Large-scale estimation and mapping of forest biodiversity indicators and ecological threats in the West Usambara montane forests
Loading...
Date
2024-05
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Sokoine University of Agriculture
Abstract
Assessment of the indicators of forest biodiversity, such as tree species diversity,
evenness, and richness, is crucial for monitoring and managing forest
ecosystems. Quantifying the relationship between environmental factors, such as
topography, soil, and climate, and tree species diversity and distribution is also
critical for understanding the pattern of the indicators of forest biodiversity and
informing conservation efforts.
Various studies have therefore, shown that environmental factors can significantly
influence tree species diversity and distribution in forest ecosystems, with
topography and soil moisture being important predictors of forest structure and
species richness. Such factors have also been used to develop models for large
scale predictions under different environmental conditions. However, in the recent
decades, remote sensing based had been considered to be as one among the
tools for facilitating the assessment of tree species diversity and distribution
patterns across large areas of dense montane tropical forests. By modelling the
relationship between remotely sensed data and tree species diversity, these
techniques can help identify priority areas for conservation and guide
management strategies. Furthermore, remotely sensed based techniques had
useful in assessing ecological threats such as wildfire which essentially play a
significant role in shaping the structure and composition of tropical forests.
This thesis consists of three manuscripts that assessed indicators of forest
biodiversity in the West Usambara montane forests of Tanzania. The first
manuscript aimed to determine the role of environmental factors on tree species
composition and diversity in the West Usambara montane forests. The field data
were collected through a two-phase systematic sampling approach, and
environmental data were obtained from USGS, ISRIC, and NCCS for
topographic, soil and climate respectively. The second manuscript aimed to
assess the potential use of remotely sensed data to model and monitor forest
biodiversity in the study area. The study computed field diversity, predicted
diversity using GAM and XGBoost models for Sentinel 2 and PlanetScope
imagery, and compared the efficiency of the sensors and models. The third
manuscript aimed at studying the post fire recovery of forest composition and
structure. It assessed the difference in structure and composition between burnt
and unburnt areas and lastly burnt area mapping was conducted.The results grouped the 183 identified tree species into three distinct forest
communities using cluster analysis. Indicator species analysis identified species
significantly associated with each community, like Dombeya burgessiae in the
higher elevation community. Environmental data on climate, soil properties, and
topography were compiled. Canonical correspondence analysis revealed
variables like precipitation, soil nitrogen, and elevation were influential in driving
community patterns. The communities differed significantly in diversity and
richness, with greater values in lower elevation communities. Species turnover
linked to environmental gradients was the primary contributor to beta diversity.
Overall, the study highlighted the importance of multi-scale abiotic factors in
shaping tropical montane forest communities. The findings have implications for
ecological monitoring and conservation efforts in these biodiverse yet threatened
ecosystems. An integrated assessment of climate, edaphic, and topographic
variables is therefore key to understanding the environmental forces structuring
tree communities and diversity in tropical montane systems.
Combining field-based approaches and remote sensing techniques provides
valuable insights into the factors that influence tree species diversity and
distribution in West Usambara Montane forests. The results highlight the strong
influence of soil factors such as pH and nitrogen on tree diversity and distribution,
while also demonstrating the potential of remote sensing data, particularly
PlanetScope data, for assessing and estimating forest biodiversity indicators. The
study recommends that conservation efforts should prioritize areas showing low
tree species diversity and take into account the influence of environmental factors
such as soil properties. The use of remote sensing techniques can facilitate the
identification of these priority areas and guide management strategies.
Additionally, further research is needed to explore the potential of other remote
sensing data sources and models to improve the accuracy and reliability of
biodiversity assessments.
Forest fires play a significant role in the diversity and composition patterns in the
tropical montane forests. It is therefore important to understand the manner
through which these ecosystems are affected by forest fires together with their
recovery patterns. From this study, it was observed that while fire did not
significantly alter overall tree species composition, it did reduce structural
parameters like density, basal area, biomass, and diversity compared to
unburned forest. However, these structural attributes exhibited recovery and
increased with time since fire. Indicator species analysis identified unique taxa in
burned areas. A 14-variable model integrating spectral, textural, and vegetation indices was used to map ~ 1430 ha of burned forest. The study demonstrated the
resilience of tropical montane forest composition to fire disturbance, although
more work is needed to fully understand post-fire dynamics and long-term
recovery. Overall, the study highlights the utility of integrating field measurements
and satellite data for assessing fire impacts in tropical ecosystems.
Description
MSc. in Forestry
Keywords
Large-Scale Estimation, Forest Biodiversity Indicators, Ecological Threats, Usambara Montane Forests