5. Conclusions
In agreement with many studies in the literature, we found that tree
planting on pasture was associated with a loss of SOC in the uppermost
soil layer. Whilst this study did not directly consider the cause of soil
carbon losses, two explanations are the loss of perennial grasses in the
understorey and an increase in soil respiration due to lower soil water
contents.
Fourteen years after planting, the woodland did increase carbon
storage as biomass, but about 37% of the increase in above ground carbon
storage was offset by the soil carbon losses at just 0–10 cm. Planting
the trees in a silvopastoral system also resulted in a loss of soil carbon in
the surface layer, but the greater size of the silvopastoral than the woodland
trees, meant that the soil carbon loss represented only 6% of the increase
in above ground storage. A simple analysis suggests that the
silvopastoral system was storing about 5% more carbon than the equivalent
separate areas of woodland and pasture. We suggest that the comparison
of the systems could also be further enhanced by using the
concept of carbon-time.
In this experiment we intensively sampled soil carbon over 0–
150 cm profiles, but we found no statistically significant differences in
soil carbon stock below a depth of 10 cm. Power analyses indicated
that despite the intensity of sampling, comparison of cumulative SOC
values below 0–40 cm had a N20% chance of falsely indicating nosignificant change. This low power issue needs to be recognised more
widely in the soil science community, as it casts doubt on a large body
of previous negative SOC measurements for which power analysis was
not completed. Whilst our results indicate that 30% of SOC stock in the
top 150 cm was stored below 60 cm, there remains a need to establish
method to statistically determine changes in carbon without excessive
replication.