In order to calculate sediment yields, a chronology was established for sediments accumulating in two
farm reservoirs at single coring locations in each reservoir. The chronology has been transferred to six adjacent cores in each reservoir using magnetic core correlation, particle size and visual stratigraphy. Reservoir sediment volumes corresponding to variable surface levels behind the two dams were estimated using GPS-based measurement of the contemporary reservoir in-fill, combined with coarse assumptions about underlying
reservoir geometry derived from aerial photographs and topographic maps. Upper limits on volume were
calculated for the hypothetical case of a reservoir with near-vertical sides, and lower limits for the case of
linearly-sloping sides (i.e. a v-shaped cross-section). The arithmetic mean of these two boundary cases was taken
as the best estimate. Reservoir sediment volumes, and volumes of sediment accumulating between dated
horizons in the sediment cores, were obtained from the best estimate of changing sediment volumes derived
from these assumptions. The volume of sediment accumulating between each time zone was adjusted by
combining sedimentation rates with trap efficiency estimates. These data were subsequently converted to
specific sediment yields using the average sediment density. Results from the two reservoirs to which this
methodology has been applied show that the historical timing of increased sediment yields is different and that
yields have not declined significantly over the last 50 years, despite reductions in stocking density and the
abandonment of rain-fed agriculture