Masters Dissertation
Introduction
Bees are known to be the most important group of insects. They are economically and
ecologically important to humans. Bees provide various products such as honey, bee wax
and bee venom on which human society depends for livelihood. Bees further provide
pollination services to agro-ecosystems and natural habitats without which productivity in
these ecosystems would not be possible. The diversity and distribution of bee species is
determined by type, quantity and quality of suitable habitats in an ecosystem. Several
studies have confirmed the influence of climate change on bee distribution. Bee species
are expected to respond differently to climate change, either by range shift or
disappearance due to loss of their suitable habitats. This study was aimed at assessing the
distribution of bee species across habitat types and the influence of climate change on bee
populations in the Kilombero SAGCOT cluster of Tanzania.
Methods
Stratified sampling was employed, where the stratification was based on habitat type. The
habitat types were; Closed Forests (CF), Grasslands (GL), Woodlands (WL) and Agro-
ecosystems (AE). Transects were established within the strata and plots measuring 20 m x
40 m (0.08 Ha) were laid along the transects. The distance between transects was 300 m
and between plots was 200 m. Sweep netting and pan trapping were used to capture bees
within the plots. At each plot identification and enumeration of each bee species collected
was done. The identification was done later in laboratory using taxonomical keys. Bee
species composition was determined as the list of the identified bee species. Euclidean
distance measure of similarity and Raup-Crick dissimilarity index were used to compare
the similarity/dissimilarity in bee species composition between the habitat types. Bee
richness was determined as the number of bee species encountered, relative abundance of bee species was determined as the number of individual of each species relative to the total
number for all species. Bee species diversity was determined by the Shannon-Wiener and
Simpson Diversity Indices. Bee species evenness was computed by the Pielou's measure of
species evenness, (J = H'/ln(S) where H' is Shannon Weiner diversity and S is the total
number of species in a sample, across all samples in the dataset. Chi- square test was used
to compare bee abundance and richness between habitat types.
The influence of climate change on the distribution of dominant bee species was
determined using Maximum Entropy (MaxEnt) modelling. The four most dominant bee
species selected for modelling were; Apis mellifera, Meliponula ferruginea, Hypotrigona
ruspolii and Liotrigona bottegoi. Pearson correlation was used to determine
multicollinearity between the environmental variables. The environmental variable
obtained after running multicorrelation analysis were; Bio11, Bio13, Bio15, Bio18, Bio19,
Bio3, Bio4, Bio7 and Land cover. The highest scenario for GHG emissions (RCP8.5) was
used for future prediction of bee distribution for the years 2050 and 2070. Response curves
were used to determine the relationships between the environmental variables and the
probability of occurrence of bee specie. Jackknife test was used to determine the variable
of importance and percentage contribution of variable in influencing bee specie
distribution.
Findings
A total of 818 individual bees, belonging to 169 species from five families were collected
during the study period. The dominant family was Apidae and the least dominant was
Andrenidae. Euclidean distance measure of similarity shows high similarity of bee species
composition between woodlands (WL) and closed forests (CF). The bee species found in
Grasslands (GL) were more similar to WL compared to bee species found in the Agro-
ecosystem (AE). The bee species in AE were less similar to CF, WL and GL. Pairwise
comparison show the habitat types which were significantly different in bee species
composition were; CF and AE (p = 0.03), WL and AE (p<0.001) and WL and GL
(p<0.001). Bee species abundance differ significant in all habitat types (Chi sq = 5.34; d.f
= 3; p = 0.04). Shannon-Wiener Index of Diversity (H’) was highest in GL (H’ = 3.358)
and lowest in AE (H’ = 2.012) while evenness was high in CF (E = 0.422) and lowest in
AE (E = 0.141). This study revealed that bee species diversity, richness and abundance
vary across the habitat types in Kilombero, thus potential implication on habitat
conservation and bee species composition and diversity.
Precipitation of the wettest period (Bio13) contributed to about 70% in influencing bee
distribution on current and future climate change of highest emission scenario (RCP8.5)
for the years 2050 and 2070. The study predicted that A. mellifera and M. ferruginea will
lose their suitable habitats under future climate scenario of 2050 while H. ruspolii and L.
bottegoi will experience slight gain in suitable habitats. Under the future climate scenario
of 2070, all bee species will lose their suitable habitats.
Conclusion
There is apparently high bee species composition, richness and diversity in the Kilombero
SACGOT cluster and the distribution varied across habitat types. Climate change has a
significant influence on the current and future distribution of bee species in Kilombero
SAGCOT cluster with precipitation of the wettest period being the main climatic variable
influencing the future distribution of bee species. Loss of suitable habitats for most bees is
the major future climate impact on bee population thus conservation of suitable habitats for
bees is of paramount importance.
Recommendation
Mainstreaming these observations in the Kilombero Cluster Development Framework
implementation of different cluster value chains, it is important if we should have
sustainable production systems that consider both environment and agriculture production.
There is a need to develop climate adaptation and mitigation strategies for conservation of
bee populations in Kilombero which consider conservation of the potential areas where
predictions show a reduction in habitat suitability for bees. Bee conservation under the
changing climate needs to consider habitat connectivity to allow bee migration not only
between current suitable habitats, but also to the future suitable habitats as predicted by the
model. Long term monitoring in changes in the dominant bee populations is essential for
predicting future climate change response. Given the potential influence of habitat type on
bee populations, future studies should go further on studying the influence of vegetation
composition on bee abundance and diversity and their interactions.
Dissertation Structure
This dissertation is in publishable manuscripts format, which consist of four main chapters.
Chapter one comprise of the General introduction, Problem statement, Justification of the
study and Objectives. The chapter two (Manuscript one) is on the assessment of bee
species composition and diversity across different habitat types in Kilombero SACGOT
cluster, Tanzania. The chapter three (Manuscript two) is on the influence of climate change
on the distribution of dominant bee species in Kilombero SAGCOT cluster, Tanzania. The
chapter four is on the general conclusion and recommendation of the study.
Development Partnership Project DCP under Southern Agriculture
Growth Corridor of Tanzania SAGCOT