Description of a new procedure to estimate the carbon stocks of all forest pools and impact assessment of methodological choices on the estimates

摘要

Forest ecosystems play a major role in atmosphericcarbon sequestration and emission. Comparableorganic carbon stock estimates at temporal and spatial scalesfor all forest pools are needed for scientific investigationsand political purposes. Therefore, we developed a new carbonstock (CS) estimation procedure that combines forestinventory and soil and litter geodatabases at a regional scale(southern Belgium). This procedure can be implemented inother regions and countries on condition that availableexternal carbon soil and litter data can be linked to forestinventory plots. The presented procedure includes a specificCS estimation method for each of the following forest poolsand subpools (in brackets): living biomass (aboveground andbelowground), deadwood (dead trees and snags, coarsewoody debris and stumps), litter, and soil. The total CS of theforest was estimated at 86 Tg (185 Mg ha-1). Soil up to0.2 m depth, living biomass, litter, and deadwood CSsaccount, respectively, for 48, 47, 4, and 1 % of the total CS.The analysis of the CS variation within the pools acrossecoregions and forest types revealed in particular that: (1) theliving biomass CS of broadleaved forests exceeds that ofconiferous forests, (2) the soil and litter CSs of coniferousforest exceed those of broadleaved forests, and (3) beechstands come at the top in carbon stocking capacity. Becauseour estimates differ sometimes significantly from the previousstudies, we compared different methods and theirimpacts on the estimates. We demonstrated that estimatesmay vary highly, from -16 to ?12 %, depending on theselected methods. Methodological choices are thus essentialespecially for estimating CO2 fluxes by the stock changeapproach. The sources of error and the accuracy of the estimateswere discussed extensively.