Greater ecosystem carbon in the Mojave Desert after ten years exposure to elevated CO 2

Abstract

Carbon dioxide is the main greenhouse gas inducing climate change. Increased global CO2 emissions, estimated at 8.4 Pg C yr(-1) at present, have accelerated from 1% yr(-1) during 1990-1999 to 2.5% yr(-1) during 2000-2009 (ref. 1). The carbon balance of terrestrial ecosystems is the greatest unknown in the global C budget because the actual magnitude, location and causes of terrestrial sinks are uncertain(2); estimates of terrestrial C uptake, therefore, are often based on the residuals between direct measurements of the atmospheric sink and well-constrained models of ocean uptake of CO2 (ref. 3). Here we report significant terrestrial C accumulation caused by CO2 enhancement to net ecosystem productivity in an intact, undisturbed arid ecosystem(4-8) following ten years of exposure to elevated atmospheric CO2. Results provide direct evidence that CO2 fertilization substantially increases ecosystem C storage and that arid ecosystems are significant, previously unrecognized, sinks for atmospheric CO2 that must be accounted for in efforts to constrain terrestrial and global C cycles.

Full text

LETTERS
PUBLISHED ONLINE: 6 APRIL 2014 | DOI: 10.1038/NCLIMATE2184
Greater ecosystem carbon in the Mojave Desert
after ten years exposure to elevated CO2
R. D. Evans1*, A. Koyama1,2, D. L. Sonderegger1,3, T. N. Charlet4, B. A. Newingham4,5,
L. F. Fenstermaker6, B. Harlow1, V. L. Jin1,7, K. Ogle8, S. D. Smith4and R. S. Nowak9
Carbon dioxide is the main greenhouse gas inducing climate
change. Increased global CO2emissions, estimated at
8.4 Pg C yr−1at present, have accelerated from 1% yr−1during
1990–1999 to 2.5% yr−1during 2000–2009 (ref. 1). The carbon
balance of terrestrial ecosystems is the greatest unknown in
the global C budget because the actual magnitude, location
and causes of terrestrial sinks are uncertain2; estimates of
terrestrial C uptake, therefore, are often based on the residuals
between direct measurements of the atmospheric sink and
well-constrained models of ocean uptake of CO2(ref. 3). Here
we report significant terrestrial C accumulation caused by
CO2enhancement to net ecosystem productivity in an intact,
undisturbed arid ecosystem4–8 following ten years of exposure
to elevated atmospheric CO2. Results provide direct evidence
that CO2fertilization substantially increases ecosystem C
storage and that arid ecosystems are significant, previously un-
recognized, sinks for atmospheric CO2that must be accounted
for in eorts to constrain terrestrial and global C cycles.
Arid and semiarid ecosystems are significant components of
the terrestrial C budget; they cover 47% of the terrestrial surface9,
represent the fifth largest pool of soil organic C (208–241 Pg;
ref. 10) and exhibit large increases in net primary productivity
(NPP) in response to small changes in water availability11. The
Nevada Desert Free-Air CO2Enrichment Facility (NDFF) was
established in 1997 to better understand the sensitivity of arid
ecosystems to increasing atmospheric CO2([CO2]). Soil organic C
and nitrogen are concentrated in the top 0.1 m and no significant
differences in soil C and N were observed between CO2treatments
in 1999 (ref. 4). Above- and belowground biomass and soils
to 1 m were harvested by plant-cover type after ten years of
continuous treatment. Soils were the dominant pool of C and
N and contents were significantly greater under elevated [CO2]
across all cover types (Fig. 1 and Supplementary Tables 1–5). Mean
total ecosystem organic C under elevated CO2was 1,170 g C m−2
with a 90% credible interval of 1,062–1,285 g C m−2, compared
with 1,030 g C m−2(credible interval of 937–1,130 g C m−2)under
ambient conditions. Differences were owing solely to soil organic
C; no differences were observed in plant pools. This contrasts
with more mesic grassland and forested ecosystems that observed
increases in plant biomass after two to nine years of exposure
to elevated [CO2] (ref. 12). Mass balance analysis of the carbon
isotope composition (δ13C) of C entering the soil after a change
in CO2sources in 2003 was −26.2h(Fig. 2), indicating ~70%
of accrued soil organic C originated from aboveground (−27.1h)
compared with belowground (−24.0h) sources. Comparisons of
the relative contribution of different C sources to accrued soil
organic C between elevated and ambient [CO2] treatments are not
possible, however, as the δ13 C of CO2for ambient CO2treatments
remained constant throughout the experiment.
Flux estimates of net ecosystem productivity (NEP) are problem-
atic in aridlands13, thus our harvest provides the first direct measure
of long-term enhancements to NEP stimulated by elevated [CO2].
Estimates of the spatial extent of aridlands range from 2.65×109ha
(ref. 10) to 4.89×109ha (ref. 9) and plant cover in arid biomes has
increased 11% as atmospheric CO2increased from 1982 to 2010
(ref. 14). Assuming that responses observed over this ten-year study
are representative of other arid ecosystems, then enhancements to
NEP in arid and semiarid lands caused by elevated CO2could
range from 0.37 to 0.68 Pg C yr−1. This enhancement of NEP is
equivalent to 4–8% of current global CO2emissions of 8.4 Pg C yr−1
and 15–28% of current terrestrial uptake estimates of 2.4 PgC yr−1
(ref. 1). The recent generation of representative concentration path-
ways (RCPs) for climate simulations predict that atmospheric [CO2]
will reach levels used in this experiment (513 µmol mol−1)between
2045 (RCP8.5) and 2063 (RCP4.5; ref. 15) and CO2enhance-
ment of NEP reported here could account for 4–8% and 2–4% of
predicted total emissions for RCP4.5 (9.0Pg C yr−1)and RCP8.5
(19.0 Pg C yr−1), respectively, at that time. Although extrapolations
such as this can be problematic, as evidenced by the range in
atmospheric [CO2] trajectories proposed in the RCPs and possible
interactions between elevated [CO2] and other global change factors
beyond the goals of this experiment, such as increased atmospheric
N deposition, changes in precipitation regimes and warming, they
do point out the potential for CO2stimulation of NEP in arid regions
to impact global [CO2].
Increases in total ecosystem C at the NDFF under elevated [CO2]
are the direct result of CO2fertilization effects on photosynthesis.
Plants grown under elevated [CO2] had photosynthetic rates 1.3–2.0
times greater than those grown under ambient [CO2] (ref. 5).
Further, integrated leaf-level C gain for the dominant shrub Larrea
tridentata was 170 g C m−2yr−1and 118 g C m−2yr−1greater in
wet and dry years under elevated [CO2], respectively5. Increased
1School of Biological Sciences and WSU Stable Isotope Core Facility, Washington State University, Pullman, Washington 99164, USA, 2Natural Resource
Ecology Laboratory and Department of Biology, Colorado State University, Fort Collins, Colorado 80523, USA, 3Department of Mathematics and Statistics,
Northern Arizona University, Flagsta, Arizona 86011, USA, 4School of Life Sciences, University of Nevada, Las Vegas, Nevada 89154, USA, 5College of
Natural Resources, University of Idaho, Moscow, Idaho 83844, USA, 6Division of Earth and Ecosystem Sciences, Desert Research Institute, Las Vegas,
Nevada 89119, USA, 7USDA-Agricultural Research Service, Lincoln, Nebraska 68583, USA, 8School of Life Sciences, Arizona State University, Tempe,
Arizona 85287, USA, 9Department of Natural Resources and Environmental Science, University of Nevada, Reno, Nevada 89557, USA.
*e-mail: rdevans@wsu.edu
NATURE CLIMATE CHANGE | ADVANCE ONLINE PUBLICATION | www.nature.com/natureclimatechange 1

LETTERS NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2184
Amb. Elev. Amb. Elev.
Carbon (g C m−2)
0
50
100
150
200
250
600
800
1,000
1,200
1,400
Nitrogen (g N m−2)
0
5
10
15
20
25
100
120
140
160
180
ab
Figure 1 | Ecosystem C and N under ambient and elevated CO2.
a,b, Posterior means and 90% credible intervals (error bars) for
aboveground (filled uptriangle, open uptriangle) and belowground (filled
downtriangle, open downtriangle) plant biomass, soils (filled square, open
square) and total (filled circle, open circle) ecosystem C (a) and N (b)
under ambient (Amb.) and elevated [CO2] (Elev.). Estimates are derived by
summing cover-weighted values for individual cover types (Supplementary
Information). Mean soil (Bayesian p-value = 0.002, 0.002) and total
ecosystem (p=0.021,0.004) were significantly dierent between CO2
treatments for C and N, respectively.
photosynthesis and leaf-level C gain in all years, however, translated
to increases in aboveground NPP in only wet, but not dry, years11,13.
The absence of [CO2] treatment differences in plant C and N
pools at final harvest seems to contradict the measured increases
in photosynthesis and NPP under elevated [CO2]. This observation,
in fact, highlights a primary mechanism for the observed increase
in soil organic C, as well as a fundamental difference in the
response between arid and more mesic ecosystems. Arid ecosystems
are characterized by rapid increases in NPP and biomass in
response to stochastic increases in water availability11. The greatest
enhancement in NPP and growth of plants under elevated [CO2]
at the NDFF occurred when moisture was most available. The
final harvest at NDFF occurred during a dry year, indicating
that peaks in production that occur under elevated [CO2] when
moisture is readily available cannot be sustained during intervening
drought. Hence, additional biomass senesced, increasing C inputs
into soil as litter. High rates of above- and belowground plant
biomass turn-over are common in arid ecosystems; turnover of
aboveground biomass at NDFF may occur every two to six years
based on measurements of aboveground NPP of 10–30 g C m−2s−1
at a nearby site13 and total aboveground biomass can turn over
every 1.5 yr based on NPP and standing biomass estimates in the
Chihuahuan Desert16.
The 118–170 g C m−2yr−1increase in leaf-level C gain observed
here under elevated [CO2] without consistent, concurrent increases
in aboveground NPP suggests a second mechanism for the observed
increases in soil organic C; significant increases in belowground
allocation of C. Belowground biological activity beneath shrubs, as
estimated by soil respiration, can be 60% greater under elevated
compared with ambient [CO2], but this increase is occurring
without significant differences in fine-root standing crop, turnover
rates6, or root respiration17 . Thus, this increase in belowground
biological activity is probably due to increases in soil microbial
activity or population size, and that increased rhizodeposition and
Amb. Elev. Amb. Elev.
−30
−28
−26
−24
−22
−20
−18
0
2
4
6
8
a
b
δ13C (% )
δ15N (% )
Figure 2 | Ecosystem C and N isotope composition under ambient and
elevated CO2.a,b, Posterior means and 90% credible intervals (error bars)
for aboveground (filled uptriangle, open uptriangle) and belowground
(filled downtriangle, open downtriangle) plant biomass, soils (filled square,
open square) and total (filled circle, open circle) ecosystem δ13C (a) and
δ15N (b) under ambient (Amb.) and elevated [CO2] (Elev.). Estimates are
derived by summing cover-weighted values for individual cover types
(Supplementary Information). Mean δ13C was significantly dierent
between treatments for above- and belowground biomass (Bayesian
p-value <0.0001) and total carbon (p=0.066). No significant dierences
were observed between treatments for δ15N.
subsequent assimilation and stabilization by the soil microbial
community is a significant mechanism for increased C inputs to
the soil under elevated [CO2]. Rates of rhizosphere C deposition
have been shown to increase 56–74% under elevated [CO2] in
diverse ecosystems18,19 and labile compounds immobilized into
microbial residues can be a major source of stable C and N in
soils20. A previous study8demonstrated that plant photosynthates
are assimilated by rhizosphere microbial communities within
1 h of exposing L. tridentata to 13 C-labelled CO2, highlighting
the tight linkage between plants, rhizodeposition and microbes
in this arid ecosystem as well as the ability of C to rapidly
transfer from the site of photosynthesis to the soil without
increases in plant biomass or turnover rates. The C:N ratio of
accumulated C and N observed here (5.5, total soil C:N =7.3)
is consistent with accumulation of bacterial and fungal residues
(C:N from 4:1 to 10:1), an observation supported by increased
amounts of fungal and bacterial biomarkers under elevated [CO2]
(A.K. and R.D.E., manuscript in preparation). Thus, root exudation
and microbial stabilization may be more important determinants of
belowground C balance in arid ecosystems than the input of fine-
root litter found in more mesic ecosystems12. The patterns observed
at the NDFF are congruent with observations from the semiarid
shortgrass steppe21, where elevated [CO2] stimulated aboveground
production of only 33% over five years but caused a doubling of
rhizodeposition over the same time period.
Mean organic N was 161 g N m−2(credible interval of
145–178 g N m−2) under elevated [CO2] in contrast to 136 g Nm−2
(credible interval of 122–150 g N m−2)for controls (Fig. 1). The
N cycle in arid ecosystems is open with relatively high rates of N
input that are balanced by similar rates of loss from soil emissions,
thus small changes in either inputs or losses can significantly alter
ecosystem N storage. The observed differences in N can therefore
result from increased N2fixation or atmospheric deposition,
2NATURE CLIMATE CHANGE | ADVANCE ONLINE PUBLICATION | www.nature.com/natureclimatechange

NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2184 LETTERS
greater retention of N through decreased gas emissions from
volatilization, nitrification and denitrification, or transfer of N
from below our sampling zone into the top 1 m of soil. Rates of
atmospheric deposition in this region are 0.5–1.0g N m−2yr−1,
whereas another study22 recently determined the mean rate of N2
fixation in aridlands is 1 g N m−2yr−1, strongly suggesting that the
differences in accumulation rates observed here are at the lower
region of the credible interval. Increased rates of heterotrophic
N2fixation were observed under elevated [CO2] (ref. 23), but
rates are not great enough to solely account for the observed
treatment differences. Changes in rooting dynamics and plant N
acquisition below the 1 m depth examined in this experiment may
have also contributed to the observed differences. Nitrate readily
leaches in coarse soils and the greatest concentrations are often
observed at depths of 1 m or greater24 . Nutrient acquisition by
dominant shrubs can occur to 5 m (ref. 25) and seasonal patterns
and observed treatment differences in leaf δ15N of L. tridentata7
are consistent with patterns observed with plant use of nitrate at
depth26; effectively transferring N from depth to the top 1 m of soil.
Finally, elevated [CO2] may increase total ecosystem N over time
by increasing rates of N retention by plants and microbes, thereby
decreasing rates of gaseous loss27. This hypothesis is supported by
experimentation; volatilization is the primary source of N loss at
the NDFF (refs 4,28) and experimental addition of C (ref. 27) or
elevated [CO2] (ref. 28) greatly decreased gaseous N emissions,
thus facilitating retention of N in the soil. Reliable estimates of
annual N fluxes in arid ecosystems are problematic due to their
episodic occurrence. The observed differences in ecosystem N
content observed here are probably due to a combination of each of
the above factors, and separating their relative roles requires further
experimentation beyond the goals of this study.
The progressive N-limitation hypothesis predicts increased
N limitations to NPP as ecosystems accumulate C, but this has
not yet been observed at the NDFF. [CO2] enhancement and
direct C-addition studies demonstrate that microbial activity at
the NDFF is limited by available C (refs 23,27) and increased C
inputs accelerate rates of soil N transformations, thus increasing N
mineralization and inorganic N availability7,29. Soil organic matter
in arid ecosystems is largely recalcitrant7, but increased litter and
rhizodeposition under elevated [CO2] have caused an increase in
microbial biomass and diversity, especially for fungi8that are more
efficient at using recalcitrant substrates. This is accompanied by
an increase in the diversity of substrates used by the microbial
community as well as the activities of enzymes involved in N and
C cycling29.
Assessing the location and magnitude of terrestrial C sinks
is challenging because of their spatial and temporal complexity.
Previous efforts to estimate C uptake by the terrestrial surface
often focused on easily identified sinks such as forest regrowth and
typically did not consider non-forested ecosystems or physiological
enhancements to photosynthesis and growth caused by increasing
[CO2]. Despite suggestions that the strength of global C sinks has
recently declined or remained static, recent mass balance analyses
of global C indicate that uptake of CO2by oceans and the land
surface has accelerated over the past 50 yr (ref. 2), highlighting the
uncertainties present in our knowledge of the global C cycle. Results
from this ten-year experiment clearly demonstrate two critical
areas that must be considered to develop a more comprehensive
understanding of the fates of atmospheric CO2. First, non-forested
ecosystems must be accounted for in studies of terrestrial sinks;
arid and semiarid lands are the most widespread terrestrial biomes
and the enhancements in NEP in response to elevated [CO2]
observed here indicate their importance as a significant C sink.
Second, increases in C storage observed here were the result of
[CO2] enhancements to photosynthesis, subsequent increases in
plant biomass during wetyears followed by greater senescence in dry
years and increased rhizodeposition. Thus, more mechanistic detail
is necessary in models predicting plant, rhizodeposition and NEP
responses to elevated [CO2] (ref. 30). Consideration of both factors
will in turn allow us to better constrain terrestrial C dynamics and
ultimately the global C cycle.
Methods
Free-air-CO2-enrichment (FACE) experiments allow investigators to quantify
whole-ecosystem responses to elevated [CO2] in coupled plant–soil systems. The
NDFF was the only FACE experiment located in an intact arid ecosystem. The
NDFF was located 15 km north of Mercury, (36°490N, 115°550W; elevation
965–970 m) in the northern Mojave Desert. The site consisted of nine
23-m-diameter experimental plots exposed to three fumigation treatments. Three
plots were fumigated at ambient atmospheric [CO2] (~380 µmol CO2mol−1)as a
blower control (ambient), three at ~550 µmolC O2mol−1(elevated) and three
received no fumigation (non-blower control). Treatments began in April 1997
and continued until June 2007. Fumigations were maintained continuously
throughout the experiment except when air temperatures were below 4 °C or
when wind speeds >7 m s−1for more than 5 min. Mean [CO2] concentrations
were 513 µmol mol−1and 375 µmol mol−1for elevated and ambient treatments,
respectively, over the life of the experiment. The δ13C of supplemental CO2was
−5.4huntil 10 February 2003 when the source CO2was switched to −32.0h
for the remainder of the experiment. Dilution with ambient air resulted in δ13C of
CO2in the elevated treatment of −7.3hand −18.2hbefore and after the source
switch, respectively. The δ13 C of ambient and control treatments was −8h
throughout the experiment.
Seven cover types based on the dominant species were identified in each plot.
The final harvested area for each ring was calculated from aerial photographs
using image-processing software (ENVI, Exelis Visual Information Solutions).
Vegetation and soils to 1 m depth were destructively harvested from two-thirds of
each plot at the end of the 2007 growing season. Aboveground biomass was
determined by cutting all plants at ground level and summing biomass for all
individuals of each species. Belowground biomass was measured for each cover
type using two approaches. First, root biomass was determined from excavated
soil collected in association with specific cover types. Second, roots were collected
from transects through the plot. Fine-root data were obtained from
minirhizotron tubes6. Soils were collected under the canopies of the five most
abundant plant-cover types and in plant interspaces to 1 m in depth at 0.2m
increments. Soils were collected from two microsites, centre and edge of
aboveground vegetation canopies, under all cover types except Pleuraphis rigida
(a C4bunchgrass). Rock and soil volumes and soil bulk densities were measured
by excavation. Two square pits (0.5 ×0.5 m projected area) in each plot were
excavated to 1 m in 0.2 m increments. Samples were passed through 2 mm mesh
screens to separate rocks (>2 mm) from soils (≤2 mm). Rock volume was
quantified by measuring the amount of water displaced in a 20l plastic container.
Bulk densities were used to scale soil content measurements to an aerial basis.
The mean rock and soil volumes for all plots by depth were used to correct for
rock content. Plant C and N contents and stable isotope compositions were
determined at the Washington State University Stable Isotope Core Facility
(Pullman) using an ECS 4010 elemental analyser (Costech Analytical, Valencia)
interfaced with an isotope ratio mass spectrometer (Delta PlusXP, Thermo
Finnigan). Soil samples for organic C content and stable isotope composition
were treated with 3NH3PO4to remove carbonates before analysis.
Final harvest data were analysed at the level of the cover type to address
merging of data from the soil, aboveground and belowground samples from each
sample location (see Supplementary Information for complete description of the
statistical analysis methods). All C and N content data were log transformed for
analysis, whereas δ13C and δ15 N data were not transformed. Data were analysed
through a linear mixed-effects model with cover type and [CO2] treatment as
fixed effects, and ring within treatment as a random effect. Separate models were
fitted to each response variable to obtain pool and isotope estimates for each
ring-cover type combination. Change in the relative proportion of the cover types
was not detected during the experiment, so landscape-level estimates were
calculated by summing values for soil, root and aboveground estimates for each
cover type, yielding cover-type pool totals and computing weighted averages of
these cover type totals to obtain the plot-level totals. All statistical models were
simultaneously implemented in a Bayesian framework that allowed us to
propagate uncertainty in the cover type ×pool type ×ring estimates, yielding
accurate estimates (posterior means and credible intervals) of the ring, cover type
by ring and treatment-level total pool estimates. The fixed effect coefficients and
standard deviation terms were assigned relative standard priors and a
semi-informative prior was used for the ring random effect standard deviations
due to the small number of rings per treatment.
Received 24 June 2013; accepted 5 March 2014;
published online 6 April 2014
NATURE CLIMATE CHANGE | ADVANCE ONLINE PUBLICATION | www.nature.com/natureclimatechange 3

LETTERS NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2184
References
1. Friedlingstein, P. et al. Update on CO2emissions. Nature Geosci. 3,
811–812 (2010).
2. Ballantyne, A. P., Alden, C. B., Miller, J. B., Tans, P. P. & White, J. W. C. Increase
in observed net carbon dioxide uptake by land and oceans during the past 50
years. Nature 488, 70–72 (2012).
3. Houghton, R. A., Hall, F. & Goetz, S. J. Importance of biomass in the global
carbon cycle. J. Geophys. Res. 114, G00E03 (2009).
4. Billings, S. A., Schaeffer, S. M. & Evans, R. D. Trace N gas losses and N
mineralization in Mojave Desert soils exposed to elevated CO2.Soil Biol.
Biochem. 34, 1777–1784 (2002).
5. Housman, D. C. et al. Increases in desert shrub productivity under elevated
carbon dioxide vary with water availability. Ecosystems 9, 374–385 (2006).
6. Ferguson, S. D. & Nowak, R. S. Transitory effects of elevated atmospheric CO2
on fine root dynamics in an arid ecosystem do not increase long-term soil
carbon input from fine root litter. New Phytol. 190, 953–967 (2011).
7. Billings, S. A., Schaeffer, S. M. & Evans, R. D. Soil microbial activity and N
availability with elevated CO2in Mojave Desert soils. Glob. Biogeochem. Cycles
18, GB1011 (2004).
8. Jin, V. L. & Evans, R. D. Microbial 13C utilization patterns via stable isotope
probing of phospholipid biomarkers in Mojave Desert soils exposed to ambient
and elevated atmospheric CO2.Glob. Change Biol. 16, 2334–2344 (2010).
9. Lal, R. Carbon sequestration in dryland ecosystems. Environ. Manage. 33,
528–544 (2004).
10. Jobbagy, E. G. & Jackson, R. B. The vertical distribution of soil organic carbon
and its relation to climate and vegetation. Ecol. Appl. 10, 423–436 (2000).
11. Smith, S. D., Monson, R. K. & Anderson, J. E. Physiological Ecology of North
American Desert Plants (Springer, 1997).
12. Jastrow, J. D. et al. Elevated atmospheric carbon dioxide increases soil carbon.
Glob. Change Biol. 11, 2057–2064 (2005).
13. Schlesinger, W. H., Belnap, J. & Marion, G. On carbon sequestration in desert
ecosystems. Glob. Change Biol. 15, 1488–1490 (2009).
14. Donohue, R. J., Roderick, M. L., McVicar, T. R. & Farquhar, G. D. CO2
fertilisation has increased maximum foliage cover across the globe’s warm, arid
environments. Geophys. Res. Lett. 40, 3031–3035 (2013).
15. Van Vuuren, D. P. et al. The representative concentration pathways: An
overview. Climatic Change 109, 5–31 (2011).
16. Huenneke, L. & Schlesinger, W. H. in Structure and Function of a Chihuahuan
Desert Ecosystem. The Jornada Basin Long-Term Ecological Research Site
(eds Havstad, K. M., Huenneke, L. & Schlesinger, William H.)
232–246 (Oxford Univ. Press, 2006).
17. Clark, N. M., Apple, M. E. & Nowak, R. S. The effects of elevated CO2on root
respiration rates of two Mojave Desert shrubs. Glob. Change Biol. 16,
1566–1575 (2010).
18. Cheng, W. X. & Johnson, D. W. Elevated CO2, rhizosphere processes, and soil
organic matter decomposition. Plant Soil 202, 167–174 (1998).
19. Hungate, B. A., Chapin, F. S., Zhong, H., Holland, E. A. & Field, C. B.
Stimulation of grassland nitrogen cycling under carbon dioxide enrichment.
Oecologia 109, 149–153 (1997).
20. Cotrufo, M. F., Wallenstein, M. D., Boot, C. M., Denef, K. & Paul, E. The
microbial efficiency-matrix stabilization (MEMS) framework integrates plant
litter decomposition with soil organic matter stabilization: Do labile plant
inputs form stable soil organic matter? Glob. Change Biol. 19,
988–995 (2013).
21. Pendall, E., Mosier, A. R. & Morgan, J. A. Rhizodeposition stimulated by
elevated CO2in a semiarid grassland. New Phytol. 162, 447–458 (2004).
22. Elbert, W. et al. Contribution of cryptogamic covers to the global cycles of
carbon and nitrogen. Nature Geosci. 5, 459–462 (2012).
23. Schaeffer, S. M., Billings, S. A. & Evans, R. D. Laboratory incubations reveal
potential responses of soil nitrogen cycling to changes in soil C and N
availability in Mojave Desert soils exposed to elevated atmospheric CO2.Glob.
Change Biol. 13, 854–865 (2007).
24. Hunter, R. B., Romney, E. M. & Wallace, A. Nitrate distribution in Mojave
Desert soils. Soil Science 134, 22–30 (1982).
25. Hartle, R. T., Fernandez, G. C. J. & Nowak, R. S. Horizontal and vertical zones
of influence for root systems of four Mojave Desert shrubs. J. Arid Environ. 64,
586–603 (2006).
26. Evans, R.D. & Ehleringer, J. Water and nitrogen dynamics in an arid woodland.
Oecologia 99, 233–242 (1994).
27. Schaeffer, S. M., Billings, S. A. & Evans, R. D. Responses of soil nitrogen
dynamics in a Mojave Desert ecosystem to manipulations in soil carbon and
nitrogen availability. Oecologia 134, 547–553 (2003).
28. McCalley, C. K. & Sparks, J. P. Controls over nitric oxide and ammonia
emissions from Mojave Desert soils. Oecologia 156, 871–881 (2008).
29. Jin, V. L., Schaeffer, S. M., Ziegler, S. E. & Evans, R. D. Soil water availability
and microsite mediate fungal and bacterial phospholipid fatty acid biomarker
abundances in Mojave Desert soils exposed to elevated atmospheric CO2.
J. Geophys. Res. 116, G02001 (2011).
30. Smith, N. G. & Dukes, J. S. Plant respiration and photosynthesis in global-scale
models: Incorporating acclimation to temperature and CO2.Glob. Change Biol.
19, 45–63 (2013).
Acknowledgements
The authors gratefully acknowledge grant support from the Department of Energy’s
Terrestrial Carbon Processes Program (DE-FG02-03ER63650, DEFG02-03ER63651) and
the NSF Ecosystem Studies Program (DEB-98-14358 and 02-12819). We also
acknowledge the DOE’s National Nuclear Security Administration for providing utility
services and undisturbed land at the Nevada National Security Site (formerly Nevada
Test Site) to conduct the FACE experiment. We also thank S. Chung at WSU for advice on
current emission inventories.
Author contributions
S.S. and R.N. conceived the study. R.E., S.S., R.N., T.C., B.N. and L.F. designed the final
harvest. R.E. and A.K. collected soils data, S.S., T.C. and B.N. collected aboveground
plant data and R.N. collected belowground plant data. L.F. and T.C. determined plot area
and species composition. B.H. provided elemental and isotopic analyses, and D.S. and
K.O. analysed the data. All authors wrote the paper.
Additional information
Supplementary information is available in the online version of the paper. Reprints and
permissions information is available online at www.nature.com/reprints.
Correspondence and requests for materials should be addressed to R.D.E.
Competing financial interests
The authors declare no competing financial interests.
4NATURE CLIMATE CHANGE | ADVANCE ONLINE PUBLICATION | www.nature.com/natureclimatechange