Biodiversity Analysis of True Bug Assemblages (Hemiptera, Heteroptera) in Four Habitats in the Garraf Natural Park (Barcelona, Spain)

Abstract

With the aim of gaining insight into true bugs assemblages, and of giving tools to enhance environmental manager tasks, an analysis was made of the alpha and beta diversity of true bug assemblages in the NE Iberian Peninsula. The study took place during the years 1999, 2000, and 2001. The diversity of true bug assemblages among four typical Mediterranean plant communities (dry grassland, calcicolous rosemary scrub, kermes oak scrub, and evergreen oak forest) in a protected natural area was compared. In each plant community, a stratified sampling, taking into account plants with highest coverage percentage, was performed. Collections were performed monthly using direct observation plus beating or sweeping of the entomological net. Objectives were 1) to assess whether true bug assemblages are different in each plant community surveyed, by means of the analysis of their alpha and beta diversity, 2) to assess if any of the four true bug assemblages may be prioritized according to the biodiversity parameters studied. In total, 3,071 specimens, belonging to 12 families and 77 species were identified. Fifty percent of specimens collected belonged to only three species, and ∼33% of species was represented by only one or two individuals. In each Heteroptera assemblage, the distribution pattern of species frequency classes followed a lognormal model. Thirty-five percent of species represented by three or more specimens were found in only one plant community. Abundance, species richness, and diversity were found to be different for each true bug assemblage and along the year. Abundance was highest in the evergreen oak forest and in spring. Species richness and diversity were highest in the kermes oak scrub and in early summer. Nonparametric species richness estimators showed that completeness of species inventory was >80%. In the study zone, true bug diversity may be considered low medium. Our results show that true bugs assemblages are characteristic of specific plant communities, and thus, true bugs fit as candidate monitoring group in environmental and conservation management.

Insect biodiversity is very important in ecosystems, and their unique biological traits and ecological roles make them valuable organisms for assessing, monitoring, and managing biodiversity in natural ecosystems. Site-specific biotic and abiotic characteristics determine insect biodiversity to be specific of a particular site. Natural resource agencies need baseline, site-specific biodiversity data that should be a basic guide for all other resource management programs ( Kim 2009 ).

The Heteroptera represent the largest and most diverse group of hemimetabolous insects ( Schuh and Slater 1995 ), with ∼42,000 described species ( Henry 2009 ). Heteroptera may be found in a large variety of niches, where they perform a range of ecological functions and services, thus affecting nearly every aspect of our environment ( Henry 2009 ). Terrestrial bugs are associated to plants, and floristic composition and vegetation structure are the environmental factors, which explain the best true bug biodiversity and distribution patterns. Plant vegetation communities, characterized by their floristic composition and vegetation structure, are a good study unit of true bug assemblages.

The importance of Heteroptera in the ecosystems makes them useful in conservation biology ( Henry 2009 ). However, very few ecological studies concerning true bug assemblages are available to date. In the Mediterranean region, some works relate true bug species to sampling locality or plant community ( Rizzotti Vlach 1994 ; Bacchi and Rizzotti Vlach 1997 , 1998 ), whereas others perform some quantifiable or statistical analysis on true bug assemblages ( Ribes et al. 2000 , 2001 ; Goula et al. 2010 ). Ecological importance of Heteroptera in ecosystems and reliable species identification for true bug Mediterranean fauna due to the availability of good and updated identification keys give ground to develop biodiversity research on true bug assemblages at a local scale providing the baseline information needed by nature resource managers to support nature conservation policy.

Garraf Natural Park (NE Iberian Peninsula) is a very dry area, composed by a typical Mediterranean landscape, in which green oak forest should be the climax vegetation type. However, frequent wildfires enhance vegetation types preceding the development of green oak forest (i.e., dry pastures, shrubs, and scrubs) to be commonly found within the Park, in which a mosaic landscape is observed. Inclusion of data on ecosystem relevant true bug biodiversity may help Responsible for Park to take better grounded management decisions.

On the basis of true bug species association to plants, and of their response to floristic composition and vegetation structure, it is expected that alpha and beta diversity of true bug assemblages will be different when studying different vegetation types. To assess this hypothesis, the goals of the research were 1) to describe and compare alpha and beta diversity of Heteroptera assemblages, collected in dry pasture, calcicolous rosemary shrub, kermes oak scrub, and green oak forest, the four plant communities characterizing the ecological succession stages in the evergreen oak forest series in dry Mediterranean habitats; and 2) to assess if any of the four true bug assemblages may be prioritized according to their biodiversity parameters. The achievement of those goals will provide local environment managers with tools for a better decision making, concerning whether any of the vegetation types considered deserves special conservation measures according to the conservation value of their true bug assemblages.

Materials and Methods

Description of the Study Area

The Heteroptera assemblages were studied in the Garraf Natural Park (41° 17′29.15″ N. 01° 52′30.06″ E), extending 10,638 ha within the Garraf massif, which is located in the Catalan Coastal Ranges ( Fig. 1 ). The study was conducted in a karstic area with limestone bedrock, in which surface water is limited. The soil unit corresponds to Calcisols ( (FAO) Food and Agriculture Organization 2006 ). The area belongs to the Medium Mesomediterranean ( Rivas-Martínez et al. 2002 ) bioclimate, characterized by average temperatures from 13 to 17°C. In the area, annual average rainfall ranges from 550 to 730 mm ( Miño 1986 ), which mainly concentrates in March–May and in September–October. From the phytogeographical point of view, the Garraf Natural Park belongs to the Mediterranean region ( Bolòs 1987 ). Their climax vegetation corresponds to the Medium Mesomediterranean bioclimate and belongs to the Landscape Unit 21: Catalan and Valencian Coastal Ranges, subunit 21.05: Garraf and Ordal Range ( Sanz et al. 2003 ).

Fig. 1.

Location of area of study, Garraf Natural Park (Barcelona, Spain).

Garraf Natural Park is under drought Mediterranean stress, and wildfires easily occur. Five large fire episodes have taken place in the Garraf area in the last 50 yr. Burnt area varied from 40 ha (2001) to 10,000 ha (1982). In the last 13 yr (data obtained from unpublished “Diputació de Barcelona” fire prevention program annual reports, between years 2000 and 2013; J. Torrentó, personal communication), there has been a sum of 70 forest fires declared in the natural reserve area, burning a total of 360 ha, with an average ratio of 27.7 ha/yr. Most of these events affected <1 ha each one. Recurrent fires in the Garraf natural reserve have severely damaged the vegetation cover and consequently both habitats and fauna. On the other hand, the Park is under a very strong pressure due to human frequentation, as it is within the Barcelona metropolitan region, with a population ∼3 million people ( Generalitat de Catalunya 2010 ). In natural protected areas, human frequentation is a threat to conservation, and management plans have to make compatible human educational and recreational uses with natural ecosystem services.

Description of Plant Communities

The Garraf Massif is the border between the Boreomediterranean evergreen oak forest of Quercus ilex L. subsp. ilex and the Austromediterranean scrubland of Quercus coccifera L. and Pistacia lentiscus L. Because of anthropogenic pressure and common wildfires, potential vegetation is replaced by seral plant communities, including dense to light scrubs, Aleppo pine ( Pinus halepensis Miller) woodland, and dry grasslands ( Hoyo 1992 ).

Four plant communities were selected: dry grassland, calcicolous rosemary scrub, kermes oak scrub, and evergreen oak forest. These communities are characterized in Table 1 , by means of their plant species abundance, and following Vigo et al. (2005) . Dry grassland ( Fig. 2 ) is an open, herbaceous to subshrubby community that barely reaches 25 cm in height. The calcicolous rosemary scrub ( Fig. 3 ) and kermes oak scrub ( Fig. 4 ) bushes reach 0.5–1.0 m height; the first is light and patchy, mostly formed by rosemary ( Rosmarinus officinalis L.) and restricted to calcium-rich substrata, whereas the second is more dense and uniformly dominated by kermes oak ( Q. coccifera ). The evergreen oak forest ( Fig. 5 ) is mainly composed by a tree canopy of holm oak ( Q. ilex ) above 2 m height and a shrubby understory (of Buxus sempervirens L. ). Only homogeneous plant formations excluding disturbed patches were selected. For each plant community, three plots were taken as replicates. The size, altitude, and localization of these plots are summarized in Table 2 . Plot size (10 by 10 m in dry grassland, calcicolous rosemary scrub, and kermes oak scrub, and 20 by 20 m in the evergreen oak forest) correspond to those used in other vegetation studies ( Mueller-Dombois and Ellenberg 1974 ).

Table 1.

Size and location of plant community plots

Plant community	Plot measures (m)	Height (m a.s.l)	UTM (ED50)
Dry grassland	10 by 10	350	31TDF0771
Calcicolous rosemary scrub	10 by 10	425	31TDF0670
Kermes oak scrub	10 by 10	500	31TDF0971
Evergreen oak forest	20 by 20	400	31TDF0375

ED, European Datum 1950.

Fig. 2.

Dry grassland.

Fig. 3.

Calcicolous rosemary scrub.

Fig. 4.

Kermes oak scrub.

Fig. 5.

Evergreen oak forest.

Table 2.

Plant species and plant individuals studied in the four vegetation communities

Family	Scientific (Latin) name	Common name	Dry grassland (D)		Calcicolous rosemary scrub (C)		Kermes oak scrub (K)		Evergreen oak forest (E)
			Number of plants	Plant individuals studied	Number of plants	Plant individuals studied	Number of plants	Plant individuals studied	Number of plants	Plant individuals studied
Poaceae	Ampelodesmos mauritanica (Poiret) Durand & Schinz	Diss grass	25	2	27	2
Ericaceae	Arbutus unedo L.	Strawberry tree							5	0
Liliaceae	Asparagus acutifolius L.	South European asparagus							42	2
Aspleniaceae	Asplenium adiantum- nigrum L.	Black oak-fern							5	0
Poaceae	Brachypodium retusum (Persoon) Beauvois	Mediterranean False-brome	437	9	185	5	455	9	10	0
Buxaceae	Buxus sempervirens L.	Common box							728	8
Arecaceae	Chamaerops humilis L.	Dwarf fan palm	6	0	2	0	12	0
Cistaceae	Cistus albidus L.	White-leaved rock rose					15	2
Cistaceae	Cistus salvifolius L.	Salvia cistus	2	0	7	0
Ranunculaceae	Clematis flammula L.	Sweet-scented virgin's bower							10	0
Thymelaeaceae	Daphne gnidium L.	Flax-leaved daphne					2	0
Asteraceae	Dorycnium hirsutum (L.) Ser.in DC.	Hairy canary clover					9	0
Asteraceae	Dorycnium pentaphyllum Scopoli	Spear pea					18	2
Ericaceae	Erica multiflora L.	Many-flowered heath			32	2
Euphorbiaceae	Euphorbia characias L.	Mediterranean Spurge							3	8
Euphorbiaceae	Euphorbia flavicoma DC.	Flavicoma spurge					34	2
Rubiaceae	Galium lucidum L.	Shining Bedstraw	24	2
Geraniaceae	Geranium robertianum L.	Herb robert							7	0
Plantaginaceae	Globularia alypum L.	Alypo globe daisy			19	2
Araliaceae	Hedera helix L.	Common ivy							12	0
Cupressaceae	Juniperus oxycedrus L.	Brown-berried juniper			4	0			3	0
Cupressaceae	Juniperus phoenicea L.	Phoenicean Juniper							33	2
Caprifoliaceae	Lonicera implexa Aiton	Minorca honeysuckle							80	2
Fabaceae	Ononis minutissima L.	Pygmy restharrow	6	0
Oleaceae	Phillyrea latifolia L.	Jasmine box			11	0			73	2
Lamiaceae	Phlomis lychnitis L.	Lampwick plant	15	2			1	0
Pinaceae	Pinus halepensis Miller	Aleppo pine							19	0
Anacardiaceae	Pistacia lentiscus L.	Lentisk	35	2	50	2	64	3	96	2
Polypodiaceae	Polypodium vulgare L.	Ever fern							28
Rosaceae	Prunus mahaleb L.	St Lucie cherry							10	0
Lamiaceae	Psoralea bituminosa L.	Treacle clover					3	0
Fagaceae	Quercus coccifera L.	Kermes oak	13	0			450	9	24	0
Fagaceae	Quercus ilex L.	Evergreen oak							749	8
Rhamnaceae	Rhamnus lycioides L.	Black hawthorn			1	0
Lamiaceae	Rosmarinus officinalis L.	Rosemary			199	5			37	2
Rubiaceae	Rubia peregrina L.	Evergreen clover	1	0			9	0	23	0
Liliaceae	Ruscus aculeatus L.	Kneeholy							187	3
Crassulaceae	Sedum sediforme Jacquin	Pale stonecrop	6	0					2	0
Smilacaceae	Smilax aspera L.	Prickly ivy	3	0					205	3
Poaceae	Stipa offneri Breistr.	Esparto grass			84	3
Lamiaceae	Teucrium chamaedrys L.	Ground oak	6	0
Lamiaceae	Thymus vulgaris L.	Common thyme	36	2	61	3	3	0	3	0
	Total individuals		615	19	682	24	1,075	27	2,394	34
	Total plant species		6		8		6		10

Data Collection

Heteroptera were collected following a stratified sampling methodology according to the relative average plant coverage in each plant community, expressed as percentage. Using transects, in each plot plants were identified, and the average coverage percentage of each plant species was calculated. Percentages ranged from 75.8 to 0.2%. As sampling was planned to be performed in individual plant species, a presence of those plants species along the 3-yr field survey and in each of the three plots of a given plant community was desirable. Only plant species with a percentage coverage higher than 2.5% assuring a good representation within the community fit those requirements and were thus considered. In all, 22 plant species were surveyed out of the total 41 species in the four plant communities. The number of individuals to be sampled of a plant species in each vegetation type was established according to the coverage percentage of that plant species in that vegetation type: two (2.5–10%), three (11–20%), four (21–30%), five (31–40%), six (41–50%), seven (51–60%), eight (61–70%), and nine (71–80%). Plant species to be sampled in each plant community are summarized in Table 1 . The different number of samples in each vegetation type will determine data analysis. On the basis that true bugs are associated to floristic composition and structure of vegetation, the use of plant coverage, that is the basis of phytosociology to define vegetation units, seemed to be a good approach to plan the sampling of plant species to collect true bugs living on them. Whether this new approach fits the objectives of our research or is advisable for conservation purposes is discussed. In all, 312 samples were collected monthly, with a total 11,232 samples in the 36 months of study.

Field work was carried out during the years 1999, 2000, and 2001 on a monthly basis. Number of sampling days was the same for all plant communities. Exploration of plots was at random. Terrestrial Heteroptera are most frequently found on the leaves of plants, thus foliage exploration is the most widespread collecting method used for terrestrial Heteroptera studies. In this study, beating, sweeping, and direct observation on the vegetation were performed, as most productive techniques for terrestrial true bug collection ( Coscarón et al. 2009 ), and also, as the best procedure to relate bug species to their host plant species. After beating and sweeping, the specimens fall into a clap net, were collected with a sucking flask, and preserved in glass vials with cork with ethyl acetate. On herbaceous plants, a sample was the sum of true bugs observed with the naked-eyed plus those obtained from the same plant species after three sweepings of the entomological net. In trees or bushes, a sample was composed of the specimens collected after beating branches three times. Specimens that could not be identified directly in the field were brought back to the laboratory in ethyl acetate and identified to species level with the aid of the stereomicroscope. As long as they could be identified to species level, nymphs were also included in this study.

Data Analysis

Abundance was compared between true bug assemblages. As abundance is influenced by sampling effort, which in our study is not the same in the four plant communities, the abundance per sample was chosen as comparative basis to perform statistical analysis.

Alpha and beta diversity were measured. The spatial scale used refers to the following diversity levels: alpha = sampled site/vegetation type; beta = turnover among vegetation types.

Alpha Diversity

Alpha diversity was evaluated by the analysis of species richness, abundance model, and diversity of true bug assemblages. The most simple species richness value in a sampled site is the absolute number of species in that site. However, the number of species depends on sampling effort ( Gaston 1996 ). To compensate for different sampling effects, we measured species richness of each true bug assemblage through Margalef’s diversity index, of easy calculation and intuitively meaningful ( Magurran 2004 ). To better evaluate the significance of species richness differences between plant assemblages, monthly mean values of Margalef’s index were calculated on which statistical analysis were run. To assess the completeness of the inventory of each sampling site, the observed species richness was compared with the expected species richness according to the Bootstrapping estimator. To provide a common abundance level on which to make comparisons of the species richness, rarefaction curves were calculated. In our research, the application of the same sampling techniques, and the focus on only one insect type, fit the requirements of rarefactions application ( Magurran 2004 ). To determine the sampling effort and improve performance, logarithmic interpolation of the number of species with respect to the minimum number of individuals observed was calculated. Estimations used aggregate data (one row per species) and discarded variability between samples. Fit between empirical and estimated data was assessed by means of R² values. An estimated rarefaction curve of all assemblages using logarithmic interpolation and the equation y  = 15.008ln( x ) − 46.467 ( R² = 0.9673) was obtained. Nonparametric estimators (bootstrap, rarefaction) were calculated with 100 resamplings, in the case of each plant community bug assemblage or 50 resamplings in the case of whole bug assemblage.

The structure of true bug assemblages was assessed by analyzing the distribution pattern of species frequency classes and by the calculation of Shannon diversity index on base e , H ’. Also monthly means of Shannon’s index e ^H’ as basis for statistical analysis to overcome the difficulty of giving significance of direct comparison of the index were used ( Magurran 2004 ).

Alpha diversity analysis was performed with the Biodiversity (V2) professional software ( McAleece et al. 1997 ). MS Excel was used to calculate specific richness and to numerically interpolate (nonlinear regression) using the rarefaction curves. The software PAST (V18) ( Hammer et al. 2001 ) was used to calculate total rarefaction and the 95% confidence interval. EstimateS (V8.2) was used to calculate nonparametric richness estimators ( Colwell 2009 ).

Beta Diversity

Beta diversity (i.e., true bug biodiversity turnover between plant communities) was assessed by measuring the differences of species composition of true bug assemblages among plant communities. Correspondence analysis (CA) was used to compare the similarity of the bug species assemblages in the four vegetation types. To elaborate the cluster analysis, the Bray–Curtis index ( Bray and Curtis 1957 ) was employed.

Beta biodiversity analysis was performed with the Biodiversity (V2) professional software ( McAleece et al. 1997 ).

Statistical Analysis

Analysis of variance (ANOVA) was used to assess differences between mean abundance of biotopes and to contrast paired monthly means for specific richness indices calculated (Margalef, Shannon, and Williams’ Alpha). As the Heteroptera abundance and richness does not fit a normal distribution according to the Kolmogorov’s test ( P  < 0.001), and have no homocesticity between groups, for comparative analysis, the nonparametric Mann–Whitney ANOVA test for abundance and Wilcoxon ANOVA test for richness were used ( P  < 0.05). The SPSS statistical software (V17) was used to describe and statistically analyze the differences between biotopes and between months.

Results

Alpha Diversity

In total, 3,071 specimens were identified, corresponding to 77 species in 12 families ( Table 3 ).

Table 3.

Absolute and mean abundance of Heteroptera species collected in dry grassland, calcicolous rosemary shrub, kermes oak scrub, and evergreen oak forest

	Dry grassland		Calcicolous rosemary shrub		Kermes oak scrub		Evergreen oak forest
	n	Mean ± SE	n	Mean ± SE	n	Mean ± SE	n	Mean ± SE
Tingidae
Hyalochiton colpochilus (Horváth 1897)	43	20.96  ± 5.39
Tingis (Tropidocheila) alberensis Péricart 1979	1	0.49  ± 0.49
Tingis (Tropidocheila) trichonota (Puton 1874)	441	214.91  ± 61.10
Microphysidae
Loricula ruficeps (Reuter 1884)					1	0.34  ± 0.34
Miridae
Deraeocoris ( Camptobrochis ) serenus Douglas & Scott 1868			2	0.77  ± 0.77	1	0.34  ± 0.34
Deraeocoris ( Knightocapsus ) lutescens (Schilling 1836)							2	0.54  ± 0.39
Macrolophus costalis Fieber 1858					6	2.06  ± 0.97
Campyloneura virgula (Herrich-Schäffer 1835)							3	0.82  ± 0.47
Phytocoris ( Exophytocoris ) fieberi Bolívar 1881							10	2.72  ± 1.09
Phytocoris ( Compsocerocoris ) sanctipetri Carapezza 1985							1	0.27  ± 0.27
Phytocoris ( Compsocerocoris ) juniperi Frey-Gessner 1865			25	9.65  ± 2.34
Phytocoris ( Ktenocoris ) vittiger Reuter 1896					1	0.34  ± 0.34
Phytocoris ( Ktenocoris ) varipes Boheman 1852			1	0.39  ± 0.39	3	1.03  ± 0.59
Phytocoris ( Ktenocoris ) flammula Reuter 1875	55	26.80  ± 5.88	7	2.70  ± 1.16	1	0.34  ± 0.34
Calocoris trivialis (A. Costa 1852)					6	2.06  ± 1.28	5.6	156.86  ± 13.54
Hadrodemus m-flavum (Goeze 1778)	1	0.49  ± 0.49
Dichrooscytus nanae Wagner 1957							1	0.27  ± 0.27
Taylorilygus apicalis (Fieber 1861)			3	1.16  ± 0.86	1	0.34  ± 0.34
Lygus maritimus (Wagner 1949)	2	0.97  ± 0.69			1	0.34  ± 0.34
Orthops (Orthops) kalmii (L. 1758)							1	0.27  ± 0.27
Pinalitus cervinus (Herrich-Schäffer 1842)							56	15.25  ± 2.57
Camptozygum aequale (Villers 1789)							1	0.27  ± 0.27
Capsodes flavomarginatus (Donovan 1798)					35	12.00  ± 3.04	8	2.18  ± 0.77
Strongylocoris cicadifrons Costa 1852					6	2.06  ± 0.97
Heterotoma diversipes Puton 1876					1	0.34  ± 0.34	3	0.82  ± 0.47
Brachynotocoris parvinotum (Lindberg 1940)							8	2.18  ± 0.94
Orthotylus (Pachylops) virescens (Douglas & Scott 1865)					3	1.03  ± 1.03
Orthotylus (Pinocapsus) gemmae Gessé and Goula 2004							6	1.63  ± 0.77
Mimocoris rugicollis (Costa 1852)					6	2.06  ± 1.45	14	3.81  ± 1.15
Chlamydatus ( Eurymerocoris ) evanescens (Boheman 1852)	1	0.49  ± 0.49
Criocoris piceicornis Wagner 1950	3	1.46  ± 0.84
Heterocapillus tigripes (Mulsant 1852)					51	17.49  ± 4.50
Heterocapillus validicornis (Reuter 1876)					45	15.43  ± 3.37
Compsidolon ( Chamaeliops ) crotchi (Scott 1870)	1	0.49  ± 0.49	532	205.25  ± 22.96			211	57.46  ± 8.87
Psallus ( Psallus ) aurora (Mulsant & Rey 1852)							3	0.82  ± 0.47
Psallus ( Phylidea ) dichrous Kerzhner 1962					10	3.43  ± 1.88	3	0.82  ± 0.61
Pachyxyphus lineellus (Mulsant & Rey 1852)					106	36.35  ± 7.97
Nabidae
Himacerus ( Aptus ) mirmicoides (O. Costa 1831)					1	0.34  ± 0.34
Himacerus ( Anaptus ) major (A. Costa 1842)					1	0.34  ± 0.34
Nabis ( Tropiconabis ) capsiformis Germar 1838					1	0.34  ± 0.34
Anthocoridae
Anthocoris nemoralis (Fabricius. 1794)	21	10.23  ± 4.70			14	4.80  ± 1.75	70	19.06  ± 2.65
Orius ( Orius ) niger Wolff 1811			1	0.39  ± 0.39			2	0.54  ± 0.39
Brachysteles parvicornis (Costa 1847)					1	0.34  ± 0.34	4	1.09  ± 0.54
Cardiastethus fasciiventris (Garbiglietti 1869)							142	38.67  ± 4.43
Cardiastethus nazarenus Reuter 1884							3	0.82  ± 0.47
Reduviidae
Ploiaria putoni Noualhier 1895					3	1.71  ± 0.77
Rhynocoris cuspidatus Ribaut 1921	2	0.97  ± 0.69	1	0.39  ± 0.39	1	0.34  ± 0.34
Sphedanolestes sanguineus (Fabricius. 1794)							5	1.36  ± 0.72
Lygaeidae
Spilostethus pandurus pandurus (Scopoli 1763)	5	2.44  ± 2.44
Nysiuscymoides (Spinola 1837)			2	0.77  ± 0.55	3	1.03  ± 0.59
Kleidocerys ericae (Horvath 1909)			1	0.39  ± 0.39	1	0.34  ± 0.34	2	0.54  ± 0.39
Geocoris ( Piocoris ) erythrocephalus (Le Peletier & Serville 1825)					7	2.40  ± 1.03
Macroplax fasciata (Herrich-Schäffer 1835)					11	3.77  ± 1.33
Oxycarenus lavaterae (Fabricius. 1787)					1	0.34  ± 0.34
Heterogaster artemisiae Schilling 1829	20	9.75  ± 2.83	2	0.77  ± 0.55	2	0.69  ± 0.48
Scolopostethus decoratus (Hahn 1833)							2	0.54  ± 0.39
Thaphropeltus andrei (Puton 1877)			1	0.39  ± 0.39
Rhyparochromus tristis (Fieber 1861)					1	0.34  ± 0.34
Stenocephalidae
Dicranocephalus agilis (Scopoli 1763)					3	1.03  ± 0.59
Coreidae
Gonocerus acuteangulatus (Goeze 1778)							112	30.50  ± 3.60
Gonocerus insidiator (Fabricius. 1874)	57	27.78  ± 7.82	10	3.86  ± 1.33	4	1.37  ± 1.37	11	3.00  ± 0.98
Gonocerus juniperi (Herrich-Schäffer 1839)							12	3.27  ± 1.39
Enoplops scapha (Fabricius. 1803)					1	0.34  ± 0.34
Rhopalidae
Brachycarenus tigrinus Schilling 1829							1	0.27  ± 0.27
Rhopalus ( Rhopalus ) subrufus (Gmelin 1790)							1	0.27  ± 0.27
Liorhyssus hyalinus (Fabricius. 1794)	1	0.49  ± 0.49			3	1.03  ± 0.59
Myrmus miriformis (Fallen 1807)	24	11.70  ± 2.47	3	1.16  ± 0.67	71	24.35  ± 3.38
Plataspidae
Coptosoma scutellatum (Geoffroy 1785)					1	0.34  ± 0.34
Pentatomidae
Sciocoris ( Neosciocoris ) maculatus Fieber 1851	15	7.31  ± 3.41	6	2.31  ± 0.94			3	0.82  ± 0.61
Dyroderes umbraculatus (Fabricius. 1775)							1	0.27  ± 0.27
Staria lunata (Hahn 1834)					1	0.34  ± 0.34
Carpocoris mediterraneus atlanticus Tamanini 1958	1	0.49  ± 0.49
Brachynema germarii (Kolenati 1846)	1	0.49  ± 0.49
Rhaphigaster nebulosa (Poda 1761)							30	8.17  ± 1.53
Piezodorus lituratus (Fabricius. 1794)					3	1.03  ± 0.59	13	3.54  ± 1.41
Acrosternum millierei (Mulsant & Rey 1866)	18	8.77  ± 2.17	4	1.54  ± 0.77	1	0.34  ± 0.34	3	0.82  ± 0.47
Picromerus nigridens (Fabricius. 1803)			1	0.39  ± 0.39	3	1.03  ± 0.59	6	1.63  ± 0.67
Total specimens	713		602		425		1,331
Total species	20		17		42		36

Mean values are raw number of specimens per sample multiplied by 1,000.

Family Miridae outstand with 33 species (43% of total true bug species found), followed by Lygaeidae (10 species, 13%) and Pentatomidae (9 species, 11.7%). Microphysidae, Stenocephalidae, and Plataspidae were poorly represented (one species each, 1.3% of the total). The remaining families ranged 6.5% (five species, Anthocoridae) to 4% (three species each, Tingidae, Reduviidae, and Nabidae). Tingidae in the dry grassland, and Anthocoridae and Coreidae in the evergreen oak forest were more numerous than Lygaeidae.

Fifty percent of individuals collected belonged to either one of the following three species: Closterotomus trivialis (A. Costa) (Miridae), Compsidolon crotchi (Scott) (Miridae), and Tingis trichonota (Puton) (Tingidae), in decreasing order of abundance. Twenty-five rare species (32.5% of total number of species identified), represented by one or two individuals among the collected specimens, were found ( Table 3 ). The distribution pattern of frequency classes of the species in each Heteroptera assemblage fitted the lognormal distribution model ( P  > 0.05; calcicolous rosemary scrub: P  = 0.4173; χ ²  = 0.658. Dry grassland: P  = 0.2022; χ ²  = 4.616. Evergreen oak forest: P  = 0.133; χ ²  = 7.056. Kermes oak scrub: P  = 0.2238; χ ²  = 2.994. High P value cannot be taken to imply a good fit. A low value does, however, imply a bad fit; Fig. 6 ). Orthotylus ( Pinocapsus ) gemmae (Gessé and Goula) (Miridae) was described as new species from specimens in this investigation ( Gessé and Goula 2004 ). Also interesting were the second record for the Iberian Peninsula Loricula ( Loricula ) ruficeps (Reuter) (Microphysiade) and rare species of the NE Iberian Hyalochiton colpochilus (Horváth) (Tingidae), Tingis ( Tropidocheila ) trichonota (Tingidae), Brachysteles parvicornis (Costa) (Anthocoridae), Sciocoris ( Neoscyocoris ) maculatus Fieber (Pentatomidae), Acrosternum millierei (Mulsant & Rey) (Pentatomidae), and Picromerus nigridens (F.) (Pentatomidae) ( Gessé and Goula 2006 ).

Fig. 6.

Distribution of species abundance, according to a lognormal distribution model.

Of the 77 species recorded ( Table 3 ), 27 (35%) represented by 3 or more individuals were collected only in one plant community, thus considered as exclusive of that plant community. The highest percentages of exclusive species against the total amount of species in the study occur in the evergreen oak forest (27.27%) and the kermes oak scrub (25.97%). Only two species occur in all four plant communities: Gonocerus insidiator (F.) (Coreidae) and A.millierei (Pentatomidae), both living on P.lentiscus L. (Terebintaceae), which was the only plant species sampled in the four communities; 30% of species occur in two or three plant communities.

The Wilcoxon signed-rank test showed that abundance per sample ( Table 3 ) was significantly higher in the evergreen oak forest than in the calcicolous rosemary scrub ( P  = 0.0326, n  = 12, z  = 2.941, Wilcoxon test), in the evergreen oak forest than in the kermes oak scrub ( P  = 0.0352, n  = 12, z  = 2.105), in the calcicolous rosemary scrub than in the kermes oak scrub ( P  = 0.0401, n  = 12, z  = 2.052, in the evergreen oak forest than in the dry grassland ( P  = 0.00267, n  = 12, z  = 3.001), and calcicolous rosemary than in dry grassland ( P  = 0.0194, n  = 12, z  = 2.337).

Abundance per sample varied along the year for each one of the four Heteroptera assemblages ( Fig. 7 ). In the dry grassland, the maximum density value appeared in May, caused by the massive occurrence of Tingis trichonota (Puton) (Tingidae), whereas minimum density values occurred in August and during the winter. The calcicolous rosemary scrub showed a maximum mean abundance value in March, caused by the massive presence of Co. crotchi (Miridae). In the kermes oak scrub, means of abundance were more even all year round, with maximum values in spring and minimum values in winter. This is due to the fact that this community did not host any particularly very abundant species ( Table 3 ). In the evergreen oak forest, abundance per sample showed highest values in March and April because of the massive occurrence of C.trivialis (Miridae). It was the only plant community where specimens were collected along the 36 sampling months. Summarizing, the highest Heteroptera abundance per sample values occurred in the spring, particularly in March or May.

Fig. 7.

Mean abundance (specimens/sample) per month, in each of the four biotopes studied. The results of the 3 yr of study are considered altogether per each month. C, calcicolous rosemary scrub; D, dry grassland; E, evergreen oak forest; K, kermes oak scrub.

According to species richness, measured by means of Margalef’s index ( Table 4 ), the true bug assemblages may be decreasingly ordered as follows: kermes oak scrub, evergreen oak forest, dry grassland, and calcicolous rosemary scrub. The Wilcoxon signed-rank test showed that Margalef’s index monthly means ( Table 4 ) were significantly different in most of all paired comparison among the four Heteroptera assemblages ( P  < 0.05; evergreen oak forest and dry grassland, P  = 0.00765, z  = 2.667, n  = 12; evergreen oak forest and calcicolous rosemary scrub P  = 0.00221, z  = 3.059, n  = 12; calcicolous rosemary scrub and kermes oak scrub P  = 0.02079, z  = 2.312, n  = 12; calcicolous rosemary scrub and dry grassland P  = 0.04986, z  = 1.961, n  = 12), except when comparing kermes oak scrub and evergreen oak forest ( P  = 0.11652, z  = 1.57, n  = 12), and kermes oak scrub and dry grassland ( P  = 1, z  = 0, n  = 12).

Table 4.

Indices of species richness (Margalef’s index) and diversity ( H ’)

	Dry grassland	Calcicolous rosemary scrun	Kermes oak scrub	Evergreen oak forest
Margalef's index	2.892	2.5	6.775	4.865
Margalef's index monthly means ± standard deviation	1.438 ± 0.549	0.912 ± 0.811	1.714 ± 1.258	2.222 ± 0.714
Diversity ( H ')	1.514	0.614	2.583	2.012
Monthly mean values of diversity ( H ’) ± standard deviation	1.187 ± 0.493	0.695 ± 0.635	1.227 ± 0.616	1.63 ± 0.376

Monthly absolute species richness values considering the four Heteroptera assemblages together may be seen in Fig. 8 . Species richness peaked in May and June, declined in August, and showed a moderate increase towards autumn, peaking again in November. Lowest species richness values occurred during the winter.

Fig. 8.

Monthly distribution of absolute species richness. Number of species, 77; number of samples, 11,232; number of individuals, 3,071.

Species inventories at the different plant communities were 83.5–90.7% complete according to nonparametric bootstrapping species-richness estimator. Rarefaction curves estimating the number of Heteroptera species expected in each biotope are shown in Fig. 9 . A good fit between empirical and estimated species richness ( R² > 0.9) occurred in all Heteroptera assemblages. Rarefaction curves confirmed the ranking of Heteroptera assemblages according to their Margalef’s species richness index, i.e., from most to least: kermes oak scrub, evergreen oak forest, dry grassland, and calcicolous rosemary scrub.

Fig. 9.

Rarefaction curves for each of the four biotopes. Solid lines: rarefaction curves; dashed lines: logarithmic mathematical interpolation. The function value and the determination coefficient R² are also stated. C, calcicolous rosemary scrub; D, dry grassland; E, evergreen oak forest; K, kermes oak scrub; n , total number of specimens observed; np, estimated number of species using logarithmic regression model.

Diversity values ( H ’) ( Table 4 ) ranged from 0.614 (calcicolous rosemary scrub) to 2.583 (kermes oak scrub), and their monthly means from 0.695 (calcicolous rosemary scrub) to 1.63 (evergreen oak forest; Table 4 ). The Wilcoxon test of paired-data showed that monthly average values of H ’ were not significantly different in the comparison between kermes oak scrub and evergreen oak forest ( P  = 0.084 , z  = 1.726, n  = 12), and between kermes oak scrub and dry grassland ( P  = 0.8139, z  = 0.2353, n  = 12). All other possible paired comparisons of the four Heteroptera assemblages were significantly different ( P  < 0.05 evergreen oak forest and dry grassland P  = 0.0096, z  = 2.586, n  = 12; evergreen oak forest and calcicolous rosemary scrub P  = 0.00222, z  = 3.059, n  = 12; calcicolous rosemary scrub and kermes oak scrub P  = 0.026231, z  = 2.223, n  = 12; and calcicolous rosemary scrub and dry grassland P  = 0.0186, z  = 2.353, n  = 12).

Beta Diversity

CA was conducted taking into account species density. The cumulative percentage explained by the first two axes was 58.8% ( Fig. 10 ). The true bug species present in the dry grassland strongly separated this biotope by axis 1 from those found in the evergreen oak forest. The Heteroptera assemblage of kermes oak scrub strongly separated this biotope by axis 2 from those found in the dry grassland and the rosemary calcicolous scrub. The similarity in bug species composition decreased with succession stage in the vegetation succession series of the evergreen oak forest.

Fig. 10.

CA for species. Solid triangles represent the position of the centroid of exclusive species of each community, whose number is stated in brackets. Each axis shows the percentage of variation explained in each one of them.

The similarity analysis ( Fig. 11 ) showed that the Heteroptera assemblages clustered in two groups: dry grassland plus calcicolous rosemary scrub cluster in one group and kermes oak scrub plus evergreen oak forest in a second. Affinity of assemblages within each cluster was lower than 50%, although it was slightly higher in the second cluster.

Fig. 11.

Affinity dendrogram, according to Bray–Curtis qualitative similarity index test. C, calcicolous rosemary scrub; D, dry grassland; E, evergreen oak forest; K, kermes oak scrub.

Discussion

Heteroptera assemblages could be specifically related to the plant community in which they live. These assemblages are characterized by good number of species exclusively found in each plant community. As shown by the correspondence and similarity analysis, each true bug community was set apart and did reflect the different plant species composition of each biotope considered. The CA of the first year’s results ( Gessé and Goula 2003 ) did already point out a higher affinity among plots of the same biotope, than among plots of different biotopes, as was also stated in Park of Collserola ( Ribes et al. 2000 , 2001 ) and in the Pyrenees ( Goula et al. 2008 ). The core of exclusive species (67.5% of total species found) is enough to characterize the Heteroptera assemblages and explains why they sort apart when analyzed in an affinity dendrogram. Thus, the plant communities shaped the Heteroptera assemblages and those could be intrinsically associated to the vegetation type from which they were collected.

Heteroptera assemblages showed a set of common features that strongly characterize them. Families Miridae, Lygaeidae, and Pentatomidae contributed the most to the species richness of the pooled assemblage ( Table 3 ), as it was already observed in the Park of Collserola ( Ribes et al. 2000 ). This result is expected by the relative abundance of these threes families within the Heteroptera ( Henry 2009 ).

The analysis of the frequency classes of species showed that true bug assemblages studied in the Garraf Massif fitted the lognormal model ( Fig. 6 ). Thus, in each true bug assemblage studied a single species outstands for its high abundance, followed by a bulk of poorly represented species ( Table 3 ). The same pattern was stated in other studies on Heteroptera assemblages in NE Iberian Peninsula ( Goula et al. 2010 , E. Ribes, personal communication) and in other European regions ( Bryja and Kula 2000 , Sobek et al. 2009 ). The log normal model is a universal structure pattern related to undisturbed insect assemblages ( Magurran 2004 ), regulated by a large number of factors that are distributed randomly ( Moreno 2001 ). Environmental protection in the Garraf Natural Park area, enhancing undisturbed insect assemblages, makes it logical to expect the fitting to a lognormal structure pattern.

A common feature in Garraf Heteroptera assemblages was that a good number of species were found only in one plant community, to which they may be considered as exclusive. Percentage of exclusive species in Heteroptera assemblages was 62.1% species in the a Natural Park close to Barcelona city ( Ribes et al. 2000 ), ∼50% in the Pyrenees ( Goula et al. 2010 ) and Italy ( Carapezza et al. 1995 , Bacchi and Rizzotti Vlach 1997 ) or 86% in Italy ( Rizzotti Vlach 1994 ). The 67.5% of exclusive species in this work is within the rank of Heteroptera assemblages in the Mediterranean.

Species richness and abundance of true bug assemblages in the Garraf Natural Park were characterized by a peak in spring and early summer. A lower autumn peak was observed as well in species richness yearly distribution. Ribes et al. (2000) found that abundance rose in spring and summer, but an additional autumnal peak occasionally may occur. Mediterranean climate is characterized by a severe late summer drought. As a result, in the Mediterranean insect, assemblages show a typical bimodal curve of abundance and species richness peaking in spring and autumn, as it was observed in Garraf Natural Park.

The diversity values found in the Heteroptera assemblages in Garraf Natural Park ranged 2.5–6.8 for Margalef index, and 0.6–2.6 for Shannon index, that were lower than the index diversity values found in central Europe ( Bryja and Kula 2000 ). In comparing the four habitats addressed here, it may be seen that the kermes oak scrub had the greatest species richness, accounting for 55% of the total of species identified, when compared with 47% found for evergreen oak forest, 26% for dry grassland, and 22% for calcicolous rosemary scrub. Kermes oak scrub lacked any high outstanding species and showed the least abundance by comparison to the rest of habitats. This was reflected in the diversity indices values that were at highest at the kermes oak scrub.

The similarity analysis showed that Heteroptera assemblages clustered in two groups. One included dry grassland plus calcicolous rosemary scrub, which are the first two communities of the vegetation succession series of the evergreen oak forest. The other included kermes oak scrub plus evergreen oak forest, which are the last two communities of this succession. Thus, true bug assemblages in the area of study clustered as expected according to their host plant communities.

The true bug species found in each vegetation type constitute Heteroptera assemblages with particular biodiversity traits that specifically characterize each one of the assemblages. None of the four assemblages studied may be neglected or prioritize after their alpha and beta biodiversity parameters, and all of them are interesting due to their singularity. Thus, the importance and specificity of true bug assemblages increase the value and supports the decision of preserving different vegetation types within the Park, as invaluable habitats to host specific insect fauna.