TESTATE AMOEBAE AND POLLEN AS PALEO-INDICATORS OF SURFACE MOISTURE CHANGES AT LOBOI SWAMP, KENYA

The East African Rift Valley is an arid to semi-arid region, with several large, well studied, alkaline lakes and abundant, small freshwater wetlands that have been commonly overlooked as paleoenvironmental and paleoclimatic records. These wetland systems are locally important freshwater sources and contain valuable archives, as the fragile ecology of these wetlands can be affected by even minor changes in the hydrologic budget (P-ET). Loboi Swamp is a 1.5 km² freshwater wetland located on a large alluvial plain between Lakes Bogoria and Baringo just north of the equator. P is ~ 700 mm/yr whereas potential ET is > 2500 mm/yr. Analysis of 25-years of precipitation data from local weather stations indicates that significant positive precipitation anomalies occur during El Niño years. Radiocarbon, pollen, and diatom data from a sediment core taken in the Typha region of Loboi Swamp indicates that the current wetland developed during the Little Ice Age on the uneven surface of the Loboi River flood plain. The swamp is dominated by Typha domingensis (~70%) on higher, seasonally wet sites and by a floating mat of Cyperus papyrus L. in topographic lows.

Sediment surface samples were collected for pollen, seeds, testate amoebae, TOC (loss-on-ignition) and carbon isotopic analysis. Environmental variables (water depth, pH, DO, conductivity and alkalinity) were collected from springs and swamp water and plant diversity and distribution were recorded along 15 transects throughout the swamp, with high diversity, 31 species, in the Typha zone and a monoculture in the Papyrus zone. The relationship of surface assemblage composition of pollen and testate amoebae to environmental parameters was established using ordination techniques. Results indicate significant qualitative differences in microfossil distribution within geographic regions of the marsh. These variations appear to correlate with changes in surface water depth related to the relict topography of the floodplain. A transfer function was applied to the fossil data to model hydrologic changes within the Loboi swamp over the last 700 years using data extracted from wetland cores. These results provide a preliminary, but promising high-resolution analysis of the spatial and temporal variability of swamp ecology as a record of climate change.