The water table treatment had, as expected, a significant effect on the CO2 flux
from shallow cores; fluxes from cores with a low water table where higher than those from cores with a high water table. However, in the deep cores there was no significant effect of water table treatment. This could be because the C in these deeper layers has become highly recalcitrant, due to the drainage of the sites which has led to long term aeration in the field (Laiho, 2006). Fluxes from the long cores were also not significantly different between the low and changed water table treatments. Other studies have also shown higher CO2 fluxes from
cores with lower water table than from cores with high water table, however these studies did not look at different core depths (e.g. Dinsmore et al. 2008; Estop- Aragonés et al. 2016; Blodau et al. 2004; Moore & Roulet 1993). The contrasting flux response to water table depth (and hence aeration of pore spaces in peat) indicate some fundamental differences in peat from superficial or deeper soil layers. Particularly at our sites, where trees had been present over preceding years (or in case of forestry sites where still present), bulk density has been affected by layers of needle litter on the surface. This lower bulk density in superficial peat depths is likely to allow a much stronger aeration effect from lowered water table compared to higher peat bulk density at greater depth, so that the oxygenation of peat pores in response to a lower water table may have a much smaller effect here. There were no significant differences in CH4 flux across all three core depths
145 expected and with the literature where studies have found higher CH4 fluxes in
high water table treatments than in low water table treatments (Aerts and Ludwig, 1997; Dinsmore et al., 2008; MacDonald and Fowler, 1998; Moore and Dalva, 1993) and where a change in water table from low to high has led to a pulse of CH4 flux (Dinsmore et al., 2008). It is possible that a short-term flush of CH4 was
missed in our study (1-2 days after water table change), but overall, the lack of CH4 flux response is surprising. This could potentially be because the average
water table depth in the field for the forest plantations is -40 cm and -10 cm in the bog (Table 2.1), this means that the low water table in the incubation study is not really that low and this could have led to the lack of water table treatment response in the CH4 fluxes. White et al. (2008) also did not find a significant
effect of water table treatment in their bog mesocosms, but they did find a significant effect in their fen mesocosms. Field results from the same sites show increasing CH4 fluxes from the forest plantation to the near pristine bog, with the
restoration sites in-between. Here we hypothesised that this was due to the increasing water table from the forest plantations to the near pristine sites (chapter 2), but this lab incubation study shows that most likely there are different drivers as well.
4.6 Conclusion
We show that forest plantations have altered the quality of the peat and nutrient availability in the pore water. Different CO2 fluxes between sites under the same
temperature and water table indicate that the chemical and physical legacies of the forest plantations shape the biogeochemical processes in peatlands. For CH4
fluxes only very few differences between sites emerged, with only two of the restoration sites displaying significant differences, which indicates that on its own (and in absence of biotic interactions under field conditions), forestry effects on CH4 flux are limited. We have found both generic and some site-specific
predictors for both CO2 and CH4 fluxes, but it was difficult to interpret consistent
changes in peat composition and water table depth in light of flux responses. It appears that site-specific conditions, possibly linked to detailed management during periods of forestry, or linked to the method of forest removal seem to override global controls, which makes prediction of the data challenging.
146 However, the complicated results could also be a statistical artefact and more robust statistical testing is needed to determine the relationships between biochemicals and fluxes.
147
5 General discussion
Peatlands are a globally important C store (Stocker et al., 2013), which can be compromised by drainage and afforestation (Lindsay, 2010). A better understanding and awareness of the importance of peatlands for ecosystem services has led to a change in land management (Andersen et al., 2016; Anderson et al., 2016), and an increasing number of afforested peatlands are now being restored to enable recolonization of peatland species and a return to ecosystem functioning (Andersen et al., 2016; Lunt et al., 2010).
There is only very limited data on GHG fluxes of peatland restoration sites in the literature (e.g. Rowson et al. 2010; Abdalla et al. 2016), and only one study from a forest-to-bog restoration site, which focuses on CO2 fluxes only (Hambley, 2016).
There is also very limited knowledge on how afforestation alters the peat biochemically and how this in itself influences the GHG fluxes of restored sites. The rate of peat decomposition under forest plantations on naturally treeless peatlands is also unknown and knowing this can help us understand and model the effects of drainage in afforested peatlands on peat oxidation rates in boreal peatlands.
The work presented here attempts for the first time to produce a GHG flux balance of forest-to-bog restoration in the UK, addressing an important land use policy question. GHG emissions are reported for the UK under the terms of the United Nations Framework Convention on Climate Change (UNFCC). GHG emissions have to be reported in climate change mitigation reports to show what kind of attempts are made to achieve the targets of GHG emissions to reduce global warming, agreed on by countries around the world in the Kyoto protocol (Morison, 2012). Large areas of afforested peatlands are undergoing restoration in the UK; since 2000, forest-to-bog restoration was conducted at a rate of 500 ha per year and more will be restored in the future (Anderson et al., 2016) as government-funded grant schemes are now in place to restore peatland habitats impacted mainly by drainage and afforestation. However, until now the UK was unable to provide net GHG numbers for the forest-to-bog restoration sites.
148 The main findings of this thesis are:
1) Forest-to-bog restoration impacts mainly on CH4 flux, while both CO2
respiration and N2O fluxes are unchanged over a chronosequence of
restoration sites. Net CH4 fluxes are lowest in forest plantations and increase
with restoration age, being highest in the near pristine bog.
2) Peat decomposition rate under the forest plantations is 126.8 ± 14.7 g C m-2 y-1, which is 44% of the total soil respiration. Hence, 56% of the total soil
respiration came from the tree roots (autotrophic flux).
3) Forest plantations have altered the quality of the peat and nutrient availability in the pore water. Different CO2 fluxes between vegetation free
peat cores from different sites for the same climatic conditions show that this shapes the biogeochemical processes in the peatlands. However there were very few differences in CH4 fluxes between vegetation free peat cores from
the different sites under the same temperature and water table level, indicating that on its own (and in absence of biotic interactions under field conditions), forestry effects on CH4 flux are limited.