The use of modi®ed soil lysimeters has been shown to provide reliable estimates of soil evaporation from either bare soil, or soil in which tree and/or crop roots are actively depleting surface soil water. In addition, the modi®cations to the lysimeter technique described in this paper demonstrate that it is possible to overcome some of the dif®culties usually associated with taking soil cores for lysimetry without disturbing the soil crust typical of soils in many semi-arid regions.
We were able to con®rm the hypothesis that soil evaporation was reduced when trees were incorporated into the cropping systems. Comparisons between rates of Es at two distances from the tree, under varying conditions of shade imposed by the tree canopy, showed a considerable in¯uence of tree proximity on both daily and cumulative soil evaporation rates. The nature of the changes in tree canopy size and hence ground cover was demonstrated, emphasising
This study determined the effect of two tree planting methods (woodland and a silvopastoral agroforestry system)
on the soil bulk density and organic carbon content of a grassland site in lowland England. Soil organic carbon
was measured in pasture, silvopastoral tree, and woodland treatments at six depths representative of 0–
150 cm. Fourteen years after tree planting, the organic carbon content in the surface soil layer (0–10 cm) was
greatest in the pasture (6.0 g 100 g−1) and least in the woodland (4.6 g 100 g−1); the value (5.3 g 100 g−1)
below the silvopastoral trees was intermediate. In the 10–20 cm layer, the organic carbon content in the woodland
was 13% lower than the pasture. No treatment effects on soil carbon were detected below 20 cm. Possible
reasons for the decline in surface soil carbon include a decline in grass cover and reduced soilwater content.Measurements
of above ground carbon storage by the trees indicated that tree planting increased overall carbon storage,
with the silvopastoral system predicted to achieve a higher level of carbon storage than equivalent areas of
separate woodland and pasture. A power analysis indicates that a prohibitively large number of replicates is
needed to ensure a lower than 20% risk of falsely concluding no treatment differences at individual depth increments
below 10 cm and cumulative depths extending below 40 cm.