Predicting carbon sequestration of woody biomass following land restoration

Land restoration through planting native species or facilitating natural regeneration may provide opportunities for sequestration of biomass carbon in many regions where woody vegetation has been cleared or largely supressed. Australia provides a good case study of how sequestration from these activities may be simulated at national- and project-scales using an empirical full carbon accounting model; FullCAM. Recent work has resulted in 1.6-fold increase in calibration sites available, while the underlying input layer of site productivity has also been recently refined. These developments provided an opportunity to expand the domain of application and capability of the model. We collated 2340 observations of above-ground biomass in planted or naturally regenerated stands across Australia and applied a novel technique to correct these observations for a baseline biomass of existing remnant trees or shrubs within these stands (typically < 5 Mg DM ha−1 and < 10% of individuals). Five different sets of parameters were estimated based on five categories of stands for the FullCAM model, based on landscape position, land use, and configuration and/or density of the stand: (i) stands accessing water in addition to rainfall; (ii) belt plantings of high (>1500 stems per hectare) stand density; (iii) belt plantings of lower stand density; (iv) blocks of plantings or natural regeneration on land used to deliver environmental services, and; (v) blocks of natural regeneration occurring on land used for livestock production. Compared to blocks of plantings or natural regeneration, yields of biomass after 30 years of growth were 1.3- to 2.2-times higher in belt plantings, and stands of mesic regions accessing ground or surface water, e.g. riparian or floodplain zones. After 30 years of natural regeneration, when compared to stands fenced-off for conservation, yields of biomass were 1.5-times lower on stands that continued to be managed for livestock production. In revising the model calibrations, the trade-off between model accuracy and utility was considered. By accounting for only stand age, site productivity and the five stand categories, the overall model efficiency of biomass prediction was 68%. Additional explanatory variables such as stocking densities of stands with a block configuration, stand establishment method, species or species mix, and belt width were also tested, but their addition to the model was considered unwarranted, given the additional resources required to account for these inputs would be substantial, and they would deliver <10% improvement in prediction efficiencies. Although predictions were unbiased overall, they may nonetheless provide erroneous predictions at a given site. Nonetheless, given the absence of bias, individual over- and under-predictions will tend to cancel out for carbon accounting at the national-scale; while at the project-scale where carbon offsets are monetarised, discounting for uncertainty may be implemented.