Species composition and abundance of rodents in Kafta-Sheraro National Park , Ethiopia : preliminary results

An ecological survey was conducted to assess the species composition and abundance of rodents in Kafta-Sheraro National Park, Ethiopia from June 2017 to April 2018 covering wet and dry seasons. Rodents were trapped using Sherman live traps from five selected habitat types—natural forest, bushland, grassland, farmland, and human settlement. A total of 209 individuals of rodents belonging to seven species were captured over 2940 trap nights, and three species of rodents were only observed. The trapped rodent species include Mastomys natalensis, Stenocephalemys albipes, Rattus rattus, Mastomys awashensis, Acomys cahirinus, Arvicanthis niloticus dembeensis, and Mastomys erythroleucus. Lemniscomys striatus, Hystrix cristata, and Xerus rutilus were only sighted during the study. Mastomys natalensis was the most abundant species (41.1%), followed by Stenocephalemys albipes (26.3%) and Rattus rattus (18.2%), whereas Mastomys erythroleucus (1.4%) was the least abundant species. The abundance of female (59.8%) was higher than male (40.2%). Of the total trapped rodents, 52.2% were adults, 32% sub-adult and 15.8% young. The abundance of rodents was varied among habitat types and between seasons. The results of the study revealed that the park has diverse species of fauna including rodents.


INTRODUCTION
Globally, rodents are the most diverse and abundant group of mammals. Over 32 extant families, 468 genera and 2,277 species of rodents are recognized throughout the world, which represents 42% of the mammalian species (Wilson & Reeder 2005;Wolff & Sherman 2007). Similarly, in eastern Africa rodents are the most abundant and numerous, which account for 28% of the total recognized species of mammals (Afework 1996;Kingdon 1997). Ethiopia is among the countries with a high resource of biodiversity and endemism due to its wide variation in altitude ranges, geographical positions and climatic conditions (Leykun 2000;Lavrenchenko et al. 2007). A total of 320 mammalian species were recorded in Ethiopia, of which 84 species are rodents with 21% endemic (Afework 1996;Lavrenchenko et al. 2007).
Rodents are successful mammals in every continent and able to exploit wide ranges of habitat types (Lange et al. 2004;Workneh et al. 2006). Habitat structure, disturbance, and other important environmental factors affect the distribution pattern and abundance of rodents (Tadesse & Afework 2008;Demeke & Afework 2014). The knowledge on global distribution and diversity of mammals are not complete, especially so for small mammals including rodents, insectivores, and bats, as many taxa are still being discovered (Kingdon 1997;Wilson & Reeder 2005). Many investigations were conducted to study the distribution and abundance of rodents in different parts of Ethiopia (Messele & Afework 2012;Adugnaw & Messele 2016). Many areas of the country, mainly the northern highlands and lowlands, are underexplored due to the remoteness, inaccessibility, harsh conditions, and political instability in protected areas such as Kafta-Sheraro National Park (KSNP), Dessa'a National Priority Forest Area and other local forests. KSNP is the newly established national park of the country and comprises diverse fauna and flora; however, scientifically its wildlife, particularly small mammal diversity including rodent community, is not entirely known. The present study aims to assess the species composition and abundance of rodents in the unexplored KSNP.

Study area
The study was conducted in KSNP (14.055-14.464 0 N & 36.695-37.675 0 E), northern Ethiopia (Fig. 1). KSNP is located in the lowland part of western and northwestern Tigray regional state known for the highest peak of the country-Siemen Mountain, Ras Dashn. The altitudinal range of KSNP is between 568m and 1,163m with an average elevation of 870m. The national park was formerly established as Shire Wildlife Reserve area in 1968 and recognized as a national park in 2007 to safeguard the African Elephant and other biological resources. Initially the total area of the protected area was 5,000km 2 but currently, an area of 2,176.43km 2 has been accorded protected area recognition. The national park is bordered with three districts of the region; on the west is Kafta-Humera, on the east Tahtay-Adyabo, and on the south Welkayit. Eritrea borders the northern part of the national park. The agro-climate of the national park is identified with warmer temperature inclined to semi-arid climate. The national park is characterized by a uni-modal type of rainfall regime with one wet season and one dry season. It receives an annual average rainfall of 600mm, and high rainfall is during the wet season from the middle of May to the beginning of September. The average monthly temperature of the national park ranges from 18 O C to 37.5 O C. KSNP consists of spectacular topographic features that support diverse flora and fauna. The vegetation of the national park belongs to Combretum-Terminalia, Acacia-Commiphora, savanna grassland, and riparian woodlands type. KSNP is a host for 42 species of mammals, 200 species of birds, nine species of reptile, and various unstudied species of different animals and plants (Ethiopian Wildlife Conservation Authority 2015).

Grid design and trapping
Before the field survey, rodent collection permission was acquired from Ethiopian Wildlife Conservation Authority, Government of Ethiopia, and a preliminary survey was carried out to select study sites. Relevant information such as topography, climatic conditions, flora, fauna and others were gathered. Besides, habitat types of the park were identified based on the vegetation classification of the country (Yalden & Largen 1992) and five habitat types (Image 1), namely, natural forest, bushland, grassland, farmland, and human settlements, were selected for the study. A permanent live trapping grid was established at each of these five habitat types, and the trapping studies were conducted both during the wet and dry seasons. Each grid constituted an area of 4,900m 2 (70 x 70 m). A total of 49 Sherman live traps (7.6 x 8.9 x 22.9 cm) were set per each grid at 10m intervals between points. In each of the habitats studied, 49 traps were set for three consecutive days

Rodents in Kafta-Sheraro National Park
Assefa & Srinivasulu amounting for 147 trap nights per habitat per season. During the present study 2,940 trap nights were spent in trapping efforts (147 trap nights x 2 seasons x 5 habitats). Peanut butter was mixed with crushed maize and rolled into small balls and then placed in the trap as bait. The traps were partially covered with leaves and grasses to prevent damage and death of trapped individuals. In order to locate the traps, each trap location was marked with yellow colored plastic tags on a branch of a nearby tree. Traps were checked early in the morning (07.00-09.00 h) and in the late afternoon (17.00-18.00 h) of each day for three consecutive days, and the bait was replenished as necessary to increase trapping success. Each trapped rodent was collected and relevant data such as season, habitat type, station numbers, species type, body weight, sex, approximate age, and reproductive conditions were recorded. Sex of trapped rodents was determined based on their nipples, testes, and the distance between their anus and urethral opening. Their age was classified into three categories-juvenile, sub-adult, and adult-and was determined based on their body size and weight, pelage color, and reproductive conditions (Taylor & Green 1976;Afework 1996). Reproductive conditions were determined based on the opening of the vagina (imperforated or perforated), size of nipples, and body weight for females, and the position of testicles (scrotal/ abdominal testes) for males (Taylor & Green 1976). Furthermore, necessary information such as morphometric measurements including body weight, lengths of head and body, hind foot, ear, tail, and other morphological characteristics of each trapped rodents were recorded. Representative species were collected as vouchers, and the voucher specimens were deposited in the collection of the Zoological Science Natural History Museum of Addis Ababa University, Addis Ababa. All trapped rodents were marked by toe-clipping and released at their capture site for further capture-markrecapture studies. Species identification was following standard references (Afework 1996;Kingdon 1997). The species identification was further checked using craniodental morphology and morphometric measurements. Voucher specimens were compared with the specimens in the collection of the Zoological Natural History Museum of Addis Ababa University, Addis Ababa.
The collected data were analyzed using SPSS version 22 computer software programme. Two-way ANOVA was used to test the significant variation of rodent species trapped across different habitat types, seasons, sex, and age groups. Species richness and diversity were computed using Shannon-Weaver diversity index (Shannon & Weaver 1949). Shannon-Weaver diversity

Image 1. Habitat types of Kafta-Sheraro National Park (KSNP), Ethiopia where rodent surveys were carried out; A & B -Natural forest | C -Grassland | D -Bushland | E -Farmland (Maize) | F -Farmland (Sorghum) | G -Human settlements (Adigoshu Military Camp) | H -Human settlements (Adebay Town). © Alembrhan Assefa.
index was calculated using the formula H΄= ∑ (Pi) ln (pi). Where, H' -Shannon-Weaver diversity index, Pi -proportion of the total individuals belonging to i th species in the sample, ln -the natural logarism and ∑ -sum of the calculations. Evenness (E) was calculated using the formula, E = H΄/ H΄ max . Where, H΄ max = ln (S), S is the number of species and H' is Shannon diversity index. We used PAST software (Hammer et al. 2001) to prepare the species accumulation curve (Mao Tau function) to assess representativeness of the sampling effort of rodents during the study.
M. natalensis and R. rattus were recorded in all habitat types (Fig. 2), while S. albipes and M. awashensis were captured in four habitats and absent from grasslands, A. cahirinus and A. n. dembeensis were captured in three habitats and absent from natural forests and farmlands, and M. erythroleucus was captured only from natural forest and farmland habitats. In the natural forests, S. albipes was the most captured species followed by R. rattus, while in all the other habitat types M. natalensis dominated (Table 2). There was significant variation in total number of captures with respect to species (Two-Factor ANOVA: F 6,4 = 19.339, p < 0.05) but not among habitat types (Two-Factor ANOVA: F 4,24 = 1.507, NS).
The species richness, diversity, and evenness of rodents varied among habitat types (Table 3). The highest species richness was recorded from bushland and human settlement habitats (6 species each), while the least was in grassland (4 species). The highest Shannon-Weaver diversity index was observed in human settlement (H' = 1.41), followed by bushland (H' = 1.25) and the least was in grassland (H' = 1.11) habitats. Grassland recorded the highest evenness (E = 0.800) and bushland had the lowest evenness (E = 0.695). The species accumulation curve for all habitat types (Fig.  3) has fully reach asymptote indicating that no more sampling effort is needed to capture all the expected rodent species of the park.

DISCUSSION
A total of 10 species of rodents were identified. Relatively similar representations of rodents were recorded previously in different protected areas of Ethiopia (Tilahun et al. 2012;Getachew & Afework 2015). High numbers of rodent species were reported from Alatish National Park, Chebera-Churchura National Park and Chilalo-Galama Mountain range (Tadesse & Afework 2008). Various levels of habitat productivity, habitat structure and complexity, risk of predation, food availability, human activities, and cattle grazing influence species richness. KSNP is highly disturbed by anthropogenic activities such as gold mining and forest destruction for agricultural expansion and charcoal production. Along with these biotic interventions livestock grazing is also rampant. These may be detrimental in rodent distribution within the park.
M. natalensis was the most abundant and widely distributed species in all habitat types studied, and was recorded in different habitat types including human settlements, and is considered as the most common pest species (Demeke et al. 2007;Tadesse & Afework 2008). Our observation on the abundance of this species agrees with earlier studies (Tilahun et al. 2012). However, higher abundance of M. natalensis has been reported in different areas (Manyingerew et al. 2006;Mohammed & Afework 2011;Mulungu et al. 2013). The high abundance of M. natalensis in different environments and habitats in its range could be due to its high reproductive potential, large litter size, ecological adaptability, and opportunistic food habit.
S. albipes was the second most abundant rodent species, is highly associated and abundant in natural forest habitats. Yalden & Largen (1992) reported S. albipes as one of the most abundant and common endemic rodent species of Ethiopia. Similar abundance was reported from other areas in Ethiopia (Mohammed & Afework 2011), however, low abundance of this species was also reported by others (Tadesse & Afework 2008;Tilahun et al. 2012). Rattus rattus was the third most abundant rodent species in the study area, and was mostly associated with human settlements, an observation that agrees with Moussa et al. (2015). This species was as least abundant Wonji sugarcane plantation and Chebera-Churchura National Park (Serkebirhan et al. 2011);however, Meheretu et al. (2014) reported this species to be abundant.
The abundance of A. cahirinus in this study was consistent with earlier reports (Tilahun et al. 2012;Getachew & Afework 2015). Unlike the present study, high abundance of A. n. dembeensis were reported in different parts of Ethiopia (Tilahun et al. 2012). M. erythroleucus is the most and important maize pest in eastern Africa (Odhiambo et al. 2005), and it was the least abundant species recorded in the study. Interestingly, many individuals of this species were trapped from natural forests. Similar findings were reported by Tadesse & Afework (2008). Afework & Leirs (1997) reported the highest abundance of this species in Central Ethiopia. Variations in the abundance of these species in different areas might be due to the variations in vegetation cover, food supply, and predators.
The highest species richness and diversity of rodents in bushland and human settlement habitat types contradicts with the findings of Demeke et al. (2007) and Moges & Dessalegn (2015), who reported that the bushland is structurally simple and comprises low biotic diversity. As compared to other habitat types of KSNP, bushland and human settlement habitats deliver relatively high shelter and food sources for rodents throughout the year.
The total number of captures and abundance of rodents was higher during the wet season than dry season, and agrees with those of Dawit & Afework (2008), (Serkebrerhan et al. 2011), and(Tilahun et al. 2012). Since the park is characterized by a uni-modal type of rainfall regime, high rainfall occurs during the wet season, as a result, there may be adequate shelter and food resources, which can contribute to high number of trapped individual rodents than the dry season. Several studies revealed that seasonal variation in weather conditions like rainfall affects the nutritional aspects, life strategies, and reproductive potential of rodents (Makundi et al. 2006;Tadesse & Afework 2008).
The sex ratio was found to be female biased, with more number of females captured than the males. Similar observations were reported by Tadesse & Afework (2008), however, in some areas male biased sex ratio was also reported (Manyingerew et al. 2006;Dawit & Afework 2008). This variation might be due to factors such as availability of food and microclimatic niches favouring females than males. Males being more abundant than the females could be due to their increased activity (Getachew & Afework 2015).
Out of the total trapped individuals, most captures were that of adults, and the least were that of juveniles. The probability of capturing more adults is may be due to their large home range, active movement, and fast capture ability. Adult animals have a higher social ranking, and cover wide home ranges, which causes higher capture rate than young individuals (Shanker 2001;Workneh et al. 2006). Higher numbers of individuals of all age groups were recorded during the wet seasons than dry seasons, and agree with findings of Makundi et al. (2006) and Getachew & Afework (2015). The seasonal variation among age groups might be due to the relationship between rainfall, seasonality in reproduction and seasonal based movement of rodents outside their home. During the present study, pregnant females of only four species were trapped. Higher pregnant females were recorded during the wet season than dry seasons, indicating that the more species reproduce more often in the wet season that may related with increased availability of food resources. Makundi et al. (2006) notified that seasonal variations in weather condition, mainly the rainfall, determine the nutritious features, which influences rodent reproduction.
During the present study, we have also observed that some rodent species including H. cristata, A. n. dembeensis, S. albipes, and M. natalensis are consumed as food locally by the indigenous Kunama ethnic group. Tadesse & Afework (2008) and Assogbadjo et al. (2005) also reported the economic importance of rodents as food source for Gumuz people, Ethiopia and Benin, respectively. This could be due to the high nutrient content in meat of rodents and hunting-based feeding habit of the people. The present study reveals that rodent diversity in KSNP is a reflection of the habitat quality and their composition and abundance in different habitat types and seasons varied. Long-term monitoring and detailed studies on ecological aspects of the rodents is suggested for future to provide more scientific information and insights on the rodent diversity in one of the least-explored and eco-sensitive protected areas of Ethiopia.