Spatial and temporal variation in small mammal abundance and diversity under protection, pastoralism and agriculture in the Serengeti Ecosystem, Tanzania

Land use is an important factor influencing animal abundance, species richness and diversity in both protected and human-dominated landscapes. Increase in human population and activities intensify changes in habitat structure and hence abundance, species richness and diversity. We investigated the influences of land use and seasonality on small mammal abundance, species richness and diversity in 10 habitat types distributed over protected, agricultural and pastoral landscapes in the Serengeti ecosystem in Tanzania. We used live traps (n = 141) and capture-recapture methods in each of 10 fixed plots distributed across three landscapes for a total of 28,200 trap nights of effort. Trapping was carried out in the wet and dry seasons for two consecutive years (April 2017 to October 2018). Small mammal abundance was higher in the pastoral than in the protected and in the agricultural landscape. Abundance was higher in the dry than the wet season across all the three landscapes. Species richness and diversity were higher in the protected, middling in the agricultural and lowest in the pastoral landscape. The high abundance in the pastoral landscape was due to the numerical dominance of two species, namely A. niloticus in the shrubland and M. natalensis in the cropland habitat, resulting in low species richness and diversity. Abundance was more evenly distributed across all habitats in the protected area due to less disturbance. The low abundance in the agricultural landscape, likely reflects disturbance from cultivation. High species richness and diversity in the protected area indicate high habitat heterogeneity while high species diversity in the agricultural landscape was likely due to high food availability during and soon after harvests. These findings emphasize the importance of protection in maintaining habitat heterogeneity for wildlife. They also reaffirm the need for buffer zones around protected areas to cushion them from intensifying human activities.

landscapes for a total of 28,200 trap nights of effort. Trapping was carried out in the wet and 23 dry seasons for two consecutive years (April 2017 to October 2018). Small mammal abundance 24 was higher in the pastoral than in the protected and in the agricultural landscape. Abundance 25 was higher in the dry than the wet season across all the three landscapes. Species richness and 26 diversity were higher in the protected, middling in the agricultural and lowest in the pastoral 27 landscape. The high abundance in the pastoral landscape was due to the numerical dominance 28 of two species, namely A. niloticus in the shrubland and M. natalensis in the cropland habitat, 29 resulting in low species richness and diversity. Abundance was more evenly distributed across 30 all habitats in the protected area due to less disturbance. The low abundance in the agricultural 31 landscape, likely reflects disturbance from cultivation. High species richness and diversity in

Introduction 42
Human influence on ecosystems is increasing worldwide due to rapid population 43 growth and increasingly resource-consuming life styles (1). This influence has become so 44 important that mankind is now and, will likely remain for years to come, the main global driver 45 of ecological change (2, 3). Human-altered ecosystems made of various settlements, agro-46 pastoral and protected areas dominate the terrestrial biosphere, covering more than three 47 quarters of the total ice-free land areas (4). These alterations to ecosystems have resulted in a 48 global biodiversity crisis that threatens the world's species and ecosystems (5-7). Today, most 49 protected areas, set aside to safeguard the remaining global biodiversity, are surrounded by 50 different human activities making them isolated "islands". This change raises fundamental 51 questions concerning whether all protected areas will last into the far future given the current 52 rate of increase in human population and activities (8)(9)(10)(11)(12). Human activities that cause land use change also act as drivers of biodiversity loss (13). 55 Agriculture is the dominant land-use activity on the planet and is responsible for altering and 56 endangering wildlife communities on a massive scale (14,15). It has transformed native 57 vegetation into monocultures thereby decreasing biodiversity by homogenising habitats (16). 58 Although, agricultural activities can provide food to some wildlife species, a leading 59 conservation concern is that agricultural lands alter wildlife communities, favouring generalists 60 at the expense of specialists (17,18). On the other hand, livestock grazing, apart from 61 promoting vegetation regrowth and nutrient enhancement, causes mechanical disturbance, 62 reduces plant biomass and changes vegetation composition (19). The changes in vegetation 63 structure can have several knock-on effects on critical ecosystem functions, such as provision 64 of shelter and food for wild animals, species composition and richness (20)(21)(22). 65 Small mammals have long been used as bioindicators and model organisms to study patterns 66 of species abundance and diversity along different land use gradients (23-26). These studies 67 show that both grazing and farming activities differentially influence small mammal 68 community characteristics, such as species richness, diversity and abundance (14,18,(27)(28)(29)(30)(31).

69
In particular, in the Serengeti National Park in Tanzania, small mammal studies have focussed   70 on species and biotope, diversity and abundance in different habitats and along altitudinal 71 gradients (32, 33); human-small mammal conflicts (34) and influence of small mammals on 72 their predator abundances (35). A few studies have also compared protected areas with their 73 adjacent human-dominated habitats to infer the influence of anthropogenic activities on small 74 mammal species diversity and abundance (17,27,32). Assessments of the influence of human 75 activities on small mammal species diversity, richness and abundance have produced mixed 76 results, ranging from positive, negative to neutral effects. This is unsurprising given the 77 complex and dynamic interactions among ecological, historical, and evolutionary processes 78 shaping rodent diversity (36).
79 Surprisingly, few studies have sampled small mammals simultaneously between protected 80 areas and the adjoining human-inhabited areas across seasons (17,32). This study aims at 81 expanding upon the earlier studies by assessing spatial and temporal variation in small mammal 82 species diversity, richness and abundance in the protected and adjoining human-dominated 4 83 livestock grazing and agricultural landscapes in the Serengeti ecosystem. We address the 84 following two objectives. First, we quantify the species richness, diversity, abundance and 85 composition of small mammals in 10 habitats distributed across the three land use types. 86 Second, we analyse temporal variation (seasonal and interannual) in small mammal abundance 87 and diversity across the 10 habitats and three land use types. We anticipate that if disturbance 88 reduces structural and functional habitat heterogeneity then small mammal species diversity 89 and population density should be highest inside the protected areas, intermediate in the pastoral 90 lands and lowest in the cultivated areas. In addition, since small mammals exhibit pronounced 91 reproductive seasonality such that more juveniles are produced during the early dry season 92 (June and July) we expect to find a higher density of most of the species in the dry than the wet 93 season because of elevated food abundance linked to higher rainfall in the wet season.

94
Nevertheless, we anticipate that species should respond to human disturbance in contrasting 95 ways, such that habitat generalists should be able to colonize disturbed areas faster than habitat 96 specialists. Thus, we expect the abundance of habitat generalists to be higher than those of 97 specialists in the more disturbed pastoral and cultivated lands than the protected land.  The study covered mainly the northern part of the Serengeti ecosystem within three main 111 blocks located along the Mto wa Mbu-Musoma road. This area was selected because it contains 112 contrasting land use types, including agricultural areas (south west), pastoral and limited 113 agricultural areas in the south east and the Serengeti National Park situated in-between these 114 two blocks (Fig. 1). The Mto wa Mbu-Musoma road bisects each of the three blocks, resulting 115 in 6 sub-blocks; three sub-blocks on either side of the road. Based on habitat type, we selected 116 two study plots from each of the 6 sub-blocks resulting in 12 study plots. However, only 10 117 plots were included in the study because the other two plots (wooded grassland and grassland),    The study design was approved by Tanzania Wildlife Research Institute (TAWIRI) 135 and the permit to conduct research was obtained from Tanzania National Parks (TANAPA).

136
For the private land, the permit was issued by District Executive Office: Serengeti and 137 Ngorongoro Districts. All captured small mammals were handled according to the approved 138 permit and released immediately at the point of capture after observation. relatively low rainfall and so gets drier early compared to the western part. The same pattern 155 was followed except for one season (wet season 2018) due to logistical constraints, which 156 forced us to set traps in the protected area after the agricultural landscape.

158
Pitfall lines and trap lines were installed to capture mostly shrews and rodents, respectively.

159
Each plot was assigned one pitfall line consisting of 11 buckets, placed 5 m apart, and buried Checking of traps was done twice a day, early in the morning and evening. Equal amounts of 178 time were allocated to both methods, so we use 'trap-night' (one trap in operation for one 24-179 hr period, 0700 to 0700 hrs, to quantify sampling effort). We refer to the success rate of capture it has been generalized to diversity measures that incorporate species abundances and those 202 that take into account the evolutionary history among species (48). Hutcheson-t test was used 203 to test the significance of differences in diversity across the three-land use types and habitat 204 types.

205
Chi-square goodness-of-fit tests were used to test whether the observed abundances differed 206 significantly from expectation assuming a uniform distribution. Chi-square tests were followed by the chisq.multcomp post hoc test from the RVAideMemoire package (49). Abundance in 208 each of the three land use types were corrected for differences in trapping efforts and the results 209 presented as the number of small mammals/ 100 trap nights. In this study, the significance level 210 of 0.05 was adopted. In addition, evenness was high in both the AG (85%) and NP (60%) landscapes but low in the 221 PA (30%) landscape an indication of lower dominance in the NP and AG (S1Table). Species richness and diversity also varied noticeably across different habitats in the same land 229 use type and across the same habitat in different land use types. Specifically, in the NP species richness was the highest in the wooded grassland followed by the forest, grassland and 231 shrubland habitats, in decreasing order (Fig. 3a & 3b). However, these apparent differences in 232 species richness were statistically insignificant ( = 2, = 0.56). In contrast to richness,  (Table A1).   The abundance of small mammals also varied between different habitats within each land use 301 type, but the pattern of the differences was inconsistent across the three land use types. For the 302 NP landscape, the abundance of small mammals varied across habitats ( = 26.8, P = < 0.001) 2 3 303 such that it was lower in the grassland than in the wooded grassland, shrubland and forest 304 habitats (Fig. 5). For the AG, the abundance of small mammals differed significantly across 305 the four habitats ( = 19.6, P < 0.001) and was higher for the shrubland than for the other 2 3 306 habitats (Fig. 6). However, there was no difference in the abundance of small mammals among 307 the wooded grassland, cropland and grassland habitats or between the cropland and shrubland 308 habitats ( 0.1, P = 0.7). The latter two habitats had the highest abundance of small abundance in the habitat (78% of the total captures were of a single species), than it did in the 329 NP or AG landscapes (Table A2). For the cropland habitat, the abundance of small mammals 330 was higher in the PA than the AG landscape ( 67.6, P < 0.001). Likewise, abundance was 2 1 = 331 higher in the wooded grassland habitat in the NP than the AG landscape ( 30, P = 0.001). Small mammal abundance also varied interannually and seasonally (Fig. 7). Across all species, 334 abundance was higher in 2018 than 2017 37.7, P < 0.001) and in the dry than the wet One plausible mechanism is that grazing increases shrub cover and patches and hence nesting 378 and refuge sites for small mammals but reduces vegetation diversity and ground cover (54  In aggregate, these results support the notion that human activities, such as grazing and 452 agriculture, homogenize habitats. This is demonstrated by the higher abundance of A. niloticus 453 in the shrubland and M. natalensis in the cropland habitat. Both species are habitat generalists 454 able to expand their home ranges depending on seasonal food availability and to persist in 455 disturbed areas (57,68,69) .This conforms with the general view that human-dominated 456 habitats should harbour many generalist small mammal species (Byrom et al., 2015). Thus, by creating habitats that favor generalists at the expense of specialist species, human activities 458 modify ecosystem function and suitability for small mammal communities.