Impact of Changes in Land Use and Climate on Agricultural Grassland in the European Alps

Impact of changes in land use and climate on agricultural grassland in the European Alps

A thesis submitted to the Faculty of Biology, University of Innsbruck for the degree of Doctor of Philosophy by Georg Niedrist

Supervised by

Ulrike Tappeiner Institute of Ecology University of Innsbruck, Austria

Erich Tasser Institute for Alpine Environment, EURAC Bozen / Bolzano, Italy

November 2015

1 2 Contents

1. Extended Summary ...... 7 1.1 Introduction ...... 7 1.2 Overview on mountain grassland communities in the European Alps ... 8 1.3 Plant diversity responses to land use changes ...... 9 1.4 Plant responses to changing climatic variability ...... 10 1.5 Mountain grassland responses to changing climate: A modelling approach ...... 11 1.6 Mountain grassland responses to changing climate: An experimental approach ...... 13 1.7 Conclusion and outlook ...... 13 1.8 References ...... 16 2. Plant communities of mountain grasslands in a broad cross-section of the Eastern Alps ...... 19 3. Plant diversity declines with recent land use changes in European Alps ...33 4. A plant’s perspective of extremes: Terrestrial plant responses to changing climatic variability ...... 51 5. Modeling changes in grassland hydrological cycling along an elevational gradient in the Alps ...... 98 6. Down to future: Transplanted mountain meadows react with increasing phytomass and shifting species composition ...... 145 7. Curriculum Vitae ...... 181 8. Additional publications, posters and presentations ...... 182 8.1 Peer reviewed publications ...... 182 8.2 Book contributions ...... 183 8.3 Oral presentations ...... 184 8.4 Posters ...... 185 9. Contribution statement ...... 186 10. Acknowledgments/Danksagung ...... 188

3 4 Abstract

Mountain pastures and hay meadows represent both a source of livelihood for mountain agriculture as well as a biodiversity hot spot. This PhD thesis investigates the impacts of changes in land use and climate on species composition, water balance and productivity of agricultural grassland. The methods include field investigations (paper 1 and 2), literature reviews (paper 1, 2 and 3), modelling (paper 4) and field experiments (paper 4 and 5). All five studies are based on data from agricultural grasslands in the European Alps; paper 3 also includes data from other ecosystems. In general, the results highlight the diversity of these grasslands both at the plot level (α-diversity, max. 63 species per plot) and landscape level (β-diversity, 39 grassland communities). However, the vulnerability of these ecosystems was also visible, especially to land use intensification (-63.4% loss of species richness). Site conditions such as elevation or inclination were found to be less decisive. We found that increasing temperature influenced species composition, especially at lower elevations. Total species number, however, was not affected, but changes are probably only detectable with long-term observation. Productivity increased significantly with a +3°K scenario at the subalpine elevation, whereas no increase was observed at the foothill elevation. Modelling as well as in-situ measurements indicate that the higher (potential) evapotranspiration at lower elevations causes more and longer drought periods (that override the potential advantage of higher temperature at lower elevations). Late spring seem to be particularly sensitive to such drought events. Extreme events and the effect on plant physiology and phenology are still not fully understood and necessitate interdisciplinary approaches that combine experiments, modeling and long-term observation.

5 6 1. Extended Summary

1.1 Introduction

Grassland ecosystems comprise a large proportion of all mountain landscapes worldwide. Grasslands cover two-thirds (i.e. 1.5 million km²) of the Tibetan plateau (Cui and Graf 2009), 37% of Tropical Andes (Tovar et al. 2013). In the European Alps 18% are covered by grasslands (Flury et al. 2013), many of them are agriculturally influenced. Though, the patterns of use are as varied as the global extent; for example, uses range from up to 7 times mown and fertilized meadows in in the pre-alpine lowlands to near-natural grasslands in remote, nutrient-poor areas, between the timberline and nival belt. Grasslands fulfill a wide spectrum of fundamental ecosystem services, depending on their location (inclination, elevation) and management intensity. Intensively used meadows and pastures are valuable for provisioning services such as hay and pastureland (O'Mara 2012). More diverse, colorful, flowering grasslands provide cultural services such as aesthetic aspects (Schirpke et al. 2013). Species-rich grasslands on steep slopes and high elevations fulfill crucial regulating services such as soil stabilization (Tasser et al. 2003) and soil water content regulation (Gross et al. 2014, Obojes et al. 2015). Currently, there are two main drivers that affect grassland species composition and thus their ecological function: 1) Modern changes in land use started in Europe during the industrial revolution and became more substantial after World War II when sociocultural changes, mechanization and rural flight led to radical changes in the structure of mountain agriculture (MacDonald et al. 2000). Consequently, grasslands on steep and poorly accessible sites lost importance for fodder production, while favorable sites (e.g. valley bottom) became more and more intensively managed in terms of cutting frequency and fertilizing (Tappeiner et al. 2008). 2) Rising greenhouse emissions affect grasslands rather indirectly via increasing global temperature and changing precipitation patterns than by rising CO2 concentrations (Dukes et al. 2005). In the European Alps, where temperature has increased by 2°K since 1900, scenarios predict a further increase of 1.5– 4.8°K by the end of the 21st century (IPCC 2013, Gobiet at al. 2014). In contrast to changing land use intensity, the consequences of climate change seem to be

7 less immediate, though they affect grasslands worldwide, from lowlands up to the alpine belt (Song et al. 2012, Cornelius et al. 2013). Both the drivers and the impact of such changes on grasslands have been extensively studied for decades (for reviews see Stoate et al. 2009, Queiroz et al. 2014, Dumont et al. 2015), but many questions are still unanswered. For example, studies on climate and land use change are often based on a single type of grassland or few study sites and are thus limited in their representativeness. Furthermore, it is not clear how extreme events and constant warming affect plant growth (Bahn et al. 2014) or to which extent the effects of higher temperature get compensated or even enhanced by lower soil water availability. This doctoral thesis consists of five studies that investigate some of these open issues. Paper 1 is an overview of the status quo (distribution and abundance) of grassland communities in the European Alps. Paper 2 builds on this database and compares diversity indicators among different land use types, i.e. very intensively used grassland, extremely species-rich meadows, pastures and abandoned areas. Paper 3, while maintaining the focus on grassland, also brings the effects of climate change into question. More in detail, it depicts differences in plant responses between changing mean climatic conditions and extreme events and provide an overview of suitable combined approaches to discriminate between changing climate mean and changing climate variability. The last two papers make use of these combined approaches. In paper 4 we investigate the effect of rising temperature on grassland (hay meadows) vegetation, bringing together a hydrological model with transect observations. In paper 5 we used field experiments along an altitude gradient to study grassland diversity and productivity in response to global warming. Finally, the extended abstract contains concluding remarks and an outlook on remaining open issues.

1.2 Overview on mountain grassland communities in the European Alps

Despite or maybe even because of the increasing importance of remote sensing technologies (Franke et al. 2012), reliable ground data remain fundamental for grassland classification and distribution. The study in paper 1 is based on 1883 vegetation relevés (Braun-Blanquet method) distributed over an area of 18000km². 1502 vegetation relevés were taken from 40 literature sources (papers, master and PhD theses) and were supplemented with 375 new surveys.

8 Data were collected on vegetation (species abundances) and site conditions (slope, altitude, land use and pH). The results confirmed that agricultural grasslands are one of the most diverse ecosystems, both in terms of α-diversity (769 plant species) and β-diversity (39 plant communities). The lowest mean species number per community was found either in near natural, acid grassland (Hygrocaricetum curvulae, 14.4 species per 4x4m) or in over-used habitats (in terms of fertilizing and grazing) such as the Rumicetum alpine (12.5 species) and the Lolium multiflorae (15.8 species). The most species-rich communities were found in lightly used grassland (pastures and meadows) such as the Campanulo scheuchzeri-Festucetum noricae (44.2 species per 4x4 m) or the Seslerio-Caricetum sempervirentis (44.8 species), both of which are located mainly on calcareous bedrock. Some of the plant communities were found only in specific regions e.g., the Crepido- Cynosuretum, which depends on high precipitation, or the Seslerio-Caricetum sempervirentis, which relies on calcareous bedrock. A clear pattern emerged regarding site conditions. Grasslands that are steeper than 20° are abandoned or sporadically used as pastures. Though this is not surprising, land use maps from the 1970s prove that most of them were used regularly, which demonstrates how fast land is abandoned in unfavorable areas. Altitude, in contrast, is less determining for land use intensity. Fertilized meadows can be found even in the upper subalpine zone (~2200 m a.s.l.) if they are not steep and are easily accessible. This study involved collecting data, digitalizing historic data, controlling the quality of data and harmonizing datasets (method, nomenclature). Overall, it is descriptive rather than analytic, but to our knowledge, it is one the most exhaustive datasets on managed grassland vegetation in the Alps.

1.3 Plant diversity responses to land use changes

Based on the above dataset further analyses were carried out to study the impact of land use intensities on species composition and diversity. Ten land use types were compared for botanical aspects (species number, characteristic species) and ecological processes (interspecific competition). A wide range of agricultural grasslands is represented, from meadows that are frequently cut (up to 5 times a season) and fertilized, to pastures and abandoned grassland. This variety enabled a relative comparison between the effects of land use intensification and land abandonment. For a numerical analysis, land use intensity was quantified by counting all human interventions per year (cutting, fertilizing).

9 Plant species richness and land use intensity were strongly correlated (r²=0.903). Surprisingly, the average species richness in abandoned land was not significantly lower than in lightly used grassland (pastures and meadows), which were the most species-rich land use type. For this study, “abandoned” was defined as an area left <30 years ago. However, intensively used (i.e. cut and fertilized) meadows are far less diverse than abandoned meadows (14.3 versus 36.5 species per 4x4m), which indicates that land use intensification has a much more pronounced impact on plant species richness than land abandonment. Species composition and interspecific competition are also strongly dependent on land use intensity. On frequently mown grassland, species abundance is distributed much more evenly than in lightly used areas or abandoned land, suggesting that light competition gets overruled by human disturbance on intensively used areas. A second intention of this study was to describe characteristic species composition for the most common land use types. By conducting a discriminant analysis, we found that grasses were a useful indicator for specific land use types. Deschampsia caespitosa and Nardus stricta were typical for pastures, while Dactylis glomerata and Alopecurus pratensis were reliable indicators for continuously disturbed grasslands. Agrostis capillaris was typical for fertilized meadows with low cutting frequency and Briza media appeared commonly in unfertilized meadows. The hypothesis for this work was that a characteristic species list could serve as the scientific basis for certification of sustainably produced hay and could be used as an additional criteria for agro-environmental payments or quality certifications in wellness applications (Dalla Via et al. 2004).

1.4 Plant responses to changing climatic variability

Land use change in mountain grasslands does not only affect species diversity and composition, but plays a crucial role in how grasslands react to rising global temperature (Theurillat and Guisan 2001, Buetof et al. 2012). Intensively used grasslands seem to be particularly affected by climate change and climatic extremes (Walter et al. 2013). Thus, in addition to land use changes it is important to evaluate and discriminate the effects of changes in mean temperature and in temperature variability on vegetation and to design adequate experimental approaches to assess the consequences on phenological, physiological and interacting physiological processes. We approached these questions with a literature review (paper 3). Phenological studies revealed that increases in mean temperature provoke a prolongation of

10 the growing season as well as an earlier onset of leaf unfolding and first flowering. The few studies on effects of changing climate variability, however, report elevated risk of frost damage, second flowering and shifted flowering time. One of the conclusions of the literature review was that knowledge on this aspect is still superficial. In comparison, the physiological response of a plant to climate variability has been more extensively investigated. Nevertheless, there is ongoing discussion whether isodydric plants (i.e. close their stomata during drought and risk carbon starvation) or anisohydric plants (i.e. keep their stomata open and risk drying out) are more successful. The issue is even more complex when looking at the community level compared to the species level. The combination (i.e. mitigation and enhancement) of different species-specific responses renders it almost impossible to differentiate whether the observed effects are caused by changes in mean or in variability. Overall, many studies suggest that changing climate variability may be even more important than nutrient availability as the limiting factor in some situations. The complex interaction of changing mean and changing variability along different scales necessitates suitable approaches for future climate change research. In this review we concluded that observational studies are more appropriate to detect effects of changing means, while experimental studies can investigate a specific factor but risk losing a broader context. A clear and realistic definition of duration of the extreme event is crucial, regardless of the type of experimental set-up. Models can be a useful instrument; however, the output for detecting effects of climate variability are often biased by uncertainties of global/regional climate models, and input climate variables should have an appropriate time resolution to smooth climatic extremes. Finally, this literature review emphasizes the value of combined approaches (e.g. observation with experiments; experiments with models) to differentiate between plant-related effects of changing climate mean and changing climate variability.

1.5 Mountain grassland responses to changing climate: A modelling approach

The recommendation to use combined approaches was the motivation behind establishing a test area in 2009 in Matsch/Mazia (Italian Alps). We used a 3- point elevation transect (1000 m, 1500 m and 2000 m a.s.l.) to simulate climate warming (2.8°K) within a range that reflects climate scenarios for the inner-

11 alpine region for the year 2100. In-situ measurements, models and field experiments were done along the transect to assess the impact of rising temperature on grassland productivity and diversity. In paper 4 we measured the main components of the hydrological cycle in grasslands (i.e. soil water content, evapotranspiration, aboveground biomass as the basis for evapotranspiration, water use efficiency and snow water equivalent) and compared them with the output of a hydrological model named GEOtop-dv. We considered data from two contrasting years to cover meteorological variability (2010 was cool and wet, 2011 was warm and dry). Furthermore, 22 years of meteorological data were used to analyze long-term climate variability. Not surprisingly, soil water content and snow water equivalent decreased with decreasing elevation and rising temperature. However, the comparison between the two years as well as the 22 year simulation demonstrated that this trend is much more pronounced in warmer years than drier years. At 2000 m a.s.l. almost no drought periods could be observed in the upper 5 cm. At the lowest site at 1000 m a.s.l., in contrast, drought periods of several days/weeks occurred in all years (which is the reason for the high demand of additional irrigation at this elevation). Evapotranspiration reached its maximum at 1500 m a.s.l. (500– 600 mm a-1). At 2000 m it was limited by temperature and growth period length, whereas at 1000 m it was limited by water. Those results refer to not irrigated conditions. Under irrigated conditions, however, the highest evapotranspiration occurred at 1000m (1000–1100 mm a-1) and followed the pattern of potential evapotranspiration. A very similar behavior was observed for aboveground biomass and water use efficiency. From a hydrological point of view these results demonstrate that the Alps do not naturally act as a water tower. There is a kind of threshold elevation in inner- alpine regions below which most of the precipitation input (on grasslands) gets lost by evapotranspiration and above which the water balance is positive. With ongoing climate change it is likely that we can expect an upwards shift of this threshold, which in the inner-alpine valleys is currently around 1500 m a.s.l. From a plant perspective, however, the results demonstrate that grasslands at higher elevations may benefit from higher temperature as long they are not water limited. At lower elevations, the advantage of higher temperature is limited by higher (potential) evapotranspiration and consequently more frequent drought periods.

12 1.6 Mountain grassland responses to changing climate: An experimental approach

Similar results were obtained by conducting a field experiment along the transect described in paper 5. In this work, we used a “space for time” transplantation to study the effects of future, warmer climate on plant species diversity, composition and aboveground productivity. We transplanted 0.7 x 0.7m monoliths of moderately used hay meadows from 2000 m a.s.l. to 1500 m a.s.l. and from 1500 m a.s.l. to 1000 m a.s.l. Control monoliths were transplanted locally, i.e. at the same elevation. Aboveground phytomass, several biodiversity indicators (e.g. total number of plant species, evenness index) and species composition were recorded for three consecutive years. Phytomass data confirmed the results from vegetation modelling in paper 4—all functional groups profited significantly from transplanting from 2000m to 1500 m a.s.l. (legumes +213.6%, herbs +128.2%, graminoids +51.7%), but no change was seen when transplanting from 1500 m to 1000 m a.s.l. Species number (between 9 and 23 species per plot) as well as species abundance distribution (evenness index between 0.7 and 0.9) showed no change. Ordination results (NMDS) showed that species composition was relatively stable in the 2000 m to 1500 m transplantation, whereas a significant shift could be detected after the third year in the monoliths transplanted from 1500 to 1000 m. Monoliths originally located at 1500 m changed significantly to reflect the local vegetation at 1000 m. A further analysis on a species level revealed that the observed shift could be attributed to both methodological artifacts (invasion from the surrounding vegetation) and warmer climate. The minor changes in the upper transplantation were caused by the transplantation process itself, mainly by root damage of deep-rooting plants. Taken together, the results suggest that the reaction of mountain hay meadows to climate warming change considerably with altitude.

1.7 Conclusion and outlook

The results of this doctoral thesis reveal that both direct (land use changes) and indirect (climate change) human activities have a strong impact on mountain grassland vegetation. Overall, the impact on species diversity was more apparent with land use change, especially management intensification. Certainly, regarding climate change, most studies predict a long-term shift in plant composition especially in higher altitudes and latitudes (Pauli et al. 2012), an aspect that we did not investigate in these short-midterm experiments.

13 Although research on grassland ecosystems and processes under changing conditions has intensified over the last decades, there are still several open issues. Many of these issues also emerged in this thesis and—depending on available funding—will be the subject of my future research:  Do grassland types (as described in paper 1) react differently to climate and land use change, and if so, how? There are numerous single, isolated studies, but these are often carried out under lab conditions (De Boeck et al. 2008) or by applying diverging climate and land use scenarios (e.g. Izaurralde et al. 2011, Rustad et al. 2001). Comparative studies of many grassland types would be necessary to gain insight into the vulnerability of a landscape (as a mosaic of different grassland types) to ongoing changes.  What is the relative importance of land use and climate change in their impact on grassland ecosystems? As this is expected to vary considerably from region to region, this information could be used to develop scale- and region-dependent priorities in protection policy (nature and human!). The answer is certainly not simple because of the multitude of factors, including nitrogen input and trade-off/feedback processes between all drivers (Baron et al. 2000, Gerdol et al. 2008).  Which role do longer vegetation periods play in relation to higher temperature during the vegetation season? This seems to be especially important for ecosystems with long snow cover periods. Existing studies seem to indicate that shorter snow cover is crucial (Gerdol et al. 2013) and locally even more important for species composition and productivity than elevated temperature during the vegetation period (Ernakovich et al. 2014).  As suggested by the results in paper 3, extreme events and their consequences on mountain grasslands may represent be one of the widest gaps in current environmental sciences. Uncertainties about the frequency and magnitude of future droughts, heatwaves or floods render it difficult to impose realistic extremes on ecosystems (Smith 2011).

Finally, I would like to make a personal reflection at the end of this Philosophiae Doctor-thesis. Looking at the lines above, “change” is among the most frequently repeated words. More than 26000 environmental ISI-publications in the last 15 years contain this word in their title (as of 10.08.2015). Considering humans as a part of nature, this leads logically to the question whether human-induced change should not be seen as something natural. Is it reasonable to hold

14 desperately onto the status quo even though Heraclitus postulated the panta rhei concept already 2500 years ago? In other words, can this be called sustainability preserving (artificially created) species-rich meadows in times where farmers would need much more forage/income to survive economically? It might depend on the respective point of view. However, from the experience gained during this research, I am convinced that conservation and development is possible at the same time. And as a scientist it is my responsibility to serve both.

15 1.8 References

Baron J, Rueth H, Wolfe A, Nydick K, Allstott E, Minear J, Moraska B (2000) Ecosystem responses to nitrogen deposition in the Colorado Front Range. Ecosystems 3:352-368.

Buetof A, von Riedmatten LR, Dormann CF, Scherer-Lorenzen M, Welk E, Bruelheide H (2012) The responses of grassland plants to experimentally simulated climate change depend on land use and region. Global Change Biol 18:127-137.

Cornelius C, Leingaertner A, Hoiss B, Krauss J, Steffan-Dewenter I, Menzel A (2013) Phenological response of grassland species to manipulative snowmelt and drought along an altitudinal gradient. J Exp Bot 64:241-251.

Cui X, Graf H (2009) Recent land cover changes on the Tibetan Plateau: a review. Clim Change 94:47-61.

De Boeck HJ, Lemmens CMHM, Zavalloni C, Gielen B, Malchair S, Carnol M, Merckx R, Van den Berge J, Ceulemans R, Nijs I (2008) Biomass production in experimental grasslands of different species richness during three years of climate warming. Biogeosciences 5:585-594

Dukes J, Chiariello N, Cleland E, Moore L, Shaw M, Thayer S, Tobeck T, Mooney H, Field C (2005) Responses of grassland production to single and multiple global environmental changes. Plos Biology 3:1829-1837.

Dumont B, Andueza D, Niderkorn V, Luescher A, Porqueddu C, Picon-Cochard C (2015) A meta-analysis of climate change effects on forage quality in grasslands: specificities of mountain and Mediterranean areas. Grass Forage Sci 70:239-254.

Ernakovich JG, Hopping KA, Berdanier AB, Simpson RT, Kachergis EJ, Steltzer H, Wallenstein MD (2014) Predicted responses of arctic and alpine ecosystems to altered seasonality under climate change. Global Change Biol 20:3256-3269.

Flury C, Huber R, Tasser E (2013) Future of Mountain Agriculture in the Alps. The Future of Mountain Agriculture. Springer Geography: 105-126.

Franke J, Keuck V, Siegert F (2012) Assessment of grassland use intensity by remote sensing to support conservation schemes. Journal for Nature Conservation 20:125-134.

16 Gerdol R, Siffi C, Iacumin P, Gualmini M, Tomaselli M (2013) Advanced snowmelt affects vegetative growth and sexual reproduction of Vaccinium myrtillus in a sub-alpine heath. Journal of Vegetation Science 24:569-579.

Gerdol R, Bragazza L, Brancaleoni L (2008) Heatwave 2003: high summer temperature, rather than experimental fertilization, affects vegetation and CO(2) exchange in an alpine bog. New Phytol 179:142-154.

Gobiet A, Kotlarski S, Beniston M, Heinrich G, Rajczak J, Stoffel M, 2014: 21st century climate change in the European Alps—A review. Sci total environ 493:1138-1151.

Gross N, Robson T M, Lavorel S, Albert C, Le Bagousse-Pinguet Y, Guillemin R, 2008) Plant response traits mediate the effects of subalpine grasslands on soil moisture. New Phyt 180:652–662.

IPCC (2013) Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 1535 pp.

Izaurralde RC, Thomson AM, Morgan JA, Fay PA, Polley HW, Hatfield JL (2011) Climate Impacts on Agriculture: Implications for Forage and Rangeland Production. Agron J 103:371-381.

MacDonald D, Crabtree JR, Wiesinger G, Dax T, Stamou N, Fleury P, Gutierrez Lazpita J, Gibon A (2000) Agricultural abandonment in mountain areas of Europe: Environmental consequences and policy response. J Environ Manage 59:47-69.

Obojes N, Bahn M, Tasser E, Walde J, Inauen, N, Hiltbrunner E, Saccone P, Lochet J, Clément J C, Lavorel S, Tappeiner U, Körner C (2015) Vegetation effects on the water balance of mountain grasslands depend on climatic conditions. Ecohydrol 8:550–567.

O'Mara FP (2012) The role of grasslands in food security and climate change. Annals of Botany 110:1263-1270.

Pauli H, Gottfried M, Dullinger S, Abdaladze O, Akhalkatsi M, Benito Alonso JL, Coldea G, Dick J, Erschbamer B, Fernandez Calzado R, Ghosn D, Holten JI, Kanka R, Kazakis G, Kollar J, Larsson P, Moiseev P, Moiseev D, Molau U, Molero Mesa J, Nagy L, Pelino G, Puscas M, Rossi G, Stanisci A, Syverhuset AO, Theurillat J, Tomaselli M, Unterluggauer P, Villar L, Vittoz P, Grabherr G

17 (2012) Recent Plant Diversity Changes on Europe's Mountain Summits. Science 336:353-355.

Queiroz C, Beilin R, Folke C, Lindborg R (2014) Farmland abandonment: threat or opportunity for biodiversity conservation? A global review. Front Ecol Environ 12:288-296.

Rustad L, Campbell J, Marion G, Norby R, Mitchell M, Hartley A, Cornelissen J, Gurevitch J, GCTE-NEWS (2001) A meta-analysis of the response of soil respiration, net nitrogen mineralization, and aboveground plant growth to experimental ecosystem warming. Oecologia 126:543-562.

Schirpke U, Tasser E, Tappeiner U (2013) Predicting scenic beauty of mountain regions. Landscape Urban Plann 111:1-12.

Smith MD (2011) The ecological role of climate extremes: current understanding and future prospects. J Ecol 99:651-655.

Song L, Fan J, Harris W, Wu S, Zhong H, Zhou Y, Wang N, Zhu X (2012) Adaptive characteristics of grassland community structure and leaf traits along an altitudinal gradient on a subtropical mountain in Chongqing, China. Plant Ecol 213:89-101.

Stoate C, Baldi A, Beja P, Boatman ND, Herzon I, van Doorn A, de Snoo GR, Rakosy L, Ramwell C (2009) Ecological impacts of early 21st century agricultural change in Europe - A review. J Environ Manage 91:22-46.

Tappeiner U, Tasser E, Leitinger G, Cernusca A, Tappeiner G (2008) Effects of Historical and Likely Future Scenarios of Land Use on Above- and Belowground Vegetation Carbon Stocks of an Alpine Valley. Ecosystems 11:1383-1400.

Tasser E, Mader M, Tappeiner U (2003) Effects of land use in alpine grasslands on the probability of landslides. Basic Appl Ecol 4:271-280.

Theurillat J, Guisan A (2001) Potential Impact of Climate Change on Vegetation in the European Alps: A Review. Clim Change 50:77-109.

Tovar C, Seijmonsbergen AC, Duivenvoorden JF (2013) Monitoring land use and land cover change in mountain regions: An example in the Jalca grasslands of the Peruvian Andes. Landscape urban plan 112:40-49.

Walter J, Hein R, Beierkuhnlein C, Hammerl V, Jentsch A, Schaedler M, Schuerings J, Kreyling J (2013) Combined effects of multifactor climate change and land-use on decomposition in temperate grassland. Soil Biology & Biochemistry 60:10-18.

18 2. Plant communities of mountain grasslands in a broad cross-section of the Eastern Alps

Lüth C., Tasser E., Niedrist G., Dalla Via J., Tappeiner U. (2010): Flora 206(5):433-443.

19 20 Flora 206 (2011) 433–443

Contents lists available at ScienceDirect

Flora

journal homepage: www.elsevier.de/flora

Plant communities of mountain grasslands in a broad cross-section of the Eastern Alps

Christian Lüth a, Erich Tasser b, Georg Niedrist b, Josef Dalla Via c, Ulrike Tappeiner a,∗ a Institute of Ecology, University of Innsbruck, Sternwartestr. 15, 6020 Innsbruck, Austria b European Academy of Bolzano/Bozen, Drususallee 1, 39100 Bozen, Italy c Research Centre for Agriculture and Forestry Laimburg, Laimburg 6, 39051 Pfatten/Auer, Italy article info abstract

Article history: Apart from forests, the landscape of the Alps is dominated by grasslands, where they account for up to Received 2 March 2010 40% of the agricultural area. This study focuses on the main man-made grassland plant communities Accepted 27 July 2010 of the Eastern Alps, shows their current spatial distribution and examines how strongly the influence of land use and site factors determines the communities. Discriminant analysis was used to harmonize Keywords: the phytosociological classification of 1502 vegetation relevés from the literature and 375 own recorded Land-use types inventories from Western Austria and Northern Italy. Land-use intensity, altitude, slope and pH were also Land-use intensity recorded, in order to assess the impact of the factors to plant communities, as calculated in nonmetric Site factors Meadows multidimensional scaling. We identified 39 plant communities and generated a table with the main eco- Pastures logical and floristic parameters as well as a map showing their present spatial distribution. Contrary Abandoned areas to the literature, the pasture communities Crepido-Festucetum commutatae, Deschampsio cespitosae- Poetum alpinae and Rumicetum alpini occur also in fertilized meadows. On the other hand we found meadow communities occurring in pastures, such as the Angelico-Cirsietum oleracei, the Pastinaco- Arrhenatheretum, the Ranunculo repentis-Alopecuretum pratensis and the Trisetetum flavescentis. The most species-rich communities – the Caricetum ferruginei and the Seslerio-Caricetum sempervirentis – occur in unfertilized meadows above calcareous bedrock. Further species-rich communities – the Cam- panulo scheuchzeri-Festucetum noricae, the Gentianello anisodontae-Festucetum variae, the Pulsatillo alpinae-Festucetum noricae, the Trifolio thallii-Festucetum nigricantis and the Hypochoerido uniflorae- Festucetum paniculatae – are endangered: they are regionally restricted and depend on the absence of fertilizer and on mowing once annually or every second or third year. Therefore agri-environmental measures should focus on unfertilized mountain meadows, in order to conserve these rare grassland communities. © 2010 Elsevier GmbH. All rights reserved.

Introduction unfertilized and mown annually or every second to third year. Gen- erally only the easily accessible meadows are fertilized and allow Grasslands represent one of the most diverse man-made land- two cuts of hay in one year. Depending on land use, typical plant scape formations in the Alps (Ellenberg, 1996; Kremer, 1991; communities have become established over the past centuries and Maurer et al., 2006). The percentage of meadows and pastures in millennia (Dullinger et al., 2003; Liira et al., 2008; Marcos et al., the subalpine-alpine belt lies between 5% and 40% of the agricul- 2003; Oomes, 1992; Tasser et al., 2003; von Arx et al., 2002). On the turally used area, depending on the region (Tasser et al., 2009). alpine belt meadows typically support a large percentage of herbs Generally, due to the shorter growing season, the unfavourable cli- as well as dwarf-shrubs (Ellenberg, 1996; Ellmauer, 1996; Marini mate and topographic conditions, farmers have tended very early et al., 2007) and show a pronounced species-richness (Grabherr and to a marginal use of these areas (Niedrist et al., 2008). Mucina, 1993; Mucina et al., 1993). Alpine grasslands are mainly used as summer pastures Besides land-use management, small changes in site factors (Ellenberg, 1996), but a small proportion are meadows, usually (altitude, slope angle and soil pH) increase ecosystem diversity (Kampmann et al., 2007; Marini et al., 2007). This is demonstrated impressively over large areas in alpine pastures, where animals are left more or less unattended (Tasser et al., 2003). Grazing animals Abbreviations: DA, discriminant analysis; NMS, nonmetric multidimensional prefer the gentle slopes of the pastures with their dry and undis- scaling; a.s.l., above sea level. ∗ Corresponding author. Tel.: +43 512 507 5923; fax: +43 512 507 2975. turbed soils. Depending on topsoil pH, communities develop with E-mail address: [email protected] (U. Tappeiner). Nardus stricta on acid soil and with Sesleria albicans on neutral to

21 434 C. Lüth et al. / Flora 206 (2011) 433–443 alkaline soils (Grabherr and Mucina, 1993; Mucina et al., 1993; and 10◦08–12◦45 E(Fig. 1). Average annual precipitation ranges Myklestad and Sætersdal, 2004). Plane places are used as animal from 700 mm to 2000 mm, with maximum rainfall observed from resting places, where the increase in local nutrient load leads to June to July (Fliri, 1998). Mean annual temperature ranges from the development of typical trodden swards. On steep pastures, the 0 ◦Cto9◦C. Strong climatic distinctions are caused by the fact that free range style of grazing in the Alps greatly reduces the frequency relevés were taken from 650 m to 2680 m a.s.l. and that two climatic of use by grazing cattle and thus the level of nutrient deposition. regions – Continental and Atlantic climate – affect the vegetation The resulting vegetation covers exhibit a dwarf-shrub invasion or in the research area (Fliri, 1998). The bedrock of the research area typical high alpine gramineous-rich communities. is comprised of calcareous sedimentary rocks in the northern and In the Alps land use combined with site factors results in a the southern regions and of primary rocks in the central massive, large variety of plant communities (Pykälä, 2000) and creates a sometimes with superimposed calcareous isles (Bögel and Schmidt, landscape with high diversity (Tappeiner et al., 2008). Hence, this 1976). The pH of the topsoil (0–10 cm), which ranges from 3.7 to region is a ‘hot spot’ of biodiversity in Europe (World Wildlife 7.8 (Niedrist et al., 2008), is either affected by the bedrock (Scheffer Fund, 2004). Unfortunately, traditional land-use practices have et al., 2002) or else has been modified as a result of fertilizer appli- become less important over the last few decades in the Alps cations (Seeber and Seeber, 2005). (Streifeneder et al., 2007; Tasser et al., 2003). About 20% – in some areas even much more – of the agricultural land has been Data collection abandoned (Macdonald et al., 2000; Tappeiner et al., 2008; Tasser et al., 2007). The crucial factor here is that steep areas and small We first collected 1507 vegetation relevés from the literature plots create higher production costs, thus mountain agriculture (see Appendix), which had been recorded following the method of is hardly competitive on national and international markets. Braun-Blanquet (1964). This collection comprised meadows, pas- Consequently marginal productive areas in the Alps have been tures and abandoned areas from 60 different sites. Thereby the increasingly abandoned since 1950. plot size ranges from 12 m2 to 25 m2. All relevés included detailed However, the severity of this decline varies significantly, information on land use, geographical coordinates and the site depending on the region (Tasser et al., 2007, 2008): in the region factors for altitude and slope angle. For regions where no litera- ‘Südtiroler Unterland/Überetsch’ – one of the most productive ture data were found, we consulted local experts (farmers, park regions of the Alps – only 6% of formerly used agricultural areas rangers and agrarian decision-makers of the particular districts) are currently abandoned, while in the region ‘Innsbruck-Land’ they about the location of remaining mountain hay meadows and pas- amount to 37%, and in the region ‘Carnia’ even to 67%. On the tures. From this information we then recorded 265 vegetation other hand, an intensification of land use can also be observed relevés from 84 sites. Furthermore, at 30 sites we recorded 110 veg- (Krausmann et al., 2003; Mottet et al., 2006), due to the mas- etation relevés of intensively used meadows from valley regions sive extension of forest roads and farm tracks, which leads to a (between 650 and 1200 m a.s.l.), in order to compare them with higher accessibility, mechanisation and manuring of agricultural lightly used ones. Fieldwork was carried out between 2005 and areas in the alpine and subalpine altitudinal belt (Liira et al., 2008; 2007. Vegetation relevés were recorded according to the method Pavlu˚ et al., 2005; Stöcklin et al., 2007). All these changes lead of Braun-Blanquet (1964): the minimum area ranged from 6 m2 to a reduction of semi-natural grasslands (Bakker and Berendse, (in intensively fertilized meadows from the valley) up to 20 m2 (in 1999; Lavorel et al., 1998; Tappeiner et al., 1998), and an old and mountain meadows without fertilization), depending on the het- invaluable cultural landscape is gradually disappearing. erogeneity of the grassland area; most of the relevés covered an Several studies exist on grasslands in the Eastern Alps (Bischof, area of 12 m2 (i.e. 4 m × 3 m). At least two relevés were recorded 1981; Kohler et al., 2004; Maurer et al., 2006; Tasser and Tappeiner, for each study site, the variables altitude, slope, and pH (CaCl2) 2002; Vonlanthen et al., 2006; Zimmermann and Kienast, 1999)as of the topsoil (0–10 cm) were measured and the managing farm- well as numerous diplomas and Ph.D. theses (see Appendix) con- ers were interviewed to obtain exact information on land use. For taining field relevés of grasslands. However, they often refer only statistical calculations we transformed the indications of Braun- to small regions and are not related to one another. Unfortunately, Blanquet (1964) to dominance values in percent (according to their descriptions of plant communities do not focus on form and Tasser and Tappeiner, 2004): r = 0.1%, + = 0.3%, 1 = 2.8%, 2m = 4.5%, intensity of land use, even though plant communities are predomi- 2a = 10%, 2b = 20.5%, 3 = 38%, 4 = 63%, and 5 = 88%. nantly the result of differences in land-use management (Burnside Land use was divided into three main groups (Table 1): (1) et al., 2007; Studer, 2001). It is therefore about time that an inven- meadows, which were subdivided into (a) unfertilized mountain tory which characterizes the remaining grassland plant communi- meadows (UM) – mown every year or infrequently every second ties is created, in order to have a database for prospective decisions. to third year, (b) fertilized mountain meadows (MM) – mown once With a huge and unique data set of vegetation relevés in the East- a year, but grazed by animals before and/or after mowing and (c) ern Alps, the present study aims at (1) providing a general overview fodder meadows (FM) – mown two to five times a year, mostly of of the man-made grassland plant communities and their distribu- valley regions; (2) pastures, which were subdivided into (a) lightly tion with a focus on montane to alpine regions, (2) clarifying how used pastures (LP) with unattended grazing animals during the veg- land use and site factors affect their establishment and (3) find out etation period and (b) intensively used pastures (IP) in fenced sites the most species-rich communities and how to maintain them. or near stables with high stocking of grazing animals; (3) young abandoned areas (AA), which were formerly mown and having lain fallow for not more than 30 years; older abandoned areas were Materials and methods excluded, and the year of the last mowing was based either on literature or interviews. Research area In alpine grasslands a smooth transition from fertilized meadows (mostly with cattle dung manuring) to meadows without To obtain a cross-section of meadows through the Eastern Alps, fertilization, abandoned areas and pastures of different grazing vegetation relevés were taken from Tyrol (Austria) and South Tyrol intensities is observed. Therefore, land-use intensity (LUi) was (Italy) together with a number of further relevés from western classified according to Tappeiner et al. (1998): For meadows we Vorarlberg, eastern Salzburg (both Austria) and northern Trentino summed every human impact Ih (mowing, fertilization) and divided (Italy). The location of the region lies between 47◦36–46◦02 N it by the frequency of these interferences in years a. The same pro-

22 C. Lüth et al. / Flora 206 (2011) 433–443 435

Fig. 1. Location of the research area Tyrol (Western Austria) and South Tyrol (Northern Italy) and the sites of the vegetation relevés. Dark grey lines: rivers and valleys, light grey areas: limestone or calcareous mica schist regions, white areas: silicate regions. cedure was applied for abandoned areas: thereby we summed the set and homogenized it; thereby the process of reviewing consisted years from the last human impact until now and took the reciprocal of the following steps: value of it. −1 1. The species names had to be standardized, as the incorporated LUi = Ih × a data from literature spanned a period of more than 50 years (e.g. The pasture utilization was indicated as ordinal data type, rang- Thimm, 1950, in Appendix), using the nomenclature of Fischer ing from 1 (low grazing intensity) to 4 (intensive grazing) (for et al. (2005). Taxa of species and subspecies mostly had to be details see Tasser et al., 1998). Thereby scaling is similar to the met- contracted to aggregates, if they had not been precisely classiﬁed ric data of the land-use intensity for meadows and abandoned areas by each author. This resulted in a data set of 769 species. (i.e. high grazing intensity has a similar impact as intensively used 2. Syntaxonomic names of plant communities also had to be stan- meadows with 4 impacts per year). dardized, based on Grabherr and Mucina (1993) and Mucina et al. (1993); three newer association names belong to Ellenberg Data analysis (1996), Wallossek (1999) and Grabner and Heiselmayer (2002).

The vegetation relevés from the literature were already clas- We then grouped our own recorded vegetation relevés with a siﬁed by the corresponding authors (Appendix). As single authors hierarchical cluster analysis and integrated them into the data set classify communities in a slightly different way (due to subjectivity from literature by comparing their species composition. Afterwards of estimating the abundance values), we reviewed the whole data we used a discriminant analysis (DA) to obtain an automated clas-

Table 1 Main land-use types of grasslands in the Eastern Alps, their respective abbreviation codes (LU), number of relevés, land-use intensity quotient (LUi), site factors (altitude, slope angle and pH of the topsoil) and number of species.

LU Name and land-use No. of relevés LUi Altitude Slope angle (◦) pH (0–10 cm) No. of species type (m a.s.l) mean±s.d. mean ± s.d. mean ± s.d. mean ± s.d.

UM Unfertilized mountain 560 0.33–1 1797 ± 223 18.8 ± 11.3 5.07 ± 0.67 39.8 ± 13.0 meadows, mown every year, seldom every 2nd or 3rd year MM Fertilized mountain 260 2–3 1737 ± 260 13.9 ± 10.2 5.37 ± 0.49 31.9 ± 11.3 meadows, mown once a year and mostly grazed in autumn FM Fodder meadows, 364 4–7 1246 ± 323 14.1 ± 11.8 5.83 ± 0.41 24.7 ± 6.9 mown two to ﬁve times a year LP Lightly used pastures, 335 1–2 2049 ± 303 20.5 ± 14.4 5.16 ± 0.86 31.3 ± 12.9 with unattended grazing animals IP Intensively used 86 3–4 1752 ± 386 11.9 ± 11.0 5.52 ± 0.63 21.9 ± 9.8 pastures, with high stocking of grazing animals AA Abandoned areas, 194 0.03–0.1 1866 ± 243 22.8 ± 11.2 4.87 ± 0.67 41.7 ± 12.2 abandoned for not more than 30 years

23 436 C. Lüth et al. / Flora 206 (2011) 433–443

Table 2 Classification result of the DA of 1882 relevés shown in %; predicted group: 39 plant communities (see Table 3), independent variables: abundance values of 544 species. Overall 88.3% of the relevés are correctly classified. Identification numbers of the communities (ID) given in Table 3. P r e d i c t e d c l a s s i f i c a t i o n o f p l a n t c o m m u n i t i e s o f t h e D i s c r i m i n a n t A n a l y s i s ID-No. 1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627282930313233343536373839 1100------2-100------3--100------4---90------1.41.4------7-- 5----82------9.1------9.1------6-----89------2.7-5.4-2.7 7------100------8------100------9------80------20------10------100------11------82------9.1-9.1 12------73-18---9.1------13------855.42.2--1.1------1.1------5.4-- 14------1.987---1.9------7.7-1.9 15------93------7.4-- 16------89------11-- 17------17--83------18------2.9-----79------2.915-- 19------98------2.5-- 20------92------4------4 21------87------13-- 22------75------25-- 23------4.8------86------4.84.8- 24------40------60-- 25------79--3.4------17 26------100------

Original classiﬁcaon of plant communies plant of classiﬁcaon Original 27------2.6-82---5.3------11 28------1.5------79------20 29------100------30------88------13-- 31------8614------32--9.1------73------18 33------100------34------9.1------91----- 35------83-11-5.6 36------0.7------1.4------0.7------915.8-- 37 - - - 0.4 - - - 0.4 0.4 - - - 0.4 0.2 0.2 0.6 - 2.8 ------0.2 - - - 0.2 94 - 0.2 38------6.7------1380- 39--0.5--0.2-0.2-----0.5---1.2-0.7----0.70.21---0.51.4---0.21.2-91

sification of all (1882) relevés, based on species composition. To relevés, because they just had short species lists, so that a clear reduce the noise from rare species, we included only species occur- classification to a defined plant community was not possible. ring more than four times in the data set or with more than 60% Since the number of relevés varies greatly between individ- abundance in a single community, giving an overall species pool of ual plant communities (Table 3), we used the mean abundance of 544 species. The DA explained 88.3% of all vegetation relevés (i.e. species belonging to a community for a new data set. This data 1662 of 1882 relevés) as correctly classified (Table 2), only 11.7% set was used in further nonmetric multidimensional scaling (NMS) were misclassified. to highlight the correlation between plant communities and the In the next step, the misclassified relevés calculated by the DA factors land-use intensity, altitude, slope and pH of the topsoil. were reviewed thoroughly by comparing the species composition As the pH is rarely indicated in literature, we calculated the miss- of these relevés with the typical species composition from the ing pH values by using the R-values (=soil reaction) of Ellenberg literature. As a consequence, we changed the classification of 83 et al. (1992) adapted for Austria by Karrer (Englisch et al., 1991; relevés, according to the predicted syntaxonomic community of Karrer and Kilian, 1990; Karrer, 1992; Pichler and Karrer, 1991). the DA. The reason of changing these relevés was the high abun- According to Warmelink et al. (2005) we calculated a mean R- dance values of species belonging to the predicted community value for each relevé on the basis of the species presented in the of the DA. Moreover, we could not detect in these relevés any relevé. A calibration curve was given for the measured pH-values character species given in literature, which confirmed the former of the 375 own recorded relevés, from which the missing pH- stated plant community. The classification of other 54 relevés values were afterwards predicted by linear regression. Thereby we was left unchanged, because the character species and differential achieved a highly significant correlation between R-value and soil species – even if their abundance values were low – confirmed the pH (R2 = 0.575, p < 0.001), which is even higher than the one stated original phytosociological classification. Finally, we eliminated 83 in Diekmann (2003).

24 C. Lüth et al. / Flora 206 (2011) 433–443 437

Table 3 Grassland plant communities of the Eastern Alps with main ecological and ﬂoristic characteristics and their respective identiﬁcation numbers (ID); Abbreviation codes of land-use types (LU) given in Table 1; bold letters describe the main land-use type of each community.

Id-No. Community name Land use Data set No. of No. of Altitude (m a.s.l) Slope angle (◦) pH (0–10 cm) No. of species types (LU) relevés sites mean ± s.d. mean±s.d. mean ± s.d. mean ± s.d.

1 Alchemillo-Poetum supinaeb IP Lit. & own 5 3 1428 ± 296 3.4 ± 3.8 5.79 ± 0.26 19.8 ± 5.9 2 Alnetum viridisa UM; LP Literature 7 1 1760 ± 23 27.6 ± 4.2 4.45 ± 0.16 30.0 ± 3.6 3 Angelico-Cirsietum oleraceib AA; LP Literature 6 1 1100 ± 0 2.2 ± 3.9 5.73 ± 0.12 20.3 ± 7.3 4 Campanulo AA; UM Lit. & own 71 13 2035 ± 153 27.2 ± 14.5 5.69 ± 0.46 44.2 ± 10.1 scheuchzeri-Festucetum noricaea 5 Caricetum curvulaea UM Literature 11 2 2349 ± 87 4.5 ± 4.2 3.63 ± 0.28 16.2 ± 7.4 6 Caricetum davallianaea AA; UM Lit. & own 37 13 1827 ± 209 11.9 ± 8.8 5.64 ± 0.44 28.9 ± 8.4 7 Caricetum ferrugineaea AA; UM; LP Literature 9 1 1903 ± 33 17.1 ± 5.7 5.54 ± 0.14 46.8 ± 3.1 8 Caricetum sempervirentisa UM Lit. & own 13 4 1953 ± 162 26.5 ± 13.1 4.43 ± 0.63 30.8 ± 6.8 9 Carici curvulae-Nardetuma UM; MM Literature 5 2 2064 ± 340 12.6 ± 8.1 4.22 ± 0.50 36.8 ± 11.8 10 Carlino acaulis-Brometumb LP; IP Literature 7 1 1699 ± 63 20.6 ± 14.9 5.81 ± 0.19 33.4 ± 5.7 11 Chaerophyllo-Ranunculetum UM; MM; Lit. & own 11 6 1577 ± 250 10.0 ± 5.5 5.42 ± 0.50 22.0 ± 7.0 aconitifoliib LP; IP 12 Crepido-Cynosuretumb AA; UM; Lit. & own 11 2 1301 ± 193 2.9 ± 4.6 6.35 ± 0.55 28.9 ± 7.4 MM; LP;IP 13 Crepido-Festucetum LP Lit. & own 93 19 1916 ± 181 14.6 ± 11.9 5.69 ± 0.49 27.6 ± 12.2 commutataeb Deschampsio cespitosae- 14 Poetum alpinaeb AA; UM; LP Lit. & own 52 22 1950 ± 170 10.5 ± 12.3 5.05 ± 0.57 23.5 ± 7.2 15 Elyno-Caricetum rosaea LP Literature 27 1 2483 ± 24 23.1 ± 13.6 5.63 ± 0.54 18.0 ± 8.7 16 Empetro-Vaccinietum AA; UM; Lit. & own 18 7 1826 ± 130 15.0 ± 8.9 4.46 ± 0.31 42.3 ± 12.8 gaultherioidisa MM; FM; LP 17 Festucetum picturataea AA; UM; LP Literature 6 3 2162 ± 107 15.0 ± 10.9 4.75 ± 0.46 31.0 ± 8.2 18 Festuco-Agrostietumc UM; MM; Lit. & own 34 17 1724 ± 250 20.0 ± 15.9 4.91 ± 0.41 32.3 ± 13.1 FM 19 Gentianello AA; UM Literature 40 2 2080 ± 103 22.6 ± 7.1 4.40 ± 0.31 39.5 ± 9.4 anisodontae-Festucetum variaee 20 Geranio lividi-Trisetetumb AA; LP Literature 25 3 1426 ± 147 22.7 ± 7.8 6.12 ± 0.41 41.2 ± 9.2 21 Gymnadenio-Nardetumb AA; UM Literature 15 1 1425 ± 52 9.7 ± 6.5 5.23 ± 0.40 33.6 ± 4.7 22 Hygrocaricetum curvulaea LP Literature 8 2 2440 ± 170 6.6 ± 6.9 3.78 ± 0.27 14.4 ± 7.7 23 Hypochoerido FM Lit. & own 21 6 2106 ± 133 18.9 ± 12.1 5.03 ± 0.39 43.7 ± 10.9 uniflorae-Festucetum paniculataea 24 Loiseloirio-Caricetum AA; UM Literature 5 1 2460 ± 0 7.6 ± 5.3 4.11 ± 0.25 17.6 ± 11.6 curvulaea 25 Lolietum multifloraeb MM; FM; IP Lit. & own 29 7 900 ± 207 7.9 ± 10.6 6.20 ± 0.24 15.8 ± 3.4 26 Onobrychido MM; FM Own 3 1 1240 ± 10 9.3 ± 5.1 6.60 ± 0.20 32.0 ± 2.6 viciifoliae-Brometumb 27 Pastinaco-Arrhenateretumb UM; LP Lit. & own 38 15 1026 ± 209 15.5 ± 15.4 6.17 ± 0.39 20.7 ± 5.3 28 Poo-Trisetetumb UM Lit. & own 65 10 954 ± 232 4.2 ± 2.1 5.66 ± 0.41 26.3 ± 5.1 29 Potentillo MM; FM; IP Own 10 4 1372 ± 157 18.4 ± 17.6 5.90 ± 0.36 35.4 ± 10.1 erectae-Brachypodietum pinnatib 30 Pulsatillo alpinae-Festucetum FM; IP Literature 8 1 1832 ± 31 23.8 ± 5.2 5.68 ± 0.27 44.5 ± 6.6 noricaed 31 Ranunculo LP Literature 7 1 1100 ± 0 26.1 ± 10.9 5.81 ± 0.28 25.7 ± 6.8 bulbosi-Arrhenatheretumb 32 Ranunculo-repentis- MM; IP Lit. & own 11 5 1048 ± 135 3.4 ± 4.8 5.92 ± 0.21 17.0 ± 4.5 Alopecuretum pratensisb 33 Rhododendretum ferrugineia AA;UM; Literature 4 2 1918 ± 223 17.5 ± 15.5 4.64 ± 0.44 28.2 ± 12.3 MM 34 Rumicetum alpinia AA;UM; Literature 11 3 2047 ± 197 8.4 ± 6.6 5.43 ± 0.86 12.5 ± 3.5 MM; LP 35 Selino-Molinetum caeruleaeb AA; UM; Literature 18 2 1486 ± 195 7.1 ± 3.0 5.61 ± 0.50 23.1 ± 12.7 MM; LP 36 Seslerio-Caricetum AA; UM; LP Lit. & own 139 15 1817 ± 252 28.4 ± 10.1 6.06 ± 0.56 44.8 ± 12.9 sempervirentisa 37 Sieversio-Nardetum strictaea UM; MM; Lit. & own 468 74 1876 ± 219 17.6 ± 8.9 4.86 ± 0.62 38.7 ± 12.8 FM; LP; IP 38 Trifolio thalii-Festucetum IP Literature 30 8 2016 ± 182 35.7 ± 17.3 5.56 ± 0.44 38.1 ± 12.5 nigricantisa 39 Trisetetum flavescentisb UM; LP Lit. & own 421 58 1559 ± 289 17.55 ± 11.6 5.72 ± 0.50 29.9 ± 10.1

a Grabherr and Mucina (1993). b Mucina et al. (1993). c Ellenberg (1996). d Grabner and Heiselmayer (2002). e Wallossek (1999).

25 438 C. Lüth et al. / Flora 206 (2011) 433–443

Arrhenatheretum (No. 27), which is typical for fertilized meadows mown 2–3 times a year, does also establish in meadows mown once a year with pasturing in autumn and in intensively used pastures. Finally the Trisetetum flavescentis (No. 39) occurs in every land-use type except in abandoned areas. Relating to the mean number of species, biodiversity is mainly affected by land use, but is also affected to a lesser extent by altitude and bedrock. Consequently, the most species-rich communities are found in mainly unfertilized mountain meadows, mown once a year or less frequently every second to third year as well as in rather young abandoned areas (Nos. 4, 7, 23, 30, 36). As a rule, these communities occur on limestone or calcareous mica schist bedrock (Fig. 3), which explains a pH of the topsoil of >5. On the other hand, we found lowest biodiversity in fertilized meadows, intensively used pastures and high alpine lightly used pastures above silicate bedrock (pH < 3.8), where just a few species (e.g. Carex curvula) are able to create a close vegetation cover (Nos. 5, 22, 25, 32, 34). The result of the NMS classification recommends a two- dimensional solution (Fig. 2) with a final stress of 15.5% for all 39 plant communities. Land-use intensity and pH have an inverse relationship to altitude and slope, whereby among all factors slope has the lowest influence on plant communities of all site factors. Referring to site factors, the 39 plant communities can be characterized as follows (Fig. 2). Highly fertilized grasslands with either frequent mowing per year (mostly three to five times) or intensively used pastures are concentrated in the valleys (Nos. 1, 3, 25, 32, 34) or near flat plateaus. Due to fertilization the soil offers high pH val- Fig. 2. NMS output graph of 39 communities and correlation between the fac- ues. Furthermore, plant communities mown twice a year (Nos. 27, tors land-use intensity (LUi), altitude, pH and slope: arrows indicate direction and strength of correlation of variables with R2 = 0.2. Identification numbers of the com- 28, 31) are clearly located at higher altitudes than intensively used munities (ID) given in Table 3. Abbreviations of main land-use types given in Table 1. ones, but at lower altitudes than lightly used grasslands. Fertilized meadows, mown once a year and grazed in autumn, when cattle The NMS calculation was performed with the slow and thorough return from the high alpine pastures occur in a mountain range autopilot mode (for parameter settings and options see McCune between 1300 m and 2000 m a.s.l. (e.g. Nos. 11, 20, 39). Unfertil- and Mefford, 1999) by using the Relative Sørensen distance mea- ized meadows were found in the northern limestone Alps at the sure; the cutoff R2 value for site factors was set at 0.2. For the lowest altitude of about 1200 m a.s.l. (No. 26, Fig. 3). Their species diagram (Fig. 2) the stress in relation to dimensionality is stated, composition is very similar to fertilized grasslands from the valleys, which visualizes the departure from monotonicity in the plot of but is distinguished by the absence of species that occur with the distance – the lower the stress, the better the fit to monotonicity use of fertilizer. In the central massive the unfertilized meadows (McCune and Grace, 2002). spread up to 2000 m a.s.l. Consequently, communities growing on Calculations were done with SPSS 15.0.1 (SPSS Inc., 2006) and calcareous bedrock (Nos. 6, 7, 10, 21, 29, 30) stand out from those PC-ORD 5.01 (McCune and Mefford, 1999). Map drawings were on siliceous or mica schist bedrock (Nos. 4, 18, 37). Locations at plotted in ArcView 3.3 (ESRI Inc., 2002). high altitudes in the Eastern Alps, where grazing is possible, are characterized by pastures with the dominant taxon Carex curvula Results s.l. (Nos. 5, 9, 15, 22, 24). Abandoned areas are located on the steepest sites, with a mean slope of 22.5◦ (Table 1). They can be divided Phytosociological communities into hydrophilous abandoned areas on calcareous bedrock (No. 35), herb-rich ones (Nos. 8, 23, 36, 38) and abandoned areas domi- On the basis of floristic composition we distinguished 39 com- nated by shrubs (No. 2), which are similar to pastures dominated munities, listed in Table 3 with a summary of their main ecological by shrubs (No. 33). and floristic characteristics. We could detect that 23 communities occur in unfertilized Distribution patterns of plant communities meadows: ten of them were established mainly in unfertilized meadows and three (Nos. 7, 10, 30) were restricted to this land- The large data set of 1799 vegetation relevés provides detailed use type. Furthermore we found that some communities, stated information about distribution patterns of current grassland com- in literature as pasture communities, were not just restricted to munities over a broad cross-section of the Eastern Alps (Fig. 3). grazed sites: The Rumicetum alpini (No. 34) occurs also on fertilized Besides fertilized grassland communities from the valleys, the most meadows mown once a year and grazed in autumn. This is similar frequent plant communities of mountain grasslands are the usually to the Crepido-Festucetum commutatae (No. 13), which addition- fertilized Trisetetum flavescentis (No. 39, 421 relevés) and the usually establishes also in unfertilized meadows. The Deschampsio ally unfertilized Sieversio-Nardetum strictae (No. 37, 468 relevés). cespitosae-Poetum alpinae (No. 14) occurs in every type of land Some communities are locally restricted (Fig. 3), based on the fact use except in fodder meadows from valley regions. On the other that some species only grow on specific bedrocks and the distribu- hand, we found meadow communities which occur in pastures: tion area of some species is peripherally situated in the research The Angelico-Cirsietum oleracei (No. 3) occurs in pastures and the area: Ranunculo repentis-Alopecuretum pratensis (No. 32) in fertilized The Crepido-Cynosuretum (No. 12) occurs only in northern meadows as well as in intensively used pastures. The Pastinaco- limestone pastures. The Onobrychido viciifoliae-Brometum (No.

26 C. Lüth et al. / Flora 206 (2011) 433–443 439

Fig. 3. Distribution pattern of plant communities. Dark grey lines: rivers and valleys, light grey areas: limestone or calcareous mica schist regions, white areas: silicate regions. Identiﬁcation numbers of the communities (ID) given in Table 3.

26) is located in the northern limestone alps, in montane altitude Classification zones (at about 1200 m a.s.l.), similar to the Potentillo erectae- Brachypodietum pinnati (No. 29), which is found in steep slopes The identified 39 grassland communities were clearly classi- of Northern Tyrol in unfertilized, annually mown meadows. The fied by the DA from their plant composition. Thereby mismatches Seslerio-Caricetum sempervirentis (No. 36) is found in limestone in the diagonal matrix (Table 2) are explained by the absence or calcareous mica schist regions over the entire Eastern Alps. of character species or incomplete species lists in single relevés. The Campanulo scheuchzeri-Festucetum noricae (No. 4), how- Most of these mismatches belong to the Sieversio-Nardetum stric- ever, is specific for the southern limestone and mica schist areas. tae. The Sieversio-Nardetum strictae is one of the dominant plant The Festucetum picturatae (No. 17), the Gentianello anisodontae- communities in alpine pastures and meadows, with a wide range Festucetum variae (No. 19) which occurs in the far south, the of growing conditions (Grabherr and Mucina, 1993). Therefore, a Hypochoerido uniflorae-Festucetum paniculatae (No. 23), the Tri- broad spectrum of species can be found, for which species with folio thalii-Festucetum nigricantis (No. 38), as well as the Pulsatillo regional restrictions are responsible. This is similar to the mis- alpinae-Festucetum noricae (No. 30) are endemic in South Tyrol or matched relevés of the Trisetetum flavescentis. The community is East Tyrol, because the centers of distribution of the eponymous the second most common grassland community in the Eastern Alps taxa Festuca picturata, Festuca varia, Festuca paniculata, Festuca and displaces the Sieversio-Nardetum strictae in middle mountain nigricans and Festuca norica are in southern and/or south-eastern range levels (Mucina et al., 1993). The community becomes estab- regions of the Eastern Alps. lished when grasslands are fertilized with manure once a year and mown once or twice a year, but grazed in autumn after the return of livestock from high alpine summer pastures (Knapp and Knapp, Discussion 1952; Mucina et al., 1993); in certain cases the community also occurs in lightly to intensively used pastures. In the Alps the geomorphology forces man to many different The link between fertilized and unfertilized mountain meadows types of agricultural land use, resulting in a high number of grass- is represented by the Festuco-Agrostietum (Ellenberg, 1996), which land plant communities (Maurer et al., 2006). The communities was also found in nearly all land-use types (Table 3). Furthermore, reflect the anthropogenic influence and topographic, edaphic and we found that the communities Crepido-Festucetum commutatae, climatic factors (Marini et al., 2007) and they show the dependency Rumicetum alpini and Deschampsio cespitosae-Poetum alpinae, on those factors, which have a discriminative influence on them. which are associations of pasture alliances (Grabherr and Mucina,

27 440 C. Lüth et al. / Flora 206 (2011) 433–443

1993; Mucina et al., 1993), are not just restricted to pastures, but be found in pastures with intensive grazing or in meadows and occur also in (predominantly fertilized) meadows. On the contrary abandoned areas above calcareous bedrock. On the other hand, low we observed the Angelico-Cirsietum oleracei, usually described as pH-values (about 3.5–5.5) occur in unfertilized grasslands above a meadow community with upright forbs (Mucina et al., 1993), silicate substrate. but here observed as a hydrophilous pasture community. We also identified the communities Pulsatillo alpinae-Festucetum noricae, Carlino acaulis-Brometum and Caricetum ferruginae belonging Plant diversity exclusively to unfertilized meadows. The same land-use type is necessary for the Caricetum davallianae, Caricetum sempervirentis, We found grassland communities which refer to specific taxa Gymnadenio-Nardetum and Hypochoerido uniflorae-Festucetum (e.g. of the genus Festuca sp.), but with local restriction (Pils, 1980) paniculatae. When the mowing is abandoned these communities or belonging to a specific bedrock (Fischer et al., 2005), repre- can still be found in young abandoned areas, but they disappear in sented by the communities Campanulo scheuchzeri-Festucetum older ones. noricae, Gentianello anisodontae-Festucetum variae, Pulsatillo alpinae-Festucetum noricae, Trifolio thalii-Festucetum nigricantis Impact of environmental factors and Hypochoerido uniflorae-Festucetum paniculatae. These rarely occurring communities are characterized by high biodiversity and Key factor in forming specific grassland plant communities in low land-use intensity (Table 3). The most species-rich meadows meadows and pastures is human influence. The higher the human are characterized by soil of high pH (Myklestad and Sætersdal, impact, the fewer species and community types occur. In favourable 2004) and do also apply to unfertilized mountain meadows above areas (flat areas in lowlands) low diversity is a result of mow- calcareous bedrock in the Eastern Alps, represented by the Carice- ing several (3–5) times a year (Marini et al., 2007; Oberdorfer tum ferruginei, the Campanulo scheuchzeri-Festucetum noricae and Müller, 1993), which allows only those species to grow that and the Seslerio-Caricetum sempervirentis (Table 3). have either very short life cycles or can clonally reproduce (Bahn According to the mean number of species we found that species- et al., 1994). This is similar to highly stocked and fertilized pastures rich communities must not be related in any case to altitude and pH that result in grassland communities with a few species that resist of the topsoil. However, land-use intensity and slope significantly trampling (Ellenberg, 1996). A strong dependency on land-use affect the number of species and correlate reciprocally to each intensity is also revealed in fertilized meadows mown twice a year. other: species-poor communities occur in flat areas and with high Since accessibility is one of the major factors in fertilizing grass- anthropogenic influence, whereas on steep sites land-use intensity lands (Macdonald et al., 2000; Tasser and Tappeiner, 2002), poorly is low and communities provide high vascular plant biodiversity, accessible unfertilized meadows form another group of meadow which is confirmed by Niedrist et al. (2008). communities. Traditional hay meadow management preserves sites with high Due to the shorter growth period, decrease of summer temper- biodiversity (Garcia, 1992; Maurer et al., 2006; Myklestad and atures and increase of rainfall with increasing altitude (Ellenberg, Sætersdal, 2004), which has been confirmed by our studies. The 1996; Körner, 2007), the altitude site factor plays a decisive highest number of species and the highest ecosystem diversity are role in forming specific communities. Especially in unfertilized found in unfertilized alpine meadows and young abandoned areas meadows the high number of 23 different plant communities (Tables 1 and 3). is explained by the extensive amplitude of elevation, which Since agri-environmental measures are among the most impor- ranges from about 1250 to 2500 m a.s.l. Here we found the tant instruments for the promotion of environmentally adapted Onobrychido-viciifoliae-Brometum and the Potentillo erectae- agricultural land use (Matzdorf et al., 2008), a register of the still Brachypodietum pinnati from valley regions depending on low remaining grassland communities is essential for such measures land-use intensity and the absence of fertilization (Mucina et al., in the Eastern Alps. Moreover, unfertilized mountain meadows 1993). In the highest located meadows the communities Cam- are nowadays increasingly abandoned (Macdonald et al., 2000; panulo scheuchzeri-Festucetum noricae and Sieversio-Nardetum Tasser et al., 2001, 2007), nevertheless they play an important role strictae establish. In lightly stocked pastures and abandoned in the culture of the Alps mountain regions (Ender and Grabner, areas the species composition is similar to unfertilized mead- 1997). Therefore, currently unfertilized meadows must be retained ows, as long as low anthropogenic influence and similar site in order to ensure that rare communities – being endangered – factors affect the areas. The grassland communities at highest alti- will continue to be found today in the Eastern Alps and also in the tudes can be separated clearly by means of species, in that only future. few specific taxa (e.g. Carex curvula) generate a close vegetation cover (Choler and Michalet, 2002), and if at all, they are used as pastures. Acknowledgements Slope is found to be relevant in meadows and abandoned areas. Meadow communities on steep slopes present a species compo- We thank Prof. Dr. Robert Crawford, Dr. Avril Arthur-Goettig sition similar to xerophilous grasslands, due to drainage (Seeber (BioScript International) and Daniela Dellantonio for language and Seeber, 2005). However, on steep slopes land use is no longer revision. We also thank the provincial government of Tyrol profitable and grasslands are nowadays often abandoned or infre- for analyzing the soil samples. This work was funded by the quently managed (Giupponi et al., 2006; Tasser et al., 2007), which European Union within the scope of the Interreg IIIA-Project “DNA- explains why abandoned areas were found on steepest sites. Near Characterisation for certification and valorisation of mountain and below the timberline abandonment leads to an increase of hay”, supported by the provincial governments of Tyrol and South shrubs which displace herbs, represented by the Alnetum viridis, Tyrol. the Empetro-Vaccinietum gaultherioidis or the Rhododendretum ferruginei. The pH value of the topsoil is affected by either fertilizing or Appendix. by bedrock. High pH values result from the absence of humic acids (Marcos et al., 2003), due to intensive fertilizing (Seeber and Seeber, Sources for information on the vegetation relevés from litera- 2005). Apart from fertilizing, higher pH-values (about 5.5–7.2) can ture.

28 C. Lüth et al. / Flora 206 (2011) 433–443 441

Author Title Type of publication Year of publication Location No. of relevés

Brunner, B. Die Vegetation von Diploma thesis, University 1999 Obernberg, North Tyrol 118 Bergmähdern im of Innsbruck (AT) Landschaftsschutzgebiet Nößlachjoch-Obernberg- Tribulaune Dalla Torre, M. Die Vegetation der Dissertation, University of 1982 St. Christina, Gröden, 11 subalpinen und alpinen Innsbruck (AT) South Tyrol Stufe in der Puez-Geisler Gruppe Dierschke, H. Grünland-Gesellschaften im Phytocoenologia 6: 1979 Galtür, Paznaun, North 54 oberen Paznauner Tal 287–303 Tyrol Dirrhammer, H. Die Vegetation im oberen Diploma thesis, University 2008 Elbigenalp, Lechtal, 18 Lechtal of Innsbruck (AT) North Tyrol Duelli, M. Die Vegetation des Dissertation, University of 1977 Gaißbergtal, Obergurgl, 27 Gaißbergtales. Ein Versuch, Innsbruck (AT) North Tyrol das Datenmaterial mit Hilfe der EDV-Anlage zu bearbeiten Ebner, S. Die Wirtschaftswiesen des Diploma thesis University 1996 Vinschgau, South Tyrol 73 oberen Vinschgaues of Innsbruck (AT) (Südseite) und ihre Bewirtschaftung Egger, G. Die Vegetation vom Personal unpubl. relevés 2006 Hohe Tauern, East 86 Nationalpark Hohe Tauern Tyrol in Tirol Ender, M. Vegetation von gemähten Diploma thesis, University 1997 Tannberg, Vorarlberg 63 Bergwiesen und deren of Innsbruck (AT) Sukzession nach Auflassung der Mahd Flecker, K. Die Vegetation von Diploma thesis, University 1996 Lech, Vorarlberg 29 Schipisten und of Innsbruck (AT) angrenzenden Bergmähdern im Raum Hochtannberg Florian, C. Die Lärchenwiesen im Diploma thesis University 1995 Bozner Unterland, 35 Nationalpark Trudner Horn of Innsbruck (AT) South Tyrol – Pflanzensoziologische Untersuchungen in verschieden bewirtschafteten Wiesen und deren Vergleich mit aufgelassenen Flächen Gander, M. Die alpine Vegetation des Diploma thesis University 1984 Defreggental, East 13 hinteren Defreggentales of Innsbruck (AT) Tyrol Grabner, S., Heiselmayer, P Diversity of mountain Razprave IV. Razreda Sazu 2002 Virgental, East Tyrol 8 meadows in the inner alpine 43/3: 167–184 valley Virgental/Eastern Tyrol Gufler, R. Analyse der Vegetations- Diploma thesis, University 1999 Kasertatt, Stubaital, 28 und Erosionsverteilung in of Innsbruck (AT) North Tyrol Abhängigkeit von Bewirtschaftungsänderun- gen am Beispiel Kaserstattalm Gumpelmayer F. Die Vegetation und ihre Dissertation, University of 1967 Leoganger Steinberge, 10 Gliederung in den Innsbruck (AT) Salzburg Leoganger Steinbergen Hellriegl, S. Wirtschaftswiesen der Diploma thesis University 1996 Vinschgau, South Tyrol 77 Nordhänge und Tallagen im of Innsbruck (AT) oberen Vinschgau aus vegetationskundlicher und futterbaulicher Sicht Keim, K. Die Vegetationsverhältnisse Dissertation, University of 1967 Pflersch, South Tyrol 26 des Pflerschertales Innsbruck (AT) Kirchmeir, H. Auswirkungen des Diploma thesis University 1996 Oberes Gericht, North 19 Pistenschilaufes auf die of Vienna (AT) Tyrol Pflanzengesellschaften der Komperdellalm (Tirol) Lechner, C. Die Vegetation im Bereich Diploma thesis, University 1995 Nauders, Noth Tyrol 31 des Dreiländerecks bei of Innsbruck, (AT) Nauders am Reschenpass Lechner, G. Die Vegetation der inneren Dissertation, University of 1969 Pfunders, South Tyrol 26 Pfunderer Täler Innsbruck (AT) Lüth, Ch. Vegetation der Personal unpubl. relevés 2002 Trafoital, South Tyrol 2 Wirtschaftswiesen von Trafoi am Stilfser Joch

29 442 C. Lüth et al. / Flora 206 (2011) 433–443

Appendix A (Continued )

Author Title Type of publication Year of publication Location No. of relevés

Mayer, Ch. Landschaftsentwicklung in Diploma thesis University 2004 Passeiertal, South Tyrol 13 der Gemeinde St. Leonard in of Innsbruck (AT) Passeier (Südtirol, Italien) unter besonderer Berücksichtigung der floristischen Biodiversität und der Bodendurchwurzelung Mayer, R. Die Vegetation der Diploma thesis, University 2002 Valsertal, Noth Tyrol 118 Bergmähder im Valsertal of Innsbruck (AT) und ihre Dynamik Meurer, M. Die Vegetation des Grödner Giessener geogr. Schriften 1980 Grödnertal, South Tyrol 7 Tales/Südtirol 47: 287 Mulser, J. Analyse der Diploma thesis University 1998 Walten, Passeiertal, 40 Vegetationsverteilung in of Innsbruck (AT) South Tyrol Abhängigkeit der Bewirtschaftungsänderung auf den Waltner Mähdern Nieder-brunner, F. Vegetation der Sextener Dissertation, University of 1975 Sexten, South Tyrol 11 Dolomiten (subalpine und Innsbruck (AT) alpine Stufe) Noichl, M. Vegetationskundliche Diploma thesis University 1997 Leukental, North Tyrol 50 Untersuchung, Kartierung of Innsbruck (AT) und Bewertung der Kulturlandschaft Kitzbühel-Aurach Oberhammer M. Die Vegetation der alpinen Dissertation, University of 1979 Prags, Dolomites, South 14 Stufe in den östlichen Innsbruck(AT) Tyrol Pragser Dolomiten Putzer, J. Pflanzengesellschaften im Dissertation, University of 1967 Brixen, South Tyrol 8 Raum von Brixen mit Innsbruck (AT) besonderer Berücksichtigung der Trockenvegetation Raffl, E. Die Vegetation der alpinen Diploma thesis University 1982 Vinschgau, 18 Stufe der Texelgruppe of Innsbruck (AT) Texelgruppe, South Tyrol Smettan, H. Die Pflanzengesellschaften Dissertation, University of 1981 Kaisergebirge, North 6 des Kaisergebirges/Tirol Innsbruck (AT) Tyrol Steinmair, V. Die Vegetation von Diploma thesis, University 1999 Plätzwiese, Prags, 92 unterschiedlich genutzten of Innsbruck (AT) South Tyrol Almflächen auf der Plätzwiese Tasser, E. Die Vegetation der Personal unpubl. relevés 2004 Stubaital, North Tyrol 26 Talwiesen des Stubaitales Tasser, E. Vegetationsaufnahmen von Personal unpubl. relevés 2005 Mt. Bondone, Trentino, 22 Bürstlingsrasen des Monte Italy Bondone Thimm, I. Die Vegetation der alpinen Dissertation, University of 1950 Rofan, North Tyrol 20 und subalpinen Stufe des Innsbruck (AT) Sonnwendgebirges Thomaser, J. Die Vegetation des Veröff Museum 1967 Gardertal, South Tyrol 4 Peitlerkofels in Südtirol. Ferdinandeum 47: 67–119 Unterhofer, Ch. Welche Landschaftsskala Diploma thesis University 2006 Pfunderertal, South 55 eignet sich zur Klärung der of Innsbruck (AT) Tyrol Fließgewässerqualität? Unterlug-gauer, P. Die Vegetation in Vent und Diploma thesis University 2003 Ötztal, North Tyrol 65 Rofen (Ötztal, Tirol) of Innsbruck (AT) Vorhauser, K. Vegetationskundliche Diploma thesis University 1998 Eggenberg, South Tyrol 116 Untersuchungen im Bereich of Innsbruck (AT) der Eggentaler Alm (Südtirol) Wallossek, C. Vegetationskundlich- Dissertationes Botanicae 1990 Lafatscher Joch, 27 ökologische 154 (IT) Latemar, South Tyrol Untersuchungen in der alpinen Stufe am SW-Rand der Dolomiten Winkler, J. Populationsbiologische Diploma thesis University 1992 Ahrntal, Hasental, 41 Untersuchungen an zwei of Innsbruck (AT) South Tyrol eng verwandten Sippen in alpinen Rasen

30 C. Lüth et al. / Flora 206 (2011) 433–443 443

References McCune, B., Grace, J.B., 2002. Analysis of Ecological Communities. MjM Software Design, Glenden Beach. Bahn, M., Cernusca, A., Tappeiner, U., Tasser, E., 1994. Wachstum krautiger Arten auf McCune, B., Mefford, M.J., 1999. PC-ORD. Multivariate Analysis of Ecological Data, einer Mähwiese und einer Almbrache. Verh. Ges. Ökol. 23, 23–30. Version 5.01. 1995–2005. MjM Software Design, Glenden Beach. Bakker, J.P., Berendse, F., 1999. Constraints in the restoration of ecological diversity Mottet, A., Ladet, A., Coqué, N., Gibon, A., 2006. Agricultural land-use change and in grassland and heathland communities. Trends Ecol. Evol. 14, 63–68. its drivers in mountain landscapes: a case study in the Pyrenees. Agric. Ecosyst. Bischof, N., 1981. Pflanzensoziologische Untersuchungen von Sukzessionen aus Environ. 114, 296–310. gemähten Magerrasen in der subalpinen Stufe der Zentralalpen. Beitr. Geobot. Mucina, L., Grabherr, G., Ellmauer, T., 1993. Die Pflanzengesellschaften Österreichs, Landesaufn. Schweiz 60, 1–127. Teil 1, Anthropogene Vegetation. G. Fischer, Jena, Stuttgart, New York. Bögel, H., Schmidt, K., 1976. Kleine Geologie der Ostalpen. Ott, Thun. Niedrist, G., Tasser, E., Lüth, Ch., Dalla Via, J., Tappeiner, U., 2008. Plant diver- Braun-Blanquet, J., 1964. Pflanzensoziologie. Grundzüge der Vegetationskunde, 3 sity declines with recent land use changes in European Alps. Plant Ecol. 202, Aufl. Springer, Wien. 195–210. Burnside, N.G., Joyce, C.B., Puurmann, E., Scott, D.M., 2007. Use of vegetation classi- Myklestad, Å., Sætersdal, M., 2004. The importance of traditional meadow manage- fication and plant indicators to assess grazing abandonment in Estonian coastal ment techniques for conservation of vascular plant species richness in Norway. wetlands. J. Veg. Sci. 18, 645–654. Biol. Conserv. 118, 133–139. Choler, P., Michalet, R., 2002. Niche differentiation and distribution of Carex curvula Oberdorfer, E., Müller, T., 1993. Süddeutsche Pflanzengesellschaften, Pt. 3, 3. Aufl. G. along a bioclimatic gradient in the southwestern Alps. J. Veg. Sci. 13, 851–858. Fischer, Stuttgart, New York. Diekmann, M., 2003. Species indicator values as an important tool in applied plant Oomes, M.J.M., 1992. Yield and species density of grasslands during restoration ecology – a review. Basic Appl. Ecol. 4, 493–506. management. J. Veg. Sci. 3, 271–274. Dullinger, S., Dirnböck, T., Greimler, J., Grabherr, G., 2003. A resampling approach for Pavlu,˚ V., Hejcman, M., Pavlu,˚ L., Gaisler, J., Nezerková,ˇ P., Andaluz, M.G., 2005. Vege- evaluating effects of pasture abandonment on subalpine plant species diversity. tation changes after cessation of grazing management in the Jizerské Mountains J. Veg. Sci. 14, 243–253. (Czech Republic). Ann. Bot. Fenn. 42, 343–349. Ellenberg, H., Weber, H.E., Düll, R., Wirth, V., Werner, W., Paulißen, D., 1992. Zeiger- Pichler, F., Karrer, G., 1991. Comparison of different ecological indicator value werte von Pflanzen in Mitteleuropa, 3. erweiterte Auflage. Goltze, Göttingen. systems. In: Horvath, F. (Ed.), Poster Abstracts. 34th IAVS Symposium. Eger, Ellenberg, H., 1996. Die Vegetation Mitteleuropas mit den Alpen. 5. Aufl. Ulmer, Hungary, pp. 102–104. Stuttgart. Pils, G., 1980. Systematik, Verbreitung und Karyologie der Festuca violacea-Gruppe Ellmauer, T., 1996. Die Bedeutung von Wiesengesellschaften für Biodiversität und (Poaceae) im Ostalpenraum. Plant Syst. Evol. 136, 73–124. Naturschutz in Österreich. Verh. Zool. – Bot. Ges. Österr. 132, 13–60. Pykälä, J., 2000. Mitigating human effects on European biodiversity through tradi- Ender, M., Grabner, S., 1997. Vegetation von gemähten Bergwiesen und deren tional animal husbandry. Conserv. Biol. 14, 705–712. Sukzession nach Auflassung der Mahd im Tannberg-Gebiet (Vorarlberg, Öster- Scheffer, F., et al., 2002. Lehrbuch der Bodenkunde. 15. Aufl. Spektrum, Heidelberg- reich). BAL Gumpenstein 2–3, 117–122. Berlin. Englisch, M., Karrer, G., Wagner, H., 1991. Bericht über den Zustand des Waldbodens Seeber, J., Seeber, G.U.H., 2005. Effects of land-use changes on humus forms on alpine in Niederösterreich. Forstliche Bundesversuchsanstalt/Amt der Niederösterr, pastureland (Central Alps. Tyrol). Geoderma 124, 215–222. Landesregierung, Wien. Stöcklin, J., Bosshard, A., Klaus, G., Rudmann-Maurer, K., Fischer, M., 2007. Land- ESRI Inc., 2002. ArcView Gis 3.3. Environmental System Research Institute. nutzung und biologische Vielfalt in den Alpen. Vdf Hochschulverlag AG an der Fischer, M.A., Adler, W., Oswald, K., 2005. Exkursionsflora von Österreich, Liecht- ETH Zürich. enstein und Südtirol. 2. Aufl. Land Oberösterreich. Biologiezentrum der OÖ Streifeneder, T., Tappeiner, U., Ruffini, F.V., Tappeiner, G., Hoffmann, C., 2007. Landesregierung, Linz. Selected aspects of agro-structural change within the Alps – a comparison of Fliri, F., 1998. Naturchronik von Tirol, Oberpinzgau, Vorarlberg, Trentino. Beiträge harmonised agro-structural indicators on a municipal level in the Alpine Con- zur Klimatologie von Tirol. Universitätsverlag Wagner, Innsbruck. vention area. Rev. Geogr. Alp. 95, 27–52. Garcia, A., 1992. Conserving the species-rich meadows of Europe. Agric. Ecosyst. SPSS Inc., 2006. SPSS for Windows. Version 15.0.1. SPSS, Chicago. Environ. 40, 219–232. Studer, S., 2001. The influence of management on the floristic composition of hay Giupponi, C., Ramanzin, M., Sturaro, E., Fuser, S., 2006. Climate and land-use changes, meadows. Ph.D. thesis, Swiss Fed. Inst. of Technology ETH Zurich. biodiversity and agri-environmental measures in the Belluno province, Italy. Tappeiner, U., Tasser, E., Tappeiner, G., 1998. Modelling vegetation patterns using Environ. Sci. Policy 9, 163–173. natural and anthropogenic influence factors: preliminary, experience with a GIS Grabner, S., Heiselmayer, P., 2002. Diversity of mountain meadows in the inner based model applied to an Alpine area. Ecol. Model. 113, 225–227. alpine valley Virgental/Eastern Tyrol. Razprave IV. Razreda Sazu 43 (3), 169–184. Tappeiner, U., Gramm, D., Pecher, C., Tasser, E., Lintzmeyer, F., Marzelli, S., Tappeiner, Grabherr, G., Mucina, L., 1993. Die Pflanzengesellschaften Österreichs, Teil 2, Natür- G., 2008. Typology of the Alps based on social, economic and environmental liche waldfreie Vegetation. G. Fischer, Jena, Stuttgart, New York. aspects. Diamont 4, 1–209. Kampmann, D., et al., 2007. Mountain grassland biodiversity: impact of site condi- Tasser, E., Tappeiner, U., 2002. Impact of land use changes on mountain vegetation. tions versus management type. J. Nat. Conserv. 16, 12–25. Appl. Veg. Sci. 5, 173–184. Karrer, G., 1992. Österreichische Waldboden-Zustandsinventur. Teil VII: Vegetation- Tasser, E., Tappeiner, U., 2004. New model to predict rooting in diverse plant com- sökologische Analysen. Mitt. Forstl. Bundesversuchsanst. Wien 168, 193–242. positions. Ecol. Model. 185, 195–211. Karrer, G., Kilian, W., 1990. Standorte und Waldgesellschaften im Leithagebirge Tasser, E., Tappeiner, U., Mulser, J., Tappeiner, G., Mayr, V., Ostendorf, B., 1998. Revier Sommerein. Mitt. Forstl. Bundesversuchsanst. Wien 165, 1–244. Kausalanalyse von Vegetationsveränderungen im Gebirge in Abhängigkeit von Knapp, G., Knapp, R., 1952. Über Goldhafer-Wiesen (Trisetetum flavescentis) im Landnutzung und Standortfaktoren. Verh. Ges. Ökol. 28, 107–115. nördlichen Vorarlberg und im Allgäu. Landwirtsch. Jahrb. Bayern. 29, 239–256. Tasser, E., Tappeiner, U., Cernusca, A., 2001. Südtirols Almen im Wandel. Ökologische Kohler, F., Gillet, F., Gobat, J.-M., Buttler, A., 2004. Seasonal vegetation changes in Folgen von Landnutzungsänderungen. EURAC, Bozen. mountain pastures due to simulated effects of cattle grazing. J. Veg. Sci. 15, Tasser, E., Mader, M., Tappeiner, U., 2003. Effects of land use in alpine grasslands on 143–150. the probability of landslides. Basic Appl. Ecol. 4, 271–280. Körner, Ch., 2007. The use of ‘altitude’ in ecological research. Trends Ecol. Evol. 22, Tasser, E., Walde, J., Tappeiner, U., Teutsch, A., Noggler, W., 2007. Land-use changes 569–574. and natural reforestation in the Eastern Central Alps. Agric. Ecosyst. Environ. Krausmann, F., Haberl, H., Schulz, N.B., Erb, K.-H., Darge, E., Gaube, V., 2003. Land- 118, 115–129. use change and socio-economic metabolism in Austria – Part I: driving forces of Tasser, E., Ruffini, F.V., Tappeiner, U., 2009. An integrative approach for analysing land-use change: 1950–1995. Land Use Policy 22, 1–20. landscape dynamics in high diverse cultural and natural mountain areas. Land- Kremer, B.P., 1991. Wiesenblumen. G. Unzer, München. scape Ecol. 24, 611–628. Lavorel, S., Touzard, B., Lebreton, J.D., Clement, B., 1998. Identifying functional von Arx, G., Bosshard, A., Dietz, H., 2002. Land-use intensity and border structures as groups for response to disturbance in an abandoned pasture. Acta Oecol. 19, determinants of vegetation diversity in an agricultural area. Bull. Geobot. Inst. 227–240. ETH 68, 3–15. Liira, J., et al., 2008. Plant functional group composition and large-scale species Vonlanthen, C.M., Bühler, A., Veit, H., Kammer, P.M., Eugster, W., 2006. Alpine richness in European agricultural landscapes. J. Veg. Sci. 19, 3–14. plant communities: a statistical assessment of their relation to microclima- Macdonald, D.V., et al., 2000. Agricultural abandonment in mountain areas of tological, pedological, geomorphological and other factors. Phys. Geogr. 27, Europe: environmental consequences and policy report. J. Environ. Manage. 59, 137–154. 47–69. Wallossek, C., 1999. The acidophilous taxa of the Festuca varia group in the Alps: Marcos, E., Calvo, L., Luis-Calabuig, E., 2003. Effects of fertilization and cutting on new studies of taxonomy and phytosociology. Folia Geobot. Phytotaxon. 34, chemical composition of vegetation and soil of mountain heathlands in Spain. J. 47–75. Veg. Sci. 14, 417–424. Warmelink, G.W.W., Goedhart, P.W., van Dobben, H.F., Berendse, F., 2005. Plant Marini, L., Scotton, M., Klimek, S., Isselstein, J., Pecile, A., 2007. Effects of local factors species as predictors of soil pH: replacing expert judgement with measurements. on plant species richness and composition of Alpine meadows. Agric. Ecosyst. J. Veg. Sci. 16, 461–470. Environ. 119, 281–288. World Wildlife Fund, 2004. Die Alpen. Das einzigartige Naturerbe. World Wildlife Matzdorf, B., Kaiser, T., Rohner, M.S., 2008. Developing biodiversity indicator to Found Deutschland, Frankfurt. design efficient agri-environmental schemes for extensively used grassland. Zimmermann, N.E., Kienast, F., 1999. Predictive mapping of alpine grass- Ecol. Indic. 8, 256–269. lands in Switzerland: Species versus community approach. J. Veg. Sci. 10, Maurer, K., Weyand, A., Fischer, M., Stöcklin, J., 2006. Old cultural traditions, in addi- 469–482. tion to land use and topography, are shaping plant diversity of grasslands in the Alps. Biol. Conserv. 130, 438–446.

31 32 3. Plant diversity declines with recent land use changes in European Alps

Niedrist G., Tasser E., Lüth C., Dalla Via J., Tappeiner U. (2009):. Plant Ecology 202:195-210.

33 34 Plant diversity declines with recent land use changes in European Alps

Niedrist Georg · Tasser Erich · Lüth Christian · Dalla Via Josef · Tappeiner Ulrike

G. Niedrist (e-mail) · E.Tasser · U. Tappeiner European Academy of Bolzano / Bozen, Drususallee 1, 39100 Bozen, Italy C. Lüth · U. Tappeiner Institute of Ecology, University of Innsbruck, Sternwartestr. 15, 6020 Innsbruck, Austria J. Dalla Via Research Centre for Agriculture and Forestry Laimburg, Laimburg 6, 39051 Pfatten/Auer, Italy e-mail: [email protected]

1 Keywords: Land use intensification · Land abandonment · Alpine meadows · Discriminant analysis · Evenness index

Abstract

Against a background of increasing land use intensification on favorable agricultural areas and land abandonment on less arable areas in the Alps, the aim of this investigation was to detect whether and how ten differently used types of grassland can be distinguished by site factors, plant species composition and biodiversity. By using a very large number of vegetation surveys (936) that were widely distributed in the Central Alps, site parameters and species composition of the different land use types were compared by discriminant analyses and various biodiversity indices. Results showed that land use is a significant factor affecting the development of different grassland communities with site factors playing a subordinate, yet still important role. The ten land use types studied can be clearly differentiated from one another by single species as well as by species composition. Our study found that the number of plant communities along with the number of species decreases constantly and significantly with increasing land use intensity and on abandoned land. For example, on average, extensively used meadows have more than three times as many species as intensively used meadows. Further, the most even distribution of species (Evenness index) is reached in intensively used meadows, whereas on pastures and abandoned land, some species become dominant forcing other species to recede. The results confirm that due to current trends in agriculture, such as land abandonment and land use intensification, plant diversity in the Alps is decreasing considerably.

2 Introduction

For decades, the agriculture in European mountain areas has been undergoing major changes: as a result of increasing mechanization and labor shortage, areas that are difficult to manage have been abandoned while fertile favorable agricultural areas have been used more and more intensively. This polarization of land use and especially land abandonment can be found all over Europe from England (Hodgson et al. 2005) through the Czech Republic (Pavlů et al. 2005) and Germany (Poschold and WallisDeVries 2002) to Spain (Gómez-Limón and Fernandez 1999). In the Alps, this trend was especially strong in the period between the two World Wars and again from the beginning of the 1950s onwards (Bätzing 1996), and can best be observed in the French and Carnic Alps where mountain agriculture was given up almost throughout the entire area (Cernusca et al. 1999; Tappeiner et al. 2003). However, in some regions, e.g. western Austria, parts of Switzerland and South Tyrol, this trend has not prevailed mainly because of governmental intervention and subsidies (Baur et al. 1999). Nevertheless, a significant decline in the use of unfavorable agricultural areas (almost 40% of alpine meadows and pastures or steep valley slopes) and increasing land use intensity in favorable areas (on ca. 30% of the agriculturally used area (ASTAT 2002)) has also been observed in these regions in the past decades (Tappeiner et al. 2006; Tasser et al. 2007). Both tendencies, i.e. increasing land abandonment as well as land use intensity, have far- reaching consequences in terms of ecological, agricultural and socio-economic aspects of the regions concerned. Several studies have shown that abandoned land has a higher surface runoff, stronger erosion activity and an increased risk of avalanche (Newesely et al. 2000; Tasser et al. 2003). In addition, the plant species diversity of these areas is also subject to major changes. This is of grave concern since for thousands of years, the vegetation of these semi-natural ecosystems has been developing alongside the mountain agriculture and even today, these semi-natural ecosystems still represent one of the most species-rich habitats in Europe (Väre et al. 2003). In the past decades, studies on land abandonment and its consequences on vegetation biodiversity have mainly been carried out (e.g. Tasser and Tappeiner 2002, Pavlů et al. 2005, Marini et al. 2007). However, most of those studies are limited by the fact that they were performed either on a small region or they only include a relative small number of surveys. Therefore, this paper gives an overview of biodiversity as well as plant species composition for a large area using a huge number of surveys and taking also information on the site factors (inclination and elevation) into consideration. 3 Besides providing a detailed list of plant communities found in the different land use types, we also address the following hypotheses: - The different site factors found at the ten land use types have a significant influence on land use intensity. - There are significant differences in the composition of species within the ten different land use types. - Plant diversity decreases considerably with increasing land use intensity as well as with land abandonment. The peculiarity of this investigation consists in the statistical analysis via discriminant analysis with a special focus on the plant species composition. Based on 936 vegetation surveys with reliable land use information distributed over a large region and the diverse conditions found within our study site (intersection of several floral elements and broad range of geological and climatic conditions), our data and conclusions are also of supra-regional importance.

4 Methods

Study site

- The Autonomous Province of Bolzano/Bozen, South Tyrol (7400 km²), is the northernmost province of Italy and is situated in the center of the Alpine range (lat. 46°22’- 47°05’, long. 10°24’- 12°25’). Despite its mountainous landscape and the decline in agriculturally used mountain areas in Europe, approximately 44% of the total surface area is still used for agricultural purposes. Most of the agricultural area (almost 2500 km²) comprises subalpine to alpine pastures and alpine meadows (ASTAT 2002). Compared to the neighboring regions, the mountain agriculture is relatively intact so that many different land use types are represented (Baur et al. 1999; Pasquali et al. 2002). Furthermore, the Germanic and Romanic laws of succession have been in effect on the study site for centuries (Baur et al. 1999), which additionally has led to a good cross-section of alpine agricultural structures in the study region. - The studied alpine grassland lies at an elevation of between 1600 and 2500 m a.s.l. offside the permanent settlement area. In addition, we included surveys on meadows in the valleys with several levels of land use intensity in order to compare them with extensively used areas. These fertilized meadows, which are cut several times a year, are situated at much lower elevations varying between 500 and 1400 m a.s.l. - Due to geographical preconditions and historical development in the study area, more than ten different floral elements from the Eastern, Southern, Central and even some from the Western Alpine region overlap in this area (Fischer et al. 2005). - The climate in South Tyrol is continentally influenced. The surrounding mountain chains reduce the effects of cold surges and, together with the mass warming of the mountains, they are responsible for above-average temperatures on the one hand and relatively low precipitation on the other. Apart from the intensively used meadows in the valleys, all of the studied areas are situated in the subalpine and alpine stages. Here, the average annual temperature ranges between 0 and 4°C, with the highest temperatures in July/August and the lowest in January. The annual precipitation fluctuates between 800 and 1500 mm. The annual snow cover duration largely depends on the aspect of the area varying between 5 and 7 months. - The study area is situated at the intersection of two continents: the north of the region belongs to the European continent; the south belongs to the Adriatic microcontinent, which 5 is part of the African Plate. Three geological regions can be distinguished. In the Penninic nappes and the Austroalpine nappes mainly neutral to acid slates, orthogneisses and paragneisses, can be found. They both belong to the European continent. In the Southern Alps, which are geologically considered as part of Africa, phyllites as well as basic sedimentary rocks such as limestones and dolomites are predominant. - The type of soil strongly depends on site conditions: in well-developed soils with deep profiles Cambisols (“Braunerden”, “Braunlehme”) and Podzols are predominant. In exposed locations and in the alpine belt, mainly weakly developed AC soils such as Ranker and Rendzina (Leptosols) can be found. Depending on the source rock, the pH of the top soils strongly varies and ranges from 3.7 in soils with silicates to 7.8 in calcareous parent rocks. Fig. 1 Niedrist et al.

Data collection

We started our research with an extensive literature search for vegetation surveys carried out on meadows and pastures in South Tyrol in the last three decades. Only high quality surveys with exact details of land use and that were also made by the method of Braun-Blanquet (1964) were taken into account. For areas for which no data were found, 176 new vegetation surveys were carried out using the same method. The size of the survey areas was 4 x 4 m and the land use details were determined through a free response interview with the farmers. Table 1 Niedrist et al.

The catalog of vascular plants of South Tyrol (Wilhalm et al. 2006) served as a basis for the nomenclature of the species. The nomenclature of plant associations is mostly based on Grabherr and Mucina (1993) and Mucina et al. (1993) and in special cases on Braun-Blanquet (1949), Ebner (1996), Ellenberg (1996), Marschall and Dietl (1974), Walloseck (1999) and Wegener and Reichhof (1988). In total, we collected 936 vegetation surveys distributed over 98 areas with reliable information on land use (for a detailed list and distribution see Table 1, Table 2 and Fig. 1). About two thirds of the surveys belong to moderately to extensively used alpine meadows and pastures as well as on abandoned land offside the permanent settlement area, and the 6 remaining are surveys from more intensively used meadows at lower elevations. We identified a total of 670 different vascular plant species, which corresponds to about one fourth of the total number of vascular plants in South Tyrol (Wilhalm et al. 2006).

Data analysis

All of the collected data were compiled in a single table for analysis. TWINSPAN (Hill 1979) and a hierarchical cluster analysis carried out with the program SORT (Ackermann and Durka 1998) were used to assign the vegetation surveys to 31 plant communities. Two discriminant analyses were used in order to calculate the probability of a survey belonging to one of the ten different types of land use. We regarded the hit rate as a quality criterion for the differentiation of the different land use types. For the calculation of the discriminant function, all cases (i.e. surveys) were taken into consideration. In the first analysis, we determined the relation between land use type and site factors using elevation, slope inclination and aspect as factors. In the second discriminant analysis, all species with more than three incidences in the 936 surveys were considered as independent variables. This approach thus allowed to determine the discriminant function, significances and the probability of a survey of anyone of the ten different types of land use as well as characteristic species that could be used to define different land use types. The plant diversity in the different types of land use was described and compared by the average number of species and dominance conditions (Shannon-Wiener index, Evenness- index). To compare the average number of species per land use type, we used a significance test based on the Bonferroni method. The Shannon-Wiener index (H’) takes into consideration the number of species as well as the species distribution in a survey area with the latter carrying more weight in the calculation. The index is calculated as follows (Tremp 2005):

s H'-p*ln(p) (1) i1 where: n p  (2) N H’ = Shannon index; s = total number of species; N = sum of abundances of all species; n = abundance of the species i; p = relative abundance of each species

7 The Evenness index only depends on the relative distribution of the species in a given survey area. It is the ratio between the Shannon index H’ and the most regular distribution Hmax for the same number of species: H' E (3) Hmax

Land use types

In South Tyrol, agriculturally used grassland is either utilized for the production of hay and silage or as pasture for livestock. The ten land use types we investigated differ from one another with regard to the way in which they are used (meadows, pastures or abandoned land), or for land use intensity (number of grazings or mowings per year) and fertilization (fertilized/unfertilized, see Table 2). The livestock on the pastures of the survey area consisted mostly of dairy cows and young cattle in the valleys up to the subalpine stage, and usually sheep and sometimes goats in the less fertile land at higher elevation. Pastures that are grazed more than two times a year independently from their altitude were termed intensively used pastures, areas that are grazed one to two times per year are referred to here as extensively used pastures. Meadows in close proximity to farms are most intensively used. They are mown two to five times per year, usually moderately to intensively fertilized and frequently irrigated. Some of these areas are additionally grazed in spring and fall. Meadows which are at a distance from the permanent settlement area are traditionally named mountain meadows (“Bergmähder”) and are in general mown only once a year or even less. They are either not fertilized at all or only with manure (“Dungmähder”), and the vast majority of meadows are not irrigated. These meadows are usually managed mechanically, and only a few are still mown manually, most of them being grazed once in the fall. Areas that had not been used for ten to twenty years from the time the vegetation surveys were carried out were referred to as abandoned land. Some of these areas are overgrown with shrubs, but they are still treeless. Areas that have been abandoned for an even longer period of time were intentionally not included in our study. Table 3 Niedrist et al.

8 Results

I Relationship between site factors and land use type

Concerning elevation, especially differently used meadows show clear gradations. For example, meadows mown five or four times (5F, 4F) are found in the lowest elevations (730 m a.s.l.), followed by fertilized meadows mown three or two times or only once (1075 m, 1370 m and 1770 m, Fig. 2). Unfertilized meadows (1U, SU) are mainly located at higher elevations than fertilized, once mown meadows (1860 m and 1940 m, respectively). However, the difference was not significant (maximum elevation for mown alpine meadows: 2500 m a.s.l.). Abandoned areas, however, differ significantly from managed areas in terms of elevation: 90% of abandoned areas are situated 2000 m a.s.l. or higher. There was no significant difference between extensively and intensively used pastures (2100 m and 1980 m, respectively). We found a strong link between average slope inclinations and land use: Not surprisingly, the results revealed a decrease in land use intensity with rising steepness. While the five different degrees of fertilized meadows (5F, 4F, 3F, 2F, 1F) exhibit only a small increase in slope inclination, i.e. between 5.4 and 12.2°, there is a sharp rise to unfertilized areas (18.1°). In addition, we found only slight differences between unfertilized alpine meadows that are mown annually and those that are mown only sporadically. The highest values were found on abandoned land with an average slope inclination of 23.1°. Significant differences were also found between extensively used pastures (22.8°) and less steep intensively used pastures (15.7°). We also compared the aspect of the various land use types. In our study, almost two thirds of the areas face south-east to south-west. However, no significant differences between the single land use types were found. Finally, we conducted a discriminant analysis to establish how many surveys could be correctly assigned to a land use type using only three site factors (elevation, slope inclination and aspect) (Table 3, Table 4). The intensively used meadows in the valleys were predominantly correctly assigned (46.2-100%). Most of the wrongly classified surveys were assigned to other types of fertilized meadows. The percentage of correctly classified abandoned areas was slightly lower but still high at 57.1%. Surveys that were wrongly classified were mostly assigned to extensively used pastures and unfertilized alpine meadows. More than one third of the surveys on pastures and fertilized alpine meadows were classified 9 correctly (35.4% and 40.2%, respectively). The lowest percentage of correct classifications was found for unfertilized alpine meadows (17.9% and 14.1%, respectively). Wrongly classified surveys were mainly assigned to extensively used pastures, once-mown fertilized alpine meadows and abandoned land. Overall, 43% of surveys could be classified correctly using the abovementioned three site factors. Fig. 2 Niedrist et al. Table 3 Niedrist et al. Table 4 Niedrist et al.

II Impact of land use intensity on plant communities and composition of species

In ten different land use types, we found 31 plant communities, whereby extensively used and abandoned areas showed a significantly higher number compared to intensively used meadows (Table 5). The greatest number of different associations by far (20) was found on annually mown, unfertilized alpine meadows, while fertilized alpine and valley meadows were inhabited only by a small number of different plant communities. Remarkably, there was not a single association that could be found in all land use types we surveyed. The yellow oat grassland association (Trisetetum flavecentis) showed the greatest prevalence: it was identified in as many as 6 of the ten studied land use types. These results point to the theory that various land use types should differ significantly in their plant species composition. This was confirmed by the discriminant analysis (Table 6, Table 7). The intensively fertilized valley meadows particularly show a markedly different species composition compared to the other land use types, and they can thus be classified correctly to a high degree. Surveys that were classified wrongly were for the most part assigned to other intensively used meadows (3F-5F). Pastures were also correctly classified to a high degree (78.8% and 84%). More than half of the surveys of unfertilized alpine meadows were assigned to the correct land use type; most of the others were at least classified as another unfertilized land use type. Overall, 72% of all vegetation surveys could be correctly classified on the basis of their plant species composition. Using discriminant analysis, we identified 67 significant species, the most common of which are listed in Table 6. Especially the intensively used valley meadows show a number of characteristic species which can be found almost exclusively in the respective land use type only, where they show high cover rates (e.g. Heracleum sphondylium, Poa trivialis or 10 Anthriscus sylvestris). For the extensive land use types, herbaceous plants such as Anthyllis vulneraria ssp. alpestris, Pedicularis elongata or also Nardus stricta were the most common character species. The latter is also very frequently found on pastures in combination with Deschampsia cespitosa or Poa alpina. Areas that were only recently abandoned are characterized by grasses such as Festuca varia and Calamagrostis villosa or Molinia caeruleae in moister areas. Table 5 Niedrist et al. Table 6 Niedrist et al. Table 7 Niedrist. et al.

III Relationship between land use and plant diversity

We found that a decrease in land use intensity was paralleled by a marked rise in the average numbers of plant species. Fertilized valley meadows mown five times a year proved to be the land use type with the lowest number of plant species (14.3 ± 6.4); meadows mown four, three or two times showed a slight but not significant rise in the number of plant species. Fertilized once-mown mountain meadows showed a distinctive jump in values; at an average of 28.5 ± 2.7 species per survey, they present a similar value as extensively used pastures. The number of species rises for areas that are used even more lightly and peaks for meadows which are mown every two or three years (average of 39.3 ± 4.5 species; minimum 11 species (Caricetum goodenowii); maximum 63 (Campanulo scheuchzeri - Festucetum noricae)). We found a slight but not significant decline in plant species numbers in the abandoned areas (Table 8). The correlation (r²) between land use type and number of species was considerably high. If the levels of land use intensity were brought onto a metric scale (ratio of number of human interventions (fertilization, mowing, grazing) and the number of years required (Tappeiner et al. 1998)), decreasing land use intensity would result in a linear regression of r²=0.90 (Fig. 4).

Fig. 4 Niedrist et al.

The values of the Shannon-Wiener index show a tendency similar to the results regarding the number of species: the more extensively an area is used, the higher the index is; the maximum is reached in sporadically mown alpine meadows at H’= 2.85 ± 0.07 (absolute maximum H’= 11 3.71 for Campanulo scheuchzeri - Festucetum noricae). Abandoned areas show a decline that is more pronounced than the decline we identified for the number of species. Intensively used pastures show the lowest value (H’= 2.08 ± 0.073, absolute minimum H’= 1.47 for Rumicetum alpini) with a low number of species as well as uneven distribution being found in this form of management. The Evenness index shows a different picture: the highest value and thus the most regular distribution of species can be found on the intensively used valley meadows (E = 0.81-0.85). This meadows differ significantly from all other land use types. Once-mown meadows reach values of between E = 0.77 and 0.79. No significant difference between fertilized and unfertilized once-mown meadows nor between intensively and extensively used pastures, which both showed lower values, was found. The lowest evenness of all land use types surveyed was found on abandoned areas with E = 0.73 (absolute minimum E= 0.51 for Gentianello anisodontae - Festucetum variae). Fig. 3 Niedrist et al. Table 8 Niedrist et al.

Discussion

I Relationship between site factors and land use type

In the discriminant analysis, as many as 43% of all cases were classified correctly using only three input variables. This finding led to the conclusion that site factors have great impact on determining how the land is used. Tappeiner et al. (1998) and Breitenberger (2007) achieved higher results in similar analyses (54.6% respective 73.5%); however, this is due the fact that these studies also included other variables such as quality yield and accessibility parameters. In our study, we found slope inclination to have the greatest influence on an area’s land use intensity as land is almost exclusively fertilized with agricultural machinery, and hence, fertilization is restricted to slopes with less than 20 degrees inclination. Meadows with steeper slopes are used more lightly, irrespective of their elevation. This is also true for pastures: steeper pastures are used less intensively because livestock prefer less inclined terrain resulting in an accumulation of feces in these areas. Abandoned land has the strongest slope inclination of all land use types investigated; this strong inclination seems to be a key-reason why these areas are no longer managed. These findings are corroborated by Tappeiner et al. 12 (1998), Tasser and Tappeiner (2002) and Giupponi et al. (2006), who also report high slope inclination to be a crucial parameter for land being abandoned. In contrast, sea level does not appear to influence whether an area is fertilized or not. Rather, the shorter getting vegetation period on higher elevations decrease a meadow’s productivity and thus the number of mowings per year (Körner 2003). Interestingly, contrary to Bischof (1981), we did not find aspect to be a factor that influences an area’s land use type. Farmers have traditionally always tried to lay out all their agricultural areas to face south independently from the form of management since the soil and canopy temperatures rise with increased sun exposure (Cernusca and Seeber 1989). This also has the effect of extending the vegetation season due to shorter snow cover duration. The percentage of correctly classified cases differs for each land use type (see Table 3); similar to Tappeiner et al. (1998) and Breitenberger (2007), intensively used areas achieve especially high percentages because only areas with low elevation and small slope inclination are suited for intensive land use. In contrast, unfertilized alpine meadows (1U, SU) achieve relatively low percentages of correctly ordered cases and cannot be clearly distinguished from extensively used pastures or abandoned land. In these areas, other factors such as the availability of water for grazing livestock and accessibility are decisive for the respective form of management. These results lead to the conclusion that land use types are strongly influenced by site factors, but cannot be explained exclusively by them. To better determine land use types in agriculturally used alpine areas, other factors such as the distance from the farmhouse, accessibility for agricultural machinery or social circumstances must also be taken into consideration (Tappeiner et al. 1998; Tasser and Tappeiner 2002; Breitenberger 2007).

II Impact of land use intensity on plant communities and species composition

Most plant communities are restricted to two or three land use types (Table 5). The fact that Trisetetum flavescentis can be found on several land use types, i.e. on fertilized areas with different levels of land use intensity as well as sometimes on extensively used areas, has already been described several times in the literature. For example, Knapp and Knapp (1952) mentioned the presence of Trisetetum flavescentis on once to three-times mown and fertilized meadows, Mucina et al. (1993) described it in fertilized as well as unfertilized areas, and Ellenberg (1996) even described a co-existence of Trisetetum flavescentis and natural alpine grassland communities such as the Seslerio-Caricetum sempervirentis on calcareous soil. The

13 reason for this broad spectrum of habitation probably lies in the fact that the yellow oat-grass association can be found as numerous different subassociations and variations. The large number of plant communities found on pastures can be explained by the high heterogeneity created by the selective feeding behavior of grazing livestock (Rook et al. 2004). Both extensively and intensively used pastures, characterized by a high proportion of graminoides as well as specialists resistant to browsing and trampling, reached high percentages of rightly ordered cases in the discriminant analysis. It is remarkable that the Crepido-Cynosuretum, which is consistently mentioned in the scientific literature on pastures in the Alps, does not appear at all in our studies. Indeed, we found the charakteristic species Cynosurus cristatus only twice in all 936 vegetation surveys. However, this finding is supported by Schubiger et al. (1999) and Mertz (2000) who reported that Crepido- Cynosuretum mainly grows in the northern Alps where precipitation is high. The discriminant analysis shows that the species composition of extensively used pastures and unfertilized alpine meadows does not always enable a clear distinction between the two land use types (see Table 6). This is due to the fact that many unfertilized alpine meadows are used as pastures in the fall (personal observation, interviews with farmers), which creates conditions similar to areas used as pastures only. Intensively used meadows can be clearly distinguished from other land use types by plant communities and species composition. If we took all types of intensive land use together (5F- 2F), the percentage of rightly ordered cases would even reach the very high percent of 96.4%. Intensive land -use areas can be distinguished from extensively used alpine meadows because of the absence of subalpine and alpine species and, in accordance with Myklestad and Sætertsdal (2004), the high incidence of generalist species such as Dactylis glomerata. In abandoned areas, the most common plant communities were Nardetum callunetosum and Hypochoerido-Festuceutum paniculatae, which correspond respectively to grass and scrub phrases as defined by Surber et al. (1973). As these meadows were only abandoned 10 to 20 years before the vegetation survey was performed, the plant communities normally considered typical for advanced successional stages in the subalpine stage (Larici Piceetum or Alnetum viridis; Bischof 1981;, Tasser et al. 2007) have not yet prevailed. Especially the grasses Molinia coerula and Calamagrostis villosa turned out to be character species in these ‘long grass meads’ (cf. Spatz et al. 1978). This finding corresponds to Spatz et al. (1978) and Grabner (1997) who also observed more incidences of these species on abandoned meadows.

III Relationship between land use and plant diversity 14

In accordance with comparable studies (Fischer and Wipf 2002; Tasser and Tappeiner 2002; Baur et al. 2006; Maurer et al. 2006), we also found the highest number of species per vegetation survey on unfertilized, sporadically mown alpine meadows. This finding supports the Intermediate Disturbance Hypothesis (Connell 1978) according to which the highest numbers of species are to be found in ecosystems that are lightly disturbed. The average number of species in our study is a little lower than in most of the abovementioned studies because under sporadically mown, unfertilized alpine meadows, we included surveys not only of neglected grassland but also of sporadically mown subalpine moist areas, which are relatively poor in species (average number of species in neglected grassland: 44.3 ± 5.4; moist areas: 25.1 ± 4.1). Moderately fertilized, once-mown alpine meadows have much lower average numbers of species. Due to faster and higher growth in fertilized areas most of the sunlight is already absorbed in the upper canopy layer and only a little amount of light gets through to the lower layers (Cernusca and Seeber 1989; Ellenberg 1996). As a result, a few competitive species come to dominate, while stress-tolerant species disappear (Marini et al. 2007). This trend of decreasing numbers of species with increasingly intensive land use can be seen through all land use types we tested with the five times mown and fertilized meadows showing the lowest average number of species. Frequent mowing acts as a continuous disturbance allowing only those plant species to grow that have either very short life cycles or can clonally reproduce (Bahn et al. 1994). Regarding species diversity on abandoned alpine meadows, there is no consensus of opinion in the scientific literature. For example, Hard (1976) and Grabner and Heiselmayer (2002) found no change in the number of species, or even a small increase during the first years of abandonment. However, most other authors of similar studies (e.g. Bischof 1981; Jacquemyn et al. 2003; Baur et al. 2006; Maurer et al. 2006) describe a decrease in species diversity right from the start of abandonment. Without mowing or grazing, the plant species composition undergoes fundamental change: tall competitive species become dominant forcing slow- growing species in the lower canopy layers gradually to recede (Cernusca and Seeber 1989; Pavlů et al. 2005). In addition, the inhibition of seedling recruitment by a litter layer may be a reason for decreasing species number in abandoned meadows and pastures (Jensen and Meyer 2001). Opinions on species diversity of pastures also differ. Austrheim et al. (1999), for instance, found more species in pastures compared to mown areas, while Beltman et al. (2003) and 15 Fischer and Wipf (2002) observed a clear reduction of species richness in grazed areas that were formerly used as meadows. The reason for these contrasts, which were also pointed out by Maurer et al. (2006) and Jacquemyn et al. (2003), is that the influence that grazing has on species diversity heavily depends on plant communities, site factors and especially on the intensity of land use (Chemini and Rizzoli 2003). Our findings corroborate this statement: all intensively used meadows investigated are less species-rich than pastures, while the number of species on unfertilized meadows (1U, SU) is significantly higher than on pastures. Principally, the number of species on pastures (both lightly and intensively grazed ones) is significantly different from abandoned land (see Table 8). This finding is in contrast to Tracy and Sanderson (2000) who found grazing not to have an influence on species diversity. The high Shannon-Wiener indices of unfertilized alpine meadows correspond to the results of comparable studies (e.g. Fischer and Wipf 2002). Contrary to Hobhom (2000) and Tremp (2005), who found a strong correlation between the Shannon-Wiener index and species distribution, our values clearly correlate with species numbers. This could be explained by differences in average species numbers within the different land use types, which are much greater than the differences in the Evenness index. The highest Evenness values were found in meadows in the valleys where disturbances are frequent, thereby creating stable conditions for competition. In contrast, the selective feeding behavior of grazing livestock causes uneven conditions (Rook et al. 2004), which helps graminoids and dwarf shrubs to become dominant, while other plants, especially herbaceous ones, are selectively browsed (Fischer and Wipf 2002). Abandoned areas show the most irregular distribution of plant species with only a few dominant species prevailing as management no longer acts as a leveling factor for species dominance.

Conclusion

Returning to the three key hypotheses of this paper, we can draw the following conclusions: - The type of land use depends on site factors. For example, intensively used grasslands are especially strongly influenced by elevation and slope inclination. In extensively used areas, other factors such as accessibility or the availability of water seem to be decisive. - The investigated types of land use differ markedly in their species composition. Thus, it is possible to deduce the land use type by analyzing the plant composition. This opens 16 up new possibilities for rapid classification of land use in mountain agriculture (e.g. in context with agri-environmental incentive payments or an application in mountain hay certification (Witting et al.; Dalla Via et al. 2004)). - The average number of plant species (alpha diversity) on a given area strongly correlates with the intensity of land use; the exceptionally high species diversity on extensively used alpine meadows was reaffirmed. Hence, the results obtained once again highlight the importance of conserving and supporting extensive mountain agriculture for the maintenance of high biodiversity in the alpine region. Measures facilitating access to alpine grassland as well as area-based subsidies would give farmers incentives to continue or even increase the sustainable use of meadows and pastures, and are therefore useful steps to take to better ensure high plant species diversity in alpine grassland ecosystems.

Acknowledgements. We thank Dr. Ruth Willmott, BioScript for her useful revision of the manuscript, the Hydrographic Agency of the Autonomous Province of Bolzano / Bozen, South Tyrol for the climate data and the Federal State of Tyrol / Austria for analyzing the soil samples. This work was funded within the framework of the INTERREG-IIIA projects “MASTA” and “DNA-Chip-Entwicklung zur Charakterisierung und Valorisierung von Bergheu”. The final publication is available at Springer via http://link.springer.com/article/10.1007/s11258-008-9487-x.

17 References

Ackermann W, Durka W (1998) Sort 4.0 Handbuch. Athors edition, München. ASTAT 2002. 5. Landwirtschaftszählung 2000. ASTAT (Autonome Provinz Bozen-Südtirol. Landesinstitut für Statistik), Bozen Austrheim G, Gunilla E, Ollson A, Grontwentd, E (1999) Land use impact on plant communities in semi-natural sub-alpine grasslands of Budalen, central Norway. Biol Conserv 87 (3):369-379 Bahn M, Cernusca A, Tappeiner U, Tasser E (1994) Wachstum krautiger Arten auf einer Mähwiese und einer Almbrache. Verhandlungen der Gesellschaft für Ökologie 23:23-30 Bätzing W (1996) Landwirtschaft im Alpenraum – unverzichtbar, aber zukunftslos? Europäische Akademie Bozen (ed) Blackwell Wissenschaftsverlag, Berlin Baur B, Cremene C, Groza G, Rakosy L, Schileyko AA, Baur A, Stoll P, Erhardt A (2006) Effects of abandonment of subalpine hay meadows on plant and invertebrate diversity in Transylvania, Romania. Biol Conserv 132:261-273 Baur P, Pezzatti M, Rieder P, Schluep, I (1999) Langfristige Entwicklung der Agrarstrukturen in Südtirol. Arbeitshefte der Europäischen Akademie Bozen, Fachbereich Alpine Umwelt, Bozen Beltman B, Van Den Broek T, Martin W, Ten Cate M, Güsewell S (2003) Impact of Mowing Regime on Species Richness and Biomass of a Limestone Hay Meadow in Ireland. Bulletin of the Geobotanical Institute ETH, 2003 (69):17-30 Bischof N (1981) Pflanzensoziologische Untersuchungen von Sukzessionen aus gemähten Magerrasen in der subalpinen Stufe der Zentralalpen. Beiträge zur geobotanischen Landesaufnahme der Schweiz 60:1-127 Braun-Blanquet J (1949) Übersicht der Pflanzengesellschaften Rätiens (IV). Plant Ecol 2:20- 34 Braun-Blanquet J (1964) Pflanzensoziologie. 3rd edn Springer, Wien Breitenberger I (2007) Standortfaktoren und Nutzung von ausgewählten Almen in Südtirol. Diploma thesis, Universität Innsbruck Cernusca A, Seeber MC, (1989) Phytomasse, Bestandesstruktur und Mikroklima von Grasland-Ökosystemen zwischen 1612 und 2030 M in den Alpen. In: Struktur und Funktion von Graslandökosystemen im Nationalpark Hohe Tauern. Österreichische Akademie der Wissenschaften, Veröffentlichungen des österreichischen MaB- Programms, Universitätsverlag Wagner, Innsbruck 13:419-461 18 Cernusca, A Tappeiner U, Bayfield N (eds) (1999) Land use Changes in European Mountain Ecosystems. Backwell Wissenschaftsverlag, Berlin Chemini C, Rizzoli A, (2003) Land use change and Biodiversity conservation in the Alps. Journal Mountain Ecology 7 (Suppl.):1-7 Connell JH (1978) Diversity in tropical rain forests and coral reefs. Science 199:1302–1310 Dietl W (1995) Wiesen und Weiden im Berggebiet. Montagna 6:1-6 Dalla Via J, Tasser E, Tappeiner U, Baric S, Mair V, Kasal A (2004) Biological-ecological fundamentals of mountain hay certification. Laimburg Journal 1:95-108 Ebner C (1996) Die Wiesengesellschaften des oberen Vinschgaus (Südseite) und ihre Bewirtschaftung. Diploma thesis, Universität Innsbruck Ellenberg H (1996) Vegetation Mitteleuropas mit den Alpen. 5th edn. Ulmer, Stuttgart Fischer M, Wipf S (2002) Effect of low-intensity grazing on the species-rich vegetation of traditionally mown subalpine meadows. Biol Conserv 104:1-11 Fischer MA, Adler W, Oswald K (2005) Exkursionsflora für Österreich, Liechtenstein und Südtirol. 2nd edn. Land Oberösterreich, Biologiezentrum der OÖ Landesmuseen, Linz Giupponi C, Ramanzin M, Sturaro E, Fuser S (2006) Climate and land use changes, biodiversity and agri-environmental measures in the Belluno province, Italy. Environmental Science and Policy 9:163-173 Gómez-Limón J, Fernandéz JdL (1999) Changes in use and Landscape preferences on the agricultural-livestock landscapes of the central Iberian Peninsula (Madrid, Spain). Landscape and Urban Planning 44:165-175 Grabherr G, Mucina L (1993) Die Pflanzengesellschaften Österreichs; Teil II. Gustav Fischer Stuttgart Grabner S (1997) Die Bergmähder des Nationalpark Hohe Tauern in Salzburg. In: Bericht über die 2. Pflanzensoziologische Tagung „Pflanzengesellschaften im Alpenraum und ihre Bedeutung für die Bewirtschaftung“, BAL Gumpenstein:109-116 Grabner S, Heiselmayer P (2002). Diversity of mountain meadows in the inner alpine valley Virgental / Eastern Tyrol. Razprave IV. Razreda Sazu, 43/3:167-184. Hard G (1976) Vegetationsentwicklung auf Brachflächen. In: Bierhals, E. et al. (1979): Brachflächen in der Landschaft. KTBL-Schrift Hill, OM (1979) TWINSPAN- a FORTRAN program for arranging multivariate data in an ordered two- way Table by classification of individuals and attributes. Cornell, University Ithaca Hobhom C (2000). Biodiversität. Quelle & Meyer, Wiebelsheim 19 Hodgson JG, Grime JP, Wilson PJ, Thompson K, Band SR (2005) The impacts of agricultural changes (1963-2003) on the grassland flora of Central England: processes and prospects. Basic Appl Ecol 6:107-118 Jacquemyn H, Brys R, Hermy M (2003) Short term effects of different management regimes on the response of calcareous grassland vegetation to increased nitrogen. Biol Conserv 111:137-147 Jensen K, Meyer C (2001) Effects of light competition and the performance of Viola palustris and on species composition and diversity of an abandoned fen meadow. Plant Ecol 155:169-181 Knapp G, Knapp R (1952) Über Goldhaferwiesen (Trisetum flavescentis) im nördlichen Vorarlberg und im Allgäu. Printed exerpt from „Landwirtschaftliches Jahrbuch für Bayern“.29, vol. 5/6:239-256 Körner C (2003) Alpine Plant Life – Functional Plant Ecology of High Mountain Ecosystems. 2nd ed. Springer, Berlin Marini L, Scotton M, Klimek S, Isselstein J, Pecile A (2007) Effects of local factors on plant species richness and composition of Alpine meadows. Agric Ecosyst Environ 119:281- 288 Marschall F, Dietl W (1974) Beiträge zur Kenntnis der Borstgrasrasen in der Schweiz. Schweiz. Landwirt. Forschung 13:115-127 Maurer K, Weyand A, Fischer M, Stöcklin J (2006) Old cultural traditions, in addition to land use and topography, are shaping plant diversity of grasslands in the Alps. Biol Conserv 130:438 – 446 Mertz P (2000) Pflanzengesellschaften Mitteleuropas und der Alpen. ecomed Landsberg/Lech Mucina L, Grabherr G, Ellmauer T (1993) Die Pflanzengesellschaften Österreichs; Teil I. Gustav Fischer, Stuttgart Myklestad Å, Sætertsdal M, (2004) The importance of traditional meadow management techniques for conservation of vascular plant species richness in Norway. Biol Conserv 118:133-139 Newesely C, Tasser E, Spadinger P, Cernusca A (2000) Effects of land use changes on snow gliding processes in alpine ecosystems. Basic Appl Ecol 1:61-67 Pavlů V, Hejcman M, Pavlů L, Gaisler J, Nežerková P, Andaluz MG (2005) Vegetation changes after cessation of grazing management in the Jizerské Mountains (Czech Republic). Ann Bot Fenn 42:343-349

20 Pasquali G, Bassetti, S, Fumai M, Ghirigato I, Morello P, Ruffini FV, Tappeiner G, Tappeiner U (2002) Erfolgsfaktoren einer Region: Das “Modell Südtirol”. Edition Raetia, Bozen Poschlod P, WallisDeVries MF (2002) The historical and socioeconomic perspective of calcareous grasslands-lessons from the distant and recent past. Biol Conserv. 104:361- 376 Raatikainen, KM, Heikkinen RK, Pykälä J (2007) Impacts of local and regional factors on vegetation of boreal semi-natural grasslands. Plant Ecol 189:155-173 Rook AJ, Dumont B, Isselstein J, Osoro K, WallisDeVries MF, Parente G, Mills J (2004) Matching type of livestock to desired biodiversity outcomes in pastures – a review. Biol Conserv 119:137-150 Schubiger FX, Dietl W, Bosshard HR (1999) Nährwert von Futterpflanzen und Weiden des Berggebietes. Montagna 6:I-VIII Spatz G, Weis B, Dolar DM (1978) Der Einfluß von Bewirtschaftungsänderungen auf die Vegetation von Almen im Gasteiner Tal. In: Cernusca A (ed) Ökologische Analysen im Gasteiner Tal. Veröffentlichungen des österreichischen MAB-Programms 2:163-180 Surber E, Amiet R, Kobert H (1973) Das Brachlandproblem in der Schweiz. EAFV -Bericht 112. Tappeiner U, Tasser E, Tappeiner G (1998) Modelling vegetation patterns using natural and anthropogenic influence factors: preliminary experience with a GIS based model applied to an Alpine area. Ecol Modelling 113 225-237 Tappeiner U, Tappeiner, G Hilbert A, Mattanovich E (eds) (2003). The EU Agricultural Policy and the Environment. Europäische Akademie Bozen, Fachbereich Alpine Umwelt. Blackwell Berlin Tappeiner U, Tasser E, Leitinger G, Tappeiner G (2006) Landnutzung in den Alpen: historische Entwicklung und zukünftige Szenarien. Alpine Space - man and environment. 1:23-39 Tasser E, Tappeiner U (2002) Impact of land use changes on mountain vegetation. Appl Veg Sci 5:173-184 Tasser, E, Mader, M. and Tappeiner U. 2003. Effects of landuse in alpine grasslands on the probability of landslides. Basic Appl Ecol 4:271-280 Tasser E, Walde J, Tappeiner U, Teutsch A, Noggler W (2007) Land use changes and natural reforestation in the Eastern Central Alps. Agric Ecosyst Environ 118:115-129 Tracy BF, Sanderson MA (2000) Patterns of plant species richness in pasture lands of the northeast United States. Plant Ecol 149:169-180 21 Tremp H (2005) Aufnahme und Analyse vegetationsökologischer Daten. Ulmer, Stuttgart Väre H, Lampinene R, Humphries C, Williams P (2003) Taxonomic diversity of vascular plants in European Alpine areas. In: Nagy L, Grabherr G, Körner C, Thompson DBA (eds) Alpine Biodiversity in Europe. Springer, Berlin:133-148 Wallosseck C (1999) The acidophilous taxa of the Festuca varia group in the Alps: new studies on taxonomy and phytosociology. Fol Geobot 34:47-75 Wegener U, Reichhoff L (1989) Zustand, Entwicklungstendenzen und Pflege der Bergwiesen. Hercynia NF 26:190-198 Wilhalm T, Niklfeld H, Gutermann W (2006) Katalog der Gefäßpflanzen Südtirols. Veröffentlichungen des Naturmuseums Südtirol. Folio, Wien-Bozen Witting B, Richter gen. Kemmermann A, Zacharias D (2006) An indicator species approach for result-orientated subsidies of ecological services in grasslands – A study in Northwestern Germany. Biol Conserv 133:186-197

Fig. 1 The study site of the Autonomous Province of Bolzano/Bozen, South Tyrol (Italy) and the geographic distribution of the vegetation surveys. Each point/star consists between 1 and 30 surveys

Fig. 2 Comparison of two site factors for differently used grassland in the Central Alps. ¯x ± s.e. For abbreviations of land use type see Table 2

Fig. 3 Comparison of three diversity indices in differently used grassland in the Central Alps. ¯x ± s.e. For abbreviations of land use types, see Table 2

Fig. 4 Relationship between the average number of vascular plants per survey and increasing land use intensity. ¯x ± s.e. For abbreviations of land use type, see Table 2 25

Table 1 References of the literature data used in this study Year of N° of Author Title Type of Publication Location Publication surveys Die Vegetation der subalpinen und alpinen Stufe in Dissertation University of St. Cristina / Dalla Torre, M. 1982 14 der Puez-Geisler Gruppe. Innsbruck (AT) Gröden Die Wirtschaftswiesen des oberen Vinschgaus Diploma thesis University Schluderns, Ebner, C. 1996 120 (Südseite) und ihre Bewirtschaftung. of Innsbruck (AT) Laas Diploma thesis University Florian, K. Die Lärchenwiesen im Naturpark Trudner Horn. 1995 Altrei, Truden 35 of Innsbruck (AT) Wirtschaftswiesen der Nordhänge und Tallagen im Diploma thesis University Hellrigl, S. Oberen Vinschgau aus vegetationskundlicher und 1996 Laas, Prad 77 of Innsbruck (AT) futterbaulicher Sicht. Dissertation University of Keim, K. Die Vegetationsverhältnisse des Pflerschertales 1967 Pflersch 38 Innsbruck (AT). Dissertation University of Lechner, G. Die Vegetation der inneren Pfunderer Täler 1969 Pfunders 26 Innsbruck (AT) Landschaftsentwicklung in der Gemeinde St. Diploma thesis University St. Leonhard Mayer, C. 2004 13 Leonhard in Passeier of Innsbruck (AT) in Passeier Giessener geogr. Meurer, M. Die Vegetation des Grödner Tales 1980 Sellajoch 3 Schriften Analyse der Vegetationsverteilung in der Diploma thesis University Walten/ Mulser, J. Abhängigkeit der Bewirtschaftungsänderungen auf 1998 39 of Innsbruck (AT) Passeiertal den Waltnern Mähdern Niederbrunner, Vegetation der Sextener Dolomiten (subalpine und Dissertation University of 1975 Sexten 11 F. alpine Stufe) Innsbruck (AT).

Oberhammer, Die Vegetation der alpinen Stufe in den östlichen Dissertation University of Fanes- 1979 18 M. Pragser Dolomiten Innsbruck (AT) Sennes-Prags

Dissertation University of Raffl, E. Die Vegetation der alpinen Stufe der Texelgruppe. 1982 Pfelders 26 Innsbruck (AT). Die Vegetation von unterschiedlich genutzten Diploma thesis University Plätzwiese/ Steinmair, V. 1998 92 Almflächen auf der Plätzwiese of Innsbruck (AT) Prags Veröffentlichungen des Thomaser, J. Die Vegetation des Peitlerkofels in Südtirol 1967 Gadertal 5 Ferdinandeums Welche Landschaftsskala eignet sich zur Erklärung Diploma thesis University Vals, Unterhofer, C. 2006 59 der Fließgewässerqualität? of Innsbruck (AT) Schalders Vegetationskundliche Untersuchungen im Bereich Diploma thesis University Vorhauser, K. 1998 Eggen 116 der Eggentaler Alm (Südtirol) of Innsbruck (AT) Vegetationskundlich-ökologische Untersuchungen in Dissertationes Botanicae Latemar, Wallossek, C. der alpinen Stufe am SW-Rand der Dolomiten (Prov. 1990 27 154 Lavazè-Joch Bozen und Trient) Populationsbiologische Untersuchuungen an zwei Diploma thesis University Haseltal/ Winkler, J. 1992 41 eng verwandten Sippen in Alpinen Rasen of Innsbruck (AT) Prettau

Table 2 Number of surveys per land use type with respective land use-ID Land use Number of Land use type ID surveys EP Extensively used pastures 183 IP Intensively used pastures 50 5F Fertilized hay meadows, mown 5 times 5 4F Fertilized hay meadows, mown 4 times 13 3F Fertilized hay meadows, mown 3 times 40 2F Fertilized hay meadows, mown 2 times 229 1F Fertilized alpine meadow, mown once 113 1U Unfertilized alpine meadows, mown once 164 SU Unfertilized alpine meadows, sporadically mown every 2-5 years 77 AL Abandoned land 60

Table 3 Classification results of the discriminant analysis. Predicted group: land use type; independent variables: elevation, slope inclination, and aspect. Average percentage of correctly classified surveys: 43.0%. For abbreviations of land use types, see Table 2 Land use EP IP 5F 4F 3F 2F 1F 1U SU AL Total EP 35.4 21.7 0.0 0.0 0.0 1.7 8.0 2.9 5.7 24.6 100.0 IP 40.0 26.0 0.0 0.0 6.0 8.0 10.0 0.0 2.0 8.0 100.0 5F 0.0 0.0 100.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 100.0 4F 0.0 0.0 30.8 46.2 15.4 7.7 0.0 0.0 0.0 0.0 100.0 3F 0.0 0.0 7.5 7.5 67.5 17.5 0.0 0.0 0.0 0.0 100.0 2F 0.0 0.0 1.3 1.3 12.6 70.9 11.7 2.2 0.0 0.0 100.0 1F 1.8 25.0 0.0 1.8 0.0 21.4 40.2 5.4 4.5 0.0 100.0 1U 4.0 28.5 0.0 0.0 0.0 4.0 24.5 17.9 6.6 14.6 100.0 SU 5.6 28.2 0.0 0.0 0.0 0.0 18.3 5.6 14.1 28.2 100.0 AL 19.0 7.1 0.0 0.0 0.0 0.0 2.4 9.5 4.8 57.1 100.0

Table 4 Addition to Table 3: Canonical discriminant functions from the discriminant analysis of land use and site factors Fcn Eigenvalue % of Variance Cum % Canonical corr Test of fcn(s) Wilks' Lambda Χ² df Sig. 1 2.0117 93.60 93.60 0.8173 1 through 3 0.291099 1131.047 27 0.0000 2 0.1092 5.08 98.68 0.3137 2 through 3 0.876711 120.590 16 0.0000 3 0.0284 1.32 100.00 0.1660 3 0.972432 25.620 7 0.001

Table 5 Weighted incidence of plant communities in different land use types. +: single incidence; I: incidence 2- 3 times; II: incidence more than 3 times. For abbreviations of land use types, see Table 2. 1 in: Grabherr and Mucina (1993); 2 in: Ebner (1996); 3 in: Mucina et al. (1993); 4 in: Ellenberg (1996); 5 in: Braun-Blanquet (1949); 6 in: Wegener and Reichhoff (1988); 7 in: Wallosseck (1999); 8 in: Dietl (1995) EP IP 5F 4F 3F 2F 1F 1U SU AL

Rumicetum alpini1 + I Festuca rupicola - Festuca valesiaca- association2 + + +

Pastinaco-Arrhenatheretum3 + + + II I

Lolietum multiflorum3 II I + Poo-Trisetetum3 + I II II + Trisetetum flavescentis3 + + + II II I

Alchemillo-Poetum supinae3 I Crepido-Festucetum commutatae3 I + + II I Caricetum curvulae1 (with Poa alpina) + Deschampsio cespitosae-Poetum alpinae3 I II +

Festuco-Agrostietum4 II I Nardetum trifolietosum5 I I Loiseleurio-Caricetum curvulae1 + Hygro-caricetum curvulae1 +

Elyno-Caricetum rosae1 + Sieversio-Nardetum strictae (typicum)1 II II II + Nardetum callunetosum6 + II II I

Gentianello anisodontae - Festucetum variae7 + + +

Seslerio-Caricetum sempervirentis1 + I + + Campanulo scheuchzeri - Festucetum noricae1 I II II Trifolio thalii - Festucetum nigricantis1 + + +

Hypochoero-Nardetum8 II I +

Hypocherido uniflorae - Festucetum paniculatae1 I I Gymnademio-Nardetum1 + + Caricetum goodenowii1 I +

Caricetum davallianae1 + I + Selino-Molinietum caeruleae3 + + Caricetum ferrugineae1 I Caricetum sempervirentis1 + +

Carici curvulae-Nardetum1 + Origano-Calamagrostietum variae1 + Number of different plant communities 12 7 3 3 4 5 8 20 9 13

28 Table 6 Classification results of the discriminant analysis. Predicted group: land use type, independent variables: all species with overall incidence of ≥ 3. Average percentage of correct classification: 72.0%. For abbreviations of land use types, see Table 2 Land use EP IP 5F 4F 3F 2F 1F 1U SU AL Total

LP 78.8 0.5 0.0 0.0 0.0 0.0 3.8 12.5 2.2 2.2 100.0

IP 6.0 84.0 0.0 0.0 0.0 2.0 6.0 2.0 0.0 0.0 100.0 5F 0.0 0.0 60.0 0.0 20.0 20.0 0.0 0.0 0.0 0.0 100.0 4F 0.0 0.0 7.7 46.2 38.5 0.0 7.7 0.0 0.0 0.0 100.0

3F 0.0 0.0 0.0 0.0 70.0 12.5 7.5 10.0 0.0 0.0 100.0

2F 0.0 0.4 0.4 1.3 7.8 80.0 8.7 1.3 0.0 0.0 100.0 1F 6.2 3.5 0.0 0.9 0.9 5.3 74.3 8.0 0.9 0.0 100.0 1U 15.9 0.6 0.0 0.0 0.0 0.6 6.7 61.0 9.8 5.5 100.0

SU 6.6 0.0 0.0 0.0 0.0 0.0 0.0 32.9 56.6 3.9 100.0 AL 0.0 0.0 1.7 0.0 0.0 0.0 1.7 25.0 8.3 63.3 100.0 - Deschampsia - Alopecurus pratensis - Agrostis - Briza media - Festuca varia Character cespitosa - Dactylis glomerata capillaris agg. - Anthyllis - Molinia - Festuca halleri - Heracleum sphondylium agg. - Rumex alpinus vulneraria ssp. caerulea species - Poa alpina - Anthriscus sylvestris - Alchemilla alpestris - Calamagrostis - Nardus stricta - Lolium perenne vulgaris agg. - Nardus stricta villosa

Table 7 Addidtion to Table 6: Canonical discriminant functions of the discriminant analysis of land use and plant species Fcn Eigenvalue % of Variance Cum % Canonical corr Test of fcn(s) Wilks' Lambda Χ² df Sig. 1 3.8356 36.87 36.87 0.8906 1 through 9 0.002604 5334.734 585 0.0000 2 2.3937 23.01 59.87 0.8398 2 through 9 0.012593 3921.843 512 0.0000 3 1.0874 10.45 70.32 0.7218 3 through 9 0.042737 2826.390 441 0.0000 4 0.7912 7.61 77.93 0.6646 4 through 9 0.089210 2166.625 372 0.0000 5 0.7184 6.90 84.83 0.6466 5 through 9 0.159795 1644.056 305 0.0000 6 0.6658 6.40 91.23 0.6322 6 through 9 0.274587 1158.716 240 0.0000 7 0.4733 4.55 95.78 0.5668 7 through 9 0.457398 701.244 177 0.0000 8 0.2768 2.66 98.44 0.4656 8 through 9 0.673880 353.852 116 0.0000 9 0.1622 1.56 100.00 0.3736 9 0.860438 134.756 57 0.0000

Table 8 Cross table with the significance values of average species number per survey in differently used grasslands in the Central Alps. (Bonferroni test, Significant differences are in bold, significance level p< 0.005). For abbreviations of land use types, see Table 2 EP IP 5F 4F 3F 2F 1F 1U SU AL IP 0.000 5F 0.075 1.000 4F 0.000 1.000 1.000 3F 0.000 1.000 1.000 1.000 2F 0.000 0.002 1.000 0.146 0.013 1F 1.000 0.000 0.137 0.001 0.000 0.029 1U 0.000 0.000 0.001 0.000 0.000 0.000 0.000 SU 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.157 AL 0.001 0.000 0.000 0.000 0.000 0.000 0.001 1.000 1.000

30 4. A plant’s perspective of extremes: Terrestrial plant responses to changing climatic variability

51 52 Global Change Biology Page 2 of 44

1 2 3 4 1 A plant’s perspective of extremes: Terrestrial plant responses to changing climatic 5 6 2 variability 7 8 9 3 Running title: Plant and changing climatic variability 10 11 C. Reyer1, S. Leuzinger2, A. Rammig1, A. Wolf2, R. P. Bartholomeus3, A. Bonfante4, F. de 12 4 13 4 5 6 7 8 2,9 10 14 5 Lorenzi , M. Dury , P. Gloning , R. Abou Jaoudé , T. Klein , T. M. Kuster , M. Martins , 15 16 6 G. Niedrist11,12, M. Riccardi4, G. Wohlfahrt12, P. de Angelis7, G. de Dato7, L. François5, A. 17 18 7 Menzel6, M. PereiraFor13 Review Only 19 20 21 8 1 Potsdam Institute for Climate Impact Research, Telegrafenberg, P.O. Box 601203 14412 22 23 9 Potsdam, Germany 24 25 26 10 2 Institute of Terrestrial Ecosystems ITES, ETH Zürich, Universitätstrasse 16, CH8092 27 28 11 Zürich, Switzerland 29 30 31 12 3 KWR Watercycle Research Institute, P.O. Box 1072, 3430 BB Nieuwegein, The 32 33 13 Netherlands 34 35 36 14 4 National Research Council of Italy, Institute for Mediterranean Agricultural and Forest 37 38 15 Systems (CNRISAFoM), via Patacca 85, 80056 Ercolano (NA), Italy 39 40 41 16 5 Unité de Modélisation du Climat et des Cycles Biogéochimiques, Université de Liège, Bât. 42 43 17 B5c, Allée du Six Août 17, B4000 Liège, Belgium 44 45 46 18 6 Chair of Ecoclimatology, Technische Universität München, HansCarlvonCarlowitzPlatz 47 48 19 2, 85354 Freising, Germany 49 50 20 7 Department for Innovation in Biological, Agrofood and Forest systems (DIBAF), 51 52 53 21 University of Tuscia, via S. Camillo de Lellis snc – 01100 Viterbo Italy 54 55 22 8 Department of Environmental Sciences and Energy Research, Weizmann Institute of 56 57 58 23 Science, Rehovot, Israel 59 60 1 53 Page 3 of 44 Global Change Biology

1 2 3 1 9 Swiss Federal Research Institute WSL, Zürcherstr. 111, CH8903 Birmensdorf, Switzerland 4 5 6 2 10 Institute of Geography and Spatial Planning (IGOT), University of Lisbon, Edifício da 7 8 3 Faculdade de Letras, Alameda da Universidade, 1600214, Lisboa, Portugal 9 10 11 4 11 Institute for Alpine Environment, European Academy of Bolzano/Bozen, Drususallee 1, 12 13 5 39100 Bolzano/Bozen, Italy 14 15 6 12 Institute of Ecology, University of Innsbruck, Sternwartestr. 15, 6020 Innsbruck, Austria 16 17 18 7 13 University of ÉvoraFor Department Review of Landscape, Environment Only and Planning, Colégio Luis 19 20 8 António Verney Rua Romão Ramalho, 7000671, Évora, Portugal 21 22 23 9 Corresponding author: C. Reyer (reyer@pikpotsdam.de); Tel.: +49 331 28820725 24 25 26 10 Keywords: climate change, plant phenology, plant physiology, observations, experiments, 27 28 11 models, combined approaches 29 30 31 12 Type of paper: Review 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 2 54 Global Change Biology Page 4 of 44

1 2 3 1 Abstract 4 5 6 2 We review here observational, experimental and model results dedicated to how plants 7 8 3 respond to extreme climatic conditions induced by changing climatic variability. 9 10 11 4 Distinguishing between impacts of changing mean climatic conditions and changing climatic 12 13 5 variability on terrestrial ecosystems is generally underrated in current studies. The goals of 14 15 6 our review are thus (1) to identify plant processes that are vulnerable to changes in the 16 17 7 variability of climatic variables rather than to changes in their mean, and (2) to 18 For Review Only 19 8 depict/evaluate available tools to quantify responses of plants to changing climatic variability. 20 21 22 9 We find that phenology is largely affected by changing mean climate but also that impacts of 23 24 10 climatic variability are much less studied but potentially damaging. We note that plant water 25 26 11 relations seem to be very vulnerable to extremes driven by changes in temperature and 27 28 12 precipitation and that changing heatwaves and flooding have stronger impacts on 29 30 31 13 physiological processes than changing mean climate. Moreover, interacting phenological and 32 33 14 physiological processes are likely to add further complexity to plant responses to changing 34 35 15 climatic variability. Phenological and physiological processes and their interactions culminate 36 37 16 in even more sophisticated responses to changing mean climate and climatic variability at the 38 39 17 species and community level. Generally, observational studies are well suited to study plant 40 41 42 18 responses to changing mean climate, but less suitable to gain a mechanistic understanding of 43 44 19 plant responses to climatic variability. Experiments seem best suited to simulate extreme 45 46 20 events and temporal resolution and model structure are crucial in models to capture plant 47 48 21 responses to changing climatic variability. We highlight that a combination of experimental, 49 50 51 22 observational and /or modeling studies have the potential to overcome important caveats of 52 53 23 the respective individual approaches. 54 55 24 56 57 58 59 60 3 55 Page 5 of 44 Global Change Biology

1 2 3 1 1. Introduction 4 5 6 2 Although the spatial and temporal extent of future climatic changes is still partly uncertain 7 8 3 (IPCC 2007a), it is likely that the adaptive capacity of terrestrial plants and ecosystems will 9 10 4 be exceeded in many regions (IPCC 2007b). Already today, changes in ecological responses 11 12 5 to climate change can be observed for individual species and ecosystems (e.g. Allen & 13 14 6 Breshears 1998; Gitlin et al. 2006) but also across species and organizational scales (e.g. 15 16 17 7 Walther et al. 2002; Allen et al. 2010; Lindner et al. 2010). Climate change may manifest 18 For Review Only 19 8 itself in two fundamentally different ways: in a change in the mean of for example 20 21 9 temperature or precipitation, or a change in their variability (i.e. variance or distribution (Fig. 22 23 10 1); Rummukainen 2012; Seneviratne et al. 2012). We define extreme events from this strictly 24 25 26 11 climatological perspective as increase in climatic variability (increasing variance or 27 28 12 distribution) in opposition to change in mean climate. Our aim is to emphasize the generally 29 30 13 unrecognized distinction between impacts of changing mean climate and changing climatic 31 32 14 variability on terrestrial ecosystems. 33 34 35 15 We center but do not limit our synthesis on a plant’s perspective of temperature and 36 37 16 precipitation extremes, since these are the most important climatic determinants of plant 38 39 17 growth and survival globally (e.g. Boisvenue & Running 2006). Observations since 1950 40 41 42 18 show that the length of warm spells and heat waves increased (e.g. Barriopedro et al. 2011; 43 44 19 Rahmstorf & Coumou 2011; Seneviratne et al. 2012). More intense and longer droughts are 45 46 20 observed but at the same time the number of heavy precipitation events increased 47 48 21 (Seneviratne et al. 2012 and references therein). Future projections on changes in climatic 49 50 51 22 variability show strong spatial and temporal heterogeneity (Giorgi et al. 2004; Orlowsky & 52 53 23 Seneviratne 2012) and are highly uncertain (Seneviratne et al. 2012). Using multimodel 54 55 24 experiments, Barriopedro et al. (2011) for instance found that the probability of summer 56 57 25 heatwaves may increase by a factor of 510 in the future while Schär et al. (2004) predict that 58 59 60 4 56 Global Change Biology Page 6 of 44

1 2 3 1 temperature variability will increase by a factor of 2 in Europe. Projected changes in extreme 4 5 2 precipitation events (droughts or flooding) are even more uncertain. Orlowsky & Seneviratne 6 7 3 2011 derived from their simulations with an ensemble of Global Circulation Models (GCMs) 8 9 10 4 robust projections on increasing droughts over the Mediterranean and increasing heavy 11 12 5 precipitation over the Northern high latitudes. 13 14 6 While changes in the mean values are important, there is evidence that plants respond to 15 16 17 7 extreme rather than to average conditions (Chapin et al. 1993; Knapp et al. 2002; van Peer et 18 For Review Only 19 8 al. 2004; Weltzin et al. 2003; Bokhorst et al. 2007; Jentsch & Beierkuhnlein 2008). 20 21 9 Additionally to that, different physiological processes at the species, community or ecosystem 22 23 10 level affect the response of plants to climatic variability (Fig. 2). To account, e.g., for 24 25 26 11 changing precipitation distributions, Knapp et al. (2002) decreased precipitation frequency 27 28 12 but not its total amount in a mesic grassland leading to more intense precipitation events. 29 30 13 They found reduced carbon turnover but increased species diversity. Thus, impacts on 31 32 14 physiological processes influence carbon and water cycles at local and regional scales. The 33 34 35 15 carbon cycle is sensitive to drought (e.g. Ciais et al. 2005; van der Molen 2011; Wu et al. 36 37 16 2012). The water cycle may also be strongly impacted by drier conditions because under 38 39 17 drought evapotranspiration tends to decrease, which leads to lower evaporative cooling 40 41 18 (Teuling et al. 2010). In combination, warming and drought can therefore lead to additional 42 43 19 warming of an ecosystem (Seneviratne et al. 2006; Fischer et al. 2007; Kuster et al. 2012). 44 45 46 20 In addition to the impacts of changing climatic variability, the physiological and ecological 47 48 21 response of terrestrial plants depends also on interactions between species (Thorpe et al. 49 50 51 22 2011) and natural adaptation and acclimation. The water available for plants depends on the 52 53 23 water holding capacity of the soil (Kramer & Boyer 1995; Porporato et al. 2004; Leuzinger & 54 55 24 Körner 2010; RazYaseef et al. 2010), competition with other plants (Casper & Jackson 1997) 56 57 25 and precipitation patterns (Knapp et al. 2008). The latter has different effects on soils with 58 59 60 5 57 Page 7 of 44 Global Change Biology

1 2 3 1 high or low water holding capacity (i.e. a stronger or weaker buffer against drought; Knapp et 4 5 2 al. 2008) or on flood occurrence, which is an important driver of plant distribution (Crawford 6 7 3 1992; Colmer & Flowers 2008; Parolin & Wittmann 2010). Furthermore, interactions 8 9 10 4 between changing climatic variables as well as thereby induced community shifts may affect 11 12 5 the response of plants to new conditions (Langley & Megonigal 2010; de Boeck et al. 2011). 13 14 6 For example, a drier and warmer climate will exert stronger constraints on plant growth than a 15 16 7 warmer but also wetter climate; or rising CO2 may alleviate the impact of drought (Morgan et 17 18 8 al. 2004; Holtum &For Winter 2010). Review Moreover, more prolonged Only dry periods will alternate with 19 20 21 9 more intensive rainfall events, both within and between years, which will change soil 22 23 10 moisture dynamics (Weltzin et al. 2003; Porporato et al. 2004; Fay et al. 2008; Knapp et al. 24 25 11 2008; Bartholomeus et al. 2011a). Eventually, it is also crucial how quickly plant 26 27 12 communities adapt genetically to the imposed changes. The IPCC (2007b) concluded that the 28 29 30 13 rate of natural adaptation will be slower than the rate of climate change. Natural adaptation 31 32 14 differs in between species: while species with short generation times may adapt within years, 33 34 15 e.g. Rehfeldt et al. (2001) estimate that it will take 212 generations (an equivalent of 200 35 36 16 1200 years) for a coniferous trees species to show genetic adaptation in response to climatic 37 38 17 change. All these factors determine whether plants at a specific site will experience changing 39 40 41 18 climatic variability as extreme or not. 42 43 19 Thus, the vulnerability of terrestrial plants to climate change will, besides changes in the 44 45 46 20 mean, largely depend on the changes in the climatic variability and the occurrence of extreme 47 48 21 events. The understanding of this difference in experiments and model simulations requires 49 50 22 very good knowledge of the baseline or control climate (especially the background variability 51 52 23 to which plants are adapted to). This complies with the fact that extreme conditions per se 53 54 55 24 have shaped ecosystems for a long time (Körner 1998, 2003) and may also foster adaptation 56 57 25 and thus decrease sensitivity (Hegerl et al. 2011). The plant’s response to specific 58 59 60 6 58 Global Change Biology Page 8 of 44

1 2 3 1 environmental conditions produces their specialized set of traits which allows them to prevail 4 5 2 over competitors and occupy a specific habitat (Körner 1998, 2003). We use the term ‘stress’ 6 7 3 throughout this review according to Lortie et al. (2004) to refer to situations in which plants 8 9 10 4 experience critical environmental conditions beyond what they experience normally (Chapin 11 12 5 1991) such that damage to vital function occurs (see Gaspar et al. 2002). 13 14 6 In this paper we strive to answer the following questions: 15 16 17 7 • Which plant processes are vulnerable to changes in the variability of climatic drivers 18 For Review Only 19 8 rather than to changes in their mean? 20 21 22 9 • How can we quantify responses of plants to changing climatic variability? 23 24 25 10 We present evidence from experiments, observations and modeling studies that help to 26 27 11 understand the current and future responses of individuals and communities to changing 28 29 12 variability, with a particular focus on temporal and spatial patterns. These examples also help 30 31 32 13 to identify important research gaps. We do not aim to cover the literature on these topics 33 34 14 systematically. 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 7 59 Page 9 of 44 Global Change Biology

1 2 3 1 2. Which plant processes are vulnerable to changes in the variability of climatic drivers 4 5 6 2 rather than to changes in their mean? 7 8 3 The vulnerability of plants refers to their susceptibility to adverse effects of environmental 9 10 11 4 change (IPCC 2007b). Estimates of vulnerability depend on the definitions (e.g. the definition 12 13 5 of death (Zeppel et al. 2011)) and the spatiotemporal scale considered. The ultimate limit to 14 15 6 withstanding environmental stress from an individual plant’s perspective is mortality due to 16 17 7 physiological failure (“You can only die once”) but at the community level, already 18 For Review Only 19 8 reductions in growth and subsequently competitiveness may constitute a limit to species 20 21 22 9 fitness. For commercial crops it may even be a critical reduction in productivity so that 23 24 10 cultivation is discontinued. 25 26 27 11 In the following sections, we discuss the vulnerability of phenological and (individual and 28 29 12 interacting) physiological processes to changes in the climatic variability rather than the mean 30 31 13 of climatic drivers and we highlight how these play out at the species and the community 32 33 14 level (see schematic overview in Fig. 2). Our list of examples is not exhaustive but meant to 34 35 15 illustrate this important difference between changes in climatic variability rather than the 36 37 38 16 mean. 39 40 41 17 2.1. Phenological processes 42 43 44 18 One of the wellstudied responses of plant species or communities to environmental change is 45 46 19 phenology, which tracks seasonal events in generative and vegetative plant growth. Given the 47 48 49 20 predominant influence of climate (with the important exception of photoperiodism, see 50 51 21 Körner & Basler 2010), phenology has emerged as a key tool in identifying fingerprints of 52 53 22 anthropogenic climate change in nature (Menzel et al. 2006). Observed largescale 54 55 23 phenological changes such as an earlier onset of leaf unfolding/ flowering (Menzel & Fabian 56 57 24 1999; Walther et al. 2002; Parmesan & Yohe 2003; Root et al. 2003; Menzel et al. 2006) are 58 59 60 8 60 Global Change Biology Page 10 of 44

1 2 3 1 mainly driven by changes in mean climatic conditions especially temperature (Vitasse et al. 4 5 2 2009; Polgar & Primack 2011; see also Table 1). 6 7 8 3 Phenological changes in response to changing climatic variability are much less studied 9 10 4 although they clearly interact with phenological changes induced by changing mean climate. 11 12 5 For example, in the temperate and boreal zones which are often temperature limited, a central 13 14 6 tradeoff revolves around maximizing the vegetation period while avoiding frost damage 15 16 17 7 (Kramer et al. 2010). An untimely response to early warm spells may be fatal but can bring 18 For Review Only 19 8 enormous advantages for early successional or opportunistic species (rstrategists, Leuzinger 20 21 9 et al. 2011a). In contrast, longlived, late successional species often have chilling 22 23 10 requirements and photoperiodic safety mechanisms (Heide 1993) and thus may be in a 24 25 26 11 position to avoid increasing risks of late frost due to changing climatic variability but would 27 28 12 also benefit less from early warm spells. This is supported by the fact that the risk of damage 29 30 13 due to late frost events has not increased so far for several coniferous and broadleaved 31 32 14 species in Central Europe (Scheifinger et al. 2003; Menzel et al. 2003). Besides this example, 33 34 35 15 there is further evidence, that extreme events may alter phenological responses depending on 36 37 16 their timing and strength (e.g. Jentsch et al. 2009; Menzel et al. 2011). This can lead to 38 39 17 unexpected effects such as second flowering in autumn or extended flowering until the 40 41 18 beginning of winter for some species (Luterbacher et al. 2007). Moreover, extreme warm 42 43 19 spells decreased the differences in spring phenology between urban and rural sites (Jochner et 44 45 46 20 al. 2011). Furthermore, only half of the trees reached leaf maturity in an extreme drought 47 48 21 experiment in the Mediterranean (Misson et al. 2011). Overall, the response of phenology to 49 50 22 climatic variability seems to be less well understood than to changing mean climate although 51 52 23 increasing climatic variability may have a strong damaging potential. 53 54 55 56 57 58 59 60 9 61 Page 11 of 44 Global Change Biology

1 2 3 1 2.2. Physiological processes 4 5 6 2 We here focus on the response of plant water relations such as transpiration to climatic 7 8 3 variability (drought/heat waves and excess water). Increasing temperatures and/or heat waves 9 10 4 combined with less or more variable precipitation events lead to prolonged dry periods and 11 12 5 high atmospheric demand for plant transpiration, which determine drought stress of plants 13 14 6 beyond changes in mean climate (Schimper 1903; Porporato et al. 2004). Barriopedro et al. 15 16 17 7 (2011) predict such an increase in drought events for the 21st century and the consequences 18 For Review Only 19 8 for plant physiology are well documented (e.g. Leuzinger et al. 2005; Bréda et al. 2006; 20 21 9 Granier et al. 2007) although not all mechanism are fully understood. There is an ongoing 22 23 10 debate about two competing response strategies to drought: Isohydric plants may respond by 24 25 26 11 closing their stomates thus reducing their water loss but eventually facing carbon starvation, 27 28 12 whereas anisohydric plants keep their stomates open thus running the risk of hydraulic failure 29 30 13 (Mc Dowell et al. 2008; Sala et al. 2010; Zeppel et al. 2011). Furthermore, Craine et al. 31 32 14 (2012) highlighted the importance of the timing of an extreme event for grassland 33 34 35 15 productivity. The response of plants to drought is of such an importance that Hartmann (2011) 36 37 16 refers to it as a ”change of evolutionary forces” from competition for light to competition for 38 39 17 water and carbon. The responses of plants to climatic variability and particularly drought have 40 41 18 important consequences for net primary productivity (NPP) and hence carbon cycling even at 42 43 19 large spatial scales such as Europe (Ciais et al. 2005; Dury et al. 2011). Thus, plant responses 44 45 46 20 to increasing drought events and heat waves influence plant functioning across spatial and 47 48 21 temporal scales. 49 50 51 22 Also climatic variability resulting in excess water (i.e. flooding or waterlogging), can induce 52 53 23 important physiological responses by terrestrial plants. Due to waterlogging O2 diffusion and 54 55 24 supply to the roots is reduced, and the oxygen demand of plant roots, (i.e. root respiration – 56 57 25 oxygen consumption in the roots), a process that increases with rising temperatures, cannot be 58 59 60 10 62 Global Change Biology Page 12 of 44

1 2 3 1 fulfilled (Lloyd & Taylor 1994; Blom & Voesenek 1996; Kozlowski 1997; Amthor 2000). 4 5 2 This results in waterlogging/ oxygen stress, i.e. lack of oxygen due to high soil moisture 6 7 3 contents (Bartholomeus et al. 2008). Both the oxygen supply and demand may be affected by 8 9 10 4 a more extreme climate, due to more intense precipitation and higher temperatures, 11 12 5 respectively. Therefore, to analyze the effects of low soil oxygen availability on species 13 14 6 performance, it is necessary to integrate the soil physical and plant physiological processes, 15 16 7 thus accounting for both the oxygen supply to and oxygen demand of plant roots 17 18 8 (Bartholomeus et al.For 2011b). BesidesReview reduced root respiration Only rates, the decrease of water 19 20 21 9 absorption due to waterlogging stress causes sensitive plants to wilt in a similar way to 22 23 10 drought (Jackson & Drew 1984). Many species already growing in floodprone habitats have 24 25 11 developed different strategies to survive hypoxia, by producing aerenchyma and/or 26 27 12 adventitious roots in response to an increase in the concentration of ethylene and auxin (Blom 28 29 30 13 & Voesenek 1996). Flooding can also give rise to detrimental effects at leaf level, by inducing 31 32 14 stomatal closure and, consequently, limiting gas exchange and plant growth (Kramer 1951; 33 34 15 Chen et al. 2005; Rengifo et al. 2005; Fernandez 2006). Thus, similarly to drought, extremes 35 36 16 of excess water, in combination with higher temperatures, strongly alter plant physiological 37 38 17 processes. 39 40 41 18 In conclusion, we note that plant water relations seem to be very vulnerable to increasing 42 43 19 variability in temperature and precipitation and that changing heatwaves and flooding have 44 45 46 20 stronger impacts on physiological processes than changing mean climate (see also Table 1). 47 48 49 21 2.3. Interacting physiological processes 50 51 52 22 The interaction of physiological processes may strongly affect the response of plants to 53 54 23 changing climatic variability. Interactions among several global change drivers or between 55 56 24 global change drivers and other environmental variables, may result in other growthlimiting 57 58 59 60 11 63 Page 13 of 44 Global Change Biology

1 2 3 1 factors (e.g. soil type) becoming less important. Drought periods for example may have the 4 5 2 potential to not only determine growth or mortality in an ecosystem but also to cause shifts in 6 7 3 growthlimiting factors, e.g. nutrient limitations. For example, in an experiment of Kuster et 8 9 10 4 al. (2012) oaks were grown on two different soil types with different nutrient availabilities. 11 12 5 Under wellwatered conditions, growth on one soil was lower due to nutrientlimiting 13 14 6 conditions, whereas under repeated drought periods these differences disappeared. This shows 15 16 7 that growthlimiting factors such as nutrient availability can become less important under 17 18 8 changing climatic For variability, whileReview they may persist Only if only changes in mean climate are 19 20 21 9 considered. There are many other examples of interacting processes under changing climatic 22 23 10 variability such as ozone stress during periods of high temperature (Matyssek et al. 2010; 24 25 11 Pretzsch & Dieler 2011). 26 27 28 12 The interactions of physiological processes can however be even more intriguing: In coastal 29 30 13 habitats (i.e. the interface of terrestrial and aquatic habitats) which are not only saline, but are 31 32 14 also prone to flooding (e.g. mangroves and salt marshes) (Colmer & Flowers 2008) Tamarix 33 34 35 15 africana Poir., for example, showed a reduction of CO2 assimilation rates only in young 36 37 16 Tamarix africana Poir. leaves after 45 days under continuous flooding with saline water (200 38 39 17 mM), while old leaves and the aboveground relative growth rate were not affected by the 40 41 18 treatment (Abou Jaoudé et al. 2012). Thus while parts of the plants actually responded to 42 43 19 flooding, this was not the case for the entire plant. This example is rather related to changes in 44 45 46 20 mean climatic conditions (i.e. temperatureinduced rising sea levels) but it highlights that 47 48 21 changing climatic variability is likely to add even another level of complexity to already 49 50 22 complex interactions of physiological processes. 51 52 53 54 55 56 57 58 59 60 12 64 Global Change Biology Page 14 of 44

1 2 3 1 2.4. Species-level processes 4 5 6 2 At the species level, responses of different genotypes to climate provide information how a 7 8 3 species may react to changing climatic variability. Since genotypic variation results in 9 10 4 different sensitivity thresholds of distinct ecotypes to changing climatic variability it can 11 12 5 partly substitute lacking data of changing climatic variability for a specific genotype. In an 13 14 6 ecotype study (Klein et al. submitted) that included all three climate types (meso 15 16 17 7 Mediterranean (MM), thermoMediterranean (TM), and semiarid (SA) within the natural 18 For Review Only 19 8 distribution of the forest tree Pinus halepensis Mill. (and hence three very different 20 21 9 combinations of mean climate and climate variability), two major physiological adjustments 22 23 10 were identified: (1) shortening of the growing season length (from 165 to 100 days) to match 24 25 26 11 a shorter rainy season and (2) increasing water use efficiency (from 80, to 95, to 110 mol 27 28 12 CO2 mol1 H2O under MM, TM, and SA climates respectively). However the sensitivity 29 30 13 threshold differed in between ecotypes: Northern ecotypes mainly responded to the change 31 32 14 MM to TM, whereas Southern ecotypes responded to the change TM to SA. At the species 33 34 35 15 level, the study showed that higher xylem sensitivity to embolism in specific ecotypes 36 37 16 matched previous reports (Atzmon et al. 2004; Schiller et al. 2009) of significantly higher 38 39 17 mortality rates in these ecotypes under yet harsher conditions. These observations suggest that 40 41 18 while hydraulic constraints in response to climatic variability limited the distribution of a tree 42 43 19 species, plasticity in water use efficiency and growth phenology enabled its success under a 44 45 46 20 wide range of climatic conditions. 47 48 49 21 2.5. Community-level processes 50 51 52 22 At the community level, phenological, physiological and specieslevel processes as well as 53 54 23 their interaction culminate in complex responses to changing mean climate and climatic 55 56 24 variability (Fig. 2). Species range shifts have been associated with changes in mean climate 57 58 59 60 13 65 Page 15 of 44 Global Change Biology

1 2 3 1 (Lenoir et al. 2008; Harsch et al. 2009) but also with changing climatic variability (Kelly & 4 5 2 Goulden 2008; Doak & Morris 2010). They lead to a disruption of ecological communities 6 7 3 and species interactions due to different dispersal speed and success. These processes differ 8 9 10 4 between the trailing and the leading edge of a population (Kramer et al. 2010; Doak & Morris 11 12 5 2010). From a community’s perspective such range shifts may entail positive (e.g. release 13 14 6 from competition) and negative (e.g. loss of important pollinator) consequences. Despite 15 16 7 these importance consequences of range shifts, it is yet unclear whether changing mean 17 18 8 climate or changingFor climatic variability Review will be the more Only important driver of range shifts. 19 20 21 9 At community level, for annual plants, the variability of rainfall is important for the success 22 23 10 of germination. Increasing climate variability can have both negative and positive effects on 24 25 26 11 species persistence and thus plant population dynamics (Levine et al. 2008). Climatic 27 28 12 fluctuations, for example, may enable species to avoid interspecific competition if species 29 30 13 differ in the years in which they perform (e.g. reproduce or grow) best (Levine & Rees 2004). 31 32 14 Dormancy and germination biology determine whether temporal variability favors or inhibits 33 34 35 15 species persistence (Levine & Rees 2004) and can thus be limiting for a species (Godefroid et 36 37 16 al. 2011). Temporal variation in resource availability as induced by climatic variability may 38 39 17 reduce the effects of competitive exclusion, allowing more species to coexist (Knapp et al. 40 41 18 2002). 42 43 44 19 A combination of extremes/multiple stresses may not only hamper performance but may also 45 46 20 drive extinctions (Smith & Huston 1989; Niinemets & Valladares 2006). Although functional 47 48 21 tradeoffs exist in adjusting to multiple environmental limitations (Holmgren et al. 1997; 49 50 51 22 Silvertown et al. 1999), adapting to one stressor may go at the cost of adapting to another 52 53 23 (Holmgren et al. 1997; Niinemets & Valladares 2006). This tradeoff among the tolerances to 54 55 24 multiple environmental limitations hampers niche differentiation (Niinemets & Valladares 56 57 25 2006). Bartholomeus et al. (2011a) demonstrated that the interaction between both the wet 58 59 60 14 66 Global Change Biology Page 16 of 44

1 2 3 1 and dry extremes of plant water stress (oxygen/waterlogging and drought stress) is 4 5 2 particularly detrimental to the survival of specialists and of endangered plant species. Both 6 7 3 wet and dry weather extremes may increase due to changing climatic variability, thus 8 9 10 4 increasing the risk of a combination of these stressors to occur at a site (Knapp et al. 2008; 11 12 5 Bartholomeus et al. 2011a). This may favor generalists over specialists and rare species and 13 14 6 thus change vegetation dynamics and associated ecosystem services in response to changing 15 16 7 climatic variability at the community level. 17 18 For Review Only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 15 67 Page 17 of 44 Global Change Biology

1 2 3 1 3. How can we quantify responses of plants to changing climatic variability? 4 5 6 2 Just as responses to global change in general (Rustad 2008), the responses of plants to 7 8 3 changing climatic variability can be assessed in observational, experimental and modeling 9 10 11 4 studies and combinations of these approaches (Fig. 2). All these approaches have their 12 13 5 limitations in assessing a plant’s perspectives of extremes: on the one hand, observational 14 15 6 studies are by definition ‘opportunistic’ in the sense that extreme conditions such as a long 16 17 7 lasting drought can not be planned (Smith 2011). On the other hand, scaling and higherorder 18 For Review Only 19 8 interactions are an important issue in experimental and modeling studies (Leuzinger et al. 20 21 22 9 2011b; Wolkovich et al. 2012). Furthermore, it is crucial for any type of study that claims to 23 24 10 assess climate variability to report whether changing mean climate and/or changing climatic 25 26 11 variability have truly been measured and what the background variability of the system is. We 27 28 12 qualitatively show this in Table 2 for a number of studies cited above as a first attempt to 29 30 31 13 foster consistent reporting of studies dealing with climatic variability. 32 33 34 14 3.1. Observational studies 35 36 37 15 Observational studies elucidate plant’s perspectives of extremes, if by chance they cover 38 39 16 extremes. This makes them inherently opportunistic (Smith 2011). Similarly to experimental 40 41 17 studies, space can be substituted with time in observations. Thus, observations from ‘extreme’ 42 43 44 18 (from a plant’s perspective) sites (e.g. from the leading and trailing edge of population (Doak 45 46 19 & Morris 2010)) can help us learning about the limits and coping range of plants. To this end, 47 48 20 GIS mapping of ‘extreme’ sites within a species’ distribution requires careful interpolation of 49 50 21 weather/climate data collected at appropriately distributed climate stations (see also Sect. 3.3). 51 52 22 However, ‘extreme’ sites are sometimes only poorly studied since they represent marginal 53 54 55 23 ecosystems, whose services are not fully valued by society and have thus been outside the 56 57 24 main focus of researchers. The psamophilic plants and vegetation of the beaches and dunes of 58 59 60 16 68 Global Change Biology Page 18 of 44

1 2 3 1 the Portuguese coast, for example, are highly adapted to very specific environmental 4 5 2 conditions and directly exposed to sea level rise, storms and severe erosion processes. Unless 6 7 3 their ecological requirements, functioning as communities and most influential physical 8 9 10 4 drivers are understood, it will be difficult to study their responses to future climate change 11 12 5 (Martins et al. 2011). 13 14 6 Generally, observational studies are well suited to study plant responses to changing mean 15 16 17 7 climate, since longterm ecological data can be matched with increasingly available climatic 18 For Review Only 19 8 observations. They are less suitable to gain a mechanistic understanding of plant responses to 20 21 9 climatic variability since usually too many factors are involved and not all are measured. 22 23 24 25 10 3.2. Experimental studies 26 27 28 11 Experiments allow for controlled conditions and factorial experiments in the field and 29 30 12 laboratory, have a long history in ecological research and are of crucial importance for global 31 32 13 change studies (Luo et al. 2011)). When quantifying climate change impacts however, field 33 34 14 experiments can usually only test a limited number of factors and their combinations due to 35 36 15 financial and logistic constraints (Templer & Reinmann 2011). Therefore, interactions can 37 38 39 16 often not be fully assessed (e.g. Wolkovich et al. 2012). Furthermore, to provide answers to 40 41 17 the question of how extreme climatic events impact on ecosystems, experimenters should 42 43 18 make sure the applied treatment is indeed ‘extreme’ beyond the current background 44 45 19 variability of the system, running the risk of killing plants (Leuzinger & Thomas 2011; Beier 46 47 48 20 et al. 2012). 49 50 21 Also, the temporal scale influences the outcome of an experiment. A comparable set of factors 51 52 22 and a minimal experimental duration, for example, for all drought experiments would 53 54 55 23 therefore be desirable. However, even then, most experiments would have to stop after few 56 57 24 years. This raises the question whether the experiment actually simulates extreme situations 58 59 60 17 69 Page 19 of 44 Global Change Biology

1 2 3 1 or longterm change and whether the system recovers after the experiment ends. The high 4 5 2 diversity in the response of growth parameters of oaks to drought as discussed in Kuster et al. 6 7 3 (2012), shows that in experimental conditions, e.g. treatment duration and intensity, tree age 8 9 10 4 or experimental set up, have to be considered in the evaluation of drought effects on trees. 11 12 5 Thus it is crucial to assess what degree of change and what temporal scale experiments cover 13 14 6 if we want to evaluate whether they actually simulate responses to changing climatic 15 16 7 variability, or rather to changing mean climate. 17 18 For Review Only 19 8 In a transplantation study, for example, the effect of a drying and warming trend was obtained 20 21 9 by comparing tree performance in Rome (Italy), Tel Aviv (Israel) and Yatir (Israel) along a 22 23 10 precipitation gradient (Klein et al. submitted). The sites differed significantly in their mean 24 25 26 11 annual precipitation, each representing a different climate type, but the responses were 27 28 12 interpreted as drought acclimation. Results from this study captured many plant adjustments 29 30 13 that were induced by both phenotypic plasticity and locally adapted ecotypes. Such 31 32 14 transplantation experiments along altitudinal or latitudinal gradients do not require 33 34 35 15 manipulation of the environment and may be an alternative to laboratory/greenhouse 36 37 16 experiments. So far, transplantation experiments have not been considered in comparative 38 39 17 studies of different artificial warming methods (e.g. Aronson & McNulty 2009). However, 40 41 18 such experiments seem to be well adapted especially for long term experiments, as they 42 43 19 project a realistic simulation of future climate conditions considering also the length of the 44 45 46 20 growing period, one of the most important limiting factors in alpine plant growth (Jonas et al. 47 48 21 2008). Similar to laboratory/greenhouse experiments it is crucial that the results are 49 50 22 interpreted in terms of changing mean climate and changing variability. 51 52 53 54 55 56 57 58 59 60 18 70 Global Change Biology Page 20 of 44

1 2 3 1 3.3. Modeling 4 5 6 2 Models can be used as diagnostic and predictive tools that integrate results from experiments 7 8 3 and observation to gain mechanistic understanding and allow testing hypothesis generated 9 10 4 from field data, experiments and theory (Leuzinger & Thomas 2011; Luo et al. 2011). Models 11 12 5 have to be designed for a specific purpose and here we discuss which are suitable to simulate 13 14 6 plant responses to changing climate variability. This is a highly relevant question, since 15 16 17 7 models that account for extremes may require a different structure, e.g. an appropriate time 18 For Review Only 19 8 resolution to capture an extreme precipitation event. Many forest models for example use 20 21 9 monthly input data and are thus unable to account for shortterm extreme events (e.g. 22 23 10 Bugmann 2001). Forcing such a model with daily weather instead with monthly climate data 24 25 26 11 improved its performance (Stratton et al. 2012). Zimmermann et al. (2009) argue that for 27 28 12 capturing some ecosystem responses even daily climate data may be insufficient since they 29 30 13 smooth meteorological extremes. 31 32 33 14 Generally, effects of climate change on ecosystems are analyzed by driving simulation 34 35 15 models with output from global and regional circulation models (GCMs and RCMs). To 36 37 16 account for the uncertainty of climate change projections, besides different scenarios, also 38 39 17 several GCM/RCMs (e.g. Buisson et al. 2010) and different realizations of a scenario may be 40 41 42 18 used. Many models do not use the original GCM/RCM data at hourly resolution (which may 43 44 19 also not always be available) but only daily or monthly aggregations and thus strictly 45 46 20 speaking miss some of the meteorological variability. The CARAIB dynamic vegetation 47 48 21 model (Otto et al. 2002; Laurent et al. 2008; Dury et al. 2011), for example, derives daily 49 50 51 22 values of meteorological variables, as usual in largescale simulations, from monthly mean 52 53 23 outputs from GCM/RCMs using a stochastic weather generator (Hubert et al. 1998). The 54 55 24 sequences of daily temperature or precipitation produced by the stochastic generator are 56 57 25 renormalized to the monthly values generated by the RCMs. Thus the precise daytoday 58 59 60 19 71 Page 21 of 44 Global Change Biology

1 2 3 1 sequence of an extreme event in the model, such as a drought period or a succession of heat 4 5 2 wave days (Beniston et al. 2007; Déqué 2007), depends on the distribution functions used in 6 7 3 the stochastic generator, although the monthly values of the climate model are not altered. 8 9 10 4 While evidently it is challenging for such large scale modeling efforts to integrate high 11 12 5 frequency climate variability, these studies are necessary to assess different feedbacks of 13 14 6 vegetation types (e.g. feedbacks of ecosystem response to drying on nearsurface temperature 15 16 7 differ between forest and grassland ecosystems (Teuling et al. 2010) at the global scale. 17 18 For Review Only 19 8 Also, species distribution models face the challenge of including changing climate variability. 20 21 9 Usually, they use information on species distribution (both potential from expert knowledge 22 23 10 or forest communities, and actual from inventories and landcoverdata) together with climate 24 25 26 11 data to construct bioclimatic ranges (also called climate envelopes). They show a two 27 28 12 dimensional frequency distribution of e.g. temperature and precipitation, indicating the mean 29 30 13 climatic range, in which the analyzed species (potentially) exist. Extrapolation of this 31 32 14 information allows identifying regions with comparable climate to e.g. estimate a (extended) 33 34 35 15 potentially occupied habitat (Guisan & Zimmermann 2000) or new growing areas outside the 36 37 16 recent (actual or potential) distribution (Miller et al. 2004; Peters et al. 2004). Also the match 38 39 17 of actual and future suitable ranges can be identified, classifying species into tolerant or 40 41 18 intolerant to expected climatic conditions (Dunk et al. 2004; Gibson et al. 2004). This 42 43 19 provides further understanding about expanding or shrinking habitats under changing climate 44 45 46 20 (Erasmus et al. 2002; Midgley et al. 2006). Usually, climate envelopes are derived from mean 47 48 21 values (e.g. mean temperature) and are thus designed to assess impacts of changes in mean 49 50 22 climate. Consequently especially regions at the edge of the distribution range may appear 51 52 23 suitable, but in reality maximum or minimum precipitation or temperature may determine the 53 54 55 24 distribution range (or other, nonclimatic factors such as soil type or herbivory). This can 56 57 25 partly be circumvented by including standard deviations as variables (Zimmermann et al. 58 59 60 20 72 Global Change Biology Page 22 of 44

1 2 3 1 2009), and species distribution models could also be built with extremes (e.g. maximum 4 5 2 temperature or minimum precipitation) to enhance the predictive power. Zimmermann et al. 6 7 3 (2009) for example found that incorporating climatic extremes slightly improved models of 8 9 10 4 species range limits, since it corrected local over and underprediction, but they also argue 11 12 5 that climate variability rather complements the response to mean climate. Thus including 13 14 6 climate variability is one uncertainty of species distribution models that has to be considered 15 16 7 to assess compliance of climate envelopes (Gloning et al. in prep.). 17 18 For Review Only 19 8 While generally processbased modeling is required to derive climaterobust relationships to 20 21 9 predict vegetation characteristics (Franklin 1995; Guisan & Zimmermann 2000; Schwalm & 22 23 10 Ek 2001; Botkin et al. 2007; Suding et al. 2008; Hajar et al. 2010), this is even more evident 24 25 26 11 when considering changing climate variability in particular. Bartholomeus et al. (2011b) 27 28 12 demonstrated that, in contrast to processbased relationships between site factors and 29 30 13 vegetation characteristics, relations based on indirect site factors produce systematic 31 32 14 prediction errors when applied outside their calibration rate, and so cannot be used for climate 33 34 35 15 projections. Mean groundwater level, for example, is only an indirect site factor related to 36 37 16 plant performance, as it is the interaction between soilwaterplantatmosphere that essentially 38 39 17 determines if plants suffer from drought stress or oxygen/waterlogging stress. When, for 40 41 18 example, soil moisture availability is too low to meet the water demand for transpiration, a 42 43 19 plant suffers from drought stress (Reddy et al. 2004; Schimper 1903). This socalled 44 45 46 20 physiological drought (Schimper 1903), implies that not only water availability but also 47 48 21 vegetation’s demand for water has to be considered. Instead, more processbased explanatory 49 50 22 variables are needed to predict the effects of changing climate variability on the species 51 52 23 composition of the vegetation. These explanatory variables should consider the interacting 53 54 55 24 meteorological, soil physical, microbial, and plant physiological processes in the soilplant 56 57 25 atmosphere system. Bartholomeus et al. (2011a) did so for water related stressors, by 58 59 60 21 73 Page 23 of 44 Global Change Biology

1 2 3 1 simulating respiration reduction (reflecting the combined effect of high temperature and low 4 5 2 oxygen availability), and transpiration reduction (reflecting the combined effect of high 6 7 3 atmospheric water demand and low water availability) for a reference vegetation. The 8 9 10 4 simulated stress for reference vegetation acts as a habitat characteristic, i.e. a measure for the 11 12 5 moisture regime of the soil to which the actual vegetation will adapt. The use of reference 13 14 6 vegetation improves the applicability of models in which stress measures are implemented, 15 16 7 especially in predicting climate change effects (Dyer 2009). 17 18 For Review Only 19 20 8 3.4. Combined approaches 21 22 23 9 Combined approaches unite experimental, observational and/or modeling studies. A recent 24 25 10 metaanalysis shows that the temperature sensitivity of phenology in warming experiments is 26 27 11 underestimated in comparison to observations (Wolkovich et al. 2012). It highlights that 28 29 12 observational studies are crucial to test whether experimental results match observations in 30 31 32 13 natural systems. A combination of laboratory and field studies is necessary to determine 33 34 14 whether thresholds detected in the laboratory, are also likely to occur in the field. This is 35 36 15 especially relevant when calculating the effects of changing climatic variability. We take leaf 37 38 16 gas exchange and ecosystem flux measurement data from Brilli et al. (2011) as an example of 39 40 17 how to link experiments and observation at different scales and how an experiment can 41 42 43 18 complement observations to study plant responses to climate variability. Fig. 3 shows that 44 45 19 evapotranspiration measured in the field with the eddy covariance method, was insensitive to 46 47 20 soil drying over the range of soil water contents occurring in the field. The leaf gas exchange 48 49 21 measurements during the laboratory drought experiment when extended to much drier 50 51 52 22 conditions showed that the plant species occurring at this site start to downregulate stomatal 53 54 23 conductance at soil water contents close to the wilting point – conditions that have never been 55 56 24 reached in the field during the observational period of 20012009. Backoftheenvelope 57 58 25 calculations suggest that ca. 10 additional rainfree days would have been required even 59 60 22 74 Global Change Biology Page 24 of 44

1 2 3 1 during the 2003 and 2006 droughts in order for plants at this site to experience gas exchange 4 5 2 limitations. Such information is crucial to assess whether responses to changing mean climate 6 7 3 or to changing climate variability are measured. 8 9 10 4 Moreover results can be extended to a larger spatial scale, by combining simulation models 11 12 5 with research tools like raster GIS (Minacapilli et al. 2009; Bonfante et al. 2011) and Digital 13 14 6 Elevation Model (DEM) derived analysis (MacMillan et al. 2000). Furthermore, studies that 15 16 17 7 combine observational or experimental results at field scale with simulation models of 18 For Review Only 19 8 hydrothermal regime at landscape scale allow to quantify the effects of changing climate 20 21 9 variability (Bonfante et al. 2010). Riccardi et al. (2011) assessed the adaptive capacity of 22 23 10 olive cultivars to future climate by means of a data base of cultivars’ climatic requirements, 24 25 26 11 combined with a spatially distributed model of the soil–plant–atmosphere system. They set up 27 28 12 a database on climatic requirements and defined critical environmental conditions using two 29 30 13 quantitative indicators of soil water availability (the relative evapotranspiration deficit, i.e. the 31 32 14 ratio of actual to maximum evapotranspiration of the crop, and the relative soil water deficit, 33 34 35 15 i.e. the ratio between the actual and the maximum volume of soil water available to plants 36 37 16 taking into account the water retention characteristics, to get a comparable indicator across 38 39 17 soil types). The response in terms of yield of several olive cultivars to these indicators was 40 41 18 determined through the reanalysis of experimental data derived from scientific literature 42 43 19 (Moriana et al. 2003; Tognetti et al. 2006). This database on cultivars’ requirements was used 44 45 46 20 in combination with a plantsoilatmosphere model (SWAP, van Dam et al. 2008). The model 47 48 21 was used to describe the soil water regime at landscape scale under future climate scenarios 49 50 22 from statistically downscaled GCMs, resulting in several realizations (Tomozeiu et al. 2007). 51 52 23 The indicators of soil water availability were thus determined in different soil units, and were 53 54 55 24 compared with the limits set for each cultivar. A cultivar was considered tolerant to expected 56 57 25 climatic conditions when the indicator values resulted above critical values in at least 90% of 58 59 60 23 75 Page 25 of 44 Global Change Biology

1 2 3 1 realizations. While Riccardi et al. (2011) did not further specify the climate scenarios and 4 5 2 realizations in terms of changing mean or climate variability, such analysis could be easily 6 7 3 linked to the soil water availability indicators and the related limits for cultivars under climate 8 9 10 4 change. 11 12 13 14 15 16 17 18 For Review Only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 24 76 Global Change Biology Page 26 of 44

1 2 3 1 4. Conclusions 4 5 6 2 In this review, we have emphasized that changing climatic variability and the resulting 7 8 3 extreme (climatic) conditions are highly relevant for different plant processes at different 9 10 11 4 scales in comparison to changes in mean climate. We have also shown how to quantify 12 13 5 responses of plants to changing climate variability: While experiments seem to be wellsuited 14 15 6 to study the effects of changing climatic variability it is important to remember that they only 16 17 7 control a limited number of factors. For modeling studies we stress that the model structure 18 For Review Only 19 8 should allow integrating extreme events (e.g. by having the appropriate temporal resolution). 20 21 22 9 These points highlight the importance of linking experiments, observations, and modeling 23 24 10 studies. We also identified the several research gaps. While knowledge of plant responses to 25 26 11 changing climatic variability for individual processes has to be consolidated, we still lack 27 28 12 knowledge on how interactions of these processes and other environmental variables play out 29 30 31 13 at different hierarchical levels and in combination with changing mean climatic conditions. 32 33 14 Similarly, while there is room to improve individual methods to study changing climatic 34 35 15 variability, there is a particular need to integrate observations, experiments and models results 36 37 16 across scales. 38 39 40 17 Ultimately, the information on extremes and corresponding vulnerability of plants are crucial 41 42 18 to identify which species and regions (and thus which ecosystem services and functions) are 43 44 19 most at risk from future climate change. Moreover, designing ecosystembased adaptation 45 46 47 20 strategies to climate change relies on understanding the interactions between species natural 48 49 21 adaptive capacity and climate change. Analyzing plant responses to climate variability is 50 51 22 important to determine drivers of ecosystem dynamics over time (slow vs. fast processes) and 52 53 23 highlights the importance of extremes to assess the impacts of environmental change on 54 55 24 socioecological systems. 56 57 58 25 59 60 25 77 Page 27 of 44 Global Change Biology

1 2 3 1 Acknowledgement 4 5 6 2 This review synthesizes the results from a session which was held during the 2011 European 7 8 3 Geoscience Union (EGU) general assembly (BG2.7). CR acknowledges funding from the EC 9 10 4 FP7 MOTIVE project (grant agreement no. 226544). SL was funded by EC FP7 ACQWA. 11 12 5 AR acknowledges funding from the EU FP7 project CARBOExtreme (grant agreement no. 13 14 6 226701). The work of FdL carried out within the Italian national project AGROSCENARI 15 16 17 7 (MIPAAF, D.M. 8608/7303/2008). 18 For Review Only 19 8 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 26 78 Global Change Biology Page 28 of 44

1 2 3 1 References 4 5 6 2 Abou Jaoudé R, De Dato G, Palmegiani M, De Angelis P (in press) Impact of fresh and saline 7 8 3 water flooding on leaf gas exchange in two Italian provenances of Tamarix africana Poiret. 9 4 Plant Biology. 10 11 5 Albert KR, RoPoulsen H, Mikkelsen TN, Michelsen A, Van Der Linden L, Beier C (2011) 12 13 6 Effects of elevated CO2, warming and drought episodes on plant carbon uptake in a temperate 14 7 heath ecosystem are controlled by soil water status. Plant, Cell & Environment, 34, 1207 15 16 8 1222. 17 9 Allen CD, Breshears DD (1998) Droughtinduced shift of a forest woodland ecotone: Rapid 18 For Review Only 19 10 landscape response to climate variation. Proceedings of the National Academy of Sciences, 20 21 11 95, 1483914842. 22 23 12 Allen CD, Macalady AK, Chenchouni H et al. (2010) A global overview of drought and heat 24 13 induced tree mortality reveals emerging climate change risks for forests. Forest Ecology and 25 26 14 Management, 259, 660684. 27 28 15 Amthor JF (2000) The McCree–de Wit–Penning de Vries–Thornley respiration paradigms: 30 29 16 Years Later. Annales of Botany, 86, 120. 30 31 17 Aronson EL, Mcnulty SG (2009) Appropriate experimental ecosystem warming methods by 32 33 18 ecosystem, objective, and practicality. Agricultural and Forest Meteorology, 149, 1791–1799. 34 19 Atzmon N, Moshe Y, Schiller G (2004) Ecophysiological response to severe drought in Pinus 35 36 20 halepensis Mill. trees of two provenances. Plant Ecology, 171, 1522. 37 38 21 Barriopedro D, Fischer EM, Luterbacher J, Trigo RM, GarcíaHerrera R (2011) The Hot 39 22 Summer of 2010: Redrawing the Temperature Record Map of Europe. Science, 332, 220224. 40 41 23 Bartholomeus R, Witte JP, Van Bodegom P, Van Dam J, Aerts R (2011a) Climate change 42 43 24 threatens endangered plant species by stronger and interacting waterrelated stresses. Journal 44 25 of Geophysical Research, 116, G04023. 45 46 26 Bartholomeus R, Witte JP, Van Bodegom P, Van Dam J, De Becker P, Aerts R (2011b) 47 48 27 Processbased proxy of oxygen stress surpasses indirect ones in predicting vegetation 49 28 characteristics. Ecohydrology. 50 51 29 Bartholomeus R, Witte JP, Van Bodegom P, Van Dam J, Aerts R (2008) Critical soil 52 53 30 conditions for oxygen stress to plant roots: Substituting the Feddesfunction by a process 54 31 based model. Journal of Hydrology, 360, 147165. 55 56 32 Beier C, Beierkuhnlein C, Wohlgemuth T et al. (2012) Precipitation manipulation 57 58 33 experiments – challenges and recommendations for the future. Ecology Letters. 59 60 27 79 Page 29 of 44 Global Change Biology

1 2 3 1 Beniston M, Stephenson D, Christensen O et al. (2007) Future extreme events in European 4 5 2 climate: an exploration of regional climate model projections. Climatic Change, 81, 7195. 6 3 Bertoldi G, Della Chiesa S, Niedrist G, Hell V, Tasser E, Tappeiner U (2011) Modelling the 7 8 4 water and energy budget in alpine environments at plot and catchment scale. Geophysical 9 10 5 Research Abstracts, 13, EGU201110152. 11 6 Blom C, Voesenek L (1996) Flooding: the survival strategies of plants. Trees, 11, 290295. 12 13 7 Blumwald E, Aharon G, Apse M (2000) Sodium transport in plant cells. Biochimica et 14 15 8 Biophysica Acta, 1465, 140151. 16 9 Boisvenue C, Running S (2006) Impacts of climate change on natural forest productivity – 17 18 10 evidence since the Formiddle of the Review 20th century. Global ChangeOnly Biology, 12, 862882. 19 20 11 Bokhorst S, Huiskes A, Convey P, Aerts R (2007) The effect of environmental change on 21 12 vascular plant and cryptogam communities from the Falkland Islands and the Maritime 22 23 13 Antarctic. BMC Ecology, 7, 15. 24 25 14 Bonfante A, Basile A, De Lorenzi F, Langella G, Terribile F, Menenti M (2010) The 26 15 adaptative capacity of a viticultural area (Valle Telesina, Southern Italy) to climate changes. 27 28 16 In: Proceedings VIII International Terroir Congress, June 14-18 Soave (VR) Italy, 96101 pp. 29 30 17 Bonfante A, Basile A, Langella G, Manna P, Terribile F (2011) A physically oriented 31 18 approach to analysis and mapping of terroirs. Geoderma, 167-168, 103117. 32 33 19 Botkin DB, Saxe H, Araujo MB et al. (2007) Forecasting the Effects of Global Warming on 34 35 20 Biodiversity. Bioscience, 57, 227236. 36 21 Bréda N, Huc R, Granier A, Dreyer E (2006) Temperate forest trees and stands under severe 37 38 22 drought: a review of ecophysiological responses, adaptation processes and longterm 39 40 23 consequences. Annals of Forest Science, 63, 625644. 41 24 Brilli F, Hörtnagl L, Hammerle A, Haslwanter A, Hansel A, Loreto F, Wohlfahrt G (2011) 42 43 25 Leaf and ecosystem response to soil water availability in mountain grasslands. Agricultural 44 45 26 and Forest Meteorology, 151, 17311740. 46 27 Brooks P, Vivoni E (2008) Mountain ecohydrology: quantifying the role of vegetation in the 47 48 28 water balance of montane catchments. Ecohydrology, 1, 187–192. 49 50 29 Bugmann H (2001) A Review of Forest Gap Models. Climatic Change, 51, 259305. 51 30 Buisson L, Thuiller W, Casajus N, Lek S, Grenouillet G (2010) Uncertainty in ensemble 52 53 31 forecasting of species distribution. Global Change Biology, 16, 11451157. 54 55 32 Casper B, Jackson R (1997) Plant Competition Underground. Annual Review of Ecology and 56 33 Systematics, 28, 545570. 57 58 34 Chapin III F (1991) Integrated Responses of Plants to Stress. Bioscience, 41, 2936. 59 60 28 80 Global Change Biology Page 30 of 44

1 2 3 1 Chapin III F, Rincon E, Huante P (1993) Environmental responses of plants and ecosystems 4 5 2 as predictors of the impact of global change. Journal of Biosciences, 18, 515524. 6 3 Chen H. Qualls R, Blank R (2005) Effect of soil flooding on photosynthesis, carbohydrate 7 8 4 partitioning and nutrient uptake in the invasive exotic Lepidium latifolium. Acquatic Botany, 9 10 5 82, 250268. 11 6 Ciais P, Reichstein M, Viovy N et al. (2005) Europewide reduction in primary productivity 12 13 7 caused by the heat and drought in 2003. Nature, 437, 529533. 14 15 8 Colmer T, Flowers T (2008) Flooding tolerance in halophytes. New Phytologist, 179, 964 16 9 974. 17 18 10 Craine JM, NippertFor JB, Elmore ReviewAJ, Skibbe AM, Hutchinson Only SL, Brunsell NA (2012) Timing 19 20 11 of climate variability and grassland productivity. Proceedings of the National Academy of 21 12 Sciences. 22 23 13 Crawford R (1992) Oxygen availability as an ecological limit to plant distribution. Advances 24 25 14 in Ecological Research, 23, 93185. 26 15 De Boeck HJ, Dreesen FE, Janssens IA, Nijs I (2011) Wholesystem responses of 27 28 16 experimental plant communities to climate extremes imposed in different seasons. New 29 30 17 Phytologist, 189, 806817. 31 18 Déqué M (2007) Frequency of precipitation and temperature extremes over France in an 32 33 19 anthropogenic scenario: Model results and statistical correction according to observed values. 34 35 20 Global and Planetary Change, 57, 1626. 36 21 Doak DF, Morris WF (2010) Demographic compensation and tipping points in climate 37 38 22 induced range shifts. Nature, 467, 959962. 39 40 23 Dunk J, Zielinski W, Preisler H (2004) Predicting the occurrence of rare mollusks in northern 41 24 California forests. Ecological Applications, 14, 713729. 42 43 25 Dury M, Hambuckers A, Warnant P, Henrot A, Favre E, Ouberdous M, François L (2011) 44 45 26 Responses of European forest ecosystems to 21st century climate: assessing changes in 46 27 interannual variability and fire intensity. iForest, 4, 8299. 47 48 28 Dyer J (2009) Assessing topographic patterns in moisture use and stress using a water balance 49 50 29 approach. Landscape Ecology, 24, 391403. 51 30 Erasmus BFN, Van Jaarsveld AS, Chown SL, Kshatriya M, Wessels KJ (2002) Vulnerability 52 53 31 of South African animal taxa to climate change. Global Change Biology, 8, 679693. 54 55 32 Fay P, Kaufman D, Nippert J, Carlisle J, Harper C (2008) Changes in grassland ecosystem 56 33 function due to extreme rainfall events: implications for responses to climate change. Global 57 58 34 Change Biology, 14, 16001608. 59 60 29 81 Page 31 of 44 Global Change Biology

1 2 3 1 Fernandez M (2006) Changes in photosynthesis and fluorescence in response to flooding in 4 5 2 emerged and submerged leaves of Pouteria orinocoensis. Photosynthetica, 44, 3238. 6 3 Fischer E, Seneviratne S, Vidale P, Luthi D, Schar C (2007) Soil moisture Atmosphere 7 8 4 interactions during the 2003 European summer heat wave. Journal of Climate, 20, 50815099. 9 10 5 Franklin J (1995) Predictive vegetation mapping: geographic modelling of biospatial patterns 11 6 in relation to environmental gradients. Progress in Physical Geography, 19, 474499. 12 13 7 Gaspar T, Franck T, Bisbis B, Kevers C, Jouve L, Hausman J, Dommes J (2002) Concepts in 14 15 8 plant stress physiology. Application to plant tissue cultures. Plant Growth Regulation, 37, 16 9 263285. 17 18 10 Gibson LA, WilsonFor BA, Cahill ReviewDM, Hill J (2004) Modelling Only habitat suitability of the swamp 19 20 11 antechinus (Antechinus minimus maritimus) in the coastal heathlands of southern Victoria, 21 12 Australia. Biological Conservation, 117, 143150. 22 23 13 Giorgi F, Bi X, Pal J (2004) Mean, interannual variability and trends in a regional climate 24 25 14 change experiment over Europe. II: climate change scenarios (2071–2100). Climate 26 15 Dynamics, 23, 839–858. 27 28 16 Gitlin A, Sthultz C, Bowker M et al. (2006) Mortality gradients within and among dominant 29 30 17 plant populations as barometers of ecosystem change during extreme drought. Conservation 31 18 Biology, 20, 1477–1486. 32 33 19 Gloning P, Taeger S, Seifert H, Schäffler U, Kölling C, Schilcher M, Menzel A (in 34 35 20 preparation) Comparison of four different types of European tree species distribution maps 36 21 and their correlation to weather and climate. 37 38 22 Godefroid S, Rivière S, Waldren S, Boretos N, Eastwood R, Vanderborght T (2011) To what 39 40 23 extent are threatened European plant species conserved in seed banks? Biological 41 24 Conservation, 144, 1494–1498. 42 43 25 Granier A, Reichstein M, Bréda N et al. (2007) Evidence for soil water control on carbon and 44 45 26 water dynamics in European forests during the extremely dry year: 2003. Agricultural and 46 27 Forest Meteorology, 143, 123145. 47 48 28 Guisan A, Zimmermann NE (2000) Predictive habitat distribution models in ecology. 49 50 29 Ecological Modelling, 135, 147186. 51 30 Hajar L, François L, Khater C, Jomaa I, Déqué M, Cheddadi R (2010) Cedrus libani (A. 52 53 31 Rich) distribution in Lebanon: Past, present and future. Comptes Rendus Biologies, 333, 622 54 55 32 630. 56 33 Harfouche A, Meilan R, Altman A (2011) Tree genetic engineering and applications to 57 58 34 sustainable forestry and biomass production. Trends in Biotechnology, 29, 917. 59 60 30 82 Global Change Biology Page 32 of 44

1 2 3 1 Harsch MA, Hulme PE, McGlone MS, Duncan RP (2009) Are treelines advancing? A global 4 5 2 metaanalysis of treeline response to climate warming. Ecology Letters, 12, 10401049 6 3 Hartmann H (2011) Will a 385 million yearstruggle for light become a struggle for water and 7 8 4 for carbon? – How trees may cope with more frequent climate changetype drought events. 9 10 5 Global Change Biology, 17, 642655. 11 6 Hegerl GC, Hanlon H, Beierkuhnlein C (2011) Climate science: Elusive extremes. Nature 12 13 7 Geosciences, 4, 142143. 14 15 8 Heide OM (1993) Dormancy release in beech buds (Fagus sylvatica) requires both chilling 16 9 and long days. Physiologia Plantarum, 89, 187191. 17 18 10 Holmgren M, SchefferFor M, Huston Review MA (1997) The Interplay Only of Facilitation and Competition 19 20 11 in Plant Communities. Ecology, 78, 19661975. 21 12 Holtum J, Winter K (2010) Elevated CO2 and forest vegetation: more a water issue than a 22 23 13 carbon issue? Functional Plant Biology, 37, 694702. 24 25 14 Hubert B, Francois L, Warnant P, Strivay D (1998) Stochastic generation of meteorological 26 15 variables and effects on global models of water and carbon cycles in vegetation and soils. 27 28 16 Journal of Hydrology, 212, 318334. 29 30 17 IPCC (2007a) Climate Change 2007: The Physical Science Basis. Contribution of Working 31 18 Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. 32 33 19 (eds Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt K, Tignor M, Miller H), 34 35 20 Cambridge University Press, Cambridge, UK and New York, USA, 996 pp. 36 21 IPCC (2007b) Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of 37 38 22 Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on 39 40 23 Climate Change. (eds Parry M, Canziani O, Palutikof J, Van Der Linden P, Hanson C), 41 24 Cambridge University Press, Cambridge, UK, 976 pp. 42 43 25 Jackson M, Drew M (1984) Effect of flooding on herbaceous plants. In: Flooding and plant 44 45 26 growth. (ed Kozlowski T), 47128 pp. Academic Press, London. 46 27 Jentsch A, Beierkuhnlein C (2008) Research frontiers in climate change: Effects of extreme 47 48 28 meteorological events on ecosystems. Comptes Rendus Geoscience, 340, 621628. 49 50 29 Jentsch A, Kreyling J, BoettcherTreschkow J, Beierkuhnlein C (2009) Beyond gradual 51 30 warming: extreme weather events alter flower phenology of European grassland and heath 52 53 31 species. Global Change Biology, 15, 837849. 54 55 32 Jochner SC, Beck I, Behrendt H, TraidlHoffmann C, Menzel A (2011) Effects of extreme 56 33 spring temperatures on urban phenology and pollen production: a case study in Munich and 57 58 34 Ingolstadt. Climate Research, 49, 101112. 59 60 31 83 Page 33 of 44 Global Change Biology

1 2 3 1 Jonas T, Rixen C, Sturm M, Stoeckli V (2008) How alpine plant growth is linked to snow 4 5 2 cover and climate variability. Journal of Geophysical Research, 113, G03013. 6 3 Kelly A, Goulden M (2008) Rapid shifts in plant distribution with recent climate change. 7 8 4 Proceedings of the National Academy of Sciences, 105, 11823–11826. 9 10 5 Klein T, Di Matteo G, Rotenberg E, Cohen S, Yakir D (submitted) Existing traits’ plasticity 11 6 will sustain growth and survival of a major pine species under predicted climate change in the 12 13 7 Mediterranean. 14 15 8 Knapp A, Beier C, Briske D et al. (2008) Consequences of more extreme precipitation 16 9 regimes for terrestrial ecosystems. Bioscience, 58, 811821. 17 18 10 Knapp A, Fay PA, ForBlair JM et al.Review (2002) Rainfall Variability, Only Carbon Cycling, and Plant 19 20 11 Species Diversity in a Mesic Grassland. Science, 298, 22022205. 21 12 Körner C (1998) Alpine plants: stressed or adapted? In: Physiological Plant Ecology (eds 22 23 13 Scholes Jd, Barker Mg), 297311 pp. Blackwell Science, Oxford, UK. 24 25 14 Körner C (2003) Limitation and stress always or never? Journal of Vegetation Science, 14, 26 15 141143. 27 28 16 Körner C, Basler D (2010) Phenology Under Global Warming. Science, 327, 14611462. 29 30 17 Kozlowski T (1997) Responses of woody plants to flooding and salinity. Tree Physiology 31 18 Monograph, 1, 129. 32 33 19 Kramer K, Degen B, Buschbom J, Hickler T, Thuiller W, Sykes MT, De Winter W (2010) 34 35 20 Modelling exploration of the future of European beech (Fagus sylvatica L.) under climate 36 21 changeRange, abundance, genetic diversity and adaptive response. Forest Ecology and 37 38 22 Management, 259, 22132222 39 40 23 Kramer P (1951) Causes of injury to plants resulting from flooding of the soil. Plant 41 24 Physiology, 26, 722736. 42 43 25 Kramer P, Boyer J (1995) Water relations of plant and soils. Academic Press, San Diego, 495 44 45 26 pp. 46 27 Kuster TM, Arend M, Bleuler P, GünthardtGoerg MS, Schulin R (2012) Water regime and 47 48 28 growth of young oak stands subjected to airwarming and drought on two different forest soils 49 50 29 in a model ecosystem experiment. Plant Biology. 51 30 Langley JA, Megonigal JP (2010) Ecosystem response to elevated CO2 levels limited by 52 53 31 nitrogeninduced plant species shift. Nature, 466, 9699. 54 55 32 Laurent J, Francois L, BarHen A, Bel L, Cheddadi R (2008) European bioclimatic affinity 56 33 groups: Datamodel comparisons. Global and Planetary Change, 61, 2840. 57 58 59 60 32 84 Global Change Biology Page 34 of 44

1 2 3 1 Lenoir J, Gégout J, Marquet P, De Ruffray P, Brisse H (2008) A Significant Upward Shift in 4 5 2 Plant Species Optimum Elevation During the 20th Century. Science, 320, 17681771. 6 3 Leuzinger S, Hartmann A, Körner C (2011a) Water relations of climbing ivy in a temperate 7 8 4 forest. Planta, 233, 10871096. 9 10 5 Leuzinger S, Körner C (2010) Rainfall distribution is the main driver of runoff under future 11 6 CO2concentration in a temperate deciduous forest. Global Change Biology, 16, 246254. 12 13 7 Leuzinger S, Luo Y, Beier C, Dieleman W, Vicca S, Körner C (2011b) Do global change 14 15 8 experiments overestimate impacts on terrestrial ecosystems? Trends In Ecology & Evolution, 16 9 26, 236241. 17 18 10 Leuzinger S, ThomasFor R (2011) ReviewHow do we improve Earth Only system models? Integrating Earth 19 20 11 system models, ecosystem models, experiments and longterm data. New Phytologist, 191, 21 12 1518. 22 23 13 Leuzinger S, Zotz G, Asshoff R, Körner C (2005) Responses of deciduous forest trees to 24 25 14 severe drought in Central Europe. Tree Physiology, 25, 641650. 26 15 Levine J, Rees M (2004) Effects of Temporal Variability on Rare Plant Persistence in Annual 27 28 16 Systems. The American Naturalist, 164, 350363. 29 30 17 Levine JM, Mceachern AK, Cowan C (2008) Rainfall effects on rare annual plants. Journal of 31 18 Ecology, 96, 795806. 32 33 19 Lindner M, Maroschek M, Netherer S et al. (2010) Climate change impacts, adaptive 34 35 20 capacity, and vulnerability of European forest ecosystems. Forest Ecology and Management, 36 21 259, 698709. 37 38 22 Lloyd J, Taylor J (1994) On the temperature dependence of soil respiration. Functional 39 40 23 Ecology, 8, 315323. 41 24 Lortie CJ, Brooker RW, Kikvidze Z, Callaway RM (2004) The value of stress and limitation 42 43 25 in an imperfect world: A reply to Körner. Journal of Vegetation Science, 15, 577580. 44 45 26 Luo Y, Melillo J, Niu S et al. (2011) Coordinated approaches to quantify longterm 46 27 ecosystem dynamics in response to global change. Global Change Biology, 17, 843854. 47 48 28 Luterbacher J, Liniger MA, Menzel A, Estrella N, DellaMarta PM, Pfister C, Rutishauser T, 49 50 29 Xoplaki E (2007) Exceptional European warmth of autumn 2006 and winter 2007: Historical 51 30 context, the underlying dynamics, and its phenological impacts. Geophysical Research 52 53 31 Letters, 34, L12704. 54 55 32 Macmillan R, Pettapiece W, Nolan S, Goddard T (2000) Generic procedure for automatically 56 33 segmenting landforms into landform elements using DEMS, heuristic rules and fuzzy logic. 57 58 34 Fuzzy Sets and Systems, 113, 81109. 59 60 33 85 Page 35 of 44 Global Change Biology

1 2 3 1 Martins M, Neto C, Costa J (in press) The meaning of mainland Portugal beaches and dunes’ 4 5 2 psammophilic plant communities: a contribution to tourism management and nature 6 3 conservation. Journal of Coastal Conservation. 7 8 4 Matyas C (1996) Climatic adaptation of trees: rediscovering provenance tests. Euphytica, 92, 9 10 5 4554. 11 6 Matyssek R, Wieser G, Ceulemans R et al. (2010) Enhanced ozone strongly reduces carbon 12 13 7 sink strength of adult beech (Fagus sylvatica) Resume from the freeair fumigation study at 14 15 8 Kranzberg Forest. Environmental Pollution, 158, 25272532. 16 9 Mcdowell N, Pockman WT, Allen CD et al. (2008) Mechanisms of plant survival and 17 18 10 mortality during drought:For why doReview some plants survive whileOnly others succumb to drought? New 19 20 11 Phytologist, 178, 719739. 21 12 Menzel A, Fabian P (1999) Growing season extended in Europe. Nature, 397, 659659. 22 23 13 Menzel A, Jakobi G, Ahas R, Scheifinger H, Estrella N (2003) Variations of the 24 25 14 climatological growing season (1951–2000) in Germany compared with other countries. 26 15 International Journal of Climatology, 23, 793812. 27 28 16 Menzel A, Seifert H, Estrella N (2011) Effects of recent warm and cold spells on European 29 30 17 plant phenology. International Journal of Biometeorology, 55, 921932. 31 18 Menzel A, Sparks TH, Estrella N et al. (2006) European phenological response to climate 32 33 19 change matches the warming pattern. Global Change Biology, 12, 19691976. 34 35 20 Midgley GF, Hughes GO, Thuiller W, Rebelo A G. (2006) Migration rate limitations on 36 21 climate changeinduced range shifts in Cape Proteaceae. Diversity And Distributions, 12, 555 37 38 22 562. 39 40 23 Miller JR, Turner MG, Smithwick EAH, Dent CL, Stanley EH (2004) Spatial Extrapolation: 41 24 The Science of Predicting Ecological Patterns and Processes. Bioscience, 54, 310320. 42 43 25 Minacapilli M, Agnese G, Blanda F et al. (2009) Estimation of actual evapotranspiration of 44 45 26 Mediterranean perennial crops by means of remotesending based surface energy balance 46 27 models. Hydrology and Earth System Sciences, 13, 10611074. 47 48 28 Misson L, Degueldre D, Collin C, Rodriguez R, Rocheteau A, Ourcival JM, Rambal S 49 50 29 (2011) Phenological responses to extreme droughts in a Mediterranean forest. Global Change 51 30 Biology, 17, 10361048. 52 53 31 Morgan JA, Pataki DE, Körner C et al. (2004) Water relations in grassland and desert 54 55 32 ecosystems exposed to elevated atmospheric CO2. Oecologia, 140, 1125. 56 33 Moriana O, Orgaz F, Fereres E, Pastor M (2003) Yield responses of a mature olive orchard to 57 58 34 water deficits. Journal of the American Society of Horticultural Science, 128, 425431. 59 60 34 86 Global Change Biology Page 36 of 44

1 2 3 1 Neppel L, Desbordes M, Masson J (1997) Spatial extension of extreme rainfall events: return 4 5 2 period of isohyets area and influence of rain gauges network evolution. Atmospheric 6 3 resaearch, 45, 183199. 7 8 4 Niinemets U, Valladares F (2006) Tolerance to shade, drought, and waterlogging of temperate 9 10 5 northern hemisphere trees and shrubs. Ecological Monographs, 76, 521547. 11 6 Orlowsky B, Seneviratne S (2012) Global changes in extreme events: regional and seasonal 12 13 7 dimension. Climatic Change, 110, 669696. 14 15 8 Orlowsky B, Seneviratne S (2011) Investigating spatial climate relations using CARTs: An 16 9 application to persistent hot days in a multimodel ensemble. Journal of Geophysical 17 18 10 Research, 116, D14106.For Review Only 19 20 11 Otto D, Rasse D, Kaplan J, Warnant P, Francois L (2002) Biospheric carbon stocks 21 12 reconstructed at the Last Glacial Maximum: comparison between general circulation models 22 23 13 using prescribed and computed sea surface temperatures. Global Planetary Change, 33, 117 24 25 14 138. 26 15 Parmesan C, Yohe G (2003) A globally coherent fingerprint of climate change impacts across 27 28 16 natural systems. Nature, 421, 3742. 29 30 17 Parolin P, Wittmann F (2010) Struggle in the flood: tree responses to flooding stress in four 31 18 tropical floodplain systems. AoB Plants, plq003. 32 33 19 Pasolli L, Notarnicola C, Bruzzone L et al. (2011) Estimation of Soil Moisture in an Alpine 34 35 20 Catchment with RADARSAT2 Images. Applied and Environmental Soil Science, Article ID 36 21 175473. 37 38 22 Peters D, Herrick J (2004) Strategies for ecological extrapolation. Oikos, 106, 627636. 39 40 23 Polgar CA, Primack RB (2011) Leafout phenology of temperate woody plants: from trees to 41 24 ecosystems. New Phytologist, 191, 926–941. 42 43 25 Porporato A, Daly E, RodriguezIturbe I (2004) Soil water balance and ecosystem response to 44 45 26 climate change. American Naturalist, 164, 625632. 46 27 Pretzsch H, Dieler J (2011) The dependency of the sizegrowth relationship of Norway spruce 47 48 28 (Picea abies [L.] Karst.) and European beech (Fagus sylvatica [L.]) in forest stands on long 49 50 29 term site conditions, drought events, and ozone stress. Trees – Structure and Function, 25, 51 30 355369. 52 53 31 Rahmstorf S, Coumou D (2011) Increase of extreme events in a warming world. Proceedings 54 55 32 of the National Academy of Sciences, 108, 1790517909. 56 33 Ramadan T (1998) Ecophysiology of salt excretion in the xerohalophyte Reaumuria hirtella. 57 58 34 New Phytologist, 139, 273–281. 59 60 35 87 Page 37 of 44 Global Change Biology

1 2 3 1 RazYaseef N, Yakir D, Rotenberg E, Schiller G, Cohen S (2010) Ecohydrology of a semi 4 5 2 arid forest: partitioning among water balance components and its implications for predicted 6 3 precipitation changes. Ecohydrology, 3, 143154. 7 8 4 Reddy A, Chaitanya K, Vivekanandan M (2004) Droughtinduced responses of 9 10 5 photosynthesis and antioxidant metabolism in higher plants. Journal of Plant Physiology, 11 6 161, 11891202. 12 13 7 Rehfeldt G, Wykoff W, Ying C (2001) Physiologic Plasticity, Evolution, and Impacts of a 14 15 8 Changing Climate on Pinus Contorta. Climatic Change, 50, 355376. 16 9 Rengifo E, Tezara W, Herrera A (2005) Water relations, chlorophyll a fluorescence, and 17 18 10 contents of saccharidesFor in tree speciesReview of a tropical forest Only in response to flood. 19 20 11 Photosynthetica, 43, 203210. 21 12 Riccardi M, Bonfante A, Mencuccini M, Di Tommasi P, De Lorenzi F, Menenti M (2011) 22 23 13 Assessing the adaptive capacity of some olive cultivars to future climate. Geophysical 24 25 14 Research Abstracts, 13, 10688. 26 15 Rigon R, Bertoldi G, Over T (2006) GEOtop: a distributed hydrological model with coupled 27 28 16 water and energy budgets. Journal of Hydrometeorology, 7, 371–388. 29 30 17 Root TL, Price JT, Hall KR, Schneider SH, Rosenzweig C, Pounds JA (2003) Fingerprints of 31 18 global warming on wild animals and plants. Nature, 421, 5760. 32 33 19 Rummukainen M (2012) Changes in climate and weather extremes in the 21st century. Wiley 34 35 20 Interdisciplinary Reviews: Climate Change, 3, 115129. 36 21 Rustad LE (2008) The response of terrestrial ecosystems to global climate change: Towards 37 38 22 an integrated approach. Science of The Total Environment, 404, 222235. 39 40 23 Sala A, Piper F, Hoch G (2010) Physiological mechanisms of droughtinduced tree mortality 41 24 are far from being resolved. New Phytologist, 186, 274281. 42 43 25 Saxe H, Cannell MGR, Johnsen Ø, Ryan MG, Vourlitis G (2001) Tree and forest functioning 44 45 26 in response to global warming. New Phytologist, 149, 369399. 46 27 Schär C, Vidale P, Lüthi D, Frei C, Häberli C, Liniger M, Appenzeller C (2004) The role of 47 48 28 increasing temperature variability in European summer heatwaves. Nature, 427, 332336. 49 50 29 Scheifinger H, Menzel A, Koch E, Peter C (2003) Trends of spring time frost events and 51 30 phenological dates in Central Europe. Theoretical and Applied Climatology, 74, 4151. 52 53 31 Schiller G, Atzmon N (2009) Performance of Aleppo pine (Pinus halepensis) provenances 54 55 32 grown at the edge of the Negev desert: A review. Journal of Arid Environments, 73, 1051 56 33 1057. 57 58 59 60 36 88 Global Change Biology Page 38 of 44

1 2 3 1 Schimper A (1903) Plant-geography upon a physiological basis. Clarendon Press, Oxford 4 5 2 839 pp. 6 3 Schwalm CR, Ek AR (2001) Climate change and site: relevant mechanisms and modeling 7 8 4 techniques. Forest Ecology and Management, 150, 241257. 9 10 5 Seneviratne SI, Luthi D, Litschi M, Schar C (2006) Landatmosphere coupling and climate 11 6 change in Europe. Nature, 443, 205209. 12 13 7 Seneviratne SI, N Nicholls, D Easterling et al. (2012) Changes in climate extremes and their 14 15 8 impacts on the natural physical environment. In: Managing the Risks of Extreme Events and 16 9 Disasters to Advance Climate Change Adaptation. A Special Report of Working Groups I and 17 18 10 II of the IntergovernmentalFor Panel Review on Climate Change (IPCC)Only (eds Field CB, Barros V, 19 20 11 Stocker TF, Qin D, Dokken DJ, Ebi KL, Mastrandrea MD, Mach KJ, Plattner GK, Allen SK, 21 12 Tignor M, Midgley PM), 109230 pp. Cambridge University Press, Cambridge, UK, and New 22 23 13 York, USA. 24 25 14 Silvertown J, Dodd M, Gowing D, Mountford J (1999) Hydrologically defined niches reveal a 26 15 basis for species richness in plant communities. Nature, 400, 6163. 27 28 16 Smith MD (2011) An ecological perspective on extreme climatic events: a synthetic 29 30 17 definition and framework to guide future research. Journal of Ecology, 99, 656663. 31 18 Smith T, Huston M (1989) A theory of the spatial and temporal dynamics of plant 32 33 19 communities. Plant Ecology, 83, 4969. 34 35 20 Stratton T, Price DT, Gajewski K (2012) Impacts of daily weather variability on simulations 36 21 of the Canadian boreal forest. Ecological Modelling, 222, 32503260. 37 38 22 Suding KN, Lavorel S, Chapin FS et al. (2008) Scaling environmental change through the 39 40 23 communitylevel: a traitbased responseandeffect framework for plants. Global Change 41 24 Biology, 14, 11251140. 42 43 25 Taylor K (2001) Summarizing multiple aspects of model performance in a single diagram. 44 45 26 Journal of Geophysical Research, 106, 71837192. 46 27 Templer PH, Reinmann AB (2011) Multifactor global change experiments: what have we 47 48 28 learned about terrestrial carbon storage and exchange? New Phytologist, 192, 797800. 49 50 29 Teuling AJ, Seneviratne SI, Stockli R et al. (2010) Contrasting response of European forest 51 30 and grassland energy exchange to heatwaves. Nature Geosciences, 3, 722727. 52 53 31 Thorpe A, Aschehoug E, Atwater D, Callaway R (2011) Interactions among plants and 54 55 32 evolution. Journal of Ecology, 99, 729–740. 56 57 58 59 60 37 89 Page 39 of 44 Global Change Biology

1 2 3 1 Tognetti R, D’andria R, Lavini A, Morelli G (2006) The effect of deficit irrigation on crop 4 5 2 yield and vegetative development of Olea europaea L. (cvs. Frantoio and Leccino). European 6 3 Journal of Agronomy, 25, 356–364. 7 8 4 Tomozeiu R, Cacciamani C, Pavan V, Morgillo A, Busuioc A (2007) Climate change 9 10 5 scenarios for surface temperature in EmiliaRomagna (Italy) obtained using statistical 11 6 downscaling models. Theoretical and Applied Climatology, 90, 2547. 12 13 7 Van Dam JC, Groenendijk P, Hendriks RFA, Kroes JG (2008) Advances of modeling water 14 15 8 flow in variably saturated soils with SWAP. Vadose Zone Journal, 7, 640653. 16 9 Van Der Molen MK, Dolman AJ, Ciais P et al. (2011) Drought and ecosystem carbon 17 18 10 cycling. AgriculturalFor and Forest Review Meteorology, 151, 765773. Only 19 20 11 Van Peer L, Nijs I, Reheul D, De Cauwer B (2004) Species richness and susceptibility to heat 21 12 and drought extremes in synthesized grassland ecosystems: compositional vs physiological 22 23 13 effects. Functional Ecology, 18, 769778. 24 25 14 Vitasse Y, Porté A, Kremer A, Michalet R, Delzon S (2009) Responses of canopy duration to 26 15 temperature changes in four temperate tree species: relative contributions of spring and 27 28 16 autumn leaf phenology. Oecologia, 161, 187198. 29 30 17 Waisel Y, Eshel A, Agami M (1986) Salt balance of leaves of the mangrove Avicennia 31 18 marina. Physiologia Plantarum, 67, 6772. 32 33 19 Walther GR, Post E, Convey P et al. (2002) Ecological responses to recent climate change. 34 35 20 Nature, 416, 389395. 36 21 Weltzin JF, Loik ME, Schwinning S et al. (2003) Assessing the response of terrestrial 37 38 22 ecosystems to potential changes in precipitation. Bioscience, 53, 941952. 39 40 23 Wolkovich EM, Cook BI, Allen JM et al. (2012) Warming experiments underpredict plant 41 24 phenological responses to climate change. Nature. 42 43 25 Wu J, Van Der Linden L, Lasslop G, Carvalhais N, Pilegaard K, Beier C, Ibrom A (2012) 44 45 26 Effects of climate variability and functional changes on the interannual variation of the carbon 46 27 balance in a temperate deciduous forest. Biogeosciences, 9, 1328. 47 48 28 Zeppel MJB, Adams HD, Anderegg WRL (2011) Mechanistic causes of tree drought 49 50 29 mortality: recent results, unresolved questions and future research needs. New Phytologist, 51 30 192, 800803. 52 53 31 Zimmermann NE, Yoccoz NG, Edwards TC, Meier ES, Thuiller W, Guisan A, Schmatz D. R, 54 55 32 Pearman PB (2009) Climatic extremes improve predictions of spatial patterns of tree species. 56 33 Proceedings of the National Academy of Sciences, 106, 19723–19728. 57 58 34 59 60 38 90

et et Page 40 of 44

et al. et et al. et al. et et al. et et al. et 2007; 2007; Reference 2011 2011 Breda 2006; McDowell 2008 Leuzinger Leuzinger 2011a; Luterbacher al. Jentsch 2009 Goebel

Effect / Effect Response carbon starvation starvation carbon plants) (isohydric failure hydraulic plants) (anisohydric Frost damage, damage, Frost fatal damage possibly to opportunistic second or species, flowering, extended mid advanced decreased flowering, length flowering to lead repellency reduced soil of decomposition matter organic Stomatal closure and and closure Stomatal

Changing variability Changing Early and late frosts, frosts, late and Early spells, drought, warm heavy rain Droughts/heatwaves water soil in Increase Drought

et et al. et 2009 2009 et al. et 2006; Polgar & Polgar 2006; et al. et 2011 2011

2001 2001 91 et al. et et al. et 2003; Menzel Menzel 2003; eference Menzel & Fabian 1999; Walther Walther 1999; & Fabian Menzel R 2002; Parmesan & Yohe 2003; Root 2003; Yohe & Parmesan 2002; Vitasse 2011; Primack al. Saxe Saxe Albert Global Change Biology

Effect / Effect Response Prolongation of growing season, season, growing of Prolongation unfolding leaf of onset earlier flowering, and first senescence leaf of delay Forsoil in increase Potentially decomposition matter organic stomatal in increases Slight Review conductance Only

Changing mean Changing Increase in mean mean in Increase temperature mean in Increase temperature night in Increase (and time warming temperature) mean

cess Water relations Water relations Pro Table 1. Examples of observed plant vulnerabilities to changes in the mean climate and climate variability. climate variability. and climate the mean changes in to plant vulnerabilities Tableof observed Examples 1. Phenology Soil organic matter decomposition 1 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

40 Testing climate variability? Testing Locally unclear but over the the over but Locally unclear range species distribution yes probably Unclear Unclear (background (background Unclear further not variability likely but specified) Unclear but testing amplitude amplitude testing but Unclear in variability than larger much control Yes, 100yearevent Yes, 100yearevent standard 3 Yes, +/ mean from deviations 100year than less Yes, but event Yes, 100yearevent Yes, 100yearevent

et et et et et et al. et et al. et al. et et al. et et al. et 2005 2005 2009 2009 2009 2009 Reference 2012 2012 Jaoudé 2012 2012 2012 al. 2011 2011 2002 al. al.

Study type Study Experiment Experiment Abou Transplantation Transplantation Klein Experiment Experiment Kuster Observation Leuzinger Experiment Experiment Knapp Observation Observation Menzel Experiment Experiment Jentsch Experiment Experiment Jentsch

92 Global Change Biology Background variability Background Not explicitly mentioned, plant survived 45 days of of 45 days survived plant mentioned, explicitly Not flooding Longterm mean (differences in mean climate are very are climate in mean (differences mean Longterm background high equals amplitude testing hence large variability climate of testing explicit no but variability Compared to the longterm mean of the site, the the site, the of mean the longterm to Compared in 16% lower was control the in irrigation of amount in 2009 higher 30% and in 2008 higher 26% 2007, mean Longterm The large size and low frequency of precipitation precipitation of frequency and low size large The well are treatment precipitation altered the in events regimes precipitation documented of the range within this region in years 100 past the of Longterm mean mean Longterm precipitation < 1mm), 33 days of drought in 1976 in 1976 drought of 33 days 1mm), < precipitation 152mm of precipitation over 14 days in 1977 in 1977 14 days over precipitation of 152mm Local 100year extreme drought (number of days with with days of (number drought extreme 100year Local

For Review Only Continuous soil flooding with fresh and and fresh with flooding soil Continuous days 45 during saline water Precipitation in Rome, 766+156mm; Tel Tel 766+156mm; in Rome, Precipitation 279+88mm treated stands was 60% lower than the 60% lower was stands treated mm (728 precipitation mean longterm to April from season growing the during in 2008 lower and 43% 2007 October) in were droughts Experimental and 2009. several for irrigation stopping by imposed selected during weeks consecutive season growing the in periods 10 the of 50% Seasonal precipitation: 1999, to 1989 from mean year the mean, below Spring precipitation: exceeded temperatures: monthly Mean + (e.g., (1989–1999) mean the longterm June). for 6.8 °C Testing amplitude Testing 68 large precipitation events per per events precipitation 68 large = 42 event per season (mean growing mm) +1.5 (warm), +3 (very warm), −1.5 (cold) (cold) −1.5 warm), (very +3 (warm), +1.5 deviations standard cold) (very and −3 the at mean the from longterm and warm classify to point grid respective spells cold Drought: 32 days days 32 Drought: days 14 over 170mm Precipitation: extreme, rainfall 100year Local Aviv, 557+184mm, and Yatir (semiarid), (semiarid), Yatir and Aviv, 557+184mm, Amount of irrigation water in drought in water irrigation of Amount

), 4 4 ),

Poiret Poiret Tamarix africana Tamarix Pinus halepensis Pinus (3 stands 5 sites, contrasting provenances) Mixed broadleaved Mixed broadleaved in forest Central Europe provenances each) each) provenances Grasslands stands Young oak (3 species ( Quercus robur, petraea, Quercus Quercus European plant plant European phenology Study system Study Table 2. Are we measuring the impact of mean climate or climate variability? Nonexhaustive list of the studies cited testing in amplitude in the comparison text to the and background their variability in the respective study system. The last column indicates in a qualitative way European & heath grassland species European & heath grassland species how well the testing amplitude accounts for climatic variability in terms of the background variability. the background of in terms variability for climatic accounts amplitude testing how the well pubescens 1 1 2 2 3 Page 41 of 44 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Global Change Biology Page 42 of 44

1 2 1 Figure Legends: 3 2 4 3 Figure 1. The three theoretical cases of changing climatic drivers (from top to bottom): (1) 5 6 4 changes in the mean but not the variance, (2) changes in the variance but not the mean of a 7 8 5 variable and (3) both the mean and the background variability remain comparable, but 9 6 individual extreme events become more frequent. 10 11 12 7 Figure 2. Conceptual overview of the different processes and scales affected by extremes and 13 8 the methods to study them. 14 15 9 Figure 3. Evapotranspiration measured in the field with the eddy covariance method (black 16 17 10 filled dots) over the range of soil water contents (grey bars) occurring in the field and stomatal 18 For Review Only 19 11 conductance measured in a laboratory experiment (black open dots). Data from and further 20 12 descriptions available in Brilli et al. (2011). SWC = Soil Water Content. 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 41 93 Page 43 of 44 Global Change Biology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Review Only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Figure 1. The three theoretical cases of changing climatic drivers (from top to bottom): (1) changes in the 33 mean but not the variance, (2) changes in the variance but not the mean of a variable and (3) both the 34 mean and the background variability remain comparable, but individual extreme events become more 35 frequent. 36 190x142mm (300 x 300 DPI) 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

94 Global Change Biology Page 44 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Review Only 19 20 21 22 23 24 25 26 27 28 29 30 Figure 2. Conceptual overview of the different processes and scales affected by extremes and the methods to study them. 31 110x74mm (300 x 300 DPI) 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

95 Page 45 of 44 Global Change Biology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Review Only 19 20 21 22 23 24 25 26 27 28 29 Figure 3. Evapotranspiration measured in the field with the eddy covariance method (black filled dots) over 30 the range of soil water contents (grey bars) occurring in the field and stomatal conductance measured in a 31 laboratory experiment (black open dots). Data from and further descriptions available in Brilli et al. (2011). 32 SWC = Soil Water Content. 33 635x423mm (72 x 72 DPI) 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

96 97 5. Modeling changes in grassland hydrological cycling along an elevational gradient in the Alps

Della Chiesa, S., Bertoldi, G., Niedrist, G., Obojes, N., Endrizzi, S., Albertson, J. D., Wohlfahrt, G., Hörtnagl, L., Tappeiner, U. (2014): Ecohydrol 7:1453–1473.

98 99 Page 1 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 4 1 Modeling changes in grassland hydrological cycling 5 6 7 2 along an elevational gradient in the Alps 8 9 3 10 4 S. Della Chiesa1, G. Bertoldi1, G. Niedrist1, N. Obojes1, S. Endrizzi2, J. D. Albertson3, G. 11 5 Wohlfahrt4, L. Hörtnagl4, U. Tappeiner1,4 12 6 13 14 7 1 15 8 Institute for Alpine Environment, EURAC research, Bolzano, Italy. 16 2 17 9 Department of Geography, University of Zurich. 18 For Peer Review 19 10 3Department of Civil and Environmental Engineering, Pratt School of Engineering, Duke 20 11 University, North Carolina, USA. 21 22 12 4Institute of Ecology, University of Innsbruck, Austria. 23 24 13 25 26 14 Abstract 27 15 28 The effects of elevation on surface water fluxes in dry alpine grassland ecosystems were 29 16 investigated along an elevational transect between 1,000 and 2,000 m a.s.l. established in the 30 17 Vinschgau/Venosta valley, a relatively dry region in the Italian Alps. The GEOtopdv 31 18 hydrological model was employed in pointscale mode to model the effects of the elevation 32 19 gradient on snow water equivalent (SWE), soil water content ( ) evapotranspiration (ET), above 33 20 ground biomass (B ) and water use efficiency (WUE) in different climatic conditions. Results 34 ag 35 21 show that SWE decreased strongly with decreasing elevation, but was also affected by the 36 22 interannual variability of meteorological drivers. During warmer years the magnitude of changes 37 23 in SWE was mitigated at higher altitudes, while exacerbated below 1,500 m. dynamics 38 24 indicated that water stress conditions for vegetation currently occur at 1,000 m each year, while 39 25 only a warmer and drier year caused drought at 1,500 m and no water stress was found at 2,000 40 41 26 m. 42 27 ET, Bag and WUE did not decrease with elevation, but showed a maximum at an intermediate 43 28 elevation around ca. 1,500 m, because of the contrasting trends of a shorter vegetation season at 44 29 higher elevations and water stress at lower elevations, where, in fact, irrigation is needed to 45 30 maintain grassland productivity. A simulation based on longterm climatic conditions in 46 31 combination with a sensitivity analysis of precipitation change showed that this effect is more 47 48 32 pronounced during drier years, while for the wettest years ET tended to decrease with increasing 49 33 elevation. Taking these findings together, this study suggests that in relatively dry climatic 50 34 conditions, mountain areas generally act as “water towers” above 1,500 m. Quantifying this 51 35 critical threshold and its likely future variation under climate change scenarios is a challenge for 52 36 water resource research in the Alpine region and can help stakeholders in planning future 53 54 37 mitigation strategies. 55 38 56 39 57 40 58 59 60 1 John Wiley & Sons, Ltd 100 http://mc.manuscriptcentral.com/ecohydrology Page 2 of 44

1 2 3 1 4 5 2 6 3 7 4 8 5 9 6 1. Introduction 10 7 11 12 8 Hydrological processes in mountain regions are considered to be particularly sensitive to 13 9 climate change (CC) (Price and Barry, 1997; Beniston, 2003; IPCC, 2007), as they respond with 14 10 altered snow melt dynamics (Barnett et al., 2005), runoff production (Adam et al., 2009), timing 15 11 of maximum river discharge (Viviroli et al., 2007), amount of evapotranspiration (Jones et al., 16 12 2007), with shortterm effects on vegetation dynamics (Theurillat and Guisan, 2001) and long 17 18 13 term impacts on elevationalFor distribution Peer (Lenoir etReview al., 2008) and species composition (Kardol et 19 14 al., 2010). In mountain regions, hydrology and vegetation exhibit rapid changes, because of the 20 15 large variations in slope, aspect and elevation across short distances (Schär et al., 1998; Barry, 21 16 2007). In particular, elevation can be seen as a proxy for climate change and an elevational 22 17 transect represents an ideal natural experimental design for CC investigations (Becker et al. 23 18 24 2007; Körner, 2007). 25 19 Along an elevation gradient, fundamental drivers such as air temperature ( ) and 26 20 precipitation (P), affect different components of the water balance such as soil water content ( ), 27 21 runoff (Q) and evapotranspiration (ET). Air temperature gradients in Alpine regions typically 28 22 range from 0.55 to 0.58 K/100 m with distinct diurnal and seasonal variations (Rolland, 2003; 29 23 30 Barry, 2007). Precipitation exhibits a more heterogeneous behavior (Lauscher, 1976; Körner, 31 24 2003). Generally it increases with elevation in most temperate zones (Wastl and Zäng, 2008) as 32 25 well as the fraction falling as snow, and therefore snow amount and duration increase in line with 33 26 elevation (Hantel et al., 2000; Durand et al., 2009). Soil water content is a dynamic variable 34 27 primarily controlled by snow melt, precipitation and evapotranspiration (Verbunt et al., 2003; 35 28 Litaor et al., 2008; Rößler and Löffler 2010), and is therefore also affected by elevation. 36 37 29 Moreover, topography, vegetation and soil characteristics contribute strongly to the control of 38 30 soil moisture patterns (Wilson et al., 2004; Bertoldi et al., 2006). Temperature and radiation 39 31 differences caused by complex mountain topography control the snow melt (Beven, 2012) which 40 32 in turn strongly affects the runoff amount and seasonality (Gurtz et al., 1999; Gurtz et al., 2003; 41 33 Verbunt et al., 2003; Adam et al., 2009). In mountain regions, annual ET generally decreases 42 34 43 with higher elevation primarily because of the shorter growing season, and also due to the lower 44 35 air temperature and vapor pressure deficit (VPD), with a consequent reduction in aboveground 45 36 biomass (Körner, 2003; Körner, 2007; Wieser et al., 2008, Van den Bergh et al., 2013). Since 46 37 soil moisture generally increases with elevation, water stress rarely occurs at higher elevations 47 38 (Calanca et al., 2006; Rößler and Löffler, 2010). For example, severe droughts such as the 2003 48 39 European heat wave, or water limitation experiments, do not necessarily limit growth and 49 50 40 evapotranspiration in mountain grasslands or forests (Jolly, 2005; Hammerle et al., 2008; Brilli et 51 41 al., 2011) at higher elevations. Therefore ET is expected to increase with higher temperatures 52 42 due to CC on a seasonal and annual basis at higher elevations. By contrast, at lower elevations 53 43 severe soil moisture deficits and consequential seasonal and annual reductions in ET might occur 54 44 (Calanca et al., 2006). This is common in drier mountain regions (i.e. Mediterranean mountains) 55 45 56 where soil moisture strongly controls ET, and semiarid climatic conditions are present at lower 57 46 elevations and more suitable moisture conditions for vegetation development occur only at 58 59 60 2

John Wiley & Sons, Ltd 101 Page 3 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 1 higher elevations (Gallart et al., 2002; LópezMoreno et al., 2008). Therefore, the contrasting 4 5 2 natural trends of decreasing temperature and increasing precipitation with elevation can have 6 3 different effects on water balance and ecosystem productivity depending on whether or not 7 4 precipitation exceeds the water usage of vegetation, because intensively managed grassland 8 5 generally consists of “waterspending” plants which downregulate stomatal conductance only 9 6 when the soil moisture deficit becomes severe (Hammerle et al., 2008; Brilli et al., 2011). 10 7 11 During the snowfree period in humid mountain ecosystems, evapotranspiration is around 12 8 50% of precipitation, rising to up to 90% during dry years (Wieser et al., 2008), while in dry 13 9 mountain regions, depending on the interannual climate variability, ET can easily approach 14 10 100% of P below a certain elevation threshold, and agriculture relies heavily on river runoff from 15 11 the higher part of the watershed for irrigation. In fact, dry inner alpine regions such as 16 12 Vinschgau or Valais (Switzerland) have a long tradition of irrigation practice (Leibundgut, 17 18 13 2004). Moreover, CCFor is expected Peer to increase theReview demand for agricultural irrigation over the 19 14 whole alpine region (Bogataj and Sušnik, 2007). This, combined with changes in runoff 20 15 seasonality, is likely to cause problems related to conflicting water uses (Beniston, 2012), with 21 16 possible critical consequences for the mountain water budget that should be investigated and 22 17 properly modeled (Messerli et al., 2004; Viviroli and Weingartner, 2004; Viviroli et al., 2007; 23 18 24 Verbunt et al., 2003). We focus in particular on the role of elevation on the water balance, since, 25 19 depending on the complex interplay of climate conditions, elevation, and local soil and 26 20 vegetation properties, a certain elevation belt can switch from source to sink in terms of runoff 27 21 production and water storage. 28 22 In order to model the interactions between soil, vegetation and hydroclimatic conditions, 29 23 several soil vegetation atmosphere transfer (SVAT) models have been developed (see review 30 31 24 Arora, 2002)). In order to model hydrological processes in mountain regions properly, models 32 25 should include a detailed description of snow (Garen and Marks, 2005) and freezing/thawing 33 26 processes (Viterbo et al., 1999) and of the effects of complex topography on water and energy 34 27 fluxes (Blöschl et al., 1991). Finally, to model short term feedback loops between hydrological 35 28 processes, it is essential to include vegetation as a dynamic component (RodriguezIturbeI, 2000; 36 29 37 Eagleson, 2002). However, few physicallybased hydrological models account for vegetation 38 30 dynamics (Ivanov et al., 2008; Fatichi et al., 2012). The GEOtopdv model is suitable for 39 31 analyzing water fluxes along mountain slopes, since it is a hydrological model which takes into 40 32 account the root zone water budget, hydrological processes in cold regions (Endrizzi and Gruber, 41 33 2013), complex topography and includes a simple treatment of vegetation dynamics. 42 34 In this paper, we will model the effects of elevation on surface water fluxes for a dry alpine 43 44 35 grassland ecosystem along an elevational transect between 1,000 and 2,000 m. a.s.l. established 45 36 in the Vinschgau valley, a relatively dry region located in the Italian Alps. The GEOtopdv 46 37 model was employed to achieve the following objectives: 47 38 48 39 1. Show the capability of the GEOtopdv model to reproduce the observed seasonal 49 40 50 dynamics of water, energy fluxes and vegetation along an elevation gradient in a 51 41 mountain area; 52 42 2. Investigate the effects of the elevation gradient on snow, soil moisture, ET, above ground 53 43 biomass (Bag) and water use efficiency (WUE); 54 44 3. Evaluate the effects of current irrigation practice on soil moisture, ET, Bag and WUE. 55 45 4. Investigate the effect of climate variability on the change of ET and surface water fluxes 56 57 46 with elevation. 58 59 60 3

John Wiley & Sons, Ltd 102 http://mc.manuscriptcentral.com/ecohydrology Page 4 of 44

1 2 3 1 The first part of the paper presents a description of the GEOtopdv model, the study area and the 4 5 2 experimental data. The second part evaluates the model’s performance in simulating energy and 6 3 water fluxes. Thereafter, the effects of the elevation gradient on snow, soil moisture, 7 4 evapotranspiration and vegetation dynamics are analyzed by considering both natural 8 5 precipitation and irrigated conditions. Finally, the implications for the surface water budget of 9 6 mountain grassland ecosystems due to climatic variability are discussed performing a long term 10 7 11 simulation and a synthetic experiment in reducing and increasing precipitation to analyze the 12 8 impact on the ET elevation pattern. 13 9 14 10 2. Data and methods 15 11 2.1. The GEOtop-dv model 16 12 17 18 13 GEOtopdv is an extensionFor of thePeer GEOtop hydrological Review model (Endrizzi and Gruber, 2013; 19 14 Rigon et al., 2006), which includes grassland vegetation canopy as a dynamic component. In 20 15 this paper, only a brief overview of the model’s capabilities is presented. The GEOtopdv model 21 16 is well suited to address the present research questions as it considers vegetation dynamics with a 22 17 simple carbon assimilation approach based on Montaldo et al. (2005), simulates ET with a dual 23 18 24 layer canopy approach to represent turbulent fluxes above the canopy (Rigon et al., 2006) and 25 19 inside the canopy (Endrizzi and Marsh, 2010), solves the heat and water flow equations for 26 20 temperature and moisture in the soil with a coupled threedimensional numerical scheme, 27 21 including thawing soil dynamics (Dall’Amico et al., 2011), and simulates snow accumulation 28 22 and snow melt with a multilayer energybased approach (Endrizzi, 2007; Zanotti et al., 2004). 29 23 Furthermore, the model considers radiation correction for complex topography including 30 31 24 shadowing of direct solar radiation by surrounding mountains and the effects of topography on 32 25 diffuse radiation (Iqbal, 1983; Bertoldi, 2004). 33 26 The model solves the soil water balance: 34 27 35 36 28 , (1) 37 . , , 38 29 39 30 and the soil energy balance: 40 31 41 42 32 , (2) 43 . , 0 44 33 where the vectors are in bold, (m3/m3) is the volumetric water content, x (m) is the position, t 45 34 (s) is the time, q (ms1) is the sum of the surface runoff discharge and the lateral subsurface 46 35 fluxes regulated by Richards equation, where the nonlinear relation between soil moisture and 47 48 36 pressure hydraulic head is calculated according to van Genuchten, (1980). The term S includes 49 37 the exchanges between atmosphere and soil (ET, which includes evaporation, transpiration and 50 38 evaporation from wet vegetation), according to Eqs. (3, 5, 6 and 7). 51 39 U (Jm3) is the internal energy density in the soil, snow, ice and the canopy atmosphere, while g 52 40 (Wm2) is the energy flux, which, at the surface, includes the energy fluxes between soil, 53 54 41 vegetation and atmosphere, i.e. sensible (H), latent (LE) heat fluxes and net radiation (Rn), see 55 42 (Endrizzi and Marsh, 2010) for more details. 56 43 The ET is given by 57 44 58 59 60 4

John Wiley & Sons, Ltd 103 Page 5 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 1 (3) 4 1 5 6 2 where E is the actual rate of bare soil evaporation, is the canopy fraction which is a function of 2 2 7 3 GAI (Green Area Index) (m m ) according to Nouvellon, (2000), is the evaporation of the 8 4 wet vegetation, T is the actual rate of transpiration. In this duallayer canopy scheme, vapor and 9 5 heat fluxes from the canopy and from the soil surface are balanced by the heat and vapor fluxes 10 6 above the vegetation respectively (Endrizzi and Marsh 2010). The green area index GAI is a 11 12 7 dynamic variable, which simulates plant development following Montaldo et al., (2005). 13 8 The components of total evapotranspiration are computed as function of the potential 14 9 evapotranspiration PET that is, 15 10 16 11 (4) 17 ∗ 18 12 where (Jkg1) is theFor latent heat Peer of vaporization, Review (kgm3) is the air density, is the bulk 19 1 1 20 13 turbulent water vapor transfer coefficient, u (ms ) is the wind speed, ∗( ) (kgkg ) is the 1 21 14 saturation specific moisture at the reference surface temperature and ( ) (kgkg ) is the 22 15 atmospheric specific moisture. 23 16 Bare soil evaporation is computed as a function of water content of the first layer of soil, 24 17 25 26 18 (5) 27 28 19 where (sm1) is aerodynamic resistance and (sm1) is soil resistance depending on the water 29 20 content of the first layer of soil according to Feddes et al. (2001). 30 31 21 Evaporation from wet vegetation is calculated as, 32 22 33 23 (6) 34 35 24 where is the wet canopy fraction, according to Deardorff, (1978) 36 25 Transpiration is calculated as, 37 26 38 39 27 (7) 40 ∑ 41 42 28 43 29 The root fraction of each soil layer decreases linearly from the surface to a maximum root 44 30 depth . 45 31 Although stomata respond to changes in VPD (Mott and Peak 2010) and leaf water potential 46 32 (Buckley et al. 2003) and only indirectly to air temperature, radiation and soil moisture, in 47 48 33 GEOtopdv canopy resistance for each soil layer accounts for environmental stresses as in 49 34 Jarvis (1976). 50 35 51 36 52 53 37 , (8) 54 ,1.0 10 55 38 56 1 39 The minimum stomatal resistance m (s ) is a constant parameter, ( ) , ( ), (DPV ) 57 r , T 58 40 and ( ) are the limiting stress functions related to water content, air temperature, VPD and 59 60 5

John Wiley & Sons, Ltd 104 http://mc.manuscriptcentral.com/ecohydrology Page 6 of 44

1 2 3 1 solar radiation respectively. The functions , , decrease from 1 in optimal condition to 0 as 4 f f f 5 2 the stress on the vegetation increases. is set to if the stress functions tend to zero. 1.0 10 6 3 The stress function for water content in the root zone ( ) is computed according to Wigmosta 7 4 et al. (1994) f 8 5 9 10 11 0, for 12 6 (9) 13 f , for 14 15 1, for 16 7 where is the moisture content above which soil conditions do not restrict transpiration while 17 8 below limitations occur. is the soil water content at the permanent wilting point (the level at 18 For Peer Review 19 9 which plants are no longer able to extract water from the soil matrix). and can be 20 10 specified for each soil layer and depend on vegetation type. The temperature effect follows the 21 11 formulation of Dickinson et al., (1991) and Best, (1998). 22 12 23 24 25 13 0, and (10) f 26 / 27 28 14 29 15 The stress functions for VPD has been taken from Dickinson et al., (1991) and Best, (1998) 30 16 31 17 (11) 32 1.0 33 34 18 35 19 36 20 where (hPa) is the VPD and . The function for solar radiation is taken from 37 21 Best, (1998) 40 hPa 38 22 39 40 23 (12) 41 42 . 43 24 44 25 With 45 26 250 Wm 46 27 The vegetation module accounts for change in biomass in three different state variables, above 47 28 2 48 ground biomass (gDMm ), belowground biomass and aboveground necromass . 49 29 Since we are studying managed alpine meadows, which are regularly fertilized, we assume that 50 30 nutrients are not a limiting factor (Larcher, 2003). The biomass components are simulated at a 51 31 daily time step (Nouvellon, 2000; Cayrol, 2000; Montaldo et al., 2005), 52 32 53 54 33 (13) 55 56 34 57 58 59 60 6

John Wiley & Sons, Ltd 105 Page 7 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 4 1 5 14 6 2 7 8 3 , (15) 9 10 11 4 2 12 5 where (gDMm ) is the gross photosynthesis , and are the allocation coefficients to 13 6 leaves and roots respectively. is the roottoshoot translocation occurring at the beginning of 14 7 the vegetation season. 15 8 16 and are the respiration rates of leaves and root biomass, and are the senescence rates 17 9 of leaf and root biomass, is the litter fall. Gross photosynthesis is computed using the 18 10 simplified approach ofFor Montaldo etPeer al., (2005), Review 19 11 20 21 12 .., (16) 22 .. 23 13 where (gDMJ1) is the leaf photochemical efficiency, 24 14 2 25 (mol m s) is the photosynthetically active radiation and is the fraction of 26 15 absorbed by the canopy and is the minimum canopy resistance and is 27 16 calculated in Eq. 8. , / 28 17 Equations and references for the terms in (13, 14 and 15) can be found in Table 1. 29 18 is finally computed through a linear relation with (Nouvellon, 2000; Montaldo et al., 30 31 19 2005), 32 20 (17) 33 34 21 where (m2gDM1) is the bulk specific green area (analogous to the more familiar specific 35 22 leaf area, SLA). 36 23 For a more exhaustive description of the different model components see (Endrizzi and Gruber, 37 38 24 2013; Montaldo et al., 2005; Rigon et al., 2006). 39 25 40 26 3. Site description 41 27 This study was conducted in the emerging LongTerm Ecological Research (LTER) site 42 28 ‘Mazia/Matsch Valley’, in the Province of Bolzano/Bozen, South Tyrol, Italy (see Figure 1). 43 29 44 Due to the sheltering effect of the surrounding mountains the Vinschgau/Venosta Valley is one 45 30 of the driest areas in the Alps with a yearly average rainfall of 550 mm for the elevation of 1,500 46 31 m in the period 1980–2010 (source: Hydrographic Office South Tyrol). The area has a typical 47 32 continental alpine precipitation regime, with low total annual precipitation, a summer 48 33 characterized by convective rainfall events and a relatively cold winter with weak Atlantic 49 34 frontal systems and little precipitation (source: Autonomous Province of Bolzano/Bozen). 50 51 35 Since regional climate scenarios predict a temperature increase of about 1.5–5.8 K for the 52 36 21st century (Houghton, 2009) a transect of three micrometeorological stations was set up with 53 37 an elevation difference of 500 m and an average temperature gradient of 2.7 K between the 54 38 stations. 55 39 Hereafter the three stations are called B1000, B1500 and B2000, according to their elevation at 56 40 57 approximately 1,000, 1,500 and 2,000 m, respectively. The transect is situated along a 58 59 60 7

John Wiley & Sons, Ltd 106 http://mc.manuscriptcentral.com/ecohydrology Page 8 of 44

1 2 3 1 south/west slope and the linear distance from the highest to the lowest point is 4.2 km. Thus, 4 5 2 weather conditions and timing of precipitation can be assumed to be uniform for all three sites. 6 3 All sites have a homogeneous topography and land use (see Table 2). During the measuring 7 4 period in the years 2010 and 2011, January was the driest month along the transect, (see Figure 8 5 2), while the maximum precipitation occurred in August. The mean temperature lapse rate for 9 6 the transect during the two years was 0.54 K/100 m, while the precipitation lapse rate was 12 10 7 11 mm/100 m. The year 2010 was in general colder with higher precipitation compared to the year 12 8 2011 (see Table 3). Geologically, the Upper Vinschgau valley belongs to the Austroalpine 13 9 OetztalStubai crystalline complex, where impermeable rock formations such as gneiss, phyllite 14 10 and mica schist prevail (Habler et al., 2009). Throughout the valley light to medium soils 15 11 dominate, with texture classes varying from loamy sands to sandy loams. Deep Dystric 16 12 Cambisols characterize the valley basin and the adjacent mountain slopes (especially the exposed 17 18 13 southern slopes). TopsoilFor with Peer high humic contentReview can be found in intensively used deep 19 14 Cambisols due to fertilization (topsoil 10–20% and subsoil 3–10%). All sites were managed as 20 15 hay meadows, which were phytosociologically assigned to the MolinioArrhenatheretea class. 21 16 The lower site B1000 was cut three times per year, the middle site B1500 was cut twice while 22 17 the upper site B2000 was cut once per year. After mowing, rapid plant regrowth took place, 23 18 24 thus grass canopies undergo multiple growing cycles within a single growing season. The fields 25 19 were typically grazed in autumn, and so the ground in all the field stations was additionally cut in 26 20 midOctober to simulate grazing inside the instrumented area that had been fenced off. 27 21 28 2 4.2 Field measurements 29 2 4.1.3 Transect station measurements 30 31 24 Each micrometeorological station had an identical instrumental setup, data were acquired with a 32 25 1minute timestep and averaged on an hourly scale, and include: 33 26 global short wave radiation , photosynthetically active radiation above the canopy 34 27 and below the canopyR in order to gain information on snow cover 35 36 28 duration and vegetation growth), air temperature ( ), relative humidity , wind speed 37 29 measured at 2 m height, precipitation , soil water content at 5 cm depth and snow 38 30 height . Net radiation was only measured at station B1500. Gaps in meteorological 39 31 data were spatially and temporally interpolated from nearby stations of the Hydrographic Office 40 32 of the Autonomous Province of Bozen. Meadows were irrigated at the sites B1000 and B1500, 41 33 42 while the B2000 site was not irrigated. Irrigation took place regularly following local 43 34 management practices. Natural precipitation and artificial irrigation events were measured 44 35 separately. 45 36 46 3 74.2. Eddy covariance measurements 47 38 The net exchanges of latent and sensible heat were quantified by means of the eddy covariance 48 49 39 (EC) method (Aubinet et al., 2000; Baldocchi et al., 1988). One EC tower was installed in the 50 40 B1500 site and data starting from April 2011 until October 2011 were used for model calibration. 51 41 The three wind components and the speed of sound were measured with a threedimensional 52 42 ultrasonic anemometer (CSAT3, Campbell Scientific, UK) and the water vapor mole density 53 43 with an openpath infrared gas analyzer (Li7500, LiCor, USA). The open volume of the sonic 54 55 44 anemometer was mounted at 2 m above ground with the infrared gas analyzer displaced by 0.1 m 56 45 laterally. Serial data (SDM protocol) of both instruments were acquired at 20 Hz by a data 57 46 logger (Cr3000, Campbell Scientific, USA). Halfhourly average latent and sensible heat fluxes 58 59 60 8 John Wiley & Sons, Ltd 107 Page 9 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 1 were calculated using the postprocessing software EdiRe as the covariances between the 4 5 2 turbulent fluctuations of vertical wind speed and water vapor mole density and sonic temperature 6 3 respectively, which were calculated by removing the respective arithmetic means during each 7 4 averaging period. Lags in the sonic anemometer and infrared gas analyzer data due to signal 8 5 processing were accounted for by the data logger program. A planarfit coordinate rotation was 9 6 applied on the wind data (Wilczak et al., 2001). Frequency response corrections were applied on 10 7 11 the raw covariances to account for highpass and lowpass filtering following Moore (1986) and 12 8 Aubinet et al. (2000) using a sitespecific reference cospectrum as described in Wohlfahrt et al. 13 9 (2005). Finally, the buoyancy flux was converted to the sensible heat flux following Schotanus 14 10 et al. (1983) and latent heat fluxes were corrected for density fluctuations following Webb et al. 15 11 (1980). The net fluxes of latent and sensible heat were then calculated as the sum of the 16 12 corrected eddy covariance and storage flux, the latter being calculated as the height integrated 17 18 13 time rate of change For at the measurement Peer height Review (Haslwanter et al., 2009). Positive fluxes 19 14 represent transport from the surface to the atmosphere, negative fluxes the reverse. 20 15 21 1 4.3.6 Vegetation and soil measurements 22 17 In addition to the micrometeorological measurements, green area index, , and were 23 18 Bag 24 measured. Harvesting took place according to traditional mowing methods on the respective 25 19 elevations (35 replicates). Cutting height was 3 cm above the soil surface while the 26 20 information between 0 and 3 cm was taken on comparable plots outside the measuring areaB inag 27 21 order to enable future regrowth. Plant material was separated into functional groups (grasses, 28 22 forbs without legumes, legumes and dead plant material). was measured using a WinDIAS 29 23 30 S3 plant area meter (DeltaT devices, UK). After measurement all the material was dried in 31 24 an oven at 80°C until the weight was constant. Finally, the bulk specific green area 32 25 was calculated for each cut and a final averaged value was used in the model 33 26 parameterization/ for each site. The soil hydraulic properties of residual moisture content 34 27 and saturated moisture content were measured during field surveys with the gravimetric 35 36 28 method and used to calibrate soil moisture sensors, while the wilting point moisture content 37 29 was estimated using the soil moisture retention curve (pF curve). 38 30 39 3 14.4. Simulations setup 40 32 4.4.1. Calibration and model performance 41 42 33 43 34 The GEOtopdv was used in a pointscale mode and requires as input the following 44 35 meteorological data at hourly resolution: air temperature ( ), relative humidity , 45 36 precipitation , wind speed , solar shortwave and photosynthetically active radiation 46 37 . Table 4 lists the soil and vegetationR parameters, indicating whether they were 47 48 38 manually calibrated, observed or taken from literature. At each site, topographical properties 49 39 such as elevation, slope and aspect were obtained from a Lidarderived digital elevation model. 50 40 For the sites, a soil depth of 1 m, was assigned to the station B1000, B1500 and B2000 51 41 respectively, based on field observations. The model was parameterized for 8 model soil layers 52 42 with increasing depths from 0.01 m up to 1 m (0.01 m, 0.01 m, 0.015 m, 0.03 m, 0.075 m, 0.10 53 54 43 m, 0.16 m, 0.20 m, 0.60 m and 1.00 m).Soil hydraulic parameters were assumed to be the same 55 44 at the three study sites and constant throughout the investigation period. In the calibration phase, 56 45 observed and simulated time series of , , , H, LE were compared to minimize root 57 46 mean square error (RMSE), bias and to maximize the correlation coefficient ( for the 58 59 60 9 John Wiley & Sons, Ltd 108 http://mc.manuscriptcentral.com/ecohydrology Page 10 of 44

1 2 3 1 intermediate station B1500, (see Table 6). Stations B1000 and B2000 were used to asses model 4 5 2 performance, in irrigated (B1000) and not irrigated (B2000) conditions, comparing the observed 6 3 and simulated time series of and in terms of RMSE, bias and . Moreover, to assess 7 4 model performance in simulating vegetation dynamics, vegetation measurements of Bag and GAI 8 5 at each cutting event were compared with the corresponding simulated values. After each 9 2 6 cutting the variable Bag was set back to initial condition ( = 300 gDMm ) to simulate the cut. 10 Bag 11 7 Since irrigation and natural precipitation were separately measured, another set of simulations 12 8 was conducted with the same parameterization but only natural precipitation as input for the 13 9 B1000 and B1500 stations. This allowed a cleaner assessment of the influence of the elevation 14 10 gradient on snow, soil moisture, and evapotranspiration and vegetation dynamics with 15 11 comparable natural precipitations amounts. As cutting dates for not irrigated grassland were not 16 12 17 available, it was decided to mimic the farmer´s decision on the cutting period with a simple rule 18 13 depending on GAI development and stress condition limiting the growth: (i) to perform a cut a For Peer2 Review 19 14 minimum value of = 300 gDMm was required and (ii) in the event of severe stress 20 15 conditions occurring Bforag a period longer than three weeks a cut was performed overriding the 21 16 previous rule. Sensitivity analyses (results not shown here) on these two simple rules showed 22 17 negligible variations in total annual water and energy fluxes. 23 24 18 25 19 4.4.2. Drought experiment and climate variability simulation setup 26 20 In order to analyze the influence of climate variability along the transect a longterm simulation 27 21 spanning 22 years was conducted based on daily meteorological data from 19902011 derived 28 22 from nine surrounding stations at different elevations within the Vinschgau/Val Venosta region 29 23 30 (source: Hydrographic Office of the Province of Bozen/Bolzano (see Appendix A). The 31 24 measured daily values of P, and SWin from these stations were available for the 32 25 simulation. Constant average values of U were assumed based on our measurements in 2010 and 33 26 2011, ranging from 1.5 ms1 at B1000 to 2.5 ms1 at B1500 and 3.5 ms1 at B2000. For the 34 27 vegetation module the daily values of were assumed to be based on the linear 35 28 0.45 36 relationship, on a daily basis, between PhAR and global SWin (Ross and Sulev, 2000). As cutting 37 29 dates for the long term simulation were not available, the harvesting period was based on the 38 30 automatic method mentioned in the previous section 4.4.1. 39 31 To assess whether changes in P affect the elevationdriven pattern of ET along the transect, we 40 32 repeated the longterm simulation, keeping the original number of events while reducing or 41 33 increasing the amount per event by 30% from May to August, but leaving the other 42 43 34 meteorological variables unchanged. The reduction setup was chosen on the basis of future 44 35 climate scenarios for the region which predict a reduction of summer precipitation of around 45 36 20% with a maximum decrease of 40% (Frei et al., 2006) and based on Gilgen and Buchmann, 46 37 (2009) who showed reduction of May to August precipitation by 45% would simulate an extreme 47 38 summer drought in the future for grassland ecosystems. The 30% precipitation increase provided 48 39 49 a further insight to assess the sensitivity of ET along the elevation gradient. 40 50 51 4 15. Results 52 4 5.1.2 Model calibration and performance 53 43 54 44 The ability of the model to predict energy, water, and vegetation dynamics is shown in Figures 3, 55 45 4, 5 and 6 and overall performance is shown in Table 6. 56 57 46 The comparison between observed and modeled ET is shown in Figure 4. Note that this refers 58 47 only to the time periods where observed values are available. Modeled and observed ET agreed 59 60 10 John Wiley & Sons, Ltd 109 Page 11 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 1 well, with a bias of 6.12 Wm2 and RMSE of 58.39 Wm2. Also on the seasonal and daily scale, 4 5 2 the model accurately reproduced field measurements, except for two periods after cutting at the 6 3 end of June 2011 and at the beginning of August 2011. In fact, Figure 3 shows that the model 7 4 tends to overestimate ET slightly, when the canopy is fully closed, while it was underestimated 8 5 after cutting 9 6 Soil moisture dynamics were reasonably well simulated along the transect, with a mean bias of 10 7 11 less than 0.01 (Table 6). This value can be considered satisfactory, considering the uncertainties 12 8 in soil moisture sensor calibrations and the differences among the different sensors at the same 13 9 depth (RMSE=0.08). Note that for freezing and snow covered conditions, soil moisture 14 10 observations were discarded. 15 11 The model also reproduced quite accurately along the transect with a constant bias of 0.01 m 16 12 for all the stations (which can be related to an offset to the snow depth sensor), while RMSE 17 18 13 ranges from 0.04 and 0.05For m (see PeerTable 6 and Figure Review 5 d), e), f)). 19 14 The performance of the vegetation module in simulating Bag and GAI along the elevation 20 15 gradient is shown in Figure 6. 21 16 observation at the different cutting dates agreed better with the simulated values for the 22 17 Bdifferentag sites than GAI. In summary, the validation exercise suggested that the model is able to 23 24 18 capture changes in hydrological cycling across the elevational gradient reasonably well and we 25 19 thus proceed to an analysis of the drivers. 26 20 27 21 5.2. Trends of SWE , soil moisture ET, Bag and WUE with respect to elevation over the 28 22 two study years 29 23 30 In this paragraph, the trends of SWE, soil moisture, ET, Bag and WUE are analyzed with respect 31 24 to elevation during the two study years, which were characterized by quite different climatic 32 25 conditions, 2011 being warmer and drier than 2010. 33 26 In Figure 7 the differences in the temporal dynamics of SWE, cumulative ET, and the soil 34 27 moisture frequency distribution for the three stations and for the years 2010 and 2011 are 35 28 displayed. 36 37 2 9In Figure 8 the relative role of the different environmental stress factors limiting ET 38 30 (temperature, soil moisture, VPD as in Eqs. (9, 10 and 11) is displayed in terms of the sum of 39 31 stress days during the season. The different curves are the time integrals of the factors (1-f1), (1- 40 32 f2 ) and (1-f3) as in Eqs. (9, 10 and 11). As the factor functions are normalized, the relative 41 33 magnitude of their time integrals can be compared over the course of the growing season. In 42 34 43 Figure 9 e), f) the vertical profile of ET, Bag and WUE are presented. The effects of elevation in 44 35 the two years on SWE, soil moisture ET, Bag and WUE are presented in detail below. 45 36 46 37 5.2.1. Effects of elevation on SWE 47 38 48 39 Figure 7a) shows the dynamics of SWE for the years 2010 and 2011 along the transect. As 49 50 40 expected, SWE and snow cover duration increased with elevation for both years. However, the 51 4 SWE1 pattern along the elevation differed from year to year. The wetter and colder year 2010 52 42 showed a linear increase of SWE with elevation. In contrast, the drier year 2011 was 53 43 characterized by a shorter snow cover period for stations B1000 and B1500, while B2000 54 44 exhibited relatively high SWE and a long snow cover period relatively similar to 2010. 55 45 56 Over the two years the difference in snow cover duration showed a reduction in 2011 of about 57 46 18% at B2000, 34% at B1500 and 20% at B1000. Along the transect in 2010, the snow cover 58 59 60 11 John Wiley & Sons, Ltd 110 http://mc.manuscriptcentral.com/ecohydrology Page 12 of 44

1 2 3 1 duration showed a reduction of 32% from B2000 to B1500 and a reduction of 61% from B2000 4 5 2 to B1000. In 2011, the snow cover period showed a reduction of 45% from B2000 to B1500 and 6 3 a reduction of 63% from B2000 to B1000. In 2011 at the end of winter and in early spring the 0 7 4 °C isotherm frequently fluctuated between 1,500 m and 2,000 m, thus affecting the precipitation 8 5 partitioning into liquid water or snow, leading to a larger reduction of SWE from B2000 to 9 6 B1500 compared to the previous year. 10 7 11 12 8 5.2.2. Effects of elevation on soil moisture 13 9 14 10 Figure 7c) shows the frequency distribution of modelled for the two years along the 15 11 transect under natural precipitation and irrigated conditions. Notice that B1000 was irrigated in 16 12 both years, whereas B1500 only in 2011, while B2000 was never irrigated. 17 18 13 The frequency distributionFor of Peer without irrigationReview exhibited a clear pattern related to 19 14 different hydroclimatic conditions along the elevation gradient, moving from drier conditions at 20 15 the lower site towards wetter conditions at the highest site. In 2010 the lower site B1000 showed 21 16 a positive skewness with a high frequency of lower soil moisture. This was due mainly to the 22 17 short water stress periods in midJune and the first ten days of July (see Figure 8a)). In 2011, the 23 24 18 relatively dry spring combined with the longer snowfree period, depleted soil water storage 25 19 through deep percolation and evaporation, triggering drought stress in April and May (see Figure 26 20 8a). The B1500 site showed a more bellshape frequency distribution centered on mean values 27 21 for both years. In 2010 water stress was limited to a short period in July (Figure 8a)). In 2011 a 28 22 larger variance with a more pronounced negative skewness implied more frequent water limited 29 23 30 conditions, and Figure 8a) indicates a drought period from midApril until midMay. The B2000 31 24 site showed a similar frequency distribution for both years, higher averaged moisture conditions 32 25 and negative skewness, showing higher frequencies of moist conditions near field capacity. This 33 26 was due to a combination of higher summer precipitation, lower air temperature and a shorter 34 27 growing season (see figures 8a) and b)). However, the warmer 2011 clearly showed signs of 35 28 larger variance with a rather small occurrence of dry conditions. 36 37 29 The frequency distributions under irrigated conditions showed a quasinormal frequency 38 30 distribution in both years. In 2010 the lower site B1000 showed high frequency of soil moisture 39 31 at intermediate with a quasinormal distribution. However, in 2011 soil moisture was more 40 32 variable with some short periods of water limitation in April. The B1500 site under irrigated 41 33 conditions showed a frequency distribution comparable with the frequency distribution under 42 43 34 natural precipitation conditions in 2011, indicating less than optimal irrigation practice. 44 35 45 36 5.2.3. Effects of elevation on ET 46 37 47 38 Without irrigation (Figure 7b)) a different pattern of ET emerged for both years. ET at the lower 48 39 site B1000 appeared to be water limited, while at the upper site B2000 ET was limited by 49 50 40 temperature and a short snowfree season. Higher ET occurred at the middle site B1500. At this 51 41 elevation, the vegetation apparently found the best conditions of soil water availability, 52 42 temperature, VPD and length of the snowfree season. Despite a later start to the growing season 53 43 in 2010, cumulative ET at B1500 caught up with B1000 in the middle of July due to the short 54 44 water stress conditions combined with higher VPD at B1000 which limited ET (see Figure 8a), 55 56 45 c)) showing similar total values at the end of the season. In 2011 due to the relatively shorter 57 46 drought period, cumulative ET at B1500 overtook B1000 in May and showed a steeper slope 58 59 60 12

John Wiley & Sons, Ltd 111 Page 13 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 1 with higher total values at the end of the vegetation season. The B2000 site, in contrast to the 4 5 2 lower sites, showed higher total ET during the warmer and drier year 2011, which indicates that 6 3 the low temperatures were the major limiting factor of ET at 2,000 m. 7 4 Under irrigated conditions, cumulative ET showed a clear decrease with higher elevation in 8 5 annual ET during both years, (see Figure 7b), 9a)). In 2010 at B1000, irrigation prevented ET 9 6 limiting soil moisture conditions and compensated for the higher VPD (Figure 8a), c)). In 2011, 10 7 11 despite the relative longer water stress (Figure 8a)), cumulative ET at B1500 closely followed 12 8 cumulative ET at B1000 (Figure 7b). Despite the large difference in the amount of irrigation 13 9 between B1000 and B1500 (see Figure 9a) total ET at B1000 and B1500 had similar values. 14 10 The overall trend of ET with elevation can be seen by looking at Figure 9b), where the vertical 15 11 profiles of actual annual ET (under both natural and irrigated conditions) and potential ET (PET) 16 12 along the transect are reported. 17 18 13 As previously noted, For actual ET Peerwithout irrigation Review followed a divergent pattern due to water 19 14 limitation at lower elevations, showing a maximum at the B1500 site, especially for the drier 20 15 year 2011, while PET generally decreased with elevation. This demonstrates that the reduction 21 16 of ET at higher elevation was related to temperature limitation and lower VPD (also affecting 22 17 PET), while the decrease at lower elevation was related to water limitation and higher VPD (see 23 18 24 Figure 8 a), b), c) where the relative magnitude of the time integrals of soil moisture, temperature 25 19 and VPD stresses are compared. 26 20 Under natural conditions, the ET/PET ratio along the transect (Figure 9 c)) ranged between 55% 27 21 and 65% at the B2000 and B1500 sites for the two years. The ET/PET ratio at the lower B1000 28 22 site was not sensitive to the varying conditions over the different years showing an identical 29 23 value (50%). The increase of ET/PET during the drier and warmer year at B1500 indicates that 30 31 24 ET at 1,500 m was limited more by temperature than by water stress. A clear difference in ET/P 32 25 occurred at all sites in the two years (figure 9d): at B1000 the grassland field evapotranspired 33 26 around 88% of precipitation in the wetter year 2010 and up to 98% in the drier year 2011, while 34 27 at 1,500 m the ET/P reached 80% in 2010 and 97% in the drier and warmer year 2011. At 2,000 35 28 m the ET/P ratio ranged between 45% in 2010 and 58% in 2011. Actual ET for irrigated 36 29 37 conditions decreased with elevation, following a similar trend in PET. The ET/PET ratio (Figure 38 30 9c)) at the B1500 site was insensitive with respect to irrigation, while at the lower B1000 site the 39 3 ET/PET1 ratio reacted to irrigation and increased to similar values to those of the upper sites. 40 32 Irrigation at B1500 and B1000 reduced ET/P (Figure 9d)). 41 33 42 34 5.2.4. Effects of elevation on B and WUE 43 ag 44 35 In Figure 9 e) and f) Bag and WUE are shown along the elevation gradient for the two years 2010 45 36 and 2011. Bag followed a similar elevation pattern as ET in irrigated and nonirrigated 46 37 conditions. 47 38 In natural precipitation condition, it displayed a maximum at intermediate elevation while Bag 48 39 and WUE was limited by water stress at the lowest site (and by temperature at the highest site). 49 40 50 In irrigated conditions WUE increases and followed a decreasing trend with increasing elevation. 51 41 Despite the small increase in WUE, irrigation practice proved to be highly inefficient, since in 52 42 2010 an increase of 350 mm (+60%) of water supply increased above ground biomass by 150 53 43 gDMm2 (+17%). In 2011 an increase of 380 mm (+61%) of water supply increased above 54 44 ground biomass by 250 gDMm2 (+28%). 55 45 56 57 46 5.2.5. Effects of long-term climate variability along the elevation gradient 58 59 60 13 John Wiley & Sons, Ltd 112 http://mc.manuscriptcentral.com/ecohydrology Page 14 of 44

1 2 3 1 4 5 2 Figure 10) presents the results of the 22 year longterm simulation (19902011) for SWE, ET and 6 3 and compares them to the detailed (hourly meteorological forcing) simulation of the years 7 4 20102011. 8 5 Compared with the two year simulation, the longterm simulation indicated a larger difference in 9 6 terms of snow cover duration along the transect, while the standard deviation in SWE differed 10 7 11 along the transect (Figure 10 a, b, c). In fact at 2,000 m the simulations of the two study years 12 8 were well within the longerterm variability range of SWE and snow duration, while at 1,500 m 13 9 and 1,000 m the detailed simulation of 2010 showed higher SWE while the 2011 rather lower 14 1 0SWE compared with the long term simulation. This difference indicated that the years 2010 and 15 11 2011 cover a large difference in terms of winter precipitation and temperature. In Figure 10 d), 16 12 e) and f) the results for ET and its partitioning into E and T are compared. Notice that the E 17 18 13 component includes evaporationFor fromPeer the bare soilReview fraction, soil below the canopy fraction and 19 14 water evaporated from the canopy after precipitation interception (eq. 6). 20 15 A comparison of ET between the longterm and the two year simulation along the transect 21 16 revealed that at 2,000 m the long term simulation had overall higher values of ET compared to 22 17 the two year simulation. The use of constant wind speed for each site in the long term simulation 23 18 24 could partly explain this difference. However, the general trend with respect to elevation of the 25 19 long term simulation is similar to what had been previously observed. 26 20 Interestingly, at 1,500 m the difference between ET values for 2010 and 2011was more than 27 21 50%, which is larger than the standard deviation of the long term simulation. Total ET of 2011 28 22 was about 150 mm higher than the 22 yearaverage, while in 2010, total ET was slightly below 29 23 the average. This indicates that the 2010 and 2011 simulation covers a large climate variability. 30 31 24 At 1,000 m the difference between ET of 2010 and 2011 was lower than the standard deviation 32 25 of the long term simulation, in the 2010 the total ET was perfectly within the 22 yearaverage 33 26 while in the 2011 total ET was slightly above the upper standard deviation. 34 27 The ET partitioning showed an increase of E/ET ratio with the increasing elevation from 40% up 35 28 to 57%: this is mainly due to the decreasing GAI and canopy fraction with increasing altitude, 36 37 29 and also due to the ET/PET ratio as seen in Figure 9 c). 38 30 Figure 10 g), h) and i) clearly show that overall the dynamics of 2010 and 2011 were inside the 39 31 longterm ranges. 40 32 Interestingly, the frequency distribution of the longterm simulation showed a peak below 41 3 3 =0.15 for all the three stations (most pronounced at the lowest B1000 station), which indicates 42 34 43 water stress conditions. This suggests that 2011 is well representative of a dry year and not an 44 35 isolated extreme event. 45 36 Figures 11 a), b) and c) analyze in more detail the yearly mean of along the whole soil profile 46 37 0100 cm for the longterm simulation for stations B2000, B1500 and B1000 respectively. 47 38 At B2000, showed relatively stable wet conditions along the whole soil profile during the year 48 39 49 with values ranging on average from 0.3 to 0.4. reached its highest values around DOY=155 50 40 (Day of Year = 155) and snow melt affected the soil profile up to 80 cm. The lower soil 51 41 boundary showed a stable value of about 0.3. 52 42 At B1500, showed a larger variability compared to B2000, ranging from 0.2 up to 0.35. The 53 43 effect of the snow melt on caused a maximum around DOY=125. At this site the effect of the 54 44 55 vegetation is evident as the lowest values of were reached in the main rooting zone during a 56 45 period with high growth rates/biomass around DOY=240 .At the lowest depth, soil moisture was 57 46 stable around =0.23. 58 59 60 14 John Wiley & Sons, Ltd 113 Page 15 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 1 The soil moisture within the soil profile at B1000 ranged from 0.16 up to 0.3. While the upper 4 5 2 sites showed a single clear maximum of in late spring due to snow melt, the site B1000 6 3 showed multiple days during late winter/early spring with maxima around 0.3. This is due to the 7 4 large variability of snow cover duration at low elevation (see Figure 10 c). 8 5 At B1000 the effect of the snowmelt on soil moisture was limited to a depth of about 50 cm and 9 6 until DOY=125. Afterwards, from DOY 150 to DOY 180, showed the driest conditions from 10 11 7 0 to 30 cm. This indicates that at the lowest elevation late spring was the most droughtsensitive 12 8 period, in contrary of the site B1500. Moreover, from DOY 150 till DOY 300 around a depth of 13 9 0.3 m, was constantly drier compared to the upper and lower soil layers, with a value of about 14 10 0.16. This probably due to the fact that in summer, because of canopy interception and strong 15 11 ET, convective storms mostly affect the surface soil layers with no effect on the deeper soil. 16 12 17 Despite snow melt in spring and summer thunderstorms, has a constant value of 0.17 in the 18 13 deepest soil layer fromFor 0.8 m to 1 Peerm throughout theReview whole year. 19 14 The precipitation change experiment presented in Figure 12 showed that a reduction of 20 15 precipitation of about 30% compared to the original 19902011 simulation increased the 21 16 diverging pattern of ET along the elevation gradient. Average ET of the 22 years was reduced 22 17 the most at 1,000 m (51 mm) and significantly at the 1,500 m site (23 mm) but stayed nearly 23 24 18 the same at 2,000 m. An increase of precipitation by about 30% showed an increase of ET at 25 19 B1000 (+22 mm) but no effects at B1500 and B2000. This result shows how the maximum of 26 2 ET0 found at the intermediate elevation was a pattern typical of dry years. 27 21 28 22 6. Discussion 29 2 3 30 6.1. Comparison between model results and observation 31 24 The GEOtop model demonstrates an overall good ability to reproduce the different hydrological 32 25 processes along an elevation gradient in challenging mountain conditions. 33 26 However, a comparison of observed and modeled ET rates shows that these appear to differ 34 27 slightly for a short period after each grass cut. This is probably due to a combination of several 35 28 factors: i) the intrinsic difficulty in simulating accurately the abrupt change in energy and water 36 37 29 fluxes after each cut as the controls energy partitioning (Hammerle et al., 2008); ii) the 38 30 accuracy of EC measurementsGAI are affected by many factors, and to assure robust data the energy 39 31 imbalance was assigned to ET (Wohlfahrt et al., 2010); iii) in the model applied at point scale 40 32 the harvest was assigned to a discrete day, while the farmers traditionally mowed the larger 41 33 meadows over the course of a couple of days. The EC tower fetch would thus appear patchy for 42 43 34 some days. This would mainly explain the discrepancy between simulation and observation 44 35 immediately after the harvest. The authors’ choice in limiting (specific green area index) 45 36 parameterization purposely reduced agreement between observation and simulated values. 46 37 In fact, during the various growing and cutting phases large differences in the ratio of grasses 47 38 and forbs were measured, and therefore varies seasonally (data not shown). Thus the results 48 49 39 suggest that implementing a dynamic parameterization might be needed to further improve 50 40 the performance of the GEOtopdv model. 51 41 52 4 26.2. Implications on SWE dynamics 53 43 Simulation of SWE along the transect indicates that even during the warmer year SWE decreases 54 44 more at the lower stations at 1,500 m and at 1,000 m than at the upper one B2000. 55 56 45 This is also confirmed by the long term simulations, where below 1,500 m in an extremely dry 57 46 and warm year snow can disappear completely. This is in line with previous studies stating that 58 59 60 15 John Wiley & Sons, Ltd 114 http://mc.manuscriptcentral.com/ecohydrology Page 16 of 44

1 2 3 1 at higher elevations winter precipitation is more important than winter temperature for SWE or 4 5 2 snow cover duration (Beniston, 2003; Laghari et al., 2012). Recently, MoránTejeda et al. 6 3 (2013) found that, in the Swiss Alps, snow cover is more sensitive to temperature variation 7 4 below an altitude of about 1,500 m.a.s.l. Similar patterns have been found in other mountain 8 5 regions in Europe (Scherrer, 2004; Marty, 2008; LópezMoreno et al., 2009; Durand, 2009). 9 6 Beniston, (2003) reported that the snow line rises by about 150 m for every °C increase in 10 7 11 temperature. This is generally true; however the magnitude of this shift is stronger around the 12 8 0°C isotherm. In fact, our simulations indicate that below ca. 1,500 m the snow line is strongly 13 9 sensitive to warmer and drier conditions, and thus the snow line will rise accordingly. On 14 10 average we found a reduction of snow cover of about 17 days for each degree along the transect. 15 11 This is perfectly in agreement with other studies (Beniston, 2003b). Seasonal variation of SWE 16 12 is the primary control of initial water availability at the beginning of the growing season and of 17 18 13 spring runoff along theFor elevation Peergradient, as becomes Review clear when comparing the SWE plots in 19 14 Figure 10 with the profiles of Figure 11. Earlier snowmelt might initially increase soil 20 15 evaporation and deep seepage while an earlier onset of vegetation might deplete soil water 21 16 storage. Thus, warmer winters are likely to trigger potential water stress in early spring even at 22 17 1,500 m. 23 18 24 25 19 6.3. Implications of soil moisture on ET, Bag and WUE 26 20 27 21 The temporal dynamics of showed drought conditions at 1,000 m in both years. The drier year 28 22 2011 also showed drought occurring at 1,500 m, while at 2,000 m, was rather insensitive to 29 23 30 drought. This is also characteristic of the long term simulation. Moreover, on a long term basis 31 2 4 showed an even larger variance in the frequency distribution, thus the years 2010 and 2011 are 32 25 representative of rather average and dry conditions, but even more extreme (moister or drier) 33 26 states are possible. 34 27 The mean temporal dynamics of along the soil profile analyzed in the long term simulation 35 28 clearly highlighted the increased water storage in the soil profile with increasing elevation. The 36 37 29 lower site at 1,000 m showed a general low in the root zone in the early spring, trigging the 38 30 possibility of severe drought conditions. Therefore irrigation is needed to prevent early spring 39 31 drought at low elevations. The dry steady state conditions in the deeper soil layers indicate little 40 32 potential of groundwater recharge at the lowest site. 41 33 At 1,500 m the lowest in the root zone occurred in late spring, thus the later start of the 42 43 34 growing season mitigated and delayed the decrease of . This supports the hypothesis that soil 44 35 moisture conditions are more suitable for plant growth or agricultural production etc. at middle 45 36 elevations. However, spring droughts might still occur in extreme years with a combination of 46 37 early snow melt and low precipitation. Irrigation is therefore also needed at 1,500 m to prevent 47 38 losses in grassland productivity. At 2,000 m the soil is rather moist along the whole soil profile, 48 39 49 providing water storage and ground water recharge. It is only at this elevation that the 50 40 Vinschgau mountain region will act as a “water tower”. Similar patterns were found at high 51 41 elevations in drier mountain regions such as the Pyrenees (Gallart et al., 2002; LópezMoreno, 52 42 2008). 53 43 In alpine regions evapotranspiration generally decreases with increasing elevation (Körner, 2003; 54 44 Wieser et al., 2008). However, our study suggests a divergent trend of evapotranspiration with 55 56 45 elevation. This is evident for the drier 2011 year and also for several years of the longterm 57 46 simulation (Figure 12). Annual ET did not change linearly with elevation, but showed a 58 59 60 16 John Wiley & Sons, Ltd 115 Page 17 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 1 maximum at an intermediate elevation around approximately 1,500 m. This is caused by a 4 5 2 frequent limitation of ET due to low soil moisture at 1,000 m, which offsets the shorter growing 6 3 season, and a lower daily maximum evapotranspiration. Irrigation generally increased ET at 7 4 B1000 and B1500 in both years. However, despite the large difference in the total irrigation 8 5 amount between B1000 and B1500, ET totals at both sites were similar. This clearly indicates a 9 6 poor irrigation strategy at B1500 where most of the irrigation was lost in runoff and deep 10 7 11 percolation, while better irrigation timing would have limited water stress at B1000. Here 12 8 further practical and theoretical work regarding irrigation practices are needed to improve the 13 9 efficiency of water use (Vico and Porporato, 2010). 14 10 The numerical precipitation change experiment we performed (Figure 12) shows that a reduction 15 11 of precipitation of about 30% increases the divergent pattern of ET along the elevation gradient 16 12 with strong reduction of ET at 1,000 m and a smaller one at the 1,500 m site. By contrast, an 17 18 13 increase of about 30%For only caused Peer a small increase Review of ET at B1000. Interestingly, the B1500 19 14 site was insensitive to increased precipitation. This is in accordance with the detailed simulation 20 15 which showed that B1500 is mainly temperature limited. Overall this supports the hypothesis 21 16 that in an even drier climate the divergent elevationdriven pattern of ET becomes stronger, 22 17 while in an increased precipitation scenario evapotranspiration follows the more general 23 18 24 elevation trend of alpine regions (Körner, 2003; Wieser et al., 2008). 25 19 The same considerations we made for ET are also valid for WUE and above ground biomass. 26 20 Generally they both decrease along an upward elevation gradient in the Alps (Körner, 2003; 27 21 Cernusca and Seeber, 1981). 28 22 Our results showed a divergent pattern in natural precipitation condition. This was observed also 29 2 in3 a similar semiarid region in China by Han et al., (2013). This indicates that changes in WUE 30 31 24 do not buffer ET variability enough to stabilize above ground productivity, and irrigation practice 32 25 prevents water limitation and increased WUE at lower elevation. Simulation proved that the 33 26 irrigation practice is highly inefficient due to the large amount of water supply with a rather 34 27 small increment of above ground biomass. 35 28 Our results clearly indicate there is an elevation threshold with a particular combination of P and 36 29 37 Ta above which precipitation exceeds the water needs of vegetation, whereas below a certain 38 30 elevation threshold, water limitation reduces above ground biomass and ET. 39 31 Only a small number of climate scenario studies reported drought stress in the Alps (Jasper et al., 40 32 2004; Zierl and Bugmann, 2005; Calanca et al., 2006; Jasper et al., 2006; Rößler et al., 2012) and 41 33 to our knowledge no studies investigated in detail the surface water balance in dry inner alpine 42 34 valleys which are currently water limited. The Vinschgau/Venosta valley has a long tradition of 43 44 35 irrigation to mitigate summer drought. However, increase in temperature and reduction of 45 36 precipitation will exacerbate longer dry periods. In the Alps future CC is likely to increase 46 37 evapotranspiration, while changes in runoff timing and volume will deplete water reserves, 47 38 causing water shortages. Irrigation will become a common practice for sustaining agriculture in 48 39 the Alps (Falloon and Betts, 2010). 49 40 50 51 41 6.4. Elevation gradient as a proxy of climate change 52 42 53 43 The elevation gradient can be seen as a proxy for CC as it implies a shift of the hydrological 54 44 cycle similar to what is expected from CC, which will lead to a shorter snow cover period 55 45 (Barnet et al., 2005), lower soil water content (Jasper et al., 2006), and if soil moisture is not 56 57 46 limiting, higher rates of ET (Calanca et al., 2006). The downward annual mean temperature 58 59 60 17 John Wiley & Sons, Ltd 116 http://mc.manuscriptcentral.com/ecohydrology Page 18 of 44

1 2 3 1 increase of about 2.7 K/500 m is within the projected range of temperature increase of about 4 5 2 1.55.8 °C for the 21st century (Houghton, 2009). 6 3 With CC it is likely that there will be an upward shift of the vertical distribution of ET (of 7 4 roughly 150 m for every 1 K of temperature increase). This implies a reduction of mountain 8 5 areas which act as “water towers” (Viviroli et al., 2007) with implications for the availability of 9 6 water resources. 10 7 11 Climate scenarios suggest diverging patterns of ET, with a decrease of ET south of the Alps, but 12 8 an increase in the central Alpine region and, at low elevations, in the northern forelands (Jasper 13 9 et al., 2004; Calanca et al., 2006). Our simulations are clearly in line with the prediction for the 14 10 inner alpine area, highlighting the Vinschgau valley as an early warning region for likely future 15 11 developments in what the scenarios predict is already occurring. 16 12 However, regional climate scenarios in the Alpine region are uncertain, especially for 17 18 13 precipitation, where manyFor scenarios Peer suggest a reductionReview of summer precipitation (Rotach et al., 19 14 1997; Smiatek et al., 2009), but show different trends with respect to yearly totals. Moreover, 20 15 the elevation dependency of precipitation changes in the context of CC are highly uncertain (Frei 21 16 et al., 2003; Kotlarski et al., 2012). 22 17 Thus, our transect illustrates only one possible future scenario. The combination of increased 23 18 24 temperatures and reduced precipitation is likely to increase the occurrence of severe drought over 25 19 larger areas of the Alps with consequences for vegetation growth and agricultural productivity 26 20 (Loreto et al., 2004; Calanca et al., 2006). 27 21 28 22 7. Conclusion 29 23 30 31 24 The GEOtopdv model offers a reliable representation of soilatmosphere and vegetation 32 25 interactions and allowed a detailed study on mountain grassland hydrological cycling along an 33 26 elevation transect for a relatively dry alpine region. 34 27 In this paper we performed a detailed study of the impacts on surface water fluxes, irrigation and 35 28 productivity of a mountain grassland ecosystem along an elevation gradient in a continental 36 29 37 alpine climate regime. We analyzed in detail two years with different climatic conditions for 38 30 which detailed validation data were available, and then we considered the interannual climatic 39 31 variability with a 22 yearlong simulation. Finally, we evaluated the impact of extreme 40 32 conditions with a synthetic precipitation change experiment. Our study suggests that in a 41 33 grassland ecosystem in a dry inner alpine valley, below a certain elevation threshold, water stress 42 34 Tlimits E and B in a natural precipitation regime, while above this elevation threshold the 43 ag 44 35 temperature acts as a limiting factor on ET and Bag. The different climatic conditions exert 45 36 further control on the different hydrological processes, exacerbating or mitigating the variability 46 37 of SWE, soil moisture, ET and consequently in Bag. For these mountain grassland ecosystems 47 38 there exists an intermediate elevation below which most of the precipitation is used for ET, and 48 39 above which an increasing part of precipitation can become runoff and seepage. This elevation 49 40 50 also corresponds to maxima in Bag and WUE. This indicates that changes in WUE do not buffer 51 41 ET variability enough to stabilize above ground productivity, and that only irrigation prevents 52 42 water limitation at lower elevations. This effect becomes more pronounced during dry years. In 53 43 other words, in relatively dry climatic conditions mountain areas act as “water towers” only 54 44 above a certain elevation. Although irrigation limits water stress and loss of productivity, 55 45 irrigation strategies are far from optimal and future summer droughts might increasingly exhaust 56 57 46 the water reserves and cause water shortages. Our simulations indicate that irrigation is highly 58 59 60 18

John Wiley & Sons, Ltd 117 Page 19 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 1 inefficient due to the large amount of water used with only a rather small increment in above 4 5 2 ground biomass. 6 3 Quantifying the critical variation of water storage and above ground productivity in alpine 7 4 regions is a relevant challenge for ecohydrological research and ecosystem service assessment. 8 5 Robust estimates can help stakeholders and managers take decisions on mitigation strategies to 9 6 cope with expected future changes in the Alpine region. 10 7 11 12 8 13 9 14 10 Acknowledgments 15 11 This work has been supported by the research grants “Climate Change in South Tirol” and 16 12 “HydroAlp” of the Provincia Autonoma di Bolzano, Alto Adige, Ripartizione Diritto allo studio, 17 18 13 Universita e ricerca scientifica.For Peer The participation Review of J. D. Albertson in this study was also 19 14 supported by the grant “Mobilità di ricercatrici e ricercatori” financed by the Provincia 20 15 Autonoma di Bolzano, Alto Adige, Ripartizione Diritto allo studio, Università e ricerca 21 16 scientifica. 22 17 The institutions involved are part of the interdisciplinary research center ‘Ecology of the Alpine 23 18 24 Region’ within the research area ‘Alpine Space – Man and Environment’ at the University of 25 19 Innsbruck. We would like to thank the Hydrographic Office of the Autonomous Province Bozen 26 20 for providing data. We thank Elisabeth Mair for the Meteo data preparation and special thanks 27 21 to Francis Howlett for final editorial services on the manuscript. 28 22 29 23 30 31 24 References 32 33 25 34 26 Adam JC, Hamlet HF, Lettenmaier DP. 2009. Implications of global climate change for 35 27 snowmelt hydrology in the twentyfirst century. Hydrological Processes 23: 962–972. 36 28 Arora V. 2002. Modeling vegetation as a dynamic component in soilvegetationatmosphere 37 29 transfer schemes and hydrological models. Reviews of Geophysics 40: 38 30 doi:10.1029/2001RG000103. 39 40 31 Aubinet M, Grelle A, Ibrom A, Rannik U, Noncrieff J, Foken T, Kowalski AS, Martin PH, 41 32 Berbigier P, Bernhofer C, Clement R, Elbers J, Granier A, Grunwald T, Morgenstern K, 42 33 Pilegaard K, Rebmann C, Snijders W, Valentini R, Vesala T. 2000. Estimates of the annual 43 34 net carbon and water exchange of forests: The EUROFLUX methodology. Adv. Ecol. Res. 30: 44 35 113175. 45 46 36 Baldocchi DD, Hincks BB, Meyers TP, 1988. Measuring biosphereatmosphere exchanges of 47 37 biologically related gases with micrometeorological methods. Ecology 69(5): 13311340. 48 38 Barnett TP, Adam JC, Lettenmaier DP. 2005. Potential impacts of a warming climate on water 49 39 availability in snowdominated regions. Nature 438: 303309. doi:10.1038/nature04141. 50 40 Barry RG. 2007. Mountain Weather and Climate. Cambridge University Press. 51 41 52 Becker A, Körner C, Brun JJ, Guisan A, Tappeiner U. 2007. Ecological and land use studies 53 42 along elevational gradients. Mountain Research and Development 27(1): 5865. 54 43 Beniston, M. 2003. Climatic change in mountain regions: A review of possible impacts. Climatic 55 44 9Change 5 (12): 531. doi:http://dx.doi.org/10.1023/A:1024458411589 56 57 58 59 60 19 John Wiley & Sons, Ltd 118 http://mc.manuscriptcentral.com/ecohydrology Page 20 of 44

1 2 3 1 Beniston M, Keller F, Koffi B, Goyette S. 2003b. Estimates of snow accumulation and volume in 4 5 2 the Swiss Alps under changing climatic conditions. Theoretical and Applied Climatology 6 3 76 (34): 125140. doi:10.1007/s0070400300165. 7 4 Beniston M. 2012. Impacts of climatic change on water and associated economic activities in the 8 5 Swiss Alps. Journal of Hydrology 412: 291296. 9 6 Bertoldi G. 2004. The water and energy balance at basin scale: a distributed modeling approach. 10 7 11 PhD thesis. 12 8 Bertoldi G, Rigon R, Over TM. 2006. Impact of watershed geomorphic characteristics on the 13 9 energy and water budgets. Journal of Hydrometeorology 7(3): 389403. 14 10 Best M. J. 1998. A model to predict surface temperatures. Boundary-Layer Meteorology 88(2): 15 11 279306. 16 12 Beven KJ. 2012. Rainfallrunoff modelling: the primer (Vol. 15, p. 360). Chichester: Wiley. doi: 17 18 13 10.1002/9781119951001.For Peer Review 19 14 Blöschl G, Kirnbauer R, Gutknecht D. 1991. Distributed Snowmelt Simulations in an Alpine 20 15 Catchment 1. Model Evaluation on the Basis of Snow Cover Patterns. Water Resources 21 16 Research7 2 (12): 31713179. 22 17 Bogataj LK, Sušnik A. 2007. Challenges to agrometeorological risk management—regional 23 18 24 perspectives: Europe. In managing weather and climate risks in agriculture (pp. 113124). 25 19 Springer Berlin Heidelberg. 26 20 Brilli F, Hörtnagl L, Hammerle A, Haslwanter A, Hansel A, Loreto F, Wohlfahrt, G. 2011. Leaf 27 21 and ecosystem response to soil water availability in mountain grasslands. 28 22 Buckley TN, Mott KA, Farquhar GD. (2003). A hydromechanical and biochemical model of 29 23 stomatal conductance. Plant Cell and Environment 26: 17671785. 30 31 24 Agricultural and Forest Meteorology 151(12): 17311740. doi:10.1016/j.agrformet.2011.07.007. 32 25 Calanca P, Roesch A, Jasper K, Wild M. 2006. Global Warming and the Summertime 33 26 Evapotranspiration Regime of the Alpine Region. Climatic Change 79(12): 6578. 34 27 doi:10.1007/s1058400691039. 35 28 Cayrol P. 2000. Grassland modeling and monitoring with SPOT4 VEGETATION instrument 36 29 37 during the 19971999 SALSA experiment. Agricultural and Forest Meteorology 105(13): 38 30 91115. doi:10.1016/S01681923(00)00191X. 39 31 Cernusca A, Seeber M. C. 1981. Canopy structure, microclimate and the energy budget in 40 32 different alpine plant communities. In Plants and their Atmospheric Environment. 21st 41 33 Symposium of the British Ecological Society. Blackwell, Oxford. 7581. 42 34 Dall’Amico M, Endrizzi S, Gruber S, Rigon R. 2011. A robust and energyconserving model of 43 44 35 freezing variablysaturated soil. The Cryosphere 5(2): 469484. doi:10.5194/tc54692011. 45 36 Deardorff JW. 1978. Efficient Prediction of Ground Surface Temperature and Moisture, With 46 37 Inclusion of a Layer of Vegetation. Journal of Geophysical Research 83(C4): 18891903. 47 38 doi:10.1029/JC083iC04p01889. 48 39 Dickinson R, Hendersonsellers A, Rosenzweig C, Sellers P. 1991. Evapotranspiration models 49 40 50 with canopy resistance for use in climate models, a review. Agricultural and Forest 51 41 Meteorology 54(24): 373388. doi:10.1016/01681923(91)90014H. 52 42 Durand Y, Giraud G, Laternser M, Etchevers P, Mérindol L, Lesaffre B. 2009. Reanalysis of 47 53 43 years of climate in the French Alps (19582005): Climatology and trends for snow 54 44 cover. Journal of applied meteorology and climatology 48(12): 24872512. 55 45 Eagleson PS. 2002. Ecohydrology: Darwinian expression of vegetation form and function. 56 57 46 Cambridge University Press, New York. 58 59 60 20 John Wiley & Sons, Ltd 119 Page 21 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 1 Endrizzi S. 2007. Snow cover modelling at a local and distributed scale over complex terrain. 4 5 2 PhD thesis. 6 3 Endrizzi S, Marsh P. 2010. Observations and modeling of turbulent fluxes during melt at the 7 4 shrubtundra transition zone 1: point scale variations. Hydrology Research 41(6): 471491 8 5 doi:10.2166/nh.2010.149. 9 6 Endrizzi S, Gruber S, Dall’Amico M, Rigon R. 2013. Geotop: Simulating the Combined Energy 10 7 11 and Water Balance at and Below the Land Surface Accounting for Soil Freezing, Snow Cover 12 8 and Terrain Effects. Geosci. Model Dev. Discuss. 6: 62796341. doi:10.5194/gmdd66279 13 9 2013 14 10 Falloon P, Betts R. 2010. Climate impacts on European agriculture and water management in the 15 11 context of adaptation and mitigation–the importance of an integrated approach. The Science of 16 12 the total environment. 408(23): 56675687. doi:10.1016/j.scitotenv.2009.05.002. 17 18 13 Fatichi, S., Ivanov, V. ForY., Caporali, Peer E. 2012. A MechanisticReview Ecohydrological Model to 19 14 Investigate Complex Interactions in Cold and Warm Watercontrolled Environments: 1. 20 15 Theoretical Framework and Plotscale Analysis. Journal of Advances in Modeling Earth 21 1 Systems6 4 (May 30): M05002. doi:10.1029/2011MS000086. 22 17 Feddes RA, Hoff H, Bruen M, Dawson T, de Rosnay P, Dirmeyer P, Pitman AJ. 2001. Modeling 23 18 24 root water uptake in hydrological and climate models. Bulletin of the American 25 19 Meteorological Society 82(12): 27972809. 26 20 Frei C, Christensen JH, Déqué M, Jacob D, Jones RG, Vidale PL. 2003. Daily precipitation 27 21 statistics in regional climate models: Evaluation and intercomparison for the European 28 22 Alps. Journal of Geophysical Research: Atmospheres 108(D3): 1984 29 23 2012. doi: 10.1029/2002JD002287. 30 31 24 Gallart F, Llorens P, Latron J, Regüés D. 2002. Hydrological processes and their seasonal 32 25 controls in a small Mediterranean mountain catchment in the Pyrenees. Hydrology and Earth 33 26 System Sciences 6(3): 527537. doi:10.5194/hess65272002. 34 27 Garen D, Marks D. 2005. Spatially distributed energy balance snowmelt modelling in a 35 28 mountainous river basin: estimation of meteorological inputs and verification of model 36 29 37 results. Journal of Hydrology 315(14): 126153. doi:10.1016/j.jhydrol.2005.03.026. 38 30 Gilgen A. K, Buchmann N. 2009. Response of temperate grasslands at different altitudes to 39 31 simulated summer drought differed but scaled with annual precipitation. Biogeosciences 6: 40 32 25252539, doi:10.5194/bg625252009, 2009. 41 33 Gurtz J, Baltensweiler A, LangH. 1999. Spatially distributed hydrotope based modelling of 42 34 evapotranspiration and runoff in mountainous basins, Hydrological Processes, 13: 27512768. 43 44 35 Gurtz J, Zappa M, Jasper K, Lang H, Verbunt M, Badoux A, Vitvar T. 2003. A comparative 45 36 study in modelling runoff and its components in two mountainous catchments. Hydrological 46 37 Processes 17(2): 297311. doi:10.1002/hyp.1125. 47 38 Habler G, Thöni M, Grasemann B. 2009. Cretaceous metamorphism in the Austroalpine Matsch 48 39 Unit (Eastern Alps): The interrelation between deformation and chemical equilibration 49 40 50 processes. Mineralogy and Petrology 97(34): 149171. doi:10.1007/s007100090094x. 51 41 Hammerle A, Haslwanter A, Tappeiner U, Cernusca A, Wohlfahrt G. 2008. Leaf area controls on 52 42 energy partitioning of a temperate mountain grassland. Biogeosciences 5(2): 421431. 53 43 doi:10.5194/bg54212008. 54 44 Han Q, Luo G, Li C, Ye H, Chen Y. 2013. Modeling grassland net primary productivity and 55 45 wateruse efficiency along an elevational gradient of the Northern Tianshan Mountains. 56 57 46 Journal of Arid Land, 5(3): 354365. 58 59 60 21 John Wiley & Sons, Ltd 120 http://mc.manuscriptcentral.com/ecohydrology Page 22 of 44

1 2 3 1 Hantel M, Ehrendorfer M, Haslinger A. 2000. Climate sensitivity of snow cover duration in 4 5 2 Austria. International Journal of Climatology 20(6): 615640. 6 3 Haslwanter A, Hammerle A, Wohlfahrt G. 2009. Openpath vs. closedpath eddy covariance 7 4 measurements of the net ecosystem carbon dioxide and water vapour exchange: A longterm 8 5 perspective. Agricultural and Forest Meteorology 149(2): 291302. 9 6 Houghton JT. 2009. Global Warming: The Complete Briefing. Cambridge University Press: 10 7 11 Cambridge. 12 8 IPCC. 2007. Climate change: the physical science basis. Cambridge University Press: 13 9 Cambridge. 14 10 Iqbal M. 1983. An introduction to solar radiation. Academic Press: Orlando. 15 11 Ivanov, V. Y., Bras, R. L., and Vivoni, E. R. 2008. Vegetation hydrology dynamics in complex 16 12 terrain of semiarid areas: 1. A mechanistic approach to modeling dynamic feedbacks. Water 17 18 13 Resour. Res. 44: W0For3429, doi:10.1029/2006WR005588. Peer Review 19 14 Jarvis PG. 1976. The Interpretation of the Variations in Leaf Water Potential and Stomatal 20 15 Conductance Found in Canopies in the Field, Philosophical Transactions of the Royal 21 16 Society. Biological Sciences 273(927): 593–610. doi:10.1098/rstb.1976.0035. 22 17 Jasper K, Calanca P, Gyalistras D, Fuhrer K. 2004. Differential impacts of climate change on the 23 18 24 hydrology of two alpine river basins. Climate Research 26: 113129. doi:10.3354/cr026113. 25 19 Jasper K, Calanca P, Fuhrer J. 2006. Changes in summertime soil water patterns in complex 26 20 terrain due to climatic change. Journal of Hydrology 327(34): 550563. 27 21 doi:10.1016/j.jhydrol.2005.11.061. 28 22 Jolly WM. (2005). Divergent Vegetation Growth Responses to the 2003 Heat Wave in the Swiss 29 23 Alps. Geophysical Research Letters 32(18): L18409. doi:10.1029/2005GL023252. 30 31 24 Kardol P, Courtney E.C, Souza L, Norby J.R, Weltzin F.J, Classen T.A. 2010. Climate Change 32 25 Effects on Plant Biomass Alter Dominance Patterns and Community Evenness in an 33 26 Experimental OldField Ecosystem. Global Change Biology 16 (10): 2676–2687. 34 27 doi:10.1111/j.13652486.2010.02162.x. 35 28 Körner C. 2003. Alpine Plant Life - Functional Plant Ecology of High Mountain Ecosystems. 36 29 37 Springer: Heidelberg. 38 30 Körner C. 2007. The use of ‘altitude’ in ecological research. Trends in ecology & evolution 39 3 221 (11): 569–74. doi:10.1016/j.tree.2007.09.006. 40 32 Kotlarski S, Bosshard T, Lüthi D, Pall P, Schär C. 2012. Elevation gradients of European climate 41 33 change in the regional climate model COSMOCLM. Climatic Change 112(2): 189215. 42 34 doi:10.1007/s1058401101955. 43 44 35 Laghari AN, Vanham D, Rauch W. 2012. To what extent does climate change result in a shift in 45 36 Alpine hydrology? A case study in the Austrian Alps. Hydrological Sciences Journal 57 (1): 46 37 103117. 47 38 Larcher W. 2003. Physiological Plant Ecology. Springer: New York. 48 39 Lauscher F. 1976. Weltweite typen der höhen abhängigkeit des niederschlag. Wetter und Leben 49 4 0 50 28: 8090. 51 41 Leibundgut C. 2004. Historical meadow irrigation in Europea basis for agricultural 52 42 development. IAHS PUBLICATION 286: 7787. 53 43 Lenoir J, Gegout JC, Marquet PA, de Ruffray P, Brisse H. 2008. A significant upward shift in 54 44 plant species optimum elevation during the 20th century. Science 320(5884): 176871. 55 45 doi:10.1126/science.1156831. 56 57 58 59 60 22 John Wiley & Sons, Ltd 121 Page 23 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 1 Litaor MI, Williams M, Seastedt TR. 2008. Topographic controls on snow distribution, soil 4 5 2 moisture, and species diversity of herbaceous alpine vegetation, Niwot Ridge, Colorado. 6 3 Journal of Geophysical Research 113(G2): 110. doi:10.1029/2007JG000419. 7 4 LópezMoreno J, Beniston M, GarcíaRuiz J. 2008. Environmental change and water 8 5 management in the Pyrenees: Facts and future perspectives for Mediterranean mountains. 9 6 Global and Planetary Change. 61(34): 300312. doi: 10.1016/j.gloplacha.2007.10.004. 10 7 11 LópezMoreno J, Goyette S, Beniston M. 2009. Impact of climate change on snowpack in the 12 8 Pyrenees: Horizontal spatial variability and vertical gradients. Journal of Hydrology 374 (3 13 9 4): 384396. doi:10.1016/j.jhydrol.2009.06.049. 14 10 Loreto F, Centritto M. 2004. Photosynthesis in a changing world. Plant Biology 6(03): 239241. 15 11 doi:10.1055/s2004820885. 16 12 Marty C. 2008. Regime shift of snow days in Switzerland. Geophysical Research Letters 35 17 18 13 (12): L12501. doi:10.1029/2008GL033998.For Peer Review 19 14 Messerli B, Viviroli D, Weingartner R. 2004. Mountains of the world: vulnerable water towers 20 15 for the 21st Century. Ambio 2934. 21 16 Montaldo N, Rondena R, Albertson JD, Mancini M. 2005. Parsimonious modeling of vegetation 22 17 dynamics for ecohydrologic studies of waterlimited ecosystems. Water Resources 23 18 24 Research 41(10): W10416. doi:10.1029/2005WR004094. 25 19 Moore CJ. 1986. Frequency response corrections for eddy correlation systems. Boundary-Layer 26 20 Meteorology 37: 1735. 27 21 MoránTejeda E, LópezMoreno JI, Beniston M. 2013. The Changing Roles of Temperature and 28 22 Precipitation on Snowpack Variability in Switzerland as a Function of Altitude. Geophysical 29 23 Research Letters 40: 16. doi:10.1002/grl.50463. 30 31 24 Mott K. A., Peak, D. 2010. Stomatal responses to humidity and temperature in darkness. Plant, 32 25 Cell & Environment 33: 10841090. 33 26 Nouvellon Y. 2000. Modelling of daily fluxes of water and carbon from shortgrass steppes, 34 27 Agricultural and Forest Meteorology 100(23): 137153. doi:10.1016/S01681923(99)00140 35 28 9. 36 29 37 Price MF, Barry RG. 1997. Climate Change, in Mountain of the World: A Global Priority, edited 38 30 by Messerli B, Ives JD. Parthenon publishing group: New York. 39 31 Rigon R, Bertoldi G, Over TM. 2006. GEOtop: A Distributed Hydrological Model with Coupled 40 32 Water and Energy Budgets. Journal of Hydrometeorology 7(3): 371388. 41 33 doi:10.1175/JHM497.1. 42 34 RodriguezIturbe I. 2000. Ecohydrology: A hydrologic perspective of climatesoilvegetation 43 44 35 dynamies. Water Resources Research 36(1): 3. doi:10.1029/1999WR900210. 45 36 Rolland C. 2003. Spatial and Seasonal Variations of Air Temperature Lapse Rates in Alpine 46 37 Regions. Journal of Climate 16: 10321046. 47 38 Ross J, Sulev M. 2000. Sources of errors in measurements of PAR. Agric Forest Meteorol 100: 48 39 103–125. 49 40 50 Rößler O, Löffler J. 2010. Potentials and limitations of modelling spatiotemporal patterns of soil 51 41 moisture in a high mountain catchment using WaSiMETH. Hydrological Processes 24(15): 52 42 21822196. doi:10.1002/hyp.7663. 53 43 Rößler O, Diekkrüger B, Löffler J. 2012. Potential drought stress in a Swiss mountain catchment; 54 44 Ensemble forecasting of high mountain soil moisture reveals a drastic decrease, despite major 55 45 uncertainties. Water Resources Research 48(4): W04521. doi:10.1029/2011WR011188. 56 57 58 59 60 23

John Wiley & Sons, Ltd 122 http://mc.manuscriptcentral.com/ecohydrology Page 24 of 44

1 2 3 1 Rotach MW, Marinucci MR, Wild M, Tschuck P, Ohmura A, Beniston M. 1997. Nested regional 4 5 2 simulation of climate change over the Alps for the scenario of a doubled greenhouse 6 3 forcing. Theoretical and applied climatology 57(34): 209227. 7 4 Schär C, Davies T, Frei C, Wanner H, Widmann M, Wild M, Davies HC. 1998. Current Alpine 8 5 Climate, in Views from the Alps: Regional perspectives on climate change. Edited by Cebon 9 6 P, Dahinden U, Davies HC, Imboden DM, Jäger CC. MIT Press: Boston. 10 7 11 Scherrer SC. 2004. Trends in Swiss Alpine snow days: The role of local and largescale climate 12 8 variability. Geophysical Research Letters 31(13): L13215. doi:10.1029/2004GL020255. 13 9 Schotanus P, Nieuwstadt F, de Bruin H. 1983. Temperature measurement with a sonic 14 10 anemometer and its application to heat and moisture fluxes. Boundary Layer Meteorology 26: 15 11 8193. 16 12 Smiatek G, Kunstmann H, Knoche R, Marx A. 2009. Precipitation and temperature statistics in 17 18 13 high resolution regionalFor climate Peer models: Evaluation Review for the European Alps. Journal of 19 14 Geophysical‐ Research: Atmospheres 114: D19107. doi:10.1029/2008JD011353 20 15 Laghari AN, Vanham D, Rauch W. 2012. To what extent does climate change result in a shift in 21 16 Alpine hydrology? A case study in the Austrian Alps. Hydrological Sciences Journal 57(1): 22 17 103117. 23 18 24 Theurillat JP, Guisan A. 2001. Potential impact on climate change on vegetation in the european 25 19 alps: a review. Climatic Change 50(12): 77109. 26 20 Jones PD, Trenberth KE, Ambenje PG, Bojariu R, Easterling DR, Klein TG, Zhai P. 2007. 27 21 Observations: surface and atmospheric climate change. IPCC, Climate change 235336. 28 22 Van den Bergh T, Inauen N, Hiltbrunner E, Körner C. 2013. Climate and Plant Cover Co 29 23 Determine the Elevational Reduction in Evapotranspiration in the Swiss Alps. Journal of 30 31 24 Hydrology 500(September): 75–83. doi:10.1016/j.jhydrol.2013.07.013. 32 25 van Genuchten MT. 1980. Closedform Equation for Predicting the Hydraulic Conductivity of 33 26 Unsaturated Soils. Soil Science Society of America Journal 44(5): 892–898. 34 27 Vanham D, Fleischhacker E, Rauch W. 2009. Impact of an extreme dry and hot summer on 35 28 water supply security in an alpine region. Water Science and Technology 59(3): 469477. 36 37 29 doi:10.2166/wst.2009.887, 2009. 38 30 Verbunt M, Gurtz J, Jasper K, Lang H, Warmerdam P, Zappa. 2003. The hydrological role of 39 31 snow and glaciers in alpine river basins and their distributed modeling. Journal of Hydrology 40 3 2282 (14): 3655. doi:10.1016/S00221694(03)002518. 41 33 Vico, G. and Porporato, A. 2010. “Traditional and Microirrigation with Stochastic Soil 42 34 Moisture”. Water Resources Research 46(3): (March 9) doi:10.1029/2009WR008130. 43 44 35 http://doi.wiley.com/10.1029/2009WR008130. 45 36 Viterbo P, Anton B, Mahfouf JF, Teixeira J. 1999. The representation of soil moisture freezing 46 37 and its impact on the stable boundary layer. Quarterly Journal of the Royal Meteorological 47 38 Society 125: 24012426. 48 39 Viviroli D, Weingartner R. 2004.The hydrological significance of mountains: from regional to 49 50 40 global scale. Hydrology and Earth System Sciences 8(6): 10171030. doi:10.5194/hess8 51 41 10172004. 52 42 Viviroli D, Dürr HH, Messerli B, Meybeck M, Weingartner R. 2007. Mountains of the world, 53 43 water towers for humanity: Typology, mapping, and global significance. Water Resources 54 44 Research3 4 (7): 113. doi:10.1029/2006WR005653. 55 56 57 58 59 60 24 John Wiley & Sons, Ltd 123 Page 25 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 1 Wastl C, Zängl G. 2008. Analysis of mountainvalley precipitation differences in the Alps, 4 5 2 Meteorologische Zeitschrift. 17(11): 311321, doi:http://dx.doi.org/10.1127/0941 6 3 2948/2008/0291. 7 4 Webb EK, Pearman GI, Leuning R. 1980. Correction of flux measurements for density effects 8 5 due to heat and water vapour transfer. Quarterly Journal of the Royal Meteorological Society 9 6 106: 85100. 10 7 11 Wieser G, Hammerle A. Wohlfahrt G. 2008. The Water Balance of Grassland Ecosystems in the 12 8 Austrian Alps. Artic, Antarctic and Alpine Research. 40(2): 439445, doi:10.1657/1523 13 9 0430(07039). 14 10 Wigmosta MS, Vail LW, Lettenmaier DP. 1994. A distributed hydrologyvegetation model for 15 11 complex terrain. Water Resources Research, 30(6): 16651679. 16 12 Wilson DJ, Western AW, Grayson RB. 2004. Identifying and quantifying sources of variability 17 18 13 in temporal and spatialFor soil moisture Peer observations. Review Water Resources Research, 40(2). 19 14 doi:10.1029/2003WR002306. 20 15 Wilczak JM, Oncley SP, Stage SA. 2001. Sonic anemometer tilt correction algorithms. Boundary 21 16 Layer Meteorology 99: 127150. 22 17 Wohlfahrt G, Anfang C, Bahn M, Haslwanter A, Newesely C, Schmitt M, Cernusca A. 2005. 23 18 24 Quantifying nighttime ecosystem respiration of a meadow using eddy covariance, chambers 25 19 and modelling. Agricultural and Forest Meteorology 128(3): 141162. 26 20 Wohlfahrt G, Irschick C, Thalinger B, Hörtnagl L, Obojes N, Hammerle A. 2010. Insights from 27 21 Independent Evapotranspiration Estimates for Closing the Energy Balance: A Grassland Case 28 22 Study. Vadose Zone Journal 9(4): 10251033, doi:10.2136/vzj2009.0158. 29 23 Zanotti F, Endrizzi S, Bertoldi G, Rigon R. 2004. The GEOtop snow module. Hydrological 30 31 24 Processes8 , 1 (18): 36673679, doi:10.1002/hyp.5794. 32 25 Zierl B, Bugmann H. 2005. Global change impacts on hydrological processes in Alpine 33 26 catchments. Water Resources Research, 41 (2): W02028, doi:10.1029/2004WR003447. 34 27 35 28 36 29 37 38 30 39 31 40 32 41 33 42 34 43 44 35 45 36 46 37 47 38 48 39 49 40 50 51 41 52 42 53 43 54 44 55 45 56 57 46 58 59 60 25 John Wiley & Sons, Ltd 124 http://mc.manuscriptcentral.com/ecohydrology Page 26 of 44

1 2 3 1 4 5 2 6 3 7 4 8 5 9 6 10 7 11 12 8 13 9 14 15 10 Tables 16 17 11 18 For Peer Review 19 12 20 21 13 Table 1. Equation of the terms in (13), (14), (15) and (16). 22 23 Term Equations Source 24 25 Photosynthesis 1 26 27 4 28 29 2 30 1 31 Translocation 3 32 33 Respiration 1 34 35 1 36 37 1 38 / 39 40 Senescence 41 42 2 43 44 Litterfall 2 45 46 14 Source: 1, Montaldo (2005); 2, Charles-Edwards et al. (1986); 3, Nouvellon et al. (2000); 4, Eagleson (2002). Notice Bag and Bbb are the above 47 15 and below ground biomass respectively. 48 16 49 17 50 51 18 52 53 19 54 55 20 56 57 21 58 59 60 26 John Wiley & Sons, Ltd 125 Page 27 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 1 4 5 2 6 7 3 8 9 4 10 11 5 12 13 6 14 15 7 Table 2. Field stations topographic characteristic. 16 Parameter B1000 B1500 B2000 17 18 Latitude 46°39’38”For N Peer 46°41’10” NReview 46°41’30” N 19 Longitude 10°35’24” E 10°34’46” E 10°35’30” E 20 Elevation 950 m 1450 m 1950 m 21 Slope 10° 10° 10° 22 Aspect South west South west South west 23 24 Soil depth 100 cm 80 cm 40 cm 25 8 26 27 9 28 29 10 30 31 11 32 12 33 34 13 35 36 14 37 38 15 39 40 16 41 42 17 43 44 18 45 19 46 47 20 48 49 21 50 51 22 52 53 23 54 55 24 56 57 25 58 59 60 27 John Wiley & Sons, Ltd 126 http://mc.manuscriptcentral.com/ecohydrology Page 28 of 44

1 2 3 1 4 5 2 6 7 3 8 9 4 10 11 5 12 13 6 14 15 7 Table 3. General climatic condition during the years 2010 and 2011 along the elevation transect. 16 Year 2010 Year 2011 17 18 P SnowfallFor PeerMean VPD Review P Snowfall [mm] Mean VPD 19 [mm] [mm] Temp [KPa] [mm] Temp [KPa] 20 [°C] [°C] 21 B1000 636 66 (10% of P) 8.1 5.03 580 23.5 (4% of P) 9.5 5.82 22 B1500 686 136 (20% of P) 5.3 3.77 632 85.7 (14% of P) 7.1 4.66 23 B2000 706 193 (27% of P) 2.5 2.65 620 150 (24% of P) 3.5 3.36 24 25 8 26 27 9 28 29 10 30 11 31 32 12 33 34 13 35 36 14 37 38 15 39 40 16 41 42 17 43 18 44 45 19 46 47 20 48 49 21 50 51 22 52 53 23 54 24 55 56 25 57 58 59 60 28 John Wiley & Sons, Ltd 127 Page 29 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 1 4 5 2 6 7 3 8 9 4 10 11 5 12 13 6 14 15 7 Table 4. GEOtopdv model parameters for the case study. 16 Parameter Description Value Source 17 18 For Peer Review 19 minimum stomatal resistance 60 1 20 , sm minimum temperature limitation 273.15 1 21 °C max temperature limitation 323.15 1 22 °C wilting point 23 0.08 cal 24 limiting soil moisture for vegetation 0.15 cal 25 saturated soil moisture 0.52 cal 26 residual soil moisture 0.04 cal 27 saturated hydraulic conductivity 3 1x10 cal 28 29 root zone depth 0.25 obs specific green area for the site B1000 30 0.014 obs 31 m gDM specific green area for the site B1500 0.013 obs 32 m gDM specific green area for the site B2000 0.011 obs 33 m gDM specific leaf area of the dead phytomass 0.02 cal 34 m gDM 35 PhAR extinction coefficient 0.5 2 36 maintenance respiration coefficient for green biomass 0.012 3 37 maintenance respiration coefficient for root biomass 0.007 cal 38 growth respiration coefficient for green biomass 0.25 4 39 growth respiration coefficient for root biomass 40 0.1 cal 41 temperature coefficient in the respiration process 2.5 5 42 death rate of green biomass 0.0064 cal 43 d death rate of root biomass 0.005 cal 44 d allocation coefficient of shoot to root biomass 0.33 1 45 d allocation coefficient of root to shoot biomass 46 0.66 1 d 47 leaf photochemical efficiency 0.0045 1 48 8 Source as follow: 1, Larcher(2003); 2. Eagleson (2002); 3, Amthor (1984); 4, Nouvellon (2000); 5, Aber and Melillo (2001); cal, value from 49 9 model calibration; obs, value from observation. 50 10 51 52 11 53 12 54 55 13 56 57 14 58 59 60 29 John Wiley & Sons, Ltd 128 http://mc.manuscriptcentral.com/ecohydrology Page 30 of 44

1 2 3 1 4 5 2 6 7 3 8 9 4 10 11 5 12 13 6 14 15 7 Table 5. Average value of soil hydraulic parameter. 16 Parameter Range observed 17 18 0.08 0.12 For Peer Review 19 0.42 0.55 20 0.02 0.08 21 2 4 2x10 5x10 22 8 Hydraulic soil parameters were measured at each stationms in order to constrain model calibration within measured ranges. Average values of all 23 9 sites were used for model calibration as soil hydraulic properties were assumed constant. 24 25 10 26 27 11 28 12 29 30 13 31 32 14 33 34 15 35 36 16 37 38 17 39 40 18 41 19 42 43 20 44 45 21 46 47 22 48 49 23 50 51 24 52 53 25 54 26 55 56 27 57 58 59 60 30

John Wiley & Sons, Ltd 129 Page 31 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 1 4 5 2 6 7 3 8 9 4 10 11 5 12 13 6 14 15 7 Table 6. Model calibration on station B1500 and performance on stations B1000 and B2000. 16 Variable Bias RMSE 17 18 B1000 For Peer Review 19 0.01 0.04 0.82 20 1.66 2.9 0.55 21 B1500 cm 22 0.01 0.05 0.78 23 24 1.1 5.9 0.57 cm 25 6.12 58.39 0.93 26 Wm 1.12 47.80 0.88 27 Wm 3.35 66.47 0.81 28 Wm 29 B2000 30 0.01 0.04 0.80 31 1.9 7.1 0.66 32 8 cm 33 34 9 35 36 10 37 38 11 39 12 40 41 13 42 14 43 15 44 16 45 17 46 47 18 48 19 49 20 50 21 51 22 52 23 53 54 24 55 25 56 26 57 27 58 59 60 31 John Wiley & Sons, Ltd 130 http://mc.manuscriptcentral.com/ecohydrology Page 32 of 44

1 2 3 1 4 5 2 6 3 7 4 8 5 9 6 10 7 11 12 8 13 9 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 32

John Wiley & Sons, Ltd 131 Page 33 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 Figure 1. Study area in the upper Vinschgau/Venosta valley in South Tyrol, Italy. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 John Wiley & Sons, Ltd 132 http://mc.manuscriptcentral.com/ecohydrology Page 34 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 Figure 2. Measured monthly mean air temperature and monthly sums of precipitation along the elevation 22 transect for the years 2010 and 2011: station B1000 (red), station B1500 (black), station B2000 (blue). 23 41x17mm (300 x 300 DPI) 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 John Wiley & Sons, Ltd 133 Page 35 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 Figure 3. Observed versus modeled net radiation (Rn), Latent heat (LE) and sensible heat (H) for station B1500, where EC tower data was available (from end April 2011 until October 2011). 48 137x368mm (300 x 300 DPI) 49 50 51 52 53 54 55 56 57 58 59 60 John Wiley & Sons, Ltd 134 http://mc.manuscriptcentral.com/ecohydrology Page 36 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 Figure 4. Observed and modeled time series of cumulative ET at station B1500 for the period where EC 25 tower data was available (from end April 2011 until October 2011). The dashed lines refer to the cutting 26 dates. Missing values were discarded. 49x24mm (300 x 300 DPI) 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 John Wiley & Sons, Ltd 135 Page 37 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 Figure 5. Precipitation and irrigation in transect stations B2000, B1500 and B1000 respectively (panels a), 28 b), c)); comparison of simulated and observed snow depth along the transect, panels d), e), f); comparison 29 of simulated and observed θ5 cm, where the black line and the shaded grey are the mean and standard 30 deviation of the different sensors respectively (panels, g), h), i)). Notice that station B1500 was the 31 calibration site and B1000 and B2000 were the validation sites. 32 88x52mm (300 x 300 DPI) 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 John Wiley & Sons, Ltd 136 http://mc.manuscriptcentral.com/ecohydrology Page 38 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 29 Figure 6. Comparison of simulated (red lines) and observed (black error bars) above ground biomass 30 (panels a), b), c)) and GAI (panels d), e), f)) for the stations B2000, B1500 and B1000 respectively. Due to 31 the harvest, grass canopies undergo multiple growing cycles within a single vegetation period. Notice that station B1500 is the calibration site and B1000 and B2000 are the validation sites. 32 68x44mm (300 x 300 DPI) 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 John Wiley & Sons, Ltd 137 Page 39 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 Figure 7. Effects of the elevation gradient for the two different years on modeled SWE a), cumulative ET b) 29 and frequency distribution of θ 5cm depth c). Red lines refer to station B1000, black lines to station B1500, 30 blues lines to station B2000, while the green dashed line represents the water limitation point assumed in ET 31 the model. Notice that and θ results refer to the snow free period only. 81x50mm (300 x 300 DPI) 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 John Wiley & Sons, Ltd 138 http://mc.manuscriptcentral.com/ecohydrology Page 40 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 Figure 8. The different curves are the time integrals of the factors (1-f1), (1-f2) and (1-f3) as in Eqs. (9, 10 29 and 11) respectively. As the stress functions are normalized, the relative magnitude of their time integrals 30 can be compared over the course of the growing season. Red refers to station B1000, black to station 31 B1500, blue to station B1000. Dashed lines refer to irrigated conditions. a) Cumulative stress for soil f1 32 moisture (1- ). It increases when soil moisture is limiting, while it is flat in optimal conditions. b) Cumulative stress for temperature (1-f2). It increases when temperature is limiting, while it is flat with 33 optimal temperature conditions. c) Cumulative vapor pressure deficit stress (1-f3). 34 78x49mm (300 x 300 DPI) 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 John Wiley & Sons, Ltd 139 Page 41 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 Figure 9. Vertical profiles along the elevation gradient in natural precipitation (dashed lines) for the years 28 2010 (black) and 2011 (red). Black asterisks and red squares refer to the irrigated conditions for the year 29 2010 and 2011, respectively. a) Annual natural precipitation and precipitation+irrigation for the years 2010 30 and 2011, b) Actual ET in natural and irrigated conditions and PET (continuous line). c)ET/PET. d)ET/P in 31 natural and irrigated condition. e) Bag. f) Water use efficiency (WUE = Bag/ET). 32 72x43mm (300 x 300 DPI) 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 John Wiley & Sons, Ltd 140 http://mc.manuscriptcentral.com/ecohydrology Page 42 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Figure 10. Longterm simulation of SWE, ET and surface θ at 5 cm depth along the transect (reference 33 period 19902011). The blue dashed line represents the average of the 20 years and shaded blue areas 34 correspond to the standard deviation (std). The black continuous line and the black dashed line refer to the detailed simulation for the years 2010 and 2011 respectively. Panels a), b) and c) show SWE for station 35 B2000, B1500, B1000 respectively. Panels d), e) and f) show ET for station B2000, B1500 and B1000 36 respectively. The shaded green areas correspond to std and the green dashed line to the time average of T. 37 The shaded red areas correspond to the std and the red dashed line to the time average of E. Panels g), h) 38 and i) represent the frequency distribution of θ 5 cm depth for station B2000, B1500, B1000 respectively. 39 95x70mm (300 x 300 DPI) 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 John Wiley & Sons, Ltd 141 Page 43 of 44 http://mc.manuscriptcentral.com/ecohydrology

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 Figure 11. Yearly mean θ profile for stations B2000, B1500 and B1000 respectively, averaged over the reference period 19902011. 43 74x77mm (300 x 300 DPI) 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 John Wiley & Sons, Ltd 142 http://mc.manuscriptcentral.com/ecohydrology Page 44 of 44

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For Peer Review 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 Figure 12. Vertical profile of ET along the transect. The dashed black line and shaded grey area refer to the 45 mean and standard deviation for the year 1990-2011. The red dashed line refers to the vertical profile of 46 average ET with reduced precipitation (-30%). The blue dashed line refers to the vertical profile of average ET with added precipitation (+30%). The experiment showed that reduced P increased the divergent pattern 47 of ET while the higher P affected mainly B1000. 48 88x98mm (300 x 300 DPI) 49 50 51 52 53 54 55 56 57 58 59 60 John Wiley & Sons, Ltd 143 144 6. Down to future: Transplanted mountain meadows react with increasing phytomass and shifting species composition

Niedrist, G., Tasser, E., Bertoldi, G., Della Chiesa, S., Obojes, N., Egarter- Vigl, L., Tappeiner, U. (2015). Submitted to: Flora

145 146 1

1 Down to future: Transplanted mountain meadows react with increasing phytomass and shifting species

2 composition

4 Georg Niedrista*, Erich Tassera, Giacomo Bertoldia, Stefano Della Chiesaa, Nikolaus Obojesa, Lukas

5 Egarter-Vigla, Ulrike Tappeinera,b

7 a Institute for Alpine Environment, European Academy of Bolzano/Bozen, Drususallee 1, 39100

8 Bolzano/Bozen, Italy

9 b Institute of Ecology, University of Innsbruck, Sternwartestraße 15, 6020 Innsbruck, Austria

10 *Corresponding author: [email protected], Tel. +390471055317

147 2

12 Abstract 13 14 Manipulative approaches under natural conditions are fundamental for understanding impacts of climate

15 warming on grassland (agro-) ecosystems. In this paper we present 3 years of data from two

16 simultaneously conducted transplantation experiments, where meadow monoliths were transplanted

17 downwards along a elevation gradient from the subalpine to the montane belt (2,000 m to 1,500 m a.s.l.),

18 and in parallel from the montane belt to the foothill zone (1,500 m to 1,000 m a.s.l.) respectively. Each

19 downward transplantation simulated a temperature increase of 2.8 K. Control and downward

20 transplanted mesocosms were compared regarding aboveground phytomass, phytodiversity, and species

21 composition. Downward transplanted mesocosms from the upper transplantation reacted significantly

22 to warming in terms of aboveground phytomass (legumes +213.6%, herbs +128.2.6%, graminoids

23 +51.7%, total aboveground phytomass +66.2%), but not with regard to species composition. The lower

24 transplantation, however, induced the complete opposite effect, while average species number and

25 species evenness remained unaffected on all treatments. Further analysis based on five plant traits

26 indicated that the observed shifts were both a consequence of warming and methodological artifacts.

27 Interestingly, the relative importance of warming, artifacts and unaffected species changed with

28 elevation: At the higher transplantation 81.2% of the species remained stable in their abundance, 17.5%

29 were affected by the transplantation, and almost no warming effect could be detected. At the lower

30 transplantation percentage of artifact- and warming- affected species increased consistently (37.5%

31 respectively 44.3%). The results showed that transplantation experiments along elevation transects are

32 an appropriate approach to detect warming impact on agriculturally used grassland at different

33 elevations. Nevertheless, the increasing influence of method-caused side effects became more and more

34 evident over time and with decreasing elevation, underlining the importance of quantifying artifacts in

35 in vivo experiments.

148 3

36 Keywords:

37 Climate Change; biodiversity; plant traits; elevation gradient; multidimensional scaling. 38

149 4

39 Introduction

40 Climate scenarios predict an increase in global mean temperature of 1.5 to 4.8 K by the end of the 21st

41 century due to rising atmospheric greenhouse gas concentration (IPCC 2013). This trend has initiated

42 numerous studies aimed at detecting, quantifying and predicting ecological impacts on grassland

43 ecosystems (for reviews see Dieleman et al., 2012; Izaurralde et al., 2011; Reyer et al., 2013; Rustad et

44 al., 2001). The observed responses vary widely depending on the scale of observation, type of biome

45 and climate region (Elmendorf et al., 2012; Shaver et al., 2000). Thus, rising temperatures affect the N-

46 and C budget of grassland ecosystems by causing higher photosynthesis and mineralization rates (Körner

47 and Larcher, 1988), elongated growing seasons (Myneni et al. 1997) and increased nutrient uptake

48 (Bassirirad, 2000) in areas where nutrients are not limited. On the other hand, in water-limited areas

49 warming can lead to drought effects in terms of reduced respiration and decreased above- and

50 belowground productivity (Hoeppner and Dukes, 2012; Xu et al., 2004). Climatic extreme events are

51 expected to have even more severe consequences (Fuhrer et al., 2006), hence a detailed knowledge of

52 impacts, thresholds and trade-offs of a rising temperatures is fundamental in order to fit agriculture for

53 a warmer future.

54 Field experiments are an effective and common way of studying potential impacts on grassland

55 productivity and species composition (Rustad, 2008). While techniques for precipitation and CO2

56 manipulation are relatively uniform (using Free Air Carbon Dioxide Enrichment and rainout shelters,

57 (Rustad, 2008; Weltzin et al., 2003; Wu et al., 2011) temperature manipulation experiments differ widely

58 regarding their set up, effort and validity. Shaver et al. (2000), and Aronson and McNulty (2009)

59 discussed the strengths and weaknesses of the most frequently used techniques, including field

60 chambers, passive night time warming, overhead infrared heating or active soil warming using heat

61 resistance cables. However, all manipulative experiments lead to an unintended manipulation of the

62 microclimate (Dunne et al., 2004).

150 5

63 Interestingly, space for time substitutions in the form of transplantation experiments are relatively rare in

64 the context of CC issues (De Frenne et al., 2011; Sebastià et al., 2008). Shaver et al., (2000) list two main

65 limitations for transplantation experiments:

66 1) Multiple environment changes impede a clear deduction from a specific treatment to an

67 observed response.

68 To solve this problem an ideal transect/gradient is needed, where site conditions such as exposure,

69 inclination, soil properties and type of management are uniform. Furthermore, meteorological conditions

70 must be comparable except for the intended temperature/precipitation gradient (Koerner, 2007). This

71 refers in particular to the timing of precipitation, which is a determining factor for grassland productivity

72 especially in water limited areas (Robertson et al., 2009).

73 2) In vivo transplanting activities lead to (unwanted) disturbance effects.

74 For realistic conditions entire natural communities (mesocosms) should be transplanted to minimize root

75 damage and keep the natural community structure (Gross et al., 2009). The monoliths should be

76 reinserted safely in their new environment, ideally with direct contact to the surrounding area in order

77 to ensure soil water fluxes and bioturbation processes remain undisturbed. This in turn provokes a high

78 probability of generative and vegetative species invasion from the surrounding area (Bruelheide, 2003;

79 De Frenne et al., 2011). Although preventing, individuating or quantifying an artificially caused species

80 turnover is fundamental for a correct interpretation of the results, many transplantation studies lack of a

81 species monitoring.

82 Otherwise, if a clean temperature gradient can be found and such methodological artifacts can be

83 excluded or at least quantified, there are strong arguments which favor transplant experiments in

84 comparison to other warming methods. The intended elevated temperature affects aboveground and

85 belowground processes evenly without creating any unnatural temperature gradients (Harte and Shaw,

86 1995). Furthermore, realistic climatic conditions are ensured as differences between day and night or

87 sunny and cloudy days are better considered. The greatest advantage of the method is a realistic

88 simulation of future vegetation period length, a factor which is crucial especially for productivity in

151 6

89 mountain grassland ecosystems (Jonas et al., 2008). Not surprisingly, current large scale projects such

90 as the TERENO project in Germany (Zacharias et al., 2011) take advantage of these positive aspects of

91 the transplantation approach and necessitate to understand the effective driving forces for the species

92 turnover in such experiments.

93 Here we present results from a parallel transplantation experiment along an elevation gradient in the

94 Central Alps, where meadow monoliths were transplanted 500 m downwards exposing them for three

95 years to 2.8 K higher temperatures, which is in line with regional climate models for the end of the 21st

96 century (IPCC A2 scenario, Beniston, 2006). Two particularities characterize this approach. At first, a

97 homogeneous elevation transect along one mountain slope from the subalpine belt (2,000 m) to the

98 foothill zone (1,000 m a.s.l.), which ensures comparable site and weather conditions apart from the

99 temperature lapse rate. Second, two identical transplantation steps were conducted in parallel on top of

100 each other, which allows an elevation-resolved interpretation of the results. This is essentially important

101 as CC impacts on grassland are known to vary highly within space and time, and therefore require

102 integrated approaches combining plot-scale experiments with transect observations (De Frenne et al.,

103 2011; Dunne et al., 2004; Rustad, 2008). Thus, we hypothesized that 1) transplanted grassland

104 mesocosms react in positively to warming in terms of aboveground phytomass and phytodiversity and

105 2) the effects are more pronounced in higher elevations. Finally the aim of this study was to quantify

106 method-caused effects on species composition and to test whether transplantations can be reasonably

107 used to study CC impacts on agriculturally used grassland ecosystems.

108

152 7

109 Methods

110 Site description

111 The study was conducted along a homogeneous elevation transect in the Matsch/Mazia-Valley in the

112 Autonomous Province of Bozen/Bolzano (Central Alps, Italy), where three south-west inclined hay

113 meadows areas are located below each other (4.2 km linear distance from the highest to the lowest site).

114 In respect to their elevation they are labeled in this study as B2000, B1500 and B1000 (Figure 1).

115 Microclimate stations were installed directly at the sites, sampling the following parameters on a quarter-

116 hourly scale: air temperature and humidity, wind speed and direction, soil moisture at 5 and 20 cm depth,

117 global shortwave radiation, photosynthetic active radiation above the grassland vegetation canopy and

118 at the soil surface within the canopy, precipitation, and snow height. Mean annual temperature (year

119 2010-2012) ranged from 9.0 °C on the foothill zone (B1000) to 3.4 °C in the subalpine belt (B2000,

120 Table 1). Mean temperature difference during the growing season (1. April-31. October) between both

121 donor and receiving sites was 2.8 C, Mean annual precipitation (2010-2012) reached 568 mm at B1000

122 and increased with 12.2 mm per 100 m elevation. As precipitation scenarios do not predict substantial

123 changes for the central alps (Beniston, 2006) we compensated the natural lapse rate with artificial

124 irrigation (see Table 1).

125 Management of the sites followed traditional land-use at the respective elevation ranging from two cuts

126 at the B2000 site to four cuts at the B1000 site. All three sites were fertilized once a year with 3 kg cow

-2 -2 -2 127 dung per m², corresponding ca. to 15 gm Ntot, 12 gm P2O5 and 22 gm K2O. Soil type on all three

128 sites was determined as loamy-sandy (dystric) cambisoils, pH value decreased slightly with elevation,

129 however, as all three sites were fertilized, none of the main nutrients was assessed to be limiting. (Lair

130 et al. in preparation).

131

132 Treatments

153 8

133 Two parallel transplantations were conducted from B2000 to B1500 and from B1500 to B1000. Control

134 monoliths were also transplanted locally and labeled with “co” after the elevation label, whilst

135 downward transplanted monoliths were given the suffix “tp” (Figure 1). The grassland monoliths have

136 been chosen randomly in the area at the respective site and had a dimension of 0.7 m x 0.7 m, which is

137 in line with other transplantation experiments (Kiehl et al., 2010). Measurements took place within a

138 0.5 m x 0.5 m core area and were surrounded by a 0.1 m wide buffer strip to reduce edge effects.

139 Monoliths depth was ca. 0.25 m which includes most of the roots in temperate meadows (Schenk and

140 Jackson, 2002).

141 The transplantation was conducted manually in order to ensure the most cautious treatment possible and

142 to avoid unnecessary root damage or changed preferential flows. After removing the surrounding soil

143 the monolith was detached by pulling a steel plate under the sample with a hand pulley block. The 180kg

144 monolith was lifted, wrapped with foil to prevent disintegration and spring-cushioned transported to the

145 respective receiving site. Finally the monoliths were lowered flush into prepared holes, and the steel

146 plate was carefully removed. Small gaps (0-3cm) on the edges were filled with soil from the respective

147 origin site of each plot. In contrast to similar studies (e.g. Gavazov et al., 2013) the monoliths were not

148 surrounded by containers in order to keep lateral and vertical soil water fluxes undisturbed. The

149 transplantation process was carried out within the time slot between snowmelt or soil thawing and the

150 start of the vegetation growth (end of March - beginning of April 2010). The transplantation process

151 was successful as neither apparent damage nor a delay in growth compared with the undisturbed

152 vegetation were observed. An additional 0.3 m border round the monoliths was continuously mown (ca.

153 once per week) in order to slow down plant invasion from the surrounding vegetation.

154

155 Data collection and analysis

156 Plant species frequency was counted within a 0.5 m x 0.5 m frame with 25 0.1 m x 0.1 m sub-squares.

157 The assessment took place before the first cut, at a comparable phenological state (flowering of Agrostis

158 capillaris). Monoliths which were transplanted to a lower elevation, were cut at 0.03 m height at the

154 9

159 same time as the control monoliths at the respective donor site. Harvested plant material was divided

160 into functional groups (legumes, herbs and graminoids) and oven dried for 7 days at 80 °C.

161 Downward transplanted monoliths were compared with the respective control monoliths by phytomass

162 (divided into functional groups), species richness, species evenness index and by mean weighted F- and

163 T-Ellenberg indicator values, assuming them as metric values (Diekmann, 2003; Ellenberg, 1974).

164 Control and treated mesocosms were compared by a simple t-test. For not normal distributed (Shapiro–

165 Wilk test) or homoscedastic (Levene's test) data comparison was performed by a non-parametric Mann–

166 Whitney U test. Three-years trend of each treatment was assessed by a linear regression analysis. For

167 the first (2010) and the last year (2012) a multidimensional scaling (MDS) analysis was performed in

168 order to detect changes at community level (species composition).

169 Finally, for interpretation of the observed changes, we hypothesized five driving factors: 1) higher

170 temperatures, 2) decreasing water availability, 3) root damage during the transplantation process,

171 4) increase of ruderal species on bare soil patches and 5) invasion phenomena (i.e. plants from the

172 receiving site which invade the transplanted monoliths). To test this, the following five indicators were

173 calculated for each species: T- and F-Ellenberg indicator values (Ellenberg, 1974), root depth (taken

174 from literature cited in Tasser and Tappeiner, 2005), ruderality and invasion probability. Ruderality was

175 determined by using Grime’s CSR-concept (Grime, 1977). To convert the respective strategy type into

176 numbers a simple linear interpolation was used as described in (Hill et al., 2002); e.g. ruderality of R-

177 species = 1.0, ruderality of CR-species = 0.5 etc.). To quantify the invasion probability an index was

178 calculated summing up the relative and the absolute difference of a species i between the frequency

179 within the transplanted monolith and the control plot at the respective receiving site.

0, 180 100 100 ,

181 : mean frequency [%] of the species i within B1500tp/B1000tp monoliths (year 2010)

182 : mean frequency [%] of the species i within B1500co/B1000co monoliths (years 2010-2012)

155 10

183

184 The index is dimensionless and has a maximum of 200 for the highest invasion probability and a

185 minimum of 0 in the case where a species i is more frequent inside than outside the monolith.

186 Species were clustered into groups with the same 3-year-performance for the control and the

187 transplanted monoliths (“performance groups” see Table 3 and Appendix A and B. A trend was assumed

188 to be relevant with a relative change of ≥ 33% and an absolute change in mean frequency of ≥ 3%

189 between the first and the last year). Finally, for a comparison among the performance groups, the plant

190 traits were species-frequency-weighted and studentized as follows:

191 ∑ ²

192 : studentized plant trait of the performance group p

193 : plant trait value of the performance group p

194 : plant trait value of the performance group which did not change at any elevation (↔↔ group)

195 n : sample size (=6, performance groups)

196 : mean plant trait value of all performance groups

197 The motivation for this kind of normalization was that we were explicitly interested in the comparison

198 between the performance group of species, which were stable (↔↔) and the other groups that showed

199 any change throughout the investigation period. Hence, the value zero in Figure 4 refers to the mean of

200 the ↔↔ group. All analyses and graphs were carried out using the statistical package R 3.0.1.

156 11

201 Results

202

203 Effects of transplantation on phytomass, phytodiversity and species composition

204

205 Even though both transplantations simulated an identic increase in 2.8 C effects between upper

206 (B2000coB1500tp) and the lower (B1500coB1000tp) transplantation differed consistently. This

207 refers especially to phytomass production: The higher transplantation induced a significant increase in

208 legumes (+213.6%), herbs (+128.2%) and graminoids (+51.7%, total aboveground phytomass +66.2%)

209 whereas at the lower transplantation none of the three groups showed a significant increase (Table 2).

210 Regarding the trend throughout the three years under investigation a significant decrease of the legumes

211 on the B2000co, B1500co and B1000tp monoliths was noticed.

212 By contrast, the species richness of the monoliths was not affect by any of the transplantations (Figure

213 2 a+e). The number of species ranged from an absolute minimum of 9 (B1500tp in 2010 and B2000co

214 in 2011) to a maximum of 23 (B1500co in 2010). Evenness (Figure 2 b+f) as well was not affected and

215 remained stable to a considerably high degree throughout the study period (0.7 - 0.9). Similar patterns

216 can also be observed at both transplantations regarding T- and F-values, which showed slight inter-

217 annual variation. T values ranged from 1.71±0.21 at B1500tp in 2010 to 3.18±0.9 at B1000tp in 2011.

218 F-values were generally higher at the higher transplantation and ranged from 5.36±0.16 at B 1500tp in

219 2012 to 3.81±0.28. But again, no significant differences could be detected between control and

220 transplanted monoliths, nor indeed could any changes be detected throughout the investigation period

221 (Figure 2 c,d,g,h).

222 Ordination results from Multidimensional Scaling (MDS, Figure 3) depict changes in species

223 composition between 2010 and 2012. In the first year all three transect sites were clearly separated from

224 each other. No significant difference between local and downward transplanted monoliths could be

225 found. In the last year (2012) however, one interesting pattern became evident: Species composition of

157 12

226 the B1000tp-monoliths, was significantly different from the vegetation of the donor site (B1500co) and

227 shifted towards the B1000co vegetation. At the higher transplantation step, by contrast, species

228 composition of B2000co and B1500tp monoliths, remained unchanged also in the last year of the study.

229 Species were clustered to four (high transplantation) respectively six (low transplantation) performance

230 groups according on their trend in the control and the transplanted monolith (Table 3). On both

231 transplantations most species were assigned to the group, which did not show a change on any of the

232 elevations (↔↔ group). It is notable that at the higher transplantation no species could be assigned

233 either to the group which increased on both treatments (↑ ↑) or to the group which remained stable on

234 the control monoliths and increased on the on the downward transplanted monoliths (↔ ↓).

235

236 Climatic effect versus methodological artifacts

237

238 Finally the performance groups shown in Table 3 were compared among five plant traits (Figure 4).

239 Species, which did not react to transplantation at all (↔↔) were considered as “control species” and

240 showed no remarkable pattern: T-values were higher in the upper transplantation (3.51) than in the lower

241 one (5.19) which was likely to expect. Also root depth increased at lower transplantation from 22.53 to

242 44.15cm. The maximum root length were found for species, which decreased in both the control and the

243 transplanted monoliths (↓ ↓-group), the minimum however for species which increased on both

244 treatments. Species which appeared newly on the transplanted plots (a ↑-group) emerged with similar

245 traits at both transplantations and were characterized by the by far highest T-value (6.48) and invasion

246 index (111.03) Species which remained stable in the control monolith and increased in the downward

247 transplanted monoliths (↔↑) showed similar properties to the “a ↑-group”. At the B2000co  B1500tp

248 transplantation no species increased on the down-transplanted plot while remaining stable at the control

249 plot. It is furthermore remarkable, that traits of the species which were stable at the control plot and

250 decreased at the lower elevation (↔ ↓) were very similar to those of the control group (↔↔) at the

251 lower transplantation while at the higher transplantation ruderality value and root depth appeared

158 13

252 relatively high. However, it must be considered that these values are based only on the traits of

253 Laserpitium halleri, the only species which was assigned to this performance group at the higher step.

254 Finally, species which increased both at control monoliths and at downward transplanted monoliths (↑

255 ↑) were characterized mainly by the highest ruderality-values of all groups.

256 Based on this descriptive analysis it becomes evident, that species, which showed a decrease or an

257 increase on control and transplanted plots(↓ ↓; ↑ ↑) as well as species which appeared newly on

258 transplanted plots (a ↑) must be declared as artifacts. Only the remaining two groups with significant

259 changes in species abundance on transplanted plots and no reaction in the control plot (↔↑, ↔ ↓) are

260 assumed to be a consequence of the warmer conditions. In numbers, 81.2% of the species showed no

261 reaction in their abundance at the higher transplantation, while only 18.2% remained stable at the lower

262 transplantation (Figure 5). Transplantation artifacts were present on both transplantations and

263 contributed with 17.5% (higher transplantation) and 37.5%, (lower transplantation). Only marginal

264 responses to warming were found at the higher transplantation step (1.3%), while 44.3% of the species

265 seem to react to the warmer environment at the lower elevations.

266

159 14

267 Discussion 268 269 Set-up limitations

270 The objective of this study was to investigate how phytomass, phytodiversity and species composition

271 in transplanted mountain meadows react to a 2.8 K warmer climate and whether potential shifts can be

272 assigned to methodological artifacts or the warming effect per se. Even though conducted under natural

273 conditions in the field, digging out plants and reinserting them in a new environment amounts to a major

274 disturbance event (Bruelheide and Flintrop, 2000). Hence, mesocosms were transplanted as a whole

275 instead of subdividing them into subplots in order to minimize edge effects. This in turn resulted in

276 greater labor effort, small sample size and thus low statistical power, which indeed is not uncommon in

277 this kind of experiment (Bloor et al., 2010, n=5; Bruelheide, 2003, n=1). However, results get more

278 robustness as they are based on three years of measurement. Two more aspects must be considered in

279 context with transplantation experiments:

280 1) Environmental conditions. The approach is based on the assumption that the ecological impact of

281 climate change in the future is equivalent to a shift of ecosystems either to higher altitudes or towards

282 the poles (Dunne et al., 2004). Basing on Koerner (2007) we considered most of those parameters such

283 as precipitation amount and timing, land-use and topography. However, other productivity-relevant

284 parameters such as decreasing CO2/O2-partial pressure with increasing elevation were not considered in

285 the experimental set-up. Nevertheless, several studies state that these factors are predominated by the

286 hydrologic conditions (i.e. precipitation, temperature and soil properties) at the respective site (Flanagan

287 and Johnson, 2005; Koerner, 2007).

288 2) Soil conditions. Physical and chemical soil properties can hardly be homogenized in a natural

289 environment in the way that is possible with (artificial) standard soils in in vitro experiments (e.g.

290 Lemmens et al., 2006). In the present elevation gradient all transect sites are situated on grassland sites

291 with at least 150 years of meadow management (cutting, fertilizing; pers. comm. of farmers and historic

292 land-use maps). Despite this anthropogenic homogenization along the transect we observed adaptations

160 15

293 in the downward transplanted monoliths in terms of pH and nutrient content towards the new receiving

294 site. This might have slightly enhanced phytomass production (Lair et al. in preparation), however, soil

295 response of the monoliths was limited due to the relative short duration of the experiment (3 years, see

296 Egli et al., 2004) as well as to the presence of the surrounding 0.1 m buffer strip.

297

298 Effects of transplantation on phytomass, phytodiversity and species composition

299

300 Regarding phytomass two main trends emerge from literature dealing with manipulative warming

301 experiments: An increase in moderate/colder climates (Myneni et al. 1997) and hardly any effect in

302 water-limited regions/experimental designs (De Boeck et al., 2008; Hoeppner and Dukes, 2012). Results

303 from our transect can be interpreted as a combination of both aspects: the higher, (=cooler)

304 transplantation induced a significant increase in aboveground phytomass across all functional groups,

305 whereas the lower (=warmer) transplantation had no significant effect. Here, the warming effect seems

306 to be partly compensated by the elevated evapotranspiration which lead to several short term droughts

307 at B1000 in July and August (Della Chiesa et al., 2014). Looking at the trend throughout the

308 investigation period the continuous decrease of legumes on most treatments (B2000co, B1500co,

309 B1000tp) is remarkable. As this was observed on both control and transplanted monoliths we assume

310 this trend to be an artificial effect of the transplantation process. The question whether it can be attributed

311 to direct mechanical root damage and/or disturbance of the mycorrhizal community is the subject of

312 ongoing studies (Wahl et al, in preparation).

313 There are still differing opinions on how fast grassland species composition react on climatic changes.

314 Grime et al., 2008, Lloret et al., 2004, long-term studies in the Swiss Alps (Vittoz et al., 2009) or the

315 Arctic mountains (Wilson and Nilsson 2009) report a short-midterm inertia of species composition to

316 warming. By contrast, studies from the Spanish Pyrenees depict a strong shift in species diversity after

317 only 17 weeks (Sebastià et al., 2008). This divergence might be explained by the more pronounced

318 change in temperature (Δ +9 K) in combination with the much drier conditions (Δ~800mm) in the afore-

161 16

319 mentioned experiment. Again, results from our experiment acts as synthesis of both trends with stable

320 conditions on the upper transplantation and shifting species composition on the lower one.

321 The observed shift in species composition in the lower transplantation consequently suggests changes

322 in the dominance structure, as described by Vittoz et al., 2009 and Walker et al., 2006. Both studies

323 report a decrease in evenness as a consequence of the better adaptation of single species to warmer

324 conditions and a parallel decrease of shade-intolerant species. Nevertheless, no significant difference in

325 the evenness index could be detected in our study, neither throughout the investigation period, nor

326 between the different treatments. We assume that this inconsistency can be ascribed to the prevailing

327 intensive land-use (2-4 cuts per year), which impedes potentially well-adapted plants to become

328 dominant (Niedrist et al., 2009). Future, long-term observation of the monoliths will show whether

329 impacts of warming will also become evident in such ecosystems which have been subject to strong

330 anthropogenic influences.

331

332 Climatic effect versus methodological artifacts

333 It is not surprising at all that outcomes from a in vivo transplantation experiment are biased by

334 methodological artifacts. Thus it is surprising, that only a few studies have dared to further analyze

335 whether the observed shifts derive from the change in climatic parameters, from methodological artifacts

336 or from a combination of the two. All the more, as results from this study suggest that method caused

337 artifacts vary highly with elevation i.e. climatic region. As a rare exception Bruelheide (2003) assumed,

338 that most of the observed changes came from methodological impacts such as root damage or higher

339 nutrient availability while he found little direct response to enhanced temperatures. In the present study

340 no explicit comparison between transplanted and undisturbed mesocosms has been performed,

341 nevertheless plant trait analysis suggests that all three hypothesized artificial drivers (i.e. root damage,

342 weed species and invasion of new species from the surrounding receiving site) influenced the observed

343 species shift. Not surprisingly, most of the observed artifacts on both transplantation steps were caused

344 on deep rooting species (↓ ↓) which is a common phenomenon in transplantation experiments

162 17

345 (Worthington and Helliwell, 1987). In our experiment, especially legumes (Trifolium repens and T.

346 pretense) and Apiaceae (Anthriscus sylvestris, Carum carvi) were affected. It appeared evidently, that

347 0.25m deep grassland monoliths may contain the majority of the root biomass but are selective to deep-

348 rooting species, especially at lower elevations with deep soils. The increase of annual species (↑ ↑) such

349 as Veronica serpyllifolia, Bromus hordeaceus and Capsella bursa-pastoris on control and transplanted

350 plots indicate bare-soil patches even though their importance was relatively low in contrast to other

351 transplantation experiments. Newly appeared species on the downward transplanted monoliths (a ↑)

352 were well adapted to warmer and dryer conditions (high T- and low F- Ellenberg values). Their high

353 invasion index indicates that they invaded into the transplanted plots during the three years via seedlings

354 or root shoots. However, only an explicit comparison between transplanted mesocosms and undisturbed

355 plots would definitely discriminate this effects. This, indeed, was not the primary objective of this study.

356 All in all, results of this study suggest that grasslands in colder environments react to warming with an

357 increase in phytomass but not in terms of species composition, while in warmer climates the response

358 seems to be the complete opposite. A definition of the ecological conditions which form the limit

359 between robustness and reaction still remains as an important challenge for predicting precise impacts

360 of warming in agriculturally used grasslands.

361

362 Acknowledgments

363 This research emerged from the project “Klimawandel Südtirol”, financed by the Autonomous Province

364 of South Tyrol. We express our warmest thank to all farmers at the LTER site Matsch/Mazia who

365 provided their properties for research purposes and to Francis Howlett for the language revision of the

366 manuscript.

367

163 18

368 References

369

370 Aronson, E.L., McNulty, S.G., 2009. Appropriate experimental ecosystem warming methods by ecosystem,

371 objective, and practicality. Agric. For. Meteorol. 149, 1791-1799.

372 Bassirirad, H., 2000. Kinetics of nutrient uptake by roots: responses to global change. New Phytol. 147,

373 155-169.

374 Beniston, M., 2006. Mountain Weather and Climate: A General Overview and a Focus on Climatic Change

375 in the Alps. Hydrobiologia, 562, 3-16.

376 Bloor, J.M.G., Pichon, P., Falcimagne, R., Leadley, P., Soussana, J., 2010. Effects of Warming, Summer

377 Drought, and CO2 Enrichment on Aboveground Biomass Production, Flowering Phenology and

378 Community Structure in an Upland Grassland Ecosystem. Ecosystems. 13, 888-900.

379 Bruelheide, H., 2003. Translocation of a montane meadow to simulate the potential impact of climate

380 change. Appl. Veg. Sci. 6, 23-34.

381 Bruelheide, H., Flintrop, T., 2000. Evaluating the transplantation of a meadow in the Harz Mountains,

382 Germany. Biol. Conserv. 92, 109-120.

383 De Boeck, H.J., Lemmens, C.M.H.M., Zavalloni, C., Gielen, B., Malchair, S., Carnol, M., Merckx, R., Van

384 den Berge, J., Ceulemans, R., Nijs, I., 2008. Biomass production in experimental grasslands of different

385 species richness during three years of climate warming. Biogeosciences. 5, 585-594.

386 De Frenne, P., Brunet, J., Shevtsova, A., Kolb, A., Graae, B. J., Chabrerie, O., Cousins, S. A., Decocq, G.,

387 De Schrijver, A., Diekmann, M., Gruwez, R., Heinken, T., Hermy, M., Nilsson, C., Stanton, S., Tack, W.,

164 19

388 Willaert, J., Verheyen, K., 2011. Temperature effects on forest herbs assessed by warming and transplant

389 experiments along a latitudinal gradient. Glob. Chang. Biol. 17, 3240-3253.

390 Della Chiesa, S., Bertoldi, G., Niedrist, G., Obojes, N., Endrizzi, S., Albertson, J.D., Wohlfahrt, G.,

391 Hörtnagl, L., Tappeiner, U., 2014. Modeling changes in grassland hydrological cycling along an elevational

392 gradient in the Alps. Ecohydrology :n/a-n/a. doi: 10.1002/eco.1471

393 Diekmann, M., 2003. Species indicator values as an important tool in applied plant ecology - a review.

394 Basic Appl. Ecol. 4, 493-506.

395 Dieleman, W.I.J., Vicca, S., Dijkstra, F.A., Hagedorn, F., Hovenden, M.J., Larsen, K.S., Morgan, J.A.,

396 Volder, A., Beier, C., Dukes, J. S., King, J., Leuzinger, S., Linder, S., Luo, Y., Oren, R., de Angelis, P.,

397 Tingey, D., Hoosbeek, M.R., Janssens, I.A., 2012. Simple additive effects are rare: a quantitative review of

398 plant biomass and soil process responses to combined manipulations of CO2 and temperature. Glob. Chang.

399 Biol. 18, 2681-2693.

400 Dunne, J., Saleska, S., Fischer, M., Harte. J., 2004. Integrating experimental and gradient methods in

401 ecological climate change research. Ecology. 85, 904-916.

402 Egli, M., Hitz, C., Fitze, P., Mirabella, A., 2004. Experimental determination of climate-change effects on

403 above-ground and below-ground organic matter in alpine grasslands by translocation of soil cores. J. Plant

404 Nutr. Soil Sci. 167, 457-470.

405 Ellenberg, H., 1974. Zeigerwerte mitteleuropäischer Gefäßpflanzen. In E. Goltze, Ed. Scripta Geobotanica,

406 9th ed.. Göttingen

407 Elmendorf, S.C., Henry, G.H.R., Hollister, R.D. et al., 2012. Global assessment of experimental climate

408 warming on tundra vegetation: heterogeneity over space and time. Ecol. Lett. 15, 164-175.

165 20

409 Flanagan, L. B., Johnson, B. G., 2005. Interacting effects of temperature, soil moisture and plant biomass

410 production on ecosystem respiration in a northern temperate grassland. Agric. For. Meteorol. 130, 237-253.

411 Fuhrer, J., Beniston, M., Fischlin, A., Frei, C., Goyette, S., Jasper, K., Pfister, C., 2006. Climate risks and

412 their impact on agriculture and forests in Switzerland. Clim. Chang. 79, 79-102.

413 Gavazov, K., Peringer,A., Buttler, A., Gillet,F., Spiegelberger,T., 2013. Dynamics of forage production in

414 pasture-woodlands of the Swiss Jura Mountains under projected climate change scenarios. Ecol. Soc.

415 18(1):38.

416 Grime, J.P., 1977. Evidence for the existence of three primary strategies in plants and its relevance to

417 ecological and evolutionary theory. Am. Nat. 111, 1169-1194.

418 Grime, J.P., Fridley, J.D., Askew, A.P., Thompson, K., Hodgson, J.G., Bennett, C.R., 2008. Long-term

419 resistance to simulated climate change in an infertile grassland. Proc. Natl. Acad. Sci. U.S.A. 105, 10028-

420 10032.

421 Gross, N., Bloor, J.M.G., Louault, F., Maire, V., Soussana, J., 2009. Effects of land-use change on

422 productivity depend on small-scale plant species diversity. Basic Appl. Ecol. 10, 687-696.

423 Harte, J. and Shaw, R., 1995. Shifting dominance within a montane vegetation community: Results of a

424 climate-warming experiment. Science. 267, 876–880.

425 Hill, M.O., Roy, D.B., Thompson, K., 2002. Hemeroby, urbanity and ruderality: bioindicators of

426 disturbance and human impact. J. Appl. Ecol. 39, 708-720.

427 Hoeppner, S.S., Dukes, J.S., 2012. Interactive responses of old-field plant growth and composition to

428 warming and precipitation. Glob. Chang. Biol., 18, 1754-1768.

166 21

429 IPCC, 2013. Summary for Policymakers. In Climate Change 2013: The Physical Science Basis.

430 Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate

431 Change [Stocker, T.F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S. K., Boschung, J., Nauels, A., Xia, Y.,

432 Bex, V., Midgley, P. M. (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York,

433 NY, USA.

434 Izaurralde, R.C., Thomson, A.M., Morgan, J.A., Fay, P.A., Polley, H.W., Hatfield, J.L., 2011. Climate

435 Impacts on Agriculture: Implications for Forage and Rangeland Production. Agron. J. 103, 371-381.

436 Jonas, T., Rixen, C., Sturm, M., Stoeckli, V., 2008. How alpine plant growth is linked to snow cover and

437 climate variability. Journal of Geophysical Research-Biogeosciences, 113, G03013.

438 Kiehl, K., Kirmer, A., Donath, T.W., Rasran, L., Hölzel, N., 2010. Species introduction in restoration

439 projects – Evaluation of different techniques for the establishment of semi-natural grasslands in Central and

440 Northwestern Europe. Basic Appl. Ecol. 11, 285-299.

441 Körner, C., Larcher, W., 1988. Plant Life in Cold Environments. In S.F. Long, F.I. Woodward, Eds., Plants

442 and temperature, pp. 25-57. Society for Experimental Biology, Cambridge, UK

443 Koerner, C., 2007. The use of 'altitude' in ecological research. Trends Ecol. Evol. 22, 569-574.

444 Lemmens, C.M.H.M., Boeck, H.J.D., Gielen, B., Bossuyt, H., Malchair, S., Carnol, M., Merckx, R., Nijs,

445 I., Ceulemans, R.,, 2006. End-of-season effects of elevated temperature on ecophysiological processes of

446 grassland species at different species richness levels. Environ. Exp. Bot. 56, 245-254.

447 Lloret, F., Peñuelas, J., Estiarte, M., 2004. Experimental evidence of reduced diversity of seedlings due to

448 climate modification in a Mediterranean-type community. Glob. Chang. Biol., 10, 248-258.

167 22

449 Myneni, R.B., Keeling, C.D., Tucker, C.J., Asrar, G., Nemani, R.R., 1997. Increased plant growth in the

450 northern high latitudes from 1981 to 1991. Nature. 386, 698-702.

451 Niedrist, G., Tasser, E., Lueth, C., Dalla Via, J., Tappeiner, U., 2009. Plant diversity declines with recent

452 land use changes in European Alps. Plant Ecol. 202, 195-210.

453 Reyer, C.P.O., Leuzinger, S., Rammig, A., Wolf, A., Bartholomeus, R.P., Bonfante, A., De Lorenzi, F.,

454 Dury, M., Gloning, P., Abou Jaoude, R., Klein, T., Kuster, T.M., Martins, M., Niedrist, G., Riccardi, M.,

455 Wohlfahrt, G., De Angelis, P., De Dato, P., Francois, L., Menzel, A., Pereira, M., 2013. A plant's

456 perspective of extremes: terrestrial plant responses to changing climatic variability. Glob. Chang. Biol., 19,

457 75-89.

458 Robertson, T.R., Bell, C.W., Zak, J.C., Tissue, D.T., 2009. Precipitation timing and magnitude differentially

459 affect aboveground annual net primary productivity in three perennial species in a Chihuahuan Desert

460 grassland. New Phytol. 181, 230-242.

461 Rustad, L., Campbell, J., Marion, G., Norby, R., Mitchell, M., Hartley, A., Cornelissen, J., Gurevitch, J.,

462 GCTE-NEWS, 2001. A meta-analysis of the response of soil respiration, net nitrogen mineralization, and

463 aboveground plant growth to experimental ecosystem warming. Oecologia. 126, 543-562.

464 Rustad, L.E., 2008. The response of terrestrial ecosystems to global climate change: Towards an integrated

465 approach. Sci. Total Environ. 404, 222-235.

466 Schenk, H.J., Jackson, R.B., 2002. The global biography of roots. Ecol. Monogr. 72(3, 311-328.

467 Sebastià, M., 2007. Plant guilds drive biomass response to global warming and water availability in

468 subalpine grassland. J. Appl. Ecol. 44, 158-167.

168 23

469 Sebastià, M., Kirwan, L., Connolly, J., 2008. Strong shifts in plant diversity and vegetation composition in

470 grassland shortly after climatic change. J. Veg. Sci. 19, 299-306.

471 Shaver, Gr., Canadell, J., Chapin, Fs., Gurevitch, J., Harte, J., Henry, G., Ineson, P., Jonasson, S., Melillo,

472 J., Pitelka, L., Rustad, L., 2000. Global Warming and Terrestrial Ecosystems: A Conceptual Framework

473 for Analysis. Bioscience, 50, 871-882.

474 Tasser, E.,, Tappeiner, U., 2005. New model to predict rooting in diverse plant community compositions.

475 Ecological Modelling, 185, 195-211.

476 Thuiller, W., Lavorel, S., Araujo, M., Sykes, M., Prentice, I., 2005. Climate change threats to plant diversity

477 in Europe. Proceedings of the National Academy of Science of the United States of America, 102, 8245-

478 8250.

479 Vittoz, P., Randin, C., Dutoit, A., Bonnet, F., Hegg, O., 2009. Low impact of climate change on subalpine

480 grasslands in the Swiss Northern Alps. Glob. Chang. Biol., 15, 209-220.

481 Walker, M.D., Wahren, C.H., Hollister, R.D., Henry, G.H.R., Ahlquist, L.E., Alatalo, J.M. et al., 2006.

482 Plant community responses to experimental warming across the tundra biome. Proceedings of the National

483 Academy of Sciences of the United States of America, 103, 1342–1346.

484 Weltzin, J., Loik, M., Schwinning, S., Williams, D., Fay, P., Haddad, B., Harte, J., Huxman, T., Knapp, A.,

485 Lin, G., Pockman, W., Shaw, M., Small, E., Smith, M., Smith, S., Tissue, D., Zak, J., 2003. Assessing the

486 response of terrestrial ecosystems to potential changes in precipitation. BioScience. 53, 941-952.

487 Wilson, S.D., Nilsson, C., 2009. Arctic alpine vegetation change over 20 years Glob. Chang. Biol., 15,

488 1676-1684.

169 24

489 Worthington, T.R., Helliwell, D.R., 1987. Transference of semi-natural grassland and marshland onto

490 newly created landfill. Biol. Conserv. 41, 301-311.

491 Wu, Z., Dijkstra, P., Koch, G.W., Penuelas, J., Hungate, B.A., 2011. Responses of terrestrial ecosystems to

492 temperature and precipitation change: a meta-analysis of experimental manipulation. Glob. Chang. Biol.,

493 17, 927-942.

494 Xu, L., Baldocchi, D.D., Tang, J., 2004. How soil moisture, rain pulses, and growth alter the response of

495 ecosystem respiration to temperature. Glob. Biogeochem. Cycles. 18, GB4002.

496 Zacharias, S., Bogena, H., Samaniego, L., Mauder, M., Fuss, R., Puetz, T., Frenzel, M., Schwank, M.,

497 Baessler, C., Butterbach-Bahl, K., Bens, O., Borg, E., Brauer, A., Dietrich, P., Hajnsek, I., Helle, G., Kiese,

498 R., Kunstmann, H., Klotz, S., Munch, J.C., Papen, H., Priesack, E., Schmid, H.P., Steinbrecher, R.,

499 Rosenbaum, U., Teutsch, G., Vereecken, H., 2011. A Network of Terrestrial Environmental Observatories

500 in Germany. Vadose Zone J. 10, 955-973.

170 29

546 Graphs and Tables 547

548

549 Figure 1. Location and set-up of the transplantation experiment in South Tyrol/Northern Italy. B+number

550 refers to transect site at the respective elevation in m.a.s.l. The suffix “co” refers to control monoliths,

551 locally transplanted) at the respective elevation, “tp” to monoliths transplanted 500 m downwards to the

552 respective elevation.

553

171 30

554

555 Figure 2. Comparison of control and downward transplanted grassland monoliths from 2,000 to

556 1,500 m.a.s.l., a-d) and from 1,500 to 1,000 m.a.s.l-., e-h) regarding species richness, evenness, F-and T-

557 Ellenberg values. Error bars: standard deviation. For treatment abbreviations (top panel) see Figure 1.

558

172 31

559

560 Figure 3. Ordination results of the Multidimensional Scaling, MDS), based on species frequency of the

561 years a) 2010 and b) 2012. Each point corresponds to one sample of the respective treatment. As all

562 environmental factors were excluded for this analysis, only changes of species composition in relation

563 to each other becomes visible. Method: Euclidean Distance, stress 1 of a) = 0.079, stress 1 of b) =

564 0.083. Circles mark the 95% confidence interval. For treatment abbreviations see Figure 1.

173 32

565

566 Figure 4. Comparison of the performance groups described in Table 3 among five plant traits. Traits were

567 studentized against the ↔↔ group, species which remained stable in frequency on control and transplanted

568 monoliths) throughout the investigation period. Hence, the value “0” refers to the respective trait of the

569 ↔↔ group. For treatment abbreviations, right panel) see Figure 1.

570

174 33

571 572 Figure 5. Relative abundance of three observed responses after three years of grassland transplantation to

573 lower elevations. “Stable” includes species which did not show a response in abundance either on control

574 or on transplanted monoliths, ↔↔). “Warming impact” includes species which remained stable on control

575 monoliths and reacted positively or negatively on the transplanted monoliths, ↔ ↑, ↔ ↓). “Artifacts”

576 includes species which showed a response in abundance at both elevations or newly appeared on the

577 downward transplanted plot, ↓ ↓, ↑ ↑, a ↑ group). Abundances were calculated from the average frequency

578 of the species on the respective control and transplantation plot, shown in Appendix A and B. For

579 performance group description see Table 3, for treatment abbreviations see Figure 1.

580

175 34

581

582 Table 1. Main site characteristics of the elevation transect regarding climate, land-use, soil properties and

583 plant species composition.

584 Site label B1000 B1500 B2000 Exact elevation [m.a.s.l] 950 1,450 1,950 Mean annual temperature [°C] 9.0 6.5 3.4 Mean temperature, April-Oct.) [°C] 13.9 11.1 8.3 Mean annual air humidity [%] 64.4 66.0 72.4 Total annual precipitation [mm] 586 655 708 Annual irrigation [no. of events x mm] 8 x 30 6 x 30 4 x 30 Mean soil moist., 0.05m, April-Oct) [%] 25.2 30.2 34.8 Mean snow cover duration [days] 17 61 116 Mean annual global radiation [W*m-²] 310.9 283.2 289.8 Land-use 4 cuts, 1 fertilizing yr-1 3 cuts, 1 fertilizing yr-1 2 cuts, 1 fertilizing yr-1 Soil pH 5.9 5.2 4.7 Soil C org. / N tot. [g*kg-1] 40.0 / 3.2 95.7 / 7.5 124.1 / 9.7 Dominant plant species Trifolium repens Trifolium repens Trifolium repens Trifolium pratense Trisetum flavescens Festuca rubra agg. Taraxacum officinale agg. Taraxacum officinale agg. Agrostis capillaris Poa pratensis Anthriscus sylvestris Carum carvi Ranunculus bulbosus Dactylis glomerata Phleum rhaeticum 585

176 35

586

587 Table 2. Phytomass and relative change of three functional groups in control and downward transplanted

588 grassland monoliths. Significant differences/changes, p<0.05) are bold. Relative changes were calculated

589 on the basis of a linear correlation between the respective functional group and the years. For treatment

590 abbreviations see Figure 1.

mean of the years 2010-2012 [g*m-2] relative change 2010-2012 [%] treatment B2000co B1500tp p-value B2000co B1500tp legumes 15.5 ± 10.3 48.6 ± 9.5 0,043 -100.0 13.4 herbs 60.6 ± 7.9 136.9 ± 53.4 0,000 25.3 -6.5 graminoids 406.9 ± 49.2 617.4± 145.4 0,025 14.9 -19.8

treatment B 1500co B 1000tp p-value B1500co B1500tp legumes 66.9 ± 30.0 48.8 ± 16.7 0,217 -39.1 -49.7 herbs 371.4 ± 98.8 391.4 ± 144.2 0,914 -5.4 -54.8 graminoids 603.8 ± 113.9 741.9 ± 254.8 0,421 1.9 41.5 591

592

177 36

593 Table 3. Mean plant traits for six performance groups. Performance groups resulted from a trend analysis

594 of all species, Appendix A and B). A trend was considered as relevant if the relative change of a species

595 between 2010 and 2012 was ≥ 33% and the absolute change ≥ 3%. For treatment abbreviations see Figure

596 1.

B2000co → B1500tp B1500co → B1000tp

no. of root no. of root 2010-2012 trend F- invasion T- F- invasion Symbol species T-value depth ruderality species depth ruderality description value index value value index assigned [cm] assigned [cm]

no significant changes ↔↔ either on control or on 20 3.51 5.55 23.53 21.31 0.16 11 5.19 5.08 44.15 26.73 0.11 transplanted monoliths

steady on control and ↔ ↓ significant decrease on 1 3.00 4.00 40.00 0.00 0.33 6 5.00 5.02 38.43 13.04 0.25 transplanted monoliths

significant decrease on ↓ ↓ control and transplanted 4 n.v. 5.00 49.00 5.87 0.22 5 4.91 5.00 74.66 23.77 0.31 monoliths

steady on control and ↔ ↑ significant increase on 0 n.v. n.v. n.v. n.v. n.v. 2 5.00 4.31 13.71 82.35 0.00 transplanted monoliths

increase on control and ↑ ↑ 0 n.v. n.v. n.v. n.v. n.v. 4 5.17 5.50 20.52 14.97 0.52 transplanted monoliths

absent on control and a ↑ new appearance on 3 5.46 5.00 16.64 90.15 0.00 2 6.48 4.45 22.66 111.03 0.00 transplanted monoliths

178 37

597 Appendix A. Effects of the downward-transplantation from 2,000 m to 1,500 m.a.s.l. of grassland monoliths

598 on species frequency for the years 2010 to 2012. n.v.: no value

599

mean 2010-2012 trend 2010-2012 [frequency %] [change %]

Species B2000co B1500tp B2000co B1500tp trend B2000co trend B2000co trend B1500tp Festuca rubra agg. 94.7 83.6 -15.6 -17.5 ↔ ↔ Agrostis capillaris 84.9 72.8 17.1 -15.9 ↔ ↔ Phleum rhaeticum 17.9 31.6 -17.4 6.5 ↔ ↔ Poa pratensis 17.3 11.5 0.0 16.9 ↔ ↔ Alchemilla vulgaris 14.2 58.7 10.4 -1.0 ↔ ↔ Crocus albiflorus 6.7 1.3 0.0 58.1 ↔ ↔ Carum carvi 5.3 11.0 -10.7 -25.3 ↔ ↔ Veronica chamaedris 4.9 0.0 39.4 n.v. ↔ ↔ Cerastium arvense 4.4 2.0 50.8 -100.0 ↔ ↔ Achillea millefolium 4.3 1.9 43.0 92.3 ↔ ↔ Carex pallescens 3.6 0.0 257.1 n.v. ↔ ↔ Campanula scheuchzeri 3.4 9.9 459.6 -24.8 ↔ ↔ Leontodon hispidus 3.2 0.8 135.3 0.0 ↔ ↔ Trisetum flavescens 1.8 4.8 0.0 -25.6 ↔ ↔ Tragopogon pratensis 1.8 3.6 0.0 -43.5 ↔ ↔ Geranium sylvaticum 1.8 3.0 544.2 -45.6 ↔ ↔ Ranunculus acris 1.6 10.1 124.1 -16.4 ↔ ↔ Rhinanthus glacialis 1.3 3.7 -100.0 -46.9 ↔ ↔ Centaurea pseudophrygia 0.8 6.3 125.6 21.1 ↔ ↔ Silene dioica 0.7 3.2 0.0 -48.2 ↔ ↔

Laserpitium halleri 8.2 1.5 34.2 -100.0 ↔ ↓

Trifolium repens 16.4 28.2 -100.0 -99.1 ↓ ↓ Rumex acetosa 12.4 10.9 -70.0 -57.7 ↓ ↓ Taraxacum officinale agg. 6.3 27.1 -100.0 -61.0 ↓ ↓ Trifolium pratense 2.7 12.6 -100.0 -100.0 ↓ ↓

Vicia cracca 0.0 4.4 n.v. 733.3 a ↑ Lolium perenne 0.0 3.7 n.v. 820.0 a ↑ Dactylis glomerata 0.0 3.1 n.v. 888.9 a ↑ 600

179 38

601 Appendix B. Effects of the downward-transplantation from 1500m to 1000 m.a.s.l. of grassland monoliths

602 on species frequency from the years 2010 to 2012. n.v.: no value.

mean 2010-2012 [species frequency] trend 2010-2012 [%]

B1000 Species B1500co B1000tp B1500co tp trend 1500co trend B1000tp Rumex acetosa 34.1 18.7 28.9 24.0 ↔ ↔ Phleum pratense 17.2 7.6 32.5 -21.3 ↔ ↔ Lolium perenne 13.0 16.3 -14.2 -32.8 ↔ ↔ Pimpinella major 10.5 4.9 59.1 33.4 ↔ ↔ Poa trivialis 4.3 4.9 -67.7 0.0 ↔ ↔ Poa pratensis 4.1 0.4 72.2 415.4 ↔ ↔ Ranunculus bulbosus 3.3 1.8 132.7 101.9 ↔ ↔ Silene dioica 3.3 0.0 45.1 n.v. ↔ ↔ Anthoxanthum odoratum 3.1 0.0 -12.1 n.v. ↔ ↔ Myosotis sylvatica 3.1 0.0 240.9 n.v. ↔ ↔ Crocus albiflorus 3.1 0.0 236.4 n.v. ↔ ↔

Trisetum flavescens 61.4 41.8 7.1 -34.5 ↔ ↓ Taraxacum officinale agg. 41.5 28.9 -8.3 -54.1 ↔ ↓ Dactylis glomerata 26.2 21.3 61.9 -46.7 ↔ ↓ Agrostis capillaris 19.2 22.0 44.6 -49.8 ↔ ↓ Heracleum sphondylium 15.8 5.7 7.9 -90.4 ↔ ↓ Festuca pratensis 3.1 2.7 675.0 -100.0 ↔ ↓

Achillea millefolium 19.6 33.0 7.4 55.8 ↔ ↑ Arrhenatherum elatius 12.7 18.7 30.9 79.1 ↔ ↑

Anthriscus sylvestris 39.6 34.7 -36.1 -33.6 ↓ ↓ Trifolium repens 34.4 35.4 -92.9 -74.9 ↓ ↓ Vicia sepium 13.2 3.7 -68.7 -82.5 ↓ ↓ Carum carvi 8.3 8.9 -56.9 -74.7 ↓ ↓ Trifolium pratense 6.8 11.1 -74.0 -87.8 ↓ ↓

Veronica serpyllifolia 35.4 31.1 34.2 41.1 ↑ ↑ Bromus hordeaceus 9.9 8.9 950.2 -43.9 ↑ ↑ Silene vulgaris 6.7 4.9 144.8 209.3 ↑ ↑ Capsella bursa-pastoris 4.7 8.0 242.9 101.9 ↑ ↑

Lolium multiflorum 0.0 7.8 n.v. 1243.3 a ↑ Ranunculus acris 0.0 3.1 n.v. 888.9 a ↑ 603

180 7. Curriculum Vitae

Name: Georg Niedrist Born: 21.01.1981 (Bozen, Italy)

2000-2006: Study of Ecology at the University of Innsbruck (Austria). 2004 Trainee at the „Research Centre for Agriculture and Forestry Laimburg, Department Viticulture 2004-2005: Practical training within the framework of EU-SOKRATES in Italy, Spain and Scotland: Training on Sustainable Development of Mountain Ecosystems. 2006: Diploma thesis within the framework of the Interreg-project “DNA-Chip-Entwicklung zur Charakterisierung und Valorisierung von Bergheu" on the vegetation composition of South Tyrolean hay meadows. Since 2006: Scientific collaborator at the Institute for Alpine Environment (EURAC research, Bolzano/Bozen, Italy), 2006-2008: Collaboration within the Interreg-project “Measures and strategies for a sustainable mountain agriculture (MASTA)” 2008 ongoing: Setup and site management of the emerging LTER-site “Matsch/ Mazia valley” 2010-2016: PhD student at the University of Innsbruck, Faculty of Natural Sciences, Institute of Ecology 2010 and 2015 Leading collaborator of the project “Wiesenmeisterschaft Südtirol” (meadow championship) 2015: Visiting scientist at the Colorado State University, Natural Resource Ecology Laboratory (U.S., 2 months).

Areas of expertise:

Good knowledge about alpine flora and alpine vegetation ecology. Profound knowledge about mountain agriculture in the European Alps (policy as well as reality). Experienced in micro-meteorological measurements with a particular focus on plant productivity and soil moisture measurements.

181 8. Additional publications, posters and presentations

8.1 Peer reviewed publications

Niedrist G., Tasser E., Lüth C., Tappeiner U. (2009): Botanisch-ökologische Untersuchungen des Wirtschaftsgrünlandes in Südtirol unter besonderer Berücksichtigung der Bergmähder. Gredleriana, 9: 11-32. Lüth C., Tasser E., Niedrist G., Dalla Via J., Tappeiner U. (2010): Classification of the Sieversio montanae-Nardetum strictae in a cross-section of the Eastern Alps. Plant Ecology , doi10.1007/s11258-010-9807-9 . Pasolli, L., Notarnicola, C., Bruzzone, L., Bertoldi, G., Della Chiesa, S., Hell, V., Niedrist, G., Tappeiner, U., Zebisch, M., Del Frate, F., Vaglio Laurin, G. (2011): "Estimation of Soil Moisture in an Alpine Catchment with RADARSAT2 Images," Applied and Environmental Soil Science, vol. 2011, Article ID 175473, 12 pages, 2011. doi:10.1155/2011/175473 Tasser, E., Lüth, C., Niedrist, G., Tappeiner, U. (2011): Bestimmungsschlüssel für landwirtschaftlich genutzte Grünlandgesellschaften in Tirol und Südtirol. Gredleriana 10, 11-62. Pasolli L., Notarnicola C., Bruzzone L., Bertoldi G., Della Chiesa S., Niedrist G., Tappeiner U., Zebisch M. (2011): Polarimetric radarsat2 imagery for soil moisture retrieval in alpine areas, Canadian Journal of Remote Sensing, 37(5), 535-547, doi: 10.5589/m11-065. Mair E., Bertoldi G., Leitinger G., Della Chiesa S., Niedrist G., Tappeiner U. (in review): ESOLIP – estimate of solid and liquid precipitation at sub-daily time resolution by combining snow height and rain gauge measurements, Hydrol. Earth Syst. Sci. Discuss., 10, 8683-8714, doi:10.5194/hessd-10-8683-2013 Bertoldi G., Della Chiesa S., Notarnicola C., Pasolli L., Niedrist G., Tappeiner U. (2014): Estimation of soil moisture patterns in mountain grasslands by means of SAR RADARSAT2 images and hydrological modeling. Journal of Hydrology doi:10.1016/j.jhydrol.2014.02.018. Mao L., Dell’Agnese A., Hiuncache C., Penna D., Engel M., Niedrist, G., Comiti F. (2014): Bedload hysteresis in a glacier-fed mountain river. Earth Surface Processes and Landforms, DOI: 10.1002/esp.3563 Pasolli, L., Notarnicola, C., Bertoldi, G., Della Chiesa, S., Niedrist, G., Bruzzone, L., Tappeiner, U. and Zebisch, M. (2014), Soil moisture monitoring in mountain areas by using high-resolution SAR images: results from a feasibility study. European Journal of Soil Science. doi: 10.1111/ejss.12189

182 Mair E., Leitinger G., Della Chiesa S., Niedrist G., Tappeiner U., Bertoldi G. (2015): A simple method to combine snow height and meteorological observations to estimate winter precipitation at sub-daily resolution. Hydrological Sciences Journal. DOI: 10.1080/02626667.2015.1081203 Pasolli L., Notarnicola C., Bertoldi G., Bruzzone L., Remegaldo R., Niedrist G, Della Chiesa S., Tappeiner U., Zebisch M. (2015): Estimation of Soil Moisture in Mountain Areas Using SVR Technique Applied to Multiscale Active Radar Images at C-Band. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 8(1), 262–283. doi:10.1109/JSTARS.2014.2378795 Greifeneder F., Notarnicola C., Bertoldi G., Niedrist G., Wagner W. (2015): From point to pixel scale: An upscaling approach for in-situ soil moisture measurements Submitted to Vadose Zone Journal. Wahl A.L., Tillet L., Niedrist G., Lair G., Tappeiner U., Spiegelberger T. (2015): Temperature increase affects root colonisation by arbuscular mycorrhizal fungi in a subalpine but not in a montane grassland after three years of climate change simulation. Submitted to Journal of Ecology. Wohlfahrt G., Hammerle A., Niedrist G., Scholz K., Tomelleri E., Zhao P. (2015): On the energy balance closure and net radiation in complex terrain. Submitted to Agricultural and Forest Meteorology.

8.2 Book contributions

Tasser E., Niedrist G., Zimmermann P., Tappeiner U. (2010): Species richness in space and time as an indicator of human activity and ecological change. In: Jorgensen S.E., Xu L., Costanza R. (eds.): Handbook of Ecological Indicators for Assessment of Ecosystem Health, Second Edition. CRC Press, Washington, D.C., USA, 147-167 Elmi M., Hoffmann C., Niedrist G., Pechlaner H., Pedoth L., Pinzger S., Pistocchi A., Tappeiner U., Tasser E., Zebisch M. (2011): Klimareport Südtirol Niedrist, G., Tasser, E., Tappeiner, U.(2011): Landwirtschaft. In: Elmi M., Hoffmann C., Niedrist G., Pechlaner H., Pedoth L., Pinzger S., Pistocchi A., Tappeiner U., Tasser E., Zebisch M. (2011): Klimareport Südtirol Niedrist G., Tasser E., Tappeiner U.(2011): Forstwirtschaft. In: Elmi M., Hoffmann C., Niedrist G., Pechlaner H., Pedoth L., Pinzger S., Pistocchi A., Tappeiner U., Tasser E., Zebisch M. (2011): Klimareport Südtirol

183 8.3 Oral presentations

Niedrist G., Tasser E., Lüth C., Baric S., Letschka T., Dalla Via J., Tappeiner U.: Botanische Grundlagen einer Bergheuzertifizierung. 5. Tagung zur Zoologische und botanische Forschung in Südtirol, 05.09.2008. Niedrist G., Lüth C., Tasser E., Letschka T., Dalla Via J., Baric S., Tappeiner U. (2009): Una base botanica per la certificazione del fieno alpino. XIX Congresso S.It.E., 15.-18.09.09, Bolzano. Niedrist G., Tasser E., Tappeiner U. (2010): Sukzessionsuntersuchungen am Beispiel artenreicher Goldschwingelwiesen. 6. Tagung „Zoologische und botanische Forschung in Südtirol“, 02-03.09.2010, Bozen (I). Niedrist G., Tasser E., Tappeiner U. (2010): Wiesenmeisterschaft Südtirol: ein Streifzug durch die Vielfalt unserer Wiesen. 6. Tagung „Zoologische und botanische Forschung in Südtirol“, 02-03.09.2010, Bozen (I). Niedrist G., Tasser E., Tappeiner U. (2011): Abandonment and re-mowing in species-rich Festuca paniculata-grassland. 41st Annual Meeting of the Ecological Society of Germany, Austria and Switzerland, 5-9.09.2011, Oldenburg (D). Niedrist G., Zebisch M. (2011): Climate Change research: Global-Alps-South Tyrol. 17.11.2011, Research Centre for Agriculture and Forestry – Laimburg (I). Niedrist G., Tasser E., Rüdisser J., Tappeiner U. (2012): Development and implementation of biodiversity indicators from ecosystem level to an alpine- wide scale. Workshop of the Global Mountain Biodiversity assessement, 23.06.2012, Kazbegi (Georgia). Niedrist G., Tappeiner U.: Sustainability research in mountain regions. Graubünden forscht, Academica Raetica, 13.09.2012, Davos (CH). Niedrist G., Bertoldi G., Della Chiesa S., Engel M., Obojes N., Tasser E., Tappeiner U. (2013): Down to the future: Insights from manipulative Climate Change experiments along an elevation transect in dry, mountain grasslands (Matsch valley, Italy) IRSTEA, Grenoble (FR) Niedrist G., Tasser E., Bertoldi G., Della Chiesa S., Obojes N., Egarter-Vigl L., Tappeiner U. (2014), 3 years of transplantation experiments in the Vinschgau valley: Impact of elevated temperatures on phyto-diversity and productivity of grassland meadows. Zoologische und botanische Forschung in Südtirol, 4-5.09.2014, Bozen (Italy). Niedrist G., Tasser E., Bertoldi G., Della Chiesa S., Obojes N., Egarter-Vigl L., Tappeiner U. (2014); Shifting plant species composition in transplanted mountain grasslands: Is it all Climate Change? GfÖ, September 08th-12th, Hildesheim (D) Niedrist G.(2015): Wiesenmeisterschaften in Südtirol – Wettbewerb und Methodik zur Bonitur für ungedüngte Extensivwiesen und gedüngte

184 Intensivwiesen. Grünlandkolloquium 13.-14.04.2015 Oberelsbach, (Germany, invited presentation).

8.4 Posters

Niedrist, G., Bertoldi, G., Della Chiesa, S., Hell, V., Tasser, E., Tappeiner, U. (2011): Transplantation experiments in an inner-alpine dry valley to predict Climate Change effects on agriculturally used grassland ecosystems. European Geosciences Union 2011, 03 – 08.04.2011, Vienna (A). Niedrist G., Bottarin R., Bertoldi B., Della Chiesa S., Obojes N., Tasser E., Tappeiner U. (2012): A trans-disciplinary approach to quantify possible future impacts of climate change on mountain regions. XXII Congresso S.It.E., 10.- 13.09.12, Alessandria (I). Niedrist G., Obojes,N, Bertoldi G., Della Chiesa S., Tasser E.,Tappeiner U. (2013) Grassland transplantations along an elevation gradient: Evaluating impacts of a simulated temperature increase on phytodiversity and aboveground productivity. Conference Moutains under Watch, 20.-21.02.2013, Bard, Aosta (I). Niedrist G., Obojes N, Bertoldi G., Della Chiesa S., Tasser E.,Tappeiner U. (2013) Spatiotemporal variability of increasing temperature impacts on grassland vegetation along an elevation transect in the Alps. European Geosciences Union, General Assembly 07.-12.04.2013, Vienna (AUT) Niedrist G., Bottarin R., Tasser, E., Zerbe, S., Tappeiner U. (2014) LTER-Fläche Matscher Tal. IALE Jahrestagung 15.-17.10.2014, Bolzano/Bozen (I) Niedrist G., Bertoldi G., Della Chiesa S., Fontana, V., Tasser, E., Tappeiner U. (2015): Long Term Ecological Research Matsch /Mazia (Italian Alps): An interdisciplinary open-air labratory between apple orchards and glaciers. LTER All scientists meeting Estes Park (CO) United States, 30.08-03.09.2015.

185 9. Contribution statement

I hereby declare that I contributed as follows to the scientific publications present in this PhD-thesis:

Lüth C., Tasser E., Niedrist G., Dalla Via J., Tappeiner U. (2010): Plant communities of mountain grasslands in a broad cross-section of the Eastern Alps. Flora 206(5):433-443. G Niedrist contributed to study conception, did parts of the vegetation survey and commented on earlier version of the manuscript.

Niedrist G., Tasser E., Lüth C., Dalla Via J., Tappeiner U. (2009): Plant diversity declines with recent land use changes in European Alps. Plant Ecology 202:195- 210. G Niedrist contributed to study conception, vegetation surveying, did the data elaboration and wrote the manuscript.

Reyer, C., Leuzinger, S., Rammig, A., Wolf, A., Bartholomeus, R. P., Bonfante, A., de Lorenzi, F., Dury, M., Gloning, P., Abou Jaoudé, R., Klein, T., Kuster, T. M., Martins, M., Niedrist, G., Riccardi, M., Wohlfahrt, G., de Angelis, P., de Dato, G., François, L., Menzel, A. & Pereira, M. (2012): A plant’s perspective of extremes: Terrestrial plant responses to changing climatic variability. Glob Change Biol 19:75–89. G Niedrist participated at the EGU session which was the basis for this review, contributed to the concept development by commenting on former draft versions and wrote small parts of the manuscript.

Della Chiesa, S., Bertoldi, G., Niedrist, G., Obojes, N., Endrizzi, S., Albertson, J. D., Wohlfahrt, G., Hörtnagl, L., Tappeiner, U. (2014): Modeling changes in grassland hydrological cycling along an elevational gradient in the Alps. Ecohydrol 7:1453–1473 G Niedrist contributed to study conception, provided validation data, discussed the results and commented on earlier versions of the manuscript.

186 Niedrist, G., Tasser, E., Bertoldi, G., Della Chiesa, S., Obojes, N., Egarter-Vigl, L., Tappeiner, U. (2015) Down to future: Transplanted mountain meadows react with increasing phytomass and shifting species composition. Submitted to: Flora G Niedrist designed the study, did the sampling and the data elaboration and wrote the manuscript.

Leopold‐Franzens‐Universität Innsbruck

Eidesstattliche Erklärung

Ich erkläre hiermit an Eides statt durch meine eigenhändige Unterschrift, dass ich die vorliegende Arbeit selbständig verfasst und keine anderen als die angegebenen Quellen und Hilfsmittel verwendet habe. Alle Stellen, die wörtlich oder inhaltlich den angegebenen Quellen entnommen wurden, sind als solche kenntlich gemacht.

Die vorliegende Arbeit wurde bisher in gleicher oder ähnlicher Form noch nicht als Magister- /Master-/Diplomarbeit/Dissertation eingereicht.

Datum Unterschrift

187 10. Acknowledgments/Danksagung

Mein erster Dank geht an Ulrike und Erich: Ihr habt mir nach dem Studium die Möglichkeit geboten an der EURAC zu arbeiten. Ihr habt mich gefördert, mir viel Spielraum und damit Vertrauen gegeben und mich durch zahlreiche -auch intensive- Diskussionen mit der Welt der Wissenschaft vertraut gemacht.

I would like to thank my working colleagues, especially those who are crossing my daily life since many years now: Roberta, Chiara, Caro, Barbara, Giacomo, Lukas, Klaus, Ale and all other colleagues: It were your ideas, your support and your critics that brought me forward! Un ringraziamento speciale a te Stefano per il tuo aiuto, per i Winkl e per avermi introdotto alla cultura Romana…

Als ein mit öffentlichen Mitteln bezahlter Wissenschaftler ist es mir ein Anliegen mich stellvertretend bei der Südtiroler Landesregierung zu bedanken. Durch die Unterstützung von Forschungseinrichtungen wie der EURAC gibt sie jungen (und nicht mehr so jungen) Forschern die Möglichkeit und das Vertrauen Forschung im Dienst der Allgemeinheit durchzuführen.

„Ohne (Arbeits-)Pause ist nichts von Dauer.“ (Ovid) Danke für viele unterhaltsame, ablenkenden und manchmal auch anstrengende Mittags- und Urlaubspausen an die BHRLBM?-Gruppe!

Die Hypothese, dass die Variablen Freundschaft und Entfernung miteinander korrelieren, konnte ich zusammen mit dir, Martin, erfolgreich widerlegen!

Keine Rebe kann ohne Stütze nach oben wachsen. In meiner Familie hatte ich eine Stütze, die immer da war und mich auf meinem Weg bestärkt hat. Donkschean!

Geduldiger Boxsack wenns fuxt, sicheres TomTom wenns düster wird, zuverlässige Bremse wenn sich das Hamsterrad zu schnell dreht und trotzdem steter Antrieb, nicht zuletzt was den Abschluss dieser Arbeit anbelangt. Ich danke dir Sylvia.

188