**2. Methods and materials**

#### **2.1 Climate and vegetation types**

The rainfall pattern is unimodal spanning from late November to May with a mean annual rainfall of 800–1200 mm in a north–south gradient. The mean annual temperature is 21°C, following the Köppen system [17]. The area consists of extensive miombo woodland, including *Brachystegia* sp., *Julbernardia* sp., *Isoberlinia* sp., *Afzelia quanzensis*, *Pterocarpus angolensis*, and rare and threatened plant species such as *Dalbergia melanoxylon,* which forms dense miombo along the hills and rivers [18]. Also, there are seasonal and permanent wetlands (swamps), riverine forests along numerous perennial and seasonal streams. Due to the increasing

anthropogenic activities, the area currently has farmlands and patches of wooded with scattered trees and grazing land.

Where A = number of species found in both communities, B = number of species only found in community 1 and C = number of species found in community 2. The equation returns a number between 0 and 1, where a number close to 1 indicates a higher similarity in species composition [23]. We then multiplied J by 100 to obtain

*Avifauna in Relation to Habitat Disturbance in Wildlife Management Areas of the Ruvuma…*

A total of 156 avian species representing 18 orders and 61 families were recorded

habitat types ranged from 2.28–4.08, except for dense miombo woodland which had H<sup>0</sup> = 1.69 (**Table 1**). Riverine forest habitat had higher species richness (n = 101 species), representing almost 45% of the total recorded individuals (**Table 1**). Avian species diversity was highest in riverine forest and lowest in dense miombo woodlands (**Table 1**; **Figure 2**). The Shannon Index of diversity revealed that species evenness for the five habitats surveyed was relatively low ranging from 0.29–0.59

Values bearing different letters within column are significantly different (p < 0.05) and values with similar letters within column are not significantly

Dense miombo 14 105 7.50 � 3.91 1.69 0.39 Farmland 40 580 14.50 � 5.82 2.46 0.29 Open miombo 98 1338 13.65 � 2.08 3.9 0.51 Riverine forest 101 759 7.52 � 0.97 4.08 0.59 Swamp areas 20 188 9.40 � 3.26 2.28 0.49

*Avian species diversity, abundance, and evenness in different habitats of WMAs in Ruvuma landscape*

**Mean abundance Shannon**

**diversity (H**<sup>0</sup>

**)**

) for all the

**Shannon evenness (EH)**

a percent, to easily interpret the results.

*DOI: http://dx.doi.org/10.5772/intechopen.97332*

**3.1 Avian species diversity, distribution, and richness**

in the five WMAs. The overall avian species Shannon diversity (H0

**Overall abundance**

**3. Results**

(**Table 1**).

**Table 1.**

**Figure 2.**

**75**

*Avian species diversity in different habitats.*

*(*� *standard error).*

**Habitat type Number of**

**avian species**

## **2.2 Sampling design**

Five sites of 200 m x 200 m were established in each WMA, making a total of 25 sites. We selected different habitat types for each of the five sites, namely miombo woodland (open and dense), farmland, swamps, and riverine forest.

#### **2.3 Avifauna survey**

Each site was sampled using three complementary methods to maximize the sample size. First, in each habitat type, avifauna counts were carried out using the point transects technique [6, 19]. This method consists of standing at a particular point or walking slowly across the site back and forth several times, to detect cryptic and skulking species in the area. These counts were repeated for 3 days, based on results from our pilot study, and the numbers for each site were averaged. A 20 minute counting period was used at each site, and the starting time (between 6:30 and 10:30 h) was rotated among the sites to reduce bias. Avifauna was identified by both sight and call, and numbers were recorded [20].

Secondly, the transect method was used. Three transects 40 km in length each were established in every WMA using existing roads. The locations of all transects were based on accessibility and were sampled using a vehicle driven at a speed of 20 km/hr. or less that stopped for each individual or group of birds encountered [21]. Two observers sighted and recorded all avifauna on either side of the vehicle and notes on habitat type were also taken [21].

Thirdly, mist-netting was used to the targeted cryptic, understory, and lower canopy avian species. Nets were erected and checked every 15 min in the early morning (between 6:30–10:30 h) and late afternoon (between 16:00–18:00 h). The total number of each species caught, and the associated habitat type was recorded. Each bird was marked with a drop of red permanent spray paints at the base of its toes on the right tarsi for verification, if recaptured, to avoid double counting [22].

### **2.4 Statistical analysis**

The biodiversity indices in different habitats or within these WMAs were obtained following Magurran [23]. This index uses three biodiversity indices including, diversity, richness, and abundance. A non-parametric Kruskal-Wallis test was used to assess whether there were significant differences in mean species abundance among five WMAs, and across each habitat type [24]. Differences in mean bird numbers between habitats in each WMA were tested using Mann–Whitney tests to assess whether the number of species was significantly lower in humanencroached habitat (farmland), i.e., farmland, compared to riverine forest, and dense and open miombo woodland habitats. Statistical tests were computed using the software package PAST [24]. For all these analyses, farmland habitat in this study represented human encroachment into protected areas and was used to compare with other habitat types found in the WMAs. We further calculated the Jaccard similarity index (Ji) between different habitat types to determine the level of similarities in species composition using the formulae [24]:

$$\text{Jaccard similarity coefficient (J);}\\\text{J} = \text{A/(A+B+C)}\tag{1}$$

*Avifauna in Relation to Habitat Disturbance in Wildlife Management Areas of the Ruvuma… DOI: http://dx.doi.org/10.5772/intechopen.97332*

Where A = number of species found in both communities, B = number of species only found in community 1 and C = number of species found in community 2. The equation returns a number between 0 and 1, where a number close to 1 indicates a higher similarity in species composition [23]. We then multiplied J by 100 to obtain a percent, to easily interpret the results.
