Ecol Res (2015) 30: 989–1003 DOI 10.1007/s11284-015-1300-4

O R I GI N A L A R T IC L E

Joa˜o Rocha • Samantha Jane Hughes Paulo Almeida • Isabel Garcia-Cabral Franciso Amich • Anto´nio L. Crespı´

Contemporary and future distribution patterns of fluvial vegetation under different climate change scenarios and implications for integrated water resource management Received: 4 November 2014 / Accepted: 1 August 2015 / Published online: 29 August 2015  The Ecological Society of Japan 2015

Abstract Knowledge of plant community structure and how it can confer resistance to climate change effects is required for the management of fluvial ecosystems. Findings from such studies can be applied in decision making processes to implement measures to maintain, conserve or improve fluvial quality. Floristic and environmental data from 100 sample stations located in three River Basin Districts in northern Portugal were gathered as part of the 2010 Water Framework Directive monitoring program carried out on mainland Portugal. Three habitat types were characterized based on the flow dynamic level: the wetted channel, the bankfull area and the riparian gallery. Hierarchical cluster analysis of environmental data revealed three distinct environmenElectronic supplementary material The online version of this article (doi:10.1007/s11284-015-1300-4) contains supplementary material, which is available to authorized users. J. Rocha Department of Biology, Faculty of Sciences, University of Porto, Rua Campo Alegre s/n, 4169-007 Porto, Portugal S. J. Hughes CITAB-UTAD, Centre for Research and Technology of AgroEnvironment and Biological Sciences, Department of Forestry and Landscape Architecture, University of Tra´s-os-Montes and Alto Douro, 5001-801 Vila Real, Portugal P. Almeida Æ I. Garcia-Cabral Herbarium, University of Tra´s-os-Montes e Alto Douro, 5001-801 Vila Real, Portugal F. Amich Evolution, Taxonomy and Conservation Group (ECOMED), Department of Botany, University of Salamanca, 37007 Salamanca, Spain A. L. Crespı´ (&) CITAB-UTAD, Centre for Research and Technology of AgroEnvironment and Biological Sciences, University of Tra´s-osMontes and Alto Douro, Vila Real Herbarium, 5001-801 Vila Real, Portugal E-mail: [email protected] Tel.: +351-259-350223

tal groups of sites. Floristic data were organized by these environmental groups characterized by altitudinal, temperature and precipitation data variables. The combination of taxonomic diversity and species frequency reflect functional differences for these habitats, here explained by a resistance and resilience approach. More low-frequency species and higher levels of functional diversity occurred at stations with more variable environmental conditions. Predictive modelling of the future distribution of the three environmental groups under two different climate scenarios supported the relevance of low-frequency traits in conferring resistance to climatic change effects. Keywords Fluvial vegetation Æ Species frequency Æ Floristic-structural characterization Æ Environmental variability Æ Climate change

Introduction Fluvial zones are a fundamental component of freshwater ecosystems, playing vital roles in groundwater recharge, nutrient cycling, buffering against erosion and pollutant transfer and providing habitat stability, complexity and connectivity (Naiman and De´camps 1997; Rodewald and Bakermans 2006). The resistance, resilience and the numerous fundamental processes associated with fluvial systems must be harnessed and guaranteed to support aquatic resource policy and Water Framework Directive River Basin Management Plans (WFD, European Commission 2000) to ensure the sustainability of aquatic resources and ecosystem services, especially under conditions of climate and land use change (Seavy 2009). A better understanding is necessary of how fluvial plant community composition, structure and function contribute to resilience and how contributes to ecological process and structure and promote the importance of ecotonic

990

systems within the setting of the landscape (Peterson 2002, Jiang et al. 2012). The associations between structure and resilience are complex, but community composition is a vital component in assessing responses to environmental disturbances (Taylor et al. 2010). Studies on the floristic characterization of Portuguese fluvial habitats are relatively scarce (Aguiar et al. 2000; Crespı´ et al. 2001; Aguiar and Ferreira 2006; Feio et al. 2014). Most studies have focused on the ecological integrity or quality of these systems (Aguiar and Ferreira 2005; Ferreira et al. 2005; Aguiar et al. 2007, 2009; Bernez and Ferreira 2007; Duarte et al. 2007) or have included them in models to predict ecological quality (Feio et al. 2007). Improving our understanding of the dynamics of the flora and vegetation of fluvial habitats is vitally important for effective management of these ecotones. This are often degraded because of changes in land use (e.g. agriculture and urbanization) but, at the same time, are recognized as fundamental ecosystems with an important role in meeting the demands of key European Directives such as the WFD the Pesticides Framework Directive, the Nitrates Directive and the Habitats Directive (European Commission 2015). Other types of anthropogenic pressures also affect fluvial ecosystems in northern Portugal (Aguiar & Ferreira 2005; Aguiar et al. 2007; Feio et al. 2014). The national dam building program ‘‘Programa Nacional de Barragens com Elevado Potencial Hidroele´ctrico’’ or PNBEPH (Ministe´rio do Ambiente, Ordenamento do Territo´rio e Energia 2015) is an example of these disturbances. The aim of this program is to reduce national dependency on imported energy by promoting alternative endogenous domestic wind and hydroelectric energy production by building several hydroelectric dams across northern Portugal (Hughes et al. 2012). However, hydroelectric infrastructures disturb the river continuum and the natural flow regimes, and consequently fluvial vegetation (Jansson et al. 2005; Bernez and Ferreira 2007; Bombino et al. 2008). All these anthropogenic alterations result in changes in the regional taxonomic diversity (Sabo et al. 2005). Based on data from sites located across three River Basin Districts in northern Portugal, this study focuses on species diversity and frequency patterns in relation to environmental heterogeneity in fluvial systems. The role of low-frequency species in the structural organization of vegetation is assessed. Traditionally, low-frequency species have been considered more variable in terms of their abundance (May 1974; Tilman 1996; Chapin III et al. 2000), and are the main reason for non-linear ecosystem effects (Holling 1973; Pimm 1984; Ives and Carpenter 2007). Non-linearity in ecosystems raises questions with regard to the existence of stability in ecosystem organization and function (Buiatti and Longo 2013), as well as the redundancy of species functionality (Gitay et al. 1996; McGrady-Steed et al. 1997; Suding et al. 2008).

This study focuses on the analysis of taxonomic diversity and floristic-structure variability measured through the frequency of species. It models fluvial communities under climate change scenarios to describe correlations between the diversity, resistance and resilience of these vegetation communities. The effect and importance of low-frequency species in fluvial habitats and floristic-structural responses is analyzed in relation to environmental variability (in this case, elevation, temperature and precipitation data). Based on their environmental characteristics, future climate change scenarios are modelled for the sampling stations to describe changes in the correlation between species frequency and environmental variability. High frequency species diversity per vegetation formation is usually less than that of low-frequency species. Phytosociological studies carried out in Portugal by Pereira (2013) revealed that 87 % of fluvial flora had frequencies lower than 50 %, and that 70 % of the flora had frequencies between 10 and 20 %. Honrado (2003) found that 70 % of species had frequencies lower than 50 %, and that 30 % of species had frequencies between 10 and 20 %. Thus, the description of low-frequency species patterns is vital for describing the ecological functionality of these communities, since the probability of ecological function overlap becomes greater (Walker 1995; Allen et al. 2005). This functional redundancy is directly correlated with community resilience and resistance (Peterson et al. 1998; Allison and Martiny 2008). However, more recently the lack of correlation between resistance and resilience has been discussed in terms of the Yachi and Loreau insurance hypothesis (Allison 2004). The results of this study are discussed in terms of the relationship between functionality (here analyzed using the relationship between taxonomic diversity and species frequency), and the structural and floristic resistance and resilience of fluvial vegetation. Such findings provide fundamental information on how climate change can affect these systems and the ecosystem services they can potentially provide, allowing the development of mitigation, adaptation and restoration initiatives.

Methods Study area Field data were collected in 2010 from 100 fluvial sample stations distributed across northern Portugal (Fig. 1; Table S1) as part of the first mandatory nationwide WFD compliant monitoring programme. Sample stations were previously selected by the Portuguese National Water Institute (Instituto Nacional de A´gua, INAG), the national governmental body as a part of the WFD national monitoring programme (INAG 2008). The geographical area (25,271 km2) covered three River Basin Districts (RBD): RBD 1 (Minho and Lima), RBD

991

Fig. 1 The study area in northern Portugal, comprising River Basin Districts 1 Minho-Lima), 2 (Ca´vado-Ave) and 3 (Douro), according to WFD proceedings (European Commission 2009). The black dots indicate the WFD monitoring program sample stations

2 (Ca´vado, Ave and Lec¸a) and RBD 3 (the Douro). These RBD include the catchments and sub catchments of three transnational rivers, the Minho, Douro and Lima, and contain over 3,500,000 inhabitants (approximately 35 % of the total population of Portugal according to the Portuguese National Institute of Statistic 2012). The sample stations for all three RBD covered a wide range of environmental conditions, ranging from reference to highly degraded states. The tributaries of the major transnational rivers are mostly small (between 200 and 20 km long), flowing through deep, narrow canyons while the lower littoral reaches of these river systems are intensely urbanized and degraded. Remaining areas are dominated by agroforestry activities with low population density levels (Cortes et al. 2013). Climacic fluvial vegetation in Northern Portugal is characterized by a compact gallery forest, dominated by the Alno-Padion Alliance (Rivas-Martı´ nez and de la Fuente 1986), with a dense tree and shrub canopy dominated by Alnus glutinosa (L.) Gaertn., Fraxinus angustifolia Vahl, Salix spp., Frangula alnus Mill., Rubus spp., Sambucus nigra L. and Erica arborea L. Aquatic herbaceous vegetation is characterized by Ranunculus subgenus Batrachium (DC.) A. Gray and Potamogeton spp., while immediate bankside vegetation is typically dominated by Oenanthe croccata L., Carex spp. and several Juncaceae and Cyperaceae species. The riparian gallery tends to comprise Glyceria spp., Brachypodium spp., Agrostis spp., Trifolium spp., and Crepis spp. (Andre´s et al. 1986; Molina 1996; Crespı´ et al. 2001; Hoelzer 2003; Aguiar and Ferreira 2005; Dias Pereira 2009). A predominantly herbaceous and seasonal bankside community emerges between gallery and aquatic vegetation that is characterized by a diverse number of important families and species (Compositae; Rubiaceae—Galium spp., Lythraceae—Lythrum spp., Polygonaceae—Polygonum spp., Juncaceae—Juncus spp., Cyperaceae -Cyperus spp., Plantaginaceae -Veronica spp., Poaceae -Poa spp., Paspalum spp., (systematics

in accordance with Angiosperm Phylogeny Website, version 13, Stevens 2001 onwards). Collection of field data Fluvial communities located within the first 3 m from the edge of the wetted channel (Fig. 2a, b) were characterized at each sample station within a 1000 m2 area following the INAG WFD-compliant field protocol (INAG 2008). The criterion of three meters into the wetted channel was applied to standardize the location of sample stations. Thus, the sampling area for aquatic habitat was the same for all the sample stations, while the area of the bankside and gallery habitats varied according to the ecological dynamics of the stream (i.e. a more stable water level within the sampling season results in a smaller bankside habitat area and greater ecological relevance of the gallery habitat). Sampling was carried out during the late spring and early summer of 2010 when plant growth was at its peak. Three separate habitats were sampled within this area (Fig. 2b), namely (A) the wetted channel with permanent flow and floating or rooted vegetation within the bankfull area of the sample station, (B) the bankside area—where herbaceous and ephemeral vegetation emerges between the wetted channel and the gallery and (C) the bankside gallery, comprising perennial forests and shrub communities situated beyond the immediate bankfull area. The extent of these three habitat types is directly influenced by temperature and precipitation regimes within the catchment area. These are in turn related to the North Atlantic Oscillation (Trigo et al. 2004; Santos et al. 2005) which influences annual and interannual thermic and pluviometric variability. Drier periods occur between the end of spring and the beginning of autumn while precipitation tends to occur during the winter months (Trigo and DaCamara 2000). These habitats and their interactions with the dynamic flow of streams have been described by several authors

992

Fig. 2 A photograph of a typical fluvial sample station and the 3 habitat types (a). Diagram of the applied sampling method indicating habitats A, B and C within a quadrat of 1000 m2 for each sample station, based on the INAG sampling protocol (http://dqa.inag.pt/dqa2002/port/docs_apoio/doc_nac/Manuais/

Macrofitos.pdf) (b). Three different habitat were proposed for the riparian ecosystems: habitat A is the wetted channel, habitat B is the bankfull width with vegetation in dry lateral sections of the riverbed and habitat C comprises the bankside gallery

(Corenblit et al. 2007; Gurnell et al. 2012; North and Davidson 2012). The existence and extent of habitats A and B are directly correlated with weather regimes. Habitat B is supported by lotic sediment transport and patterns of erosion and deposition (Coelho et al. 2009; Garel et al. 2009), while thermic and pluviometric variation will cause change to flow levels for habitat A (Trigo et al. 2004). Species presence detected per habitat was compiled into a binary data floristic matrix 1 (presence)/0 (absence) floristic matrix, per type of habitat and sample station (as proposed by Crespı´ et al. 2001, 2005). No abundance or cover values (number of individuals or covering space per species) were used, because of the difficulty in applying these values across extensive sample stations.

measured objects. City Block Standardized Manhattan distances are appropriate for this amalgamation tool to correct the Euclidean error exposed by the shortest similarity distance between sample stations (Cha 2008). Groups were identied by discriminant function analysis based upon the similarity results. discriminant canonical analysis (DCA) was used to detect underlying environmental trends in data sets. This multivariate approach identifies the best descriptors for characterizing the frequency numerical matrix (Niemi and McDonald 2004). We used a forward stepwise DCA to select the most strongly discriminating environmental variables using the F statistic (F-remove) and P values to describe the distribution of the variable. Wilk´s Lambda was applied to explain variance between variables (1 minus the squared canonical correlation) and tolerance (1 minus the squared multiple correlation -1-Toler. R-Sqr.-). The most discriminant group, based on these statistical tests, was selected to describe environmental groups. All statistical analyses were carried out using STATISTICA 9.0.

Environmental characterization A CORINE land cover shape file was used to characterize the land uses surrounding the sample stations. Land use data (CORINE Land Cover; CLC) was obtained from Direc¸a˜o Geral do Territo´rio (2006). This socioeconomic information supported the environmental characterization obtained by Worldclim Global Climate Data (http://www.worldclim.org/), proposed by Hijmans et al. (2005). The Worldclim Global Climate database was used to obtain elevation, bioclimatic, thermal and precipitation data for each sample station (Table 1). A hierarchical cluster analysis was calculated to identify groups of sample stations, based on the environmental data, using Unweighted Pair Group Method with Arithmetic Mean, UPGMA and City Block Standardized Manhattan distances (STATISTICA version 9.0, Statsoft Ltd., http://www.statsoft.com). The UPGMA amalgamation method is highly efficient for evaluating the distances of clusters (Farris 1969) based on the average distance between all pairs of

Floristic characterization The binary presence-absence data of the numerical matrices obtained per habitat type were transformed into a continuous matrix, by replacing the presence (1) value of each species with its respective average value per habitat (Crespı´ et al. 2001, 2005, 2013), which describes the frequency of each species in habitat A, B or C. The higher the average presence, the higher frequency of that species at a given sample station. A frequent species can have an average presence value close to 1 (100 %) for one habitat type, but can be rarer in another habitat type and thus be attributed a value closer to 0. A UPGMACity Block Standardized Manhattan distances cluster analysis was carried out on a similarity matrix of sample

993 Table 1 Environmental variables used by Worldclim (http:// worldclim.org/) and applied for the environmental characterization of sample stations Name

Description

Alt Bio1 Bio2

Elevation Annual mean temperature Mean diurnal range [mean of monthly (max temp– min temp)] Isothermality (Bio2/Bio7) (·100) Temperature seasonality (standard deviation · 100) Max temperature of warmest month Min temperature of coldest month Temperature annual range (Bio5–Bio6) Mean temperature of wettest quarter Mean temperature of driest quarter Mean temperature of warmest quarter Mean temperature of coldest quarter Annual precipitation Precipitation of wettest month Precipitation of driest month Precipitation seasonality (coefficient of variation) Precipitation of wettest quarter Precipitation of driest quarter Precipitation of warmest quarter Precipitation of coldest quarter Precipitation of month 1 Precipitation of month 2 Precipitation of month 3 Precipitation of month 4 Precipitation of month 5 Precipitation of month 6 Precipitation of month 7 Precipitation of month 8 Precipitation of month 9 Precipitation of month 10 Precipitation of month 11 Precipitation of month 12 Max temperature of month 1 Max temperature of month 2 Max temperature of month 3 Max temperature of month 4 Max temperature of month 5 Max temperature of month 6 Max temperature of month 7 Max temperature of month 8 Max temperature of month 9 Max temperature of month 10 Max temperature of month 11 Max temperature of month 12 Min temperature of month 1 Min temperature of month 2 Min temperature of month 3 Min temperature of month 4 Min temperature of month 5 Min temperature of month 6 Min temperature of month 7 Min temperature of month 8 Min temperature of month 9 Min temperature of month 10 Min temperature of month 11 Min temperature of month 12

Bio3 Bio4 Bio5 Bio6 Bio7 Bio8 Bio9 Bio10 Bio11 Bio12 Bio13 Bio14 Bio15 Bio16 Bio17 Bio18 Bio19 Prec1 Prec2 Prec3 Prec4 Prec5 Prec6 Prec7 Prec8 Prec9 Prec10 Prec11 Prec12 Tmax1 Tmax2 Tmax3 Tmax4 Tmax5 Tmax6 Tmax7 Tmax8 Tmax9 Tmax10 Tmax11 Tmax12 Tmin1 Tmin2 Tmin3 Tmin4 Tmin5 Tmin6 Tmin7 Tmin8 Tmin9 Tmin10 Tmin11 Tmin12

stations to group species according to the following descriptors: life forms, systematic, and biogeographic classes. Thus, each habitat was described by groups of life forms frequencies, systematic frequencies, and biogeographical frequencies. Biogeographical and systematic information (botanic families, in this case) were also

used to characterize the species functionality. Life forms classification was carried out using Raunkier´ s typification, adapted from Braun-Blanquet (1979). A total of nine general biogeographic distribution classes were derived and applied to the data across the broadest to the narrowest distribution areas. The classes were cosmopolitan, sub cosmopolitan, Euroasiatic, occidental Euroasiatic, Mediterranean, Atlantic, sub-endemic (west of the Mediterranean basin), endemic (exclusive to the Iberian Peninsula and north-western Morocco), and allochthonous (Occhipinti-Ambrogi 2004). The systematic approach employed subclasses belonging to the APGIII classification (The Angiosperm Phylogeny Group 2009). Frequency curves were drawn up between the average presence of the species (species frequency) and the distribution of average presences. The average presence of each species/number of repetitions of that average presence was calculated for each environmental group and habitat type (A, B and C), both obtained by similarity analyses. The distribution of frequencies of the life forms, biogeographic and systematic behaviors was analyzed to describe the functional diversity per environmental group and habitat. Quadratic curves can usefully describe the population distribution along environmental variations (Agras and Chapman 1999; Fryxell et al. 2005; Cinner et al. 2009). In the present study the environmental gradient was replaced by species frequency to describe the structural organization of individuals per environmental group or habitat type. These curves represented the trend between frequency values and number of species with those frequency values. The extension and concavity of the curve was used to explain the relationship (the more concavity the higher the number of low-frequency species). This approach allowed the establishment of frequency distributions, from the lowest to the highest frequencies. Each frequencies under 25 % of presences was classified as low-frequencies species. Potential distribution patterns under climate change scenarios Climate predictors were derived from a general circulation model (CCCMA: CGCM2) for the year 2080 under IPCC emission scenarios (SRES) A2a and B2a, to predict potential future habitat distribution patterns (http:// gisweb.ciat.cgiar.org/GCMPage; Ramirez and Jarvis 2008). The B2a scenario has a lower rate of global warming and less intense changes in temperature and precipitation than the A2a scenario (http://forest. jrc.ec.europa.eu/climate-change/future-trends). Maxent software (version 3.3) for species habitat modelling (Baldwin 2009) was used to estimate the probability of potential future suitable habitat occurrence at the sample stations with values varying from 0 (lowest probability) to 1 (highest probability) (Phillips et al. 2006). Results from this analysis described the potential distribution patterns under future climate

994

change scenarios, as well variation in temperature and precipitation values for the sample stations. Maxent requires only species presence data and environmental variables in GIS shapefiles for the study area. Receiveroperating characteristic (ROC) plots (Fielding and Bell 1997) and area under curve (AUC) methodologies (Phillips et al. 2006) were used to validate model accuracy and determine the probability that a location, indicating the presence of a sample station, was ranked higher than the random background probability. Locations with a random background probability served as pseudo-absences for all analysis in Maxent (Phillips et al. 2004, 2006). The following criteria were applied to all models: 10 repetitions with cross-validation, standard regularization multiplier (affects output distribution fit) and 500 iterations. For further details on these parameters see Phillips (2010). The obtained output (in ASCII format) were introduced into ArcGIS software version 9.2 (ESRI, Redlands, California, USA, http://www.esri.com) as floating-point grids (Peterson et al. 2007) and the occurrence probability of the species at each site was mapped.

Results Environmental characterization Results of the UPGMA cluster analysis of the similarity matrix of environmental data (Fig. 3a) and DCA revealed three distinct groups of sample stations (Fig. 3b), separated primarily along an altitudinal gradient (Table 2). Geographic distribution patterns (Fig. 3c) revealed a western mountain group (Group 1, n = 20 sample stations), a western group (Group 2, n = 30 sample stations) and a eastern group (Group 3, n = 50 sample stations). Characterization of land cover at each sample station, organized by environmental group (Table 3), showed that Group 1 sites were low in land use diversity compared with Groups 2 and 3, which had very heterogeneous landscapes. The DCA summary statistics (Table 2) revealed the distinct environmental character of the three geographic groups and indicated that elevation and highest average temperature for May were the most strongly discriminant parameters between groups. Group 1 sample stations had the highest average elevation value (alt, 632–906 m), the lowest average maximum temperature in May (Tmax5, 16.3–17.7 C), highest annual temperature range (Bio17, 9.8–12.2 C), and lowest isothermal values (Bio3, 3.8–4.0 C). In contrast, Group 2 stations had both the lowest average elevation (3–373 m) and Tmax5 (18.3–21.4 C), and the highest Bio3 values (4.0–4.5 C). Group 3 sample stations were intermediate in character with the exception of annual temperature range, where the lowest value was recorded (alt 81–845 m, Tmax5 17.4–22.8 C, and Bio3 3.5–4.1 C).

Fig. 3 An Unweighted Pair Group Method with Arithmetic Mean (UPGMA) cluster analysis dendrogram based on environmental data for sample station. Group 1 comprises western mountain stations, Group 2 comprises western littoral sites and Group 3 comprises inland eastern stations (a). A discriminant canonical analysis (DCA) ordination plot based on the environmental data from the three groups of sample stations (b). Distribution of sample stations by UPGMA derived environmental groups across the study area (c)

Summary statistics of the DCA discriminant variables are given in Table 4. Environmental Group 2 (western) and Group 3 (eastern) stations both had higher numbers of species with lower frequency values (Fig. 4b, c) compared with Group 1 sites (western mountains) (Fig. 4a). The relationship curve obtained for Group 1 sites was lower and longer compared with those for Group 2 sites (Fig. 4b) and Group 3 sites (Fig. 4c), indicating a greater probability of finding the same species among Group 1 sample stations (Table S3). The high number of low-frequency species found in Group 2 and Group 3 sites resulted in higher curves that were skewed to the left (a lower

995 Table 2 Numerical DCA (discriminant canonical analysis) values for selected environmental variables (Toler. = tolerance by coefficient of determination R2 or R-Sqr) Wilks’ Alt Tmax5 Bio17 Bio3

0.1836 0.1225 0.1028 0.0903

Lambda

F

P value

1-Toler. (R-Sqr.)

53.959 20.378 9.535 2.669

<0.0001 <0.0001 0.0002 0.0745

0.7184 0.7599 0.3833 0.2283

Elevation of sample stations (alt, F = 53.958 and P value <0.0001), average maximum temperature during May (Tmax5, F = 20.378 and P value <0.0001) are the most strongly discriminant variables (alt = elevation, Tmax5 = maximum average temperature for May; Bio17 = precipitation of driest quarter; Bio3 = annual isothermality) according to the environmental variables used in Worldclim, http://www.worldclim.org/formats) Table 3 Land uses per environmental group obtained from CORINE land cover shape (http://sniamb.apambiente.pt/clc/frm/), based on the land cover around each sample station (values in km2)

Urban Arable land Permanent crops Heterogeneous agricultural areas Forests Scrub and/or herbaceous vegetation associations Open spaces with little or no vegetation Inland waters

Group 1

Group 2

Group 3

0 0 0 0 0 80

5.1 10.3 0.0 53.8 15.4 7.7

1.8 7.1 12.5 41.1 14.3 23.2

20

2.6

0.0

0

5.1

0.0

Environmental Groups 2 and 3 show the highest land uses variability Table 4 Average values, amplitudes and standard deviation (SD) for the principal selected discriminant variables for the environmental characterization (alt = elevation, Tmax5 = highest average temperature for May, Bio17 = precipitation of driest quarter, Bio3 = annual isothermality) Alt

Tmax5 Bio17 Bio3

Group 1 Average 808.0 16.9 Western Mediterranean Maximum 906.0 17.7 Minimum 632.0 16.3 SD ±121.8 5.2 Group 2 Average 91.5 20.0 Atlantic Maximum 373.0 21.4 Minimum 3.0 18.3 SD ±86.4 6.7 Group 3 Average 477.6 20.2 Continental Maximum 845.0 22.8 Minimum 81.0 17.4 SD ±179.7 12.8

11.0 12.2 9.8 11.1 8.4 11.3 6.9 9.5 6.7 9.8 4.7 11.9

3.9 4.0 3.8 0.9 4.2 4.5 4.0 1.3 3.9 4.1 3.5 1.3

probability of finding the same species among sample stations for both environmental groups). Floristic characterization Four hundred and sixty-four taxa (Table S2) were detected across the three habitat types, with a noticeable

increase in species richness occurring from the stream bed community to the bankside gallery community (77 taxa for habitat A at 87 sample stations, 179 taxa for habitat B at 40 sample stations and 424 taxa for habitat C at 100 sample stations). The lower number of species in habitats A and B was probably due to fluctuations in water level (for habitat B). Floristic differences between these habitats were also confirmed by low frequencies of shared common species. Just 9 % of taxa detected for all the sample stations were observed along the three habitats. Common species between habitats A and B made up 24 % of the taxonomic diversity detected for both habitats, 13 % between A and C, and 33 % between B and C (Table S3). The average number of species per environmental group revealed broad differences between habitats A and B (very low average number of species per sample station), compared with habitat C (Table 5). On the other hand, although there were only slight difference in average number of species between environmental groups, Groups 2 and 3 were clearly more diverse than Group 1 in terms of the diversity of systematic groups (botanic families), life forms and biogeographic classes. Based on the functional redundancy hypothesis, these results indicated higher redundancy for the Group 1 resulting in floristic-structure resistance, possibly because of environmental limitations. The DCA ordination for the floristic matrix grouped by habitat type (Fig. 5a) revealed distinct differences between habitats A and C. The floristic information for habitat B obtained from the sample stations was intermediate relative to habitats A and C. Systematic floristic variables, in this case the frequency of Commelinidae and Helophytes, were the most strongly discriminant floristic variables (Table 6). This indicates a higher concentration of low-frequency species in the herbaceous layer, since all Commelinidae and Helophytes species are herbaceous. Biogeographical variables were not significantly discriminant with the exception of distribution of allochthonous taxa which were more frequent in habitats C and B. The greater dispersal of habitat C sites in the ordination space indicates a greater range of floristic frequencies in comparison with habitats A and B. This was also reflected in their respective frequency curves (Fig. 5b, d). In these frequency curve distributions the highest values for low-frequency species occurred in habitats B and C, in contrast to habitat A, where there were marked divergences between lowand high-frequency species. Table 5 lists the number of species, families, life forms, and biogeographic classes per habitat as well as their frequencies. A common pattern was observed across all habitats, namely over 70 % of species were present at less than 30 % of sample stations, and all the functionalities observed for each habitat occurred in these lower frequencies. These results clearly highlighted the focus of habitat ecological functionality in the lowest frequencies. Habitat C had a more even distribution of frequencies, but with a higher number of low-frequency species. These results agree

996

change scenarios A2a and B2a, based on the most strongly discriminant environmental variables (elevation and highest average temperature for May). The potential area for Group 1 sites was restricted to areas of higher elevation in the west of the region (Fig. 6a). The potential area for Group 2 sites was limited to lower elevation areas (Fig. 6d) also in the west of the study area, while the potential areas for Group 3 sites extended along the central lowlands and the mountains east of the study area at higher elevations (Fig. 6g). The future scenario outputs highlighted the better resistance of Group 2 sites to climate change, maintaining a significant presence in 2080 under the harsher A2a climate scenario (Fig. 6e) albeit with a more restricted distribution pattern compared with contemporary distribution patterns. Group 3 stations disappeared altogether under the A2a climate scenario (Fig. 6h) along with most of the Group 1 stations (Fig. 6b). Future distribution patterns under the more benign B2a climate scenario for 2080 increase the potential distribution for Group 2 stations (Fig. 6f), while there is no apparent alteration in the distribution of Group 3 stations (Fig. 6i). The environmental conditions for Group 1 stations will no longer exist under this future climate change scenario (Fig. 6c). Analysis of the environmental variables (with the exception of elevation, which remained constant along the time interval) revealed significant variations in maximum temperature for May (Tmax5), and precipitation of driest quarter (Bio17), especially harder for Groups 1 and 2 (Fig. 7). These future climate change scenarios indicate a reduction in annual precipitations and increases of annual temperature in spring in these western regions.

Discussion

Fig. 4 Frequency curves between number of species (Nº spp) and their frequencies (average presence for all sample stations) for environmental Group 1—western mountain sites (a), Group 2—western sites (b), and Group 3—eastern sites (c). The number of species with the same average presence is described along these curves

with the curves of frequency per environmental group and habitat. Potential distribution patterns under climate change scenarios The results of Maxent species habitat modelling predicted the potential areas for the three environmentally distinct groups of sites in the year 2080 under climate

Analysis of environmental data and floristic patterns of fluvial vegetation communities, collected across northern Portugal, clearly identified three distinct groups of sample stations. The macrogeomorphology of mountain chains along northern Portugal results in a narrow bioclimatic strip running along the Atlantic coast to the west, separated from the central and eastern Mediterranean areas by an extensive mountain system (Martins et al. 2007). The environmental groups identified in this study coincide with these latter bioclimatic zones. Group 1 stations occurred along the western mountain system, while Group 2 stations were distributed west of the mountain system along the littoral Atlantic strip, and Group 3 stations were distributed along the Mediterranean area. The floristic behavior of these three areas is distinct, as confirmed by other authors (Rozeira 1944; Costa et al. 1998). However, seasonal lotic flow dynamics are an important influence on the environmental variability of the studied fluvial ecosystems under study (Trigo et al. 2004; Coelho et al. 2009; Garel et al.

997 Table 5 Number of species per habitat, and environmental group per interval of presence classified in eleven classes (between 0–9 % and 100 % of presence per sample station) 0–9 % 10–19 % 20–29 % 30–39 % 40–49 % 50–59 % 60–69 % 70–79 % 80–89 % 90–99 % 100 % Habitat A Group 1 N = 4 Spp Fam L.F. Biog Group 2 N = 2 Spp Fam L.F. Biog Group 3 N = 4 Spp Fam L.F. Biog Habitat B Group 1 N = 3 Spp Fam L.F. Biog Group 2 N = 4 Spp Fam L.F. Biog Group 3 N = 6 Spp Fam L.F. Biog Habitat C Group 1 N = 30 Spp Fam L.F. Biog Group 2 N = 31 Spp Fam L.F. Biog Group 3 N = 30 Spp Fam L.F. Biog

0 0 0 0 0 0 0 0 56 29 9 9 0 0 0 0 61 29 10 6 142 29 10 8 0 0 0 0 141 39 7 5 251 50 10 6

8 8 4 6 25 20 8 7 5 5 5 4 11 9 5 4 6 4 1 3 8 5 6 1 0 0 0 0 43 19 4 2 51 8 1 2

0 0 0 0 7 7 6 3 0 0 0 0 0 0 0 0 3 3 1 1 0 0 0 0 56 25 4 0 22 7 1 2 30 7 1 2

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 0 0 0 0 15 5 1 2 5 1 1 1

1 1 1 1 2 2 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 23 12 4 3 9 4 1 2 5 1 1 1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 2 1 1 2 2 1 1

0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 6 4 1 2 2 2 1 1 1 1 1 1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1

1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 3 2 1 1 1 1 1 1 1 1 1

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0

The average number of species per environmental group (N) is also specified (Spp. = species; Fam. = botanic families; L.F. = life forms; Biog = biogeographic classes)

2009). Altitudinal, thermic and pluviometric factors can be classified as stress processes (sensu Rykiel 1985) that regulate both the existence and persistence of habitats and their floristic structure. The development of the floristic structure of the habitats we analyzed influenced species frequency behavior in each habitat. Development of habitats A (where persistence depends on the annual precipitation variation) and B (which emerges when the flow levels decrease) within the river channel is restricted by strong environmental filters related to spatiotemporal variations in flow (Pearson et al. 1992; Hupp and Osterkamp 1996; Dixon 2003; Hughes et al. 2009). Habitat C, the riparian bankside vegetation, is clearly more temporally stable compared with habitat A and B. The extent and persistence of habitats A and B depends directly on annual precipitation patterns, which are a major driver of stream flow levels (Trigo et al. 2004). The seasonal decrease in flow from the end of spring until the autumn

is a major environmental filter controlling the development of vegetation in these habitats. Flow persistence and channel depth for habitat A (Dixon 2003), and the short colonization period for habitat B are critical environmental factors for the existence of flora. A prolonged decrease in flow results in a longer temporal existence of habitat B as well as greater taxonomic richness and structural complexity, when compared with habitat A. These environmental filters result in lower taxonomic diversity for habitats A and B, compared with habitat C (77 and 179 species for habitat A and B, respectively, in contrast with 424 specific taxa for habitat C). These environmental factors also influence the floristic structure of the three habitats described in this study namely the distribution of species frequencies. In contrast with habitat A, most species found in habitat B had lower frequencies and there was a marked divergence between the lowest and highest frequencies. Habitat C species were mostly concentrated at the lowest

998 Fig. 5 DCA plots of floristicstructural combinations per sample station and type of habitat (a), and frequency curves between number of species and respective average presence per habitat: for habitat A—wetted channel (b), habitat B—bankfull width (c), and habitat C—bankside gallery (d)

Table 6 Numerical DCA (discriminant canonical analysis) values for selected floristic variables (Toler. = tolerance by coefficient of determination R2 or R-Sqr) Wilks’ Commelinidae Helophyte Hamamelididae Allochthonous Atlantic Hidrophyte Chloranthidae

Lambda

0.4177 0.4015 0.3506 0.3375 0.3278 0.3171 0.3132

F

P value

1-Toler. (R-Sqr.)

50.1492 42.5529 18.7472 12.5836 8.0648 3.0543 1.2276

<0.0001 <0.0001 <0.0001 <0.0001 0.0004 0.0487 0.2945

0.5712 0.5457 0.0617 0.0579 0.1639 0.0173 0.0021

The frequency of Commelinidae (F = 50.1492 and P value < 0.0001) and the frequency of Helophytes (F = 42.5529 and P value <0.0001) are the most strongly discriminant variables

frequencies, however the distribution of the higher-frequency species tended to be continuous (i.e. three species were present at more than 40 % of the sample stations, and eight species were present between 30 % and 40 %). According to the facilitation theory (Bruno et al. 2003), the existence of dominant species facilitates exclusion of competitors and a greater presence of low-frequency species. This was especially evident for habitat A, where there was a notable lack of continuity in species frequency, illustrated by the presence of two high frequency species while over 90 % of remaining species had very low frequencies. Lower levels of environmental stress in habitat C facilitate greater species diversity, as well as a continuous distribution of frequencies. Higher frequencies (more than 25 % of recorded presences) were clearly higher than those in the other two habitats. Flow related

environmental limitations could explain the concentration of habitat B species frequencies at lower values (less than 25 % of presences). For example, warmer and drier periods of the year, resulting in lower flow levels, will allow greater development of vegetation in habitat B. Under such environmental conditions, no dominant species will be established, i.e. these environmental restrictions will force species to have low frequency values. The presence and extent of the three habitat types at a given sample station is the principal driver for increased floristic-structural variability, shaped by patterns of environmental variability. Distinct floristic-structural combinations were observed for each habitat. However, continuity between these habitats was not observed. The distinct ecological character of each habitat could be the

999

Fig. 6 Actual and predicted distributions under the A2a and B2a for 2080 climate change scenarios based on the most discriminant variables (environmental variables of elevation and highest average temperature for May) per environmental Groups 1, 2, and 3: a actual potential distribution for Group 1; b predictive distribu-

tion under A2a scenario; and c under B2a scenario; d actual potential distribution for Group 2; e predictive distribution under A2a scenario; and f under B2a scenario; g actual potential distribution for Group 3; h predictive distribution under A2a scenario; and i under B2a scenario

main reason for this low floristic-structural continuity. The percentage of common species between habitats was remarkable low, and functional diversity was also different (i.e. lower in habitat A compared with habitat C), indicating significant functional differences. These results can be explained by referring to Elton´s theory (Naeem et al. 2000) and Yachi and Loreau´s insurance hypothesis (Allison 2004). According to Elton´s theory, wild vegetation is more resistant to environmental change (originally described as allochthonous invasions). Thus, environmental conditions will be the

principal driver of floristic-structural variability. Describing differences in terms of resistance and resilience responses, the Yachi and Loreau insurance hypothesis proposes that there is no direct correlation between resistance, resilience and taxonomic diversity which support the results of this study. Resistance under restricted environmental conditions, such as habitat A, did not allow diverse floristic-structural combinations to develop, reflected in the low taxonomic diversity and the marked differences in species frequencies. Conversely, the wider range of environmental conditions in habitat C

1000

Fig. 7 Graphic representation of the most discriminant environmental variables per Group (1, 2, and 3), from the current (Tmax5, bio17, and bio3) to the 2080 conditions, for the B2a (Tmax5 B2a, bio17 B2a, and bio3 B2a) and A2a (Tmax5 A2a, bio17 A2a, and bio3 A2a) scenarios. An average increase of five Celsius degrees is verified for each environmental group, in this period

resulted in higher taxonomic diversity and floristicstructural combinations illustrated by higher levels of species diversity and continuous species frequencies. Habitat B exemplifies a resilient plant community. When the environmental stress decreases (i.e. when flow levels decrease) habitat functionality will be based on larger number of species than habitat A. Sharp fall in flow over a short period of time will prevent high species frequencies from developing, and the functionality of habitat B will be supported by low-frequency taxa. The lack of correlation between taxonomic diversity, resistance and resilience has been also suggested for other ecological systems (Downing and Leibold 2010; Van Ruijven and Berendse 2010). We used this environmental approach to predict and describe contemporary and future vegetation trends under different climate change scenarios over an 80-year period. Environmental Group 1 sample stations had the highest level of functional redundancy, where fewer families, life forms and biogeographic classes were represented. The role of low-frequency species under future climatic scenarios is valuable in explaining the structural-floristic diversity and functionality for describing scenarios. Information on the predicted distribution of fluvial stands at the selected sample stations within the three environmental groups under different future climate change scenarios provides important information for natural resource managers, allowing the implementation of successful mitigation measures such as restoration projects to sustain essential ecosystem services linked with aquatic ecosystems. It also highlights the important role of environmental heterogeneity in maintaining habitat persistence. Group 2 and 3 sample stations were more environmentally heterogeneous and had higher levels of low-frequency species. Under the harsher A2a climate scenario, the extinction of the all three environmental groups was practically certain; only

Group 2 sample stations showed some persistence, albeit over a much reduced area. The potential distribution for environmental Group 2 increased under the less harsh B2a climate change scenario, probably as a response to the extinction of the environmental Group 1 sites, while the extent of the Group 3 area remained practically the same. The concentration of precipitation events and the warmer annual average temperatures under both future climate change scenarios resulted in a decrease in environmental variability for Groups 1 sites, along with the essential services they provide. The most realistic scenario (A2a) reduces environmental moisture as a result of increasing average temperature and the concentration of precipitation. This type of environmental stress will directly influence low-frequency species, as observed for habitat B vegetation. Thus, fluvial floristic-structural variability will be sustained by the presence of low-frequency species. The results of this study provide greater insight into both the characterization and dynamics of fluvial plant communities, and how they will be affected by climate change under distinct climate scenarios across northern Portugal. Without appropriate, typologically relevant, catchment level forward planning such as the WFD programs of measures to mitigate or adapt to such changes, the fundamental resistance and resilience of these ecosystems and the many ecological services they provide will be lost (Carpenter et al. 1992; Meyer et al. 1999; Heller and Zavaleta 2009). These results agree with those obtained by Ferna´ndez Gonza´lez et al. (2005) for the vegetation of the Iberian Peninsula, and also with other more general approaches proposed by other authors (Thuiller et al. 2005; Parmesan 2006; Khavhagali 2011). Important changes in thermal and precipitation regimes will drive change in the resistance and resilience of fluvial systems (Seavy et al. 2009). Climatic changes are essentially disturbance processes (White 1979; Rykiel 1985; Pickett et al. 1989), and the resistance and resilience of these vegetation communities will be directly influenced by their structural organization and their capacity to maintain ecological functionality (Gunderson 2000; Lake 2000). Our results show that fluvial vegetal community resistance and resilience to climate change effects will depend on a combination of taxonomic diversity and low-frequency species, where most functional diversity was focused. Higher levels of environmental variability result in higher taxonomic diversity and floristic-structural variability, which will clearly decrease as a result of environmental limitations. We found non-linear effects on resistance and resilience of vegetation, appearing to support the insurance theory of Yachi and Loreau. At the same time, the information produced using the approach adopted in this study should underpin planning for mitigation or adaptation measures such as those listed within WFD River Basin Management Plans. This approach constitutes a tool for developing and using functional indicators of provision of ecosystem services (Hering et al. 2010).

1001

In summary, the diversity and functional ecosystem characterization approach applied in this study has potential for management applications (Poff 1997), particularly if it is coupled with GIS platforms and models for aquatic resource management decision support systems. Outputs would contribute to the development and implementation of sustainable adaptation, mitigation and restoration measures linked to large scale impacts such as climate change and changes in land use (Benda et al. 2004, 2007; Ferna´ndez et al. 2012; Hughes et al. 2012). Such a tool would need to include other sources and types of data such as hydrological data and characteristics of the catchment ranging factors such as slope to land use in order to produce typologically and regionally relevant results. Acknowledgments This work was carried out as part of the ‘‘Assessment of Ecological Sustainability of Fluvial Habitats of Protected or Classified Areas in the Minho/Lima Hydrographical Region V01/2010’’ contract 2010–2011 between the Northern River Basin District Administrative Body and the University of Tra´s-osMontes e Alto Douro. The authors would like to thank FCTFoundation of Science and Technology Science for Joa˜o Rocha´s doctoral grant SFRH/BD/43167/2008. Samantha Jane Hughes is SUSTAINSYS funded post-doctoral fellow—North-07-0124FEDER-0000044, financed by the Regional Operational Program North (ON. 2—The New North), under the National Strategic Framework (NSRF), through the European Regional Development Fund and PIDDAC via the Foundation for Science and Technology. This work was also supported by European Union Funds (FEDER/COMPETE—Operational Competitiveness Programme) and by national funds (FCT—Portuguese Foundation for Science and Technology) under the project FCOMP-01-0124FEDER-022692.

References Agras J, Chapman D (1999) A dynamic approach to the environmental kuznets curve hypothesis. Ecol Econ 28:267–277 Aguiar FC, Ferreira MT (2005) Human-disturbed landscapes: effects on composition and integrity of riparian woody vegetation in the Tagus River basin, Portugal. Environ Conserv 32:1–12 Aguiar FC, Ferreira MT (2006) Riparian and aquatic vegetation in Mediterranean-type streams (western Iberia). Limnetica 25:411–424 Aguiar FC, Ferreira MT, Moreira IS, Albuquerque A, Boatman ND, Clay DV, Pywell RF (2000) Riparian types on a Mediterranean basin. Asp Appl Biol 58:221–232 Aguiar FC, Ferreira MT, Albuquerque A, Moreira I (2007) Alien and endemic flora at reference and non-reference sites in Mediterranean-type streams in Portugal. Aquat Conserv Marine Freshwater Ecosyst 17:335–347 Aguiar FC, Ferreira MT, Albuquerque A, Rodrı´ guez-Gonza´lez P, Segurado P (2009) Structural and functional responses of riparian vegetation to human disturbance: performance and spatial scale-dependence. Fundam Appl Limnol 175:249–267 Allen CR, Gunderson L, Johnson AR (2005) The use of discontinuities and functional groups to assess relative resilience in complex systems. Ecosystems 8:958–966 Allison G (2004) The influence of species diversity and stress intensity on community resistance and resilience. Ecol Monogr 74:117–134 Allison SD, Martiny JB (2008) Resistance, resilience, and redundancy in microbial communities. Proc Natl Acad Sci 105(Supplement 1):11512–11519

Andre´s FN, Zapatero MG, Martı´ n FG, Rosello´ JAE, Anta MA´S, Blanco LL (1986) Alisedas salmantinas y zamoranas. Studia botanica 5:39–52 [in Spanish] Baldwin RA (2009) Use of maximum entropy modeling in wildlife research. Entropy 11:854–866 Benda L, Poff NL, Miller D, Dunne T, Reeves G, Pess G, Pollock M (2004) The network dynamics hypothesis: how channel networks structure riverine habitats. Bioscience 54:413–427 Benda L, Miller D, Andras K, Bigelow P, Reeves G, Michael D (2007) NetMap: a new tool in support of watershed science and resource management. Forest Science 53:206–219 Bernez I, Ferreira T (2007) River macrophytes in regulated Mediterranean-type rivers of southern Portugal. Belg J Bot 140:136–139 Bombino G, Gurnell AM, Tamburino V, Zema DA, Zimbone SM (2008) Sediment size variation in torrents with check dams: effects on riparian vegetation. Ecol Eng 32:166–177 Braun-Blanquet J (1979) Fitosociologı´ a. Bases para el estudio de las comunidades vegetales. Blume Madrid (in Spanish) Bruno JF, Stachowicz JJ, Bertness MD (2003) Inclusion of facilitation into ecological theory. Trends Ecol Evol 18:119–125 Buiatti M, Longo G (2013) Randomness and multilevel interactions in biology. Theory Biosci 132:139–158 Carpenter SR, Fisher SG, Grimm NB, Kitchell JF (1992) Global change and freshwater ecosystems. Annu Rev Ecol Syst 23:119–139 Cha SH (2008) Taxonomy of nominal type histogram distance measures. City 1:325–330 Chapin FS III, Zavaleta ES, Eviner VT, Naylor RL, Vitousek PM, Reynolds HL, Hooper DU, Lavorel S, Sala OE, EHSE Sarah, Mack MC, Dı´ az S (2000) Consequences of changing biodiversity. Nature 405:234–242 Cinner JE, McClanahan TR, Daw TM, Graham NA, Maina J, Wilson SK, Hughes TP (2009) Linking social and ecological systems to sustain coral reef fisheries. Curr Biol 19:206–212 Coelho C, Silva R, Veloso-Gomes F, Taveira-Pinto F (2009) Potential effects of climate change on northwest Portuguese coastal zones. ICES J Marine Sci J Du Conseil 66:1497–1507 Commission European (2000) Directive 2000/60/EC of the European Parliament: establishing a framework for Community action in the field of water policy. Off J Eur Commun L327:1–72 Commission European (2009) Common implementation strategy for the water framework directive (2000/60/EC). River Basin Manag Chang Climate 24:141 Corenblit D, Tabacchi E, Steiger J, Gurnell AM (2007) Reciprocal interactions and adjustments between fluvial landforms and vegetation dynamics in river corridors: a review of complementary approaches. Earth Sci Rev 84:56–86 Cortes R, Hughes SJ, Pereira VR, Varandas S (2013) Tools for bioindicator assessment in rivers: the importance of spatial scale, land use patterns and biotic integration. Ecol Ind 34:460–477 Costa JC, Aguiar C, Capelo JH, Lousa˜ M, Neto C (1998) Biogeografia de Portugal continental. Quercetea 0:5–56 [in Portuguese] Crespı´ AL, Pereira A, Fernandes CP, Cortes R, Oliveira S, Ribeiro JA (2001) Descripcio´n de la estrutura vegetal de comunidades riparias del Noroeste de Portugal. Boletı´ n de la Real Sociedad Espan˜ola de Historia Natural Sec Biol. 96:69–80 (in Spanish) Crespı´ AL, Bernardos S, Castro A, Fernandes CP, Amich F (2005) Phytostructural characterization of several vegetation types in northern Portugal. II. The structural expressivity and the resistance of the vegetation. Plant Biosyst 139:387–398 Crespı´ AL, Ferreira M, Fonseca TF, Pacheco Marques C (2013) Structural and floristic behaviors in East Timor forest vegetation. Ecol Res 28:1081–1090 Direc¸a˜o Geral do Territo´rio (2006) Carta Corine Land Cover. http://www.dgterritorio.pt/cartografia_e_geodesia/mapas_on_ line/. Accessed 29 Jul 2015

1002 Dixon M (2003) Effects of flow pattern on riparian seedling recruitment on sandbars in the Wisconsin River, Wisconsin, USA. Wetlands 22:125–139 Downing AL, Leibold MA (2010) Species richness facilitates ecosystem resilience in aquatic food webs. Freshw Biol 55:2123–2137 Duarte P, Azevedo B, Ribeiro C, Pereira A, Falca˜o M, Serpa D, Bandeira R, Reia J (2007) Management oriented mathematical modelling of Ria Formosa (South Portugal). Transit Waters Monogr 1:13–51 European Commission (2015). Introduction to the new EU Water Framework Directive. http://ec.europa.eu/environment/water/ water-framework/info/intro_en.htm. Accessed 29 Jul 2015 Farris JS (1969) On cophenetic correlation coefficients. Syst Zool 18:279–285 Feio MJ, Reynoldson TB, Ferreira V, Grac¸a MAS (2007) A predictive model for freshwater bioassessment (Mondego river, Portugal). Hydrobiologia 589:55–68 Feio MJ, Aguiar FC, Almeida SFP, Ferreira J, Ferreira MT, Elias C, Serra SRS, Buffagni A, Cambra J, Chauvin C, Delmas F, Do¨rflinger G, Erba S, Flor N, Ferre´ol M, Germ M, Mancini L, Manolaki P, Marcheggiani S, Minciardi MR, Munne´ A, Papastergiadou E, Prat N, Puccinelli C, Rosebery J, Sabater S, Ciadamidaro S, Torne´s E, Tziorztis I, Urbanicˇ G, Vieira C (2014) Least disturbed condition for European Mediterranean rivers. Sci Total Environ 476:745–756 Ferna´ndez D, Barquı´ n J, I´varez-Cabria MA, Pen˜as FJ (2012) Delineating riparian zones for entire river networks using geomorphological criteria. Hydrol Earth Syst Sci Discus 9:4045–4071 Ferna´ndez-Gonza´lez F, Loidi J, Moreno JC (2005) Impacts on plant biodiversity. In: Moreno JM (ed) Impacts on climatic change in Spain. OCCE, Ministerio de Medio Ambiente, Madrid, pp 183–248 Ferreira MT, Aguiar FC, Nogueira C (2005) Changes in riparian woods over space and time: influence of environment and land use. For Ecol Manage 212:145–159 Fielding AH, Bel JF (1997) A review of methods for the assessment of prediction errors in conservation presence/absence models. Environ Conserv 24:38–49 Fryxell JM, Wilmshurst JF, Sinclair AR, Haydon DT, Holt RD, Abrams PA (2005) Landscape scale, heterogeneity, and the viability of Serengeti grazers. Ecol Lett 8:328–335 Garel E, Pinto L, Santos A, Ferreira O´ (2009) Tidal and river discharge forcing upon water and sediment circulation at a rock-bound estuary (Guadiana estuary, Portugal). Estuar Coast Shelf Sci 84:269–281 Gitay H, Wilson JB, Lee WG (1996) Species redundancy: a redundant concept? J Ecol 84:121–124 Gunderson LH (2000) Ecological resilience in theory and application. Annu Rev Ecol Syst 31:425–439 Gurnell AM, Bertoldi W, Corenblit D (2012) Changing river channels: the roles of hydrological processes, plants and pioneer fluvial landforms in humid temperate, mixed load, gravel bed rivers. Earth Sci Rev 111:129–141 Heller NE, Zavaleta ES (2009) Biodiversity management in the face of climate change: a review of 22 years of recommendations. Biol Conserv 142:14–32 Hering D, Borja A, Carstensen J, Carvalho L, Elliott M, Feld CK, Heiskanen AS, Johnson RK, Moe J, Pont D, Solheim AL, van de Bund W (2010) The European Water Framework Directive at the age of 10: a critical review of the achievements with recommendations for the future. Sci Total Environ 408:4007–4019 Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A (2005) Very high resolution interpolated climate surfaces for global land areas. Int J Climatol 25:1965–1978 Hoelzer A (2003) Vegetation ecological studies at the lower course of Sabor river (Tra´s-os-Montes, NE Portugal). MsD These, University of Bremen, Bremen Holling CS (1973) Resilience and stability of ecological systems. Annual Rev Ecol Evolu Sys 4:1–23

Honrado J, Alves P, Aguiar C, Ortiz S, Barreto Caldas F (2003) De Vegetatio Lusitana Notae-I: 1. Juresian riparian birch woodlands: Carici reuterianae—Betuletum celtibericae as. Nova. Silva Lusitana 11:237–241 [in Portuguese] Hughes S, Santos J, Ferreira MT, Carac¸a R, Mendes AM (2009) Ecological assessment of an intermittent Mediterranean river using community structure and function: evaluating the role of different organism groups. Freshw Biol 54:2383–2400 Hughes SJ, Cabecinha E, Santos J, Mendes Andrade CM, Lopes DM, Trindade HM, Cabral JA, Santos M, Lourenc¸o M, Aranha JT, Fernandes L, Morais MM, Leite SM, Oliveira PC, Cortes R (2012) A predictive modelling tool for assessing climate, land use and hydrological change on reservoir physicochemical and biological properties. Area 44(4):432–442 Hupp CR, Osterkamp WR (1996) Riparian vegetation and fluvial geomorphic processes. Geomorphology 14:277–295 INAG (2008) Insituto Nacional da A´gua. Protocolo de amostragem e ana´lise para os macro´fitos. In Manual para a avaliaca˜o biolo´gica da qualidade da a´gua em sistemas fluviais segundo a Directiva Quadro da A´gua. http://www. apambiente.pt/dqa/assets/protocolo-de-amostragem-e-an%C3 %A1lise-para-os-macr%C3%B3fitos.pdf. Accessed 29 Jul 2015, (in Portuguese) Ives AR, Carpenter SR (2007) Stability and diversity of ecosystems. Science 317:58–62 Jansson R, Zinko U, Merritt DM, Nilsson C (2005) Hydrochory increases riparian plant species richness: a comparison between a free-flowing and a regulated river. J Ecol 93:1094–1103 Jiang J, Gao D, DeAngelis DL (2012) Towards a theory of ecotone resilience: coastal vegetation on a salinity gradient. Theor Popul Biol 82:29–37 Khavhagali VP (2011) Climate change impacts on plants. Grassroots 11:9–13 Lake PS (2000) Disturbance, patchiness, and diversity in streams. J North Am Benthol Soc 19:573–592 Martins A´, Crespı´ AL, Castro A, Fernandes CP, Rocha J, Bernardos S, Aguiar C, Amich F (2007) Contribuicio´n para la caracterizacio´n florı´ stico-ambiental del norte de Portugal. Botan Complut 31:99–111 (in Spanish) May RM (1974) Stability and complexity in model ecosystems, 2nd edn. Princeton University Press, New Jersey McGrady-Steed J, Harris PM, Morin PJ (1997) Biodiversity regulates ecosystem predictability. Nature 390:162–165 Meyer JL, Sale MJ, Mulholland PJ, LeRoy Poff N (1999) Impacts of climate change on aquatic ecosystems functioning and health. J Am Water Resour Assoc 35:1373–1386 Ministe´rio do Ambiente, Ordenamento do Territo´rio e Energia 2015 Programa Nacional de Barragens de Elevado Potencial Hidroele´trico (PNBEPH). http://www.apambiente.pt/?ref= 16&subref=7&sub2ref=9&sub3ref=835. Accessed 29 Jul 2015, (in Portuguese) Molina JA (1996) Sobre la vegetacio´n de los humedales de la Penı´ nsula Ibe´rica (1. Phragmiti-Magnocaricetea). Lazaroa 16:27–88 (in Spanish) Naeem S, Knops JM, Tilman D, Howe KM, Kennedy T, Gale S (2000) Plant diversity increases resistance to invasion in the absence of covarying extrinsic factors. Oikos 91:97–108 Naiman RJ, De´camps H (1997) The ecology of interfaces: riparian zones. Annual Rev Ecol Evol Sys 28:621–658 Niemi GJ, McDonald ME (2004) Application of ecological indicators. Annual Rev Ecol Evol Sys, pp 89–111 North CP, Davidson SK (2012) Unconfined alluvial flow processes: Recognition and interpretation of their deposits, and the significance for palaeogeographic reconstruction. Earth Sci Rev 111:199–223 Occhipinti-Ambrogi A, Galil BS (2004) A uniform terminology on bioinvasions: a chimera or an operative tool? Mar Pollut Bull 49:688–694 Parmesan C (2006) Ecological and evolutionary responses to recent climate change. Annu Rev Ecol Evol Sys, pp 637–669

1003 Pearson TN, Li HW, Lamberti GA (1992) Influence of habitat complexity on resistance to flooding and resilience of stream fish assemblages. Trans Am Fish Soc 121:427–436 Pereira MCDMD (2009) A flora e vegetac¸a˜o da serra de Monfurado (Alto Alentejo-Portugal). Guineana 15:153–166 (in Portuguese) Pereira EMP (2013) Ana´lise Geobotaˆnica dos Bosques e Galerias Ripı´ colas da Bacia Hidrogra´fica do Tejo em Portugal. PhD thesis, University of Lisbon, Instituto de Geografia e Ordenamento do Territo´rio. Lisbon Peterson GD (2002) Estimating resilience across landscapes. Conservation Ecology 6: 17. http://www.consecol.org/vol6/ iss1/art17/. Accessed 29 Jul 2015 Peterson G, Allen CR, Holling CS (1998) Ecological resilience, biodiversity, and scale. Ecosystems 1:6–18 Peterson AT, Papes M, Eaton M (2007) Transferability and model evaluation in ecological niche moedeling: a comparison of GARP and Maxent. Ecography 30:550–560 Phillips SJ (2010) Species’ distribution modeling for conservation educators and practitioners. Exercise. American Museum of Natural History. Lessons in Conservation. http://ncep.amnh. org/linc. Accessed 15 Nov 2013 Phillips SJ, Dudı´ k M, Schapire RE (2004) A maximum entropy approach to species distribution modeling. In: Proceedings of the 21st International Conference on Machine Learning. ACM Press, New York Phillips SJ, Anderson RP, Schapire RE (2006) Maximum entropy modeling of species geographic distributions. Ecol Model 190:231–259 Pickett STA, Kolasa J, Armesto JJ, Collins SL (1989) The ecological concept of disturbance and its expression at various hierarchical levels. Oikos 54:129–136 Pimm SL (1984) The complexity and stability of ecosystems. Nature 307:321–326 Poff NL (1997) Landscape filters and species traits: towards mechanistic understanding and prediction in stream ecology. J North Am Benthol Soc pp 391–409 Portuguese National Institute of Statistic (2012). Censos 2011. Resultados definitivos. http://www.ine.pt/scripts/flex_ definitivos/Main.html. Accessed 29 Jul 2015 (in Portuguese) Ramirez J, Jarvis A (2008) High resolution statistically downscaled future climate surfaces. Int Centre Trop Agricult, CIAT Rivas-Martı´ nez S, de la Fuente Garcı´ a V (1986) Alisedas mediterra´neo-iberoatla´nticas en la Penı´ nsula Ibe´rica. Studia Botanica 5:9–38 (in Spanish) Rodewald AD, Bakermans MH (2006) What is the appropriate paradigm for riparian forest conservation? Biol Conserv 128:193–200 Rozeira A (1944) A flora da provı´ ncia de Tra´s-os-Montes e Alto Douro. Memo´rias da Sociedade Broteriana 3:110–118 (in Portuguese)

Rykiel EJ (1985) Towards a definition of ecological disturbance. Aust J Ecol 10:361–365 Sabo JL, Sponseller R, Dixon M, Gade K, Harms T, Heffernan J, Jani A, Katz G, Soykan C, Watts J, Welter J (2005) Riparian zones increase regional species richness by harboring different, not more, species. Ecology 86:56–62 Santos JA, Corte-Real J, Leite SM (2005) Weather regimes and their connection to the winter rainfall in Portugal. Int J Climatol 25:33–50 Seavy NE, Gardali T, Golet GH, Griggs FT, Howell CA, Kelsey R, Small SL, Viers JH, Weigand JF (2009) Why Climate Change Makes Riparian Restoration More Important than Ever: Recommendations for Practice and Research. Ecol Restorat 27:330–338 Stevens PF (2001 onwards). Angiosperm Phylogeny Website. http://www.mobot.org/mobot/research/apweb/. Accessed 29 Jul 2015 Suding KN, Lavorel S, Chapin FS, Cornelissen JH, Diaz S, Garnier E, Goldberg D, Hooper DU, Jackson ST, Navas ML (2008) Scaling environmental change through the communitylevel: a trait-based response-and-effect framework for plants. Glob Change Biol 14:1125–1140 Taylor DL, Herriott IC, Stone KE, McFarland JW, Booth MG, Leigh MB (2010) Structure and resilience of fungal communities in Alaskan boreal forest soils. Can J For Res 40:1288–1301 The Angiosperm Phylogeny Group (2009) An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG III. Bot J Linn Soc 161:105–121 Thuiller W, Lavorel S, Arau´jo MB, Sykes MT, Prentice IC (2005) Climate change threats to plant diversity in Europe. Proc Natl Acad Sci USA 102:8245–8250 Tilman D (1996) Biodiversity: population versus ecosystem stability. Ecology 77:350–363 Trigo RM, DaCamara CC (2000) Circulation weather types and their influence on the precipitation regime in Portugal. Int J Climatol 20:1559–1581 Trigo RM, Pozo-Va´zquez D, Osborn TJ, Castro-Dı´ ez Y, Ga´mizFortis S, Esteban-Parra MJ (2004) North Atlantic Oscillation influence on precipitation, river flow and water resources in the Iberian Peninsula. Int J Climatol 24:925–944 Van Ruijven J, Berendse F (2010) Diversity enhances community recovery, but not resistance, after drought. J Ecol 98:81–86 Walker B (1995) Conserving biological diversity through ecosystem resilience. Conserv Biol 9:747–752 White PS (1979) Pattern, process, and natural disturbance in vegetation. Bot Rev 45:229–299