Lake Tanganyika is a complex, tropical ecosystem in East Africa, harboring an estimated 2,100 species, many of which are endemic. Extensive watershed deforestation threatens the biodiversity and ecological integrity of the lake. In my dissertation research , I employed ecological and paleoecological methods to study the distribution of invertebrate biodiversity through space and time, with particular emphasis on linkages between biodiversity and land-use patterns. As part of my early dissertation work, I collaborated with several colleagues on ecological surveys of fish, mollusc, and ostracod crustacean diversity at sites in northern Lake Tanganyika that represented different levels of watershed disturbance. This study revealed a negative correlation between lacustrine biodiversity and intensity of watershed disturbance. To elucidate the long-term relationship between disturbance and biodiversity and to address the issue of causality versus correlation in this relationship, I subsequently focused on examining paleoecological records of invertebrates collected offshore from watersheds experiencing different degrees of landscape disturbance.

First, I compared life, death, and fossil assemblages of ostracod valves to assess the reliability and natural variability inherent to the paleoecological record. These comparisons indicated that paleoecological (i.e., death and fossil) assemblages reliably preserve information on species richness, abundance, and occurrence frequency at comparable-to-annual resolution. However, unlike life assemblages, species composition of paleoecological assemblages reflects input of species from multiple habitat types. Ostracod paleoecological assemblages are characterized by spatiotemporal averaging that renders them representative of larger areas and longer time spans than life assemblages. Thus, paleoecological assemblages provide an efficient means of characterizing longer-term, site-average conditions.

Natural variability in ostracod fossil assemblages from a sediment core representing the Late Glacial to the present indicates that abundance of individual ostracod species is highly variable. Ostracod assemblages were preserved during only the most recent 2,500 years of sediment deposition. Core geochemical data (elemental and stable isotopic data for carbon and nitrogen) indicate that the coring site may have been below the oxycline for ~2,000 years, inhibiting ostracod survival and preservation. Species composition of ostracod assemblages sensitively reflected lake water depth. A water depth index based on ostracod species composition in core samples accurately reconstructed previously described, past lake level changes and contributed new information on the timing of other lake level changes. The geochemical data further illuminated the Mid-Late Holocene history of climate change in this region of East Africa, indicating that a period of increased lake stratification may have been longer than previously thought.

Finally, in an applied collaboration intended to compare the effects of land use change on offshore benthic ecology, paleoecological, sedimentological, geochronological (Pb-210, C-14), and stable isotopic data were used to reveal differences through time in biodiversity and watershed disturbance offshore from a pair of watersheds in northwestern Tanzania. Cores were collected offshore from Gombe Stream National Park and a deforested watershed just outside the park. Offshore from the deforested watershed, sedimentation rates increased, indicators of increased soil influx increased sharply, and turnover in ostracod species composition occurred during the past 50 years. Comparable changes were not observed offshore from the Gombe Stream National Park.