A Long-Term Macroecological Analysis of the Recovery of a Waterbird

A Long-Term Macroecological Analysis of the Recovery of a Waterbird

A Long-Term Macroecological Analysis of the Recovery of a Waterbird Metacommunity after Site Protection Janina Pagel1,2, Alejandro Martı´nez-Abraı´n1,...

2MB Sizes 0 Downloads 0 Views

Recommend Documents

A comparison of the structure of 2 waterbird assemblages - CiteSeerX
the Amu-Darya River is expected to decrease over time. (Agal'tseva et al., 2011). Large-scale water withdrawals from the

Journeys of Recovery - Going the Distance: Journeys of Recovery, a
Oct 27, 2010 - DaviD L. BroWn is an Emmy Award–winning documentary ... and Best Graphics and Animation in a Program) a

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

P10, Accordi, Iury A - The Waterbird Society
interactions between seabirds and the marine ecosystems that support them. ..... do Curral (Morro de Sao Paulo), Bahia d

assessment of heat recovery an recovery efficiency of a seasonal
ASSESSMENT OF HEAT RECOVERY AN RECOVERY EFFICIENCY OF A SEASONAL. THERMAL ENERGY STORAGE IN A MOIST POROUS MEDIUM. M.A.

Longterm response of a Mojave Desert winter annual plant community
Long-term response of a Mojave Desert winter annual plant community to a whole-ecosystem atmospheric CO2 manipulation (F

The Process of Divorce Recovery: A Review of the Research.
Richard W. Gastil. APPROVED: (°je10't_Z .... (Bohannon, 1970; Goode, 1956; Wiseman, 1975). ..... research reviewed (Goo

A Publicafion of the Towing & Recovery Association of Kentucky
22 state associations, up from 19 the year before, to a meeting at the Florida ... Owner-Operator Independent Drivers As

Optimisation of the Recovery Section of a Polyolefin Catalyst
Sep 20, 2010 - Polyolefin Catalyst Manufacturing Process. Master´s Thesis in the Programme Innovative and Sustainable C

THE WATERBIRD SOCIETY Die WATERBIRD SOCIETY
Entenverwandte und Schreitvögel, machen einen großen Teil der. Vogelwelt aus. Wasservögel sind vielfältig in. Erscheinun

A Long-Term Macroecological Analysis of the Recovery of a Waterbird Metacommunity after Site Protection Janina Pagel1,2, Alejandro Martı´nez-Abraı´n1,3*, Juan Antonio Go´mez4, Juan Jime´nez4, Daniel Oro1 1 Population Ecology Group, Mediterranean Institute for Advanced Studies (Consejo Superior de Investigaciones Cientı´ficas-Universitat de les Illes Balears), Esporles, Mallorca, Spain, 2 University of Applied Sciences Bremen, Bremen, Germany, 3 Universidade da Corun˜a, Departamento de Bioloxı´a Animal, Bioloxı´a Vexetal e Ecoloxı´a, Facultade de Ciencias, A Corun˜a, Spain, 4 Servicio de Vida Silvestre, Generalitat Valenciana, Conselleria de Infraestructuras, Territori i Medi Ambient, Ciutat Administrativa 9 d9 Octubre, Torre 1, Valencia, Spain

Abstract We used the so called ‘‘land-bridge island’’ or ‘‘nested-subsets’’ theory to test the resilience of a highly fragmented and perturbated waterbird metacommunity, after legal protection of 18 wetlands in the western Mediterranean. Sites were monitored during 28 years and two seasons per year. The metacommunity was composed by 44 species during breeding and 67 species during wintering, including shorebirds, ducks, herons, gulls and divers (Podicipedidae). We identified a strong nested pattern. Consistent with the fact that the study system was to a large extent a spatial biogeographical continuous for thousands of years, fragmented only during the last centuries due to human activities. Non-random selective extinction was the most likely historical process creating the nested pattern, operated by the differential carrying capacity (surface-area) of the remaining sites. We also found a positive temporal trend in nestedness and a decreasing trend in species turnover among sites (b-diversity), indicating that sites are increasingly more alike to each other (i.e. increased biotic homogenization). This decreasing trend in b-diversity was explained by an increasing trend in local (a) diversity by range expansion of half the study species. Regional (c) diversity also increased over time, indicating that colonization from outside the study system also occurred. Overall our results suggest that the study metacommunity is recovering from historical anthropogenic perturbations, showing a high long-term resilience, as expected for highly vagile waterbirds. However, not all waterbird groups contributed equally to the recovery, with most breeding shorebird species and most wintering duck species showing no geographical expansion. Citation: Pagel J, Martı´nez-Abraı´n A, Go´mez JA, Jime´nez J, Oro D (2014) A Long-Term Macroecological Analysis of the Recovery of a Waterbird Metacommunity after Site Protection. PLoS ONE 9(8): e105202. doi:10.1371/journal.pone.0105202 Editor: Claudia Mettke-Hofmann, Liverpool John Moores University, United Kingdom Received November 14, 2013; Accepted July 22, 2014; Published August 18, 2014 Copyright: ß 2014 Pagel et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: The corresponding author was supported by an Isidro Parga-Pondal postdoctoral contract from University of A Corun˜a. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * Email: [email protected]

Sebastia´n-Gonza´lez et al. [13], who found substantial nestedness in all seasons and years. Lately the old biogeographical concept of nestedness has been adopted to analyze the architecture of mutualistic networks (see e.g. [14]), but that application has nothing to do with the aims of our study. Here we analyze longitudinally the diversity pattern of a waterbird metacommunity occupying formerly much more continuous, but currently highly fragmented, Mediterranean costal areas to test if it is nested, as expected according to the so-called ‘‘landbridge island’’ paradigm, and also if nestedness has increased over time (see e.g. [15], [16]). This would be expected for a highly vagile zoological group such as waterbirds, and will suggest longterm metacommunity tendency to restore its original structure of spatial homogeneity in species composition once sources of perturbation are under control (i.e. high metacommunity resilience). An increasing trend in nestedness over time should be paralleled by a decreasing trend in species turnover rate (the socalled b-diversity) among sites, since sites become more similar among them. After approaching the long-term pattern of nestedness we will move on to explore potential (non-random) ecological processes behind it by relating nestedness with either selective colonization or selective extinction [10], [11], [17], accounting as well for other

Introduction The idea of nestedness in biogeography dates back to the 1930s, but the modern use of the concept to explain diversity patterns in animal communities and metacommunities starts with Patterson and Atmar [1], when studying the community structure of nonvolant mammalian faunas in naturally-fragmented archipelagos in montane habitats of the American Rocky Mountains. This concept of nested patterns has been widely applied to landbridge islands [2], [3], those located on the continental shelf, but also to other types of habitats which originally formed a continuous unit to become later on fragmented and isolated due to a variety of reasons, such as mountain tops [4], [5], boreal forests affected by glaciations [6], cloud forest fragments [7], national parks [8] or lake islands [9]. Applied in its original biogeographical sense, a nested pattern means that the species composition of species-poor assemblages is a subset of the species composition of richer assemblages [10], [11]. Many coastal Mediterranean wetlands are also examples of habitats relatively recently isolated and perturbated by anthropogenic activities, which have not received much attention from the macroecological and biogeographical perspectives: nested-subsets or landbridge island framework. Some exceptions are the studies by Paracuellos and Tellerı´a [12] and PLOS ONE | www.plosone.org

1

August 2014 | Volume 9 | Issue 8 | e105202

Long-Term Recovery of an Avian Metacommunity

Figure 1. Location and size of the study wetlands in Eastern Spain. doi:10.1371/journal.pone.0105202.g001

To the best of our knowledge probabilities of species detection can be considered to be constant across sites and seasons from year to year as both methodology and human team composition have remained approximately constant during the study period, and hence results derived from species-richness data are comparable among years despite the biases that differential detectability among species could introduce in the absolute estimates of our metrics of nestedness [18]. No birds were collected or samples taken. Two of the co-authors (J. A. Go´mez and J. Jime´nez) are the civil servants from the regional government in charge of coordinating field teams, and the authors have collaborated directly on the detection and count of waterbirds in the main study sites over many years. Our field work did not violate any law or invaded private land at all. Many sites in this study are protected as nature parks (i.e. Cabanes, Albufera, Pego-Oliva, Santa-Pola, Torrevieja, El Hondo) since 1986–1988 or have been protected afterwards as Important Bird Areas (IBA) (2007–2009) by the regional environmental authority. Detailed information of the sites protected as IBAs can be consulted at: http://www.docv.gva.es/datos/2009/06/09/pdf/2009_6699.pdf. Winter counts were performed simultaneously in all wetlands each year during two weeks around the second weekend of January, in coordination with the International Waterbird Census (IWC) (for further details see http://www.wetlands.org/African EurasianWaterbirdCensus/tabid/2788/Default.aspx). Wintering ducks, coots or divers (Podicipedidae) were counted from the distance and from fixed sites every year using scopes. Other wintering bird groups such as herons, gulls or shorebirds were

possible causes of nestedness such as passive sampling or habitat heterogeneity [6], [11], [18], [19]. If the pattern is nested we would expect that differential extinction, rather than differential colonization, is the major process creating the nested-subset pattern, because the original situation was one in which all wetlands roughly formed a geographical continuous which has been fragmented by human activities during the last centuries [20]. Reduction in patch size, due to anthropogenic intervention, is known to reduce faunal diversity by shrinking population size leading species to local extinction and lack of colonization [3], [21]. Additionally, selective extinction could be due, in a nonmutually exclusive way, to the losing of habitat types as patch size decreases, affecting more strongly to specialist species [11]. Finally we analyse the applied benefits and drawbacks of having a nested architecture for the long-term resistance and resilience of the system to human perturbations widening the link between nestedness and conservation biology [22], [23], [24], [25].

Methods The data set and field procedures An official data set on bird counts for a 28-year period (1984– 2011), compiled over the years by the environmental authority of Comunidad Valenciana region (i.e. Generalitat Valenciana), was used to analyse the nested pattern of a waterbird metacommunity including 44 breeding and 67 wintering species in 18 wetlands in the western Mediterranean (Eastern Spain). We show in Figure 1 the location of sites, their relative size and distance among them. PLOS ONE | www.plosone.org

2

August 2014 | Volume 9 | Issue 8 | e105202

Long-Term Recovery of an Avian Metacommunity

Figure 2. Overall nestedness of the waterbird metacommunity. Graphical representation of the qualitative maximally packed matrix during A) the breeding season and B) the wintering season (1984–2011). Black dots are species presences recorded at least once during the study period; white dots are species absences. A perfectly nested system would have a 50% fill in the upper-left corner. doi:10.1371/journal.pone.0105202.g002

counted along fixed car itineraries with variable detection band widths depending on the characteristics of each study site. Wintering marsh harriers were counted around sunset at communal roosts. Breeding season counts were not coordinated internationally and were mostly carried out by the staff of protected areas that monitor study sites. Visits to the study area were carried out almost on a daily basis over the whole breeding season (March-August) to prevent overlooking relevant information due to the lack of complete overlap in the breeding calendar of the study species. Counts were performed using specific and fixed methodologies for each species. Colonial species (herons, gulls, terns, shorebirds, and flamingo) were counted by visiting breeding colonies and counting individual nests at the peak of their breeding period. Non-colonial species (ducks, coots, Podicipedidae) were detected by inspecting water masses by means of motor boats, counting nests or birds PLOS ONE | www.plosone.org

displaying breeding behaviour or adults in the company of chicks. Species of difficult detection (rallid species, little bittern) were detected prospecting the study area in detail by means of boats propelled manually in shallow water areas. Further information on winter and summer counts for the whole study region during the period 1984–2004 is available at http://www.cma.gva.es/ webdoc/documento.ashx?id = 164402. Most study sites were former (Holocene) coastal lagoons in different stages of its natural succession towards terrestrial ecosystems, but with a high degree of human influence on all components of the structure of their animal and plant communities. Evidence for this is the high rate of loss of suitable habitat in Mediterranean coastal wetlands for many zoological groups, including birds, reported during the decades prior to habitat protection (see e.g. [26], [27]).

3

August 2014 | Volume 9 | Issue 8 | e105202

212.65

211.42 72.44

66.15 58.04

65.97 1.67 50.22 Wintering

69.27

Lower 95% CI Upper 95% CI z’ SD

2.15 62.38 Breeding

Simulated Observed

35.24

Season

WNODF

Quantitative matrix

18.21u

PLOS ONE | www.plosone.org

Qualitative Matrix: the values are based on the overall qualitative matrix (a species presence or absence in a site within 28 years). Simulated T/NODF is in each case the average of 1000 Monte Carlo simulations run in ANINHADO. SD = standard deviation of simulated T. CI = 95% confidence interval of simulated T/NODF. Z = standardized effect size of T (see text). NODFr = Relative NODF (see text). Quantitative Matrix: the values are based on the overall quantitative matrix (an average of species abundance in a site within 28 years). Simulated WNODF is in each case the average of 1000 Monte Carlo simulations run with the NODF software. SD = standard deviation of simulated WNODF. CI = 95% confidence interval of simulated WNODF. Z’ = standardized effect size of WNODF (see text). doi:10.1371/journal.pone.0105202.t001

0.16 72.63 66.83 69.73 80.89 28.38 55.34u 41.26u 3.59

NODFr

Wintering

48.30u

61.26

Upper 95%/CI Lower 95%/CI

52.72 56.99

Simulated Observed

77.75 29.26 53.26u 38.96u

Lower 95% CI Upper 95% CI z SD

3.65 Breeding

Simulated Observed

12.32u

Season

46.11u

NODF Temperature (T)

Qualitative matrix

Table 1. Analysis of the overall nestedness of the waterbird metacommunity studied during the breeding and wintering seasons.

0.36

Long-Term Recovery of an Avian Metacommunity

Overall nestedness For the nestedness analysis we first assembled presence-absence summary matrices from the dataset of bird counts obtained as described above, both for breeding and wintering seasons, with wetlands in rows and bird species in columns. A presence hence indicates that the species has been observed in a particular wetland at least once during the period 1984–2011. To quantify the degree of nestedness for the summarized qualitative matrices we calculated two metrics, a) the matrix temperature (T) [21] and b) the nested overlap decreasing fill (NODF) [28] with software ANINHADO 3.0 [28], [29]. Specifically ANINHADO uses the dispersal of unexpected presences or absences in the maximally packed matrix to derive the observed temperature of the matrix which varies from 0u (perfectly nested) to 100u [28]. The NODF metric is based on standardized differences in row and column fills and paired matching occurrences, ranging from 100 (perfectly nested) to 0 [11]. The observed temperature or NODF is then compared to the mean temperature of a frequency distribution of one thousand Monte-Carlo simulated temperatures (or NODFs) obtained under a null model selected out of four null models available. We used null model number two (CE) because it calculates the probability,that a cell  (aij) in the simulated matrix Pi Pj z =2 where Pi = number of shows a presence, as C R presences in the row i; Pj = number of presences in the column j; C = number of columns; R = number of rows, that is the assignment of a presence takes into account data-derived information (i.e. species distribution range and wetland richness), whereas all other three null models available do the assignment of presences either by columns only, rows only or at random. Hence results derived from null model 2 could be considered more restrictive. The species that are present in most wetlands are placed in the top left column, whereas wetlands with the highest number of species are placed in the topmost row. ANINHADO orders wetlands so that nestedness is maximally visualized. A perfectly nested system has a 50% fill in the upper-left corner of the packed matrix [28]. To visualize the maximally packed matrix we used the ‘‘bipartite’’ package in R software [30]. In this graphical representation of the geographical matrix black squares stand for species presences and white squares for species absences. We chose to use both metrics to a) allow comparisons with previous studies based on temperature, and b) provide information regarding the suitability of one of the metrics over the other by checking whether results from both metrics coincided or not. To compare the degree of nestedness of qualitative matrices with that of quantitative matrices we assembled quantitative matrices with the average of abundance of each species in each wetland (wetlands in rows, species in columns) both for breeding and wintering for the period 1984–2011. We used software NODF (not to confuse with the NODF metric used together with qualitative matrices) [31] to calculate the degree of nestedness of the quantitative matrices. A nested pattern with a quantitative matrix, compared to a qualitative matrix, not only means that the species composition in smaller assemblages is a subset of that in larger assemblages but that their abundances are also nested (i.e. all populations making up local assemblages have lower abundances than their conspecific populations in richer assemblages) [31]. To use NODF software the unpacked quantitative matrices were modified first by software EcoSim 7.0 [31], [32]. EcoSim changes the format of the quantitative matrix in a space delimited text file matrix after importing the quantitative matrix from Microsoft Excel. This text file is needed to run NODF software. To calculate the degree of nestedness using quantitative data NODF software uses a modification of the NODF index [28] 4

August 2014 | Volume 9 | Issue 8 | e105202

Long-Term Recovery of an Avian Metacommunity

PLOS ONE | www.plosone.org

5

August 2014 | Volume 9 | Issue 8 | e105202

Long-Term Recovery of an Avian Metacommunity

Figure 3. Inter-annual variability of degree of nestedness of the qualitative and quantitative matrices during the breeding and wintering seasons (1990–2011). This figure shows the inter-annual variability of degree of nestedness of the qualitative matrix during A) breeding season, B) wintering season and of the C) quantitative matrix (left breeding, right wintering). Note that absolute values of z9-scores were used despite z-values are negative. z. = z-score; z. = z9-score. The lines show the best fit (solid line) and 95% confidence bands (dotted line). See Table 2 for a summary of parameter estimates of the general linear models fitted to these standardized effect sizes. doi:10.1371/journal.pone.0105202.g003

gamma or regional diversity (i.e. the number of species in all our metacommunity) and a is local diversity (i.e. the arithmetic mean of the number of species in each of our study sites) [37]. To analyze the trends in b, a and c-diversity we also fitted general linear models to data in R. Given that during the years from 1984 to 1989 a smaller sampling effort was done (i.e. a smaller number of wetlands were censused) we used the time series only from 1990 on, to analyze the trend in nestedness over time. Range expansion was calculated by subtracting the average of the number of occupied sites in the second half of the time series (2001–2011) from the average number of occupied sites during the first half of it (1990–2000) (‘‘Dsites’’ hereon). The second half of the time series corresponds approximately to a period of consolidated protection of the sites, after a decade of protection by law as nature parks, chosen so that sample size is equalized with the first half of the time series.

called Weighted Nestedness metric based on Overlap and Decreasing Fill (WNODF). WNODF measures the degree of nestedness based directly on overlap and decreasing fill, and, in contrast to temperature, with a range of degree of nestedness from 100 (perfectly nested) to 0 [31]. To measure the degree of nestedness WNODF quantifies if the marginal total (i.e. incidences or richness) of a given sequence of columns or rows decrease and also if the study system loses species in an ordered way, as in the case of NODF or T. We used null model rc that assigns individuals to matrix cells proportionally to observed row and column abundance totals until, for each row and column, total abundances are reached [31]. As sorting option we used row/column abundance totals. To verify the degree of nestedness of our maximally packed qualitative and quantitative matrices and to compare them with other studies we calculated standardized effect sizes, which measure the number of standard deviations that the observed index is above or below the mean index of the simulated index [13], [33]. For temperature (T) of the qualitative matrix we calculated the standardized effect size as a z-score (observed T – mean simulated T)/standard deviation of the simulated T; for NODF we obtained a relative NODF (observed NODF – mean simulated NODF)/mean simulated NODF following MontesinosNavarro et al. [34]. For the quantitative matrix we calculated the standardized effect size (z9-score) with NODF software in the same way as the z-score for the qualitative matrix is calculated. A z-score with a value below -2.0 or above 2.0 indicates approximate statistical significance for a at the 5% a priori risk level of committing a Type I error [13], [35]. The relative NODF values cannot be compared directly with temperature values. We ordered the wetlands of the overall qualitative matrices both for breeding and wintering by their degree of nestedness calculated with BINMATNEST [36]. BINMATNEST reorders rows and columns until nestedness is maximized and unexpectedness is minimized by using a genetic algorithm that is more accurate to order rows and columns than that used by other programs [15], [36]. We explored whether selective extinction or selective colonization were the processes behind the nested pattern by calculating Spearman rank correlation coefficients. We correlated the row order of the qualitative packed matrix with the size of each wetland (extinction), and also with the distance to the nearest wetland (colonization) using the R software [30]. We verified whether results were similar when data were not subjected previously to BINMATNEST.

Results Overall nestedness The waterbird metacommunity was found to be highly nested, both during the breeding and wintering seasons, as observed temperatures were quite low (Figure 2, Table 1). The qualitative matrix showed a higher nested pattern during breeding than during wintering, as determined by a higher (negative) z-score during breeding. Higher nestedness during the breeding period was also detected when using the relative NODF metric with the qualitative matrix. The values of the standardized effect sizes for WNODF in the quantitative matrix indicated no nestedness either during wintering or breeding (Table 1 B).

Temporal dynamics in nestedness The study metacommunity showed a nested pattern in most study years, when using the qualitative but not when using the quantitative matrix (Table S1). Interestingly, the z-scores of the qualitative matrix showed a strong positive trend over time (Figure 3, Table 2), indicating increasing nestedness during both study seasons, contrarily to that found by other authors for artificial Mediterranean wetlands [13]. The values of NODFr also showed a strong positive trend during wintering, although not during breeding (Figure 3, Table 2). Regarding the quantitative matrix the increasing trend of the negative values of z9-scores during breeding indicated poorer nestedness with time (i.e. smaller observed values of the metric compared to simulations). During wintering no pattern of nestedness appeared for the quantitative matrix (Figure 3, Table 2). Regarding Dsites results indicated that roughly 50% of the species, both in wintering and summer, were expanding their ranges (Figure 4).

Inter-annual variability in nestedness and diversity To calculate the degree of inter-annual variability in nestedness we used the same software and calculations as to obtain the degree of overall nestedness, but for each year of the study period both for breeding and wintering, using all three standardized effect sizes. Since we detected that the three effect sizes used showed, as a rule, an increasing trend over time, we fitted general linear models to the observed trend in order to determine their strength (slope), statistical significance (95% confidence interval) and degree of fit (r2). In order to find out whether increasing degree of nestedness was paralleled by a decrease in beta-diversity calculated b wec  diversity both for breeding and wintering as b~ , where c is a PLOS ONE | www.plosone.org

Beta-diversity and nestedness In order to validate our results on increased nestedness over time we calculated beta-diversity. As nestedness increases bdiversity is expected to decrease in a closed system [38], [39], [40]. However, we found a decreasing trend in b-diversity in our open system during breeding (slope = 20.06; 95% CI slope = 20.09, 20.04; r2 = 0.53) (Figure 5) and wintering (slope = 20.03; 95% CI slope = 20.05, 20.01; r2 = 0.21) (Figure 5). Since b-diversity is 6

August 2014 | Volume 9 | Issue 8 | e105202

Long-Term Recovery of an Avian Metacommunity

Figure 4. Delta-sites or change in number of sites occupied by each study species over time (1990–2011). Delta-sites are shown for both A) breeding and B) wintering. Delta-sites is defined as the subtraction of the average of the number of occupied sites in the second half of the time series (2001–2011) from the average number of occupied sites during the first half of the time series (1990–2000). Positive values of the index indicate species expansion. The dotted line is the arbitrary minimum value beyond which we consider range expansion is taking place. doi:10.1371/journal.pone.0105202.g004

calculated as gamma-diversity over alpha-diversity a decrease in bdiversity can be due either to a decrease in gamma-diversity or to an increase in alpha-diversity [41], [42]. We found indeed an

PLOS ONE | www.plosone.org

increasing trend in a-diversity during breeding (slope = 0.29; 95% CI slope = 0.23, 0.35; r2 = 0.81) (Figure 5) and, wintering (slope = 0.39; 95% CI slope = 0.26, 0.52; r2 = 0.6) (Figure 5). But

7

August 2014 | Volume 9 | Issue 8 | e105202

Long-Term Recovery of an Avian Metacommunity

Table 2. General linear models fitted to the change of the standardized effect sizes over time both for the breeding and wintering seasons and for the qualitative and quantitative matrices.

Matrix

Season

Metric

Qualitative

Breeding

Quantitative

r2

Slope

Lower 95% CI

Upper 95% CI

NODFr

0.0029

20.0030

0.0088

20.0031

z-score

0.1353

0.0891

0.1815

0.6031

Wintering

NODFr

0.0251

0.0189

0.0312

0.7482

z-score

0.2291

0.1648

0.2934

0.6946

Breeding

z‘-score

0.3211

0.1418

0.5003

0.0022

z‘-score

20.0762

20.1935

0.0412

0.0286

Wintering 2

CI = 95% confidence interval of the slope. r = coefficient of determination. NODFr = Relative NODF. Z = standardized effect size of temperature. Z’ = standardized effect size of WNODF. Values in bold are statistically significant (i.e. 0 is within the 95% confidence intervals). doi:10.1371/journal.pone.0105202.t002

interestingly, we also found strong increasing trends in c-diversity during wintering (slope = 0.89; 95% CI slope = 0.68, 1.1; r2 = 0.77) and breeding (slope = 0.39; 95% CI slope = 0.29, 0.49; r2 = 0.72;) (Figure 5). Spearman rank correlations between b-diversity and nestedness were strong and statistically significant when using the z-score both for breeding (rs = 20.69; p,0.001) and wintering (rs = 20.74; p,0.001), although not when using the NODFr metric.

nestedness studies typically do not account for detection probability [18]. These false zeros would reduce the degree of nestedness incorrectly. In our case sampling artefacts are very implausible because we have used a long-term (28-years) time series for our study, and hence it is very unlikely that our summarized matrix is missing some species which are present but not detected in the study system. Sampling artefacts are not to be mistaken with the mechanism of passive sampling [11], [43]; in this case, species colonize fragmented habitats proportional to their abundance. In our study system some abundant species such as herons were lost as we moved from bigger to smaller wetlands, especially during breeding, suggesting that factors other than passive sampling, related to size of the wetland or habitat heterogeneity, were most likely acting. Differences in water quality probably did not influence the degree of nestedness of the metacommunity [6], because more or less similar efforts have been devoted to water quality restoration in all wetlands. Hence we have presently an array of wetlands in which most sites are all in a similar (although still poor) water quality condition. One particular component of habitat quality is human disturbance. Some authors have found that nestedness can be promoted by human disturbance but depending on its level and the disturbance tolerance of the species [19]. In our case differences in human disturbance are not likely a cause of nestedness because most coastal wetlands in the study region are effectively protected as nature parks, and human uses are alike. A further option to get a nested pattern is to have an array of sites with different habitat heterogeneity so that habitat type is nested in the sense that sites with smaller species assemblages have a subset of the habitats present in the richer sites. Losing habitats sequentially can lead to losing species in an ordered way [11]. According to our long experience in the study area, that factor is most likely influencing nestedness, but we have no fine-grain quantitative data available to test its influence. However habitat heterogeneity is most likely highly correlated with wetland size (area) and probably the identified influence of decreasing wetland size on the loosing of species is in fact driven by the loss of habitat heterogeneity [44]. An alternative causal factor of nestedness is the fact that the loss of species may be proportional to local abundance (i.e. population size or density).

Extinction versus colonization processes Spearman rank correlation coefficients between row order of the packed qualitative matrix and size of the wetlands were negative, strong, and statistically significant, in most years and in both seasons (Table 3, Table S2), either using or not BINMATNEST to re-order rows and columns, suggesting a role of selective extinction in creating the nestedness pattern observed (i.e. nestedness generated by ordered species loss). On the contrary, the Spearman rank correlation coefficients between row order and distance to the nearest wetland were both positive and negative, showed very low values and, as a rule, were not statistically significant (Table 3), suggesting a low influence of selective colonization on the observed pattern of nestedness.

Discussion Qualitative matrices showed that the waterbird metacommunity was highly nested, both during the wintering and breeding seasons, but nestedness was higher during the breeding season. The metacommunity also showed increasing trends of nestedness over time, both during wintering and breeding. Nestedness however was not found when using quantitative matrices. The difference between qualitative and quantitative results suggests that compositional and density changes do not follow the same structural rules. Wetlands can recover lost species but not necessarily their abundances need to remain lower in the wetland gaining the species, compared to the donating wetland. Hence presence/ absence has little to do with densities. Compared to the results for the waterbird community in completely artificial wetlands (i.e. irrigation ponds) [13] our negative z-scores were higher indicating a higher nestedness of our study system involving mostly natural wetlands or former natural wetlands transformed by human action such as salt-pans.

Selective extinction and selective colonization The theory of island biogeography [45], [46] predicts that a fragmented habitat tends to lose species as its size decreases, and that colonization decreases as a direct function of patch isolation, although there are some exceptions (see e.g. [47]). Our results suggest that selective extinction was the most likely historical cause

Processes generating nestedness According to some authors [11] a high nestedness in a metacommunity can be the result of several causes. False negatives in the qualitative matrix can be due to imperfect detection since PLOS ONE | www.plosone.org

8

August 2014 | Volume 9 | Issue 8 | e105202

Long-Term Recovery of an Avian Metacommunity

PLOS ONE | www.plosone.org

9

August 2014 | Volume 9 | Issue 8 | e105202

Long-Term Recovery of an Avian Metacommunity

Figure 5. Beta-, alpha- and gamma-diversity over time (1990–2011). Trend of b-diversity over time for A) breeding and B) wintering; adiversity over time for C) breeding and D) wintering, and c-diversity over time for E) breeding and F) wintering. Solid lines are the lines of best fit and dotted lines are the 95% confidence intervals. doi:10.1371/journal.pone.0105202.g005

determinants of b-diversity [50], [51] (but see [52], [53]). Our analysis of the change in number of sites occupied by each species indicated that 50% of the species expanded geographically over the study period (i.e. secondary colonization or/and immigration). Frequent colonization is likely to enhance nestedness [23], as it reduces the number of unexpected absences.

generating nestedness in our waterbird metacommunity, from the original situation in which the study area roughly formed a continuous to the highly fragmented pattern of today (see [48]). This result is consistent with the findings done by other authors working with waterbirds [13] since they detected pond sizedependent selective extinction as the main cause of nestedness in artificial wetlands. Selective colonization did not play a relevant role in creating nestedness in our study system most likely due to the highly vagile nature of the study group (i.e. birds with a high colonization capacity) (see e.g. [49]). Selective extinction in our study system could be related to two factors, either a) wetlands lose species as they become smaller because population size decreases below the minimum viable population size [9]. This may lead to deterministic Allee effects (i.e. deterministic problems in finding food, mates or defence against predators at low densities). Also the different species could be forced to using similar resources as wetland size decreases, and hence deterministic competitive exclusion among species might take place. Finally, demographic stochasticity could lead to loosing species just by random changes in demography (i.e. random changes in vital rates such as fecundity or survival) as wetland size and, in turn, overall population size, shrinks. Or b) wetlands lose habitat heterogeneity (i.e. nested habitat hypothesis) and hence species associated to those habitats. As already-stated we do not have detailed information on habitat type presence and abundance for each study wetland, and thus we cannot rule this factor out. The most vulnerable bird groups to reduction in wetland size were herons, divers and gulls in the breeding season and ducks and shorebirds during wintering (Figure S1).

Conservation implications The increase in nestedness over time could be initially interpreted as a negative result from a conservation viewpoint because it means increasing the biotic homogenization of the system (by losing b-diversity) [54]. However, it also has a positive interpretation. By increasing nestedness the system is showing a high resilience to recover from historical fragmentation and perturbation after only two and a half decades of legal site protection. Increased nestedness also leads to gaining overlap among wetland biotas and hence probably to increased resistance and resilience against perturbations, as the system becomes more and more redundant [23]. Thus losing one of the species in a site is not so relevant for the whole metacommunity, as it can be recovered by reshuffling of local species (i.e. secondary colonization). Hence, in summary, we can conclude that our study system is becoming more and more homogenized because of species expansion. These results may suggest that the regional system of protected wetlands studied is showing some positive results, despite the degree of fragmentation has remained approximately unchanged and extensive work remains to be done for the full recovery of water quality and habitat heterogeneity. It is a fact that this system was in a very impoverished state at the beginning of our study period (1980s) according to the rich composition of its avian communities up to the 1970s (see e.g. [55]), and hence we are most likely observing a recovery of the original metacommunity by immigration and also by range expansion of local species, during the last decades, following some improvement in environmental conditions and reduction in human pressure. This recovery of the metacommunity is likely due not only to the local protection of sites, but also to the improving conditions in wetlands outside the study system, at the regional, national and trans-national levels [56], [57], [58], [59]. Additionally, the metacommunity has gained some species by means of reintroduction programmes (i.e. Redknobbed Coots and Purple Swamphen) and probably due to increasing temperatures at the regional level, because former migrating species during the winter now remain in our study sites; clear examples are Little Tern, Squacco Herons, Black-crowned Night Heron and Black-Winged Stilt. Hence the study system is not any more within the stage of ecological relaxation (i.e. gradual losing of species by increased fragmentation). Obviously, the fact of dealing with a highly vagile animal group makes the recovery of the whole system (covering several hundred kilometres in length) more viable. However, not all bird groups contributed equally to homogenization (see e.g. [7]). During the breeding season, shorebirds, gulls and herons comprised 29%, 25% and 25% respectively of the species performing poorly in the sense of lack of expansion. In winter shorebirds and ducks represented 43% and 30% of the species not under expansion. Within groups 87% of the shorebird species were not expanding in breeding and 73% of the duck species in winter. This suggests scarcity of suitable breeding habitat during the summer for shorebirds and a poor water quality for

Between seasons variability in nestedness We found a solid difference between the compositional structure of our breeding and wintering communities, with a higher nestedness always taking place during the breeding season, regardless of the metric used (temperature or NODF). This result is coincident with the structure found in artificial wetlands located in the southern tip of our study region [13]. It is likely that the process of selective extinction is affecting more heavily waterbird species during the reproductive season. This may be so because habitat requirements are probably more demanding during breeding than wintering, because of the need of getting resources for both parents and offspring, especially in the Mediterranean region, where the highest temperatures of the annual cycle coincide with the lowest precipitations. Breeding birds need appropriate nesting habitat, quietness and enough food of high quality for their offspring.

Inter-annual variability in nestedness We found an increasing trend of nestedness in the waterbird metacommunity over time confirmed through a decrease in bdiversity in both seasons [40]. The main reason why b-diversity decreased was probably by the fact that a-diversity (the local number of species in each wetland) also increased over time due to species reshuffling among sites generating the pattern of increased nestedness with time. Actually b-diversity did not decrease faster due to increased nestedness because we also found an increasing trend in c-diversity in our open system, that is gaining species from outside by immigration (e.g. Spoonbill, Great Egret, Glossy Ibis). That could shed some light to the current debate on the PLOS ONE | www.plosone.org

10

August 2014 | Volume 9 | Issue 8 | e105202

PLOS ONE | www.plosone.org

11

20.850

20.682

20.775

20.370

20.679

20.631

20.560

20.451

20.544

20.648

20.679

20.857

20.811

20.746

20.888

20.777

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

20.035 20.123

14 16

20.007 20.083

16 18

0.116

15 0.154

0.147

15

15

14

0.040

18

16

15

15

15

13

0.196

13

13

13

14

14

0.339

13

13 20.036

15

0.030 14

10

0.049

10 15

15

20.030

15

16

15

14

15

10

16

16

16

11

20.109

20.136

15

10 12

16

20.119

20.419

14

15

0.482

0.261

10

16

20.001

16 0.133

0.342

11

16

20.067 0.270

10 12

20.885

20.521

20.578

20.543

20.600

20.644

20.718

20.684

20.534

20.121

20.647

20.850

20.771

20.903

20.809

20.824

20.723

20.688

20.650

20.716

20.652

20.664

20.687

20.745

16

16

17

15

15

16

0.069

0.077

20.264

20.211

20.086

20.141

20.158

0.064

14 16

20.002

20.192

20.192

20.157

20.071

20.215

0.077

20.094

0.024

14

15

16

17

16

16

16

13

14

0.156

20.219 14

20.716 17

20.129

0.081

0.405

0.256

0.267

0.256

20.290

20.236

16

16

17

15

15

16

16

14

14

15

16

17

16

16

16

13

14

14

17

17

17

15

13

10

10

13

13

n

Distance to the nearest wetland (km) Coefficient

17

17

15

13

10

10

13

13

n

We show in bold the statistically significant results. N = Number of rows which were used for the correlation (for the overall matrices 18 rows were used for all correlations). No data provided in the breeding season of 1984 and wintering season of 1986 for the qualitative matrix because less than 10 rows of those ordered matrices contained data. doi:10.1371/journal.pone.0105202.t003

20.837

20.611

1994

1995

20.650

20.603

1991

20.564

20.621

1990

1993

20.764

1989

1992

20.776

20.418

20.309

10

1988

0.268

1987

10

20.370

11

1986

0.123 20.670

11

20.655

20.769

1985

0.068 20.841

20.767

Wetland size (ha) Coefficient

Coefficient

n

Distance to the nearest wetland (km) Coefficient

Wetland size (ha) n

Wintering season

Breeding season

1984

Summary matrix

Year

Table 3. Spearman rank correlation coefficients (rs) for the summary matrices and annual variability of correlations between row order for the matrices packed by BINMATNEST, and wetland size as well as distance to the nearest wetland.

Long-Term Recovery of an Avian Metacommunity

August 2014 | Volume 9 | Issue 8 | e105202

Long-Term Recovery of an Avian Metacommunity

wintering ducks, especially for diving species dependent on submerged vegetation. Both of these matters (water quality and habitat heterogeneity) are the key factors to be improved in the near future to allow the full recovery of the former waterbird metacommunity. However, since immigration from outside the system also plays a role, the temporal trends of breeding shorebirds and wintering ducks should be explored at large geographical scales to make sure that the lack of local recovery of these groups is not only due to poor suitability of the study wetlands for them but also to larger-scale problems either in Africa or central and northern Europe (see e.g. [60]).

WNODF is in each case the average of 1000 Monte Carlo simulations run in NODF. SD = standard deviation of simulated WNODF. CI = 95% confidence interval of simulated WNODF. Z’ = standardized effect size of WNODF (see text). (DOC) Table S2 Order of nestedness of the overall qualitative matrix for breeding and wintering season. The order of nestedness is according to the degree of nestedness packed by BINMATNEST. (DOCX)

Acknowledgments

Supporting Information

We are grateful to Anna Traveset, Alicia Montesinos, Claudia MettkeHofmann and three anonymous referees for providing helpful comments to a draft of the manuscript. We are also thankful to our partners at the Population Ecology Group in Mediterranean Institute for Advanced Studies for their critical and constructive comments. This study would have not been possible without the anonymous work of a large number of dedicated persons who carried out the bird counts over the years. Special thanks are due to SEO-BirdLife for many years of dedicated waterbird counts in the study area. We are also thankful to Catherine Andre´s-Langa for building Figure 1. Raw field data (tables of counts for each species and year) can be supplied by the authors upon request.

Loosing species in relation to wetland size reduction. Species loss by zoological groups in relation to wetland size reduction for both A) breeding and B) wintering season. (TIF)

Figure S1

Table S1 Analysis of the annual variability of nested-

ness of the waterbird metacommunity studied during breeding and wintering. Qualitative Matrix: simulated T/ NODF is in each case the average of 1000 Monte Carlo simulations run in ANINHADO. SD = standard deviation of simulated T. CI = 95% confidence interval of simulated T/ NODF. Z = standardized effect size of T (see text). NODFr = Relative NODF (see text). Values in bold are statistically significant results (the observed temperature/NODF is not within the 95% confidence intervals). Quantitative Matrix: simulated

Author Contributions Conceived and designed the experiments: AMA. Analyzed the data: JP AMA. Contributed reagents/materials/analysis tools: JAG JJ. Wrote the paper: JP AMA. Supervised the project and provided relevent review of the manuscript: DO.

References 16. Bloch CP, Higgins CL, Willig MR (2007) Effects of large-scale disturbance on metacommunity structure of terrestrial gastropods: temporal trends in nestedness. Oikos 116: 395–406. 17. Lomolino MV (1996) Investigating causality of nestedness of insular communities: selective immigrations or extinctions? Journal of Biogeography 23: 699– 703. 18. Cam E, Nichols JD, Hines JE, Sauer JR (2000) Inferences about nested subsets structure when not all species are detected. Oikos 91: 428–434. 19. Ferna´ndez-Juricic E (2002) Can human disturbance promote nestedness? A case study with breeding birds in urban habitat fragments. Oecologia 131: 269–278. 20. Rossello´ V, Panareda JM, Pe´rez A (1998) Manual de geografı´a fı´sica. Universitat de Vale`ncia. 21. Atmar W, Patterson BD (1993) The measure of disorder in the distribution of species in fragmented habitats. Oecologia 96: 373–382. 22. Boecklen WJ (1997) Nestedness, biogeography theory, and the design of nature reserves. Oecologia 112: 12–142. 23. Cook RR, Quinn JF (1995) The influence of colonization in nested species subsets. Oecologia 102: 413–424. 24. Jonsson BG (2001) A null model for randomization tests of nestedness in species assemblages. Oecologia 127: 309–313. 25. Patterson BD (1987) The principle of nested subsets and its implications for biological conservation. Conservation Biology 1: 323–334. 26. Gonza´lez-Garcı´a R, Pe´rez-Aranda D (2011) Las aves acua´ticas en Espan˜a, 1980–2009. Madrid: SEO/BirdLife. 340 p. 27. Martı´ R, Del Moral JC (Eds.) (2002) La invernada de aves acua´ticas en Espan˜a. Direccio´n general de Conservacio´n de la Naturaleza-SEO/BirdLife. Organismo Auto´nomo de Parques Nacionales, Ministerio de Medio Ambiente, Madrid. 28. Almeida-Neto M, Guimara˜es P, Guimara˜es PR, Loyola RD, Ulrich W (2008) A consistent metric for nestedness analysis in ecological systems: reconciling concept and measurement. Oikos 117: 1227–1239. 29. Guimara˜es PR, Guimara˜es P (2006) Improving the analysis of nestedness for large sets of matrices. Environmental Modelling & Software 21: 1512–1513. 30. R Development Core Team (2008) R: a Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Available: http://www.R-project.org. Accessed 20 February 2012. 31. Almeida-Neto M, Ulrich W (2011) A straightforward computational approach for measuring nestedness using quantitative matrices. Environmental Modelling & Software 26: 173–178. 32. Gotelli NJ, Entsminger GL (2001) EcoSim: Null models software for ecology, Version 7.0. Acquired Intelligence Inc. & Kesey-Bear. Available: http:// homepages.together.net/,gentsmin/ecosim.htm. Accessed 19 March 2012.

1. Patterson BD, Atmar W (1986) Nested subsets and the structure of insular mammalian faunas and archipelagos. Biological Journal of the Linnean Society 28: 65–82. 2. Dennis RLH, Hardy PB, Dapporto L (2012) Nestedness in islands faunas: novel insights into island biogeography through butterfly community profiles of colonization ability and migration capacity. Journal of Biogeography 39: 1412– 1426. 3. Greve M, Gremmen NJM, Gaston KJ, Chown SL (2005) Nestedness of Southern Ocean island biotas: ecological perspectives on a biogeographical conundrum. Journal of Biogeography 32: 155–168. 4. Bruun HH, Moen J (2003) Nested communities of alpine plants on isolated mountains: relative importance of colonization and extinction. Journal of Biogeography 30: 297–303. 5. Cutler A (1991) Nested faunas and extinction in fragmented habitats. Conservation Biology 5: 496–505. 6. Hylander K, Nilsson C, Jonsson BG, Go¨thner T (2005) Differences in habitat quality explain nestedness in a land snail meta-community. Oikos 108: 351–361. 7. Martı´nez-Morales MA (2005) Nested species assemblages as a tool to detect sensitivity to forest fragmentation: the case of cloud forest birds. Oikos 108: 634– 642. 8. Newmark WD (1987) A land-bridge island perspective on mammalian extinctions in western North American parks. Nature 325: 430–432. 9. Wang Y, Bao Y, Yu M, Xu G, Ding P (2010) Nestedness for different reasons: the distribution of birds, lizards and small mammals on islands of an inundated lake. Diversity and Distributions 16: 862–873. 10. Patterson BD (1990) On the temporal development of nested subset patterns of species composition. Oikos 59: 330–342. 11. Ulrich W, Almeida-Neto M, Gotelli NJ (2009) A consumer’s guide to nestedness analysis. Oikos 118: 3–17. 12. Paracuellos M, Tellerı´a JL (2004) Factors affecting the distribution of a waterbird community: the role of habitat configuration and bird abundance. Waterbirds 27: 446–453. 13. Sebastia´n-Gonza´lez E, Botella F, Paracuellos M, Sa´nchez-Zapata JA (2010) Processes driving temporal dynamics in the nested pattern of waterbird communities. Acta Oecologica 36: 160–165. 14. Bascompte J, Jordano P (2007) Plant-animal mutualistic networks: the architecture of biodiversity. Annual review of Ecology, Evolution and Systematics 38: 567–593. 15. Azeria ET, Kolasa J (2008) Nestedness, niche metrics and temporal dynamics of a metacommunity in a dynamic natural model system. Oikos 117: 1006–1019.

PLOS ONE | www.plosone.org

12

August 2014 | Volume 9 | Issue 8 | e105202

Long-Term Recovery of an Avian Metacommunity

48. Kopecky M, Hedl R, Szabo P (2013) Non-random extinctions dominate plant community changes in abandoned coppices. Journal of Applied Ecology 50: 79– 87. 49. Martı´nez-Abraı´n A, Oro D, Forero MG, Conesa D (2003) Modelling temporal and spatial colony-site dynamics in a long-lived seabird. Population Ecology 45: 133–139. 50. Baselga A (2012) The relationship between species replacement, dissimilarity derived from nestedness, and nestedness. Global Ecology and Biogeography 21: 1223–1232. 51. Carvalho JC, Cardoso P, Borges PAV, Schmera D, Podani J (2013) Measuring fractions of beta diversity and their relationships to nestedness: a theoretical and empirical comparison of novel approaches. Oikos 122: 825–834. 52. Almeida-Neto M, Ulrich W (2012) Rethinking the relationship between nestedness and beta diversity: a comment on Baselga (2010). Global Ecology and Biogeography 21: 772–777. 53. Ulrich W, Almeida-Neto M (2012) On the meanings of nestedness: back to the basis. Ecography 35: 865–871. 54. Baeten L, Vangansbeke P, Hermy M, Peterken G, Vanhuyse K, et al. (2012) Distinguishing between turnover and nestedness in the quantification of biotic homogenization. Biodiversity and Conservation 21: 1399–1409. 55. Bernis F (1964) Informacio´n espan˜ola sobre ana´tidas y fochas (e´poca invernal). Madrid: Sociedad Espan˜ola de Ornitologı´a. 154 p. 56. Donland PF, Sanderson FJ, Burfield IJ, Bierman SM, Gregory RD, et al. (2007) International conservation policy delivers benefits for birds in Europe. Science 317: 810–813. 57. Galewski T, Collen B, McRae L, Loh J, Grillas P, et al. (2011) Long-term trends in the abundance of Mediterranean wetland vertebrates: from global recovery to localized declines. Biological Conservation 144: 1392–1399. 58. Oro D, Pe´rez-Rodrı´guez A, Martı´nez-Vilalta A, Bertolero A, Vidal F, et al. (2009) Interference competition in a threatened seabird community: A paradox for a successful conservation. Biological Conservation 142: 1830–1835. 59. Rendo´n MA, Green AJ, Aguilera E, Almaraz P (2008) Status, distribution and long-term changes in the waterbird community wintering in Don˜ana, south-west Spain. Biological Conservation 141: 1371–1388. 60. Tucker GM, Heath MF (1994) Birds in Europe: their conservation status. Cambridge: BirdLife International (BirdLife Conservation Series no.3). 600 p.

33. Gotelli NJ, McCabe DJ (2002) Species co-occurrence: a meta-analysis of J. M. Diamond’s assembly rules model. Ecology 83: 2091–2096. 34. Montesinos-Navarro A, Segarra-Moragues JG, Valiente-Banuet A, Verdu´ M (2012) The network structure of plant-arbuscular mycorrhizal fungi. New Phytologist 194: 536–547. 35. Ulrich W, Gotelli NJ (2007) Null model analysis of species nestedness patterns. Ecology 88: 1824–1831. 36. Rodrı´guez-Girone´s MA, Santamarı´a L (2006) A new algorithm to calculate the nestedness temperature of presence-absence matrices. Journal of Biogeography 33: 924–935. 37. Jost L (2007) Partitioning diversity into independent alpha and beta components. Ecology 88: 2427–2439. 38. Baselga A (2010) Partitioning the turnover and nestedness components of beta diversity. Global Ecology and Biogeography 149: 134–143. 39. Fischer J, Lindenmayer DB (2005) Perfectly nested or significantly nested - an important difference for conservation management. Oikos 109: 485–494. 40. Wright DH, Reeves JH (1992) On the meaning and measurement of nestedness of species assemblages. Oecologia 92: 416–428. 41. Ricklefs RE, Miller GL (1999) Ecology, 4th edn. New York: WH Freeman and Company. 822 p. 42. Whittaker RH (1960) Vegetation of the Siskiyou Mountains, Oregon and California. Ecological Monographs 30: 280–338. 43. Wright DH, Patterson BD, Mikkelson GM, Cutler A, Atmar W (1998) A comparative analysis of nested subset patterns of species composition. Oecologia 113: 1–20. 44. Yu M, Hu G, Feeley KJ, Wu J, Ding P (2012) Richness and composition of plants and birds on land-bridge islands: effects of island attributes and differential responses of species groups. Journal of Biogeography 39: 1124–1133. 45. Losos JB, Ricklefs RE (2010) The Theory of Island Biogeography Revisited. Princeton, Oxford: Princeton University Press. 494 p. 46. MacArthur RH, Wilson EO (1967) The Theory of Island Biogeography. Princeton, Oxford: Princeton University Press. 203 p. 47. Simberloff D, Levin B (1985) Predictable sequences of species loss with decreasing island area - land birds in two archipelagos. New Zealand Journal of Ecology 8: 11–20.

PLOS ONE | www.plosone.org

13

August 2014 | Volume 9 | Issue 8 | e105202