Effect of Habitat Types on Breeding Bird Assemblages in the Sidi Reghis Forests ( Oum El Bouaghi , North-Eastern Algeria )

Abderraouf Chouaib Rebbah1, Mohcen Menaa2,*, Salah Telailia3, Menouar Saheb1 and Mohamed Cherif Maazi2 1Department of Natural and Life Sciences, Faculty of Sciences, University of Oum El Bouaghi, Constantine Road, Oum El Bouaghi, Algeria 2LEAT Laboratory, Department of Biology, Faculty of Natural and Life Sciences, University of Souk-Ahras, Annaba Road, Souk-Ahras, Algeria 3Faculty of Sciences, University of El Tarf, El Tarf, Algeria


INTRODUCTION
B irds are one of the most attractive life forms on Earth, with their ability to fly and wonderful coloration.They are found in different habitat types across the globe and provide numerous ecosystem services for the suite of species that live alongside them (Sekercioglu, 2006;Whelan et al., 2008).Birds are often used as a wildlife tool for a variety of purposes that facilitate the development of management strategies in avian protection and their effective implementation in nature conservation practices (Titeux, 2006).
Multiple studies on Algerian water birds in aquatic environments have produced species lists that are useful descriptors for distributional patterns (Houhamdi and Samraoui, 2002;Samraoui and Samraoui, 2008), but there are only a few avian studies associated with Algerian forest environments that have analyzed how bird community composition varies with habitat characteristics across ecological gradients (Benyacoub, 1993;Bellatreche, 1994;Mostefai, 2011;Menaa et al., 2016).These and other descriptive studies (Bensizerara et al., 2013) have dealt with the breeding ecology of bird communities in forested habitats (Bensouilah et al., 2014;Boudeffa et al., 2015).
The mountain of Sidi Reghis is located in a mountainous region of Algeria.Due to the size of its land surface area and altitude, which varies between 800 m and 1635 m above sea level (a.s.l.), the mountain is characterized by higher rainfall and a unique vegetation cover, both of which differ vastly from the climatic conditions and plant assemblages that occur in the semiarid lowlands that surround it (Mosbah, 2007).These contrasting bioclimatic zones are clearly mirrored in the variance of the vegetation structure, which boasts diverse plant species, and consequently habitats.Natural processes, such as soil erosion from water and wind, as well as temporal changes affect plant succession on Sidi Reghis.Floral assemblages are additionally impacted by anthropogenic disturbance factors (FAO, 2012), due to the proximal location of well-developed human settlements at the base of the mountain.Anthropological impacts through the exploitation of wood, over-grazing, fires, reforestation, and the uncontrolled dumping of garbage (Mosbah, 2007), all affect the biodiversity of forested patches.
In this study, the aim is to inventory the forest avifauna of Sidi Reghis, and to explore the effects of habitat types on avian assemblages by studying community parameters (abundance, species richness and occurrence frequency); to determine the intensity of habitat selection by each species.In addition, this study also aimed to providing management recommendations that encourage forest avifauna to live and breed in different forest types of the Sidi Reghis Mountain.

Study site
This study was carried out in the mountain of Sidi Reghis in Northeastern Algeria near the town of Oum El Bouaghi.The study area covers approximately 4106 ha and forms a part of the Hracta forest (Fig. 1), which belongs to the forest zone of the Algerian-Tunisian border and extends over more than 26,000 ha (Boudy, 1955).The central locality of the Sidi Reghis Mountain is located at 35° 54 ' 10.27" N and 7° 7' 26.58" E. The main soil types are clay, calcareous-clay, and ferruginous soils, of which calcareous-clay soil dominates (BNEDER, 1997).The climate is semi-arid, with a seven-month dry season from May to November, and a rainy season from December to April.Mean annual precipitation ranges from 267.91 mm to 435.88 mm, and average monthly temperatures are 4.8°C in December and 34.33 °C in July (ONM, 2016).
According to Mosbah's (2007) description, there are two main parts in the Sidi Reghis forest; native (indigenous) tree species (Quercus ilex) and reforested (non-native) tree species (Pinus halepensis).The forest includes three major habitat types that are classified according to the dominant tree species: holm oak (Quercus ilex) forests, Aleppo pine (Pinus halepensis) stands and mixed oak and pine stands.

Bird surveys
Birds were surveyed using a point count or IPA method (Indices Ponctuels d'Abondance; see Blondel et al., 1970;Bibby et al., 1992), with two bird surveys undertaken over the breeding season in 2014, 2015 and 2016 (Drapeau et al., 1999).One survey took place from mid-March to mid-April for early breeders and the second from mid-May to mid-June for species that arrived later.
The census technique involved a count of all birds seen or heard inside or outside a 100 m radius circular plot during a 15 min period.Birds that flew over and did not land in trees or on the ground were recorded but not included in the data analysis because point counts were not considered a suitable method for these taxa (Bibby et al., 2000).Surveys were restricted to good weather conditions (no rain and wind speed lower than 20 km/h) and occurred within the four hours of sunrise, when vocal activities of diurnal birds began (Frochot and Roché, 1990).We established 126 point count stations that were distributed systematically across the entire study area.Each point count was separated by at least 250 m from all other points to minimize the probability of sampling the same bird more than once because, in forested areas, the loudest song can be heard at a maximum distance of approximately 250 m (Foucès, 1995).We used the maximum abundance of each bird species per survey point from the replicate counts, because these counts are closer to the real number of individuals and species present in each plot (Sánchez et al., 2012).

Data analyses Bird species composition
To compare diversity between habitat types, a variety of community parameters were used.The Shannon-Wiener index (diversity: H') was calculated with the parameters that affect this index, such as species richness (S), relative abundance (A), and frequency of occurrence (F%) (Anjos et al., 2010).Observed distribution of these community parameters was tested for normality using Shapiro-Wilk test (Shapiro and Wilk, 1965) and the Fligner-Killeen test (Fligner and Killeen, 1976) for homogeneity of variance among habitats.Then, one-way analysis of variance (ANOVA) tests were used to test for differences in species richness, relative abundance, and species diversity among the three main surveyed habitats (oak stands, pine stands, oak-pine mixed forests).When a significant difference was detected, Tukey's Honestly Significant Difference posthoc tests (Kramer, 1956;Keselman and Rogan, 1977) were used to determine individual mean differences (α = 0.05).

Comparison of bird assemblages among habitat types
We used non-metric multi-dimensional scaling (NMDS) to test for differences between avian assemblages and habitat types.The NMDS was constructed using a matrix of ecological dissimilarity among habitat types (Legendre and Legendre, 1998), and a probability value that was calculated based on 10,000 Monte Carlo simulations.An advantage of using NMDS is that it is based on ranked distances, which tends to linearize the relationship between environmental distance and biological distance (Legendre and Legendre, 1998).The amount of stress can be used for judging the goodness of fit of NMDS.Kruskal (1964) provided an interpretation of the stress value with respect to the goodness of fit of NMDS, indicating that a small stress value highlights a good fit (lower than 0.2).Whereas; a high value points towards a weak fit (higher than 0.2).Although, the amount of stress is informative, it has been generally accepted that stress levels only offer a vague indication of goodness of fit (Oksanen, 2013).
To analyze the differences in bird assemblage composition between habitats a Permutational Multivariate Analysis of Variance procedure (PERMANOVA) was used (Anderson, 2001).This procedure acts as a matrixbased non-parametric analysis of variance.PERMANOVA analyses and segments sums of squares using semi-metric and metric distance matrices using permutation methods (Anderson, 2005).When differences were detected, a one-way Analysis of Similarity (ANOSIM) was used to further investigate whether bird community structure (a single data matrix composed of the relative abundance of all species detected at each point count) differed among the possible pairwise combinations in the three sampled habitats (Minchin, 1987).This was done because ANOSIM tests whether the dissimilarities identified in the assembly composition are larger between groups than within groups or not; this also produces an estimated p-value based on 10,000 Monte Carlo simulations (Clarke, 1993).In addition, a Similarity Percentage (SIMPER) test was conducted to estimate overall dissimilarity among habitat types.The SIMPER test also allows to assess the relative contribution of each species to the assembly composition, both in respect of contribution to the average similarity within a group (i.e. which species at what abundance tends to characterize groups); and average dissimilarity between groups (i.e. which species at what abundance tends to separate groups) (Clarke, 1993).A Bray-Curtis pairwise distance coefficient was used in all cases to express similarities as it is less sensitive to differences among rare species, and 10,000 Monte Carlo permutations were conducted to generate the random test statistics (Bray and Curtis, 1957).
All analyses were undertaken in R (R Development Core Team, 2014) with the Community Ecology Package 'vegan' (Oksanen et al., 2010).

Bird species composition
During the breeding periods of 2014, 2015 and 2016, we conducted 252 visits (252 partials IPA).A total of 1276 pairs of birds in 53 genera and 69 species were recorded.Fifty three species were Passeriformes and the remainder (16) were non-Passeriformes.Sixty one bird species were found in pine woodlands, fifty one in oak woodlands and 34 in mixed oak-pine forests (Table I).One species was recorded only in mixed oak-pine forests (Eurasian Siskin Spinus spinus; at one point count), six species were found only in oak woodlands and 17 species were found only in pine woodlands (Table I).The six most commonly detected species in the mountain of Sidi Reghis were Common Chaffinch (155.5 pairs), European Serin (151.5 pairs), Common Blackbird (116 pairs), European Greenfinch (111.5 pairs), Spotted Flycatcher (91.5 pairs), and European Turtle Dove (45.5 pairs).These six species accounted for over half (52.62%) of all detections (Table I).
The family with the highest species richness was Muscicapidae (13 species), followed by Fringillidae (seven species), Sylviidae (six species), Accipitridae (five species) and Alaudidae (four species).These five families alone represented more than 50% of the total species richness of the community.The family that dominated the population in number of pairs was Fringillidae (446.5 pairs), followed by Muscicapidae (194 pairs), Turdidae (117 pairs), Columbidae (85 pairs), and Passeridae (74.5 pairs).They represented more than 70% of the total abundance of the entire population (Table II).
Results from the one-way ANOVA analysis for the effect of forest type on bird species richness (S), abundance (A) and diversity (H') indicated that forest bird abundance and species richness significantly differed among the three forest types (abundance: F 2.123 = 6.205, p < 0.01, adjusted R 2 = 0.076; species richness: F 2.123 = 6.059, p < 0.01, adjusted R 2 = 0.074).Abundance and species richness were significantly higher in pure pine woodlands than in mixed oak-pine forests and oak woodlands (Tukey's HSD post-hoc test: p < 0.01) (Fig. 2A, B; Table III).
Species diversity also differed significantly among the three forest types (F 2.123 = 5.108; p < 0.01, adjusted R 2 = 0.063), with significantly higher species diversity in pure pine woodlands than in mixed oak-pine forests and oak woodlands (Tukey's HSD post-hoc test: p < 0.01) (Fig. 2C; Table III).

Comparison of bird assemblages among habitat types
Avian assemblages in Sidi Reghis Mountain varied significantly between the different habitats (PERMANOVA: F 2.58 = 5.572, p < 0.001).Further, differences in bird species composition among the possible pairwise combinations in the three sampled habitat types were confirmed by the ANOSIM test (Table IV).These results were supported by the nonmetric multidimensional scaling (NMDS) analysis, which produced a good fit (0.185 stress) with a clear positive linear relationship between the observed dissimilarity and the ordination distances (for linear fit: r 2 = 0.835; Fig. 3).The NMDS plot revealed that some species were entirely restricted to a given habitat type, which shared different complements of its avifauna with other habitat types (Table V; Fig. 4).The most marked contrast in species composition was between oak woodlands and pine woodlands, with only 14 species in common (Table V; Fig. 4).They diverged considerably in their bird assemblage composition, being distinctly separated at opposite ends of the ordination diagram.The mixed oak-pine sites were similarly different, with 30 species in common with oak woodlands and/or pine woodlands (Table V), appearing to cluster between oak woodlands and pine woodlands (Fig. 4).Considering the overlapping of oak woodlands and oak-pine forests, pine woodlands hosted the most dissimilar community (Table V).Six of the seven species were responsible for the mean of 50% of dissimilarity between sampled habitats (Table V).The dissimilarity between pine woodlands and mixed oak-pine forests was about 50%, in general, and produced by differences in abundance of common chaffinch, European serin, house sparrow, European greenfinch, spotted flycatcher and common blackbird.The differences between pine woodlands and oak forests (about 60%) and between oak-pine mixed and oak woodlands (about 50%) were mainly produced by species that were present in just one sampled area, most with preference for pine woodlands.
The differences related to pine woodlands are the results from Moussier's redstart and European turtle dove presence.

Bird species composition
According to Isenmann and Moali (2000), 406 species of birds are found in Algeria, thus the species recorded at the mountain of Sidi Reghis correspond to 17% of the Algerian avifauna.
The presence of numerous protected, endangered and endemic species confirms the importance of the mountain of Sidi Reghis as a key habitat for the conservation of rare and endemic avifauna.Black kite Milvus migrans, Egyptian vulture Neophron percnopterus, common kestrel falco tinnunculus, booted eagle Hieraaetus pennatus, long-legged buzzard, on the other hand, have been taken into account because some of the pairs nest in the heart of the mountain of Sidi Reghis and feed there.
The most abundant species at the Mountain of Sidi Reghis is common chaffinch, which is a typical forest bird in North Africa and temperate Europe (Cherkaoui et al., 2007;Dronneau, 2007;Mostefai, 2011;Menaa, 2016).Muller (1985) demonstrated that this sparrow occupies the first place in all types of forests, whether hardwoods, conifers or mixed stands.
Interestingly, several species nesting in Sidi Reghis forests are mainly subservient to open areas (Calandra lark Melanocorypha calandra, Maghreb lark, Wood lark Lullula arborea, Eurasian skylark Alauda arvensis, Cirl bunting Emberiza cirlus and rock bunting Emberiza cia) and urban land (white stork Ciconia ciconia, rock dove Columba livia, European turtle dove barn swallow Hirundo rustica, common house hartin Delichon urbicum and house sparrow).This is easily explained because the Sidi Reghis Mountain contains forest edges influenced by anthropological impacts and the proximal location of welldeveloped human settlements at the base of the mountain (around the mountain there is a large urban agglomeration, especially in the south).
Our results also support the conclusions of Camprodon and Brotons (2006).We have suggested that the presence of species of grassland and open areas beside purely forest species is due to the mosaic structure of the Sidi Reghis forests (presence of clearings and scrubland) and the clearing in agro-forestry habitats that also support grassland species (because the grasslands are located adjacent to the mountain).

Comparison of bird assemblages among habitat types
Using diversity indices is one of the most important challenges in ecological studies aiming at understanding patterns of biodiversity and their underlying causes (Colwell and Coddington, 1994).
Increases in vegetation structure complexity and floristic composition are quite often related to enrichment of bird communities (Wiens, 1989;Hobson and Bayne, 2000a, b;Shochat et al., 2001;Laiolo, 2002;Machtans and Latour, 2003).However, relative abundance, species richness and species diversity of forest birds in the mountain of Sidi Reghis were on average higher in pine woodlands than oak woodlands and mixed oak-pine forests, contrary to our expectation.
On the other hand, several authors found lower species richness in coniferous compared to broadleaved forests (James and Wamer, 1982;Barbaro et al., 2005;Gil-Tena et al., 2007) or a greater association of bird communities with the latter (Berg, 1997).Nonetheless, results from previous studies are often contradictory and dependent on the scales and study areas.
Similarly, Hobson and Bayne (2000a) could not associate higher species richness to coniferous or deciduous forests.Studies conducted in the Iberian Peninsula regarding the environmental patterns associated with the distribution of forest avian communities have also pointed out this uncertainty (Tellería and Santos, 1994;Carrascal and Díaz, 2003).
Consequently, the significant increase in bird species richness in pine forests is likely to be the result of the assemblage of bird species from urban land and open area.In contrast, the significant decrease in bird species richness in mixed oak-pine forests is likely to be the result of the loss and degradation of native vegetation by land management practices in Sidi Reghis Mountain, where the native holm oak have been replaced by the introduced Aleppo pine.Because native vegetation is important for many species, numerous authors have equated 'habitat' with 'native vegetation' (Andrén, 1994).Hence, the loss of native vegetation at landscape and regional scales has been linked to the loss of native species around the world (Andrén, 1994;Kerr and Deguise, 2004).Similarly, the loss of native vegetation at the local scale tends to reduce native species richness, which is in accordance with our results.
Our study revealed some resemblance of bird communities among habitat types.These three habitats are geographically close to each other, while the whole of this mountain allows a sparse evolution of the vegetation.
Each elevation stage has its own type of flora.The lower elevation is composed of an Aleppo pine plantation (introduced for reforestation during the last two centuries) which develops to the detriment of other species.The intermediate elevation is composed of mixed woodlands of Aleppo pine and holm oak.Finally, the higher elevation consists of holm oak, the autochthonous species which is typical of the Mediterranean region (Djema and Messoudène, 2009).
In contrast, strong dissimilarity between bird communities among habitats was found in Sidi Reghis Mountain.This is probably due to the geographical (altitude) and ecological characteristics.Most of the differences are found between the lower mountain altitude part (pine woodlands) and the higher mountain altitude part (oak woodlands).In the lower altitude part, the pine woodlands connect with open areas (grasslands) and urban lands, allowing a wider range for species movement.In higher altitude, bird community of oak woodlands has its own forest characteristics which are specific to altitudes above 1500 m.The hostility of the climate and the poverty of the soil yield poor vegetation cover in this part of the mountain: the holm oak only occupies rocky spaces providing shelter from the wind (Mosbah, 2007).

CONCLUSION
The results obtained in this study significantly contribute to knowledge of breeding birds in the Sidi Reghis Mountain; help further assessing the effects of habitat types on the integrity of bird communities.This information will help also in planning future conservation activities to maintain the biodiversity in this forest ecosystem by providing a short list of some management recommendations, according to Fischer and Lindenmayer (2007): 1) Forest landscape management should focus on maintaining forest heterogeneity in order to provide a diversity of habitat types that are useful to a range of different bird species; 2) Especially for bird species which depend on native vegetation, it is very important to restore large and structurally complex patches of native vegetation in order to provide core habitat for these species and 3) Provide habitat for many species throughout the woodstands, by maintaining and/or restoring a matrix that is structurally similar to native vegetation.for providing us stock maps and other information.We would like to thank all the staff of Tour du Valat research institute, and particularly to Dr. Patrick Grillas, Ms. Roberta Fausti, Dr. Thomas Galewski, Mr. Anis Guelmami and Mr. Antoine Arnaud, for their help and support.We would like to acknowledge the valuable comments and suggestions of Dr. Paul Acker, which have improved the quality of this paper.We also thank all the members of the Algerian National Association of Ornithology (ANAO).

Statement of conflict of interest
Authors have declared no conflict of interest.

Fig. 2 .
Fig.2.The relative abundance (A), species richness (B), and species diversity of forest birds in mixed oak-pine forests (Mix), oak woodlands (Oak), and pine woodlands (Pine) (C).Left graphic represent significant (p < 0.05) pairwise differences between forest types for each community parameters.

Table I .-Bird species/families/orders and avian distribution recorded in the mountain of Sidi Reghis during the breeding period of 2014, 2015 and 2016. No. Scientific name/ Common English name Habitat Abundance F (%) IUCN red list status 2015
Oak, oak woodlands; Mix, mixed oak-pine forests; Pine, pine woodlands.P, protected; UP, unprotected (according to the National Protection status 2012); NE, not evaluated; LC, least concern; EN, endangered (according to the IUCN Red List status 2015).