This study was carried out to investigate the species diversity and habitat use of birds in the Menagesha Suba State forest and surrounding farmland. The study was conducted from July 2018 to January 2019 during the wet and dry seasons. The study area was stratified based on the dominant habitat types. A standardized survey technique was employed using systematically established point counts for all habitat types. EstimateS software (version 9.1) and Chi-square test were applied to analyze the data. A total of 122 bird species belonging to 14 orders and 49 families were identified in the study area during the two-season surveys. There was a statistically significant difference in the relative abundance of species among habitat types (χ2=81.928, df = 2, p<0.05). During both seasons, the highest bird diversity was observed in farmland (H’=3.65), followed by plantation forest (H’=3.52). The lowest and highest distributions were observed in natural forests (J=0.77) and plantation forests (J=0.89), respectively. Variations in the number of bird species were observed among the three habitats. Species similarity was highest between natural forests and plantation forests during both the dry and wet seasons. There was a statistically significant difference in habitat use of bird species among the three habitat types either when seasons were considered separately (dry season: χ2 = 22.825, df = 6, P<0.05; wet season: χ2 = 22.186, df = 6, P<0.05) or when combined (χ2 = 26.658, df = 6, P<0.05). The Menagesha Suba State forest is rich in endemic bird species to Ethiopia and shared with Eritrea, and more than 15% of bird species distributions are locally rare. There is a need for a detailed study of long duration on the diversity and other ecological aspects of forest bird species should be conducted to get exhaustive information.

1. Sampling Design

A stratified random sampling design was used to study bird species composition, abundance, habitat use, and relative distribution. The study site was stratified into three dominant habitat types: dry evergreen afro-montane forest, plantation forest, and farmland. A total of 80 sampling plots were systematically generated by a geographic information system (ESRI, 2012) using QGIS V2.18. The study area covers 9557 hectares, and the area is stratified into three dominant habitat types (Natural forest, Plantation forest, and Farmland). To avoid edge effects, sample plots were established 300 m away from the edge of each habitat type (natural forest, farmland, and plantation forest). Sampling plots were established proportional to each habitat type; as a result, in the dry evergreen Afro-montane forest 34 sampling plots, in the plantation forest 10 sampling plots, and farmland 36 sampling plots were established. The radius of each point count was fixed at 50m at the center of the sampling plots. To avoid double-counting of the birds, sampling plots were reasonably spaced out at 500 m in natural forest and plantation forest and 1000 meter in farmland (Bibby et al., 1998; Sutherland, 2006). Following Yosef Mamo et al. (2016), the Chao 1 richness estimator was computed in EstimateS to determine sampling adequacy. This estimator computes the total number of species (observed plus not observed during the survey though present in the area) expected to be present in a particular area (Colwell et al., 2012). Thus, the ratio of observed richness to Chao 1 estimated richness gives the proportional number of species recorded during the survey relative to the expected total number of species (Colwell, 2013).

2. Data Collection

Data collection was carried out in two seasons; the wet season (July and August 2018) and the dry season (November 2018 and January 2019). In each season (dry and wet) two terms comprised a total of four data collection sessions. In each session each sampling plot in each habitat type was visited three times, making up a total of 6 visits of each sampling plot per season, and each sampling plot was visited 12 times. The observers start for all dominant habitats uniform from the southern point to the northern point because of the slope of the area that easily walks one point to other points without challenges. Observers stand in the center of one point and count over nine minutes per point count before walking through the area (Bibby et al., 1998).

Data were collected early in the morning from 6:00 to 10:00 am and late in the afternoon from 3:00 to 6.30 pm when birds are more active (Bibi and Ali, 2013). There were no censuses during rainy and cloudy days because such climatic conditions significantly affect the activities of birds and make the identification of bird species difficult (Bibby et al., 1998). While the observer stands in one location the bird species observed, the number of individuals, time of observation of vegetation (habitat) characteristics (species, height, cover type) of perching trees, and perching distance and activity of birds were carefully recorded. In addition, auxiliary data such as latitude and longitude, elevation, slope, and aspect were recorded using clinometers and GPS (Garmin 60). Birds were detected with the naked eye and with the help of a pair of binocular (Nikon 10*50). For identification of species present in the study area, plumage pattern, size, shape, color, songs, and calls were important parameters (Bibby et al., 1998; Sutherland, 2006; Shimelis Aynalem and Afework Bekele, 2008). Identifications of species are confirmed by using the Field guide book (Redman et al., 2009), and photographs and videos are taken to justify the species type which difficult to identify. Finally, the identified bird species are properly grouped into their taxonomic category.

3. Data Analysis

Species diversity was computed in three approaches: species richness (number of species present in a particular area), evenness (relative abundance distribution of species), and Shannon diversity index (a heterogeneity measure that includes both species richness and evenness) (Colwell, 2013). Species richness was computed in EstimateS version 9.1 software (Colwell, 2013), using an individual-based sampling procedure. This was computed for each of the three habitat types separately for each season (i.e., dry and wet seasons) and pooled season, as well as for each season based on pooled habitat. As the treatments (habitat types or seasons) differed in sample size (i.e., number of individual birds recorded in each treatment category), both rarefaction of the observed number of species and extrapolation methods were used to calculate species richness and compare between respective treatments. Extrapolations were made up to twice the smallest (or smaller) sized sample size (i.e., twice the number of individuals recorded in the site with the smallest number of individual birds were recorded) (Addisu Asefa et al., 2017). In both rarefaction and extrapolation methods, mean richness was computed along with 95% confidence intervals based on 100 times sample randomization and used for comparisons between treatments. Following Addisu Asefa et al. (2017), two treatments were considered to be significantly different if they had non-overlapping 95% confidence intervals of species richness. Similarly, Chao 1 richness estimator was computed in EstimateS to determine sampling adequacy. This estimator computes the total number of species (observed plus not observed during the survey though present in the area) expected to be present in a particular area (Colwell et al., 2012). Thus, the ratio of observed richness to Chao 1 estimated richness gives the proportional number of species recorded during the survey relative to the expected total number of species (Colwell, 2013).

Shannon index of diversity: It was being calculated using the formula below:

H^I=-PilnPi

Where Pi: The proportion of the i^th species to total abundance value Pi= ni/Ni.

Ln Pi: the natural logarithm of Pi.

Evenness (Shannon-Wiener's Equitability) index (J): It was calculated using the formula below:

J=H^IH^max

Where; H^max= lnS Where, H^I is Shannon-Wiener's diversity index, H^max is the maximum value of H^I and S is Species richness.

Simpson’s Diversity Index ‘D’ was calculated using the formula below:

D=1-^Pi2

Where Pi: The proportion of the i^th species to total abundance value Pi= ni/Ni.

Bird Species Similarity

Sorensen’s similarity index was calculated using the following equation:

Cs=2aba+b

Where a is the number of species found in site A; b is the number of species in site B and ab is the number of species shared by the two sites (Bibi and Ali, 2013).

Relative Abundance

The relative abundance of each species in each habitat type was calculated using a simple formula:

RAi=niTni  

Where, RAi = relative abundance of species i; ni = number of individual birds (abundance) of species i recorded in a particular habitat; and Tni = the total number of individual birds recorded in that habitat.

Based on these RA values, each species was classified into four relative abundance categories: Abundant (the top abundant species whose sum of RA values equals 0.25), Common (the second top abundant species whose sum of RA values equals 0.25), Uncommon (the third top abundant species whose sum of RA values equals to 0.25), and Rare (the least abundant species whose sum of RA values equals to 0.25) (Addisu Asefa, 2014). Then, the number of species classified in each relative abundance category was computed and the difference in the number of species classified in each relative abundance category was tested using the chi-square test in SPSS version 20 software. This was conducted separately for each season and season combined.

The data sets analyzed during the current study are available from the corresponding author (H.T) upon request.