Diversity and habitat association of medium and large mammals in Gibe Sheleko National Park, Southern Ethiopia

Abstract Complete documentation on the status of mammals is indispensable for appropriate conservation measures in protected areas. However, there is inadequate information on mammalian resources in the ecosystem of Gibe Sheleko National Park (GSNP). Thus, the study aimed to assess species diversity, abundance, and habitat association of medium‐ and large‐sized mammals in GSNP. We stratified the study area into five dominant habitat types, namely dense forest, wooded grassland, grassland, riverine forest, and farmland habitat types based on land cover and vegetation structures and further employed stratified random sampling technique across each habitat type. The sample transects covered 20% of the study area. Transect width ranged from 50 m to 400 m based on vegetation cover and visibility of mammals. The main data were collected via direct observation. Data were analyzed via chi‐square test and species diversity indexes. We recorded the total of 20 mammals species' those belong to 10 families of which 8 species were large‐sized and 12 species medium‐sized mammals. There were two IUCN vulnerable species, namely Hippopotamus amphibious and Panthera pardus, and two globally near‐threatened species, particularly Litocranius walleri and Caracal caracal in the study area. Dense forest held the highest species diversity of medium‐ and large‐sized mammals (H′ = 2.28) with the highest evenness index (J = 0.84). Riverine forest had the least diversity with uneven population distribution. Papio anubis was the most abundance species, whereas Caracal caracal was the least abundant in the study area. GSNP is home for threatened and spectacular mammals species'; hence, an appropriate conservation measure is mandatory to keep existing mammals species'.


| INTRODUC TI ON
Mammals are highly versatile group that include the fastest runners, deep divers, and most agile fliers, having colonized most of the Earth's habitats. Mammals constitute a substantial proportion of terrestrial and freshwater biodiversity (Mound & Gaston, 1993).
They have a critical ecosystem functions and as a consequence must be considered in the protected area monitoring systems (Vane-Wright, 1993;Woodroffe & Ginsberg, 1998).
Mammals are fundamental elements of most ecosystems. Large carnivores frequently shape the abundance, distribution, and behavior of prey animals (Lacher et al., 2019). Large herbivores function as ecological engineers by changing the structure and species composition of the surrounding vegetation (Owen-Smith, 1988).
Furthermore, both sets of mammals profoundly influence the environment beyond direct species interaction through cascading trophic effects (Crooks & Soulé, 1999). Large mammals perform important ecological functions and are good indicators of the habitat value because they do not typically rely on specific single habitat (Lacher et al., 2019).
Ethiopia is among the world's rich biodiversity countries with high level of endemism (Yalden & Largen, 1992). The variations in climate, topography, and vegetation have contributed to the presence of numerous endemic species. Ethiopia's high fauna diversity potential reflects the existence of many species of mammals. This in turn reflects diversity of habitats, created by the different combinations of elevation rainfall, geology, soil surface, and groundwater. For conservation of important biological resources, 27 national parks, 2 wildlife sanctuaries, 6 wildlife reserves, 25 controlled hunting areas, 5 biosphere reserves, and 8 community conservation areas have been established as refuge in different parts of Ethiopia (Tessema, 2019).
Ethiopian is endowed with 311 of mammal species belong to 14 orders of which 55 are endemic to Ethiopia. This endemism is much higher than other African countries (Lavrenchenko & Bekele, 2017).
In spite of the number of mammal taxa recording for Ethiopia has been increased, still there is no complete inventory and well documentation of mammals species' in various ecosystems of Ethiopia (Tefera, 2011). However, for sufficient management and protection of protected areas, information on the status and trends of mammals is mandatory (Qufa & Bekele, 2019). The knowledge of mammals' diversity, abundance, and habitat preference is the basics for the status determination and proposing appropriate conservation measures (Gonfa et al., 2015). Moreover, determining habitat associations of mammal and environmental features important for site occupancy is quite essential to understanding the basic ecology and community organization of mammals (Stephens & Anderson, 2014).
The study on species diversity and habitat association is the common tools used for ecologists and biologists in order to understand community structure and important for conservation efforts.
Gibe Sheleko National Park (GSNP) is one of the youngest protected area of Ethiopia that has been underexplored by ecologists especially on mammals species'. Hence, there is no adequate information about the mammalian resources in GSNP. To bridge these gaps, we intended to answer the following research questions: (a) what mammals species' are dwelling in GSNP, (b) how is the species diversity of medium-and large-sized mammals in GSNP, and (c) is there a relationship of abundance of medium and large sized mammals with the season and habitat.

| Study area description
Gibe Sheleko National Park was established as a National Park in 2009 to conserve diverse mammals and bird species. It is managed by Cultural and Tourism Bureau of the Southern Nation Nationalities Regional State. It is found 176 km far from southwest of Addis Ababa and 18 km from Wolkite town. It is geographically located between 7°54′00″ to 8°21′30″ N and 37°27′00″ to 37°45′00″E (Figure 1). It covers an area of 360 km 2 , which is characterized by heterogeneous landscape, flora, fauna, and habitat types. It is situated within three districts: Cheha, Abeshigie, and Enemurena-Ener. Its average rainfall ranges from 960 to 1,400 mm and altitude ranges from 1,050 to 1,835 m above sea level. The temperature of the study area ranges from 10 to 28℃ (Tilahun et al., 2017). The study area is classified within the climate zone of upper kola and dissected by deep gorges of the Gibe and Wabe rivers.

| Reconnaissance survey and study period
Before the actual study commenced, a reconnaissance survey was conducted in November 2015 for 7 days to gather general information about the vegetation types, topography, and accessibility of the study area through direct field observation and interview of local people and experts of GSNP.
Data collection was undertaken from December 2015 to January 2017 during both wet and dry seasons. According to the rainfall distribution of the area, from December to April was considered as dry season and May to October was considered as wet season.

| Study design and data collection method
The study area is characterized by heterogeneous vegetation types and topographic features. We stratified it into five dominant habitat types based on the types of vegetation structure and land cover features. These habitats include dense forest, wooded grassland, grassland, riverine forest, and farmland habitat. We classified dense forest habitat is the compacted natural forest with large canopy trees in the study area. Wooded grassland is areas covered by grasses with mix of large trees, while grassland is notoriously covered by grasses with scattered trees and twigs. Riverine forest habitat is the habitat which is covered by evergreen trees along Wabe and Gibe rivers' sides, while farmland includes human settlement and agricultural land along buffer zone of GSNP (Table 1).
The stratified random sampling technique was used via laying sample transects across the five stratified habitat types. The sample transects covered 20% (72 km 2 ) of the study area. We placed transect lines proportional to the area coverage of each habitat type (Table 1) (Chanea & Yirga, 2014). The transect length was measured and located in the study area with the help of GPS (Mengesha & Bekele, 2008). We employed the total of 42 transect lines, and an average transect length ranged from 3.6 km to 8 km. Transect width ranged from 50 m to 400 m based upon vegetation cover and accessibility of the sample site .
Moreover, the total number of 10, 15, 7, 5, and 5 transect lines was deployed in dense forest, wooded grassland, grassland, riverine forest, and farmland habitats, respectively. The adjacent transects were 1.5 km apart, all transect lines were roughly parallel to each other, and their ends were greater than one km far from the habitat edge to minimized edge effect (Regassa & Yirga, 2013).
We did the filed survey of medium-and large-sized mammals on foot across the selected transect lines. Simultaneous transect count was deployed to avoid double counting of mammals by involving many individual observers in mammals counting the same time.
We collected the data 4 times per seasons and included data collectors from wildlife expertise, ecologist, park scouts, and zoologist.
All data collectors received initial training on field count, mammals' identification, and application of field materials and field guidebook (Grimmett et al., 1988;Khanum et al., 1980). We deployed transect sample count method in the morning time from 6:30 to 10:00 a.m. and late afternoon from 5:00 to 7:00 p.m.
in which most mammals were active in the study area (Brower et al., 1990;Pomeroy, 1992). The survey was carried out by careful observation with the aid of binoculars and mammals guide book particularly Kingdon (1997).
The mammals that weigh an average weight of 2-7 kg were medium and all above this were categorized as large-sized mammals (Emmons & Feer, 1997).
Indirect detection indices such as hair samples, tracks, pellets, burrows, scratches, and other mammal's remains, for example, horn and skin, were used to know the presence and absence of nocturnal, rare, and elusive mammals species (Chanea & Yirga, 2014;Girma, Mamo, et al., 2012). We also used a key informant interview to collect general information about mammals in the area.  (Table 2).

| Species composition of medium-and largesized mammals
The total of 20 species of medium-and large-sized mammals that belong to 10 families and 4 orders (Primate, Artiodactyla, Carnivora, and Rodentia) was recorded in the study area. Among the four identified orders, order Carnivora and order Artiodactyla had eight species each, and three species of order Primate and one species of order Rodentia were identified in the area (Table 3).
The study indicates that 8 species were large-sized and 12 species medium-sized mammals.

| Diversity and abundance of medium-and large-sized mammals
This study shows that the highest species diversity (H′ = 2.28) of medium-and large-sized mammals was recorded in dense forest with the highest evenness index (J = 0.84). The least diversity index

Formulas Description
Where H′ is Shannon-Wiener index, pi is estimated as ni/N, where ni is the proportion of the total population of the ith species and N = ∑ni, this simply used proportions rather than absolute abundance values to reduce the effects of order of magnitude difference in mammals' numbers between species (Shannon & Weaver, 1949) J� = H�

ln(S)
Where J′ is Evenness index, H′ is Shannon-Wiener index and used the formula one, and S is species richness (Jarvis & Robertso, 1999;Magurran, 1988) Relative abundance = n N * 100 Where n is the number of individuals of particular species recorded and N is the total number of individuals of the species (H′ = 1.31) and evenness index (J = 0.63) were in riverine forest (Table 4).
However, it did not show significant variation between the seasons (χ 2 = 0, df = 1, p > .05) in the study area.
We found the highest species richness of medium-and largesized mammals in wooded grassland habitat (18 species), while farmland had the lowest species richness (7 species). The highest population proportion (45.2%) of medium-and large-sized mammals found in wooded grassland habitat. The lowest proportion was (7.9%) in riverine forest during the study period. Similarly, the abundance of medium-and large-sized mammals was highest in wooded grassland (89.5 ± 50.2SE) and the least was in riverine forest (13.3 ± 9.8SE) (

| Habitat association of medium-and largesized mammals
The study shows that medium-and large-sized mammal's abundance had no association with all the habitats (χ 2 = 0, df = 4, p > .05) of the study area. Nevertheless, it shows a significant difference (p < .05) in wooded grassland and riverine forest habitat types.

| D ISCUSS I ON
We found a total of 20 species of large-and medium-sized mammals during the study period in the study area. Legese et al. (2019) reported that 12 species of medium-and large-sized mammals were found in the similar ecological gradient of the study area, in Wabe forest fragment.

TA B L E 3
Species composition, conservation status, and relative abundance of medium-and large-sized mammals  According to Diriba et al. (2020) report there was a variation of species composition of mammals in different habitats. Our results also show that the highest species richness of medium-and largesized mammals recorded in wooded grassland, while the least was in farmland in the study area. High human interference (e.g., illegal settlement, considerable agricultural activities, and charcoal production) might account for this record in farmland habitat.
The highest population distribution of medium-and large-sized mammals found in wooded grassland habitat and riverine forest recorded the least distribution in GSNP. The highest distribution of mammals species' in wood grassland could be due to the movement of many species from the peripheral area toward the inner in search of water, food, and cover, for example, wooded grassland is used as corridor to move to dense forest and watering point in the study area. This variation of mammals distribution also links to availability major habitat requirements (Girma, Mamo, et al., 2012;Kasso & Bekele, 2017).
The abundance of medium-and large-sized mammals had no association with all habitat type of the study area, but we found a TA B L E 5 The abundance of medium-and large-sized mammals across the habitat types relationship of medium-and large-sized mammal's abundance in two habitats in wooded grassland and riverine forest. The distribution and habitat association of mammals are often correlated mainly with the availability of water, food, and cover. Therefore, variation of vegetation structure and composition across the habitat and most mammal species use riverine area at least to get water might be contributed for this result regardless of human disturbance in the two habitats of the study area. A mammals' abundance association with their habitat is often linked to the availability of basic necessity like cover and food for animals (Chanea & Yirga, 2014;Girma, Mamo, et al., 2012;Kasso & Bekele, 2017).
Out of the five dominant habitat types, dense forest harbored the highest species diversity of medium-and large-sized mammals with even distribution, whereas riverine forest had the lowest species diversity with less even distribution during the study period. However, Chanea and Yirga (2014) and Gonfa et al. (2015) reported that woodland habitat had the highest species diversity and evenness indexes. The presence of illegal agricultural encroachment and human settlement had an effect on medium-and large-sized mammals obliged to stay mostly in dense forest to hide themselves. Therefore, these human interference might contribute to record the highest species diversity in dense forest habitat in the study area.
We found many skins and old and broken horns of Greater Kudu and Red Bohor from local communities during field survey. However, we could not confirm the presence of these mammals by direct and other indirect counting techniques during the study period in the area. Perhaps, they have migrated to other adjacent area or forest patch due to high illegal settlement, agricultural expansion, and deforestation in the study area.

| CON CLUS ION
This study reveals that GSNP is endowed with the total of 20 species of medium-and large-sized mammals. Gibe Sheleko National Park harbored many IUCN Red-listed species of medium-and large-sized mammals.
The occurrence of such species shows that conservation relevance of the area needs urgent conservation action in collaboration with respective stakeholders.
The dense forest supported the highest species diversity of medium-and large-sized mammals with even distribution. Hence, dense forest is the most preferred habitat for most of the medium-and large-sized mammals due to its better cover and foraging opportunities.

Despite many old and broken horns of Greater Kudu and Red
Bohor found in the hands of local communities, no other evidence showed the presence of these mammals during the study period.
This might be due to the local extinction of species from the study area or migration to other nearby forest patches of the Oromia Region. Therefore, further study shall be undertaken to confirm the presence of these species and discover other mammals in the study area.

ACK N OWLED G M ENTS
The authors would like to thank the Wolkite University for financial and field materials support to undertake this research. We would also like to forward the deepest gratitude to all experts and responsible bodies of Gibe Sheleko National Park for assisting during field survey. Our gratitude goes to Mr. Habtamu Kerebeh for his valuable contribution to produce the study area map.

CO N FLI C T O F I NTE R E S T
The authors declared that there is no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data used in the study are included in this manuscript.