No abstract

No abstract

No abstract

Rainfall controls fire in tropical savanna ecosystems through impacting both the amount and flammability of plant biomass, and consequently, predicted changes in tropical precipitation over the next century are likely to have contrasting effects on the fire regimes of wet and dry savannas. We reconstructed the long-term dynamics of biomass burning in equatorial East Africa, using fossil charcoal particles from two well-dated lake-sediment records in western Uganda and central Kenya. We compared these high-resolution (5 years/sample) time series of biomass burning, spanning the last 3800 and 1200 years, with independent data on past hydroclimatic variability and vegetation dynamics. In western Uganda, a rapid (<100 years) and permanent increase in burning occurred around 2170 years ago, when climatic drying replaced semideciduous forest by wooded grassland. At the century time scale, biomass burning was inversely related to moisture balance for much of the next two millennia until ca. 1750 ad, when burning increased strongly despite regional climate becoming wetter. A sustained decrease in burning since the mid20th century reflects the intensified modern-day landscape conversion into cropland and plantations. In contrast, in semiarid central Kenya, biomass burning peaked at intermediate moisture-balance levels, whereas it was lower both during the wettest and driest multidecadal periods of the last 1200 years. Here, burning steadily increased since the mid20th century, presumably due to more frequent deliberate ignitions for bush clearing and cattle ranching. Both the observed historical trends and regional contrasts in biomass burning are consistent with spatial variability in fire regimes across the African savanna biome today. They demonstrate the strong dependence of East African fire regimes on both climatic moisture balance and vegetation, and the extent to which this dependence is now being overridden by anthropogenic activity.

Understanding palaeotopographical variability forms the basis for understanding prehistoric societies.Alluvial and lacustrine environments, in particular, are key areas with both a high archaeological and palaeoecological potential. However, the often deep stratification of these sites, the high water table and the complex sedimentological variations can hamper a detailed reconstruction of the spatial relationship between prehistoric settlement and their environment. Combining different remote and proximal sensing techniques and coring data, can offer detailed insight into such landscapes. More specifically, the integration of mobile geophysical methods allows the collection of unprecedented continuous information on large-scale palaeolandscape variability. In this study we present a combined approach in order to map and model prehistoric landscapes and river systems in and around a Late Glacial palaeolake in north-western Belgium. Based on filtered and unfiltered digital elevation models, a survey area of 60 ha was selected, in which detailed mobile multi-receiver electromagnetic induction survey was conducted. The results allowed for the delineation of palaeochannels in the area and enabled modelling the depth of these features in the survey area, providing insight into their flow characteristics.14 C sampling enabled the dating of the evolving river system to the transition between the Late Glacial and the Early Holocene. Through additional coring, this river system could be traced further through the palaeolake area. Based on these results a detailed reconstruction was made of the palaeotopography that harboured the Final Palaeolithic andEarly Mesolithic occupation of the study site.