Breeding Bird Assemblages in Relation to Changing Forest Composition in North- Eastern Algeria: Zonation or Continuum?

1Department of Agronomy, Faculty of Nature and Life Sciences, Chadli Bendjedid University of El Tarf, PO Box 73, El Tarf, 36000, Algeria. 2Department of Biology, Faculty of Nature and Life Sciences, Mohamed Cherif Messaadia University of Souk Ahras, PO Box 1553, Souk-Ahras, 41000, Algeria 3Department of Biology, Faculty of Nature and Life Sciences and Sciences of Earth and Universe, 08 May 1945 University of Guelma, Guelma, 24000, Algeria Article Information Received 05 January 2023 Revised 22 June 2023 Accepted 24 July 2023 Available online 09 October 2023 (early access)


INTRODUCTION
B irds occupy different habitat types and ecosystems which make them a useful tool for numerous purposes by landscape services (Whelan et al., 2008;Sekercioglu, 2012;Maas et al., 2016), such as application of biodiversity maintenance and restoration, and landscapescale conservation (Sandström et al., 2006).
During the past decades, avian research scientists have long been interested in bird habitat relationships based on ecological systems features with the main objective of studying spatial distribution variability in presence/absence of bird species (MacArthur and MacArthur, 1961;MacArthur, 1964;Cody, 1985;Wiens, 1989) because habitat requirements had a great effect on bird community structure and composition. Generally, many physical and biological constraints that manage bird species and their functioning, such as the presence or availability of foraging resources and the potential bird niche space, which in turn are affected by the spatial scale of landscape patterns (MacArthur and MacArthur, 1961;MacArthur, 1964;Cody, 1985;Wiens, 1989). Therefore, the understanding of the spatial scale of landscape patterns influence is related also to the proper temporal scale (Wiens, 1989).
Results of many studies carried out earlier have O n l i n e

F i r s t A r t i c l e
considered that plant composition was the secondary determinant factor for bird assemblages (Robinson and Holms, 1984;Rotenberry, 1985;Benyakoub, 1993;Bellatreche, 1994). Other environmental factors include foliage volume (Buchanan et al., 1999), tree age (Sallabanks et al., 2006;Menaa, 2017), forest productivity (Cody, 1981), physiognomy of shrub layer (Reid et al., 2004;Diaz, 2006), plant succession (Sweeney et al., 2010), the size and structures of the habitat patch or connectivity (Henderson et al., 1985), and edge effects (McGarigal and McComb, 1995;Kuehler et al., 2001). All of these findings were in relation to the ecological system hierarchy and scale (Virkkala, 1991). Plant composition changes along gradients of ecological factors, where their ecological parameters varied significantly across altitudinal changes which are considered as the biggest determinant factor affecting floristic composition (Hemp, 2006). This important determinant factor is still a central issue whether altitudinal gradients are continuous or discontinuous. This issue is related to zonation and continuum concepts.
Many authors described vegetation communities as a continuum such as the forest stands of the tropical mountains (Hamilton, 1975, Hamilton et al., 1989Lieberman et al., 1996;Lovett, 1996Lovett, , 1998. However, other studies highlighted that landscape systems of vegetation surrounded by narrow boundaries in which there is altitudinal gradient discontinuous vary in plant composition or structure and forms zones or belts (Woldu et al., 1989;Friis and Lawesson, 1993).
The forest of Ouled Bechih is known as an important ecosystem for biodiversity and unfortunately is subject to natural and anthropogenic disturbance factors (Ganaoui et al., 2020;Guellati et al., 2022). It is located in a region characterized by an exceptional climate. This led to a clear variation in plant structure and in plant species diversity, and consequently in habitats. Despite all this significant plant diversity, this complexity of landscape has never been the subject of any serious ornithological study.
In the present study, our overarching aim was: (i) to prepare the inventory of forest avifauna of the forest of Ouled Bechih; (ii) to explore the effects of the habitat type (cork oak woodlands, mixed oak woodlands and zeen oak woodlands) on forest avifauna, by studying community parameters (abundance, species richness, and frequency of occurrence); (iii) to determine the intensity of selection of each species by their preferred habitat; and (iv) to provide management recommendations for encouraging forest avifauna in woodlands.

Study site
This study was carried out in the forest of Ouled Bechih (Mechroha municipality) in the extreme Northeastern Algeria within the territory of the province of Souk Ahras in the north, limited to the north and the northeast by the forests of Beni Salah and Fedj Laamed, to the west and the north-west by the forests of Beni Salah and Oued Ghanem and to the south by the forests of Rezgoune and Fedj Mactaa. It covers approximately 6,582 ha ( Fig.  1) and includes three major forest types classified by their dominant tree species: Zeen oak (Quercus canariensis) stands, cork oak (Quercus suber) stands and mixed cork oak and zeen oak forests. The coordinates of the central locality of the forest of Ouled Bechih are: 36° 23.415'N; 7° 52.735'E, with altitudes ranging from about 392 to 1,252 m. Principal soil types are podzol, brown forest soil and siliceous soils. Among these types, siliceous soil is predominant. The forest Ouled Bechih is located between the humid and subhumid bioclimatic stage with a dry season from mid-May to late October, and a rainy season from November to mid-May. It receives an annual average of 690 mm of precipitation and the average temperature is 16°C.

Bird surveys
Forest birds were sampled with the point-counting method or IPA method (Indices Ponctuels d'Abondance) (Blondel et al., 1970;Bibby et al., 2000) in providing two bird survey rounds (Drapeau et al., 1999)  mid-June for late breeding species. This technique involved a count of all birds seen or heard inside or outside a 100 m radius circular plot during a 15 min period; overflying birds that did not land in trees or on the ground were recorded but their data were not used in statistical analysis because point-counting method is not a suitable sampling method for these taxa (Bibby et al., 2000). Bird surveys occurred within four hours of sunrise when song activities of diurnal birds begin (Frochot and Roché, 1990), and restricted to good weather conditions. The different point count stations were distributed systematically and covered the study area; each point count was separated by at least 350 m from the other points to minimize the probability of contacting the same individual bird more than once because it can be heard at distances of 250 m (Foucès, 1995).

Forest bird species structure and composition
To compare the structure and the composition of the bird species among the three forest types, we used a variety of ecology parameters, the Shannon-Wiener index (H') and the parameters that affect this index such as species richness (S) and relative abundance (A) . The significance of differences in species richness, relative abundance and diversity among these forest habitats was checked with one-way ANOVA analysis test. The normality of the observation of the different ecology parameters were tested by the Shapiro-Wilk test (Shapiro and Wilk, 1965). The homogeneity of their variances among the main habitats was also tested using the Fligner Killeen test (Fligner and Killeen, 1976).

Forest bird assemblage structure and composition
Permutational multivariate analysis of variance (PERMANOVA) method was used to compare differences in bird assemblage composition across habitat types (Anderson, 2001), which is undertaken on a matrix-based nonparametric analysis of variance. By using permutation methods, the PERMANOVA analyses and divides the sums of squares based on semi-metric and metric distance matrices (Anderson, 2005).
One-way analysis of similarity ANOSIM (analysis of similarity) was also performed to further determine if bird community structure (a single data frame composed of the relative abundances of all bird species detected at each point count) differed significantly among the possible pairwise combinations of the three sampled forest areas (Minchin, 1987) when the PERMANOVA results gave significant differences. Furthermore, the dissimilarities in the assembly composition were checked if they were larger between combinations than inside the combinations and produces an estimated p-value based on 10,000 Monte Carlo simulations (Clarke, 1993).
A complementary non-parametric technique was used to project the variation of avian assemblages among forest habitats, using nonmetric multidimensional scaling

F i r s t A r t i c l e
(NMDS) method; the NMDS was conducted with a data frame of ecological dissimilarity among the different habitat types (Legendre and Legendre, 1998), and a probability value was calculated based on 10,000 Monte Carlo simulations. The NMDS builds upon a general principle of ranked distances and linearized relationship between environmental distance and ecological distance (Legendre and Legendre, 1998). The importance of stress value could be used in order to judge the goodness of fit of this nonparametric method (Kruskal, 1964). A low importance of the stress value (< 0.2) indicates a good fit, whereas a high importance (> 0.2) indicates a weak fit (Oksanen, 2013). Complementarily, a similarity percentage (SIMPER) test was performed to estimate overall dissimilarity between habitats and also to determine the relative contribution of each bird species to the community composition, both in respect of contributions to the average similarity within a community assemblage (i.e. identity and relative abundances of species which unite a group) and average dissimilarity between community assemblages (i.e. identity and relative abundances of species which to separate groups) (Clarke, 1993). Bray-Curtis pairwise distance coefficients were used in all analyses to express similarities and/or dissimilarities, which it is less sensitive to differences among rare species, where they were also based on 10.000 Monte Carlo permutations to generate a random test statistic (Bray and Curtis, 1957).
These analyses were all undertaken in R (R Development Core Team, 2020) with the Community Ecology Package 'vegan' (Oksanen et al., 2010) and the create elegant data visualisations using the grammar of graphics pakage 'ggplot2' (Wickham et al., 2016).

Forest bird species structure and composition
We conducted 120 visits (120 partials IPA) in the breeding period of 2018 and 2019. A total of 1119.5 pairs of birds in 50 genera and 62 species were recorded. Fortyfive species were Passeriformes and the remainder (17) was non-Passeriformes; 50 bird species were found in cork oak woodlands, 46 in zeen oak woodlands, and 44 in mixed oak forests (Table I).

Forest Bird Assemblage Structure and Composition
No.

O n l i n e F i r s t A r t i c l e
A. Guefri et al.

O n l i n e F i r s t A r t i c l e
Forest Bird Assemblage Structure and Composition 7 In contrast, four species was recorded only in mixed forests (Coturnix coturnix africana, Turnix sylvaticus, Milvus migrans migrans, Saxicola rubetra), five species were found only in cork oak woodlands (Jynx torquilla, Curruca hortensis, Curruca iberiae, Curruca undata, Phoenicurus moussieri), and six were found only in zeen oak woodlands (Alectoris barbara Barbara, Alauda arvensis, Calandrella brachydactyla rubiginosa, Phylloscopus ibericus, Ficedula semitorquata, Emberiza cirlus) ( Table I). The family with the highest species richness was Muscicapidae (eight species), followed by Sylviidae (six species), Phylloscopidae (five species), Fringillidae (four species), and Picidae (four species) (Table II). These five families alone represented more than 40% of the total species richness of the community. Paridae (dominated the population in number of pairs with 185.5 pairs, followed by Muscicapidae (120.5 pairs), Fringillidae (108 pairs), Sylviidae (93.5 pairs), and Phylloscupidae (80.5 pairs). They represented more than 52% 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 indices richness (S), abundance (A), and diversity (H') indicated that forest bird abundance, richness and species diversity did not vary significantly among the three forest types (abundance: F 2.57 = 2, p > 0.05; richness: F 2.57 = 0.156, p > 0.05; species diversity: F 2.57 = 0.213, p > 0.05) (Fig. 2a, b, c).

Forest bird assemblage structure and composition
The bird assemblages of the forest of Ouled Bechih varied significantly between the different habitats (PERMANOVA: F 2,57 = 0.0824, p < 0.001), as well as the differences in bird species assemblages among the possible pairwise combinations in the three forest habitats were confirmed by the ANOSIM test, where the significant differences have been shown only between Q. canariensis woodlands and Q. suber woodlands (Table III). These results are supported by the NMDS analysis, indicating a good fit (0.1747 stress, p < 0.05) with a clear positive linear relationship between the observed dissimilarity and the ordination distances (for linear fit: r 2 = 0.849, Fig. 3).

O n l i n e F i r s t A r t i c l e
A. Guefri et al. According to the NMDS diagram, some species were entirely restricted to a given forest, which shared different complements of its bird community with other forests (Table IV and Fig. 4). The most marked contrast in species composition was therefore between the bird assemblages of Q. canariensis stands and oak mixed forests with only 14 species in common (Table IV and Fig. 4) on the one hand, and between Q. suber stands and oak mixed forests with only 14 species in common (Table  IV and Fig. 4) on the other. They separated considerably in their bird assemblage composition, being distinctly divided at opposite ends of the ordination graph. The Q. canariensis sites were similarly distinct, with 14 species in common with Q. suber woodlands (Table IV), appearing to cluster between these woodlands (Fig. 4). As well as the most dissimilar across the communities is Q. canariensis woodlands, considering the intersecting of Q. suber stands and oak mixed forests (Table IV).

O n l i n e F i r s t A r t i c l e
A. Guefri et al.
Nine species are responsible for the mean of over 30% of dissimilarity between sampled areas (Table IV). The dissimilarity produced between Q. suber stands and oak mixed forests was in general, by the abundance difference of Curruca melanocephala, Cuculus canorus, Dendrocopo major numidus and Serinus serinus. The differences between Q. canariensis and Q. suber woodlands are produced mainly by C. teneriffae, Ficedula hypoleuca, Columba palumbus and Lullula arborea which were present with preference for a given woodlands.
The dissimilarity showed between Q. canariensis stands and oak mixed forests are due to Phylloscopus bonelli, C. canorus and C. palumbus presence.

Forest bird species structure and composition
406 species were found in Algeria (Isenmann and Moali, 2000), and the species recorded in the forest of Ouled Bechih correspond to 15% of the Algerian avifauna species.
More than 32% of birds species occurring in the study area are threatened (Jordap, 2012) The most abundant species in the studied forest areas are C. teneriffae, F. coelebs, E. rubecula witherbyi, T. murela and P. major, these forest birds are specialists of mediterranean oak woodlands that need more mature forest (Diaz et al., 1998;Santos et al., 2006), thus highlighting the suitability of this ecosystem for forest birds of the region. This confirms also to the importance of the biogeographic region when planning forest management measures (Suárez-Seoane et al., 2002).
We noted also the breeding in the studied landscape of some bird species mainly belonging to open areas, (e.g., L. arborea, Emberiza cia) and urbain land (e.g., C. ciconia ciconia, S. turtur, H. rustica rustica) because this study area contains forest edges and habitats influenced by human activities.
Our study also indicated that the presence of species of grassland and open areas beside purely forest species is due to the clear and the mosaic structure of the forest areas (presence of clearings and scrubland); the clearing of the open forests to create agro-forestry habitat that also supports grassland species, as the grasslands are located adjacent to the forest of Ouled Bechih, which confirms the conclusions of Rebbah et al. (2019).

Forest bird assemblage structure and composition
In order to apprehend the patterns of biological diversity and their fundamental causes, various diversity parameters were used. They are one of the most significant challenges used in ecological studies (Colwell and Coddington, 1994). Diversity indices are suitable ecological tools where comparisons across different habitats were needed (Begon et al., 1996).
Bird community enrichment is related to intensification in vegetation physiognomy, complexity and composition (Wiens, 1989;Hobson and Bayne, 2000a, b;Shochat et al., 2001;Laiolo, 2002;Machtans and Latour, 2003). However, no significant variation of the relative abundance, the species richness, and the species diversity of forest birds in Ouled Bechih were detected among the sampled forest areas, contrary to our expectation.
Moreover, several authors found less species richness in pure forests compared to mixed forests (James and Wamer, 1982;Barbaro et al., 2005) or a greater association of bird communities with the habitat complexity (Berg, 1997), although results from other studies are contradictory and related on the scales perception of the study area.
Furthermore, according to  more species richness was not related to heterogeneous forests, and other studies carried out in the Iberian Peninsula concerning the environmental factors associated with the distribution of forest bird communities also have emphasized this hesitation (Tellería and Santos, 1994;Carrascal and Díaz, 2003).
Our study revealed a significant resemblance of bird assemblages across forest types. These three forests are geographically closest to each other, while the entire of the forest of Ouled Bichih allows a sparse evolution of the vegetation, for each stage its own type of vegetation, the lower altitude stage consisting of cork oak stands, then in the middle, oak mixed forests, and finally to the highest altitude stage, zeen oak woodlands, these autochthonous species characterize the Mediterranean perimeter (Djema O n l i n e

F i r s t A r t i c l e
and Messoudène, 2009). However, important dissimilarity bird assemblages were observed between the two pure stands, the lower forest of Q. suber and the upper forest of the Q. canariensis. This is probably due to its geographical (altitude gradient) and their ecological characteristics where most of the differences come from the lowest forest altitude part and the highest forest altitude part; in the lowest altitude stage, the cork oak stands is in direct contact with open landscapes (grasslands) and urban lands, permitting a significant species movement, and in the highest altitude of the forest, bird assembly of the cork oak woodlands presents its specific forest characteristic for altitudes above 1000m. In addition, these results determined the altitudinal zonation of the two forest stands cited above where significant discontinuities in the bird assemblages were revealed. The findings of the present study agree with previous studies (Romdal and Rahbek, 2009;Patterson et al., 1998;Goodman and Rasolonandrasana, 2001;Hamilton, 1975).
Nevertheless, the insignificant dissimilarity in bird assemblages within the three forests between Q. canariensis stands and oak mixed forests on the one hand, and between Q. suber stands and oak mixed forests on the other, should be the result of the continuities in the plant composition communities, which occurred in parallel between the three stands of the forest are indicative of the transition from one forest to the other by means of the middle oak mixed forests. These results led to consider the contrasting community unit theory vs (individualistic) continuum concepts (Moravec, 1989).
Despite the clear limits between the three forest landscapes, this theory illustrates that the forest of Ouled Bechih correspond as continuum landscape which can by their different forest stands perceived the same bird species assemblages. The intermediate habitat of the oak mixed forests at Ouled Bechih corresponds on the whole to a continuum of the other forest stands (Q. suber and Q. canariensis). This indicated the largest response of the oak mixed forests to the assembly of intermediate bird species that preferred zeen oak forests or cork oak forests. For these species, oak mixed forests provided habitat substitution for zeen oak forests or cork oak forests, such as habitat for movement, or habitat for foraging (Lindenmayer et al., 2002). For example, Certhia brachydactyla and Pycnonotus barbatus used oak mixed forests for foraging, permitting habitat movements to be extended outside cork oak forests into oak mixed forests because they might otherwise be considered sensitive to plant composition, indicating a preference for this habitat in some aspects of their ecology (White et al., 2005;Stralberg et al., 2015).

CONCLUSION
In conclusion, in our study, the results obtained contribute significantly to understanding the breeding bird distribution across the oak woodlands of Ouled Bechih; help further assessing the effects of plant composition on the integrity of bird communities. These findings will help also in planning future conservation measures to supporting the biodiversity in this forest landscapes by providing some management recommendations such as; (1) Forest landscape management should focus on maintaining forest heterogeneity in order to provide a diversity of habitats that are useful to a range of different bird species; (2) Especially for bird species which depend on homogenous vegetation, it is very important to restore large and structurally complex patches of homogenous forests in order to provide essential habitat for these bird species.

ACKNOWLEDGEMENT
We would like to express our particular appreciation and our profound gratitude to Professor Menouar Saheb (University of Oum El Bouaghi, Algeria) for his continuous efforts in Ornithology.

Funding
None.

IRB approval
This study was conducted in accordance with ethical standards and guidelines. As the research did not involve human participants, institutional review board (IRB) approval was not required.

Ethics statement
Our research was conducted with integrity and in accordance with the highest ethical standards and applicable regulations.

Data availability statement
Data available on request from the authors.

Statement of conflict of interest
The authors have declared no conflict of interest.

O n l i n e F i r s t A r t i c l e
Forest Bird Assemblage Structure and Composition