Above Left: Example of net ecosystem exchange (NEE) measurements and light response curves (LRC) and ecosystem respiration (Reco) with a chamber connected to a LiCor 8100 CO2 analyser.

1. Top: heather plant community enclosed inside a 30 cm diameter perspex chamber (NEE) with air in and outflow and the chamber, sealing with Sphagnum moss and the light and temperature (shielded) sensor inside the chamber.

2. Below: shading using mesh to achieve two different shading levels (LRC) and a hood for complete darkness (Reco), which together with cut and uncut soil respiration fluxes will allow estimating plant above and below ground respiration.

Above Right: C export estimates for the three sites based on stream flow rates (CF: Control/Burnt; TF: Mown/Cut). Shown are average rates for the pre- versus post-treatment periods for the three sites (i.e., NMW); the annual C export losses are around 30 and 2 g C per m-2 for DOC and POC, respectively. Fluvial C-export rates are upscaled estimates based on averaging monthly samples with very few missing months due to no flow (drought or frost) at the flow weirs.

Below: Modelled average C budgets (NMW) during 2012-2024 comparing no management (uncut) to burnt and mown treatments. Shaded bars are for years with heather beetle defoliation/death (and therefore C loss). The dashed lines represent the projected C balance if no heather beetle effects had been observed (taking plots for sites with no beetle damage allowed to derive this curve). The  unmanaged (uncut) plots were largely unaffected by heather beetle (solid bars and line). Not accounting for this damage would be equal to not accounting for felling of trees during a storm (i.e., deriving the natural/potential C balance of a normal forest cycle requires excluding such data).  However, the beetle damage is important to consider as an extreme event - there are, so far, no such data of beetle damage on the C-balance (which seems substantial) - which would inform the Net Biome Carbon Balance (including extreme events). 

Above: Examples of mean monthly Light Compensation Point (LCP) values based on the Light Response Curves (LRC) models (solving to find the light (PAR) required to obtain a '0' C balance or NEE flux - for daytime period only) during 2012-2014. Note the overall seasonal patterns, inter annual differences, but particularly the generally decline in the order of Nidderdale (green) > Whitendale (purple) > Mossdale (blue). Note: a lower LCP implies greater C uptake.

Below: Interestingly, the differences in LCP (see above) seem to be more important in explaining differences in net C uptake (C budgets) than respiration as Reco (see below) was quite similar across sites for each management (and was reduced by mowing and burning) as shown for the corresponding mean monthly ecosystem respiration (Reco) values based on the Light Response Curves (LRC) models for the three sites during 2012-2015 and for the different treatments (uncut "do nothing" = dn; burning = bt; mowing = mn):

Below: The percentage flux of shoots, roots and peat (microbial) based on the total ecosystem respiration (Reco) as means for all three sites during all measurement days during 2012-2025 for un managed (uncut) and managed (burnt and mown with either leaving or removing brash) treatments. Subtracting total soil respiration (with roots) from Reco provided %shoot flux and subtracting fluxes from cut SR plots, which provided %peat (microbial) flux, from fluxes from uncut SR plots (i.e. including roots) provided %root flux. Note that no flux separation could be obtained during the pre-management period as soil fluxes from cut root locations showed increased fluxes reflecting root decomposition. Subsequent cutting was done twice a year (early spring and early autumn). 

Above: The individual C fluxes were used to predict continuous daily sums of the major carbon fluxes (NEE, GPP, Reco) based on light response curves using climate parameter regressions (see Appendix 6 in the Defra report: Heinemeyer et al., 2019) and net primary productivity (NPP) using the proportions for Reco flux components (above ground: Rab, total soil: Rs, root: Rr and decomposition: Rh) obtained from modelled hourly net ecosystem exchange (NEE) fluxes. Data show the averages for uncut (during 2012-2016) together with monthly carbon use efficiencies (CUE) (NPP/GPP) for Nidderdale (left), Mossdale (middle) and Whitendale (right).

Below: The individual carbon flux components for the uncut plots based on the above C flux figures in a summary table showing annual totals and averages for 2012-2016.

The light response curve models allowed predicting monthly and thus annual NEE C flux balance per site and comparing managements (uncut vs burnt & mown see above graphs). The sites show clear differences in net C uptake of unmanaged heather (Mossdale > Whitendale > Nidderdale) and, importantly, recovery of net C uptake post-management was faster on burnt than on mown plots, albeit heather beetle damage caused major issues, particularly at Mossdale and Whitendale.

The annual carbon flux balance (NEE) values were added up to derive cumulative C budgets, which was combined further with emissions from management (fuel used for burning and mowing management), and for burning biomass combustion and charcoal production (below table). Notably, burning showed a tendency to, over time, become a stronger C sink than mowing, after an initial higher C release. However, whilst Nidderdale (the driest site with fastest heather regrowth) showed a stronger overall C sink than mown plots, losses from biomass combustion still caused a substantial C source at Mossdale and Whitendale. The key seems to be to monitor an entire management rotation to be able to identify a possible optimum rotation length (e.g. for C storage) in relation to environmental site conditions. So far the data below show only 2012-2019:

Overall net ecosystem carbon balance (NECB) was calculated by combining NEE with fluvial C fluxes (DOC & POC) and methane emissions (below using the median flux). This was done per site (left; combining uncut and managed scenarios including the areas managed) and per individual management (right; assuming the same fluvial export for all as only partly burnt and mown catchment streams were measured). However, there is considerable uncertainty as to the inclusion of fluvial C fluxes in the NECB; notably, a considerable amount of the DOC export has been shown to be many years old whilst the CO2 emissions from rivers can be very old carbon, distorting the meaning of adding DOC export to the annual NECB in relation to the current situation (Billett et al., 2015): "Inclusion of the evasion flux term in the NECB would be justified if evaded CO2 and CH4 were isotopically  'young' and derived from a 'within-ecosystem' source, such as peat or in-stream processing of DOC. Derivation from 'old' biogenic or geogenic sources would indicate a separate origin and age of C fixation, disconnected from the ecosystem accumulation rate that the NECB definition implies".  Moreover, it is unclear how degradation at river conditions (light & temperatures) could have altered observed degradation measured at on land under air light and temperature conditions (Moody & Worrall, 2017), considering river temperatures can be different (warmer & cooler at the low and high range, respectively) and light penetration might be very different due to overall particle absorbance in a river from all angles. Therefore, the below tables show NECB values for both, including or excluding fluvial C losses (DOC & POC).

Finally, we did not consider the additional inputs of C via rainfall, as reported by Worrall et al. (2003) these are likely to be about 1.1 g C m-2 yr-1 for DIC and 3.1 g C m-2 yr-2 for DOC inputs. So, in addition to the uncertain losses of DOC and POC via fluvial export, about 4-5 g C should be added to the C sink (or subtracted from a C source). 

The below tables show the site and management values (averaged across the three sites) for 2012-2020 (note 2012 was pre-management).

Finally, the annual C budget data were correlated (below) with key environmental parameters (light, temperature and precipitation; left) and hydrological conditions (water table; right). This identified key parameters such as PAR (light) and precipitation but also a mean annual water table depth thresholds for switching from a net C sink to a net C source (for NEE about -12 cm & for NECB budgets about -8 cm):