Land Use Alters the Drought Responses of Productivity and CO2 Fluxes in Mountain Grassland

Land Use Alters the Drought Responses of Productivity and CO2 Fluxes in Mountain Grassland Ecosystems (2018) 21: 689–703 DOI: 10.1007/s10021-017-0178-0 2017 The Author(s). This article is an open access publication Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 1 2 3 Johannes Ingrisch, Stefan Karlowsky, Alba Anadon-Rosell, Roland 1 1 4 2 Hasibeder, Alexander Ko¨nig, Angela Augusti, Gerd Gleixner, and Michael Bahn * 1 2 Institute of Ecology, University of Innsbruck, Sternwartestraße 15, 6020 Innsbruck, Austria; Max Planck Institute for Biogeo- chemistry Jena, Postbox 100164, 07701 Jena, Germany; Department of Evolutionary Biology, Ecology and Environmental Sciences, University of Barcelona, Avinguda Diagonal 643, 08028 Barcelona, Spain; Institute of Agro-Environmental and Forest Biology, CNR Italy, Via G. Marconi n.2, 05010 Porano, TR, Italy ABSTRACT Climate extremes and land-use changes can have low resistance was related to high recovery from major impacts on the carbon cycle of ecosystems. drought and vice versa. In consequence, the overall Their combined effects have rarely been tested. We perturbation of the carbon cycle caused by drought studied whether and how the abandonment of tra- was larger in the managed than the abandoned ditionally managed mountain grassland changes the grassland. The faster recovery of carbon dynamics resilience of carbon dynamics to drought. In an in situ from drought in the managed grassland was associ- common garden experiment located in a subalpine ated with a significantly higher uptake of nitrogen meadow in the Austrian Central Alps, we exposed from soil. Furthermore, in both grasslands leaf intact ecosystem monoliths from a managed and an nitrogen concentrations were enhanced after abandoned mountain grassland to an experimental drought and likely reflected drought-induced in- early-summer drought and measured the responses creases in nitrogen availability. Our study shows that of gross primary productivity, ecosystem respiration, ongoing and future land-use changes have the phytomass and its components, and of leaf area index potential to profoundly alter the impacts of climate during the drought and the subsequent recovery extremes on grassland carbon dynamics. period. Across all these parameters, the managed grassland was more strongly affected by drought and Key words: Carbon cycle; Climate extreme; Gross recovered faster than the abandoned grassland. A primary productivity; Land-use change; Nitrogen; bivariate representation of resilience confirmed an Recovery; Resilience; Resistance; N labelling. inverse relationship of resistance and recovery; thus, Received 6 March 2017; accepted 14 August 2017; published online 15 September 2017 INTRODUCTION Electronic supplementary material: The online version of this article The frequency and severity of extreme climatic (doi:10.1007/s10021-017-0178-0) contains supplementary material, events are expected to increase in the near future, which is available to authorized users. with major implications for the carbon (C) cycle of Authors contributions MB and GG conceived the study, JI, SK, AAR, RH, AK and AA performed the experiment and collected data, JI and SK ecosystems and related feedbacks to the atmo- analysed the data, JI and MB led the writing of the manuscript. All au- sphere and the climate system (Reichstein and thors commented on the drafts and gave final approval for publication. others 2013; Frank and others 2015). On a global *Corresponding author; e-mail: Michael.Bahn@uibk.ac.at 689 690 J. Ingrisch and others scale, severe droughts are amongst the climate ex- Khan and others 2016; Roy and others 2016). In- tremes exerting the strongest effects on the C cycle creased leaf N concentrations can in turn promote (Ciais and others 2005; Reichstein and others 2013; photosynthetic C assimilation and thereby speed up Knapp and others 2015). The overall resilience of ecosystem recovery from drought (Roy and others an ecosystem to climate extremes can be charac- 2016). To date it is not known whether abandon- terized by the resistance, that is, the ability of an ment of mountain grasslands, associated with a ecosystem to persist during disturbance, and the reduction in N availability (Zeller and others 2000; recovery, that is, the ability of a system to return to Robson and others 2007), alters the role of plant N pre-disturbance levels (Holling 1996; Hodgson and uptake during post-drought recovery, and what the others 2015; Oliver and others 2015). Thus, a sys- consequences are for tissue N concentrations and tem may be resilient due to a high resistance, a CO uptake dynamics. high capacity to recover or both. Although the We established an experiment testing whether relationships and potential trade-offs between the and how abandonment of managed mountain components of resilience have been subject to re- grassland changes the resilience of C dynamics to cent discussions (for example, Hodgson and others an extreme early-summer drought. We analysed 2015; Yeung and Richardson 2016), few studies both drought resistance and post-drought recovery have actually tested such relationships on the C of ecosystem CO fluxes and of the phytomass and cycle responses of ecosystems to drought. its components. To understand the relationships Grasslands cover more than one-fifth of the and potential trade-offs between the components 6 2 global land surface (35 9 10 km , Dixon and of resilience in response to land-use change, we others 2014) and constitute an important carbon applied a recently proposed bivariate representa- sink (Conant and others 2001; Smith 2014). In tion of resilience (Nimmo and others 2015) and many mountain regions of Europe, grasslands also tested if the conclusions were robust across the play an important role in the production of fodder. different parameters studied. We furthermore ap- In recent decades, land-use changes have led to the plied a perturbance-based approach (Potts and abandonment of mountain meadows and pastures others 2006; Todman and others 2016) to obtain an (for example, MacDonald and others 2000; Tasser integrated quantification of the overall perturba- and others 2007), with consequences for species tion of the two grasslands by the drought event. composition (Tasser and Tappeiner 2002), produc- Our main hypothesis was that abandonment in- tivity and ecosystem C fluxes (Schmitt and others creases the resistance of C dynamics to and de- 2010), soil C and nitrogen (N) turnover (Zeller and creases their recovery from drought. We others 2000; Robson and others 2007; Meyer and furthermore tested the hypothesis that drought others 2012a; Grigulis and others 2013) and the enhances N uptake and tissue N concentrations in water cycle (Obojes and others 2015). To date, managed grassland, whereas the role of plant N surprisingly few studies have explored how land- uptake during recovery is strongly reduced in use changes affect ecosystem responses to climate abandoned grassland. extremes (Bahn and others 2014). Although man- agement intensity has been suggested to modify MATERIAL AND METHODS grassland responses of productivity and CO fluxes Study Site to precipitation variability and drought (Klumpp and others 2011; Vogel and others 2012; Zwicke The study site is located near Neustift in the Stubai and others 2013), the consequences of an aban- valley in the Austrian Central Alps and is composed donment of managed grasslands for the drought of grasslands differing in land use, including a tra- and post-drought responses of their C dynamics are ditionally managed hay meadow and an aban- largely unknown. doned grassland (Schmitt and others 2010). The Site fertility has been shown to modify grassland two subalpine grasslands are located on a southeast responses to climate change (Grime and others exposed hillside with similar inclination (ca. 20), 2000); however, the role of nutrient availability for average annual temperature (3C), annual precip- ecosystem recovery from drought is still poorly itation (1097 mm) and the same soil type (dystric understood. Recent studies suggest that soil nitro- cambisol). The soil textural fractions for clay, silt gen (N) dynamics can be altered by drought events and sand are 13.3, 36.2 and 50.2%, respectively, on and that rewetting of soil after drought can en- the meadow, and 23.4, 45.5 and 31.1%, respec- hance N mineralization and consequently lead to tively, on the abandoned grassland (Meyer and higher tissue N concentrations (Fuchslueger and others 2012b). The meadow (1820–1850 m a.s.l.; others 2014; Canarini and Dijkstra 2015; Arfin Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 691 4707¢45¢¢N, 1118¢20¢¢E) is cut once per year at grassland, respectively. On 28 June 2014, rain-out peak biomass in early August and is fertilized with shelters were removed and 50 mm of rainwater manure every 2–3 years. Additionally, light grazing was added to each of the monoliths to simulate a by sheep and cattle takes place in spring and late heavy rain event ending the drought, and to autumn. The vegetation community has been achieve a well-defined and rapid rewetting. classified as Trisetetum flavescentis and consists of perennial grasses and forbs dominated by Agrostis Measurements capillaris, Festuca rubra, Ranunculus montanus, Tri- Microclimate folium pratense and T. repens (Bahn and others 2009). The second grassland (1970–2000 m a.s.l.; A microclimate station at the common garden re- 4707¢31¢¢N, 1117¢24¢¢E) has been abandoned corded photosynthetically active radiation (PAR), since 1983. Its vegetation has been classified as precipitation, air temperature and humidity (see Seslerio-Caricetum with some dwarf shrubs and is details in Hasibeder and others 2015). During the rain exclusion, air temperature, humidity and PAR dominated by Sesleria varia, Erica carnea, Carex sem- (S-THB-M002 and S-LIA-M003, onset Computer pervirens and Poa alpina (Schmitt and others 2010; Corporation, Bourne, MA, USA) were additionally Grigulis and others 2013). Further details con- measured in two of the six rain-out shelters. Soil cerning vegetation and soils, as well as the overall water content (Decagon EC-5, 5TM, 5TE; combining nutrient supply and productivity of the two sites, SWC and temperature, logger Em50; Decagon De- can be taken from Bahn and others (2006), Schmitt and others (2010), Meyer and others (2012a), vices, Pullman, WA, USA) and soil temperature Grigulis and others (2013), Fuchslueger and others (sensors S-TMB, logger HOBO Micro Station H21- (2014) and Legay and others (2014). 002; Onset Computer corporation, Bourne, MA, USA) were measured continuously in the main Experimental Set-up rooting horizon (30-min interval) in a subset of the monoliths (SWC: n = 17, Temp: n =14) over the The drought experiment was conducted in a com- whole course of the season. In early May, before the mon garden established at the meadow site (see start of the experiments, all soil moisture sensors were above). At both the meadow and abandoned offset-calibrated in situ after a rainy period, when all grassland, 20 intact vegetation–soil monoliths were monoliths had reached field capacity. To determine extracted in June 2013. The monoliths had a the water balance, we measured the amount of lea- diameter of 25 cm and a height of 28 cm and were chates accumulated in the reservoir of each monolith fit in open-top round stainless steel cylinders over the period of rain exclusion (Obojes and others (height 40 cm), with a reservoir for leachates at the 2015). Total evapotranspiration during the drought bottom (for detailed description see Obojes and experiment was estimated for the subset of monoliths others 2015). The monoliths were installed in the equipped with soil moisture sensors by means of a common garden in a randomized factorial design water balance approach, accounting for the amount with six blocks and were left for almost a year be- of water added, the change of water storage in each fore the drought experiment started in May 2014. monolith as derived from monitored changes in vol- For the drought experiment, each of the six umetric soil water content in the main rooting hori- blocks was covered with a rain-out shelter, which zon, and the amount of leached water (Table S1; see had a base area of 3 9 3.5 m and was open at the also Obojes and others 2015). bottom (up to 0.5 m above ground) and at the top of the face sides to allow air circulation. Rain exclusion was performed with transparent UV-A Phytomass and UV-B transmissive plastic foil (Lumisol clear Aboveground phytomass was sampled destruc- AF, Folitec, Westerburg, Germany, light transmit- tively during three campaigns, at peak drought tance c. 90%). The rain exclusion lasted from 21 (‘‘resistance campaign’’ on 1 July 2014, n = 12, May 2014 until 28 June 2014. During this period, that is, 3 replicates per land use and treatment monoliths allocated to the drought treatment did combination) and twice during the recovery peri- not receive any precipitation. Control monoliths od, that is, 4 (‘‘recovery 1 campaign’’ on 24 July, were manually watered every 1–4 days with pre- n = 12) and 8 weeks (‘‘recovery 2 campaign’’ on 22 viously collected rainwater to maintain soil water August, n = 16) after termination of the experi- content above 25 vol.% (Figure 1D, E). The mental drought. Thus, at each campaign, a separate amount of water added to the controls was 180 and subset of monoliths was harvested by cutting 170 mm for the meadow and the abandoned 692 J. Ingrisch and others Figure 1. Time course of A daily sums of photosynthetically active radiation (PAR), daily means of B air temperature (T ) and C air vapour pressure deficit (vpd) and in the rain-out shelters. C, D Daily means of soil moisture (vol.%) in the air main rooting horizon in monoliths from the meadow (control n = 3, drought n = 5) and the abandoned grassland (control n = 4, drought n = 5) exposed to ambient conditions (control, solid line) and drought (dashed line). Shaded areas show the standard error of the mean. Vertical bars show daily precipitation (open = natural, shaded = manual watering). Note that during rain exclusion (horizontal black bar) only monoliths from the control treatment received water. phytomass to 2 cm aboveground. The samples were (3 cm diameter), washed, sieved to 2 mm and mi- frozen at -18C until further analysis. crowaved before transporting to the laboratory. Phytomass samples were split into four func- All plant samples were oven-dried at 60C, tional groups (forbs, grasses, legumes, dwarf ground, weighed (2–5 mg and analysed on an shrubs), and into stems, leaves, reproductive or- elemental analysis—isotope ratio mass spectrome- gans and living phytomass (hereafter biomass) and ter (EA-IRMS; EA 1100, CE Elantech, Milan, Italy; necromass (senesced plant parts) were separated. coupled to a Delta + IRMS, Finnigan MAT, Bre- For each functional group in each monolith, men, Germany). LNC was calculated based on the specific leaf area (SLA) was obtained for a subset of peak area and the known nitrogen concentration of leaves saturated with water and scanned (V700 external acetanilide standards. The d N was Photo, Epson, WinRHIZO Pro 2012, Regent determined in per mil (%) relative to the interna- Instruments) and subsequently dried at 60C for tional reference standard AIR-N using IAEA-N1 3 days. Leaf area index (LAI) was calculated from (Werner and Brand 2001). The amount of N label the leaf biomass and SLA for each functional group. recovered in roots and shoots is calculated as: Community-weighted mean (CWM) of SLA was ðÞ atom%  atom% N labelled NA pool calculated as the leaf-biomass-weighted mean of incorporated N ¼ 100% SLA of each functional group. with atom% being the N atom% of the la- labelled 15 15 Tissue Nitrogen and N Labelling belled samples, atom% being the N atom% of NA natural abundance samples and N being the pool For each sampling campaign (see above), we -2 respective nitrogen pool (mg N m ). measured the leaf nitrogen concentration (LNC) and its nitrogen isotope ratio (d N) on a subset of CO Fluxes leaves sampled in each of the monoliths. Further- We measured the net ecosystem exchange (NEE) of more, at the end of the drought experiment, we CO using closed dynamic chambers, similar to the performed a N pulse labelling experiment on the system applied by Schmitt and others (2010). The 12 monoliths sampled during the Recovery 1 15 15 chambers were transparent Plexiglas cylinders (di- campaign. 20 mg of KNO with 10% N(2 mg N ameter 25 cm, height 50 cm) which fitted airtight per monolith) dissolved in 100 ml rain water was on the steel cylinders containing the monoliths. distributed equally on the soil of the monoliths. Pressure effects on CO fluxes were avoided by a During the subsequent harvest, both the shoots and 2 hole in the top of the chamber, which was closed the roots from the uppermost 7 cm of the soil were with a plug after placing the chamber. Air inside sampled. Roots were extracted with a soil corer Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 693 the chambers was ventilated with fans. Concen- ment points is a measure of change within this trations of CO (GMP 343, Vaisala Helsinki, Fin- bivariate space, and the cumulative length of each land) and water vapour, as well as temperature trajectory is a measure of the overall perturbation (HMP 75, Vaisala, Helsinki Finland) were logged for (Potts and others 2006). 1 min with 5-s intervals. During each measure- The resistance of GPP , ER, leaf area index (LAI) sat ment, the photosynthetically active radiation and biomass was determined based on the mea- (PQS1 PAR Quantum Sensor, Kipp & Zonen, Delft, surements during peak drought. We express resis- the Netherlands) was recorded. Ecosystem respi- tance (RST) as the ratio of drought to control ration (ER) was measured by covering the chamber measurements (Kaufman 1982). A recovery index with a dark cloth, excluding any light inside the according to Nimmo and others (2015) was calcu- chamber. To obtain estimates of gross primary lated for GPP , ER, LAI and biomass. This index is sat productivity (GPP), paired measurements of NEE a measure of the post-drought change of the under sunlit and dark conditions were taken. parameter. We adapted the approach by using Monoliths were measured in randomized order in measurements of control monoliths instead of the morning hours on days with clear sky. In pretreatment measurements to account for sea- addition to sunlit and dark measurements, a series sonal changes in the controls. The recovery index of light response curves were obtained for each was calculated as D /C – RST, where D and rec rec rec treatment type, using layers of semitransparent C denote parameter values during recovery in rec cloth (Schmitt and others 2010). To obtain a con- the drought and control treatment, respectively, sistent time series of CO fluxes throughout the and RST is the resistance of the parameter (see whole study, flux measurements were taken on the above). monoliths which were harvested during the last All calculations and statistical analyses were campaign (Recovery 2). performed in R 3.2.3 (R Development Core Team CO flux rates were calculated as described by 2015). We used permutational ANOVA with the Schmitt and others (2010). Each measurement was package lmPerm (Wheeler 2010). For each sam- quality controlled based on visual inspection and pling day, we tested for the interaction of land-use quality of the linear fits as recently recommended type and drought treatment and drought effects by Pirk and others (2016). GPP was calculated as within each land-use type. the difference of the corresponding NEE and ER measurements. Throughout this study, GPP and ER RESULTS fluxes are both assigned positive signs. For each Key meteorological variables during the experi- land-use type and precipitation treatment, light mental period are presented in Figure 1A–C. Dur- response curves were obtained from pooled data by ing the rain exclusion, soil moisture declined to less fitting a rectangular hyperbolic model (Ruimy and than 20 vol.% in the monoliths from both grass- others 1995; Schmitt and others 2010). Above a -2 -1 lands which were exposed to drought (Figure 1 D, photon flux density (PFD) of 1000 lmol m s , E). The drought treatment reduced the total all light response curves levelled off and reached amount of evapotranspiration significantly 80–85% of the maximum values (Figure S1). Thus, (p < 0.001) in both grasslands (Table S1). Neither for the sake of comparability across treatments we land use nor the interaction of drought and land only present data obtained at PFD above this use had a significant effect on the evapotranspira- threshold and apply the term light-saturated GPP (GPP ). For our analysis, we only included fluxes tion. sat from measurement days for which at least three quality-controlled replicate data sets per land use Aboveground Plant Productivity and and treatment combination were available. Nitrogen Relations The meadow had a generally higher biomass and Calculations of Indices and Statistics leaf area index (LAI) compared to the abandoned To obtain normalized fluxes, values of GPP and grassland, but these differences declined in the sat ER in the drought treatment were divided by their course of the season (Figure 2A, B, G and H, Fig- respective values in the controls. The daily means ure S3), reflecting a delayed development and of the normalized fluxes define the response tra- lower aboveground net primary production of the jectory of each grassland in the bivariate space of abandoned grassland and an earlier plant senes- normalized GPP and normalized ER. The cence of the meadow plants (Figure 2C, D). During sat Euclidian distance between consecutive measure- the period of peak growth in early July, leaf area 694 J. Ingrisch and others CD EF G H Figure 2. A–F Biomass, necromass, phytomass (sum of biomass and necromass) and G, H leaf area index (LAI) of monoliths from the meadow and the abandoned grassland subjected to ambient conditions (open symbols) and drought (closed symbols) and sampled at the end of drought (1 July) and during recovery (24 July, 22 August). Error bars indicate standard errors of the mean (n = 3 for July and n = 4 for August samplings), and stars indicate significant treatment effects within land use and sampling date (p value: *** < 0.001 < ** < 0.01 < * < 0.05 < (*) < 0.1). 2 -1 index (LAI) was higher on the meadow compared significantly (meadow: 10.1 ± 1.3 m kg , aban- 2 -1 to the abandoned grassland. The fraction of grasses doned grassland: 7.3 ± 0.8 m kg ). was significantly higher in the abandoned grass- Drought reduced biomass significantly in the lands (78 ± 3%) compared to the meadow meadow (Figure 2A), but not in the abandoned (55 ± 5%, p < 0.001). The community-weighted grassland (Figure 2B, Table 1; Figure S2 and mean (CWM) of SLA for the meadow and the Tables S2 and S4). The drought-induced reduction abandoned grassland was 14.4 ± 0.8 and was persistent for 4 weeks, but disappeared later 2 -1 9.3 ± 0.6 m kg , respectively (p < 0.001). Forbs when biomass in the control monoliths declined had a significantly higher mean SLA in the mea- (Figure 2A). Drought did not immediately induce 2 -1 dow (20.3 ± 1.5 m kg ) compared to the aban- leaf senescence at either site, but significantly re- 2 -1 doned grassland (15.2 ± 1.5 m kg , p < 0.001), duced necromass in the abandoned grassland in the whereas the mean SLA of grasses did not differ late season (Table 1, Figure 2C, D). LAI was re- Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 695 Table 1. Drought and Land-Use Effects on Phytomass Parameters Campaign Weeks after rewetting Treatment Land use Treatment * land use Biomass Resistance 0 *** n.s. * Recovery 1 3.5 *** n.s. n.s. Recovery 2 8 n.s. n.s. n.s. Necromass Resistance 0 * n.s. n.s. Recovery 1 3.5 n.s. n.s. * Recovery 2 8 n.s. n.s. n.s. Phytomass Resistance 0 *** n.s. * Recovery 1 3.5 *** n.s. n.s. Recovery 2 8 n.s. n.s. n.s. LAI Resistance 0 *** *** ** Recovery 1 3.5 * (*) n.s. Recovery 2 8 * *** n.s. Biomass of forbs Resistance 0 n.s. ** * Recovery 1 3.5 ** ** * Recovery 2 8 * ** * Biomass of grasses Resistance 0 *** n.s. n.s. Recovery 1 3.5 n.s. n.s. n.s. Recovery 2 8 n.s. (*) n.s. LNC Resistance 0 n.s. n.s. Recovery 1 3.5 *** (*) n.s. Recovery 2 8 * n.s. (*) d15N Resistance 0 n.s. *** n.s. NA Recovery 2 8 n.s. *** n.s. Results of permutational ANOVA testing overall treatment effect, land-use effect and their interaction on different phytomass parameters. Biomass = living phytomass; Necromass = senescent phytomass; LAI = leaf area index; LNC = leaf nitrogen concentration; d15N = natural abundance nitrogen isotope ratio of leaves. Resistance = 1 NA July, peak drought, Recovery 1 = 24 July, Recovery 2 = 22 August. Stars indicate the significance level: *** < 0.001 < ** < 0.01 < * < 0.05 < (*) < 0.1 duced by drought in the meadow (Table 1, Fig- uptake of N during the recovery from drought ure 2G, H), but exceeded the values in the controls (Figure 5). at the Recovery 2 sampling. This was also reflected by leaf mass dynamics (Figure S2). Biomass of forbs CO Fluxes responded differently between the two grasslands During drought, GPP was progressively reduced sat (Table 1, Figure 3A, B, Figure S3) and was less to 20 and 40% of the controls in the meadow and resistant but recovered more rapidly in the mea- the abandoned grassland, respectively (Figure 6A, dow. In contrast, biomass of grasses was signifi- B); the interaction of land use and drought was cantly reduced by drought and recovered quickly in significant (Table 2). At peak drought, ER was re- both grasslands (Table 2, Figure 3C, D, Figure S3). duced by up to 60 and 25% on the meadow and During the recovery period, LNC was signifi- the abandoned site, respectively (Figure 6C, D); cantly higher in monoliths previously exposed to however, the interaction of land use and drought drought compared to the controls. The effect was was not significant. After rewetting, GPP fully sat observed in both grasslands and was more pro- recovered within 9 days in both grasslands (Ta- nounced in the meadow (Table 1, Figure 4A, B). ble 2). Its recovery rate was distinctly higher in the Leaf d N was generally higher in the meadow than meadow compared to the abandoned grassland in the abandoned grassland (Table 1, Figure 4C, D) -2 -1 (23.4 versus 14.6 lmol m s over the period of and was not affected by drought. 9 days). During the recovery, ER was significantly Following N pulse labelling meadow plants enhanced in previously drought-exposed mono- recovering from drought took up significantly more liths relative to controls in both grasslands, the ef- N label than the controls (p < 0.01) and incor- fect being more pronounced and sustained in the porated this nitrogen into shoots (+110%, meadow than in the abandoned grassland (Fig- p < 0.01), but not into roots (Figure 5). In con- ure 6C, D; Table 2). trast, the abandoned grassland did not increase its 696 J. Ingrisch and others A B Figure 3. Biomass of A, B forbs and C, D grasses of monoliths from the meadow and the abandoned grassland under ambient conditions (open symbols) and drought conditions (closed symbols), sampled at the end of the drought treatment (1 July) and during recovery (24 July, 22 August). Error bars indicate standard errors of the mean (n = 3 in July and n =4 in August), asterisks indicate significant treatment effects within land use and sampling date (p value: *** < 0.001< ** < 0.01). Effects of Land-Use Change on Integrated DISCUSSION Drought Responses Does Land Use Alter Drought Resistance To integrate the drought responses across the two and Recovery of Productivity and CO grasslands, the resistance and the recovery of Fluxes? studied parameters were related in a bivariate ap- The conceptualization of resilience in ecology has proach (see methods). Drought resistance generally led to contrasting definitions and terminologies, increased from GPP to ER to LAI to biomass and sat which have recently been under some debate (e.g. was generally lower in the meadow than in the Hodgson and others 2015; Yeung and Richardson abandoned grassland (Figure 7). In contrast, across 2016). For analysing the disturbance responses of a all parameters, the recovery index was higher in system within a given stability domain (‘‘engi- the meadow than in the abandoned grassland, neering resilience’’, sensu Holling 1996), it has yielding an overall negative relationship between been suggested to distinguish resistance and resistance and recovery. recovery as the two major underlying processes of The effect of land-use change on the pertur- resilience (Hodgson and others 2015; Nimmo and bation of the two CO flux components GPP 2 sat others 2015; Oliver and others 2015). In our study, and ER was assessed in more detail by comparing we observed that both grasslands were highly re- their drought response trajectories for the mea- silient to drought, confirming conclusions from dow and the abandoned grassland. While the earlier studies (Gilgen and Buchmann 2009; trajectories had similar shapes for both grass- Jentsch and others 2011; Hoover and others 2014), lands (Figure 8A), the cumulative length of the though it should be noted that the timing, the response trajectory for the meadow was 33% magnitude and the interannual pattern of drought larger than for the abandoned grassland (Fig- may modify specific grassland drought responses ure 8B). (Knapp and others 2015; Estiarte and others 2016; Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 697 meadow abandoned ** *** ** 2.4 2.0 1.6 1.2 CD ∂d∂ -1 -2 Jul 01 Jul 15 Aug 01 Aug 15 Jul 01 Jul 15 Aug 01 Aug 15 Control Drought Figure 4. A, B Leaf nitrogen concentrations (LNC) and C, D corresponding natural abundance d N isotope values of leaves in monoliths from the meadow and the abandoned grassland subjected to ambient conditions (open symbols) and drought (closed symbols) and sampled at peak drought (1 July) and during recovery (24 July, 22 August). Error bars indicate standard errors of the mean (n = 3 for July and n = 4 for August samplings), stars indicate significant treatment effects within land use and sampling date (p value: *** < 0.001 < ** < 0.01 < * < 0.05). Table 2. Drought and Land-Use Effects on CO Fluxes Date Days after rewetting Treatment Land use Treatment * land use Meadow Abandoned GPP Pretreatment n.s. (*) n.s. n.s. n.s. sat Resistance *** *** * *** *** Recovery 1 5 ** n.s. n.s. * *** Recovery 2 9 * (*) n.s. n.s. (*) Recovery 3 17 n.s. n.s. n.s. (*) n.s. ER Pretreatment n.s. n.s. n.s. n.s. n.s. Resistance ** (*) n.s. ** * Recovery 1 5 ** n.s. n.s. *** * Recovery 2 9 n.s. n.s. n.s. * n.s. Recovery 3 17 n.s. n.s. * n.s. * Recovery 4 30 n.s. n.s. n.s. n.s. n.s. Results of permutational ANOVA testing overall treatment effect, land-use effect and their interaction as well as within-land-use effects of drought on key dates of experiment. GPP = light-saturated rate of gross primary productivity, ER = ecosystem respiration. Dates: Pretreatment = 19 May, Resistance = 26 June, Recovery 1 = 3 July, Recovery sat 2 = 7 July, Recovery 3 = 15 July, Recovery 4 = 28 July. Stars indicate the significance level: *** < 0.001 < ** < 0.01 < * < 0.05 < (*) < 0.1 Hoover and Rogers 2016). Furthermore, we found studied C cycle parameters. This notion is con- that the role of resistance and recovery for resi- firmed when applying a bivariate approach (Fig- lience can be strongly altered by land-use change: ure 7), as recently suggested in the literature although the abandoned grassland had a distinctly (Hodgson and others 2015; Nimmo and others higher drought resistance, the managed meadow 2015): the abandoned grassland was resilient due displayed a higher recovery rate across all the to a high resistance, whereas the meadow was less LNC[%] ∂ N [‰] NA 698 J. Ingrisch and others Figure 5. Amount of N label recovered in the two grasslands in A shoots and B roots of control monoliths (open bars) and in monoliths recovering from drought (shaded bars) 3 weeks after the rewetting. Error bars indicate standard errors of the mean (n = 3). Stars B indicate significant differences between control and drought treatment (p value: ** < 0.01). Figure 6. A, B Light- saturated rates of gross AB primary productivity (GPP ) and C, D sat ecosystem respiration (ER) in monoliths from the meadow and the abandoned grassland subjected to ambient conditions (open symbols) and drought (closed C D symbols). Error bars indicate standard errors of the mean (n = 3–4). The horizontal black bars indicate the time of rain exclusion. resistant, but resilient due to a high recovery. In ment was smaller than the concurrent gain of consequence, the overall perturbation of the C resistance. cycle caused by drought was larger in the managed The consistent trade-off between resistance and than the abandoned grassland, as indicated by a recovery between the two studied grasslands was greater length of a multivariate response trajectory likely related to differences in plant community (Potts and others 2006, Figure 8). This suggests that composition and associated differences in the pre- the loss of ability to recover caused by abandon- vailing strategies of plant species to cope with Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 699 drought. Our results suggest that the different re- sponses of the grasslands were not predominantly driven by their relative composition of grasses versus forbs, as the drought and post-drought recovery response of grasses was similar in the meadow and the abandoned grassland. However, the drought response of forbs differed distinctly between the two grasslands: forb biomass was less affected by drought in the abandoned grassland than in the meadow (Figure 3), which contributed to the higher resistance of the abandoned grass- land. The forbs in the abandoned grassland were characterized by lower mean SLA compared to the meadow, a trend already previously observed both for SLA and LNC across gradients of decreasing land-use intensity (Bahn and others 1999; Grigulis Figure 7. Resistance and recovery of the investigated C and others 2013; but note that reduced grazing cycle parameters in the two grasslands (open = meadow, shaded = abandoned grassland). The resistance was cal- intensity can also favour species with higher SLA culated as the ratio of parameter performance in drought ´ and LNC, see Laliberte and others 2012). Species plots relative to the parameter performance in the control with lower SLA and LNC are associated with lower plots during peak drought. The recovery index is a growth rates (Lambers and Poorter 1992; Wright measure for the absolute recovery of the parameter after and others 2004) and are characterized by a higher the end of the drought. High values indicate high resis- tolerance to nutrient stress (Garnier and others tance and recovery, respectively. The arrow indicates the 2004; Que´tier and others 2007; Grigulis and others shift in resilience caused by abandonment. ER ecosystem 2013). These ‘‘conservative species’’ have been respiration, GPP light-saturated gross primary produc- suggested to be more resistant, but less capable of tivity, LAI leaf area index. recovering quickly from disturbance (Lambers and Poorter 1992; MacGillivray and Grime 1995; Reich 2014). Conversely, communities dominated by AB Figure 8. A The course of normalized light-saturated rates of gross primary productivity (GPP ) and ecosystem respi- sat ration (ER) before (grey points), during (open points) and after (black points) the drought experiment in the meadow (solid line) and the abandoned grassland (dotted line). Normalized fluxes were calculated as the ratio of the flux in drought monoliths to the respective flux in control monoliths. The direction of the path is given by the arrow, symbols denote the periods before (shaded), during (open) and after (closed) drought. B Cumulative Euclidian distance of the response tra- jectories of the two grasslands over the course of the drought. The Euclidian distance between two consecutive mea- surements days is a measure of the system’s change in the bivariate flux space. The cumulative Euclidian distance from beginning of the drought (pretreatment) is a measure of the overall perturbation of the grassland. The black horizontal bar indicates period of rain exclusion. 700 J. Ingrisch and others ‘‘exploitative species’’ (fast growth related to a recovery, and indicates that the effect was more higher SLA and LNC) have been shown to recover pronounced in the managed meadow (Figure 4A, better from climatic disturbances (Lepsˇ and others B). This is in line with recent observations that 1982; Grime and others 2000). In our study, all the resource pulses can be larger under intensive C cycle parameters studied were more susceptible compared to extensive management (Fuchslueger to drought but recovered more rapidly in the and others 2014; Schrama and Bardgett 2016) and managed meadow, which is more strongly domi- suggests a higher post-drought availability of N in nated by exploitative species. the meadow compared to the abandoned grassland. Vegetation phenology has been suggested to be Higher rewetting-induced resource pulses in the sensitive to climate extremes (for example, Jentsch meadow might also be reflected by a more pro- and others 2009). In our study, phenological nounced stimulation of CO release from soil dynamics during post-drought recovery likely re- (‘‘Birch effect’’, Figure 6C, D), which has fre- flected contrasting plant strategies of the two quently been associated with a rapid mineralization grasslands. The meadow built up new biomass and of organic matter (for example, Borken and Matz- increased leaf area more rapidly (Figure 2A, G), ner 2009; deVries and others 2012). reflecting its fast-growth strategy. In contrast, the Tissue N concentrations result from the uptake of lower necromass in the abandoned grassland at the N and its dilution by growth. Since during its last sampling date (Figure 2D) indicates a delayed recovery from drought, the meadow increased tis- leaf senescence of that plant community in re- sue N concentrations while producing more new sponse to the drought. Both grasslands can thereby biomass than the abandoned grassland (Figure 2), potentially maintain a higher C uptake later in the it must have taken up distinctly more N. This is season to compensate C deficits from drought confirmed by our NO labelling experiment, (Casper and others 2001). which suggests a doubling of nitrate uptake in It has been shown that with increasing time since monoliths from the meadow during recovery from abandonment the resistance of abandoned fields to drought, while no clear effect was observed for the drought (Lepsˇ and others 1982) and of shrublands abandoned grassland (Figure 5). The labelling to long-term warming and drought (Kro¨ el-Dulay experiment suggests that increased N uptake on the and others 2015) increases. As secondary succes- meadow was not only caused by more pronounced sion proceeds, an increasing dominance of woody resource pulses, but also by a higher root N uptake species (shrubs, trees) also favours plants with capacity, which is typically higher for exploitative, more conservative water use, which could buffer fast-growing species (Osone and others 2008; effects of dry periods on soil moisture (Teuling and Grassein and others 2015). Interestingly, on the others 2010; Wolf and others 2013; Gavazov and meadow the additional N taken up during recovery others 2014). As suggested by our results for the did not remain in the roots, but was used for the water balance (Table S1), such a water-sparing production of aboveground biomass. Since the strategy was not yet observed in the abandoned natural abundance d N of leaves was not affected grassland of our study, where herbaceous species by drought (Figure 4C, D), it can be assumed that prevailed. the major sources of N were not strongly altered by drought (Craine and others 2015). Collectively, our Does Land Use Alter N Uptake During results suggest that drought can increase soil N availability and enhance the potential for the re- Recovery from Drought? growth of biomass after drought, which con- Site fertility has been shown to modify grassland tributed particularly to the rapid recovery on the responses to climate change (Grime and others meadow, where the resource pulse during rewet- 2000); however, the immediate role of nutrient ting was larger and roots had a stronger capacity for availability for ecosystem recovery from drought is N uptake. still poorly understood. Recent studies have sug- gested that drought-induced increases in N turn- CONCLUSIONS over and availability can increase plant tissue N concentrations (Fuchslueger and others 2014; Ca- We conclude that the studied grasslands are highly narini and Dijkstra 2015; Arfin Khan and others resilient to extreme drought and that land-use 2016) and can thereby enhance grassland CO change has a strong potential for altering the rela- uptake dynamics during recovery (Roy and others tive contributions of resistance and recovery to the 2016). Our study supports the hypothesis that overall resilience. Our results suggest that aban- drought can increase leaf N concentrations during donment increases the resistance and decreases the Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 701 Bahn M, Reichstein M, Dukes JS, Smith MD, McDowell NG. recovery of grassland carbon dynamics across dif- 2014. Climate-biosphere interactions in a more extreme ferent carbon cycle parameters and different mea- world. New Phytol 202:356–9. sures of resilience. Rapid recovery from drought Bahn M, Schmitt M, Siegwolf R, Richter A, Bru¨ ggemann N. was supported by drought-induced increases in 2009. Does photosynthesis affect grassland soil-respired CO nitrogen availability and enhanced leaf nitrogen and its carbon isotope composition on a diurnal timescale? concentrations, which was more pronounced in New Phytol 182:451–60. the managed grassland. We conclude that managed Bahn M, Wohlfahrt G, Haubner E, Horak I, Michaeler W, Rottmar K, Tappeiner U, Cernusca A. 1999. Leaf photosyn- mountain grassland is likely prone to larger overall thesis nitrogen contents and specific leaf area of grassland perturbations from extreme early-summer species in mountain ecosystems under different land use Land droughts compared to abandoned grassland. use changes in European mountain ecosystems: ECOMONT Therefore, ongoing and future land-use changes concepts and results. Vienna: Blackwell. pp 247–55. have the potential to significantly alter impacts of Borken W, Matzner E. 2009. Reappraisal of drying and wetting climate extremes on grassland carbon dynamics. effects on C and N mineralization and fluxes in soils. Glob Change Biol 15:808–24. Canarini A, Dijkstra FA. 2015. Dry-rewetting cycles regulate ACKNOWLEDGEMENTS wheat carbon rhizodeposition, stabilization and nitrogen cy- Open access funding provided by University of cling. Soil Biol Biochem 81:195–203. Innsbruck and Medical University of Innsbruck. We Casper BB, Forseth IN, Kempenich H, Seltzer S, Xavier K. 2001. Drought prolongs leaf life span in the herbaceous desert thank Sarah Scheld, Mario Deutschmann, David perennial Cryptantha flava. Funct Ecol 15:740–7. Reinthaler and Marine Zwicke for assistance in the Ciais P, Reichstein M, Viovy N, Granier A, Oge´e J, Allard V, field. The study was conducted at the LTER-Austria Aubinet M, Buchmann N, Bernhofer C, Carrara A, Chevallier master site ‘‘Stubai Valley’’ and was financially F, de Noblet N, Friend AD, Friedlingstein P, Gru¨ nwald T, supported by the Austrian Science Fund (FWF Pro- Heinesch B, Keronen P, Knohl A, Krinner G, Loustau D, ject No I 1056) and by the German Federal Ministry Manca G, Matteucchi G, Miglietta F, Ourcival JM, Papale D, Pilegaard K, Rambal S, Seufert G, Soussana JF, Sanz M-J, of Education and Research (BMBF Project No. Schulze E-D, Vesala T, Valentini R. 2005. Europe-wide 01LC1203A) in the framework of the ERA-Net reduction in primary productivity caused by the heat and BiodvERsA project ‘‘REGARDS’’, coordinated by drought in 2003. Nature 437:529–33. Sandra Lavorel. A.A-R. was funded by a STSM of the Conant RT, Paustian K, Elliott ET. 2001. Grassland management COST Action STReESS (PF1106) and a FPU grant and conversion into grassland. Effects on soil carbon. Ecol from the Ministerio de Educacio´ n, Cultura y De- Appl 11:343–55. porte, Spain. Final data analysis and writing of this Craine JM, Brookshire ENJ, Cramer MD, Hasselquist NJ, Koba K, Marin-Spiotta E, Wang L. 2015. Ecological interpretations manuscript were supported by CLIMLUC (‘‘Climate of nitrogen isotope ratios of terrestrial plants and soils. Plant extremes and land use change—effects on ecosystem Soil 396:1–26. processes and services’’), funded by the Austrian deVries FT, Liiri ME, Bjørnlund L, Bowker MA, Christensen S, Academy of Sciences. J.I. received a Ph.D. comple- Seta¨la¨ HM, Bardgett RD. 2012. Land use alters the resistance tion grant from the University of Innsbruck. and resilience of soil food webs to drought. Nat Clim Change 2:276–80. OPEN ACCESS Dixon AP, Faber-Langendoen D, Josse C, Morrison J, Loucks CJ, Ebach M. 2014. Distribution mapping of world grassland This article is distributed under the terms of the types. J Biogeogr 41:2003–19. Creative Commons Attribution 4.0 International Estiarte M, Vicca S, Penuelas J, Bahn M, Beier C, Emmett BA, License (http://creativecommons.org/licenses/by/ Fay PA, Hanson PJ, Hasibeder R, Kigel J, Kroel-Dulay G, Larsen KS, Lellei-Kovacs E, Limousin J-M, Ogaya R, Ourcival 4.0/), which permits unrestricted use, distribution, J-M, Reinsch S, Sala OE, Schmidt IK, Sternberg M, Tielborger and reproduction in any medium, provided you K, Tietema A, Janssens IA. 2016. Few multiyear precipitation– give appropriate credit to the original author(s) and reduction experiments find a shift in the productivity-pre- the source, provide a link to the Creative Commons cipitation relationship. Glob Change Biol 22:2570–81. license, and indicate if changes were made. Frank D, Reichstein M, Bahn M, Thonicke K, Frank DC, Ma- hecha MD, Smith P, van der Velde M, Vicca S, Babst F, Beer C, Buchmann N, Canadell JG, Ciais P, Cramer W, Ibrom A, REFERENCES Miglietta F, Poulter B, Rammig A, Seneviratne SI, Walz A, Wattenbach M, Zavala MA, Zscheischler J. 2015. Effects of Arfin Khan MA, Kreyling J, Beierkuhnlein C, Jentsch A. 2016. climate extremes on the terrestrial carbon cycle. Concepts, Ecotype-specific improvement of nitrogen status in European processes and potential future impacts. Glob Change Biol grasses after drought combined with rewetting. Acta Oeco- 21:2861–80. logica 77:118–27. Fuchslueger L, Kastl E, Bauer F, Kienzl S, Hasibeder R, Ladreiter- Bahn M, Knapp M, Garajova Z, Pfahringer N, Cernsuca A. 2006. Knauss T, Schmitt M, Bahn M, Schloter M, Richter A, Szukics Root respiration in temperate mountain grasslands differing in U. 2014. Effects of drought on nitrogen turnover and abun- land use. Glob Change Biol 12:995–1006. 702 J. Ingrisch and others dances of ammonia-oxidizers in mountain grassland. Bio- Knapp AK, Carroll CJW, Denton EM, La Pierre KJ, Collins SL, geosciences 11:6003–15. Smith MD. 2015. Differential sensitivity to regional-scale drought in six central US grasslands. Oecologia 177:949–57. Garnier E, Cortez J, Billes G, Navas M-L, Roumet C, Debussche M, Laurent G, Blanchard A, Aubry D, Bellmann A, Neill C, Kro¨ el-Dulay G, Ransijn J, Schmidt IK, Beier C, de Angelis P, de Toussaint J-P. 2004. Plant functional markers capture Dato G, Dukes JS, Emmett BA, Estiarte M, Garadnai J, ecosystem properties during secondary succession. Ecology Kongstad J, Kovacs-Lang E, Larsen KS, Liberati D, Ogaya R, 85:2630–7. Riis-Nielsen T, Smith AR, Sowerby A, Tietema A, Penuelas J. 2015. Increased sensitivity to climate change in disturbed Gavazov KS, Spiegelberger T, Buttler A. 2014. Transplantation of ecosystems. Nat Commun 6:6682. subalpine wood-pasture turfs along a natural climatic gradient reveals lower resistance of unwooded pastures to climate Laliberte´ E, Shipley B, Norton DA, Scott D. 2012. Which plant change compared to wooded ones. Oecologia 174:1425–35. traits determine abundance under long-term shifts in soil resource availability and grazing intensity? J Ecol 100:662– Gilgen AK, Buchmann N. 2009. Response of temperate grass- lands at different altitudes to simulated summer drought dif- fered but scaled with annual precipitation. Biogeosciences Lambers H, Poorter H. 1992. Inherent variation in growth rate 6:2525–39. between higher plants: a search for physiological causes and ecological consequences. In: Begon M, Fitter A, Eds. Advances Grassein F, Lemauviel-Lavenant S, Lavorel S, Bahn M, Bardgett in ecological research. London, San Diego: Academic Press. p RD, Desclos-Theveniau M, Laine P. 2015. Relationships be- 187–261. tween functional traits and inorganic nitrogen acquisition among eight contrasting European grass species. Ann Bot Legay N, Baxendale C, Grigulis K, Krainer U, Kastl E, Schloter 115:107–15. M, Bardgett RD, Arnoldi C, Bahn M, Dumont M, Poly F, Pommier T, Cle´ment J-C, Lavorel S. 2014. Contribution of Grigulis K, Lavorel S, Krainer U, Legay N, Baxendale C, Dumont above- and below-ground plant traits to the structure and M, Kastl E, Arnoldi C, Bardgett RD, Poly F, Pommier T, function of grassland soil microbial communities. Ann Bot Schloter M, Tappeiner U, Bahn M, Cle´ment J-C, Hutchings M. 114:1011–21. 2013. Relative contributions of plant traits and soil microbial properties to mountain grassland ecosystem services. J Ecol Lepsˇ J, Osbornova´-Kosinova´ J, Rejma´nek M. 1982. Community 101:47–57. stability, complexity and species life history strategies. Vege- tatio 50:53–63. Grime JP, Brown VK, Thompson K, Masters GJ, Hillier SH, Clarke IP, Askew AP, Corker D, Kielty JP. 2000. The response MacDonald D, Crabtree J, Wiesinger G, Dax T, Stamou N, Fleury of two contrasting limestone grasslands to simulated climate P, Gutierrez Lazpita J, Gibon A. 2000. Agricultural abandon- change. Science 289:762–5. ment in mountain areas of Europe. Environmental conse- quences and policy response. J Environ Manag 59:47–69. Hasibeder R, Fuchslueger L, Richter A, Bahn M. 2015. Summer drought alters carbon allocation to roots and root respiration MacGillivray CW, Grime JP. 1995. Testing predictions of the in mountain grassland. New Phytol 205:1117–27. resistance and resilience of vegetation subjected to extreme events. Funct Ecol 9:640. Hodgson D, McDonald JL, Hosken DJ. 2015. What do you mean, ‘resilient’? Trends Ecol Evol 30:503–6. Meyer S, Leifeld J, Bahn M, Fuhrer J. 2012a. Free and protected soil organic carbon dynamics respond differently to aban- Holling CS. 1996. Engineering resilience versus ecological resi- donment of mountain grassland. Biogeosciences 9:853–65. lience. In: Schulze P, Ed. Engineering within ecological con- straints. Washington, D.C.: National Academy Press. p 31–44. Meyer S, Leifeld J, Bahn M, Fuhrer J. 2012b. Land-use change in subalpine grassland soils: effect on particulate organic car- Hoover DL, Knapp AK, Smith MD. 2014. Resistance and resi- bon fractions and aggregation. J Plant Nutr Soil Sci 175:401–9. lience of a grassland ecosystem to climate extremes. Ecology 95:2646–56. Nimmo D, Mac Nally R, Cunningham S, Haslem A, Bennett A. 2015. Vive la re´sistance: reviving resistance for 21st century Hoover DL, Rogers BM. 2016. Not all droughts are created equal: conservation. Trends Ecol Evol 30:516–23. the impacts of interannual drought pattern and magnitude on grassland carbon cycling. Glob Change Biol 22:1809–20. Obojes N, Bahn M, Tasser E, Walde J, Inauen N, Hiltbrunner E, Saccone P, Lochet J, Cle´ment J-C, Lavorel S, Tappeiner U, Jentsch A, Kreyling J, Boettcher-Treschkow J, Beierkuhnlein C. Ko¨ rner C. 2015. Vegetation effects on the water balance of 2009. Beyond gradual warming. Extreme weather events alter mountain grasslands depend on climatic conditions. Ecohy- flower phenology of European grassland and heath species. drology 8:552–69. Glob Change Biol 15:837–49. Oliver TH, Heard MS, Isaac NJ, Roy DB, Procter D, Eigenbrod F, Jentsch A, Kreyling J, Elmer M, Gellesch E, Glaser B, Grant K, Freckleton R, Hector A, Orme CDL, Petchey OL, Proenc¸a V, Hein R, Lara M, Mirzae H, Nadler SE, Nagy L, Otieno D, Raffaelli D, Suttle KB, Mace GM, Martı´n-Lo´ pez B, Woodcock Pritsch K, Rascher U, Scha¨dler M, Schloter M, Singh BK, BA, Bullock JM. 2015. Biodiversity and resilience of ecosys- Stadler J, Walter J, Wellstein C, Wo¨ llecke J, Beierkuhnlein C. tem functions. Trends Ecol Evol 30:673–84. 2011. Climate extremes initiate ecosystem-regulating func- tions while maintaining productivity. J Ecol 99:689–702. Osone Y, Ishida A, Tateno M. 2008. Correlation between relative growth rate and specific leaf area requires associations of Kaufman LH. 1982. Stream aufwuchs accumulation. Distur- specific leaf area with nitrogen absorption rate of roots. New bance frequency and stress resistance and resilience. Oecolo- Phytol 179:417–27. gia 52:57–63. Pirk N, Mastepanov M, Parmentier F-JW, Lund M, Crill P, Klumpp K, Tallec T, Guix N, Soussana J-F. 2011. Long-term Christensen TR. 2016. Calculations of automatic chamber flux impacts of agricultural practices and climatic variability on measurements of methane and carbon dioxide using short carbon storage in a permanent pasture. Glob Change Biol time series of concentrations. Biogeosciences 13:903–12. 17:3534–45. Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 703 Potts D, Huxman TE, Enquist BJ, Weltzin JF, Williams DG. 2006. Teuling AJ, Seneviratne SI, Sto¨ ckli R, Reichstein M, Moors E, Resilience and resistance of ecosystem functional response to Ciais P, Luyssaert S, van den Hurk B, Ammann C, Bernhofer a precipitation pulse in a semi-arid grassland. J Ecol 94:23–30. C, Dellwik E, Gianelle D, Gielen B, Gru¨ nwald T, Klumpp K, Montagnani L, Moureaux C, Sottocornola M, Wohlfahrt G. ´ ´ Quetier F, Thebault A, Lavorel S. 2007. Plant traits in a state and 2010. Contrasting response of European forest and grassland transition framework as markers of ecosystem response to energy exchange to heatwaves. Nat Geosci 3:722–7. land-use change. Ecol Monogr 77:33–52. Todman LC, Fraser FC, Corstanje R, Deeks LK, Harris JA, Pawlett R Development Core Team. 2015. A language and environment M, Ritz K, Whitmore AP. 2016. Defining and quantifying the for statistical computing. Vienna: R Foundation for Statistical resilience of responses to disturbance: a conceptual and Computing. modelling approach from soil science. Sci Rep 6:28426. Reich PB. 2014. The world-wide ‘fast-slow’ plant economics Vogel A, Scherer-Lorenzen M, Weigelt A. 2012. Grassland spectrum. A traits manifesto. J Ecol 102:275–301. resistance and resilience after drought depends on manage- Reichstein M, Bahn M, Ciais P, Frank D, Mahecha MD, ment intensity and species richness. PLoS One 7:e36992. Seneviratne SI, Zscheischler J, Beer C, Buchmann N, Frank Werner RA, Brand WA. 2001. Referencing strategies and tech- DC, Papale D, Rammig A, Smith P, Thonicke K, van der Velde niques in stable isotope ratio analysis. Rapid Commun Mass M, Vicca S, Walz A, Wattenbach M. 2013. Climate extremes Spectrom RCM 15:501–19. and the carbon cycle. Nature 500:287–95. Wheeler B. 2010. lmPerm: Permutation tests for linear models. Robson T, Lavorel S, Cle´ment J-C, Roux X. 2007. Neglect of R package version 1.1-2. Available at: http://cran.r-project. mowing and manuring leads to slower nitrogen cycling in org/web/packages/lmPerm/. subalpine grasslands. Soil Biol Biochem 39:930–41. Wolf S, Eugster W, Ammann C, Ha¨ni M, Zielis S, Hiller R, Stieger Roy J, Picon-Cochard C, Augusti A, Benot M-L, Thiery L, Dar- J, Imer D, Merbold L, Buchmann N. 2013. Contrasting re- sonville O, Landais D, Piel C, Defossez M, Devidal S, Escape C, sponse of grassland versus forest carbon and water. Environ Ravel O, Fromin N, Volaire F, Milcu A, Bahn M, Soussana J-F. Res Lett 9:89501. 2016. Elevated CO2 maintains grassland net carbon uptake under a future heat and drought extreme. Proc Natl Acad Sci Wright IJ, Reich PB, Westoby M, Ackerly DD, Baruch Z, Bongers USA 113:6224–9. F, Cavender-Bares J, Chapin T, Cornelissen JHC, Diemer M, Flexas J, Garnier E, Groom PK, Gulias J, Hikosaka K, Lamont Ruimy A, Jarvis PG, Baldocchi D, Saugier B. 1995. CO2 fluxes BB, Lee T, Lee WG, Lusk C, Midgley JJ, Navas M-L, Niinemets over plant canopies and solar radiation: a review. Adv Ecol U, Oleksyn J, Osada N, Poorter H, Poot P, Prior L, Pyankov VI, Res 26:1–68. Roumet C, Thomas SC, Tjoelker MG, Veneklaas EJ, Villar R. Schmitt M, Bahn M, Wohlfahrt G, Tappeiner U, Cernusca A. 2004. The worldwide leaf economics spectrum. Nature 2010. Land use affects the net ecosystem CO exchange and its 428:821–7. components in mountain grasslands. Biogeosciences 7:2297– Yeung AC, Richardson JS. 2016. Some conceptual and opera- tional considerations when measuring ‘resilience’. A response Schrama M, Bardgett RD. 2016. Grassland invasibility varies to Hodgson et al. Trends Ecol Evol 31:2–3. with drought effects on soil functioning. J Ecol 104:1250–8. Zeller V, Bahn M, Aichner M, Tappeiner U. 2000. Impact of Smith P. 2014. Do grasslands act as a perpetual sink for carbon? land-use change on nitrogen mineralization in subalpine Glob Change Biol 20:2708–11. grasslands in the Southern Alps. Biol Fertil Soils 31:441–8. Tasser E, Tappeiner U. 2002. Impact of land use changes on Zwicke M, Alessio GA, Thiery L, Falcimagne R, Baumont R, mountain vegetation. Appl Veg Sci 5:173–84. Rossignol N, Soussana J-F, Picon-Cochard C. 2013. Lasting Tasser E, Walde J, Tappeiner U, Teutsch A, Noggler W. 2007. effects of climate disturbance on perennial grassland above- Land-use changes and natural reforestation in the Eastern ground biomass production under two cutting frequencies. Central Alps. Agric Ecosyst Environ 118:115–29. Glob Change Biol 19:3435–48. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Ecosystems Springer Journals

Land Use Alters the Drought Responses of Productivity and CO2 Fluxes in Mountain Grassland

Free
15 pages
Loading next page...
 
/lp/springer_journal/land-use-alters-the-drought-responses-of-productivity-and-co2-fluxes-10cCphMoG6
Publisher
Springer US
Copyright
Copyright © 2017 by The Author(s)
Subject
Life Sciences; Ecology; Plant Sciences; Zoology; Environmental Management; Geoecology/Natural Processes; Hydrology/Water Resources
ISSN
1432-9840
eISSN
1435-0629
D.O.I.
10.1007/s10021-017-0178-0
Publisher site
See Article on Publisher Site

Abstract

Ecosystems (2018) 21: 689–703 DOI: 10.1007/s10021-017-0178-0 2017 The Author(s). This article is an open access publication Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 1 2 3 Johannes Ingrisch, Stefan Karlowsky, Alba Anadon-Rosell, Roland 1 1 4 2 Hasibeder, Alexander Ko¨nig, Angela Augusti, Gerd Gleixner, and Michael Bahn * 1 2 Institute of Ecology, University of Innsbruck, Sternwartestraße 15, 6020 Innsbruck, Austria; Max Planck Institute for Biogeo- chemistry Jena, Postbox 100164, 07701 Jena, Germany; Department of Evolutionary Biology, Ecology and Environmental Sciences, University of Barcelona, Avinguda Diagonal 643, 08028 Barcelona, Spain; Institute of Agro-Environmental and Forest Biology, CNR Italy, Via G. Marconi n.2, 05010 Porano, TR, Italy ABSTRACT Climate extremes and land-use changes can have low resistance was related to high recovery from major impacts on the carbon cycle of ecosystems. drought and vice versa. In consequence, the overall Their combined effects have rarely been tested. We perturbation of the carbon cycle caused by drought studied whether and how the abandonment of tra- was larger in the managed than the abandoned ditionally managed mountain grassland changes the grassland. The faster recovery of carbon dynamics resilience of carbon dynamics to drought. In an in situ from drought in the managed grassland was associ- common garden experiment located in a subalpine ated with a significantly higher uptake of nitrogen meadow in the Austrian Central Alps, we exposed from soil. Furthermore, in both grasslands leaf intact ecosystem monoliths from a managed and an nitrogen concentrations were enhanced after abandoned mountain grassland to an experimental drought and likely reflected drought-induced in- early-summer drought and measured the responses creases in nitrogen availability. Our study shows that of gross primary productivity, ecosystem respiration, ongoing and future land-use changes have the phytomass and its components, and of leaf area index potential to profoundly alter the impacts of climate during the drought and the subsequent recovery extremes on grassland carbon dynamics. period. Across all these parameters, the managed grassland was more strongly affected by drought and Key words: Carbon cycle; Climate extreme; Gross recovered faster than the abandoned grassland. A primary productivity; Land-use change; Nitrogen; bivariate representation of resilience confirmed an Recovery; Resilience; Resistance; N labelling. inverse relationship of resistance and recovery; thus, Received 6 March 2017; accepted 14 August 2017; published online 15 September 2017 INTRODUCTION Electronic supplementary material: The online version of this article The frequency and severity of extreme climatic (doi:10.1007/s10021-017-0178-0) contains supplementary material, events are expected to increase in the near future, which is available to authorized users. with major implications for the carbon (C) cycle of Authors contributions MB and GG conceived the study, JI, SK, AAR, RH, AK and AA performed the experiment and collected data, JI and SK ecosystems and related feedbacks to the atmo- analysed the data, JI and MB led the writing of the manuscript. All au- sphere and the climate system (Reichstein and thors commented on the drafts and gave final approval for publication. others 2013; Frank and others 2015). On a global *Corresponding author; e-mail: Michael.Bahn@uibk.ac.at 689 690 J. Ingrisch and others scale, severe droughts are amongst the climate ex- Khan and others 2016; Roy and others 2016). In- tremes exerting the strongest effects on the C cycle creased leaf N concentrations can in turn promote (Ciais and others 2005; Reichstein and others 2013; photosynthetic C assimilation and thereby speed up Knapp and others 2015). The overall resilience of ecosystem recovery from drought (Roy and others an ecosystem to climate extremes can be charac- 2016). To date it is not known whether abandon- terized by the resistance, that is, the ability of an ment of mountain grasslands, associated with a ecosystem to persist during disturbance, and the reduction in N availability (Zeller and others 2000; recovery, that is, the ability of a system to return to Robson and others 2007), alters the role of plant N pre-disturbance levels (Holling 1996; Hodgson and uptake during post-drought recovery, and what the others 2015; Oliver and others 2015). Thus, a sys- consequences are for tissue N concentrations and tem may be resilient due to a high resistance, a CO uptake dynamics. high capacity to recover or both. Although the We established an experiment testing whether relationships and potential trade-offs between the and how abandonment of managed mountain components of resilience have been subject to re- grassland changes the resilience of C dynamics to cent discussions (for example, Hodgson and others an extreme early-summer drought. We analysed 2015; Yeung and Richardson 2016), few studies both drought resistance and post-drought recovery have actually tested such relationships on the C of ecosystem CO fluxes and of the phytomass and cycle responses of ecosystems to drought. its components. To understand the relationships Grasslands cover more than one-fifth of the and potential trade-offs between the components 6 2 global land surface (35 9 10 km , Dixon and of resilience in response to land-use change, we others 2014) and constitute an important carbon applied a recently proposed bivariate representa- sink (Conant and others 2001; Smith 2014). In tion of resilience (Nimmo and others 2015) and many mountain regions of Europe, grasslands also tested if the conclusions were robust across the play an important role in the production of fodder. different parameters studied. We furthermore ap- In recent decades, land-use changes have led to the plied a perturbance-based approach (Potts and abandonment of mountain meadows and pastures others 2006; Todman and others 2016) to obtain an (for example, MacDonald and others 2000; Tasser integrated quantification of the overall perturba- and others 2007), with consequences for species tion of the two grasslands by the drought event. composition (Tasser and Tappeiner 2002), produc- Our main hypothesis was that abandonment in- tivity and ecosystem C fluxes (Schmitt and others creases the resistance of C dynamics to and de- 2010), soil C and nitrogen (N) turnover (Zeller and creases their recovery from drought. We others 2000; Robson and others 2007; Meyer and furthermore tested the hypothesis that drought others 2012a; Grigulis and others 2013) and the enhances N uptake and tissue N concentrations in water cycle (Obojes and others 2015). To date, managed grassland, whereas the role of plant N surprisingly few studies have explored how land- uptake during recovery is strongly reduced in use changes affect ecosystem responses to climate abandoned grassland. extremes (Bahn and others 2014). Although man- agement intensity has been suggested to modify MATERIAL AND METHODS grassland responses of productivity and CO fluxes Study Site to precipitation variability and drought (Klumpp and others 2011; Vogel and others 2012; Zwicke The study site is located near Neustift in the Stubai and others 2013), the consequences of an aban- valley in the Austrian Central Alps and is composed donment of managed grasslands for the drought of grasslands differing in land use, including a tra- and post-drought responses of their C dynamics are ditionally managed hay meadow and an aban- largely unknown. doned grassland (Schmitt and others 2010). The Site fertility has been shown to modify grassland two subalpine grasslands are located on a southeast responses to climate change (Grime and others exposed hillside with similar inclination (ca. 20), 2000); however, the role of nutrient availability for average annual temperature (3C), annual precip- ecosystem recovery from drought is still poorly itation (1097 mm) and the same soil type (dystric understood. Recent studies suggest that soil nitro- cambisol). The soil textural fractions for clay, silt gen (N) dynamics can be altered by drought events and sand are 13.3, 36.2 and 50.2%, respectively, on and that rewetting of soil after drought can en- the meadow, and 23.4, 45.5 and 31.1%, respec- hance N mineralization and consequently lead to tively, on the abandoned grassland (Meyer and higher tissue N concentrations (Fuchslueger and others 2012b). The meadow (1820–1850 m a.s.l.; others 2014; Canarini and Dijkstra 2015; Arfin Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 691 4707¢45¢¢N, 1118¢20¢¢E) is cut once per year at grassland, respectively. On 28 June 2014, rain-out peak biomass in early August and is fertilized with shelters were removed and 50 mm of rainwater manure every 2–3 years. Additionally, light grazing was added to each of the monoliths to simulate a by sheep and cattle takes place in spring and late heavy rain event ending the drought, and to autumn. The vegetation community has been achieve a well-defined and rapid rewetting. classified as Trisetetum flavescentis and consists of perennial grasses and forbs dominated by Agrostis Measurements capillaris, Festuca rubra, Ranunculus montanus, Tri- Microclimate folium pratense and T. repens (Bahn and others 2009). The second grassland (1970–2000 m a.s.l.; A microclimate station at the common garden re- 4707¢31¢¢N, 1117¢24¢¢E) has been abandoned corded photosynthetically active radiation (PAR), since 1983. Its vegetation has been classified as precipitation, air temperature and humidity (see Seslerio-Caricetum with some dwarf shrubs and is details in Hasibeder and others 2015). During the rain exclusion, air temperature, humidity and PAR dominated by Sesleria varia, Erica carnea, Carex sem- (S-THB-M002 and S-LIA-M003, onset Computer pervirens and Poa alpina (Schmitt and others 2010; Corporation, Bourne, MA, USA) were additionally Grigulis and others 2013). Further details con- measured in two of the six rain-out shelters. Soil cerning vegetation and soils, as well as the overall water content (Decagon EC-5, 5TM, 5TE; combining nutrient supply and productivity of the two sites, SWC and temperature, logger Em50; Decagon De- can be taken from Bahn and others (2006), Schmitt and others (2010), Meyer and others (2012a), vices, Pullman, WA, USA) and soil temperature Grigulis and others (2013), Fuchslueger and others (sensors S-TMB, logger HOBO Micro Station H21- (2014) and Legay and others (2014). 002; Onset Computer corporation, Bourne, MA, USA) were measured continuously in the main Experimental Set-up rooting horizon (30-min interval) in a subset of the monoliths (SWC: n = 17, Temp: n =14) over the The drought experiment was conducted in a com- whole course of the season. In early May, before the mon garden established at the meadow site (see start of the experiments, all soil moisture sensors were above). At both the meadow and abandoned offset-calibrated in situ after a rainy period, when all grassland, 20 intact vegetation–soil monoliths were monoliths had reached field capacity. To determine extracted in June 2013. The monoliths had a the water balance, we measured the amount of lea- diameter of 25 cm and a height of 28 cm and were chates accumulated in the reservoir of each monolith fit in open-top round stainless steel cylinders over the period of rain exclusion (Obojes and others (height 40 cm), with a reservoir for leachates at the 2015). Total evapotranspiration during the drought bottom (for detailed description see Obojes and experiment was estimated for the subset of monoliths others 2015). The monoliths were installed in the equipped with soil moisture sensors by means of a common garden in a randomized factorial design water balance approach, accounting for the amount with six blocks and were left for almost a year be- of water added, the change of water storage in each fore the drought experiment started in May 2014. monolith as derived from monitored changes in vol- For the drought experiment, each of the six umetric soil water content in the main rooting hori- blocks was covered with a rain-out shelter, which zon, and the amount of leached water (Table S1; see had a base area of 3 9 3.5 m and was open at the also Obojes and others 2015). bottom (up to 0.5 m above ground) and at the top of the face sides to allow air circulation. Rain exclusion was performed with transparent UV-A Phytomass and UV-B transmissive plastic foil (Lumisol clear Aboveground phytomass was sampled destruc- AF, Folitec, Westerburg, Germany, light transmit- tively during three campaigns, at peak drought tance c. 90%). The rain exclusion lasted from 21 (‘‘resistance campaign’’ on 1 July 2014, n = 12, May 2014 until 28 June 2014. During this period, that is, 3 replicates per land use and treatment monoliths allocated to the drought treatment did combination) and twice during the recovery peri- not receive any precipitation. Control monoliths od, that is, 4 (‘‘recovery 1 campaign’’ on 24 July, were manually watered every 1–4 days with pre- n = 12) and 8 weeks (‘‘recovery 2 campaign’’ on 22 viously collected rainwater to maintain soil water August, n = 16) after termination of the experi- content above 25 vol.% (Figure 1D, E). The mental drought. Thus, at each campaign, a separate amount of water added to the controls was 180 and subset of monoliths was harvested by cutting 170 mm for the meadow and the abandoned 692 J. Ingrisch and others Figure 1. Time course of A daily sums of photosynthetically active radiation (PAR), daily means of B air temperature (T ) and C air vapour pressure deficit (vpd) and in the rain-out shelters. C, D Daily means of soil moisture (vol.%) in the air main rooting horizon in monoliths from the meadow (control n = 3, drought n = 5) and the abandoned grassland (control n = 4, drought n = 5) exposed to ambient conditions (control, solid line) and drought (dashed line). Shaded areas show the standard error of the mean. Vertical bars show daily precipitation (open = natural, shaded = manual watering). Note that during rain exclusion (horizontal black bar) only monoliths from the control treatment received water. phytomass to 2 cm aboveground. The samples were (3 cm diameter), washed, sieved to 2 mm and mi- frozen at -18C until further analysis. crowaved before transporting to the laboratory. Phytomass samples were split into four func- All plant samples were oven-dried at 60C, tional groups (forbs, grasses, legumes, dwarf ground, weighed (2–5 mg and analysed on an shrubs), and into stems, leaves, reproductive or- elemental analysis—isotope ratio mass spectrome- gans and living phytomass (hereafter biomass) and ter (EA-IRMS; EA 1100, CE Elantech, Milan, Italy; necromass (senesced plant parts) were separated. coupled to a Delta + IRMS, Finnigan MAT, Bre- For each functional group in each monolith, men, Germany). LNC was calculated based on the specific leaf area (SLA) was obtained for a subset of peak area and the known nitrogen concentration of leaves saturated with water and scanned (V700 external acetanilide standards. The d N was Photo, Epson, WinRHIZO Pro 2012, Regent determined in per mil (%) relative to the interna- Instruments) and subsequently dried at 60C for tional reference standard AIR-N using IAEA-N1 3 days. Leaf area index (LAI) was calculated from (Werner and Brand 2001). The amount of N label the leaf biomass and SLA for each functional group. recovered in roots and shoots is calculated as: Community-weighted mean (CWM) of SLA was ðÞ atom%  atom% N labelled NA pool calculated as the leaf-biomass-weighted mean of incorporated N ¼ 100% SLA of each functional group. with atom% being the N atom% of the la- labelled 15 15 Tissue Nitrogen and N Labelling belled samples, atom% being the N atom% of NA natural abundance samples and N being the pool For each sampling campaign (see above), we -2 respective nitrogen pool (mg N m ). measured the leaf nitrogen concentration (LNC) and its nitrogen isotope ratio (d N) on a subset of CO Fluxes leaves sampled in each of the monoliths. Further- We measured the net ecosystem exchange (NEE) of more, at the end of the drought experiment, we CO using closed dynamic chambers, similar to the performed a N pulse labelling experiment on the system applied by Schmitt and others (2010). The 12 monoliths sampled during the Recovery 1 15 15 chambers were transparent Plexiglas cylinders (di- campaign. 20 mg of KNO with 10% N(2 mg N ameter 25 cm, height 50 cm) which fitted airtight per monolith) dissolved in 100 ml rain water was on the steel cylinders containing the monoliths. distributed equally on the soil of the monoliths. Pressure effects on CO fluxes were avoided by a During the subsequent harvest, both the shoots and 2 hole in the top of the chamber, which was closed the roots from the uppermost 7 cm of the soil were with a plug after placing the chamber. Air inside sampled. Roots were extracted with a soil corer Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 693 the chambers was ventilated with fans. Concen- ment points is a measure of change within this trations of CO (GMP 343, Vaisala Helsinki, Fin- bivariate space, and the cumulative length of each land) and water vapour, as well as temperature trajectory is a measure of the overall perturbation (HMP 75, Vaisala, Helsinki Finland) were logged for (Potts and others 2006). 1 min with 5-s intervals. During each measure- The resistance of GPP , ER, leaf area index (LAI) sat ment, the photosynthetically active radiation and biomass was determined based on the mea- (PQS1 PAR Quantum Sensor, Kipp & Zonen, Delft, surements during peak drought. We express resis- the Netherlands) was recorded. Ecosystem respi- tance (RST) as the ratio of drought to control ration (ER) was measured by covering the chamber measurements (Kaufman 1982). A recovery index with a dark cloth, excluding any light inside the according to Nimmo and others (2015) was calcu- chamber. To obtain estimates of gross primary lated for GPP , ER, LAI and biomass. This index is sat productivity (GPP), paired measurements of NEE a measure of the post-drought change of the under sunlit and dark conditions were taken. parameter. We adapted the approach by using Monoliths were measured in randomized order in measurements of control monoliths instead of the morning hours on days with clear sky. In pretreatment measurements to account for sea- addition to sunlit and dark measurements, a series sonal changes in the controls. The recovery index of light response curves were obtained for each was calculated as D /C – RST, where D and rec rec rec treatment type, using layers of semitransparent C denote parameter values during recovery in rec cloth (Schmitt and others 2010). To obtain a con- the drought and control treatment, respectively, sistent time series of CO fluxes throughout the and RST is the resistance of the parameter (see whole study, flux measurements were taken on the above). monoliths which were harvested during the last All calculations and statistical analyses were campaign (Recovery 2). performed in R 3.2.3 (R Development Core Team CO flux rates were calculated as described by 2015). We used permutational ANOVA with the Schmitt and others (2010). Each measurement was package lmPerm (Wheeler 2010). For each sam- quality controlled based on visual inspection and pling day, we tested for the interaction of land-use quality of the linear fits as recently recommended type and drought treatment and drought effects by Pirk and others (2016). GPP was calculated as within each land-use type. the difference of the corresponding NEE and ER measurements. Throughout this study, GPP and ER RESULTS fluxes are both assigned positive signs. For each Key meteorological variables during the experi- land-use type and precipitation treatment, light mental period are presented in Figure 1A–C. Dur- response curves were obtained from pooled data by ing the rain exclusion, soil moisture declined to less fitting a rectangular hyperbolic model (Ruimy and than 20 vol.% in the monoliths from both grass- others 1995; Schmitt and others 2010). Above a -2 -1 lands which were exposed to drought (Figure 1 D, photon flux density (PFD) of 1000 lmol m s , E). The drought treatment reduced the total all light response curves levelled off and reached amount of evapotranspiration significantly 80–85% of the maximum values (Figure S1). Thus, (p < 0.001) in both grasslands (Table S1). Neither for the sake of comparability across treatments we land use nor the interaction of drought and land only present data obtained at PFD above this use had a significant effect on the evapotranspira- threshold and apply the term light-saturated GPP (GPP ). For our analysis, we only included fluxes tion. sat from measurement days for which at least three quality-controlled replicate data sets per land use Aboveground Plant Productivity and and treatment combination were available. Nitrogen Relations The meadow had a generally higher biomass and Calculations of Indices and Statistics leaf area index (LAI) compared to the abandoned To obtain normalized fluxes, values of GPP and grassland, but these differences declined in the sat ER in the drought treatment were divided by their course of the season (Figure 2A, B, G and H, Fig- respective values in the controls. The daily means ure S3), reflecting a delayed development and of the normalized fluxes define the response tra- lower aboveground net primary production of the jectory of each grassland in the bivariate space of abandoned grassland and an earlier plant senes- normalized GPP and normalized ER. The cence of the meadow plants (Figure 2C, D). During sat Euclidian distance between consecutive measure- the period of peak growth in early July, leaf area 694 J. Ingrisch and others CD EF G H Figure 2. A–F Biomass, necromass, phytomass (sum of biomass and necromass) and G, H leaf area index (LAI) of monoliths from the meadow and the abandoned grassland subjected to ambient conditions (open symbols) and drought (closed symbols) and sampled at the end of drought (1 July) and during recovery (24 July, 22 August). Error bars indicate standard errors of the mean (n = 3 for July and n = 4 for August samplings), and stars indicate significant treatment effects within land use and sampling date (p value: *** < 0.001 < ** < 0.01 < * < 0.05 < (*) < 0.1). 2 -1 index (LAI) was higher on the meadow compared significantly (meadow: 10.1 ± 1.3 m kg , aban- 2 -1 to the abandoned grassland. The fraction of grasses doned grassland: 7.3 ± 0.8 m kg ). was significantly higher in the abandoned grass- Drought reduced biomass significantly in the lands (78 ± 3%) compared to the meadow meadow (Figure 2A), but not in the abandoned (55 ± 5%, p < 0.001). The community-weighted grassland (Figure 2B, Table 1; Figure S2 and mean (CWM) of SLA for the meadow and the Tables S2 and S4). The drought-induced reduction abandoned grassland was 14.4 ± 0.8 and was persistent for 4 weeks, but disappeared later 2 -1 9.3 ± 0.6 m kg , respectively (p < 0.001). Forbs when biomass in the control monoliths declined had a significantly higher mean SLA in the mea- (Figure 2A). Drought did not immediately induce 2 -1 dow (20.3 ± 1.5 m kg ) compared to the aban- leaf senescence at either site, but significantly re- 2 -1 doned grassland (15.2 ± 1.5 m kg , p < 0.001), duced necromass in the abandoned grassland in the whereas the mean SLA of grasses did not differ late season (Table 1, Figure 2C, D). LAI was re- Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 695 Table 1. Drought and Land-Use Effects on Phytomass Parameters Campaign Weeks after rewetting Treatment Land use Treatment * land use Biomass Resistance 0 *** n.s. * Recovery 1 3.5 *** n.s. n.s. Recovery 2 8 n.s. n.s. n.s. Necromass Resistance 0 * n.s. n.s. Recovery 1 3.5 n.s. n.s. * Recovery 2 8 n.s. n.s. n.s. Phytomass Resistance 0 *** n.s. * Recovery 1 3.5 *** n.s. n.s. Recovery 2 8 n.s. n.s. n.s. LAI Resistance 0 *** *** ** Recovery 1 3.5 * (*) n.s. Recovery 2 8 * *** n.s. Biomass of forbs Resistance 0 n.s. ** * Recovery 1 3.5 ** ** * Recovery 2 8 * ** * Biomass of grasses Resistance 0 *** n.s. n.s. Recovery 1 3.5 n.s. n.s. n.s. Recovery 2 8 n.s. (*) n.s. LNC Resistance 0 n.s. n.s. Recovery 1 3.5 *** (*) n.s. Recovery 2 8 * n.s. (*) d15N Resistance 0 n.s. *** n.s. NA Recovery 2 8 n.s. *** n.s. Results of permutational ANOVA testing overall treatment effect, land-use effect and their interaction on different phytomass parameters. Biomass = living phytomass; Necromass = senescent phytomass; LAI = leaf area index; LNC = leaf nitrogen concentration; d15N = natural abundance nitrogen isotope ratio of leaves. Resistance = 1 NA July, peak drought, Recovery 1 = 24 July, Recovery 2 = 22 August. Stars indicate the significance level: *** < 0.001 < ** < 0.01 < * < 0.05 < (*) < 0.1 duced by drought in the meadow (Table 1, Fig- uptake of N during the recovery from drought ure 2G, H), but exceeded the values in the controls (Figure 5). at the Recovery 2 sampling. This was also reflected by leaf mass dynamics (Figure S2). Biomass of forbs CO Fluxes responded differently between the two grasslands During drought, GPP was progressively reduced sat (Table 1, Figure 3A, B, Figure S3) and was less to 20 and 40% of the controls in the meadow and resistant but recovered more rapidly in the mea- the abandoned grassland, respectively (Figure 6A, dow. In contrast, biomass of grasses was signifi- B); the interaction of land use and drought was cantly reduced by drought and recovered quickly in significant (Table 2). At peak drought, ER was re- both grasslands (Table 2, Figure 3C, D, Figure S3). duced by up to 60 and 25% on the meadow and During the recovery period, LNC was signifi- the abandoned site, respectively (Figure 6C, D); cantly higher in monoliths previously exposed to however, the interaction of land use and drought drought compared to the controls. The effect was was not significant. After rewetting, GPP fully sat observed in both grasslands and was more pro- recovered within 9 days in both grasslands (Ta- nounced in the meadow (Table 1, Figure 4A, B). ble 2). Its recovery rate was distinctly higher in the Leaf d N was generally higher in the meadow than meadow compared to the abandoned grassland in the abandoned grassland (Table 1, Figure 4C, D) -2 -1 (23.4 versus 14.6 lmol m s over the period of and was not affected by drought. 9 days). During the recovery, ER was significantly Following N pulse labelling meadow plants enhanced in previously drought-exposed mono- recovering from drought took up significantly more liths relative to controls in both grasslands, the ef- N label than the controls (p < 0.01) and incor- fect being more pronounced and sustained in the porated this nitrogen into shoots (+110%, meadow than in the abandoned grassland (Fig- p < 0.01), but not into roots (Figure 5). In con- ure 6C, D; Table 2). trast, the abandoned grassland did not increase its 696 J. Ingrisch and others A B Figure 3. Biomass of A, B forbs and C, D grasses of monoliths from the meadow and the abandoned grassland under ambient conditions (open symbols) and drought conditions (closed symbols), sampled at the end of the drought treatment (1 July) and during recovery (24 July, 22 August). Error bars indicate standard errors of the mean (n = 3 in July and n =4 in August), asterisks indicate significant treatment effects within land use and sampling date (p value: *** < 0.001< ** < 0.01). Effects of Land-Use Change on Integrated DISCUSSION Drought Responses Does Land Use Alter Drought Resistance To integrate the drought responses across the two and Recovery of Productivity and CO grasslands, the resistance and the recovery of Fluxes? studied parameters were related in a bivariate ap- The conceptualization of resilience in ecology has proach (see methods). Drought resistance generally led to contrasting definitions and terminologies, increased from GPP to ER to LAI to biomass and sat which have recently been under some debate (e.g. was generally lower in the meadow than in the Hodgson and others 2015; Yeung and Richardson abandoned grassland (Figure 7). In contrast, across 2016). For analysing the disturbance responses of a all parameters, the recovery index was higher in system within a given stability domain (‘‘engi- the meadow than in the abandoned grassland, neering resilience’’, sensu Holling 1996), it has yielding an overall negative relationship between been suggested to distinguish resistance and resistance and recovery. recovery as the two major underlying processes of The effect of land-use change on the pertur- resilience (Hodgson and others 2015; Nimmo and bation of the two CO flux components GPP 2 sat others 2015; Oliver and others 2015). In our study, and ER was assessed in more detail by comparing we observed that both grasslands were highly re- their drought response trajectories for the mea- silient to drought, confirming conclusions from dow and the abandoned grassland. While the earlier studies (Gilgen and Buchmann 2009; trajectories had similar shapes for both grass- Jentsch and others 2011; Hoover and others 2014), lands (Figure 8A), the cumulative length of the though it should be noted that the timing, the response trajectory for the meadow was 33% magnitude and the interannual pattern of drought larger than for the abandoned grassland (Fig- may modify specific grassland drought responses ure 8B). (Knapp and others 2015; Estiarte and others 2016; Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 697 meadow abandoned ** *** ** 2.4 2.0 1.6 1.2 CD ∂d∂ -1 -2 Jul 01 Jul 15 Aug 01 Aug 15 Jul 01 Jul 15 Aug 01 Aug 15 Control Drought Figure 4. A, B Leaf nitrogen concentrations (LNC) and C, D corresponding natural abundance d N isotope values of leaves in monoliths from the meadow and the abandoned grassland subjected to ambient conditions (open symbols) and drought (closed symbols) and sampled at peak drought (1 July) and during recovery (24 July, 22 August). Error bars indicate standard errors of the mean (n = 3 for July and n = 4 for August samplings), stars indicate significant treatment effects within land use and sampling date (p value: *** < 0.001 < ** < 0.01 < * < 0.05). Table 2. Drought and Land-Use Effects on CO Fluxes Date Days after rewetting Treatment Land use Treatment * land use Meadow Abandoned GPP Pretreatment n.s. (*) n.s. n.s. n.s. sat Resistance *** *** * *** *** Recovery 1 5 ** n.s. n.s. * *** Recovery 2 9 * (*) n.s. n.s. (*) Recovery 3 17 n.s. n.s. n.s. (*) n.s. ER Pretreatment n.s. n.s. n.s. n.s. n.s. Resistance ** (*) n.s. ** * Recovery 1 5 ** n.s. n.s. *** * Recovery 2 9 n.s. n.s. n.s. * n.s. Recovery 3 17 n.s. n.s. * n.s. * Recovery 4 30 n.s. n.s. n.s. n.s. n.s. Results of permutational ANOVA testing overall treatment effect, land-use effect and their interaction as well as within-land-use effects of drought on key dates of experiment. GPP = light-saturated rate of gross primary productivity, ER = ecosystem respiration. Dates: Pretreatment = 19 May, Resistance = 26 June, Recovery 1 = 3 July, Recovery sat 2 = 7 July, Recovery 3 = 15 July, Recovery 4 = 28 July. Stars indicate the significance level: *** < 0.001 < ** < 0.01 < * < 0.05 < (*) < 0.1 Hoover and Rogers 2016). Furthermore, we found studied C cycle parameters. This notion is con- that the role of resistance and recovery for resi- firmed when applying a bivariate approach (Fig- lience can be strongly altered by land-use change: ure 7), as recently suggested in the literature although the abandoned grassland had a distinctly (Hodgson and others 2015; Nimmo and others higher drought resistance, the managed meadow 2015): the abandoned grassland was resilient due displayed a higher recovery rate across all the to a high resistance, whereas the meadow was less LNC[%] ∂ N [‰] NA 698 J. Ingrisch and others Figure 5. Amount of N label recovered in the two grasslands in A shoots and B roots of control monoliths (open bars) and in monoliths recovering from drought (shaded bars) 3 weeks after the rewetting. Error bars indicate standard errors of the mean (n = 3). Stars B indicate significant differences between control and drought treatment (p value: ** < 0.01). Figure 6. A, B Light- saturated rates of gross AB primary productivity (GPP ) and C, D sat ecosystem respiration (ER) in monoliths from the meadow and the abandoned grassland subjected to ambient conditions (open symbols) and drought (closed C D symbols). Error bars indicate standard errors of the mean (n = 3–4). The horizontal black bars indicate the time of rain exclusion. resistant, but resilient due to a high recovery. In ment was smaller than the concurrent gain of consequence, the overall perturbation of the C resistance. cycle caused by drought was larger in the managed The consistent trade-off between resistance and than the abandoned grassland, as indicated by a recovery between the two studied grasslands was greater length of a multivariate response trajectory likely related to differences in plant community (Potts and others 2006, Figure 8). This suggests that composition and associated differences in the pre- the loss of ability to recover caused by abandon- vailing strategies of plant species to cope with Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 699 drought. Our results suggest that the different re- sponses of the grasslands were not predominantly driven by their relative composition of grasses versus forbs, as the drought and post-drought recovery response of grasses was similar in the meadow and the abandoned grassland. However, the drought response of forbs differed distinctly between the two grasslands: forb biomass was less affected by drought in the abandoned grassland than in the meadow (Figure 3), which contributed to the higher resistance of the abandoned grass- land. The forbs in the abandoned grassland were characterized by lower mean SLA compared to the meadow, a trend already previously observed both for SLA and LNC across gradients of decreasing land-use intensity (Bahn and others 1999; Grigulis Figure 7. Resistance and recovery of the investigated C and others 2013; but note that reduced grazing cycle parameters in the two grasslands (open = meadow, shaded = abandoned grassland). The resistance was cal- intensity can also favour species with higher SLA culated as the ratio of parameter performance in drought ´ and LNC, see Laliberte and others 2012). Species plots relative to the parameter performance in the control with lower SLA and LNC are associated with lower plots during peak drought. The recovery index is a growth rates (Lambers and Poorter 1992; Wright measure for the absolute recovery of the parameter after and others 2004) and are characterized by a higher the end of the drought. High values indicate high resis- tolerance to nutrient stress (Garnier and others tance and recovery, respectively. The arrow indicates the 2004; Que´tier and others 2007; Grigulis and others shift in resilience caused by abandonment. ER ecosystem 2013). These ‘‘conservative species’’ have been respiration, GPP light-saturated gross primary produc- suggested to be more resistant, but less capable of tivity, LAI leaf area index. recovering quickly from disturbance (Lambers and Poorter 1992; MacGillivray and Grime 1995; Reich 2014). Conversely, communities dominated by AB Figure 8. A The course of normalized light-saturated rates of gross primary productivity (GPP ) and ecosystem respi- sat ration (ER) before (grey points), during (open points) and after (black points) the drought experiment in the meadow (solid line) and the abandoned grassland (dotted line). Normalized fluxes were calculated as the ratio of the flux in drought monoliths to the respective flux in control monoliths. The direction of the path is given by the arrow, symbols denote the periods before (shaded), during (open) and after (closed) drought. B Cumulative Euclidian distance of the response tra- jectories of the two grasslands over the course of the drought. The Euclidian distance between two consecutive mea- surements days is a measure of the system’s change in the bivariate flux space. The cumulative Euclidian distance from beginning of the drought (pretreatment) is a measure of the overall perturbation of the grassland. The black horizontal bar indicates period of rain exclusion. 700 J. Ingrisch and others ‘‘exploitative species’’ (fast growth related to a recovery, and indicates that the effect was more higher SLA and LNC) have been shown to recover pronounced in the managed meadow (Figure 4A, better from climatic disturbances (Lepsˇ and others B). This is in line with recent observations that 1982; Grime and others 2000). In our study, all the resource pulses can be larger under intensive C cycle parameters studied were more susceptible compared to extensive management (Fuchslueger to drought but recovered more rapidly in the and others 2014; Schrama and Bardgett 2016) and managed meadow, which is more strongly domi- suggests a higher post-drought availability of N in nated by exploitative species. the meadow compared to the abandoned grassland. Vegetation phenology has been suggested to be Higher rewetting-induced resource pulses in the sensitive to climate extremes (for example, Jentsch meadow might also be reflected by a more pro- and others 2009). In our study, phenological nounced stimulation of CO release from soil dynamics during post-drought recovery likely re- (‘‘Birch effect’’, Figure 6C, D), which has fre- flected contrasting plant strategies of the two quently been associated with a rapid mineralization grasslands. The meadow built up new biomass and of organic matter (for example, Borken and Matz- increased leaf area more rapidly (Figure 2A, G), ner 2009; deVries and others 2012). reflecting its fast-growth strategy. In contrast, the Tissue N concentrations result from the uptake of lower necromass in the abandoned grassland at the N and its dilution by growth. Since during its last sampling date (Figure 2D) indicates a delayed recovery from drought, the meadow increased tis- leaf senescence of that plant community in re- sue N concentrations while producing more new sponse to the drought. Both grasslands can thereby biomass than the abandoned grassland (Figure 2), potentially maintain a higher C uptake later in the it must have taken up distinctly more N. This is season to compensate C deficits from drought confirmed by our NO labelling experiment, (Casper and others 2001). which suggests a doubling of nitrate uptake in It has been shown that with increasing time since monoliths from the meadow during recovery from abandonment the resistance of abandoned fields to drought, while no clear effect was observed for the drought (Lepsˇ and others 1982) and of shrublands abandoned grassland (Figure 5). The labelling to long-term warming and drought (Kro¨ el-Dulay experiment suggests that increased N uptake on the and others 2015) increases. As secondary succes- meadow was not only caused by more pronounced sion proceeds, an increasing dominance of woody resource pulses, but also by a higher root N uptake species (shrubs, trees) also favours plants with capacity, which is typically higher for exploitative, more conservative water use, which could buffer fast-growing species (Osone and others 2008; effects of dry periods on soil moisture (Teuling and Grassein and others 2015). Interestingly, on the others 2010; Wolf and others 2013; Gavazov and meadow the additional N taken up during recovery others 2014). As suggested by our results for the did not remain in the roots, but was used for the water balance (Table S1), such a water-sparing production of aboveground biomass. Since the strategy was not yet observed in the abandoned natural abundance d N of leaves was not affected grassland of our study, where herbaceous species by drought (Figure 4C, D), it can be assumed that prevailed. the major sources of N were not strongly altered by drought (Craine and others 2015). Collectively, our Does Land Use Alter N Uptake During results suggest that drought can increase soil N availability and enhance the potential for the re- Recovery from Drought? growth of biomass after drought, which con- Site fertility has been shown to modify grassland tributed particularly to the rapid recovery on the responses to climate change (Grime and others meadow, where the resource pulse during rewet- 2000); however, the immediate role of nutrient ting was larger and roots had a stronger capacity for availability for ecosystem recovery from drought is N uptake. still poorly understood. Recent studies have sug- gested that drought-induced increases in N turn- CONCLUSIONS over and availability can increase plant tissue N concentrations (Fuchslueger and others 2014; Ca- We conclude that the studied grasslands are highly narini and Dijkstra 2015; Arfin Khan and others resilient to extreme drought and that land-use 2016) and can thereby enhance grassland CO change has a strong potential for altering the rela- uptake dynamics during recovery (Roy and others tive contributions of resistance and recovery to the 2016). Our study supports the hypothesis that overall resilience. Our results suggest that aban- drought can increase leaf N concentrations during donment increases the resistance and decreases the Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 701 Bahn M, Reichstein M, Dukes JS, Smith MD, McDowell NG. recovery of grassland carbon dynamics across dif- 2014. Climate-biosphere interactions in a more extreme ferent carbon cycle parameters and different mea- world. New Phytol 202:356–9. sures of resilience. Rapid recovery from drought Bahn M, Schmitt M, Siegwolf R, Richter A, Bru¨ ggemann N. was supported by drought-induced increases in 2009. Does photosynthesis affect grassland soil-respired CO nitrogen availability and enhanced leaf nitrogen and its carbon isotope composition on a diurnal timescale? concentrations, which was more pronounced in New Phytol 182:451–60. the managed grassland. We conclude that managed Bahn M, Wohlfahrt G, Haubner E, Horak I, Michaeler W, Rottmar K, Tappeiner U, Cernusca A. 1999. Leaf photosyn- mountain grassland is likely prone to larger overall thesis nitrogen contents and specific leaf area of grassland perturbations from extreme early-summer species in mountain ecosystems under different land use Land droughts compared to abandoned grassland. use changes in European mountain ecosystems: ECOMONT Therefore, ongoing and future land-use changes concepts and results. Vienna: Blackwell. pp 247–55. have the potential to significantly alter impacts of Borken W, Matzner E. 2009. Reappraisal of drying and wetting climate extremes on grassland carbon dynamics. effects on C and N mineralization and fluxes in soils. Glob Change Biol 15:808–24. Canarini A, Dijkstra FA. 2015. Dry-rewetting cycles regulate ACKNOWLEDGEMENTS wheat carbon rhizodeposition, stabilization and nitrogen cy- Open access funding provided by University of cling. Soil Biol Biochem 81:195–203. Innsbruck and Medical University of Innsbruck. We Casper BB, Forseth IN, Kempenich H, Seltzer S, Xavier K. 2001. Drought prolongs leaf life span in the herbaceous desert thank Sarah Scheld, Mario Deutschmann, David perennial Cryptantha flava. Funct Ecol 15:740–7. Reinthaler and Marine Zwicke for assistance in the Ciais P, Reichstein M, Viovy N, Granier A, Oge´e J, Allard V, field. The study was conducted at the LTER-Austria Aubinet M, Buchmann N, Bernhofer C, Carrara A, Chevallier master site ‘‘Stubai Valley’’ and was financially F, de Noblet N, Friend AD, Friedlingstein P, Gru¨ nwald T, supported by the Austrian Science Fund (FWF Pro- Heinesch B, Keronen P, Knohl A, Krinner G, Loustau D, ject No I 1056) and by the German Federal Ministry Manca G, Matteucchi G, Miglietta F, Ourcival JM, Papale D, Pilegaard K, Rambal S, Seufert G, Soussana JF, Sanz M-J, of Education and Research (BMBF Project No. Schulze E-D, Vesala T, Valentini R. 2005. Europe-wide 01LC1203A) in the framework of the ERA-Net reduction in primary productivity caused by the heat and BiodvERsA project ‘‘REGARDS’’, coordinated by drought in 2003. Nature 437:529–33. Sandra Lavorel. A.A-R. was funded by a STSM of the Conant RT, Paustian K, Elliott ET. 2001. Grassland management COST Action STReESS (PF1106) and a FPU grant and conversion into grassland. Effects on soil carbon. Ecol from the Ministerio de Educacio´ n, Cultura y De- Appl 11:343–55. porte, Spain. Final data analysis and writing of this Craine JM, Brookshire ENJ, Cramer MD, Hasselquist NJ, Koba K, Marin-Spiotta E, Wang L. 2015. Ecological interpretations manuscript were supported by CLIMLUC (‘‘Climate of nitrogen isotope ratios of terrestrial plants and soils. Plant extremes and land use change—effects on ecosystem Soil 396:1–26. processes and services’’), funded by the Austrian deVries FT, Liiri ME, Bjørnlund L, Bowker MA, Christensen S, Academy of Sciences. J.I. received a Ph.D. comple- Seta¨la¨ HM, Bardgett RD. 2012. Land use alters the resistance tion grant from the University of Innsbruck. and resilience of soil food webs to drought. Nat Clim Change 2:276–80. OPEN ACCESS Dixon AP, Faber-Langendoen D, Josse C, Morrison J, Loucks CJ, Ebach M. 2014. Distribution mapping of world grassland This article is distributed under the terms of the types. J Biogeogr 41:2003–19. Creative Commons Attribution 4.0 International Estiarte M, Vicca S, Penuelas J, Bahn M, Beier C, Emmett BA, License (http://creativecommons.org/licenses/by/ Fay PA, Hanson PJ, Hasibeder R, Kigel J, Kroel-Dulay G, Larsen KS, Lellei-Kovacs E, Limousin J-M, Ogaya R, Ourcival 4.0/), which permits unrestricted use, distribution, J-M, Reinsch S, Sala OE, Schmidt IK, Sternberg M, Tielborger and reproduction in any medium, provided you K, Tietema A, Janssens IA. 2016. Few multiyear precipitation– give appropriate credit to the original author(s) and reduction experiments find a shift in the productivity-pre- the source, provide a link to the Creative Commons cipitation relationship. Glob Change Biol 22:2570–81. license, and indicate if changes were made. Frank D, Reichstein M, Bahn M, Thonicke K, Frank DC, Ma- hecha MD, Smith P, van der Velde M, Vicca S, Babst F, Beer C, Buchmann N, Canadell JG, Ciais P, Cramer W, Ibrom A, REFERENCES Miglietta F, Poulter B, Rammig A, Seneviratne SI, Walz A, Wattenbach M, Zavala MA, Zscheischler J. 2015. Effects of Arfin Khan MA, Kreyling J, Beierkuhnlein C, Jentsch A. 2016. climate extremes on the terrestrial carbon cycle. Concepts, Ecotype-specific improvement of nitrogen status in European processes and potential future impacts. Glob Change Biol grasses after drought combined with rewetting. Acta Oeco- 21:2861–80. logica 77:118–27. Fuchslueger L, Kastl E, Bauer F, Kienzl S, Hasibeder R, Ladreiter- Bahn M, Knapp M, Garajova Z, Pfahringer N, Cernsuca A. 2006. Knauss T, Schmitt M, Bahn M, Schloter M, Richter A, Szukics Root respiration in temperate mountain grasslands differing in U. 2014. Effects of drought on nitrogen turnover and abun- land use. Glob Change Biol 12:995–1006. 702 J. Ingrisch and others dances of ammonia-oxidizers in mountain grassland. Bio- Knapp AK, Carroll CJW, Denton EM, La Pierre KJ, Collins SL, geosciences 11:6003–15. Smith MD. 2015. Differential sensitivity to regional-scale drought in six central US grasslands. Oecologia 177:949–57. Garnier E, Cortez J, Billes G, Navas M-L, Roumet C, Debussche M, Laurent G, Blanchard A, Aubry D, Bellmann A, Neill C, Kro¨ el-Dulay G, Ransijn J, Schmidt IK, Beier C, de Angelis P, de Toussaint J-P. 2004. Plant functional markers capture Dato G, Dukes JS, Emmett BA, Estiarte M, Garadnai J, ecosystem properties during secondary succession. Ecology Kongstad J, Kovacs-Lang E, Larsen KS, Liberati D, Ogaya R, 85:2630–7. Riis-Nielsen T, Smith AR, Sowerby A, Tietema A, Penuelas J. 2015. Increased sensitivity to climate change in disturbed Gavazov KS, Spiegelberger T, Buttler A. 2014. Transplantation of ecosystems. Nat Commun 6:6682. subalpine wood-pasture turfs along a natural climatic gradient reveals lower resistance of unwooded pastures to climate Laliberte´ E, Shipley B, Norton DA, Scott D. 2012. Which plant change compared to wooded ones. Oecologia 174:1425–35. traits determine abundance under long-term shifts in soil resource availability and grazing intensity? J Ecol 100:662– Gilgen AK, Buchmann N. 2009. Response of temperate grass- lands at different altitudes to simulated summer drought dif- fered but scaled with annual precipitation. Biogeosciences Lambers H, Poorter H. 1992. Inherent variation in growth rate 6:2525–39. between higher plants: a search for physiological causes and ecological consequences. In: Begon M, Fitter A, Eds. Advances Grassein F, Lemauviel-Lavenant S, Lavorel S, Bahn M, Bardgett in ecological research. London, San Diego: Academic Press. p RD, Desclos-Theveniau M, Laine P. 2015. Relationships be- 187–261. tween functional traits and inorganic nitrogen acquisition among eight contrasting European grass species. Ann Bot Legay N, Baxendale C, Grigulis K, Krainer U, Kastl E, Schloter 115:107–15. M, Bardgett RD, Arnoldi C, Bahn M, Dumont M, Poly F, Pommier T, Cle´ment J-C, Lavorel S. 2014. Contribution of Grigulis K, Lavorel S, Krainer U, Legay N, Baxendale C, Dumont above- and below-ground plant traits to the structure and M, Kastl E, Arnoldi C, Bardgett RD, Poly F, Pommier T, function of grassland soil microbial communities. Ann Bot Schloter M, Tappeiner U, Bahn M, Cle´ment J-C, Hutchings M. 114:1011–21. 2013. Relative contributions of plant traits and soil microbial properties to mountain grassland ecosystem services. J Ecol Lepsˇ J, Osbornova´-Kosinova´ J, Rejma´nek M. 1982. Community 101:47–57. stability, complexity and species life history strategies. Vege- tatio 50:53–63. Grime JP, Brown VK, Thompson K, Masters GJ, Hillier SH, Clarke IP, Askew AP, Corker D, Kielty JP. 2000. The response MacDonald D, Crabtree J, Wiesinger G, Dax T, Stamou N, Fleury of two contrasting limestone grasslands to simulated climate P, Gutierrez Lazpita J, Gibon A. 2000. Agricultural abandon- change. Science 289:762–5. ment in mountain areas of Europe. Environmental conse- quences and policy response. J Environ Manag 59:47–69. Hasibeder R, Fuchslueger L, Richter A, Bahn M. 2015. Summer drought alters carbon allocation to roots and root respiration MacGillivray CW, Grime JP. 1995. Testing predictions of the in mountain grassland. New Phytol 205:1117–27. resistance and resilience of vegetation subjected to extreme events. Funct Ecol 9:640. Hodgson D, McDonald JL, Hosken DJ. 2015. What do you mean, ‘resilient’? Trends Ecol Evol 30:503–6. Meyer S, Leifeld J, Bahn M, Fuhrer J. 2012a. Free and protected soil organic carbon dynamics respond differently to aban- Holling CS. 1996. Engineering resilience versus ecological resi- donment of mountain grassland. Biogeosciences 9:853–65. lience. In: Schulze P, Ed. Engineering within ecological con- straints. Washington, D.C.: National Academy Press. p 31–44. Meyer S, Leifeld J, Bahn M, Fuhrer J. 2012b. Land-use change in subalpine grassland soils: effect on particulate organic car- Hoover DL, Knapp AK, Smith MD. 2014. Resistance and resi- bon fractions and aggregation. J Plant Nutr Soil Sci 175:401–9. lience of a grassland ecosystem to climate extremes. Ecology 95:2646–56. Nimmo D, Mac Nally R, Cunningham S, Haslem A, Bennett A. 2015. Vive la re´sistance: reviving resistance for 21st century Hoover DL, Rogers BM. 2016. Not all droughts are created equal: conservation. Trends Ecol Evol 30:516–23. the impacts of interannual drought pattern and magnitude on grassland carbon cycling. Glob Change Biol 22:1809–20. Obojes N, Bahn M, Tasser E, Walde J, Inauen N, Hiltbrunner E, Saccone P, Lochet J, Cle´ment J-C, Lavorel S, Tappeiner U, Jentsch A, Kreyling J, Boettcher-Treschkow J, Beierkuhnlein C. Ko¨ rner C. 2015. Vegetation effects on the water balance of 2009. Beyond gradual warming. Extreme weather events alter mountain grasslands depend on climatic conditions. Ecohy- flower phenology of European grassland and heath species. drology 8:552–69. Glob Change Biol 15:837–49. Oliver TH, Heard MS, Isaac NJ, Roy DB, Procter D, Eigenbrod F, Jentsch A, Kreyling J, Elmer M, Gellesch E, Glaser B, Grant K, Freckleton R, Hector A, Orme CDL, Petchey OL, Proenc¸a V, Hein R, Lara M, Mirzae H, Nadler SE, Nagy L, Otieno D, Raffaelli D, Suttle KB, Mace GM, Martı´n-Lo´ pez B, Woodcock Pritsch K, Rascher U, Scha¨dler M, Schloter M, Singh BK, BA, Bullock JM. 2015. Biodiversity and resilience of ecosys- Stadler J, Walter J, Wellstein C, Wo¨ llecke J, Beierkuhnlein C. tem functions. Trends Ecol Evol 30:673–84. 2011. Climate extremes initiate ecosystem-regulating func- tions while maintaining productivity. J Ecol 99:689–702. Osone Y, Ishida A, Tateno M. 2008. Correlation between relative growth rate and specific leaf area requires associations of Kaufman LH. 1982. Stream aufwuchs accumulation. Distur- specific leaf area with nitrogen absorption rate of roots. New bance frequency and stress resistance and resilience. Oecolo- Phytol 179:417–27. gia 52:57–63. Pirk N, Mastepanov M, Parmentier F-JW, Lund M, Crill P, Klumpp K, Tallec T, Guix N, Soussana J-F. 2011. Long-term Christensen TR. 2016. Calculations of automatic chamber flux impacts of agricultural practices and climatic variability on measurements of methane and carbon dioxide using short carbon storage in a permanent pasture. Glob Change Biol time series of concentrations. Biogeosciences 13:903–12. 17:3534–45. Land Use Alters the Drought Responses of Productivity and CO Fluxes in Mountain Grassland 703 Potts D, Huxman TE, Enquist BJ, Weltzin JF, Williams DG. 2006. Teuling AJ, Seneviratne SI, Sto¨ ckli R, Reichstein M, Moors E, Resilience and resistance of ecosystem functional response to Ciais P, Luyssaert S, van den Hurk B, Ammann C, Bernhofer a precipitation pulse in a semi-arid grassland. J Ecol 94:23–30. C, Dellwik E, Gianelle D, Gielen B, Gru¨ nwald T, Klumpp K, Montagnani L, Moureaux C, Sottocornola M, Wohlfahrt G. ´ ´ Quetier F, Thebault A, Lavorel S. 2007. Plant traits in a state and 2010. Contrasting response of European forest and grassland transition framework as markers of ecosystem response to energy exchange to heatwaves. Nat Geosci 3:722–7. land-use change. Ecol Monogr 77:33–52. Todman LC, Fraser FC, Corstanje R, Deeks LK, Harris JA, Pawlett R Development Core Team. 2015. A language and environment M, Ritz K, Whitmore AP. 2016. Defining and quantifying the for statistical computing. Vienna: R Foundation for Statistical resilience of responses to disturbance: a conceptual and Computing. modelling approach from soil science. Sci Rep 6:28426. Reich PB. 2014. The world-wide ‘fast-slow’ plant economics Vogel A, Scherer-Lorenzen M, Weigelt A. 2012. Grassland spectrum. A traits manifesto. J Ecol 102:275–301. resistance and resilience after drought depends on manage- Reichstein M, Bahn M, Ciais P, Frank D, Mahecha MD, ment intensity and species richness. PLoS One 7:e36992. Seneviratne SI, Zscheischler J, Beer C, Buchmann N, Frank Werner RA, Brand WA. 2001. Referencing strategies and tech- DC, Papale D, Rammig A, Smith P, Thonicke K, van der Velde niques in stable isotope ratio analysis. Rapid Commun Mass M, Vicca S, Walz A, Wattenbach M. 2013. Climate extremes Spectrom RCM 15:501–19. and the carbon cycle. Nature 500:287–95. Wheeler B. 2010. lmPerm: Permutation tests for linear models. Robson T, Lavorel S, Cle´ment J-C, Roux X. 2007. Neglect of R package version 1.1-2. Available at: http://cran.r-project. mowing and manuring leads to slower nitrogen cycling in org/web/packages/lmPerm/. subalpine grasslands. Soil Biol Biochem 39:930–41. Wolf S, Eugster W, Ammann C, Ha¨ni M, Zielis S, Hiller R, Stieger Roy J, Picon-Cochard C, Augusti A, Benot M-L, Thiery L, Dar- J, Imer D, Merbold L, Buchmann N. 2013. Contrasting re- sonville O, Landais D, Piel C, Defossez M, Devidal S, Escape C, sponse of grassland versus forest carbon and water. Environ Ravel O, Fromin N, Volaire F, Milcu A, Bahn M, Soussana J-F. Res Lett 9:89501. 2016. Elevated CO2 maintains grassland net carbon uptake under a future heat and drought extreme. Proc Natl Acad Sci Wright IJ, Reich PB, Westoby M, Ackerly DD, Baruch Z, Bongers USA 113:6224–9. F, Cavender-Bares J, Chapin T, Cornelissen JHC, Diemer M, Flexas J, Garnier E, Groom PK, Gulias J, Hikosaka K, Lamont Ruimy A, Jarvis PG, Baldocchi D, Saugier B. 1995. CO2 fluxes BB, Lee T, Lee WG, Lusk C, Midgley JJ, Navas M-L, Niinemets over plant canopies and solar radiation: a review. Adv Ecol U, Oleksyn J, Osada N, Poorter H, Poot P, Prior L, Pyankov VI, Res 26:1–68. Roumet C, Thomas SC, Tjoelker MG, Veneklaas EJ, Villar R. Schmitt M, Bahn M, Wohlfahrt G, Tappeiner U, Cernusca A. 2004. The worldwide leaf economics spectrum. Nature 2010. Land use affects the net ecosystem CO exchange and its 428:821–7. components in mountain grasslands. Biogeosciences 7:2297– Yeung AC, Richardson JS. 2016. Some conceptual and opera- tional considerations when measuring ‘resilience’. A response Schrama M, Bardgett RD. 2016. Grassland invasibility varies to Hodgson et al. Trends Ecol Evol 31:2–3. with drought effects on soil functioning. J Ecol 104:1250–8. Zeller V, Bahn M, Aichner M, Tappeiner U. 2000. Impact of Smith P. 2014. Do grasslands act as a perpetual sink for carbon? land-use change on nitrogen mineralization in subalpine Glob Change Biol 20:2708–11. grasslands in the Southern Alps. Biol Fertil Soils 31:441–8. Tasser E, Tappeiner U. 2002. Impact of land use changes on Zwicke M, Alessio GA, Thiery L, Falcimagne R, Baumont R, mountain vegetation. Appl Veg Sci 5:173–84. Rossignol N, Soussana J-F, Picon-Cochard C. 2013. Lasting Tasser E, Walde J, Tappeiner U, Teutsch A, Noggler W. 2007. effects of climate disturbance on perennial grassland above- Land-use changes and natural reforestation in the Eastern ground biomass production under two cutting frequencies. Central Alps. Agric Ecosyst Environ 118:115–29. Glob Change Biol 19:3435–48.

Journal

EcosystemsSpringer Journals

Published: Sep 15, 2017

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off