Seasonal changes in small mammal assemblage in Kogyae Strict Nature Reserve , Ghana

The small mammal community at Kogyae Strict Nature Reserve (KSNR) in the Ashanti Region of Ghana were studied in two habitats during the wet and dry seasons to investigate seasonal changes in species richness, abundance, composition and diversity. Ninety-six individuals belonging to nine species were recorded in 720 trap-nights, giving overall trap-success of 13.33%. Species richness (Sr), trap-success (Ts) and relative abundance (Ra) were higher (Sr = 6 species; Ts = 23.1%; Ra = 86.5%) in wooded grassland than forest (Ra = 4 species; Ts = 3.6%; Ra = 13.5%). However, species diversity was higher (Shannon-Wiener index Hʹ = 1.157) in forest than in wooded grassland (Hʹ = 1.089). Mastomys erythroleucus dominated in wooded grassland (68%) and Hylomyscus alleni in forest (53.8%). The species composition was unique for both habitats, with Mus musculoides being the only species common to both habitats. Seasonal changes in community assemblages were evident in both habitats, with species richness, diversity and abundance of the dominant species being highest in the wet seasons. Sex-ratio was unity in both habitats, and remained fairly constant throughout the rainy and dry seasons. Breeding activity was evident all-year-round for most species, but peaked in the rainy season. Our findings are consistent with that of other studies in Ghana and elsewhere in the African subregion, highlighting the importance of rainfall to the ecology of tropical small mammals.


INTRODUCTION
Identification of the factors that influence the distribution of species and temporal and spatial abundance, richness and composition of communities are of central importance in ecology, biogeography and biodiversity conservation.Knowledge of how natural environmental changes impact on organisms and how they in turn, respond to these changes can be used to forecast population trends, species turn-over and potential local extinctions (Soule et al., 2003;Manning and Edge, 2008).This in turn, can reveal subtle changes in environments and provide great insights into the threats facing biodiversity (Vos et al., 2000;Zahratka and Shenk, 2008), allowing for effective conservation planning and landscape management (Attum et al., 2008).*Corresponding author.E-mail: byofori@yahoo.com.
Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License Small mammals are important contributors of biodiversity and biomass of most natural and semi-natural ecosystems (Makundi et al., 2009;Habtamu and Bekele, 2012).They are the most diverse group of mammals, with considerable diversity in life-history, morphology and habitat associations.Rodents alone comprise over 40% of all mammalian fauna globally (Wilson and Reeder, 2005).Small mammals have complex effect on the structure, composition and functional diversity of their environment through various ecological interactions.For instance, by feeding on seeds, seedlings, fungal spores and insects, and serving as important sources of food for many medium-sized mammalian predators, raptors and snakes, small mammals maintain many food webs and ecosystem balance (Angelici and Luiselli, 2005).Some small mammals are also sensitive to even small changes in the environment (Malcom and Ray, 2000), as reflected in changes in their abundance, diversity and composition.Changes in the community structure of small rodents can therefore be used as surrogates for, and a quick and cost effective way of, measuring habitat quality or environmental disturbance (Avenant, 2011).Some small mammals are pests of agriculture and carriers of zoonotic diseases, causing significant economic losses and serious health implications (Habtamu and Bekele, 2012).Thus, the impacts of environmental change on small mammal populations have been the subject of intense research globally.
The abundance, diversity and community structure of small mammals are affected by several factors, including floristic composition, productivity, resource availability and microhabitat features such as available cover from predators (Nicolas and Colyn, 2003;Garratt et al., 2012).These factors in turn, are affected by climatic variability and disturbance regimes such as fire and habitat clearance (Jackson et al., 2009).In tropical and semi-arid regions, rainfall is often the most important driver of ecosystems' productivity (Coe et al., 1976).Small mammals therefore experience seasonal and inter-annual changes in abundance, composition and distribution tied to the amount and pattern of rainfall in space and time (Nicolas and Colyn, 2003).
In most tropical regions, natural habitats are being lost and degraded at alarming rates (FAO, 2007), posing serious threats to wildlife in general, and forest-specialist small mammals in particular, as these species require specific habitat structure and quality.The proliferation and mono-dominance of opportunistic species that are able to tolerate habitat modifications may further impoverish small mammal richness and diversity.It is feared that some tropical small mammal might be exterminated before they are discovered given the current rate of habitat loss in tropical regions.Therefore, it is of utmost necessity to conduct as many surveys as possible, particularly in unsurveyed areas that have experienced no or relatively less human modifications.
The composition of small mammal communities in many habitats and regions in Ghana is incompletely known.Knowledge of the effects of changing environments on this group of animals also is limited, despite recent efforts to bridge this knowledge gap (Yeboah, 1998;Decher and Bahian, 1999;Ryan and Attuquayefio, 2000;Decher and Abedi-Lartey, 2002;Attuquayefio and Wuver, 2003;Attuquayefio and Ryan, 2006;Barriere et al., 2009;Ofori et al., 2013a).A recent review of the status and challenge for conservation of small mammal assemblages of Ghana showed that 34 species of rodent (excluding squirrels, grasscutter), 14 species of shrews and one species of hedgehog have been recorded in the country between 1975 and 2014 (Ofori et al., unpublished).Most studies have been conducted in the southern part of the country.
As yet, no small mammal study has been published from the forest-savanna transition zone and the two upper regions of Ghana (Ofori et al., unpublished).
In this study, we conducted small mammal trapping in two habitats: (i) forest and (ii) wooded grassland during the dry season and the minor and major rainy seasons to investigate seasonal changes in species richness, abundance, diversity and composition at the Kogyae Strict Nature Reserve in the Ashanti Region of Ghana.

Study area
Kogyae Strict Nature Reserve (07° 12'N 01 ° 11'W), with an area of 386 km², is located in the forest/savanna transition zone in the Sekyere Central District of the Ashanti region.The climate is typical Transitional Woodland, with annual rainfall ranging between 1,200-1,300 mm (mean: 1,254 mm) and elevation of 120-130 m (WCMC, 2006).The area is bordered by the Afram River and riparian forest along its south-western boundary, as well as small pockets of dry forest and small rocky hills in the west.Much of the reserve has lost its status of "strict nature reserve", with an increasing number of farms encroaching from the south and east, as well as logging and hunting activities (Kyerematen et al., 2014).

Study site selection and trapping protocol
Small mammals were live-trapped in forest and wooded grassland, which form the major habitats in the study area.The wooded grassland is characterized by grasses and sedges, notably Sporobolus pyramidalis, Vertiveria fulvibarbis, Panicum maximum, Andropogon gayanus and Heteropogon contortus, with scattered trees.Trees in the wooded grassland included Daniellia oliveri, Ekebergia senegalensis, and M. multinervis, T. laxiflora.L. lanceolata, P. biglobosa and the palms like B. aethiopum.In each habitat, two permanent transects, each about 210 m long were established.Twenty standard Sherman collapsible live-traps (23 x 9 x 7.5 cm; H.B. Sherman Traps Inc., Florida, USA) were placed at about 10 m intervals along the length of each transect.Traps were baited with a mixture of corn meal and peanut butter, and were set during the day at about 16:30 GMT and checked the following morning from 07:30 to 10:00 GMT.Traps were set for three consecutive nights during the minor rainy season (September 2011), the dry season (January 2012), and the major rainy seasons (June 2012).There was therefore a total trapping effort of 360 trapnights per habitat and an overall effort of 720 trap-nights.
Small mammal trapping and handling protocols followed standard methods (Wilson et al., 1996;Martin et al., 2001) and complied with the animal care and use guidelines of the American Society of Mammalogists (Gannon et al., 2007).Captured individuals were identified on the spot (when possible), weighed to the nearest gram, sexed using the anal-genital distance (which is shorter in females) (Attuquayefio and Ryan, 2006) and checked for reproductive condition (scrotal testes in males and perforate vagina, enlarged nipples and pregnancy in females).Standard morphometric data, including head and body length (HB), tail length (TL), hind foot length (HF) and ear length (EL) were recorded.Each individual captured was marked by toe-clipping, before being released at the point of capture.Voucher specimens of species that could not be identified on site were sent to the University of Ghana zoological museum for identification.

Relative abundance (Ra)
This was estimated as the number of individuals of the ith species caught per 100 individuals.Thus, (2)

Diversity
For diversity of rodents, the Shannon-Wiener index (Hʹ) (Pianka, 1966) was estimated as follows: Where Nt is number of rodents captured, Ni is the number of individuals of the ith species, Tn is the total number of trap-nights and Pi is the proportion of the ith species in the total sample.

Species composition
The similarity of small mammal composition between wooded grassland and forest was computed using Sorenson's index (CN) (Krebs, 2001) as follows: Where a and b are the number of species at the first and second habitats, respectively, and c is the number of species common to the two habitats.The value of CN may range from 0 to 1, with a value of 0 (zero) indicating that the species composition of the two sites are distinct with no common species shared between them, whereas, a value of 1 means the species composition of both habitats are identical.

Overall trap-success, species richness, relative abundance, composition and diversity
Ninety-six individuals belonging to nine species were recorded over the entire trapping session, giving an overall trap-success of 12.63% and species diversity of 1.16 for the study area (Table 1).Eighty-three individuals (86.5%) of six species were recorded in wooded grassland, while 13 individuals of four species were recorded in the forest.The total trap-success was therefore higher (23.1%) in the wooded grassland than in the forest (3.6%).The total species diversity was however, slightly higher in forest (Hʹ = 1.157) than in wooded grassland (Hʹ = 1.089) (Table 2).M. erythroleucus recorded the highest captures (68.7%) in wooded grassland, and together with T. kempi, comprised 80.7% of the total number of individuals recorded in this habitat.H. alleni was the dominant (53.9%) species in forest (Table 2).Species composition was unique for both habitats (Sorenson's index C N = 0.2), with M. musculoides being the only species common to both habitats.

Wooded grassland
Species richness was higher in the rainy seasons, with both the major and minor rainy season recording five species each (Table 3).The total number of individuals (abundance), relative abundance and trap-success were highest in the minor rainy season (Ni = 34, Ra = 35.41%,Ts = 28.3%) and lowest in the major rainy season (Ni = 22, Ra = 22.91%, Ts = 18.3%).Species diversity was however, highest (Hʹ = 1.184) in the major rainy season and lowest in the dry season (Hʹ = 0.754) (Table 3).Three species, M. erythroleucus, U. ruddi and L. striatus were recorded in the major and minor rainy season and in the dry season.M. erythroleucus, the dominant species in the wooded grassland, was most abundant in the dry season with 24 individuals and least abundant (12 individuals) in the major rainy season.3).

Forest
In the forest habitat, species richness and diversity were highest (3 species, Hʹ = 1.099) in the major rainy season and lowest (1 species, Hʹ = 0) in the dry season.Relative abundance and trap-success were highest (Ra = 5.0%, Ts = 46.1%) in the major rainy season and lowest (Ts = 1.7%, Ra = 23.4%) in the minor rainy season (Table 3).
Praomys tullbergi was recorded during the major and minor rainy seasons only, while H. alleni was recorded during the major rainy and dry seasons, but not in the minor rainy season.H. alleni was also the only species recorded during the dry season.Malacomys edwardsi was recorded in the major rainy season only, whereas M. musculoides was recorded in the minor rainy season only (Table 4).P. tullbergi, M. edwardsi and H. alleni were equally abundant (33.33% each) during the major rainy season.M. musculoides and P. tullbergi also recorded similar relative abundance (50% each) and trap success  4).

Seasonal changes in sex-ratio
For individual species, sample sizes were too small to indicate any seasonal trends in sex ratio.Sex-ratio of M. erythroleucus, the most abundant species in the study area, was female-biased (♂1:♀2) in the major rainy season, male-biased (♂1.6:♀1) in the dry season and unity (♂1:♀1) in the minor rainy season.

Seasonal changes in breeding activity
Small mammals showed signs of breeding activity during the major and minor rainy season, with 93.3% of males having scrotal testes and 77% of females with perforate vaginas during the rainy seasons.All the captured males of M. erythroleucus had scrotal testes in the major and minor rainy seasons, but only 46% had scrotal testes in the dry season.The highest percentage (75%) of M. erythroleucus females with perforate vaginas was recorded in the dry season, and the lowest (18.2%) in the minor rainy season.All the individuals recorded in forest showed signs of breeding activity during the wet season.

Relative abundance, diversity and composition
The small mammal species richness at the KSNR (9 species) compared well with that obtained in the Accra Plains of Ghana (Decher and Bahien, 1999) and the Muni-Pomadzi Ramsar site in the Central region of Ghana (Attuquayefio and Ryan, 2006).However, it fell short of the 11 species reported from the Amansie West District of the Ashanti Region (Ofori et al., 2013) and the 14 species in the Western Region (Yeboah, 1998).The low abundance and diversity of small mammals recorded in this study could be, in part, due to the low trapping effort.It normally takes considerable trapping effort to capture naturally rare species that have low population sizes and small distributional ranges (Fichet-Calvet et al., 2009a).Placing the traps on the ground only, might also have resulted in a bias of trapping effort towards forestfloor dwelling species.Species composition at the study area was typical of West African small mammal assemblage (Yeboah, 1998;Decher and Bahian, 1999;Attuquayefio and Ryan, 2006;Fichet-Calvet et al., 2009a;Nicolas et al., 2010).The distinctiveness of species composition in the two habitats corresponded well with the habitat type.P. tullbergi, H. alleni and M. edwardsi, which are known forestassociated species, were recorded in forest only, whereas M. erythroleucus, L. striatus, T. kempi and U. ruddi are typical open grassland species with some preference for farmbush and forest clearings (Happold, 1987;Decher and Abedi-Lartey, 2002;Ofori et al., 2013a).Mus musculoides and M. daltoni are widespread habitat generalists that occur in a wide range of habitats and seem to depend on leaf litter (Decher and Abedi-Lartey, 2002).The high trapping success and relative abundance of M. erythroleucus in wooded grassland was not unexpected.The species is well-documented as very dominant in open grasslands and thickets, as well as in more arid areas (Decher and Bahian, 1999;Attuquayefio and Ryan, 2006;Makundi et al., 2009).The local and regional abundance of M. erythroleucus could be explained by its opportunistic feeding on grasses, leaves, seeds, seedlings, and insects, and its ability to adapt to modified habitats from open woodland to perennial grassy habitats (Decher and Bahian, 1999).
West African forests are usually dominated by either P. tullbergi or Hylomyscus sp.(Decher and Bahian, 1999).Most small mammal studies in forests in Ghana have reported P. tullbergi as the dominant species (Cole, 1975;Jeffrey, 1977;Garshong et al., 2013;Ofori et al., 2013b).In this study however, H. alleni was the most abundant forest-specialist species, even though the numerical difference in individual numbers between these species was not as large as their relative abundance scores might suggest.

Seasonal changes in species richness, abundance, diversity and composition
Seasonal variations in rodent communities were evident in the study area, supporting previous studies (Fichet -Calvet et al., 2009b;Makundi et al., 2009;Ofori et al., 2013a).The high species richness and abundance in the rainy season could be attributed to abundance of food and vegetation cover for small mammals during the wet season (Habtamu and Bekele, 2012).The high trapsuccess and relative abundance in the dry season could be attributed to the mono-dominance of M. erythrleucus, an opportunistic feeder, in the dry season (Decher and Bahian, 1999).

Seasonal changes in sex-ratio and breeding activity
Even though the male : female ratio of small mammals in this study did not deviate very much from unity, males were generally more abundant than females.This observation is supported by several other studies (Nicolas and Colyn, 2003;Garshong and Attuquayefio, 2013;Ofori et al., 2013ab).This may be because dispersal in small mammals is male-biased, increasing the chance of males encountering traps and getting captured (Garshong and Attuquayefio, 2013;Ofori et al., 2013b).
Overall breeding pattern of rodents in the study area was seasonal and related to rainfall.It is probably because the abundance of protein-rich diets like foliage, seedlings and insects, and lush vegetation cover during the wet season provide adequate security for lactating females and their offspring (Attuquayefio and Wuver, 2003;Nicolas and Colyn, 2003;Makundi et al., 2005Makundi et al., , 2009;;Fichet-Calvet et al., 2009b;Habtamu and Bekele, 2012).The findings of this study also showed that breeding activity in M. erythroleucus continued during the dry season indicating perhaps, a year-round breeding activity in this species (Fichet-Calvet et al., 2009b).Further studies will however be needed to confirm this, even though opportunistic feeding behaviour of the species coupled with a year-long breeding activity could account for its local dominance.
The present study is the first detailed survey of small mammals at the Kogyae Strict Nature Reserve, and the forest-savanna transition of Ghana.The findings of this study therefore will serve as baseline information for future monitoring programmes, which are necessary to evaluate impacts of environmental changes on small mammals.

Table 3 .
Seasonal changes in overall species richness, abundance, diversity and trapsuccess in the different habitats in the study area.

Table 4 .
Seasonal changes in species composition, relative abundance and trap-success of small mammals in the wooded grassland.