4.2. Explaining the effect of tree planting
Possible explanations for the observed decline in SOC below the
silvopastoral trees and woodland, compared to the pasture, include
pre-experimental variation across the site, pre-planting disturbance,
changes in ground vegetation, and bulk density and soil water effects.
These are considered in turn.
No soil carbon measurements were taken before tree planting and it
was not possible to implement a split plot design, so it is impossible to
completely rule out a pre-existing gradient of soil carbon. However,
we consider this to be unlikely for two reasons. Firstly records demonstrate
that the entire site had been under the same management for at
least 200 years prior to tree planting. Secondly, it is improbable that
the 20 randomly selected silvopastoral tree blocks would show a systematic
decline in carbon compared to the pasture. Whilst treatment related
bulk density changes can cause changes in SOC stock, these effects
were mitigated by the use of the equivalent soil mass method to determine
SOC changes.
High levels of pre-planting disturbance can result in slower gains
(and even losses) of soil carbon (Laganiere et al., 2010). However, the
trees at Clapham Park were all planted by hand (Burgess et al., 2000)
and there was no pre-planting preparation of the soil. Hence the disturbance
effect is expected to be minimal.
The loss of perennial grass cover in the woodland could explain the
observed decline in SOC (Fig. 9). This is consistent with intermediate
soil carbon content below the silvopastoral trees where large amounts
of solar radiation is able to reach the understorey from the sides of the
silvopastoral blocks. Hence although there was a loss of grass cover compared
to the open pasture, a less dense grass understorey was maintained
under the silvopastoral trees throughout the year (Fig. 9). By contrast in
the woodland, the extensive ground cover primarily of Gallium aparine
L. observed in April–May had died by August–September (Fig. 9).