Improving the representation of the fate of harvested wood in global and regional carbon budgets

The rapidly shrinking carbon budget to stay within the Paris Agreement targets raises the importance of reducing CO₂ emissions from land use (E_LUC), while enhancing land-based carbon sinks. In this context, realistic estimates of E_LUC and land-based carbon dioxide removal require an accurate representation of the fate of harvested wood, including how much carbon remains on site as slash versus being removed for harvested wood products, and how long this carbon is stored.

For regional and global carbon budgets, like the Global Carbon Project’s annual Global Carbon Budget, E_LUC estimates are typically derived from bookkeeping models, which track carbon pools and their responses to land use change and land management. These models, however, currently represent the fate of harvested wood in highly simplified ways: They heavily rely on expert judgement and outdated data to determine slash and wood product pool fractions, typically only using a few generic wood product pools with lifetimes following an exponential decay, which poorly captures the dynamics of long-lived wood products. Here, we implement an improved representation of the fate of harvested wood in the bookkeeping model, BLUE, one of the three bookkeeping models used in the Global Carbon Budget. It is informed by the best available literature, FAO and IPCC data, and the LUH2 land use forcing. Our approach includes spatially explicit, partially time-varying fractions of harvested carbon allocated to slash and six wood product pools with different lifetimes, which capture specific product groups, and a computationally efficient two-step scheme that approximates more realistic gamma-function-like decay.   

Compared to the default implementation of harvested wood products, net E_LUC in our improved scheme differs substantially in regions where harvested carbon is stored for much longer periods, predominantly in the Global North: Over the last 20 years, net E_LUC is more than 30% smaller in Canada and ~10% smaller in Europe, China and Oceania. In contrast, net E_LUC is more than 40% larger in Russia, highlighting that longer storage of harvested carbon can also result in temporary emission increases, since carbon from past harvest or clearing events remains in the system for longer. In tropical regions, where soil turnover times are short and most wood products are short-lived, differences in net E_LUC are small. As tropical regions contribute most to global net E_LUC, this translates into a reduction of only ~3% at the global scale over the last 20 years and even less over longer time scales. 

Our results suggest that the simplified representation of the fate of harvested wood in current bookkeeping models is adequate for global E_LUC estimates. However, as regional and even national carbon budgets gain importance, our improved approach provides a more realistic basis for estimating E_LUC. Our approach is also transferable to land surface and Earth system models, which currently exhibit similar limitations in their treatment of harvested wood.