arid, our understanding of how human-induced global
change will affect key controls on the carbon cycle in these ecosystems
is critical, and factors affecting rates of photochemical mineralization
could have consequences for the potential of carbon sequestration in
water-limited ecosystems and the global carbon balance.
METHODS
Study site. The Instituto Nacional de Tecnologı
´
a Agropecuaria (INTA) Rı
´
o
Mayo experimental station is located in the Argentinean province of Chubut
(458 41
0
S, 708 16
0
W) at 500 m elevation. Long-term mean annual precipitation
is 152 mm, and is strongly seasonal, with .70% of the precipitation falling in
winter months. Monthly mean temperature ranges from 15 8C in January to 1 8C
in July. The vegetation is classified as semi-arid steppe, with the dominant
vegetation of perennial tussock grasses (32% basal cover) and shrubs (15% basal
cover)
26
. Soils are coarsely textured aridisols, with high gravel content and low
soil-water holding capacity. Solar irradiance for the period of study is shown in
Supplementary Fig. 2.
Field methods. For both experiments, a representative mixture of leaf litter from
the site’s dominant perennial grass species (Stipa speciosa, Poa ligularis, Stipa
humilis) was collected and separated from green and standing dead material to
include only recently-senesced litter. We established 50 plots of 1 m
2
within a
grazing exclosure. In each plot, we removed all above-ground vegetation to
minimize shading and plant-microbial competition. For the radiation/biocide
experiment, plastic frames of 20 £ 10 £ 6 cm were used to construct litter boxes.
We attached a plastic filter on the top and the northern face of the frame that
corresponded to one of the three light treatments: (1) Aclar filters, which allowed
transmission of .95% of solar radiation (full sun); (2) Mylar filters, which
attenuated all radiation below 310 nm, ‘(2)UV-B’; and (3) Mylar filters covered
with reflective aerosol paint that effectively blocked .90% of solar radiation
(blocked total). Perforations were made in the filters to allow water infiltration
by precipitation. Fibreglass mesh (2 mm) was attached to the underside and
lateral sides of the boxes, so that the litter was in direct contact with the soil but
received adequate ventilation. We placed one gram of recently senesced litter in
five litterboxes in each randomly assigned treatment plot, which was represen-
tative of the spatial distribution of standing dead and litter detritus exposed to
solar radiation in this ecosystem. The soil and litter biocide treatment combined
a general fungicide at 17.5 g m
22
(Captan, containing N-(trichloromethylthio)
phthalimide) with a bactericide at 15.0 g m
22
(Agri-micina, containing strepto-
mycin sulphate and oxytetracycline) in an aqueous suspension distributed
uniformly (2 l m
22
) over the plot area. In addition, naphthalene pellets
(100 g m
22
) were placed on the soil surface and incorporated into the surface
soil by raking. In control plots, water alone was applied at 2 l m
22
. Applications
of biocide or water were made three times per year over the experimental period.
At each sampling date, filters were evaluated for damage or age, and replaced as
necessary. Soil temperature under litterboxes was evaluated using thermistors
buried at 2 cm depth at four time points during the day.
For the substrate addition experiment, litterbags (20 £ 10 cm) of fibreglass
mesh (2 mm) filled with one gram of recently-senesced litter were placed (n ¼ 5)
on the soil surface. Before litter placement and three times per year over the
experimental period, substrate additions of labile carbon, a combination of
glucose and cornstarch (annual rate of 330 g C m
22
yr
21
), and inorganic nitro-
gen, ammonium nitrate (annual rate of 40 kg N ha
21
yr
21
), were incorporated
into the first 5 cm of the soil with gentle raking. Litterboxes and litterbags were
collected at 3.5, 7, 12 and 18 months and analysed for changes in organic matter
loss over time. Ash-free dry mass was determined for all samples to correct for
soil contamination from the field.
Characteristics and rates of soil biotic activity. Gravimetric soil water content
and inorganic N content were determined for all dates (all data not shown);
inorganic N was evaluated using 2 N KCl extracts and measured colorimetrically
with an Alpkem autoanalyser. Extractions of soil microbial biomass, potential
soil C mineralization and bacterial and fungal colonization of litter for biocide
and substrate addition treatments were completed at the last litter harvest.
Microbial C was measured using a modified chloroform fumigation-extraction
Figure 2 | Effect of soil substrate additions of labile carbon and nitrogen on
litter decomposition in the Patagonian steppe.
a, Organic matter
remaining over time. Symbols indicate different treatments of soil substrate
additions (n ¼ 5); see Methods for details. Error bars, ^s.e.m. b, Constants
of decomposition, k (n ¼ 5). Error bars, ^s.e.m.
NATURE|Vol 442|3 August 2006 LETTERS
557
© 2006 Nature Publishing Group
technique with a conversion factor of 0.45 (K
c
). Potential soil carbon mineral-
ization was determined in laboratory incubations at 25 8C for 7 days using 50 g of
fresh soil, with addition of base traps of 0.1 M NaOH, titrated with 0.15 M BaCl
2
and 0.1 M HCl. Bacterial and fungal counts of litter were completed with
subsamples of 50–100 mg of litter from the final litter collection which were
extracted in a buffer solution (0.88% NaCl); 5 ml of suspension were spread-
plated on nutritive agar for determination of bacteria (nutritive agar) and fungi
(potato-dextrose agar, with 50
m
gml
21
tetracycline and 200
m
gml
21
streptomy-
cin). Plates were incubated at 25 8C for 24 h and colony-forming units (CFUs)
were counted.
Statistics. Data for the two experiments were analysed separately; a two-factor
analysis of variance (ANOVA) was completed for evaluation of litter decompo-
sition at each sampling date, for fungal/bacterial colonization, and k constants.
Soil characteristics were evaluated using a one- or two-way ANOVA. When
necessary, data were log-transformed to correct for violations of homogeneity of
variance.
Received 16 January; accepted 28 June 2006.
1. Olson, J. S. Energy storage and the balance of producers and decomposers in
ecological systems. Ecology 44, 322–-331 (1963).
2. Montan
˜
a, C., Ezcurra, E., Carrillo, A. & Delhoume, J. P. The decomposition of
litter in grasslands of northern Mexico: a comparison between arid and non-
arid environments. J. Arid Environ. 14, 55–-60 (1988).
3. Moorhead, D. L. & Reynolds, J. F. Mechanisms of surface litter mass loss in the
northern Chihuahuan desert: a reinterpretation. J. Arid Environ. 16, 157–-163
(1989).
4. Steinberger, Y., Shmida, A. & Whitford, W. G. Decomposition along a rainfall
gradient in the Judean desert, Israel. Oecologia 82, 322–-324 (1990).
5. Whitford, W. G. et al. Rainfall and decomposition in the Chihuahuan desert.
Oecologia 68, 512–-515 (1986).
6. Meentemeyer, V. Macroclimate and lignin control of litter decomposition rates.
Ecology 59, 465–-472 (1978).
7. Swift, M. J., Heal, O. W. & Anderson, J. M. Decomposition in Terrestrial
Ecosystems (Univ. California Press, Berkeley, 1979).
8. Aber, J. D. & Melillo, J. M. Nitrogen immobilization in decaying hardwood leaf
litter as a function of initial nitrogen and lignin content. Can. J. Bot. 60,
2263–-2269 (1982).
9. Austin, A. T. Differential effects of precipitation on production and
decomposition along a rainfall gradient in Hawai’i. Ecology 83, 328–-338
(2002).
10. Gholz, H. L., Wedin, D. A., Smitherman, S. M., Harmon, M. E. & Parton, W. J.
Long-term dynamics of pine and hardwood litter in contrasting environments:
toward a global model of decomposition. Glob. Change Biol. 6, 751–-765 (2000).
11. Vossbrinck, C. R., Coleman, D. C. & Wooley, T. A. Abiotic and biotic factors in
litter decomposition in a semiarid grassland. Ecology 60, 265–-271 (1979).
12. Schlesinger, W. H. & Lichter, J. Limited carbon storage in soil and litter of
experimental forest plots under increased atmospheric CO
2
. Nature 411,
466–-469 (2001).
13. Nemani, R. R. et al. Climate-driven increases in global terrestrial net primary
production from 1982 to 1999. Science 300, 1560–-1563 (2003).
14. Kieber, D. J., McDaniel, J. & Mopper, K. Photochemical source of biological
substrates in sea water: implications for carbon cycling. Nature 341, 637–-639
(1989).
15. Miller, W. L. & Zepp, R. G. Photochemical production of dissolved inorganic
carbon from terrestrial organic matter: significance to the oceanic organic
carbon cycle. Geophys. Res. Lett. 22, 417–-420 (1995).
16. Mopper, K. et al. Photochemical degradation of dissolved organic carbon and
its impact on the oceanic carbon cycle. Nature 353, 60–-62 (1991).
17. Tarr, M. A., Miller, W. L. & Zepp, R. G. Direct carbon monoxide production
from plant matter. J. Geophys. Res. 100, 11403–-11413 (1995).
18. Anesio, A. M., Tranvik, L. J. & Grane
´
li, W. Production of inorganic carbon from
aquatic macrophytes by solar radiation. Ecology 80, 1852–-1859 (1999).
19. Schade, G. W., Hormann, R. M. & Crutzen, P. J. CO emissions from degrading
plant matter. Tellus B 51, 899–-908 (1999).
20. Johnson, D., Campbell, C. D., Lee, J. A. & Callaghan, T. V. Arctic
microorganisms respond more to elevated UV-B radiation than CO
2
. Nature
416, 82–-83 (2002).
21. Pancotto, V. A., Sala, O. E., Robson, T. M., Caldwell, M. M. & Scopel, A. L.
Direct and indirect effects of solar ultraviolet-B radiation on long-term
decomposition. Glob. Change Biol. 11, 1982–-1989 (2005).
22. Gehrke, C., Johanson, U., Callaghan, T. V., Chadwick, D. & Robinson, C. H. The
impact of enhanced UV-B radiation on litter quality and decomposition
processes in Vaccinum leaves from the subarctic. Oikos 72, 213–-222 (1995).
23. Moody, S. A. et al. The direct effects of UV-B radiation on Betula pubescens
litter decomposing at four European field sites. Plant Ecol. 154, 29–-36 (2001).
24. Moorhead, D. L. & Callaghan, T. V. Effects of increasing UV-B radiation on
decomposition and soil organic matter dynamics. A synthesis and modeling
study. Biol. Fertil. Soils 18, 19–-26 (1994).
25. Pancotto, V. A. et al. Solar UV-B decreases decomposition in herbaceous plant
litter in Tierra del Fuego, Argentina: potential role of an altered decomposer
community. Glob. Change Biol. 9, 1465–-1474 (2003).
26. Sala, O. E., Golluscio, R. A., Lauenroth, W. K. & Soriano, A. Resource
partitioning between shrubs and grasses in the Patagonian steppe. Oecologia
81, 501–-505 (1989).
27. Yahdjian, L., Sala, O. E. & Austin, A. T. Differential controls of water input on
litter decomposition and nitrogen dynamics in the Patagonian steppe.
Ecosystems 9, 128–-141 (2006).
28. Smith, S. D. et al. Elevated CO
2
increases productivity and invasive species
success in an arid ecosystem. Nature 408, 79–-82 (2000).
29. Jackson, R. B., Banner, J. L., Jobba
´
gy, E. G., Pockman, W. T. & Wall, D. H.
Ecosystem carbon loss with woody plant invasion of grasslands. Nature 418,
623–-626 (2002).
30. Thomas, D. S. G., Knight, M. & Wiggs, G. F. S. Remobilization of southern
African desert dune systems by twenty-first century global warming. Nature
435, 1218–-1220 (2005).
Supplementary Information is linked to the online version of the paper at
www.nature.com/nature.
Acknowledgements We acknowledge the late A. Soriano for establishment of a
research program at the INTA study site more than 50 years ago; C. Mazza, L.
Raiger, P. Flombaum, N. Sala, J. Vrsalovic, P. Araujo, L. Gherardi, M. Gonzalez-
Polo, V. Marchesini, A. Ferna
´
ndez-Souto, P. Rojas, M. Taglizacchi and L. Yahdjian
for field and laboratory assistance; and O. Sala, P. Vitousek, K. O’Shea, G. Pin
˜
eiro
and C. Ballare
´
for comments on the manuscript. We acknowledge financial
support from the Fundacio
´
n Antorchas and the Fundacio
´
n YPF of Argentina, the
Inter-American Institute for Global Change Research, the US National Science
Foundation, the Agencia Nacional de Promocio
´
n de Ciencia y Tecnologı
´
a
(ANPCyT) and the Universidad de Buenos Aires (UBACyT) of Argentina.
Author Information Reprints and permissions information is available at
npg.nature.com/reprintsandpermissions. The authors declare no competing
financial interests. Correspondence and requests for materials should be
addressed to A.T.A. (austin@ifeva.edu.ar).
LETTERS NATURE|Vol 442|3 August 2006
558
© 2006 Nature Publishing Group
