Detailed network descriptions of the proposed forest carbon monitoring system

 

Intensive network

High cost prevents the extension on an intensive network to a sufficient number of sites to verify national-level or district-level estimates of carbon stocks and stock changes. However, these are the only types of network which provide complete measurements for all relevant forest carbon pools. These measurements could be used to validate estimates of carbon stocks derived from carbon accounting modules or more coarse-scale monitoring networks, and as such, the contribution made by intensive networks is essential.

The Long-term Carbon Flux Network (CFN) is an elegant approach to monitoring of the net carbon flux of forests but does have a number of limitations. Firstly, the time-step at which flux assessments have been made is much finer than required for a five-year reporting period. Fine scale assessments are routinely integrated, requiring the implementation of statistically-based gap-filling procedures to replace missing data. The gap-filling procedure should be independent of any process-based model which might itself be calibrated using the flux data.

There are also issues associated with data storage and quality assurance for the CFN approach as discussed by Aubinet et al. (2000). Protocols would need to be developed to adjust raw CFN measurements to account for loss of carbon through DOC and POC as well as through removal of biomass as HWP. The most logical way to achieve this would seem to be through further integration of the CFN with the Level II (L-II) network.

The L-II network is already well placed to take on the role identified for an intensive network in validating models and coarse-scale estimates. However, inclusion of flux-tower measurements at some suitable sites would provide valuable verification of assessments made using the L-II protocols on the basis of a completely independent measurement procedure. Site selection requirements for flux towers are exacting, including an approximately homogenous 500 m fetch in relatively simple terrain. These requirements would severely limit an expansion of the CFN within the current L-II network. 

Intermediate network

The principal role of a network such as that formed by Permanent Mensuration Sample Plot Network (PSP) and UK Woodland Assurance Standard (UKWAS) plots is probably in the calibration and validation of forest carbon accounting models and underlying growth models. This role will be especially important if carbon accounting models (or benchmark estimates derived from them) are to represent the diversity of stand conditions in the UK, particularly in terms of management regime.

The emphasis should be placed on improving and verifying representation in carbon accounting models of tree and stand biomass / carbon dynamics in response to alternative silvicultural prescriptions. The presumption is that soil carbon stocks, while influenced by management of above-ground biomass, do not vary greatly in response to management interventions, except in extreme cases such as afforestation of previously bare ground or long-term denuding or complete removal of forest cover. Accordingly, protocols for the periodic assessment of litter and soil carbon at a relatively low precision may need to introduced, particularly since information on the carbon content of forest soils is poor at the national level.

The strategic development of the PSP network (Matthews et al., 2003) with less emphasis on monitoring and priority given to providing data for model development underlines the role for an intermediate network as one of support for model calibration and validation. However, there may be a secondary application for estimates obtained from an intermediate network in verifying more coarse-scale carbon stock estimates derived from models from an extensive network. Any assessments of litter and soil carbon made in an intermediate network, albeit with limited precision, would be particularly important for this type of application.

It is likely that at least a proportion of the existing EU Level I network (L-I) sites would be upgraded and incorporated within the intermediate network. Furthermore, a sub-set of the intermediate network might be selected to provide more detailed information of the soil/litter carbon pools and act as benchmark sites for validating the extensive network.

Extensive network

Compared to intensive and intermediate networks, extensive networks such as Forestry Stewardship Council (FSC) and L-I are weak in terms of the completeness and precision of the assessments for estimation of carbon stocks. However, their strength lies in their potential to cover a representative range of sites, tree species and management regimes.

High cost would prevent complete assessments of carbon stocks being made, even with limited precision, but the extensive network would permit the broad-scale verification of national and district estimates of carbon stocks derived from carbon accounting models.

For reasons elaborated in the discussion of the intermediate network, emphasis in extensive plots would be on assessment of forest biomass with only limited consideration of litter and soil carbon, possibly through collection of input data for a litter / soil carbon model such as one based on RothC (Falloon and Smith, this report), or by using soil type as a surrogate for carbon stocks based on measurements in the intermediate network. The verification process would consist of a plot-by-plot comparison of estimates made by the carbon accounting model with assessments from the extensive network. Inherent in this application would be a network fully representative of the UK forest estate and probably based upon an extension of the 16 x 16 km transnational grid.

Alternatively or additionally, depending on the configuration of the network, it may be possible to compare upscaled estimates at national or district level derived independently using the carbon accounting model and statistical analysis/synthesis of estimates made in the extensive network.

The FCS and L-I networks and the National Inventory of Woodlands and Trees (NIWT) ground-truthing plot network would require careful adaptation and harmonisation of protocols in order to provide robust estimates of carbon stocks. In the case of the FCS network, a plot-based assessment protocol would need to be incorporated, perhaps taking a modified L-I plot design as a minimum specification. The 0.25 ha assessment area used for structural assessments in NIWT ground-truth plots would be ideal, but introduction of even abbreviated mensurational assessments for all trees in a plot of this area in all situations would be prohibitively costly.

As with the FCS network, one solution might be to include a modified L-I plot within the 0.25 ha area. In general, if a unified plot design could be adopted across FCS, L-I and NIWT ground-truth plots, either based on the L-I system or otherwise, then the extensive carbon monitoring network could be formed from the combination of these three networks or from appropriate sub-sets of plots in each network. In any event, a unified protocol for assessment of mensurational (and other) variables to support carbon stock assessment would need to be developed. This could be based on assessments of tree dbh and either tree or stand height, depending on stand structure and density following protocols similar in principle to abbreviated tariffing as described by Edwards (1983).

The precision of abbreviated mensuration procedures has been estimated to be within ±20% (Edwards, 1983). Analyses carried out as part of this review suggest that stand-scale allometric relationships used as part of these procedures are subject to bias, effectively inflating uncertainty to ±30% or more. However, scope exists to improve on or refine existing procedures and underlying allometric relationships. Research could be carried out to support the revision of relevant procedures as part of their adaptation to biomass estimation. Improved allometric relationships could also be derived, eliminating bias and ensuring that precision is within ±20%. It should also be noted that although estimates of standing stem carbon are uncertain in young crops, the uncertainty falls as the expected standing carbon stock increases. This reflects the small proportion of stem biomass compared to total biomass and uncertainty in allometric relationships at early stages of development.