PhD Thesis
Rodent pests cause significant losses on several cereal crops but more so on maize and
rice crops and thus pose a threat to the food security in Uganda and the rest of East Africa.
In order to develop an effective management strategy against rodent pests, it is important
to understand the diversity, breeding patterns and population dynamics and key
demographic and ecological factors that regulate abundance and richness of the major
rodent pest species. The above aspects formed the basis of this study’s objectives and
provide the first detailed ecological information in Uganda’s rodent fauna in agro
ecosystem environments.
Two studies were conducted; in the first study a 2-year Capture Mark Release (CMR)
rodent trapping was carried out in Mayuge district with grids placed in cultivated and
fallow field habitats with the aim of establishing the species composition and community
structure of small rodents, establish the population fluctuation and breeding patterns of
key rodent species and establish the demographic traits including survival and maturation
of key small rodent species in agro ecosystem environment. Sherman live traps were used
to trap the small rodents on a monthly basis and traps set for 3 consecutive nights at each
trapping session. Trapping was done on four permanent grids of 70*70 m2 measurement,
with traps spaced at 10*10m making a total of 49 traps per grid. The second study was a
trapping survey conducted across three districts with varying cropping systems and in the
different seasons with the aim of establishing geo spatial ecological factors that influence
rodent abundance. With the aid of Geological Information System (GIS) and remote
sensing 20 sampling grids in each district were randomly identified using a randomization
tool in Quantum GIS and were georeferenced and located with Global Position System
machine for field rodent trapping. Landsat 8 images were acquired from the United States
3
Geological Survey (USGS) website for land use land cover characterization of the studied
sites with further additional data collection on farm management practices.
A total of 11 identified small rodent species and one insectivorous small mammal were
recorded in both fallow and cultivated field habitats with the Mastomys natalensis being
the most dominant in the following order; M. natalensis (60.7%), Mus triton (16.1%),
Aethomys hendei (6.7%), Lemniscomys zebra (5.2%), Lophuromys sikapusi (4.8%),
Arvicanthis niloticus (0.9%), Gerbilliscus kempi (0.1%), Graphiurus murinus (0.1%),
Steatomys parvus (0.1%), Dasymys incomtus (0.1%) and Grammomys dolichurus (0.1%).
In terms of species turnover, spatially there was a significant difference (F1,6 =9, p=0.024)
for the studied field habitats with fallow field habitats showing higher species turnover
(6±1) compared to cultivated field habitat (4±1). Temporal species turn over (βT) also
showed a significant difference (F5, 44 =18.819, p=0.0001) over the three years of the
study, with her turnover in first of trapping. In terms of community structure, higher
species diversity associated more with fallow field habitats but also with certain rare
species found only in cultivated fields.
Higher diversity and species turnover in fallow fields could be explained by the
characteristic nature of this habitat having better vegetation ground cover and less human
interaction as compared to cultivated fields. On the other hand, the higher abundance
ranking of M. natalensis in cultivated fields could be associated to its characteristic nature
of being a good colonizer of disturbed habitats like farm lands. Based on these findings,
management strategies ought to target M. natalensis and should be applied in both
cultivated and fallow fields to prevent rodent infestation of crops. Further analysis on the
population structure of the most dominant species i.e. M. natalensis was performed on the
basis of seasonal changes in breeding patterns, population density and recruitment.
4
In terms of breeding, female animals were considered to be actively breeding when they
had perforated vagina, enlarged nipples or pregnant whereas males were considered
actively breeding in case of a visible scrotum.
Data on recruitment of new individuals in the trappable population was also generated.
Data on population density, percentage actively breeding animals and percentage
recruitment were subjected to Generalized Linear Mixed Model with a Penalized Quasi
Likelihood (PQL) method. Results on the breeding patterns showed continuous breeding
of females throughout the year, with significantly higher percentages in second dry season
(June-July), which was attributed to the observed availability of drying cereals and
legumes from main planting season of the first wet season (March-May). Population
density was observed to have peaked in the second wet season (September-November)
and recruitment was significantly higher in the second dry season, with the later attributed
to the increased animal breeding observed in the same period as reported above. Results
have important applications on the timing of management practices, with possible
recommendations to apply management options in the first wet season to prevent buildup
of populations in subsequent seasons that can result into damaging populations.
The demographic traits of M. natalensis in terms of maturation (the probability for a
juvenile to become an adult) and survival (the monthly probability for an individual to
survive from one month to the next) were also examined. CMR input data were subjected
to specialized statistical modelling using software E-SURGE. Best model estimates
showed higher survival of animals when rainfall was medium (300ml) to high (500ml) in
the past month. The higher survival during periods of higher rainfall was attributed to
possible increase in ground vegetation cover which is triggered by rainfall, thus animals
are shielded from potential predatory animals. Maturation on the other hand was not
5
affected by rainfall, sex, habitat and population density a condition which could be
attributed to the mixed cropping system in the study area thus offering a diversity of food
sources for animal normal growth and development. The recapture probability was also
modelled and higher recapture probability was observed in male animals compared to
females.
In the second study, data on small rodent population abundance and richness as the
response variables were subjected to the Boosted Regression Tree (BRT) model analysis
with several predictor variables including; Normalized Vegetation Index (NDVI), total
monthly rainfall, soil texture class, farm management practices (crop field status, crop
type etc.) to determine the relative importance of these variables on predicting small
rodent abundance and richness. Overall modelling showed farm management practices
(crop field status) was the most important factor for predicting both abundance and
richness. This could be attributed to the additive influence of both crop and field states
which have influence on food availability and suitable vegetation ground cover for animal
habitation. Other important factors including NDVI, crop type and soil texture (in
particular soil silt particle proportion) were important in predicting abundance and
richness. In conclusion this study has provided important information on several aspects
of the ecology of rodents associated with agro ecosystems, knowledge that can help in the
development of adaptive pest management strategies such as concentrating efforts on
M. natalensis species, timing control in the first wet season, promotion of collective
control even in non-cereal crops in Uganda and other regions with similar climatic and farming systems.
Bill and Melinda Gates foundation