Structure of waterbird assemblages in fragmented coastal wetlands of Northeastern Algeria

Study Extent

Fieldwork was conducted in Jijel eco-complex wetlands located in north-eastern of Algeria (36° 34'-36° 52'N, 5° 33'-6° 19' E), which are approximately 60 km2 in size. This natural wetland reserve comprises two major aquatic habitats that may vary in vegetation composition and structure, namely: (i) Beni-Belaid Lake (B-B.Lake) and (ii) El-Kennar Marsh (E-K.Marsh) (see map in fig. 1). This latter freshwater swamp is dominated by herbaceous rather than woody plant species. These mosaic hydrosystems are characterized by several types, mainly lakes, marshes, reservoirs, ponds, estuarine waters, and shallow seawater. Together, they form a complex ecosystem of wetlands. It is of note that the extent of many wetlands declined significantly over the ten-year study period, and that the difference in surface area between the two main sites and their associated wetlands is large, with more than 50 % being less than 3 ha (pers. obs.). This ecoregion has a typical Mediterranean climate characterized by alternating wet and dry seasons, with a mean annual rainfall of 950 mm, mainly from November to February (Bouldjedri, 2013). The water level peaks between November-April, and drops to its lowest level over August-October.
Beni-Belaid Lake is situated in the east of the mouth of the El-Kebir River. It covers an area of about 43 ha and has brackish and alkaline water. This lake was formed by an abandoned channel of the El-Kebir River and is defined as an old delta of this River (Bouldjedri et al., 2011). During the lowest rainfall period, the water storage depth of the lake often maintains a level of some 0.5-1m in the centre. The hydrological gradient also has a temporal component, characterized by an intra-annual cyclicity linked to the summer development of the hydrophilic and halophytic communities. The lush vegetation of this site is characterized by woody and herbaceous species, which is structured in belts. It is composed mainly of white poplar (Populus alba), French tamarisk (Tamarix gallica), white broom (Retama raetam), sea clubrush (Bolbochoenus maritimes), and common clubrush (Schoenoplectus lacustris), as well as fields of grass (Cynodan dactylon) and knotgrass (Paspalum distichum). Lake communities are dominated by fennel-leaved pondweed (Potamogeton pectinatus), water horn wart (Ceratophylum demersum), stone warts (Charophyta), and the invasive species creeping water primrose (Ludwigia peploides). In the context of preservation of the avian biodiversity, Beni-Belaid Lake is listed as a natural reserve and a wetland of “international importance” by the Ramsar Convention (2003). El-Kennar Marsh is a large open water area. The depth of water storage does not exceed 0.9 m in the centre of the swamp during the drought period. It is well-known that this swamp is contaminated by urban sewage, and it is also under pressure because of intensive illegal hunting. It is an unprotected freshwater swamp of 15 ha with an eutrophic waterbody, characterized by an extensive bed of floating vegetation of European white waterlily (Nymphaea alba) and emergent vegetation such as sea clubrush, common reed (Phragmites australis), common clubrush, branched bur-reed (Sparganium erectum), and Southern cattail (Typha domingensis). A small stand of tamarisk trees is found on its eastern shore. This natural environment is a suburb of the Jijel coastal town that has undergone demographic growth over the last two decades. Since 1997, the site has been surrounded by burgeoning housing and farms. Both Beni-Belaid Lake and El-Kennar Marsh are affected by the same human activities (e.g. agriculture, hunting, grazing, and fishing).

Sampling

To investigate the spatial and temporal variations of waterbird communities, we carried out monthly surveys over ten years, from October 2007 to December 2017, at the two main sites and their related areas. Surveys were carried out during the migration season (in autumn, from September to November), the wintering period (from December to February), the breeding season (from March to July), and the post-nuptial period (from June to August). The species and the peak in total number recorded over a month were all used for analysis. Birds were identified and counted using either: 7 × 50 binoculars or 20 × 60 Meopta telescope, and field guidebooks; we selected waterbirds from the avian fauna that are obligate users of the wetland. Other species were all excluded from the analysis. At each site, all birds were counted, so that the time effort was proportional to the size of the wetland. For each main site, we simultaneously counted its related sites to avoid errors due to bird movements between adjacent areas, the observation points were always the same. Methods of bird observation are numerous and dependent on the studied species and the purpose of the study. Depending on the circumstances, we combined two main methods in our field survey, the absolute method and the relative method. In the case of the absolute method the census is called exhaustive because the population was estimated directly on its absolute value and all the individuals were counted (point counts). An individual bird count was carried out when the group contained less than 200 birds and when it was at a distance of less than 200 m. This is because the area of the related sites was small enough for us to observe the whole site from one location with no optical obstacle, and one point counting was used (Gregory et al, 2004). Concerning the relative method, the population was estimated using sampling that includes only a part of the population. We estimated a sample of average size, then divided the field of vision into several transects (i.e. the flock size was divided into small equal blocks) and counted the number of blocks. This method is commonly used in winter counts of waterfowl when the individual number exceeds 200 birds or when estimating at a distance of more than 200m, as described by the "Scan Sampling Method" (Lamotte and Bourlière, 1969; Blondel, 1975).

Quality Control

For the transect survey, the perpendicular searching distances varied depending on the shape of the site, with 0.5 and 0.8 km in El-Kennar Marsh and Beni-Belaid Lake, respectively (i.e. depending on the size, and boundary location of the site). The vegetation type (emergent, submerged, grasses, shrubs, and trees) was determined by direct observation within the consistency of the point count stations. We also estimated the depth of the water in the shallow area from the length of the legs of shorebird species. Variability in observer error was minimized by measurements being performed by the same observer wherever possible throughout the study. Table 1 lists all the species included in the analyses.

Method steps

1. The study of the avifauna ecology was approached by determining ecological indexes directly linked to the balance of populations (Shannon–Weaver, Frequency and Uniformity). The monthly mean number for species richness and the total number of individuals was calculated for the whole period. These values were then used to calculate the diversity of species for the two main sites, expressed by the alpha diversity (Hα) of Shannon–Weaver index (H′). This index was computed using the following mathematical formula (Shannon–Weaver, 1949): H’ = -∑ pi ln pi where pi = (ni / N), N is the total population (individuals of all species), ni is the size of the population of species and S denotes the total number of species composing a population (specific richness). Frequency index: Relative abundance (RA %)is computed by using the formula: RA (%) = ni / N × 100, where: ni is the number of individuals of species taken into account and N is the total number of individuals of all species.. Species relative abundance status (table 1) was estimated from the frequencies of sightings over 12 months and the categories were assigned according to the Pettingill (1969) scale: abundant (90-100 %), common (65-89 %), moderately common (31-64 %), uncommon (10-30 %) and rare (1-9 %). Uniformity Index (Pielou index): this index measures the degree of numerical balance between species represented by the Equitability index (Evenness index). ‘E’ is calculated as the ratio: E = H′obs / H′max between observed diversity H′obs and maximum diversity H′max = log2(S), S being the total number of species recorded during each month (Pielou, 1966). The equitability index ‘E’ makes it possible to assess the imbalances that the diversity index cannot detect; it varies from 0 to 1. When E < 0.5, this means that the individuals mostly tend to belong to a single species. When E > 0.5, it means that species have similar or balanced abundance, the same abundance occurs when the value is exactly one (Okpiliya, 2012). For the entire study region, the mean (± standard deviation) was calculated for each ecological index (Shannon–Weaver diversity (H′), Equitability index (E), Abundance, Specific richness and Monthly population dynamics) based on all observations. The results were compared using analysis of variance (ANOVA), the choice of a parametric test analysis (two ways ANOVA) is supported by the Shapiro-Wilk test on the ANOVA residuals for the previous ecological index comparison which finds no indication that normality is violated. After this analysis, we used a Bartlett test on the ANOVA residual because the normality and homoscedasticity conditions were met: as the p-value was largely greater than 0.05, the assumption of homogeneity of the residual variances was therefore accepted. However, ANOVA led to the conclusion that there were differences between the means in some groups. Tukey’s (HSD) multiple comparison tests were used to compare the difference between each pair of monthly means. The choice of a non-parametric test (Friedman rank-sum test) analysis in the monthly abundance of shorebirds population is referred to that the condition of residual normality is violated (p= 0.001). These tests were performed using the statistical software R version 3•2•2 (R Core Team, 2015). Correspondence analysis (CA) was applied (bi-plot species and dates) to the data collected from the entire e eco-complex. CA is a two-dimensional solution that is well-suited to describe the spatial and temporal gradients in composition occurrence, status and assembly of waterfowl found at this ecological site. This last statistical analysis was carried out using the ADE-4 software package.

1. Accipitridae
   rank: family
2. Alcedinidae
   rank: family
3. Anatidae
   rank: family
4. Ardeidae
   rank: family
5. Burhinidae
   rank: family
6. Charadriidae
   rank: family
7. Ciconiidae
   rank: family
8. Haematopodidae
   rank: family
9. Laridae
   rank: family
10. Phalacrocoracidae
    rank: family
11. Phoenicopteridae
    rank: family
12. Podicipedidae
    rank: family
13. Rallidae
    rank: family
14. Recurvirostridae
    rank: family
15. Scolopacidae
    rank: family
16. Threskiornithidae
    rank: family

Fieldwork was conducted in Jijel eco-complex wetlands located in north-eastern of Algeria (36° 34'-36° 52'N, 5° 33'-6° 19' E), which are approximately 60 km2 in size. This natural wetland reserve comprises two major aquatic habitats that may vary in vegetation composition and structure, namely: (i) Beni-Belaid Lake (B-B.Lake) and (ii) El-Kennar Marsh (E-K.Marsh) (see map in fig. 1). This latter freshwater swamp is dominated by herbaceous rather than woody plant species. These mosaic hydrosystems are characterized by several types, mainly lakes, marshes, reservoirs, ponds, estuarine waters, and shallow seawater. Together, they form a complex ecosystem of wetlands.