Longterm response of a Mojave Desert winter annual plant community

Longterm response of a Mojave Desert winter annual plant community

Global Change Biology (2014) 20, 879–892, doi: 10.1111/gcb.12411 Long-term response of a Mojave Desert winter annual plant community to a whole-ecosy...

481KB Sizes 0 Downloads 0 Views

Recommend Documents

mojave desert - Adventure To Fitness
On today's episode, we went to one of the hottest places in the world … the Mojave Desert in ... different rock layers

Conserving Desert Tortoise Population Linkages - Mojave Desert
Apr 30, 2013 - rail lines, and new airport construction. Low-mobility ... Ivanpah. 10. Ivanpah. Death Valley National Pa

Mojave Desert Ecosystem Program - State of California
Kathryn Thomas, U.S. Geological Survey (USGS), Southwest Biological Science. Center, Colorado Plateau Field Station, Fla

Miocene Arthropods from the Mojave Desert California
... Government Printing Office. Washington 25, D. C. - Price $1 (paper cover) .... the fossil fauna by a subspecies, D.

Longterm social structure of a resident community of Atlantic spotted
Little Bahama Bank (LBB) is 64 km from the east coast of Florida, and north of. West End, Grand Bahama Island (Fig. 1).

Mojave Desert Discovery - National Park Service
My Favorite Plant. 5. Be a Botanist. 5 ..... such as the gila monster, ocotillo, palo verde, and smoke tree. ... known t

Reaction of a California Annual-plant Community to Fire
T HE occurrence of fires on annual- plant ranges of California poses a question to the range manager: How will the plant

Mojave Desert Land Trust Responds to the Confirmation of David
Jul 24, 2017 - Bernhardt's career as a lobbyist, and work at Brownstein Hyatt Farber Schreck for Cadiz Inc, a company th

Changes in a California Annual Plant Community Induced By
Changes in a California Annual Plant Community Induced By Manipulation of Natural Mulch. Author(s): Harold F. Heady. Sou

desert community college district - College of the Desert
Nov 18, 2002 - Marcos Cicerone. Employment Development .... RODRIGUEZ, Darlene. Student Wprler. $6.75 per ... CERANIC, E

Global Change Biology (2014) 20, 879–892, doi: 10.1111/gcb.12411

Long-term response of a Mojave Desert winter annual plant community to a whole-ecosystem atmospheric CO2 manipulation (FACE) STANLEY D. SMITH*, THERESE N. CHARLET*, STEPHEN F. ZITZER†, SCOTT R. ABELLA‡, C H E R Y L H . V A N I E R * and T R A V I S E . H U X M A N § ¶ *School of Life Sciences, University of Nevada, 4505 S. Maryland Parkway, Las Vegas, NV 89154-4022, USA, †Division of Ecosystem and Earth Sciences, Desert Research Institute, 755 E. Flamingo Road, Las Vegas, NV 89119, USA, ‡Biological Resource Management Division, National Park Service, Washington Office, Natural Resource Stewardship and Science Directorate, 1201 Oakridge Drive, Ft. Collins, CO 80525, USA, §Ecology & Evolutionary Biology, University of California, Irvine, CA 92697-2525, USA, ¶Center for Environmental Biology, University of California, Irvine, CA 92697-1450, USA

Abstract Desert annuals are a critically important component of desert communities and may be particularly responsive to increasing atmospheric (CO2) because of their high potential growth rates and flexible phenology. During the 10-year life of the Nevada Desert FACE (free-air CO2 enrichment) Facility, we evaluated the productivity, reproductive allocation, and community structure of annuals in response to long-term elevated (CO2) exposure. The dominant forb and grass species exhibited accelerated phenology, increased size, and higher reproduction at elevated (CO2) in a wet El Ni~ no year near the beginning of the experiment. However, a multiyear dry cycle resulted in no increases in productivity or reproductive allocation for the remainder of the experiment. At the community level, early indications of increased dominance of the invasive Bromus rubens at elevated (CO2) gave way to an absence of Bromus in the community during a drought cycle, with a resurgence late in the experiment in response to higher rainfall and a corresponding high density of Bromus in a final soil seed bank analysis, particularly at elevated (CO2). This long-term experiment resulted in two primary conclusions: (i) elevated (CO2) does not increase productivity of annuals in most years; and (ii) relative stimulation of invasive grasses will likely depend on future precipitation, with a wetter climate favoring invasive grasses but currently predicted greater aridity favoring native dicots. Keywords: Bromus, desert annuals, elevated CO2, free-air CO2 enrichment, invasive species, Lepidium, Mojave Desert, primary productivity, seed bank Received 1 July 2013 and accepted 15 August 2013

Introduction Perhaps more than any other biome type, deserts have been proposed to exhibit increased annual net primary production (ANPP) in response to rising atmospheric CO2 [(CO2)] (Melillo et al., 1993; Smith et al., 2009). This prediction comes primarily from the fact that primary production in deserts is a strong function of precipitation (Smith et al., 1997), and higher photosynthesis (Long, 1991; Naumburg et al., 2003) in concert with lower stomatal conductance (Nowak et al., 2001) at elevated (CO2) results in substantially higher plant wateruse efficiency. Alternatively, severe water and nutrient limitations in the Mojave Desert may limit this ANPP response to only high-resource (wet) years (Housman et al., 2006). Functional plant types in deserts contribute differently to ANPP; woody shrubs contribute a greater Correspondence: Stanley D. Smith, tel. + 702 895 3197, fax + 702 774 4733, e-mail: [email protected]

© 2013 John Wiley & Sons Ltd

proportional amount during dry years, but during wet years the contribution of annual plants approaches that of perennials (Turner & Randall, 1989). Thus, yearto-year variation in ANPP in response to rainfall is strongly driven by the relative contribution of annuals. Winter annuals are the primary contributor to plant diversity in the Mojave Desert (Mulroy & Rundel, 1977). The flora has been invaded by the prolific exotic annual grass, Bromus rubens (cf., Bromus madritensis ssp. rubens; red brome), with many locations having high relative Bromus density (Beatley, 1966). Just as B. tectorum (cheatgrass) has strongly influenced Great Basin communities (e.g., Young & Evans, 1978), B. rubens has had an increasingly important role in the Mojave Desert (Smith et al., 1997). Bromus spp. compete with native species for soil moisture and nutrients and substantially alter wildfire frequency and intensity (Brooks & Matchett, 2006). Native annuals and invasive grasses have similar phenology – they germinate in response to cool season precipitation, develop relatively high growth 879

880 S . D . S M I T H et al. rates when daytime temperatures rise in the spring, and rapidly advance to reproductive maturity in late spring (Beatley, 1974). Elevated (CO2) alters species interactions and the resultant structure of annual plant communities (Belote et al., 2003; Reich, 2009). Desert annual communities exhibit complex structure in which temporal and spatial variation in the environment result in the coexistence of several functional types (Huxman et al., 2013), with pronounced year-to-year population variation driven largely by variable precipitation and differential species functional responses (Angert et al., 2007; Huxman et al., 2008). These dynamics are intricately tied to persistent seed banks characterized by species-specific germination cues (Pake & Venable, 1996). Strong evidence shows desert annuals already responding to climate change in the region, with a shift toward more stress-tolerant species that germinate under colder conditions (Kimball et al., 2009). Many plants, including desert annuals (Huxman & Smith, 2001), should increase biomass and seed production as a result of increased photosynthesis associated with exposure to elevated (CO2). Mojave Desert annuals are an important functional group to understand with respect to increasing (CO2), as: (i) their relative contribution to ANPP and community seed rain is an important driver for desert ecosystem structure and function; (ii) their growing season is a function of temperature and precipitation, which is changing in the southwestern desert region (Seager et al., 2007); (iii) their annual life-history strategy, in which their persistence is intricately tied to their ability to make viable seeds, may be responsive to elevated (CO2); and (iv) the interaction between invasive and native species, and exotic grass-driven changes in disturbance regimes, may be particularly important to desert ecosystem function in the future (D’Antonio & Vitousek, 1992). Here, we present results from the Nevada Desert Free-air CO2 Enrichment Facility (NDFF) that exposed an intact ecosystem to 1.5 times ambient (CO2) (to a 550 lmol mol 1 set point) for 10 growing seasons. As part of this ecosystem experiment, we evaluated annual plant community biomass production, along with allocation patterns and seed production in several dominant species. We tested the following hypotheses: (i) annual plants exposed to elevated (CO2) will increase in biomass and community-level ANPP; (ii) the increase in biomass at elevated (CO2) will be greater for invasive species as compared to native annuals (Smith et al., 1987; Ziska, 2003); (iii) increases in plant biomass at elevated (CO2) will be greater in microsites beneath perennial plants, where soil N contents are substantially higher (Schlesinger et al.,

1996); (iv) increases in biomass will be accompanied by proportional increases in reproductive allocation and seed production; and (v) higher plant production and seed rain of the invasive Bromus (relative to native annuals) will shift community species composition and seed bank toward dominance by the invasive grass. With a full decade of community composition and biomass data, plus a final analysis of the soil seed bank, this study included the inherent temporal and spatial variation in deserts as an essential component of the design.

Materials and methods

Site description The Nevada Desert Free-air CO2 Enrichment Facility (NDFF) was located on the US Department of Energy Nevada National Security Site (36°49′N, 115°55′W, 960–970 m elev.) in southern Nevada. The climate is arid, with precipitation falling mostly as rain during the cool season (October–February), although significant rainfall via unpredictable thunderstorms can occur during the summer (Rundel & Gibson, 1996). Precipitation was continuously measured with a centrally located HOBO RG3 data logging, tipping bucket rain gauge (Onset Computer Corp., Bourne, MA, USA) with 0.25 mm resolution. The community is Mojave Desert scrub (<20% perennial cover) dominated by the evergreen shrub Larrea tridentata, the deciduous shrubs Ambrosia dumosa, Lycium andersonii, and L. pallidum, and the C4 bunchgrass Pleuraphis rigida. More than 60 species of native winter annuals and 6 species of summer annuals inhabit the area, depending on seasonal rainfall (Bowers, 1987). Due to a lack of historical disturbance, biological soil crusts (cyanobacteria and lichens) cover 30–60% of the ground surface. The NDFF consisted of nine circular plots, each 25 m in diameter (491 m2), with a 1 m buffer from the plot edge to provide a 23-m-diameter sampling area (415 m2). Three plots were maintained at 550 lmol mol 1 (CO2), three had the Freeair CO2 enrichment (FACE) apparatus but were maintained at ambient (CO2) (380 lmol mol 1), and three were non-FACE controls (Jordan et al., 1999). The NDFF maintained continuous (CO2) enrichment >95% of daylight hours, except when the 5-min wind speed average exceeded 6.0 m s 1 (7.0 m s 1 in the growing season) or when air temperature was below 3 °C. Plants were accessed from an overhead moveable walkway system and attached sampling platform that prevented surface disturbance of the plots.

Community sampling Two transects (8 m long, 0.2 m wide) were established in each of the nine plots in winter 1997–1998 prior to germination of annuals. Transects originated 2 m from the plot centers and radiated toward the plot outer edge and were randomly located within two arcs on different sides of the plot, with © 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 879–892

E L E V A T E D C O 2 E F F E C T S O N D E S E R T A N N U A L S 881 each arc representing the mean percent cover of the primary microsites (evergreen shrub, deciduous shrub, bunchgrass, and open space). Transect locations were permanently marked for repeat sampling. Annual plants were counted – by species – in 32 subplots (0.5 9 0.2 m, 1000 cm2; 16 per transect), and the overstory cover type (as percentage surface area) by perennial species was estimated for each subplot using a clear plexiglass template with gridded lines. Transects were measured at least twice at preflowering and then maximum biomass and seed set, respectively, for the community dominants. In the El Ni~ no year of 1998, they were measured monthly, commencing in February.

Individual plant biomass sampling The four dominant winter annual plant species (greater than 70% of the total annual plant density) were harvested from each plot: (i) Bromus rubens, exotic grass; (ii) Vulpia octoflora, native grass; (iii) Lepidium lasiocarpum, native forb; and (iv) Eriogonum trichopes, native forb. Four microsites were sampled in each plot: (i) open space, greater than 1 m from the edge of a perennial plant; (ii) beneath the canopy of the evergreen shrub Larrea; (iii) beneath the canopy of a drought-deciduous shrub, either Ambrosia or Lycium; and (iv) within the canopy of the bunchgrass Pleuraphis. These four microsites encompassed the range of light, water, and nutrient availabilities in plots. In the 1998 El Ni~ no year, four harvests were conducted from vegetative stage to maximum seed set. In subsequent years, either harvests were conducted near anthesis and then at peak seed set (coinciding with community transects), or in low biomass years a single harvest was conducted at peak biomass and reproductive activity. At each harvest date, within each plot, five plants of each species were harvested for each microsite [n = 15 plants for each species 9 date 9 microsite 9 (CO2) treatment combination]. Plants were selected by randomly choosing a cover type and removing the individual annual plant closest to the center of that cover type for all combinations. Each plant was cut at the soil surface and all aboveground portions were removed. Roots could not be harvested to protect the long-term integrity of NDFF soil surfaces. However, because Mojave Desert annuals average a root : shoot ratio of 0.1 (Bell et al., 1979), aboveground harvests accounted for ca. 90% of total annual plant production. Total leaf area of each plant was determined with a leaf area meter (Delta-T Devices, Cambridge, UK). Harvested plants were separated into vegetative and reproductive structures and placed in an oven at 70 °C to constant dry weight. Dry mass of each plant component was determined separately. Total seed production was estimated from a subsample of seeds and accessory reproductive structures where mass had been determined. The ratio of masses of these structures was used to adjust total reproductive mass into total seed mass and total accessory reproductive structure mass. A subsample of seeds (greater than 10 per individual) chosen at random was weighed and used to estimate seed production from total seed mass.

© 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 879–892

Seed bank analysis The soil seed bank was sampled beneath three mature individuals of Larrea tridentata, Lycium spp., and Pleuraphis rigida as well as interspaces (>1 m from the nearest shrub) within each plot. Samples were collected from two points spaced 0.3 m from each other on the south and north sides of perennial plants, resulting in four samples per plant, or from four equally spaced points within a 1 m2 area in interspaces. A 100 cm3 sample of the 0- to 5-cm-depth mineral soil at each point was obtained with a hand trowel, resulting in 400 cm3 of soil per plant and 1200 cm3 composited on a microsite basis for each of the 9 plots. Samples were collected on 7 May 2007; no germination of annual plants occurred that year, so the sample quantified the seed bank carried over from previous years (Baskin & Baskin, 2001). Three seed bank assays were conducted using the emergence method (Warr et al., 1993). First, beginning the day after sample collection, 360 cm3 of seed bank soil (in a 2-cm-thick layer) from each microsite from each plot was placed on top of sterile potting soil in 15-cm-diameter pots (4 microsites 9 9 plots = 36 pots). Pots were randomly arranged on a bench in a glasshouse, together with seed contamination controls (no contamination was detected). The glasshouse received ambient lighting and was maintained at 28 °C. Samples were misted automatically with 1.5 cm of water day 1. Emerging species were inventoried every 2 weeks for 6 months. Second, the remaining sample was placed in nylon mesh bags and stored outdoors under an opaque shade structure. After 4 months of storage outdoors, we retrieved 240 cm3 of seed bank soil from each sample. We placed this soil in a 2-cm-thick layer on top of potting soil in 700 cm3 square plastic pots. The samples were placed in an incubator (Percival Scientific, Inc., Perry, IA, USA) with 10 h of light (400 lmol m 2 s 1 PPFD) at 7 °C for 6 h and 15 °C for 4 h. The 14-h night consisted of 4 h at 7 °C and 10 h at 0 °C. These conditions simulated winter day/night field conditions at the experimental site during germination (Beatley, 1974). Samples were kept moist in the incubator for 4 months. Finally, we retrieved an additional 360 cm3 of seed bank soil from each sample that had been stored outdoors (for 11 months by that time). These samples were evaluated in the glasshouse using the first assay method.

Statistical analysis Production and reproductive effort data were analyzed with ANOVAS to evaluate effects of elevated (CO2), microsite, harvest date, and their interactions, given the partial hierarchical design of the NDFF experiment (SPSS Inc., Chicago, IL, USA). The whole plot effect, (CO2), was tested over plot within (CO2). As appropriate for each dependent variable, microsite or harvest date were within-plot factors, and within-plot factors along with all interactions were tested over their interaction with plot. Plot means [n = 3 for each (CO2) 9 microsite 9 harvest date combination] were used as the raw data. If microsite effects were not significant, then models were simplified to include only (CO2) and harvest date for variables with multiple harvest dates.

882 S . D . S M I T H et al. Reproductive and vegetative allometry was evaluated using nonlinear regression of the natural logs of reproductive and vegetative mass (leaf mass) for all data for each species (Coleman et al., 1994). We eliminated the smallest and largest plants from each harvest, species, cover type, and replicate combination, which resulted in a sample size of 400 individuals per species. For this analysis, eliminating these individuals ensured that comparisons were being made between individuals on similar developmental schemes. A second-order polynomial was fit through the complete data for each species [both (CO2) treatments], then individually [each (CO2) treatment separately] and F-statistics were constructed from the residual sum of squares from each regression line following Potvin et al. (1990). The three seed bank analyses were each used to express seed densities ha 1 on a plot basis by multiplying seeds m 2 within each microsite by the area occupied by microsites. Mean seed density (square root transformed) for the four dominant seed bank species was used in a one-way ANOVA to compare (CO2) treatments. For all variables in this study, there were no significant differences between the two types of controls. Therefore, the blower controls [FACE apparatus at ambient (CO2)] were used as a comparison to the elevated FACE treatment for all significance tests.

mean (128 mm) from a nearby weather station in Mercury, NV (36°37′N, 116°02′W, 1006 m elevation). The wettest months are January and February, with intermediate months in the fall (October–November) and early spring (March–April). Two years (1998 and 2005 growing seasons) showed substantially higher (330 and 240 mm, respectively) rainfall than the 10-year mean, and 2002 was substantially lower (Fig. 1). The exceptionally wet El Ni~ no year of 1997–1998 was followed by four consecutive below-average rainfall years, with relatively average-to-wet years during the last several years of the experiment (although 2007 was extremely dry).

Species composition Fifty-one species of annuals, representing 17 families, were observed in the sampling transects over the course of the 10-year study (Table 1). The wet years of 1998 and 2005 had the most species (33 and 35, respectively), whereas the very dry years of 1999 and 2007 had no species germinate.

Growth and phenology in individual species Results

Precipitation Germination of Mojave Desert annuals is stimulated by both autumn and winter rains, so the hydrologic year is from October 1 to September 30 (Fig. 1). The 10-year mean accumulated annual rainfall was ca. 138 mm, which was slightly higher than the previous 30-year

Frequent harvests during 1998 allowed us to examine growth and phenology for the two dominant annuals, Bromus rubens and Lepidium lasiocarpum (an exotic grass and native dicot, respectively). Date of peak biomass in Bromus differed among (CO2), microsite cover type, and harvest date (F9, 48 = 1.9, P = 0.08), with peak biomass recorded 2 weeks earlier at elevated (CO2) than at ambient (CO2) (Fig. 2). Furthermore, the three-way

Fig. 1 Hydrologic year (1 October to 30 September) accumulated precipitation for the 10 years of the FACE experiment, from 1997– 1998 to 2006–2007; the 1996–1997 year is also shown as CO2 treatment began in April 1997. The mean monthly precipitation over the 11 years is shown in the bold-type black dashed line, and the extreme annual precipitation high (1997–1998) and low (1988–1989) are shown in horizontal dashed lines. © 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 879–892

E L E V A T E D C O 2 E F F E C T S O N D E S E R T A N N U A L S 883 Table 1 Winter–spring annual species sampled in permanent transects at the Nevada Desert FACE Facility. ‘X’ indicates presence in the annual flora in the year indicated (1999 and 2007 had no annuals in any transects) and in the 2007 seed bank analysis Species

Family

Amsinckia tessellata Astragalus tidestromii Bromus rubens Bromus tectorum Camissonia boothii Camissonia brevipes Caulanthus cooperi Calycoseris wrightii Chaenactis fremontii Chaenactis macrantha Chaenactis stevioides Chorizanthe brevicornu Chorizanthe rigida Cryptantha angustifolia Cryptantha circumscissa Cryptantha micrantha Cryptantha nevadensis Cryptantha pterocarya Cryptantha recurvata Cryptantha utahensis Dasyochloa pulchella Delphinium parishii Descurainea pinnata Erodium cicutarium Eriogonum deflexum Eriogonum trichopes Eriophyllum wallacei Eschscholzia californica Eschscholzia glyptosperma Gilia cana Gilia inconspicua Hazardia brickelliodes Ipomopsis polycladon Lepidium lasiocarpum Malacothrix glabrata Mentzelia albicaulis Monoptilon bellidiforme Nama demissum Oenothera primiveris Pectis papposa Pectocarya platycarpa Phacelia crenulata Phacelia fremontii Plantago ovata Rafinesquia neomexicana Salsola iberica Schismus arabicus Schismus barbatus Selinocarpus nevadensis Sonchus oleraceus Vulpia octoflora

Boraginaceae Fabaceae Poaceae Poaceae Onagraceae Onagraceae Brassicaceae Asteraceae Asteraceae Asteraceae Asteraceae Polygonaceae Polygonaceae Boraginaceae Boraginaceae Boraginaceae Boraginaceae Boraginaceae Boraginaceae Boraginaceae Poaceae Ranunculaceae Brassicaceae Geraniaceae Polygonaceae Polygonaceae Asteraceae Papaveraceae Papaveraceae Polemoniaceae Polemoniaceae Asteraceae Polemoniaceae Brassicaceae Asteraceae Loasaceae Asteraceae Hydrophyllaceae Onagraceae Asteraceae Boraginaceae Hydrophyllaceae Hydrophyllaceae Plantaginaceae Asteraceae Chenopodiaceae Poaceae Poaceae Nyctaginaceae Asteraceae Poaceae

1998

2000

2001

2002

2003

2004

2005

2006

Seed Bank

X

X

X X X

X X

X

X X X

X

X X

X X X

X

X

X X X X X

X

X

X X X X X

X

X X

X X

X X X X

X

X X

X X X X X X X

X X X

X

X X X X

X

X X

X

X

X X X X X X X X X

X

X X X X X X X

X

X

X X X X

X

X

X

X

X X

X

X X

X X

X

X X X X X X X X X X X X X X

X X

X

X

X X

X

X X

X X

X

X X X

X X

X X

X X X

X X X

X

X

X

X

X

X

X

X

X

X

X X X X X X X X X X X

X X

X X X

X

X X X X

X X

X X

X X X

X

X

X

X

X

X X

X X

© 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 879–892

X

X

X

X X

884 S . D . S M I T H et al.

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 2 Individual plant biomass (a, b), seed production per unit leaf surface area (c, d), and seed production per plant (e, f) for Bromus rubens in open interspace (a, c, e) and beneath perennial cover (b, d, f) microsites at the Nevada Desert FACE Facility in the 1998 El Ni~ no year. For plant biomass, biomass is partitioned into vegetative (Veg) and reproductive (Repro) structures. Means  1 SE are shown, and significant CO2 treatment differences (P < 0.05) at each harvest are indicated (*). The four harvest numbers are Day 97 (7 March), Day 111 (21 March), Day 125 (5 April), and Day 140 (20 April).

interaction also indicated that Bromus at elevated (CO2) produced greater total mass than at ambient (CO2) in perennial plant microsites vs. open (intercanopy) spaces. Finally, total leaf surface area per plant nearly doubled for Bromus at elevated (CO2) (28 cm2) beneath perennials vs. ambient (CO2) (15 cm2), but Bromus leaf area in interspace microsites was not significantly different between (CO2) treatments. The mass ratio of reproductive and vegetative structures was significantly influenced by microsite (F3, 6 = 32.4, P < 0.01) and an interaction between (CO2) and harvest date (F3, 48 = 3.8, P < 0.02); the interaction indicated accelerated phenology at elevated (CO2) such that Bromus reached peak reproductive biomass earlier

(Fig. 2). Peak seed production also occurred earlier for Bromus at elevated (CO2) in both microsites (Fig. 2; significant (CO2)-by-harvest date interaction; F3, 48 = 8.8, P < 0.01). A significant (CO2) and microsite interaction (F3, 48 = 5.6, P < 0.01) indicated that the greatest enhancement in seed production at elevated (CO2) was for plants growing under the canopy of perennial plants. Bromus growing in the open did not show a significant stimulation in total seed production at elevated (CO2) once phenology had been accounted for in the comparison (Fig. 2). Despite differences in total seed production in Bromus, the efficiency of reproduction, as measured by seed number per total plant leaf surface area (LSA), © 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 879–892

E L E V A T E D C O 2 E F F E C T S O N D E S E R T A N N U A L S 885 was a function of microsite location and time (F3, 6 = 4.8, P < 0.05 and F3, 6 = 18.1, P < 0.01, respectively) but not elevated (CO2). Within each microsite, (CO2) did not affect the total numbers of seeds produced per LSA, just temporally shifted the maximal value similar to seed production (Fig. 2). All of these responses in reproductive variables over time indicate that Bromus experienced an ontogenetic shift related to increased growth rates at elevated (CO2). In Lepidium there was a (CO2)-by-harvest date interaction (F3, 48 = 3.4, P < 0.03) that led to the greatest biomass enhancement due to (CO2) during the later harvest dates (Fig. 3). However, there was no significant interaction of (CO2) and microsite (F3, 48 = 1.8, P = 0.16) even though higher biomass production

occurred beneath perennial canopies. Plants in the open had low leaf biomass and a relatively low enhancement of leaf biomass in elevated (CO2). Increases in total biomass production in Lepidium at elevated (CO2) resulted from proportional increases in both leaf and reproductive biomass across all overstory cover types (no shift in allometry), with no major effects of (CO2), microsite, or higher interactions on the ratio of reproductive mass and vegetative mass. Peak seed production also occurred earlier in elevated (CO2) plants, as seed production peaked by the third harvest while plants in ambient (CO2) exhibited peak seed production at the fourth harvest. There were no effects on number of seeds produced per unit leaf surface area (Fig. 3).

(a)

(b)

(c)

(d)

(e)

(f)

Fig. 3 Individual plant biomass (a, b), seed production per unit leaf surface area (c, d), and seed production per plant (e, f) for Lepidium lasiocarpum in open interspace (a, c, e) and beneath perennial cover (b, d, f) microsites at the Nevada Desert FACE Facility in the 1998 El Ni~ no year. Legends are as in Fig. 2. © 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 879–892

886 S . D . S M I T H et al.

Reproductive allocation

Total production of annuals

Reproductive allocation (RA) in Bromus and Lepidium varied from 30 to 74% of aboveground biomass over the duration of the 10-year experiment (Fig. 4). Most of the variation in RA was temporal (F6, 22 > 15.7, P < 0.01) and not significantly different across (CO2) treatments (F1, 4 < 2.3, P > 0.20), although there was marginally significant variation in (CO2) treatment effects across years (Bromus: F6, 22 = 2.1, P = 0.09; Lepidium: F6, 22 = 2.3, P = 0.07; Fig. 4). (CO2) did not affect RA for either species during the wet years of 1998 and 2005, nor did (CO2) treatment differ in RA of either species during the dry years of 2001 and 2006. Lepidium had higher RA at elevated (CO2) in 2000 (t22 = 1.8, P = 0.09) but lower in 2004 (t22 = 2.7, P = 0.01), whereas Bromus had higher RA in elevated (CO2) in 2003 (t22 = 3.0, P < 0.01).

Peak standing biomass of the annual plant community (Fig. 5) was not on average higher at elevated (CO2) over the duration of the experiment (treatment: F1, 4 = 0.3, P = 0.63) and only marginally within any year (year 9 treatment: F6, 24 = 2.4, P = 0.06), although significant differences occurred among years (year: F6, 24 = 76.2, P < 0.01). There were also some interesting patterns within the two dominant species. In the El Ni~ no year of 1998, Bromus had both higher density at elevated compared to ambient (CO2) (peak density of 34.9  12.7 (mean  SE) and 26.8  9.8 plants m 2, respectively), and larger individual biomass at elevated (CO2) (0.49  0.09 g plant 1) than at ambient (CO2) (0.32  0.08 g plant 1). Lepidium were also larger in elevated (CO2) (0.93  0.20 g plant 1) than ambient (0.64  0.17 g plant 1), but Lepidium trended toward

(a)

(b)

Fig. 4 Reproductive allocation at peak seed set in Bromus rubens (a) and Lepidium lasiocarpum (b) at ambient (open bars) and elevated (solid bars) CO2 from the 1998 to 2007 growing seasons at the Nevada Desert FACE Facility. Means and SE with significant differences (P < 0.05) by CO2 treatment in each year indicated by (*). There were no annual plants in permanent plots in 1999, 2002, and 2007. © 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 879–892

E L E V A T E D C O 2 E F F E C T S O N D E S E R T A N N U A L S 887 (a)

(b)

(c)

Fig. 5 Peak standing biomass of Bromus rubens (a), Lepidium lasiocarpum (b) at ambient (open bars) and elevated (solid bars) CO2 from the 1998 to 2007 growing seasons, and peak standing biomass for the community as a whole (c). Ambient (white bar) and elevated (gray bar) biomass are partitioned each year into native and invasive species for the community (c). Means and 1 SE with significant differences (P < 0.05) by CO2 treatment for each year indicated by (*). No annual plants occurred in the permanent plots in 1999, 2002, and 2007.

having lower plant density at elevated (CO2) (39.5  19.9) vs. ambient (47.9  24.0). This resulted in significantly higher total standing biomass in Bromus at elevated (CO2) (P < 0.05), but comparable standing biomass in elevated and ambient (CO2) in Lepidium. After a dry year with no germination (1999), winter rains resulted in high density in 2000 (Bromus = 32.0  7.9 plants m 2, Lepidium = 6.6  2.6 plants m 2 averaged across both treatments), with no (CO2) effect in either species (t24 < 1.5, P > 0.15). However, because of the absence of spring rains (Fig. 1), plants were extremely © 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 879–892

small in 2000 (Bromus = 0.024  0.003 g plant 1, Lepidium = 0.038  0.006 g plant 1 averaged across both treatments), again with no (CO2) effect (t24 < 0.9, P > 0.38). This resulted in extremely low seed production in 2000 at the community level. Bromus then largely disappeared from the annual flora in 2001 and 2003 (11.4% and 1.1% of total annual biomass, respectively), whereas Lepidium still had modest to dominant density and biomass in 2001 and 2003 (32% and 88% of total annual biomass, respectively; Fig. 5). However, Lepidium exhibited small plant size in

888 S . D . S M I T H et al. those 2 years (0.090  0.020 and 0.073  0.003 g plant 1, respectively). In 2003, Lepidium exhibited higher standing biomass at elevated (CO2) (t24 = 2.6, P = 0.02) because plant density was greater (elevated: 323  159; ambient: 186  92 plants m 2) and plants were larger (elevated: 0.083  0.005; ambient: 0.063  0.005 g plant 1) in elevated (CO2). With the return of average to above-average rainfall in 2005 and 2006, Bromus re-emerged in the flora and increased in standing biomass, but with no (CO2) effect in 2005 (t24 = 0.7, P > 0.50) or 2006 (t24 = 1.2, P > 0.24). Lepidium also did not respond to (CO2) treatment in 2005 (t24 = 1.1, P > 0.25) or 2006 (t24 = 1.5, P > 0.15), a year with low Lepidium biomass (1.1% of all annual vegetation biomass) at both (CO2). Annual community-level biomass (Fig. 5c) was twofold higher at elevated (CO2) in 2000 (t24 = 2.0, P = 0.06) and 2003 (t24 = 2.2, P = 0.04). Interestingly, community-level biomass consisted almost entirely of native species in the intervening dry years (2001–2003), and then invasive grasses (Bromus and Schismus arabicus) exceeded 30% of all annual biomass in 2004 and 2005 and rose to 57% of community biomass in 2006 (Fig. 5).

Seed bank A total of 13 species emerged from seed bank samples (Table 1), with four species – the exotic annual grasses Bromus rubens and Schismus arabicus, the native forb Lepidium lasiocarpum, and the native annual grass Vulpia octoflora – comprising 92% of total germinants. Seed bank density of Lepidium, Vulpia, and Schismus did not differ significantly (P > 0.05) among treatments. Bromus, however, differed significantly (t7, 6 = 4.0, P = 0.01) and exhibited a severalfold increase in emergent seed density at elevated (CO2) compared to ambient (CO2) at the end of the 10-year experiment (Fig. 6).

Discussion

Plant growth and phenology Individual plant biomass increased considerably at elevated (CO2) during a wet El Ni~ no year in the two dominant Mojave Desert winter annual species at the NDFF – the exotic grass Bromus rubens and the native forb Lepidium lasiocarpum. However, the (CO2) response was species specific and influenced by microsite. Plants growing under shrub or perennial grass canopies (fertile islands) were larger and more responsive to elevated (CO2) than those growing in interspaces (Figs 2 and 3). These data are consistent with the hypotheses that plant size would increase at elevated

Fig. 6 Seed bank density of Bromus rubens, Schismus arabicus, Lepidium lasiocarpum, and Vulpia octoflora at ambient (open bars) and elevated (solid bars) CO2 at the end of the experiment (2007). Means and 1 SE; significant CO2 treatment difference (P < 0.05) noted (*) for each species.

(CO2) when sufficient water is present, and that concentrated resources in fertile islands would enhance the (CO2) stimulatory effect on growth. Total biomass enhancement at elevated (CO2) was predicted to be greatest for the invasive Bromus, but enhancement for the native Lepidium was essentially equivalent (2.0- and 1.75-fold increases, respectively). Finally, increases in plant size resulted in greater seed production at elevated (CO2) (Figs 2 and 3). This greater biomass enhancement at elevated (CO2) beneath shrub canopies was not as pronounced in intershrub spaces, locations of decreased resource availability as compared to shrub fertile islands (Schlesinger et al., 1996). Soil nitrogen pools beneath Larrea shrub islands are severalfold higher than in surrounding interspaces (Titus et al., 2002). Nutrient availability can limit growth enhancement at elevated (CO2) in many plant species (Crous et al., 2010; Reich & Hobbie, 2012), and so the relative lack of biomass enhancement by elevated (CO2) in open interspaces may be a consequence of lower soil nutrients. The strong growth response of annuals in perennial microsites could have occurred due to greater wateruse efficiency (WUE) of the overstory plants at elevated (CO2), resulting in potentially greater soil water. Although elevated (CO2) increased WUE of shrubs at this site (Housman et al., 2006), the change was most likely due to photosynthetic enhancements and not whole-plant water loss (Pataki et al., 2000). Thus, soil water at elevated (CO2) was not conserved in this system (Nowak et al., 2004). Furthermore, if the growth response was due to increased soil water, we would expect these winter–spring annuals to have increased life-cycle duration (Gutierrez & Whitford, 1987), as © 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 879–892

E L E V A T E D C O 2 E F F E C T S O N D E S E R T A N N U A L S 889 favorable soil water status would support growth longer into the summer dry season. However, we observed the opposite response, i.e., an acceleration of phenology at elevated (CO2). Thus, we conclude that improved soil water status in fertile islands was not the primary driver for this biomass enhancement.

Reproductive allocation and seed production Seed production responses of these winter annuals to elevated (CO2) were qualitatively similar between exotic and native species. Both Bromus and Lepidium showed a distinct acceleration of phenology, such that peak seed production generally occurred earlier in spring (Figs 2 and 3). This change in phenology may be related to how these plants shift resource allocation to reproductive structures earlier in the season, restricting the potential for maximum final size. Several other studies with annual plants observed greater growth early in the life cycle under elevated (CO2) but reduced stimulation as plants reach asymptotic size (Berntson et al., 1998; Thomas et al., 1999). At elevated (CO2), both Bromus and Lepidium had a decreased ability to grow vegetatively following floral initiation, most likely as a result of determinate meristematic activity. These species reached a size threshold for reproduction that was phenologically accelerated compared to ambient (CO2) because they had accelerated growth rates at earlier vegetative stages. Scaling seed production from individual plants to total landscape seed rain illustrates some of the potential effects of elevated (CO2) on native/nonnative dynamics. Although seed production per plant increased under elevated (CO2) for both species, total seed rain was not significantly increased for Lepidium [30 000 and 22 000 seeds m 2 at elevated and ambient (CO2), respectively]. In contrast, Bromus increased seed production at elevated vs. ambient (CO2) (13 000 and 4300 seeds m 2, respectively) because both seed production per plant and plant density increased. At the community level, seed production of all annuals at elevated (CO2) was generally greater than ambient (CO2) (43 000 and 29 000 seeds m 2, respectively) in the wet El Ni~ no year. Although seed rain was greater, particularly in the invasive Bromus, number of seeds produced per unit LSA was not significantly affected by (CO2) treatment in Bromus (Fig. 2) or Lepidium (Fig. 3) even though nitrogen content per unit LSA decreased (Huxman & Smith, 2001). This reduced leaf N content functionally resulted in a smaller pool size of nutrients available for reallocation to reproduction. In Bromus, although total seeds produced was greater and a prolonged seasonal investment in photosynthesis was observed (Huxman © 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 879–892

& Smith, 2001), there was an actual decrease in reproductive allocation and diminished seed quality at elevated (CO2) (Huxman et al., 1998, 1999). An explanation may be that elevated (CO2) reduced the energetic cost of biomass construction in Bromus relative to native species (Nagel et al., 2004), resulting in faster growth and more seed production earlier in the growth cycle. Despite these striking results that occurred during a wet year early in the experiment, reproductive allocation (RA) was not higher at elevated (CO2) for the remainder of the experiment (Fig. 4). Certainly, most of these years were much drier than the El Ni~ no year, with a pronounced drought cycle in years 2–6 of the experiment, which resulted in much smaller plants and shorter developmental cycles from germination to seed set. In the closely related Bromus tectorum, Dyer et al. (2012) noted that desert populations had highly constrained phenologies, whereas more mesic populations extended the growing season when presented with late-season favorable conditions; B. rubens may indeed have a flexible growth strategy in which it can exploit additional resources [such as elevated (CO2)] in wet years but have a more constrained growth pattern in drier years. Similar to our results, other studies have found little to no increase in RA of annual species at elevated (CO2) in arid/semiarid regions (Navas et al., 1997; Gr€ unzweig & K€ orner, 2000; Th€ urig et al., 2003).

Primary production of annuals Over all 10 years of the NDFF experiment, primary production of annuals was not stimulated by elevated (CO2) (Fig. 5), in contrast to our initial hypothesis but in agreement with Newingham et al.’s (2013) observation of no increase in perennial plant production at the NDFF over the 10-year experiment. Only in Bromus in the wet El Ni~ no year was production higher; in that case, higher production was achieved at elevated (CO2) through both larger plant size and higher plant density. In the drier years with shorter spring growing seasons, elevated (CO2) did not stimulate either plant size or plant density. Given no increase in total annual community biomass (Fig. 5) and no differences in reproductive allocation (Fig. 4), increases in seed rain observed under elevated (CO2) in the 1998 El Ni~ no year were not repeated at elevated (CO2) for the remaining 9 years of the experiment. Our results agree with studies finding no increase in net production at elevated (CO2) in annual grassland (Zavaleta et al., 2003), calcareous grassland (Niklaus & K€ orner, 2004), and alpine forefields (Inauen et al., 2012), whereas other studies of herbaceous communities have shown elevated (CO2) to stimulate production in

890 S . D . S M I T H et al. shortgrass steppe (Morgan et al., 2004) and a deciduous forest understory (Souza et al., 2010). In several studies, soil nutrient availability appears to influence the production response to elevated (CO2), with nutrient enhancement often required to effect a stimulation in production (Gr€ unzweig & K€ orner, 2000; Reich et al., 2006). In the Mojave Desert, shrub fertile islands (Titus et al., 2002) and soils with biological soil crust cover (DeFalco et al., 2001) have higher N content and therefore support higher annual plant biomass. However, we did not observe a pronounced elevated (CO2) effect on stand-level annual plant biomass on fertile island microsites except in the wet El Ni~ no year. Therefore, when assessed over the whole 10-year experiment, the lack of an elevated (CO2) effect on community-level production does not appear to be a soil nutrient effect. A general acceleration of the life cycle at elevated (CO2), resulting in few differences in asymptotic plant size in most years, appears to be a more plausible explanation of the overall lack of a production response over the long term. A general acceleration in phenology [at both (CO2) levels] in dry years could also be a significant factor in the lack of an elevated (CO2) growth response due to the onset of water stress, particularly in light of recent findings that elevated (CO2) may not increase drought tolerance in annual plants, confining greater production to years when higher water use can occur (Medeiros & Ward, 2013).

Community structure of annuals Assessing the response of the ephemeral community to global change is difficult due to the episodic nature of this life form and substantial population variation resulting from year-to-year differences in rainfall (Table 1; Guo & Brown, 1996; Angert et al., 2007). As in other studies that have shown one or several species to largely determine herbaceous community responses to elevated (CO2) (Berntson et al., 1998; Morgan et al., 2004), Bromus and Lepidium similarly dominated the response of the Mojave Desert annual community. The most striking response was the interannual dynamics of Bromus. Our initial hypothesis suggested that Bromus would become more dominant in the community with elevated (CO2) based on glasshouse experiments with the closely related Bromus tectorum (Smith et al., 1987; Ziska et al., 2005), early results from the NDFF (Smith et al., 2000), and differential responses of invasive species in other settings (e.g., Belote et al., 2003). But after a strong relative response of Bromus in the 1998 El Ni~ no, this exotic grass was largely absent from the aboveground community in the intervening dry years of 1999–2004 (Fig. 5). Following massive seed production in 1998 but then no plant germination in the

extremely dry year of 1999, the NDFF received sufficient rain in the early spring of 2000 to stimulate a dense germination event. However, no more rainfall occurred for the duration of the growing season (Fig. 1, red line), resulting in extremely small plants and little or no seed production in 2000. Although this anomalous year affected all annual species, it impacted the Bromus population more. Bromus produces fewer, larger seeds than do native dicots, and Bromus seeds readily germinate whereas natives maintain partial dormancy of their seed bank even after ideal germination conditions (DeFalco et al., 2003). Therefore, early and uniform germination conditions followed by drought can lead to population crashes in Bromus but not in native species (Salo, 2004). This pattern of population crash apparently occurred for Bromus at the NDFF after the anomalous rainfall year in 2000, with Bromus largely absent from the community for the next 3 years. With above-average and average rainfall returning in 2005 and 2006, respectively, Bromus exhibited a significant re-emergence in the community. Because of our inability to destructively sample soils while elevated (CO2) exposure was occurring, we were not able to quantify the soil seed bank concomitantly with aboveground community structure over the 10-year experiment. However, with the final destructive harvest at the end of the experiment, we found Bromus to be a dominant (with Lepidium) in the soil seed bank (Fig. 6). Interestingly, a second exotic grass species, Schismus arabicus, was also common in the seed bank despite little presence prior to 2006. Of particular note was the increased relative density of Bromus in the seed bank of the elevated (CO2) plots. Although a strong re-emergence of Bromus density and biomass (and therefore seed rain) occurred in 2005 and 2006, Bromus seed production was not substantially higher at elevated (CO2) in those years (Figs 4 and 5). This unexpected result cannot be adequately explained by aboveground production processes in the preceding years. In addition, glasshouse experiments have previously shown that although elevated (CO2) results in Bromus producing smaller seeds, there was no (CO2) effect on seed germination percentage or timing (Huxman et al., 1998). Given the strong increase in seed rain of Bromus at elevated (CO2) in the 1998 El Ni~ no year, it could be that an increase in seed rain at elevated (CO2) in 2005 and 2006, although not readily detected in our productivity-by-reproductive allocation analysis, could have been increased through the production of smaller seeds at elevated (CO2). Nevertheless, combining production and seed bank results, Bromus had clearly re-emerged in the community after two consecutive years of higher rainfall, and appeared to be returning to levels observed in the initial years of the experiment. © 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 879–892

E L E V A T E D C O 2 E F F E C T S O N D E S E R T A N N U A L S 891

Global change implications Early indications that the invasive Bromus would become more dominant in the Mojave Desert at elevated (CO2) (Smith et al., 2000) were not sustained in a long intervening dry cycle, although a strong recovery of Bromus occurred in later years that favored Bromus in the seed bank at the end of the experiment. Seed production by annual plants is an important process in deserts. Seed input affects future population dynamics, but seeds also affect the transfer of carbon and nitrogen to soil pools and are a primary food source for granivores. Quantitatively, the changes in total seed production may not be dramatic, but qualitatively the change in proportion of Bromus seeds in the total seed rain at elevated (CO2), at least in wet years, may impact higher trophic levels. Bromus seeds are not favored, in comparison to native seeds, in feeding trials with desert granivores (Everett et al., 1978). In addition, Bromus seeds can be smaller at elevated (CO2), and our earlier studies show that the reduction in seed size is a function of reduced nitrogen content that results in a diminished ability to grow as a seedling (Huxman et al., 1998). A reduction in quality and increase in overall proportion of Bromus seed may thus have negative impacts on granivores. Elevated (CO2), through effects on plant water-use efficiency, has been proposed to lead to increased plant cover and biomass in open intercanopy spaces (Sage, 1996). This pattern could lead to greater total plant cover across a Mojave Desert landscape that currently only supports ca. 20% perennial cover. The addition of fine fuels and concomitant increased probability of episodic fire could lead to larger and more intense fires that eliminate shrubs from the landscape (Brooks & Matchett, 2006; Abella, 2009). The lack of a large, consistent biomass response in between-shrub interspaces under elevated (CO2) implies that annual cover and biomass in interspaces may not increase and thus will not lead to additional fuel to carry fire across the landscape above that currently observed. Although our early results suggested that increased cover and hence fuel for wildfires was a strong possibility (Smith et al., 2000), the lack of a sustained productivity response at elevated (CO2) suggests that elevated (CO2) alone may not increase wildfire risk. The effect of other global change drivers, such as precipitation and nitrogen deposition, may ultimately dictate the response of annual plants to elevated (CO2) in desert ecosystems and potential changes in disturbance regimes. We propose that wetter cycles would indeed favor Bromus and also increase fire cycles in the Mojave Desert, but a drier climate, as currently predicted for the region (Seager et al., 2007), reduces the potential advantage of © 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 879–892

Bromus and other invasive grasses over the native dicots, with a concomitant reduction in wildfire risk.

Acknowledgements The NDFF coprincipal investigators – Bob Nowak, Dave Evans, Lynn Fenstermaker – and Beth Newingham provided critical comments on the manuscript. Data collection and sample processing were provided by many individuals, but Derek Babcock and Amrita De Soyza are particularly noted. Eric Knight and Dean Jordan provided excellent logistical and operations support in all phases of the study. The NDFF was supported by the DOE Terrestrial Carbon Processes Program (DE-FG02-03ER63651) and the annuals project was supported by the NSF Ecosystem Studies Program (grants DEB-98-14358 and DEB-02-12812) to SDS and a DOE-EPSCoR traineeship to TEH. The National Nuclear Security Administration’s Nevada Field Office and Brookhaven National Laboratory provided logistical support in the construction and operation of the NDFF.

References Abella SR (2009) Post-fire plant recovery in the Mojave and Sonoran Deserts of western North America. Journal of Arid Environments, 73, 699–707. Angert AL, Huxman TE, Barron-Gafford G, Gerst KL, Venable DL (2007) Linking growth strategies to long-term population dynamics in a guild of desert annuals. Journal of Ecology, 95, 321–331. Baskin CC, Baskin JM (2001) Seeds: Ecology, Biogeography, and Evolution of Dormancy and Germination. Academic Press, New York, NY, USA. Beatley JC (1966) Ecological status of introduced brome grasses (Bromus spp.) in desert vegetation of southern Nevada. Ecology, 47, 548–554. Beatley JC (1974) Phenological events and their environmental triggers in Mojave Desert ecosystems. Ecology, 55, 856–863. Bell KL, Hiatt HH, Niles WE (1979) Seasonal changes in biomass allocation in eight winter annuals of the Mojave Desert. Journal of Ecology, 67, 781–787. Belote RT, Weltzin JF, Norby RJ (2003) Response of an understory plant community to elevated [CO2] depends on differential responses of dominant invasive species and is mediated by soil water availability. New Phytologist, 161, 827–835. Berntson GM, Rajakaruna N, Bazzaz FA (1998) Growth and nitrogen uptake in an experimental community of annuals exposed to elevated atmospheric CO2. Global Change Biology, 4, 607–626. Bowers MA (1987) Precipitation and the relative abundance of desert winter annuals: a 6-year study in the northern Mojave Desert. Journal of Arid Environments, 12, 141–149. Brooks ML, Matchett JR (2006) Spatial and temporal patterns of wildfires in the Mojave Desert, 1980-2004. Journal of Arid Environments, 67, 148–164. Coleman JS, McConnaughay DKM, Ackerly DD (1994) Interpreting phenotypic variation in plants. Trends in Ecology and Evolution, 9, 187–191. Crous KY, Reich PB, Hunter MD, Ellsworth DS (2010) Maintenance of leaf N controls the photosynthetic CO2 response of grassland species exposed 9 years of free-air CO2 enrichment. Global Change Biology, 16, 2076–2088. D’Antonio CM, Vitousek PM (1992) Biological invasions by exotic grasses, the grass/ fire cycle, and global change. Annual Review of Ecology and Systematics, 23, 63–87. DeFalco LA, Detling JK, Tracy CR, Warren SD (2001) Physiological variation among native and exotic winter annual plants associated with microbiotic crusts in the Mojave Desert. Plant and Soil, 234, 1–14. DeFalco LA, Bryla DR, Smith-Longozo V, Nowak RS (2003) Are Mojave Desert annual species equal? Resource acquisition and allocation for the invasive grass Bromus madritensis subsp. rubens (Poaceae) and two native species. American Journal of Botany, 90, 1045–1053. Dyer AR, Hardison JL, Rice KJ (2012) Phenology constrains opportunistic growth response in Bromus tectorum L. Plant Ecology, 213, 103–112. Everett R, Meewig RO, Steven R (1978) Deer mouse preference for seed of commonly planted species, indigenous weed seed, and sacrifice foods. Journal of Range Management, 31, 70–73. Gr€ unzweig JM, K€ orner C (2000) Growth and reproductive responses to elevated CO2 in wild cereals of the northern Negev of Israel. Global Change Biology, 6, 631–638.

892 S . D . S M I T H et al. Guo Q, Brown JH (1996) Temporal fluctuations and experimental effects in desert plant communities. Oecologia, 107, 568–577.

Pake CE, Venable DL (1996) Seed banks in desert annuals: implications for persistence and coexistence in variable environments. Ecology, 77, 1427–1435.

Gutierrez JR, Whitford WG (1987) Chihuahuan Desert annuals: importance of water and nitrogen. Ecology, 68, 2032–2045. Housman DC, Naumburg E, Huxman TE, Charlet TN, Nowak RS, Smith SD (2006) Increases in desert shrub productivity under elevated carbon dioxide vary with water availability. Ecosystems, 9, 374–385. Huxman TE, Smith SD (2001) Photosynthesis in an invasive grass and a native forb at no year in the Mojave Desert. Oecologia, 128, 193–201. elevated CO2 during an El Ni~

Pataki DE, Huxman TE, Jordan DN et al. (2000) Water use of Mojave Desert shrubs under elevated CO2. Global Change Biology, 6, 889–898. Potvin C, Lechowicz MJ, Tardif S (1990) The statistical analysis of ecophysiological response curves obtained from experiments involving repeated measures. Ecology, 71, 1389–1400. Reich PB (2009) Elevated CO2 reduces losses of plant diversity caused by nitrogen deposition. Science, 326, 1399–1402.

Huxman TE, Hamerlynck EP, Jordan DN, Salsman KJ, Smith SD (1998) The effects of parental CO2 on seed quality and subsequent seedling performance in Bromus rubens. Oecologia, 114, 202–208. Huxman TE, Hamerlynck EP, Smith SD (1999) Reproductive allocation and seed production in Bromus madritensis ssp. rubens at elevated CO2. Functional Ecology, 13, 769–777.

Reich PB, Hobbie SE (2012) Decade-long soil nitrogen constraint on the CO2 fertilization of plant biomass. Nature Climate Change, 3, 278–282. doi: 10.1038/nclimate1694. Reich PB, Hobbie SE, Lee T et al. (2006) Nitrogen limitation constrains sustainability of ecosystem response to CO2. Nature, 440, 922–925. Rundel PW, Gibson AC (1996) Ecological Communities and Processes in a Mojave Desert

Huxman TE, Barron-Gafford G, Gerst KL, Angert AL, Tyler AP, Venable DL (2008) Photosynthetic resource-use efficiency and demographic variability in desert winter annual plants. Ecology, 89, 1554–1563. Huxman TE, Kimball S, Angert AL, Gremer J, Barron-Gafford G, Venable DL (2013) Understanding past, contemporary, and future dynamics of plants, populations and communities using Sonoran Desert winter annuals. American Journal of Botany, 100, 1369–1380.

Ecosystem: Rock Valley, Nevada. Cambridge University Press, Cambridge, UK. Sage RF (1996) Modification of fire disturbance by elevated CO2. In: Carbon dioxide, Populations, and Communities (eds K€ orner CH, Bazzaz FA), pp. 231–249. Academic Press, San Diego, CA, USA. Salo LF (2004) Population dynamics of red brome (Bromus madritensis subsp. rubens): times for concern, opportunities for management. Journal of Arid Environments, 57, 291–296.

Inauen N, K€ orner C, Hiltbrunner E (2012) No growth stimulation by CO2 enrichment in alpine glacier forefield plants. Global Change Biology, 18, 985–999. Jordan DN, Zitzer SF, Hendrey GR et al. (1999) Biotic, abiotic and performance aspects of the Nevada Desert Free-Air CO2 Enrichment (FACE) Facility. Global Change Biology, 5, 659–668. Kimball S, Angert AL, Huxman TE, Venable DL (2009) Contemporary climate change in

Schlesinger WH, Raikes JA, Hartley AE, Cross AF (1996) On the spatial pattern of soil nutrients in desert ecosystems. Ecology, 77, 364–374. Seager R, Ting MF, Held I et al. (2007) Model projections of an imminent transition to a more arid climate in southwestern North America. Science, 316, 1181–1184. Smith SD, Strain BR, Sharkey TD (1987) Effects of elevated CO2 enrichment on four Great Basin grasses. Functional Ecology, 1, 139–143.

the Sonoran Desert favors cold-adapted species. Global Change Biology, 16, 1555–1565. Long SP (1991) Modification of the response of photosynthetic productivity to rising temperature by atmospheric CO2 concentrations: has its importance been underestimated? Plant, Cell and Environment, 14, 729–739. Medeiros JS, Ward JK (2013) Increasing atmospheric [CO2] from glacial to future concentrations affects drought tolerance via impacts on leaves, xylem and their integrated function. New Phytologist, 199, 738–748.

Smith SD, Monson RK, Anderson JE (1997) Physiological Ecology of North American Desert Plants. Springer-Verlag, Berlin, Germany. Smith SD, Huxman TE, Zitzer SF et al. (2000) Elevated CO2 increases productivity and invasive species success in an arid ecosystem. Nature, 408, 79–82. Smith SD, Charlet TN, Fenstermaker LK, Newingham BA (2009) Effects of global change on Mojave Desert ecosystems. In: The Mojave Desert: Ecosystem Processes and Sustainability (eds Webb RH, Fenstermaker LK, Heaton JS, Hughson DL, McDonald

Melillo JM, McGuire AD, Kicklighter DW, Moore B III, Vorosmarty CL, Schloss AL (1993) Global climate change and terrestrial net primary production. Nature, 363, 234–240. Morgan JA, Mosier AR, Milchunas DG, LeCain DR, Nelson JA, Parton WA (2004) CO2 enhances productivity, alters species composition, and reduces digestibility of shortgrass steppe vegetation. Ecological Applications, 14, 208–219.

EV, Miller DM), pp. 31–56. University of Nevada Press, Reno, NV, USA. Souza L, Belote RT, Kardol P, Weltzin JF, Norby RJ (2010) CO2 enrichment increases forest understory biomass and accelerates successional development of an understory community. Journal of Plant Ecology, 3, 33–39. Thomas SC, Jasienski M, Bazzaz FA (1999) Early vs. asymptotic growth responses of herbaceous plants to elevated CO2. Ecology, 80, 1552–1567.

Mulroy TW, Rundel PW (1977) Annual plants: adaptations to desert environments. BioScience, 27, 109–114. Nagel JM, Huxman TE, Griffin KL, Smith SD (2004) CO2 enrichment reduces the energetic cost of biomass construction in an invasive desert grass. Ecology, 85, 100–106. Naumburg E, Housman DC, Huxman TE, Charlet TN, Loik ME, Smith SD (2003) Photosynthetic response of Mojave Desert shrubs to free air CO2 enrichment are greatest during wet years. Global Change Biology, 9, 276–285.

Th€ urig B, K€ orner C, St€ ocklin J (2003) Seed production and seed quality in a calcareous grassland in elevated CO2. Global Change Biology, 9, 873–884. Titus J, Nowak RS, Smith SD (2002) Soil resource heterogeneity in the Mojave Desert. Journal of Arid Environments, 52, 269–292. Turner FB, Randall DC (1989) Net production by shrubs and winter annuals in southern Nevada. Journal of Arid Environments, 17, 23–36. Warr SJ, Thompson K, Kent M (1993) Seed banks as a neglected area of biogeographic

Navas M-L, Sonie L, Richarte J, Roy J (1997) The influence of elevated CO2 on species phenology, growth and reproduction in a Mediterranean old-field community. Global Change Biology, 3, 523–550. Newingham BA, Vanier CH, Charlet TN, Ogle K, Smith SD, Nowak RS (2013) No cumulative effect of 10 years of elevated [CO2] on perennial plant biomass components in the Mojave Desert. Global Change Biology, 19, 2168–2181.

research: a review of literature and sampling techniques. Progress in Physical Geography, 17, 329–347. Young JA, Evans RA (1978) Population dynamics after wildfires in sagebrush grasslands. Journal of Range Management, 31, 283–289. Zavaleta ES, Shaw MR, Chiariello NR, Thomas BD, Cleland EE, Field CB, Mooney HA (2003) Grassland responses to three years of elevated temperature, CO2, pre-

Niklaus PA, K€ orner C (2004) Synthesis of a six-year study of calcareous grassland responses to in situ CO2 enrichment. Ecological Monographs, 74, 491–511. Nowak RS, DeFalco LA, Wilcox CS, Jordan DN, Coleman JS, Seemann JR, Smith SD (2001) Leaf conductance decreased under free-air CO2 enrichment (FACE) for three desert perennials in the Nevada desert. New Phytologist, 150, 449–458. Nowak RS, Zitzer SF, Babcock D et al. (2004) Elevated atmospheric CO2 does not conserve soil water in the Mojave Desert. Ecology, 85, 93–99.

cipitation, and N deposition. Ecological Monographs, 73, 585–604. Ziska LH (2003) Evaluation of the growth response of six invasive species to past, present and future atmospheric carbon dioxide. Journal of Experimental Botany, 54, 395–404. Ziska LH, Reeves JB III, Blank B (2005) The impact of recent increases in atmospheric CO2 on biomass production and vegetative retention of cheatgrass (Bromus tectorum): implications for fire disturbance. Global Change Biology, 11, 1325–1332.

© 2013 John Wiley & Sons Ltd, Global Change Biology, 20, 879–892